Metabolomics of V₂O₅ nanoparticles and V₂O₅ nanofibers in human airway epithelial BEAS-2B cells

Abstract

Vanadium is a toxic metal listed by the IARC as possibly carcinogenic to humans. Manufactured nanosize vanadium pentoxide (V₂O₅) materials are used in a wide range of industrial sectors and recently have been developed as nanomedicine for cancer therapeutics, yet limited information is available to evaluate relevant nanotoxicity. In this study we used high-resolution metabolomics to assess effects of two V₂O₅ nanomaterials, nanoparticles and nanofibers, at exposure levels (0.01, 0.1, and 1 ppm) that did not cause cell death (i.e., non-cytotoxic) in a human airway epithelial cell line, BEAS-2B. As prepared, V₂O₅ nanofiber exhibited a fibrous morphology, with a width approximately 63 ± 12 nm and length in average 420 ± 70 nm; whereas, V₂O₅ nanoparticles showed a typical particle morphology with a size 36 ± 2 nm. Both V₂O₅ nanoparticles and nanofibers had dose-response effects on aminosugar, amino acid, fatty acid, carnitine, niacin and nucleotide metabolism. Differential effects of the particles and fibers included dibasic acid, glycosphingolipid and glycerophospholipid pathway associations with V₂O₅ nanoparticles, and cholesterol and sialic acid metabolism associations with V₂O₅ nanofibers. Examination by transmission electron microscopy provided evidence for mitochondrial stress and increased lysosome fusion by both nanomaterials, and these data were supported by effects on mitochondrial membrane potential and lysosomal activity. The results showed that non-cytotoxic exposures to V₂O₅ nanomaterials impact major metabolic pathways previously associated with human lung diseases and suggest that toxico-metabolomics may be useful to evaluate health risks from V₂O₅ nanomaterials.

Graphical Abstract

Introduction

Vanadium (V) is one of the most prevalent elements in Earth’s crust, rich in soils near mines (Kelley, Scott et al. 2017), and indispensable in iron and steel manufacturing. Vanadium has electrochemical properties that are useful in many catalytic applications (Wachs 2013), and V production for commercial use increased from 30,000 kt in 1995 (Petranikova, Tkaczyk et al. 2020) to 105,000 kt in 2020 (USGS 2020). As a consequence, V emissions from anthropogenic sources increased from 86 kt/year in 1983 (Nriagu and Pacyna 1988) to 240 kt/year in 1995 (Pacyna and Pacyna 2001), with 57 kt/year in Europe and 27 kt/year in North America.

Vanadium pentoxide (V₂O₅) is the leading form of occupational and environmental V exposure. V₂O₅ causes lung irritation and bronchitis and is classified by the International Agency for Research on Cancer (IARC) as possibly carcinogenic to humans. Relatively high V content is common in cigarette smoke (Adachi, Asai et al. 1998, Pappas 2011), coal (Lin, Soong et al. 2018), crude oil and fossil fuels (Lin and Chiu 1995, Sasaki, Maki et al. 1998), and V has been found in at least 319 of the 1,699 National Priorities List sites, identified by the U.S. Environmental Protection Agency (ATSDR 2012). Previous studies in rodents, primates and humans demonstrate that V₂O₅ can cause severe lung damage and dysfunction, and regulatory guidelines are available (Knecht, Moorman et al. 1985, Knecht, Moorman et al. 1992, Irsigler, Visser et al. 1999, Bonner, Rice et al. 2000, Hauser, Eisen et al. 2002, National Toxicology Program 2002, Rondini, Walters et al. 2010).

Nanosized V₂O₅ materials have unique physicochemical properties and are gaining increased use in chromogenic devices (Le, Pham et al. 2022), electrodes (Liu, Zeng et al. 2018), batteries (Yue and Liang 2017), capacitors (Mahmood, Zulfiqar et al. 2022), and sensors (Qu, Shi et al. 2014, Alrammouz, Lazerges et al. 2021). Additionally, there is a growing body of evidence suggesting that exposures to nanoparticles and nanofibers are common (Kumar, Robins et al. 2010, Liang, Ladva et al. 2019, Sonwani, Madaan et al. 2021) and can contribute to adverse health impacts (He, Aker et al. 2015, He and Hwang 2016, He, Fu et al. 2018, He, Deng et al. 2019, Calderón-Garcidueñas and Ayala 2022). Additionally, nanosized V₂O₅ particles have recently been developed as nanomedicine for treating cancer (Roy and Patra 2020, Suma, Padmanabhan et al. 2020), including melanoma (Das, Roy et al. 2020). While U.S. standards regulate particulate matter for ambient concentrations and emissions, they do not regulate particulates explicitly at nanoscale.

