Raman spectroscopy was used to characterize the structure of pristine and acid-washed MWCNTs and to determine the degree of MWCNTs functionalization after acid treatment. Figure 1A shows the Raman spectra of pristine and acid-washed MWCNT. There are 4 bands identified in both pristine and acid-washed MWCNTs samples, i.e. D band around 1350 cm^-1 that reflects the level of disorder in the sample, the G band around 1585 cm^-1 indicative of the high degree order and well-structured samples, the G’ band around 2690 cm^-1 representing the binary disordered band and lastly the peak around 2930 cm^-1 indicative of the oxidation level of the sample being characterized. As shown, the D band was wider and had a higher frequency for the acid-washed sample when compared to the pristine MWCNTs. The shift in the D band indicates that the acid treatment minimally altered the chemical structure of MWCNTs by disrupting the structured walls and introducing additional functional groups (carboxylic acid groups) [53]. For the acid-washed MWCNTs there was also a shift of G’ band towards higher frequency; this may be due to the removal of metal catalysts, increase in the number of functional groups having electron accepting ability and decrease in the amorphous carbon. The ratio of intensity of D to G peaks indicate the degree of functionalization [54-56] and was 0.59 for pristine and 0.81 for 1 hr. acid-washed MWCNTs. This also confirms that the acid treatment increased the number of functional groups (i.e. free carboxylic acid groups) on the walls of the MWCNTs samples. Energy dispersive X-ray spectroscopy (EDX) confirmed the increase in the oxygen content due to the acid treatment and thus the increase in the MWCNTs degree of functionalization with free carboxylic acid groups as shown in Additional file 1. Further, the acid washing also reduced the catalyst content in the sample (Fe, 0.81). The content of the iron, cobalt and nickel were further analyzed by inductively coupled plasma-mass spectrometry (ICP-MS). Specifically, the MWCNT by ICP-MS contained 0.03% Fe ±0 .001, 0% cobalt, and 0% Nickel [57].

The length distribution of pristine and 1 h acid-washed MWCNT respectively is shown in Figure 1B (at least 30 individual MWCNTs were measured for each sample). AFM analysis showed that pristine MWCNT samples had an average length of 5499 ± 3009 nm while 1 h acid-washed MWCNTs had an average length of 825 ± 585 nm respectively indicating that acid treatment led to shortening of the nanotubes. The pristine and acid washed MWCNT had a diameter of 15 ± 5 nm. Moreover, acid washing also increased nanotube solubility in DMEM + FBS by two-fold compared to pristine MWCNT [58] as a result of the addition of the free carboxylic acid groups [2].

Two human epithelial cell populations were examined to determine whether MWCNT induced genetic damage. Immortalized respiratory epithelial cells (BEAS-2B) were used to determine the effects of MWCNT on the mitotic spindle. Primary respiratory epithelial cells (SAEC) were included in the analysis to determine whether MWCNT induced errors in chromosome number. Treatment with acid-washed MWCNT induced a dose dependent mitotic spindle disruption (Figure 2A). The disrupted mitotic spindles were predominantly monopolar (Figure 2B). Figure 2C shows a 20× photomicrograph of the cultured cells with three monopolar mitotic spindles in one 40× field. Only 5-10% of the disrupted mitotic spindles were multipolar (Figure 2D).

