All particles utilized in the current study were characterized by chemical analysis by NMAM #5040 and ICP-AES. Pyrograf CNF was found to be 98.6% wt. elemental carbon with iron levels of 1.4% wt. CNF diameters ranged from 80 to 160 nm. A specific surface area (SSA) of CNF was 35-45 m²/g (measured by BET). Length was determined by SEM and found to be approximately 5-30 μm (Figure 1). SWCNT were 99.7% wt. elemental carbon with 0.23% wt iron. Individual SWCNT had diameters ranging from 1 to 4 nm and were 1-3 μm in length. SWCNT were found to have a specific surface area of 1040 m²/g. As evidenced by TEM, individual SWCNT were bundled in ropes with diameters of ~65 nm (Figure 1C, inset). Crocidolite asbestos fibers lengths were within 2-30 μm range and width of 160-800 nm. Asbestos had a surface area of 8.3 m²/g. Iron levels in crocidolite asbestos were found to be 18% wt. In the suspensions of asbestos and CNF utilized in the current study, a substantial amount of particles/bundles had a conventional fibrous morphology falling under WHO definition: e.g. length > 5 micrometer, diameter < 3 micrometer (Figure 1E). Individual SWCNT bundles in our preparations also fit into the fiber definition; however, they were agglomerated into tertiary structures, which, of course, no longer possess a fibrous morphology (Figure 1C).

Geometrical analysis of the surface areas of SWCNT and CNF agglomerates were conducted as follows. The characterization of SWCNT particles using TEM indicated that a bundled rope of SWCNT (Figure 1C, inset) has an average diameter of ~65 nm. It should be noted that the manufacturer's specifications for SWCNT had a specific surface area (SSA) of 1040 m²/g and diameters ranging from 1 to 4 nm, significantly lower than what we observed. This apparent discrepancy in diameters may be due to the tendency of SWCNT to form bundles because of Van der Waals interactions. Hence, it becomes important to understand, how a bundle with N identical SWCNT arranged in the form of an agglomerated network (Figure 2A) affects the effective surface area. Assuming that each SWCNT has an average diameter of ~3 nm ((2 + 3 + 4)/3 = 3), we estimated that a SWCNT bundle of diameter ~65 nm will contain a total of ~295 SWCNT arranged in 10 layers with an effective accessible surface area equivalent to ~31 SWCNT (see Methods). Therefore, the effective surface area of SWCNT bundles can be estimated as 138.2 m²/g (using eq (4): 1315 × (31/295) m²/g). Thus, the effective surface area of SWCNT of diameter ~3 nm decreases from 1040 m²/g to 138 m²/g when it forms SWCNT bundles with a diameter of ~65 nm. Similarly, from the SEM images (Figure 1A), we determined that CNF particles had an average hollow core and outer diameters (Figure 2B) of ~53 nm and ~109 nm, respectively. Using equations 5 & 6 (see Methods), we estimated that CNF particles are made up of ~82 carbon layers, and have an effective surface area of ~21 m²/g. Based on these calculations, the effective surface area of 40 μg of SWCNT and 120 μg of CNF will correspond to 5.52 × 10^-3 m²/g and 2.52 × 10^-3 m²/g. As a result, the effective surface area of SWCNT and CNF administered is ~5.8 times and ~2.6 times higher, respectively as compared to asbestos (9.6 × 10^-4 m²/mouse).

In order to evaluate lung injury and inflammatory responses to CNF particles, comparing to those observed in SWCNT and asbestos, cell differential and total BAL cell counts, permeability of the lung epithelium (protein levels), and cell damage (LDH release) were determined 1, 7, and 28 days following pharyngeal aspiration of nanoparticles/fibers in C57BL/6 mice. Analysis of the pulmonary inflammatory response following CNF and SWCNT exposure indicated an accumulation of PMNs (150 and 700 fold vs control, respectively, Figure 3A) on day 1 followed by an influx of AMs (Figure 3B) peaking on day 7 (2.0 and 1.6 fold vs control, respectively). In comparison, exposure to asbestos induced a "delayed" inflammatory response with maximal PMN influx (675-fold vs control) occurring on day 7 post exposure. By day 28 post exposure to CNF, PMNs, and AMs in BAL fluid substantially decreased; however, the numbers still remained elevated (25-, and 1.6 fold, respectively) as compared to control.

