We applied oxidative lipidomics protocols to reveal whether pulmonary responses to SWCNT engage non-specific phospholipid peroxidation processes or selective pathways involved in phospholipid-mediated signaling. In this study, we purposely utilized non-purified commercial SWCNT with high content of transition metals, particularly iron (up to 17.7 weight %), known to effectively catalyze non-specific free radical peroxidation of polyunsaturated phospholipids. Toxicological assessment of such materials is relevant to human exposure to SWCNT in the occupational setting. C57BL/6 mice were exposed to non-purified SWCNT via inhalation (5mg/m³, whole body inhalation for 4 consecutive days, 5 hrs/day) and sacrificed 1, 7 and 28 days thereafter. In line with previous reports,^{22, 26} SWCNT caused a robust inflammatory pulmonary response with a characteristic time-dependent recruitment of neutrophils and macrophages (Figure 1a) and the formation of persistent granulomas (Figure 1b). Notably, significantly elevated levels of apoptotic cells were observed on days 1 and 7 as evidenced by TUNEL staining (Figure 1c). An increased infiltration of myeloperoxidase-positive neutrophils was detected on day 1 and day 7 post-exposure to SWCNT when compared to air exposed control mice. The increase of the number of neutrophils was more pronounced on 1 day of exposure (Figure 1d), which is in accordance with the quantification of the number of neutrophils recovered in BAL (Figure 1a). In addition, an increased infiltration of macrophages - more pronounced on day 7 post exposure (data not shown) - was also observed in SWCNT exposed mice compared to controls.

To characterize possible SWCNT-induced changes in phospholipids, we performed lipidomics/oxidative lipidomics characterization of all molecular species in seven major classes of phospholipids in total lung tissue from exposed animals. The initial 2D-high performance thin layer chromatography (2D-HPTLC) analysis of phospholipid classes did not reveal any significant changes in total phospholipid composition (without detailed MS-analysis of individual molecular species) after SWCNT exposure (Figure 2).

The rate of non-enzymatic free radical lipid peroxidation very strongly depends on the level of lipid unsaturation whereby the process propagates most effectively in phospholipids containing fatty acyls with 6, 5 and 4 double bonds, followed by phospholipids with 3 and 2 double bonds.²⁷ Mono-unsaturated and saturated phospholipids are highly resistant to peroxidation. The polar “head” of phospholipids is much less essential as a factor affecting the rate of stochastic free radical lipid peroxidation. In contrast, enzymatically-catalyzed phospholipid peroxidation is very sensitive to the structure of the polar parts of phospholipid molecules recognized by specific protein-catalysts.²⁸ By using MS analysis, we identified 135 individual molecular species of total phospholipids in the lung of control and SWCNT exposed animals. Of those, 89 individual species were represented by polyunsaturated phospholipids while 46 species of phospholipids were monounsaturated or saturated. This analysis conducted for seven major classes of phospholipids revealed that with the exception of sphingomyeline (Sph) and phosphatidylglycerol (PG) - that contained predominantly saturated and monounsaturated species - all other phospholipids included a variety of highly oxidizable polyunsaturated species (Figure 3). The ranking order of non-enzymatic free radical “oxidizability” of phospholipids defined as the product of the content of polyunsaturated molecular species times phospholipid abundance decreased in the order: phosphatidylcholine (PC) > phosphatidylethanolamine (PE) > phosphatidylinositol (PI) > phosphatidylserine (PS) > cardiolipin (CL). Thus, PC and PE represented the most likely candidates as substrates for random free radical peroxidation in the lung.

Next, we performed oxidative lipidomics analysis of the lung phospholipids in exposed vs control mice. Surprisingly, we did not detect any oxidized molecular species in the two most abundant (constituting more than 75% of all) phospholipids – PC and PE - that should have been highly susceptible to non-enzymatic free radical peroxidation. Instead, three relatively minor anionic phospholipids – CL, PS and PI - underwent peroxidation (Table S1). The magnitude of the effect was particularly pronounced in CL. To directly characterize the peroxidized CL species, the CL “spots” obtained by HPTLC separation were analyzed by 2D-liquid chromatography (LC)/MS whereby a reverse phase chromatography was utilized in the second dimension. We were able to resolve non-oxidized from peroxidized species of CL that differed significantly in their retention times. In the control samples, we detected the abundant non-oxidized CL species with only small amounts of three types of peroxidized CLs corresponding to mono-hydroperoxy- or/and dihydroxy- molecular species of CL. These oxidized CL species were also observed in the lungs from SWCNT exposed mice (day 7) at approximately a 3-fold higher intensity. An additional mono-hydroxy-mono-hydroperoxy-CL species were present in the 7 day SWCNT sample. These oxidized species could be generated from non-oxidized CL species C_18:1/C_18:2/C_18:2/C_16:0, C_18:2/C_18:2/C_18:2/C_20:4, C_18:2/C_18:2/C_18:1/C_20:4, and C_18:2/C_18:1/C_18:1/C_20:4 upon addition of 3, 2, 2, and 2 oxygens, respectively, thus representing combinations of hydroxy- and/or hydroperoxy-modifying groups (Figure 4). Profiles for the non-oxidized CL species for the control and the SWCNT exposed groups were essentially identical (Figure 4). In line with these 2D-LC/MS data, ESI-MS analysis revealed a significant decrease of polyunsaturated molecular species of CL – caused by their involvement in the peroxidation reactions – in the lung of mice exposed to SWCNT (Figure S1 and Figure S2). Oxidation of CL on day 7 post exposure was further confirmed by another type of MS analysis, MALDI-TOF-MS (Figure S3), which allows for the analysis of predominantly singly-charged species and the ability to re-analyze the sample thus alleviating the disadvantage of analyzing the oxidized species during a specified timed-event (ie. the elution of the chromatographic peak) during traditional LC/MS analysis. Significantly decreased levels of “oxidizable” species were documented by MS-analysis of two other classes of pulmonary anionic phospholipids - PS and PI – obtained from SWCNT exposed mice.