Primary SAEC cells from a normal donor were used to investigate the effects of MWCNT on the chromosome number. The normal karyotype of the primary cells made it possible to evaluate the treatment related changes in chromosome number. FISH analysis for either chromosome 1 or 4 demonstrated a 2.25 ± 1.0% aneuploidy in the untreated SAEC cells (Table 1). The frequency of the cells with abnormal chromosome number is within the range reported in adult human cells in culture [59,60]. By contrast, the MWCNT-treated SAEC cells had a level of aneuploidy that was comparable to the vanadium pentoxide-treated positive control cells (Figure 2D; Table 1). Abnormal chromosome number was significantly elevated following MWCNT treatment as follows: 62 ± 7.0%, 24 μg/cm²; 59.0 ± 6.0%, 2.4 μg/cm²; 49 ± 6.0%, 0.24 μg/cm² and 42 ± 10%, 0.024 μg/cm² compared with control incidence of 2.25 ± 1.0%. Treatment with 0.31 μg/cm² V₂0₅ resulted in 67 ± 6.0% aneuploid cells. The chromosome alterations in the MWCNT treated cells were predominantly gains of either chromosome 1 or 4 (Table 1). The chromosome losses accounted for 24%, 24 μg/cm²; 13%, 2.4 μg/cm²; 8%, 0.24 μg/cm² and 12%, 0.024 μg/cm². Chromosomal gains accounted for over 70% of the aneuploidy (Table 1). There was also a dose-dependent increase in the number of cells with gains of both chromosomes 1 and 4 indicating an increase in polyploid cells. The number of alterations of chromosome 1 was not statistically different than the alterations of chromosome 4, therefore; there was not a bias for a change of either chromosome.

The MWCNTs were 10–20 nanometers in width. Nanotubes of 10 nanometers or greater can be observed using differential interference contrast imaging. MWCNTs were observed in the cytoplasm and the nucleus (Figure 3A). The MWCNTs also had a strong association with the centrosomes as shown in Figure 3B. The high frequency of monopolar mitotic spindles allowed confirmation of the monopolar phenotype by transmission electron microscopy (TEM) as shown in Figure 3C. The 3D reconstructed image demonstrates strong physical associations between the carbon nanotubes, the microtubules and DNA and the centrosomes (Figure 4A- B). The 3D reconstruction further demonstrated that MWCNTs not only associated with the centrosome but inside the centrosomal structure (Figure 4C).

Exposure to MWCNT did not reduce viability 24 hours after treatment in either the primary SAEC or the immortalized BEAS-2B cells (Figure 5A). Vanadium pentoxide treatment resulted in reduced viability in both SAEC and the BEAS-2B cells. Seventy-two hours following exposure, the viability of the SAEC cells was significantly reduced in cells exposed to 0.024, 0.24, 2.4 or 24 μg/cm² MWCNT (Figure 5B). Three weeks following exposure, the BEAS-2B cells had a small increase in colony formation at 0.024 μg/cm² (Figure 5C). One month following exposure, the SAEC cells had a reduced number of colonies at the highest dose; however, exposure to 0.024, 0.24 and 2.4 μg/cm² resulted in a dramatic increase in colony formation (Figure 5C).

The impact of MWCNT-treatment on the cell cycle was evaluated by Click-iT EdU Flow Cytometry assay. Treatment with 24 ug/cm² MWCNT induced a statistically significant increase in the percent of cells in S phase from 32.11% (PBS-treated) to 40.1% (Table 2). When the cells in G2 phase of the cell cycle were compared, exposure to the positive control, arsenic, resulted in 32.1% of the cells in G2 compared to 18.30% of the cells in the PBS control group thus indicating an arsenic-induced block in G2 (Table 2, p < .05).

Multi-walled carbon nanotubes produced by chemical vapor deposition (Nanolab Inc. PD15L5-20) were acid-washed to remove iron catalyst. The MWCNT were suspended in a mixture of 3:1 v/v sulfuric acid (H₂SO₄) (96.4%, Fisher Scientific, Pittsburgh, PA): nitric acid (HNO₃) (69.5%, Fisher Scientific, Pittsburgh, PA) for 1 hour in a water bath sonicator (Branson 2510, Fisher, Pittsburgh, PA) over ice. The mixture was subsequently diluted in deionized water (2 L) and filtered through a 0.2 μm polycarbonate membrane filter (Millipore, USA); the filtration step was repeated 6 times to remove catalysts or impurities. All cell exposure experiments were performed with one hour acid-washed MWCNT materials.