Prior research shows that V₂O₅ nanomaterials are cytotoxic in fibroblasts (Ivanković, Musić et al. 2006), rabbit corneal fibroblasts (Fukuto, Kim et al. 2021), human corneal epithelial cells (Kim, Gates et al. 2020), mouse tracheal and human alveolar epithelial cells (Domanico, Fukuto et al. 2022), human breast cancer cell lines [3T3 (Yuvakkumar and Hong 2016); MCF-7 (Gholami-Shabani, Sotoodehnejadnematalahi et al. 2021)], and HeLa cells (Raj and Chatterjee 2016). However, whether V₂O₅ nanomaterials have adverse effects on cell functions at sub-cytotoxic exposure levels is not known. In the present study, we used liquid chromatography-high-resolution mass spectrometry (LC-HRMS), microscopy, and organelle function analyses to investigate V₂O₅ nanomaterial effects in a human airway epithelial cell line, BEAS-2B (Longhin, Capasso et al. 2016, Tran, Shi et al. 2022). Experiments were performed with V₂O₅ nanoparticles and nanofibers at sub-cytotoxic exposure levels, i.e., at doses that did not cause cell death. Because pulmonary injuries such as inflammation and fibrosis are associated with impaired bronchial epithelium function (Crosby and Waters 2010, Cabrera-Benítez, Parotto et al. 2012), measurement of metabolic responses at sub-cytotoxic levels could enhance capabilities to detect nanotoxicity in airways. Results show that V₂O₅ nanoparticles and V₂O₅ nanofibers have metabolic effects on BEAS-2B cells at sub-ppm levels. Metabolic responses differed for V₂O₅ nanoparticles and V₂O₅ nanofibers, suggesting that physicochemical properties of specific V₂O₅ nanomaterials may also contribute to toxicologic responses. An overlap of metabolic responses with those previously observed for lung diseases suggests that toxico-metabolomics may be useful to understand mechanisms in pulmonary inflammation, fibrosis and respiratory disease.

Methods and Materials

Chemicals

VOSO₄·×H₂O and KBrO₃ were purchased from Sigma-Aldrich (St. Louis, MO). Nitric acid was purchased from MilliporeSigma (Burlington, MA). V₂O₅ nanoparticles were purchased from US Research Nanomaterials, Inc. (Houston, TX). An internal standard mixture for mass spectral analysis consisted of [¹³C₆]-D-glucose, [¹⁵N]-indole, [2-¹⁵N]-L-lysine dihydrochloride, [¹³C₅]-L-glutamic acid, [¹³C₇]-benzoic acid, [3,4-¹³C₂]-cholesterol, [¹⁵N]-L-tyrosine, [trimethyl-¹³C₃]-caffeine, [¹⁵N₂]-uracil, [3,3-¹³C₂]-cystine, [1,2-¹³C₂]-palmitic acid, [¹⁵N , ¹³C₅]-L-methionine, [¹⁵N]-choline chloride, and 2’-deoxyguanosine-¹⁵N₂,¹³C₁₀-5’-monophosphate, all from Cambridge Isotope Laboratories, Inc (Andover, Pennsylvania). Ultra-pure water (Milli-Q; Sigma-Aldrich, St. Louis, MO) was used as indicated. All reagents were analytical grade or better unless otherwise stated.

V₂O₅ nanofiber synthesis

A hydrothermal synthesis was developed to fabricate V₂O₅ nanofibers. In brief, 10 mmol of VOSO₄·xH₂O and 5 mmol of KBrO₃ were carefully mixed in 40 mL Milli-Q water for 30 min at room temperature inside a fume hood. Nitric acid was added dropwise until pH reached 2. The solution was then transferred into a 50 mL Teflon-lined hydrothermal stainless-steel autoclave (Techinstro, Maharashtra, India) and maintained at 180 °C for 24 h. After cooling to room temperature, the reaction products were collected and washed with Milli-Q water at least three times, followed by anhydrous ethanol at least three times, and dried in a Refrigerated CentriVap Vacuum Concentrator (Labconco, Kansas City, MO) at 80 °C for 12 h, and kept at 4 °C for 6 h, then stored at room temperature.