Exposure to SWCNT, CNF or asbestos caused increased lung permeability, as evidenced by elevated total protein in the BAL fluid. CNF exposure (120 μg/mouse) induced a 1.86-, 1.75-, and 1.14-fold increase in BAL protein on days 1, 7, and 28 post exposure, respectively (Figure 4A). In comparison, SWCNT (40 μg/mouse) exposure resulted in a raise in protein up to 3.75, 2.5 and 2.6 folds of control on days 1, 7 and 28, respectively. Increased lung permeability was also observed following asbestos exposure with a maximal 2.05-fold increase in protein observed 7 days post exposure (Figure 4A).

The degree of pulmonary cytotoxicity elicited by SWCNT, CNF or asbestos was assessed by LDH activity in the BAL fluid recovered from mice. LDH levels were significantly elevated after exposure to CNF (120 μg/mouse; 1.8 fold vs control mice) on days 1 and 7 post exposure (Figure 4B). On day 28 post-CNF, LDH levels remained significantly (1.5 fold) elevated as compared to control mice. Similarly, the release of LDH in response to SWCNT and asbestos followed the same trends. Overall, SWCNT, CNF, and asbestos were all capable of inducing acute pulmonary cell damage with the potency as follows: SWCNT > CNF > asbestos.

Oxidative damage assessed by levels of 4-hydroxynonenol (4-HNE) and oxidatively modified proteins (protein carbonyls) in the lungs of mice exposed to CNF, SWCNT, or asbestos is presented at Figure 5. The time course of 4-HNE accumulation in the lungs following CNF aspiration revealed a significant 6 and 4-fold increase (vs. control) after 1 and 7 days post exposure, respectively (Figure 5A). On day 28 post exposure, the levels of 4-HNE in the lungs of CNF exposed mice returned to control levels. As compared to CNF, SWCNT exposure (40 μg/mouse) induced a more pronounced accumulation of 4-HNE (9.5 and 10.0 fold vs control) on 1 and 7 days post exposure, persisting through day 28 post exposure (6.0 fold vs control). Asbestos exposure did not induce 4-HNE accumulation in the lungs on days 1 and 7 post exposure; however, a marked increase in amount of 4-HNE (11 fold vs control) was observed 28 days post exposure (Figure 5A).

Additionally, levels of oxidatively modified proteins, i.e. protein carbonyls, were evaluated in the lungs following SWCNT, CNF or asbestos exposure. SWCNT induced the most significant and sustained increase of protein carbonyls (4.8, 3.5, 3.5 fold vs control) found 1, 7, and 28 days post exposure, respectively. CNF and asbestos exposure (120 μg/mouse) elicited a steady 1.8-fold increase in protein carbonyls in the lungs of exposed mice (Figure 5B). Overall, SWCNT, CNF, and asbestos caused oxidation of proteins in the lungs with the magnitude of the oxidative damage as follows: SWCNT > CNF > asbestos.

Exposure to the studied NP resulted in accumulation of pro-inflammatory cytokines in the mouse lungs (Figure 6). The cytokine response peaked on day 1 post exposure to SWCNT or CNF (Figure 6). On day 7 post-SWCNT or CNF, levels of TNF-α and IL-10 remained significantly elevated (2.5 and 0.8 fold of control, respectively) as compared to controls (Figure 6A, F). In contrast, in asbestos-exposed mice the release of inflammatory cytokines peaked on day 7 post exposure (Figure 6).

Significant elevation of TGF-β1 was found in BAL fluid of mice after SWCNT, CNF or asbestos exposure throughout the time course of the study (Figure 7A). Maximal TGF-β1 release in BAL fluid of mice was observed on day 7 post exposure. SWCNT induced the higher release of TGF-β1 on days 1, 7, and 28 post exposure (250, 450, and 125%, respectively, vs control), with the highest level found on day 7, while exposure to CNF or asbestos increased TGF-β1 up to maximum of 200% of control on day 7. A pronounced collagen accumulation was observed on day 28 post exposure in the lungs of mice treated with CNF or asbestos (3 or 2.8 -fold vs control, respectively) (Figure 7B). SWCNT exposure induced the most robust collagen buildup reaching 5.8 fold increase comparing to controls. Morphometric analysis of connective tissue stained with Sirius Red is given in Figure 8A. Deposition of collagen was observed in both granulomatous regions as well as in the areas distant from granulomas in the lung of mice exposed to SWCNT. This was established by conventional light microscopy of lung sections specifically stained with Sirius red (Figure 8E). The potency of alveolar interstitial fibrosis was as follows: SWCNT > CNF = asbestos (Figure 8).