To more completely characterize the nature of oxygenated species in anionic phospholipids, we further performed MS/MS analysis of these phospholipids as well as their hydrolysis products formed after treatment with phospholipase A₂ (PLA₂) (Figure S4 and Figure S5). These measurements revealed yet another level of specificity of the peroxidaion process - within each class of anionic phospholipids, the peroxidation pattern displayed a non-random character; only a few molecular species out of many (10 out of 23 for CL, 1 out of 7 for PS, and 3 out of 5 for PI) were converted into oxygenated species (Figure 5).

The specificity was also detected towards fatty acids of the oxidized anionic phospholipids (Figure 6). Surprisingly, among the polyunsaturated CLs and PSs, the species containing fatty acid residues with 2 double bonds – C_18:2 – underwent oxidative modification to mono- (13-OH-C_18:2 and 9-OH-C_18:2) and di-hydroxy (13-,8-diOH-C_18:2 and 9-,14-diOH-C_18:2) as well as mono-hydroperoxy (13-OOH-C_18:2 and 9-OOH-C_18:2) molecular species, whereas more polyunsaturated fatty acid residues such as C_20:4, and C_22:6 – known to be more readily “oxidizable” in non-enzymatic free radical reactions - remained non-oxidized. Small amounts of PI oxidative molecular species containing OH-C_20:4 were detected (Figure 6).

To gain an insight into possible mechanisms involved in SWCNT-triggered phospholipid oxidation in mouse lung we performed in vitro model experiments. Because we documented both increased amounts of apoptotic cells as well as CL and PS oxidation on days 1 and 7 after SWCNT exposure we reasoned that the substrate specificity of the peroxidation process might be related to the execution of the apoptotic program. Our previous work has identified cyt c/H₂O₂ as the catalyst of CL and PS peroxidation in apoptotic cells in vitro and in vivo.^{6, 29-33} Therefore, we performed LC-MS measurements in total lipids extracted from mouse lung oxidized by cyt c/H₂O₂. Intriguingly, lipidomics analysis revealed a pattern of selective oxidation of anionic phospholipids - CL, PS and PI - similar to that detected in the lung of mice exposed to SWCNT (Figure 5). No oxidation of either PE or PC was detected. This is compatible with an interpretation that apoptotic cyt c-driven oxidation reactions are involved, at least in part, in the selective peroxidation of phospholipids after SWCNT inhalation.

SWCNT were purchased from Carbon Nanotechnology (CNI, Houston, TX). The nanotubes were manufactured using the high-pressure CO disproportionation process (HiPco) and were used in the inhalation studies. The supplied SWCNT contained nanometer-scale Fe catalyst particles inherent to the HiPco process. These SWCNT contained elemental carbon (82% wt), Fe (17.7%). Trace elements present included Cu (0.16%), Cr (0.049%), and Ni (0.046%). Raman spectroscopy, near-infrared spectroscopy, and thermo-gravimetric analysis (TGA) were used for purity assessment of HiPco SWCNT. The spectra of SWCNT sample revealed distinct peaks at Raman shift of 1,586–1,591 cm⁻¹, corresponding to the characteristic G band of SWCNT as well as D and G¹ bands typically found in SWCNT spectra. The diameter of SWCNT measured by transmission electron microscopy (TEM) was 0.8–1.2 nm. The length of SWCNT was 100–1,000 nm as measured by atomic force microscopy. The surface area of SWCNT measured by the nitrogen absorption-desorption technique (Brunauer-Emmett-Teller method, BET) was 508 m²/g.