Characterization

The particle morphology and particle size distribution of nanoparticles were determined by transmission electron microscopy (TEM, JEM-1400, JEOL Company, Japan) and scanning electron microscopy (SEM, Topcon DS-150F Field Emission SEM). UV-vis spectra were measured using a NanoDrop 2000c Spectrophotometer (Thermo Fisher Scientific, Waltham, MA). The phase structures of particles were characterized by X-ray diffraction (XRD, Synergy-S and the D8 Venture Diffractometers, Billerica, MA) with Cu Kα radiation (λ = 1.541874 Å) over the 2θ range of 5–60°.

Cell culture

BEAS-2B (CRL-9609) are an SV40 large T antigen immortalized human bronchial epithelial cell line derived from noncancerous individuals, purchased from American Type Culture Collection (ATCC). Cells were cultured at 37 °C and 5% CO₂ in serum-free Bronchial Epithelium Basal Medium (BEBM) containing BEGM^® SingleQuots^® (Catalog #: CC-4175, Lonza, Walkersville, MD). At 80–90% confluency, cell medium was replaced with media containing V₂O₅ nanoparticles or V₂O₅ nanofibers at concentrations as indicated (0, 0.01, 0.1, or 1 ppm). It is noted that V₂O₅ nanoparticles and V₂O₅ nanofibers had little dissolution of V at the time of addition to cells or in culture medium over a 24-h period. V₂O₅ cytotoxicity was measured following incubation at 37 °C and 5% CO₂ for 24 h based upon previous studies (Gallardo-Vera, Diaz et al. 2016, Fallahi, Foddis et al. 2018, He, Jarrell et al. 2022). Use of this time point assured that metabolomics and other studies were performed on viable cells but do not exclude possible longer-term effects on cell viability.

Cytotoxicity test

Cytotoxicity of V₂O₅ nanoparticles and nanofibers was evaluated using the WST-1 assay following the manufacturer’s instruction. In brief, BEAS-2B cells were seeded in 96-well plate with 8 replicates for each treatment and 16 replicates for controls and grown to at least 80% confluency. The cells were treated with a wide range of V₂O₅ nanoparticles and nanofibers doses (0–200 ppm) for 24 h. After incubation, cell media were replaced with 100 μL fresh medium containing 10 μL WST-1 reagents (Roche, Catalog # 11644807001, Rotkreuz, Switzerland), cells were further incubated at 37 °C and 5% CO₂ for 4 h, and absorbance was measured at 450 nm relative to a reference wavelength at 610 nm using an ELISA plate reader (SpectraMax M2, Molecular Devices, San Jose, CA). For normalization of signal relative to cell content, cells were rinsed with PBS, stained with 100 μL 1:2500 Hoechst 33342 in PBS at 5% CO₂ and 37 °C for 5 min, rinsed with PBS, and quantified at excitation and emission wavelengths of 350 and 461 nm. Final cytotoxicity was calculated by normalizing the readings from WST-1 assay with the measurements from Hoechst 33342 staining.

Quantification of intracellular V levels

⁵¹V was measured in BEAS-2B cells with inductively-coupled plasma mass spectrometry (ICP-MS, iCap Q, ThermoFisher Scientific) following procedures to measure trace metals that conformed to accuracy (100 ± 10%) and precision standards (RSD < 12%). Briefly, cells were cultured in 10-cm plates and 1 mL cell media was collected at indicated times. Cells were then washed three times with ice-cold PBS solution and lysed by adding 410 μL pure water and applying three freeze-thaw cycles. 5 μL of lysed cells were used for bicinchoninic acid assay (BCA) of protein content. 400 μL of lysed cells or 200 μL of cell media were digested in 40% nitric acid and 9% hydrogen peroxide using a programmable 1200 W microwave (MARS 5, CEM Corp., Matthews, NC) with a rotor for 40 Teflon-lined vessels rated at 210 °C and 350 psi (HP-500 Plus, CEM Corp., Matthews, NC). V standard (1000 mg/L, 2% HNO₃) was purchased from Ricca Chemical (Arlington, TX) and diluted in series for the standard curve.