Histopathological changes in the lung of mice exposed to CNF, SWCNT or asbestos were evaluated by a board-certified veterinary pathologist. Representative micrographs from each group are shown in Figure 9. A chronic inflammatory reaction and fibrosis were observed 28 days following NP exposure (Figure 9B-D). Fibrous connective tissue was visualized in alveolar septa within the lungs of animals exposed to SWCNT, CNF or asbestos. In contrast to CNF and asbestos, SWCNT exposure revealed granulomatous inflammation, with discrete granulomas often surrounded by hypertrophied epithelioid macrophages associated with dense SWCNT agglomerates. Interfacing bundles of fibrous connective tissue were observed within discrete granulomas elicited by SWCNT agglomerates (Figure 9D).

To evaluate immune outcomes following exposure to respirable SWCNT, CNF or asbestos, we assessed the proliferative response of splenic T cells upon stimulation with concavalin A (T cell mitogen). No significant changes in proliferative response of splenic T cells from animals exposed to CNF were observed on day 7 post exposure, however a decreased proliferation was seen on day 28 post exposure. We found a decrease in proliferation of splenic T-cells obtained from SWCNT-exposed animals (~15% decrease vs control) on day 7 post exposure; however, it was returned to control values after 28 days. In contrast to CNF or SWCNT-treated mice, spleen T cells obtained from asbestos-treated animals showed an increased proliferative response on day 7 post exposure as compared to controls, while on day 28 we observed a decrease in responsiveness of the T cells to concavalin A (Figure 10).

Pearson's correlation coefficients were calculated for pairs of variables including NM dose, expressed as specific surface area (measured by BET, Figure 11A) or effective surface area (calculated as shown in Figure 2A-B) of NM per mouse (Figure 11B), and the relative values of the respective pulmonary outcomes. Several correlations with effective surface area were found to be statistically significant (P < 0.05), including PMN counts and total protein on day 1 post- exposure in BAL fluid (Figure 11B). In contrast, no statistically significant correlations were found between the specific surface area and pulmonary outcomes (Figure 11A).

Specific pathogen-free adult female C57BL/6 mice (8-10 wk) were supplied by Jackson Lab (Bar Harbor, ME) and weighed 20.0 ± 1.9 g when used. Animals were housed one mouse per cage receiving HEPA filtered air in the AAALAC-accredited NIOSH animal facilities. All animals were acclimated in the animal facility under controlled temperature and humidity for one week prior to use. Beta Chips (Northeastern Products Corp., Warrensburg, NY) were used for beddings and changed weekly. Animals were supplied with water and certified chow 7913 (Harlan Teklad, Indianapolis, IN) ad libitum, in accordance with guidelines and policy set forth by the Institute of Laboratory Animals Resources, National Research Council. All experimental procedures were conducted in accordance with a protocol approved by the NIOSH Institutional Animal Care and Use Committee.

A suspension of SWCNT (40 μg/mouse), CNF (120 μg/mouse) or asbestos (120 μg/mouse) was used for single pharyngeal aspiration of C57BL/6 mice, while the corresponding control mice were administered sterile Ca⁺² + Mg⁺²-free phosphate-buffered saline (PBS) vehicle. Mice were sacrificed on days 1, 7 and 28 following exposure. All experiments were repeated at least three times. Inflammation was evaluated by total cell counts, cell differentials, and accumulation of cytokines in the bronchoalveolar lavage (BAL) fluid. Pulmonary toxicity was assessed by elevation of LDH activity in acellular BAL fluid. Fibrogenic responses to exposed materials were assessed by alveolar wall thickness morphometry and collagen deposition. For each group six animals were used to do BAL analysis, histopathology evaluation, oxidative stress markers, and lung collagen measurement.