C57BL/6 mice were exposed to SWCNT via inhalation (5mg/m³, whole body inhalation for 4 consecutive days, 5 hrs/day).²⁶ All procedures were pre-approved and performed according to the protocols established by the NIOSH Institutional Animal Care and Use Committee and the Institutional Animal Care and Use Committee of the University of Pittsburgh.

Mice were 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, 10 min, 4 °C) and re-suspension.

The degree of inflammatory response was estimated by macrophages, and 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 256C channelizer, Coulter Electronics, Hialeah, FL). Alveolar macrophages 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 was carried out. Three hundred cells per slide were counted.

Preservation of the lung was performed using protocols previously employed to study pulmonary effects of SWCNT.⁶ Briefly, animals were deeply anesthetized with an overdose of sodium pentobarbital, the trachea was cannulated, 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 histopathologic evaluation were stained with hematoxylin and eosin and examined by a board certified veterinary pathologist for morphological alterations.

Cell apoptosis was determined by a fluorescence labeling apoptosis detection system (Promega, Madison,Wi). This system measures fragmented DNA in apoptotic cells by catalytically incorporating fluorescein 12 dUTP at the 3′OH ends using the TUNEL (TdT mediated dUTP Nick End Labeling) assay. Normal cells were fluorescently labeled red by counter-staining with propidium iodide. Positive control slides were prepared by treating sections with DNase prior to TUNEL. Negative control slides were prepared by omission of the TdT enzyme from the labeling solution. The number of TdT positive nuclei enclosed by a cell cytoplasm (apoptotic cells) per unit area were counted in 10 randomly selection fields (40×) from each animal. Cell counts were expressed as the number of apoptotic cells per unit area of section.

Lung sections were hydrated with tap water and antigen retrieval was performed in citrate buffer solution, pH 6.0. The slides were pretreated with 3% hydrogen peroxide for 20 min to inhibit endogenous peroxidase and blocked using 2.5% normal horse serum (Vector Labs, Burlingame, CA) for 30 min to prevent nonspecific reactivity. Sections were incubated with anti-myeloperoxidase primary antibody (1:100, 4°C overnight incubation) (ab9535, AbCam, Cambridge, UK). Then, sections were treated with a peroxidase conjugated secondary antibody (ImmPRESS, MP-7401, Vector Labs) for 30 min at RT. After washing, color was developed by adding AEC Chromogen for 10 min (SK-4200, Vector Labs). Finally, slides were counterstained with hematoxylin and mounted using Faramount (DAKO, Glostrup, Denmark). Negative control sections were treated in the same way but primary antibodies were omitted. For the detection of macrophages, sections were incubated with anti-rabbit F4/80 antibody (1:50, 4° C overnight) (ab100790, AbCam, Cambridge, UK).

Total lipids were extracted from lung homogenates by Folch procedure.⁵⁷ Lipid extracts were separated and analyzed by 2D HPTLC.⁵⁸ To prevent oxidative modification of phospholipids during separation plates were treated with methanol containing 1 mM EDTA, 100 μM DTPA prior to application and separation of phospholipids by 2D-HPTLC. The phospholipids were visualized by exposure to iodine vapors and identified by comparison with authentic phospholipid standards. For electrospray ionization mass spectrometry (ESI-MS) and analysis of phospholipid hydroperoxides (PL-OOH) by fluorescence high performance liquid chromatography (HPLC) using Amplex Red, the phospholipid spots on the silica plates were visualized by spraying the plates with deionized water. Subsequently, the spots were scraped from the silica plates and phospholipids were extracted in choloroform:methanol:water (10:5:1 v/v). Lipid phosphorus was determined by a micro-method.⁵⁹

Lipids extracted from lung of naïve mice were dried under N₂ and re suspended in 50 mM Na, Na-phopshate buffer pH 7.4, containing 100 μM DTPA. Lipids (500 μM) were sonicated on ice-water for 20 min using high performance ultrasonic system FS3 (Fisher Scientific, PA). Lipids were incubated in with cyt c (10 μM) in the presence of H₂O₂ (100 μM) for 30 min at 37 °C. At the end of incubation lipids were extracted using Folch procedure⁵⁷ and analyzed by LC/ESI-MS as described below.