High-resolution metabolomics

Cells from 6-well plates were used to extract metabolites in 300 μL acetonitrile: water (2:1) containing internal standards following the procedures as described previously (Go, Kim et al. 2015, Jarrell, Smith et al. 2021). Each sample was analyzed with using a High-Field Q Exactive mass spectrometer (120,000 resolution, 85–1275 m/z, Thermo Fisher), coupled with hydrophilic interaction liquid chromatography with positive electrospray ionization [HILIC (+)] and C18 reversed phase chromatography with negative electrospray ionization [C18 (−)]. Each sample was run with three technical replicates in batches of 20 samples. Mass spectral features were extracted using apLCMS (Yu, Park et al. 2009) and xMSanalyzer (Uppal, Soltow et al. 2013) and expressed as accurate mass to charge (m/z), retention time (rt), and associated intensity. Data were filtered to retain features with nonzero values in >70% in all samples and >80% in at least one group. Data were averaged among replicates, log2 transformed and quantile normalized. NIST SRM1950 and Q-Standard 3 (Qstd3) were used as reference materials as described previously (Go, Walker et al. 2015), for quality control and quality assurance.

Metabolomics data analysis

A 2-step statistical procedure was used to protect against both Type 1 and Type 2 statistical error in identification of pathways and metabolites that varied due to nanoparticle and nanofiber exposures. In this process, features with raw p <0.05 were selected in the first step to protect against Type 2 statistical error. These features were used with permutation testing in pathway enrichment analysis (see below) in a subsequent step to protect against Type 1 statistical error. Statistical significance of metabolic features between groups was computed using one-way ANOVA and linear regression-based feature selection using xmsPANDA (https://github.com/kuppal2/xmsPANDA, last accessed July 15, 2022). The selected features differentiating treatment groups (p < 0 .05) were examined by hierarchical cluster analysis (HCA), partial least squares-discriminant analysis (PLS-DA), and correlation analysis using MetaboAnalyst (Chong, Soufan et al. 2018).

Metabolic pathway enrichment analysis

Selected features (p < 0.05) were used for metabolic pathway enrichment analysis with mummichog as described earlier (Jarrell, Smith et al. 2021, Jarrell, Smith et al. 2022). Permutation testing was performed using 1000 permutations to protect against type 1 statistical error (Uppal, Walker et al. 2016). Analysis was performed using 5 ppm maximum tolerance in m/z. Significant pathways were selected at p < 0.05.

Metabolite annotation and identification

Over 400 metabolites with Level 1 identification by accurate mass m/z, rt and MS/MS matching standards (Schymanski, Jeon et al. 2014), including amino acids and metabolites from most major metabolic pathways, have been established in the laboratory using the LC-HRMS methods as used in the current study (Liu, Nellis et al. 2020). To obtain information on other potential metabolites, annotation was carried out using xMSannotator (Uppal, Walker et al. 2016). xMSannotator employees a multilevel scoring algorithm to annotate metabolic features with confidence score from 0 (only accurate mass m/z match) to 3 (high confidence annotation), based on adduct/isotope patterns and elemental or abundance ratio checks. Annotations were made against Kyoto Encyclopedia of Genes and Genomes (https://www.kegg.jp, last accessed July 15, 2022) and the Human Metabolome Database (http://www.hmdb.ca/, last accessed July 15, 2022).

Transmission Electron Microscopy (TEM)

TEM was used to investigate potential effect of V₂O₅ exposures on subcellular ultrastructure. Cells were washed with PBS and fixed in Karnovsky’s fixative at least overnight and maintained at 4°C until processing. Cells were then post-fixed with 1% osmium tetroxide in 0.1 M phosphate buffer (pH 7.4) for 1 hr. Following graded ethanol dehydration, a monolayer of cells in each well of a 6-well plate was infiltrated, embedded, and polymerized in Eponate 12 resin (Ted Pella Inc., Redding, CA). Ultrathin sections were cut using a Leica Ultracut S ultramicrotome at a thickness of 80 nm. Sections were stained with 5% uranyl acetate and 2% lead citrate and imaged at 80 kV on a JEOL JEM-1400 TEM (JEOL Ltd., Japan) equipped with a Gatan US1000 2k×2k CCD camera (Gatan, Pleasanton, CA).