CNF were obtained from Pyrograf Products, Inc. Carbon nanofibers were vapor grown (PR-24, LHT grade) and heat treated (up to 3000°C) to graphitize chemically vapor deposited carbon present on the surface of the pyrograf and to remove iron catalyst. SWCNT (Unidym Inc, Sunnyvale, CA) were manufactured using the high pressure CO disproportionation process (HiPco™) and purified with acid treatment to remove catalytic metal contaminants [50]. An UICC standard crocidolite asbestos was utilized for comparison of fiber effects. Total elemental carbon and trace metal analysis was performed by the Chemical Exposure and Monitoring Branch (DART/NIOSH, Cincinnati, OH). Elemental carbon was assessed according to the NIOSH Manual of Analytical Methods (NMAM) [51], while trace metal were analyzed by nitric acid dissolution and inductively coupled plasma-atomic emission spectrometry (ICP-AES) following NMAM method 7300 for trace metals. Raman spectroscopy, near-infrared (NIR) spectroscopy, and thermo-gravimetric analysis (TGA) were used for purity assessment of HiPco™ SWCNT. Specific surface area was measured at -196°C by the nitrogen absorption-desorption technique (Brunauer Emmet Teller method, BET) using a SA3100 Surface Area and Pore Size Analyzer (Beckman Coulter Inc, Fullerton, CA), and diameter was measured by TEM. For all size distribution measurements and animal exposures, divalent ion free PBS was utilized as a particle dispersion medium. Prior to animal exposure or size measurements, particles were ultrasonicated (30s × 3 cycles) for improved dispersion of nanoparticles using Vibra Cell (Sonics and Materials Inc., CT, USA) probe sonicator operating at 20 kHz (65% power).

Images of NP suspensions were obtained by field emission scanning electron microscopy. The size distribution of samples, including agglomerate measurements, were performed as previously described by Wang et al. [52]. In brief, the particles deposited on polycarbonate filter were viewed under a field emission scanning electron microscope (model S-4800; Hitachi, Tokyo, Japan) at 400 and 30,000 magnifications. The average length and width of the structures in each sample were determined by analysis of a minimum of 300 particles. Number concentrations of structures were estimated by analyzing 10 fields of view for every particle type/suspension. Agglomerates were counted as a single entity.

Mouse pharyngeal aspiration was used for particulate administration. Briefly, after anesthetization with a mixture of ketamine and xylazine (62.5 and 2.5 mg/kg subcutaneous in the abdominal area), the mouse was placed on a board in a near vertical position and the animal's tongue extended with lined forceps. A suspension (approximately 50 μl) of SWCNT (40 μg/mouse), CNF (120 μg/mouse), or asbestos (120 μg/mouse) prepared in divalent ion free PBS was placed posterior on the throat and the tongue which was held until the suspension was aspirated into the lungs. Control mice were administered sterile Ca⁺² + Mg⁺²-free phosphate-buffered saline (PBS) vehicle. The mice revived unassisted after approximately 30-40 min. All mice in PBS, SWCNT, CNF, and asbestos groups survived this exposure procedure. This technique provided good distribution of particles widely disseminated in a peri-bronchiolar pattern within the alveolar region as was detected by histopathology [53]. Animals treated with the particulates and PBS recovered easily after anesthesia with no behavioral or negative health outcomes. Mice were sacrificed on days 1, 7, and 28 days following the exposure.

The theoretical effective surface area (ESA) was estimated based on the geometrical analysis of carbon nanotubes and nanofibers, using the CNT surface area models developed previously [54]. Assuming that both CNT's and CNF's are generated from the base material graphene, the ESA of SWCNT bundle can be estimated as the product of specific surface area (SSA) of graphene (in m²/g) and the effective surface area of a N layered SWCNT bundle. The SSA of a graphene was estimated as 1315 m²/g. The effective surface area(N_eq-ssa) of N layered (N_L) SWCNT bundle can be obtained using the ratio between the number of individual SWCNT with a total accessible surface area equal to that of a bundle made up of N SWCNT and the total number of SWCNT's (N) in a bundle. The parameters N_L, N_eq-ssa and N of a bundle can be estimated as follows:

(1)NL=0.5×diameter of the bundle/diameter of SWCNT-0.5

(2)Neq-ssa=3×NL+1

(3)N=6.3791×NL(1..6646)

(4)ESAbundle=SSA of graphene×Neq-ssa/N=1315×Neq-ssa/N

In order to estimate the ESA of CNFs, we assumed the concentric graphitic sheets of CNFs are approximately similar to the model of MWNT proposed by Peigney et al. [54]. To calculate thickness of the walls and number of carbon layers of CNF in our study, we determined the average diameter of the hollow core (d_HC) and outer diameters (d_OD) of the CNF based on the SEM images taken (Figure 1A). Assuming that the inter-wall distance between any two carbon layers is 0.34 nm, number of layers (n-layers) in a CNF and ESA of CNF can be estimated using the following equations.