Lipid hydroperoxides were determined by fluorescence HPLC of resorufin formed in peroxidase-catalyzed reduction of specific PL-OOH with Amplex Red after hydrolysis by porcine pancreatic PLA₂ (1 U/μl) in 25 mM phosphate buffer containing 1.0 mM Ca, 0.5 mM EGTA and 0.5 mM SDS (pH 8.0 at RT for 30 min). For the peroxidase reaction, 50 μM Amplex Red and microperoxidase - 11 (1.0 μg/μL) were added to the hydrolyzed lipids, and the samples were incubated at 4 °C for 40 min. The reaction was started by addition of microperoxidase-11 and terminated by a stop reagent (100 μL of a solution of 10 mM HCl and 4 mM butylated hydroxytoluene in ethanol). The samples were centrifuged at 10,000 g for 5 min and the supernatant was used for HPLC analysis. Aliquots (5 μL) were injected into a C₁₈ reverse phase column (Eclipse XDB-C18, 5 μm, 150 × 4.6 mm) and eluted using a mobile phase composed of 25 mM KH₂PO₄ (pH 7.0)/methanol (60:40 v/v) at a flow rate of 1 mL/min. The resorufin fluorescence was measured at 590 nm after excitation at 560 nm. Shimadzu LC-100AT vp HPLC system equipped with a fluorescence detector (model RF-10Axl) and an autosampler (model SIL-10AD vp) was used.²³

LC/ESI-MS was performed using a Dionex Ultimate™ 3000 HPLC coupled on-line to ESI and a linear ion trap mass spectrometer (LXQ Thermo-Fisher). The lipids were separated on a normal phase column (Luna 3 μm Silica 100A, 150×2 mm, (Phenomenex, Torrance CA)) with flow rate 0.2 mL/min using gradient solvents containing 5 mM CH₃COONH₄ (A – n-hexane: 2-propanol: water, 43:57:1 (v/v/v) and B - n hexane: 2-propanol: water, 43:57:10 (v/v/v). Analysis of phospholipid oxidized molecular species (hydroperoxy- and hydroxy-) was performed as described.^22,23 To minimize isotopic interferences between isolated masses M+2, the spectra were acquired using an isolation width of 1.0 m/z. For identification of phopsholipids ESI-MS analysis was performed by direct infusion into linear ion-trap mass spectrometer LXQ™ with the Xcalibur operating system (Thermo Fisher Scientific, San Jose, CA) as previously described.²³

CL separated by HPTLC was analyzed by LC/MS using a Prominence HPLC system (Shimadzu, Inc.) with a reverse phase C₈ column (Luna, 5 micron, 4.6mm × 15 cm, Phenomenex, Inc.). An isocratic solvent system (2-propanol:water: triethylamine:acetic acid (450:50:2.5:2.5, v/v/v/v) was used at a flow rate of 0.4 mL/min. Spectra were analyzed on Q-TOF Premier mass spectrometer (Waters, Inc.). Parameters were as follows: capillary voltage: 2.85 kV, negative mode; source temperature, 100 °C; desolvation gas, 400 L/hr; sampling cone, 60V; extraction cone, 4.5V; ion Guide, 3.0V. Tuning was optimized for oxidized and non-oxidized CL species across CL scan range.

CL extracts were analyzed in triplicate on a Voyager DE-STR MALDI-TOF-MS (Applied Biosystems/Life Technologies, Carlsbad, CA). High resolution mass measurements were done on an Ultraflex II MALDI-TOF-MS (Bruker Daltonics, Billerica, MA). Sample preparation was slightly modified from the method of Schiller et at.^{60, 61} In brief, the CL extract was spotted onto the MALDI target, followed by 500mM 2,5-dihydroxybenzoic acid (DHB) in methanol containing 0.02% trifluoroacetic acid. Spectra were acquired in reflector-negative mode with external calibration. Masses of all detected molecular species were confirmed to within 0.02 Da.

Oxygenated fatty acids were analyzed by LC/ESI-MS after hydrolysis of major classes of phospholipids with porcine pancreatic PLA₂ (1 U/μL) in 25 mM phosphate buffer containing 1.0 mM Ca, 0.5 mM EGTA and 0.5 mM SDS (pH 8.0 at RT for 30 min). Aliquots of extracted lipids (5 μL) were injected into a C₁₈ reverse phase column (Luna, 3 μm, 150 × 2 mm) and eluted using gradient solvents (A and B) containing 5 mM ammonium acetate at a flow rate of 0.2 mL/min. Solvent A was tetrahydrofuran/methanol/water/CH₃COOH, 25:30:50:0.1 (v/v/v/v). Solvent B was methanol/water 90:10 (v/v). The column was eluted during first 3 min isocratically at 50% B, from 3 to 23 min with a linear gradient from 50% solvent B to 98% solvent B, then 23-40 min isocratically using 98% solvent B, 40-42 min with a linear gradient from 98% solvent B to 50% solvent B, 42-28 min isocratically using 50% solvent B for equilibration of the column. Hydroperoxy-fatty acids: 13S-OOH-9Z,11E-octadecadienoic acid, 9-OH-10E,12Z-octadeacadienoic acid, 15S-OOH-5Z,8Z,11Z13E-eicosatetraenoic acid from Cayman chemicals (Ann Arbor, MI) were used as standards.

The results are presented as mean ± S.D. values from at least three experiments, and statistical analyses were performed by either paired/unpaired Student’s t test or one-way ANOVA. The statistical significance of differences was set at p< 0.05.