Lysosomal Activity Measurement

Lysosomal activity in BEAS-2B cells was determined with the Lysosomal Intracellular Activity Assay Kit (Abcam, ab234622, Waltham, MA) following the manufacturer’s instructions. The fluorescence signal of the Lysosome-Specific Self-Quenched Substrate is proportional to the intracellular lysosomal activity and was imaged with fluorescent microscopy (BioTek Lionheart FX, Winooski, VT) at excitation wavelength of 488 nm and emission at 510 nm. To stain the nuclei, BEAS-2B cells were washed with PBS and then stained with Hoechst 33342 1:2500 (Invitrogen, Waltham, MA) at 5% CO₂ and 37 °C for 5 min, with quantification using ImageJ software. BEAS-2B cells incubated with bafilomycin A1 (1x) served as positive controls.

Mitochondrial Membrane Potential Assay

Mitochondrial membrane potential was assessed using JC-1 (Invitrogen, Waltham, MA). Briefly, BEAS-2B cells in 96 well plates were treated with V₂O₅ nanomaterials for 24 h, washed twice with PBS and incubated with JC-1 (10 μg/mL) for 20 min at 37 ℃. The relative ratio of red/green fluorescence (Red, JC-1 aggregate, EX/EM: 535/590; Green, JC-1 monomer, EX/EM: 485/530) was measured using a fluorescence plate reader (SpectraMax M2, Molecular Devices, San Jose, CA). JC-1 aggregate and monomers were imaged with fluorescent microscopy (BioTek Lionheart FX, Winooski, VT). BEAS-2B cells treated with H₂O₂ (1 mM) for 10 min were used as positive controls.

Statistics

Quantification data were analyzed using Matlab R2021a (MathWorks, Inc., Natick, MA) and graphs were made using OriginPro 2021b (OriginLab Corp., Northampton, MA). Results are presented as mean ± standard error. One-way ANOVA with post hoc testing using Least Significant Difference (LSD) was performed to obtain statistical significance. Six replicates were used for metabolomics, three replicates were used for ICP-MS and lysosomal activity assay, 8 replicates were used for mitochondrial membrane potential test, and 8–12 replicates were used for cytotoxicity tests. Values with p < 0.05 were considered to be significant.

Results

Particle and Fiber Characterization

TEM of the prepared V₂O₅ nanofibers revealed a fibrous morphology (Fig. 1A), with a width 63 ± 12 nm and length 420 ± 70 nm. In contrast, V₂O₅ nanoparticles showed a typical round particle morphology with a size 36 ± 2 nm, and some larger aggregates 166 ± 34 nm (Fig. 1A). Examination by SEM (Fig. 1B) showed that the V₂O₅ nanofibers had a fibrous structure corresponding to the TEM images and that the V₂O₅ nanoparticles exhibited amorphous masses, which may have been a result of aggregation. The UV–Vis optical properties of the nanofibers and nanoparticles showed that both types of nanomaterials had some light scatter (Fig. 1C). V₂O₅ nanofibers had better defined maxima at 253 and 392 nm, i.e., with a red shift relative to the nanoparticles relative to the V₂O₅ nanoparticles, which had a broad absorption peak in the range of 229 nm.

Fig. 1.

Physicochemical characterization of V₂O₅ nanofibers and nanoparticles. (A) TEM micrographs. (B) SEM micrographs. (C) UV–Vis absorption spectrum. (D) XRD pattern with (001) phase identified. (E) TEM micrographs for lattice visualization with fast Fourier transform. nf V₂O₅: V₂O₅ nanofiber; np V₂O₅: V₂O₅ nanoparticle.

The XRD patterns of the V₂O₅ nanofibers included diffraction peaks of (001), (004), (005) and (006) at Bragg’s angle 8.28, 25.56, 29.56 and 39.53 corresponding the layered hydrated V₂O₅ 00l reflections (JCPDS No. 40–1296), with a characteristic lattice spacing of 10.67 nm at the (001) reflection peak (Fig. 1D). The XRD patterns of the nanoparticles had intense reflections at Bragg’s angle (2θ) 15.47, 20.34, 21.8, 25.63, 31.12, 34.42, 41.63, 45.77, 47.52, 51.42, 55.88, and 59.35, corresponding to (200), (001), (101), (110), (400), (310), (002), (411), (600), (020), (121) and (611) diffraction planes, respectively of an orthorhombic-phase V₂O₅ (JCPDS No. 41–1426) (Figure 1D). The calculated interplanar d-spacing of 0.44 nm corresponds to the (001) lattice planes of an orthorhombic V₂O₅. Rietveld refinement of the XRD data also revealed an orthorhombic crystal structure (space group Pmn21) with simulated lattice parameters, a = 11.48 Å, b = 4.36 Å, and c = 3.55 Å. Independent examination of the lattice spacing of V₂O₅ nanofibers and the nanoparticles using high-resolution TEM showed similar lattice spacing as identified by XRD (Fig. 1E). Collectively, these data document that the V₂O₅ nanomaterials have nanometer dimensions and differ in physical properties.