(5)n-layers=((dOD)-(dHC))/(2×0.34)

(6)ESACNF=SSA of graphene×dOD/n-layers ×dOD-2×0.34 ×∑i=1n-1i

Mice were weighed and sacrificed with intraperitoneal injection of sodium pentobarbital (> 100 mg/kg) and exsanguinated. The trachea was cannulated with a blunted 22 gauge needle, and BAL was performed using cold sterile PBS at a volume of 0.9 ml for first lavage (kept separate) and 1.0 ml for subsequent lavages. Approximately 5 ml of BAL fluid per mouse was collected in sterile centrifuge tubes. Pooled BAL cells for each individual mouse were washed in PBS by alternate centrifugation (800 ×g for 10 min at 4°C) and resuspension. Cell-free first fraction BAL aliquots were stored at 4°C for LDH assays while the remainder was frozen at -80°C until processed.

The degree of inflammatory response induced by pharyngeally aspirated SWCNT, CNF or asbestos was estimated by quantitating total cells, alveolar macrophages (AMs), and polymorphonuclear leukocytes (PMNs) recovered by BAL. Cell counts were performed using an electronic cell counter equipped with a cell sizing attachment (Coulter model Multisizer II with a 256 C channelizer, Coulter Electronics, Hialeah, FL). Alveolar macrophages (AM), and PMNs, were identified by their characteristic cell shape in cytospin preparations stained with Diffquick (Fisher Scientific, Pittsburgh, PA), and differential counts of BAL cells were carried out. Three hundred cells per slide were counted.

Measurement of total protein in the BAL fluid was performed by a modified Bradford assay according to the manufacturer's instructions (BioRad, Hercules, CA) with bovine serum albumin as a standard. The activity of LDH was assayed spectrophotometrically by monitoring the reduction of nicotinamide adenine dinucleotide at 340 nm in the presence of lactate using Lactate Dehydrogenase Reagent Set (Pointe Scientific, Inc., Lincoln Park, MI).

Levels of cytokines were assayed in the acellular BAL fluid following SWCNT, CNF or asbestos aspiration. The concentrations of TNF-α, MCP-1, IL-12, IL-6, IL-10 and IFN-γ (sensitivity of assay is 5-7.3 pg/ml) were determined using the BD™ Cytometric Bead Array, Mouse Inflammation kit (BD Biosciences, San Diego, CA). The concentration of active TGF-β1, (sensitivity of assay is < 15.6 pg/ml) was determined using an ELISA kit (Biosource International Inc., Camarillo, CA).

The whole mouse lungs were separated from other tissues and weighed before being homogenized with a tissue tearer (model 985-370, Biospec Products Inc., Racine, WI) in PBS (pH 7.4) for 2 min. The homogenate suspension was frozen at -80°C until processed.

Oxidative damage to the lung following exposure to CNF, SWCNT, or asbestos was evaluated by the presence of 4-hydroxynonenol (4-HNE) and protein carbonyl formation. 4-HNE, a byproduct of lipid peroxidation, was measured in lung homogenates by ELISA using the OxiSelect HNE-His adduct kit (Cell Biolabs, Inc, San Diego, CA). The quantity of oxidatively modified proteins as assessed by measurement of protein carbonyls in lung homogenates was determined using the Biocell PC ELISA kit (Northwest Life Science Specialties). Sensitivity of the assay is < 0.1 nmol/mg protein.