Atomic force microscopy (AFM) was used to investigate the length of both pristine and acid-washed MWCNT. Commercial Si tips (Asylum Research, AC240TS, USA) were used at their original resonance frequency, varying from 50 to 90 kHz. Pristine or acid-washed nanotubes (10 μg/ml) were deposited on mica surfaces (9.5 mm diameter, 0.15-0.21 thickness, Electron Microscopy Sciences, USA) and dried overnight under vacuum. Scans of 10 μm × 10 μm were acquired using tapping mode in air. At least 30 individual MWCNTs were analyzed to determine their length.

Raman spectroscopy was used to characterize the structure of both pristine and acid-washed MWCNTs. Raman analyses were performed at room temperature using a Renishaw InVia Raman Spectrometer (CL532-100, 100 mW, USA). The excitation source used an argon ion (Ar⁺) laser operating at 514.5 nm. MWCNT (pristine or acid-washed, 1 mg) were mounted on a clean glass slide (Fisher, Pittsburgh, PA) and a 20× microscope objective was used to focus the laser beam to a spot size of < 0.01 mm² and to collect the scattered light. Low energy laser of < 0.5 mV and an exposure time of 10 sec were used to prevent unexpected heating effects of the MWCNT samples being analyzed. Detailed scans ranging from 100 to 3200 cm^-1 were acquired.

The elemental analysis of the pristine and acid-washed carbon nanotubes was examined by energy dispersive X-ray spectroscopy (EDX). Both pristine and acid-washed MWCNT (1 mg/ml in water) were vacuum-dried on silica wafers. The experiments were performed using a Hitachi S-4700 Field Emission Scanning Electron Microscope (USA) and backscattered (BSE) electron detection in a single unit and operating at 20 kV.

ICP-MS) was performed to further analyze the chemical composition of the nanotubes as described previously. Carbon nanotubes were suspended in pure H₂O (18.2 MΩ–cm) at a concentration of 1.0 mg/ml. One ml of each vortexed suspension was added to a 100 ml polytetrafluoroethylene digestion tube (CEM, Matthews, NC) along with 9.0 ml of ultrapure HNO₃ and 1.0 ml of ultrapure H₂O₂ (Fisher Optima, Fisher Scientific, Pittsburgh, PA). Three replicate samples for each nanotube type were digested in the Microwave-Assisted Reaction System (CEM, Matthews, NC) by ramping up to 200°C for 15 min., holding at 200°C for 30 minutes, then cooling to 22°C, adapting a procedure as previously described [100]. There was no visible carbonaceous material remaining in any of the samples after digestion. After suspension (1 mg/ml), the metal content of the nanotubes was analyzed by ICP-MS using the Perkin-Elmer Nexion 300D [101], using ⁵⁴Fe, ⁶⁰Ni, and ⁵⁹Co isotopes. Standards were certified multi-element standards in 1% HNO₃.

The dispersity of pristine MWCNTs and acid-washed MWCNTs in Phosphate buffered Saline (PBS, Fisher, Pittsburgh, PA) was determined by centrifuging the corresponding suspensions (initial concentration 5 mg/mL for both pristine and acid-washed MWCNTs) at 3000 rpm for 5 min. Subsequently, 0.8 mL of the supernatant mixture was filtered through a 0.2 μm filter membrane. After complete drying under vacuum, the amount of pristine MWCNTs or acid-washed MWCNTs on the filter membrane was measured and the dispersity was calculated based on the starting volumes. The obtained values do not reflect the saturation dispersity.

Two human respiratory epithelial cell populations were used to examine the potential genetic damage to MWCNT exposure. Immortalized human bronchial epithelial cells (BEAS-2B, ATCC, Manassas, VA) cultures of passage 4–6 were used to examine the mitotic spindle integrity. The high mitotic rate of the BEAS-2B cells allows examination of sufficient number of mitotic spindles following treatment. BEAS-2B cells grown in serum enriched media double every 18–20 hours and have normal mitotic spindle morphology. The high mitotic index of the BEAS-2B cells made it possible to analyze a sufficient number of mitotic spindles during the 24 hour exposure. Primary small airway respiratory epithelial cells (SAEC; Lonza, Walkersville, MD) from a normal human donor were used to determine the response of a normal cell population. In addition, the normal karyotype of the primary cells was essential for the examination of aneuploidy. The SAEC cells double every 20–24 hours which allowed analysis of a potential change in chromosome number and centrosome morphology of cells that have divided during the 24 hour exposure. The low mitotic index of the SAEC cells (0.5%) prevented the analysis of mitotic spindle integrity in this cell population. The BEAS-2B and SAEC cells were therefore analyzed 24 hours after exposure to allow a sufficient number of cells that have gone through division.