Dose-dependent Cytotoxicity Following V₂O₅ Nanoparticle and Nanofiber Exposure

To determine the cytotoxicity of V₂O₅ nanoparticle and nanofiber in BEAS-2B cells, we measured cell survival at 24 h after treatment with concentrations ranging from zero to 200 ppm. The results showed that 50% (LC₅₀) was 24 ppm for V₂O₅ nanoparticles and 36 ppm V₂O₅ nanofibers (Fig. 2A and 2D). No apparent cell death was observed at 24 h with concentrations at 1 ppm and below (Fig. 2B and 2E). To avoid difficulties in interpretation due to a variable extent of cell death, metabolomics studies were performed with 1 ppm or lower exposures. ICP-MS was used to verify actual exposure levels at 0.01, 0.1 and 1 ppm for V₂O₅ nanoparticles (Fig. S1A) and V₂O₅ nanofibers (Fig. S1B), and to measure intracellular level of V in BEAS-2B cells (Fig. 2C and 2F).

Fig. 2.

Cytotoxicity and intracellular vanadium levels in BEAS-2B cells treated with V₂O₅ nanomaterials. Panels A, B and C provide results for V₂O₅ nanoparticles and Panels D, E and F provide results for V₂O₅ nanofibers. Cytotoxicity was quantified using WST1 assay after 24 h incubation and expressed as cell survival, presented as a percentage of respective control values (A, B, D, E). Cellular V⁵¹ was measured using ICP-MS following 24 h treatment. Abbreviations: nfV₂O₅: V₂O₅ nanofiber; npV₂O₅: V₂O₅ nanoparticle.

Metabolic Responses to V₂O₅ Nanoparticle Exposure

Metabolomics of V₂O₅ nanoparticles at 0, 0.01, 0.1 and 1 ppm yielded a total of 10288 and 5870 metabolic features from HILIC(+) and C18(−) chromatography after data filtering. Linear regression analysis (p < 0.05) showed 678 associated features for HILIC(+) (Fig. 3A) and 500 for C18(−) (Fig S2A). One-way HCA of the selected HILIC(+) features (Fig. 3B) for nanoparticles showed separation into metabolite Clusters 1 and 2, with the larger Cluster 2 showing that more metabolite signals decreased with increased nanoparticle exposure. PLS-DA of the selected features showed a trajectory from control to highest exposure, with little separation at 0.01 ppm but clear separation at 1 ppm (Fig. 3C). Together with the HCA, the PLS-DA data show that the dose-response effects are largely driven by the 1 ppm dose and minimal effects occur at 0.01 and 0.1 ppm. Similarly, one-way HCA of the selected C18(−) features associated with V₂O₅ nanoparticle exposures (Fig. S2B) showed that more metabolites were decreased than increased, but the difference was not as great. PLS-DA of selected features also showed a trajectory from control to highest dose, with separation only at 1 ppm (Fig. S2C). Corresponding data for 500 selected features for C18(−) are provided in Supplemental materials (Fig S2A to S2C).

Fig. 3.

Metabolic effects of V₂O₅ nanoparticles in BEAS-2B cells. (A) Manhattan plot of 678 metabolic features selected by linear regression with V₂O₅ nanoparticle dose (p < 0.05) from HILIC(+) chromatography. (B). One-way hierarchical cluster analysis (HCA) of 678 selected metabolic features altered by V₂O₅ nanoparticles. (C) PLS-DA of 678 selected metabolites altered by V₂O₅ nanoparticles exposure at p < 0.05. np V₂O₅: V₂O₅ nanoparticle.