Preservation of the lung was achieved by vascular perfusion with a glutaraldehyde (2%), formaldehyde (1%), and tannic acid (1%) fixative with sucrose as an osmotic agent [55]. This method of fixation was chosen to prevent possible disturbances of the airspace distribution of deposited materials while maintaining physiological inflation levels comparable to that of the end expiratory volume. This was performed using protocols previously employed to study pulmonary effects of SWCNT [10]. Briefly, animals were deeply anesthetized with an overdose of sodium pentobarbital by subcutaneous injection in the abdomen, the trachea was cannulated, and laparotomy was performed. Mice were then sacrificed by exsanguination. The pulmonary artery was cannulated via the ventricle and an outflow cannula inserted into the left atrium. In quick succession, the tracheal cannula was connected to a 5 cm H₂O pressure source, and clearing solution (saline with 100 U/ml heparin, 350 mosM sucrose) was perfused to clear blood from the lungs. The perfusate was then switched to the fixative. Coronal sections were cut from the lungs. The lungs were embedded in paraffin and sectioned at a thickness of 5 μm with an HM 320 rotary microtome (Carl Zeiss, Thornwood, NY). Lung sections for histopathological evaluation were stained with hematoxylin and eosin and examined by a board certified veterinary pathologist for morphological alterations.

Total lung collagen content was determined by quantifying total soluble collagen using the Sircol Collagen Assay kit (Accurate Chemical and Scientific Corporation, Westbury, NY). Briefly, whole lungs were homogenized in 0.7 ml of 0.5 M acetic acid containing pepsin (Accurate Chemical and Scientific Corporation, Westbury, NY) with 1:10 ratio of pepsin: tissue wet weight. Each sample was stirred vigorously for 24 h at 4°C, centrifuged, and 200 μl of supernatant was assayed according to the manufacturer's instructions.

The distributions of type I and III collagen in the lung tissue were determined by morphometric evaluation of the Sirius red-stained sections. Briefly, paraffin lung sections (5-μm thick) were deparaffinized and dehydrated. To identify collagen fibers under the microscope, the sections were stained with F3BA/picric acid for 1-2 h, washed with 0.01 N HCl for 1 min, and counterstained with Mayer's hematoxylin for 2 min. The slides were then dehydrated and mounted with a coverslip [56]. Type I and III collagen stained by Sirius red was visualized, and six randomly selected areas were scored under polarized microscopy using image analysis. With this morphometric method, the average thickness of Sirius red-positive connective tissues in the alveolar wall was quantitatively measured. Volume and surface density were measured using standard morphometric analyses of points and intercept counting [57]. Average thickness of the Sirius red-positive connective tissues of the alveolar wall was computed from two times the ratio of volume density of points to the surface density of the alveolar wall.

Spleens from C57BL/6 mice following exposure to PBS, CNF, SWCNT, or asbestos were obtained on day 7 and 28 post exposure. Spleens were aseptically harvested then ground and the suspension was filtered through a cell strainer. Isolated splenocytes were then centrifuged and red blood cells were lysed utilizing red blood cell lysing buffer (Sigma).

Splenocytes were obtained from exposed (120 μg/mouse CNF, asbestos or 40 μg/mouse SWCNT) or nonexposed C57BL/6 mice (day 7 and 28 post exposure). Cells were labeled with 5-(and-6)-carboxyfluorescein diacetate at 5 μM concentration for 5 min (Invitrogen, Carlsbad, CA), counted using a hemacytometer, diluted in complete medium (1 × 10⁶ cells/mL), and stimulated with 5 μg/mL concavalin A (Sigma, St. Louis, MO) for 4 days in 24-well plates in triplicates. The proliferation response was measured using flow cytometry (BD FACSCalibur instrument, BD, NJ). Dead cells were excluded from the assay with propidium iodide staining preceding the flow cytometry. The background fluorescence readings were subtracted during the analysis. Proliferation index is the average number of cell divisions that the responding T cells underwent. Only responding T cells are reflected in the proliferation index. The proliferation indices were calculated from flow cytometry data using the Flowjo software package (Tree Star Inc., Ashland, OR).

Treatment related differences were evaluated using two-way ANOVA, followed by pair wise comparison using the Student-Newman-Keuls tests, as appropriate. Pearson's correlation coefficients (r) were calculated for pairs of variables including NP effective surface area and relative values of PMN, protein, 4-HNE, IL-6 and alveolar wall thickness. The degrees of freedom (df) for correlation calculations were considered 1. Statistical significance was considered at p < 0.05. Data are presented as Mean ± SE.