BEAS-2B cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) media supplemented with 10% serum (Invitrogen, Grand Island, NY). The SAEC cultures were cultured following manufacturer’s directions and using Cabrex media (Lonza, Walkersville, MD). The cell cultures were examined by electron microscopy and cytokeratin 8 and 18 staining to verify the epithelial phenotype of the cells as described previously [102].

The immortalized BEAS-2B and the primary SAEC were exposed in parallel culture dishes to MWCNT or to the positive control, vanadium pentoxide (Sigma St. Louis, MO). Three independent experiments were performed for each exposure for SAEC and BEAS-2B respectively. MWCNT and vanadium control were suspended in media and sonicated over ice for 5 minutes and 30 minutes respectively. The cells were seeded in dishes and exposed 0, 0.024, 0.24, 2.4 and 24 μg/cm² MWCNT or to 0.031 μg/cm² vanadium pentoxide when the cells were 70% confluent. The one milliliter culture was treated with 0.024, .24, 2.4 and 24 μg/ml respectively. Twenty-four hours after exposure all cells were prepared for analysis of apoptosis and necrosis, integrity of the mitotic spindle, as well as the centrosome and chromosome number as described below.

Triplicate cultures were prepared in 96 well plates (Becton Dickinson Franklin Lakes, NJ) for the analysis of viability using the Alamar Blue bioassay (Invitrogen, Carlsbad, CA), following manufactures directions as described previously [24]. Eight wells were performed for each treatment and dose. Three independent experiments were performed for the analysis of cellular toxicity by Alamar Blue. Parallel cultures were also prepared in duplicate in one milliliter chamber slides (Nunc Rochester, NY) for the analysis of apoptosis using the TUNEL assay following the manufacturer’s directions (Roche, Inc., Indianapolis, IN) with some modifications outlined previously [24]. A minimum of 100 cells were analyzed for each sample; experiments were repeated three times for a total of 300 cells for each treatment and dose, respectively for the analysis of apoptosis by the TUNEL assay. An additional positive control, 1.68 Molar DNase (Sigma St. Louis, MO) was used for the analysis of apoptosis. Twenty-four hours after dosing, cells in the chamber slides were fixed in 4% paraformaldehyde in phosphate buffer (Sigma St. Louis, MO) and stained with DAPI (Millipore Billerica, MA). The resulting stained samples were fluorescently analyzed using a Zeiss Axiophot fluorescent microscope (Carl Zeiss Microimaging Inc. Thornwood, NY).

BEAS-2B was cultured in 1 milliliter chamber slides as described previously. Dual chambers were prepared for each treatment and each cell type. Three independent experiments were prepared for each cell type and treatment [24]. A minimum of 100 cells of good centrosome and mitotic spindle morphology were analyzed for each sample; experiments were repeated three times for a total of 300 cells for each treatment and dose, respectively. The centrosome integrity as well as the dispersion of carbon nanotubes in the cell cultures was evaluated The spindle integrity of the BEAS-2B cells was examined using dual-label immunofluorescence for tubulin and centrin to detect the mitotic spindle and the centrosomes, respectively. Primary rabbit anti-beta tubulin (Abcam, La Jolla, CA, USA) and mouse anti-centrin antibodies (a generous gift from Dr. Jeff Salisbury), and secondary Rhodamine Red goat anti-rabbit IgG and Alexa 488 goat anti-mouse IgG antibodies (Invitrogen, Carlsbad, CA) were used. The mitotic spindle and centrosome morphology were analyzed in the BEAS-2B cells using a laser scanning confocal microscope (LSM 510, Carl Zeiss MicroImaging Inc., Thornwood, NY) as previously described [103]. Briefly, a monopolar or multipolar mitotic spindle was counted as disrupted. The location of MWCNT was determined by differential interference contrast. Because the nanotubes block the light, the nanotubes produce a black image. To determine the association of the MWCNT with the microtubules of the mitotic spindle and the centrosome, serial optical slices was obtained to create a z-stack and permit three-dimensional reconstruction using LightWave software [104] by TEM following methods outlined previously [103]. Briefly, cells were fixed in 2% glutaraldehyde in sodium phosphate buffer, pH 7.2, for 2 h, postfixed in osmium tetroxide, dehydrated through an ethanol series, and embedded in Spurr’s resin (Sigma, St Louis, MO). Silver-gold sections were stained in 2% aqueous uranyl acetate and Reynolds’ lead citrate, observed using a JEOL 1200 EX electron microscope and recorded digitally.