Metabolic Pathway Effects of V₂O₅ Nanoparticle Exposure

Pathway enrichment analysis for V₂O₅ nanoparticle exposures was performed independently with mummichog for the 678 [HILIC(+)] and the 500 [C18(−)] selected metabolic features; pathway results are discussed together for simplicity. Mummichog pathway enrichment analysis showed that 17 metabolic pathways varied (p < 0.05) with increasing V₂O₅ nanoparticle exposures. Lipid pathways showed redundancies due to mapping of some metabolites to multiple pathways (Fig. 4A). Omega-3 fatty acid (p = 0.0018) and fatty acid (p = 0.01) metabolism were two of the top three pathways associated with increased V₂O₅ nanoparticle exposure. Other fatty acid pathway associations included de novo fatty acid biosynthesis (p = 0.01), fatty acid activation (p = 0.018), fatty acid oxidation (p = 0.04) and the carnitine shuttle (p = 0.03). Additionally, glycerophospholipid (p = 0.048) and glycosphingolipid (p = 0.028) were also changed by V₂O₅ nanoparticle exposure, indicating metabolic changes related to cell membrane functions (Wakil and Abu-Elheiga 2009). Other related pathways, aminosugar metabolism (p = 0.0045) and pyrimidine metabolism (p = 0.04), illustrate an important relationship among pathways because pyrimidine pathway overlaps with lipid metabolism (CDP-choline functions in phospholipid biosynthesis), and with aminosugar metabolism (UDP-sugars function in oligosaccharide biosynthesis; (Handford, Rodriguez-Furlán et al. 2006).

Fig. 4.

Metabolic pathways altered by V₂O₅ nanoparticles exposure. (A) Pathway analysis revealed 17 pathways associated with V₂O₅ exposure. Red bars indicate upregulated pathways, blue bars indicate downregulated pathways, grey bars indicate non-specific direction. (B) Network of the major activity modules is shown with metabolites associated with those significantly altered pathways. (C) Top 25 metabolites from PLS-DA altered by V₂O₅ nanoparticles exposure using significant HILIC features with p < 0.05. Annotated metabolites are indicated by pound sign (#). Selected annotated metabolites are individually plotted in (D-H): (D). 3-Dehydroxycarnitine (m/z 146.1175, rt 34s); (E). Palmitoylglycine (m/z 314.26875, rt 23s); (F) 2-Amino-4-hydroxy-3-methylpentanoic acid (m/z 148.0969, rt 272s); (G) LysoPE(18:2, 0:0) (m/z 478.2826, rt 198s); and (H) N-Gluconyl ethanolamine phosphate (m/z 320.0756, rt 249s). Statistical significance for each condition vs. control is indicated by asterisks, with *p < 0.05, and **p < 0.01. np V₂O₅: V₂O₅ nanoparticle.

A number of other pathways also varied with V₂O₅ nanoparticle exposures (Fig 4a). Purine metabolism (p = 0.049) plays important roles in metabolic signaling and cell growth along with pyrimidines. Several amino acid pathways (methionine, cysteine, aspartate, asparagine, glycine, serine, alanine and threonine), and vitamin B3 (niacin and niacinamide) metabolism (p = 0.025) also varied with V₂O₅ nanoparticles. Niacin is a critical precursor for NAD, and other pyridine nucleotides functioning in metabolism of fatty acids, carbohydrates and amino acids (Sauve 2008). Positive t-statistic scores for 248 [HILIC(+)] and the 206 [C18(−)] of the selected metabolites (p < 0.05) indicated upregulation in all fatty acids pathways and in aminosugar, niacin and niacinamide, carnitine shuttle, and amino acid pathways (red bars in Fig. 4A, Fig. S4A). Negative t-statistic scores indicated downregulation in glycosphingolipid metabolism (blue color bars in Fig. 4A, and Fig. S4B).

An associated network map of 92 top metabolic features from HILIC(+) and 62 from C18(−) from mummichog had empirical IDs that mapped to subclusters centered around AMP (Fig. 4B). The central AMP hub was connected to fatty acyl-CoAs, free fatty acids and an N-acetylglutamate subcluster (upper right) containing N-acetylglutamate, aspartate and aspartyl-N-acetylglutamate. N-Acetylglutamate is an activator of carbamoylphosphate synthetase, an essential step in ammonia elimination from amino acid degradation and a possible link to the diverse amino acid pathway associations (Fig. 4A). A niacin subcluster on the upper left (Fig 4B) was linked to AMP through nicotinamide mononucleotide, a precursor for NAD biosynthesis and also a breakdown product from poly-ADP-ribose polymerase, an important nuclear regulatory system. This subcluster contained other niacin metabolites and S-adenosylhomocysteine, an intermediate in the transulfuration pathway leading to synthesis of the antioxidant glutathione (GSH), and also resulting from common methylation reactions controlling the epigenome and diverse metabolic products. A hydroxyproline subcluster on the lower right (Fig. 4B) connects AMP to proline, hydroxyproline and pyrroline hydroxycarboxylic acid, reactions that are likely linked to increased collagen biosynthesis and remodeling. Top metabolites correlated with V₂O₅ nanoparticle exposures from the PLS-DA (Fig. 4C) illustrate the diversity of metabolites in network associations with nanoparticle exposures: 3-dehydroxycarnitine (Fig. 4D), palmitoylglycine (Fig. 4E), 2-amino-4-hydroxy-3-methylpentanoic acid (Fig. 4F), LysoPE(18:2) (Fig. 4G), and N-gluconyl ethanolamine phosphate (Fig. 4H).