Due to the necessity of a normal diploid karyotype for the analysis of chromosome number, the SAEC cells were prepared for analysis of the chromosome number. Fluorescence in situ hybridization (FISH) for human chromosomes 1 and 4 was used to determine the chromosome number (Abbott Molecular, Des Plaines, IL) according to the guidelines of the American College of Medical Genetics [105]. Three independent experiments for a total of 300 cells were evaluated for each treatment and dose. A minimum of 100 interphase cells of good FISH morphology were analyzed to determine the number of chromosome 1 and 4. The SAEC cells were photographed using a Zeiss Axiophot microscope and Genetix Cytovision software. Cells with three copies or greater than 4 copies of chromosome 1 or 4 were recorded as a gain for that chromosome. Cells with less than two copies of chromosome 1 or 4 were recorded as a loss of that chromosome. The loss and gain of both chromosomes were added to obtain the errors in chromosome number (aneuploidy).

Triplicate cultures of SAEC cells were grown in T25 flasks. When the cells were 70% confluent they were treated with MWCNT. After 24 hours, the cells were trypsinized, counted and plated at 500 cells/well in 6-well plates for analysis of colony formation. One month following exposure, the cells were washed with PBS, stained with 10% crystal violet solution in neutral buffered formalin (Sigma, Saint Louis, MO) and colonies counted.

BEAS-2B cells were grown in six parallel T25 flasks. A total of 9 independent experiments were performed for the analysis of cell cycle. Twenty-four hours after exposure to 24 μg/cm² MWCNT or to the positive control, 5 μM arsenic (Sigma, St Louis MO), the cells were washed twice with PBS and removed from the dishes with 0.25% trypsin prior to detection of the cell cycle. The cells were stained according to (Invitrogen) manufacturer’s instructions. EdU (5-ethynyl-2′-deosyuridine) is a nucleoside analog of thymidine and is incorporated into DNA during active DNA synthesis. Detection is based on a click reaction- a copper catalyzed covalent reaction between an azide and an alkyne. Twenty-four hours after exposure to MWCNT, the cells were washed twice with PBS and incubated with EdU for 2 hours to detect cells in S-phase. Following incubation, the cells were removed from the plate using 0.25% trypsin. After fixation and Click-iT Saponin permeabilization, CuSO₄ was added to the cells to detect the EdU signal. The total amount of DNA was analyzed following incubation with 7AAD (7-aminoactinomycin D) using a LSR II flow cytometer (BD Biosciences Immunocytometry Systems, San Jose, CA). Data were analyzed and plotted using FlowJo v7.2.5 software.

All analyses were performed using SAS/STAT (Version 9.3) for Windows. Chi-square analysis was used to determine statistical significance for the scoring of the mitotic spindle abnormalities and the number of cells with abnormal chromosome number. The number of viable and apoptotic cells were analyzed using analysis of variance (ANOVA). The mean of duplicate samples were used for the analysis. For cell cycle analysis, a mixed model ANOVA was used to compare the proportion of cells in G1, S and G2/M phase across treatment groups. Experimental block was utilized as a random factor. All differences were considered statistically significant at p < 0.05.