Metabolic Responses to V₂O₅ Nanofiber Exposure

A total of 10223 HILIC(+) features and 5843 C18(−) features were identified for cells exposed to V₂O₅ nanofibers. Similar numbers of metabolic features varied with V₂O₅ nanofiber exposure as found to vary with V₂O₅ nanoparticle exposure, i.e., 631 for HILIC (Fig. 5A) and 742 for C18 (Fig. S3A). Results from 1-way HCA showed that a similar number of features increased and decreased with V₂O₅ nanofiber exposure in BEAS-2B cells. Similar to the results with nanoparticles, changes were most apparent only at 1 ppm exposures (Fig. 5B). PLS-DA showed a trajectory from low to high V₂O₅ nanofiber exposure; however, only the highest dose showed clear separation from control and lower doses (Fig. 5C).

Fig. 5.

Alterations in metabolites by V₂O₅ nanofiber exposure. (A). Manhattan plot of associated metabolic features (p < 0.05) from HILIC (+) chromatography. (B). HCA of metabolites from HILIC (+) chromatography altered by V₂O₅ nanofiber at p < 0.05. (C) PLS-DA of selected metabolites from HILIC (+) chromatography altered by V₂O₅ nanofiber exposure at p < 0.05. Warm color and positive z-score indicates higher abundance in (A). nf V₂O₅: V₂O₅ nanofiber.

Metabolic Pathway Effects of V₂O₅ Nanofiber Exposure

Combined results for pathway enrichment analysis for HILIC(+) and C18(−) data showed 15 metabolic pathways were associated (p < 0.05) with nanofiber exposure (Fig. 6A), and these pathways overlapped considerably with pathways that associated with V₂O₅ nanoparticle exposure (Fig. 4A). Among these, four fatty acid pathways and carnitine shuttle were associated with nanofibers. A notable difference from the results for nanoparticles was that glycerophospholipid and glycosphingolipid pathways were not detected as changed in association with V₂O₅ nanofibers. Other pathways that varied with V₂O₅ nanoparticles and also varied with V₂O₅ nanofibers included purine (p = 0.032), vitamin B3 (niacin and niacinamide) (p = 0.043), and several amino acid pathways. Pathways that were associated with V₂O₅ nanofibers but not with V₂O₅ nanoparticles included sialic acid (p = 0.047), carbon fixation with pentose phosphate metabolism (p = 0.03), and squalene and cholesterol metabolism. Consistent with the nanoparticle results, positive t-statistic scores indicated upregulation in all fatty acid pathways, as well as in aminosugars, sialic acid, carbon fixation, carnitine shuttle, glycine, serine, alanine and threonine for nanofibers (red bars in Fig. 6A, and Fig. S5A). However, nanofiber responses showed many more pathways with negative t-statistic scores (blue bars in Fig. 6A, and Fig. S5B). These included squalene and cholesterol, purine, vitamin B3 (niacin and niacinamide), methionine and cysteine, valine, leucine and isoleucine pathways. Thus, even though many of the same pathways were changed, the data show that the responses differ for the nanomaterials.

Fig. 6.

Metabolic pathways altered by V₂O₅ nanofiber exposure. (A) Pathway analysis revealed 15 pathways were enriched with V₂O₅ exposure. Red bars indicate upregulated pathways, blue bars indicate downregulated pathways, grey bars indicate non-specific direction. (B) Network of the major activity modules is shown with metabolites associated with those significantly altered pathways. (C) Top 25 metabolites from PLS-DA altered by V₂O₅ nanofiber exposure using significant Hilic(+) features with p < 0.05. Annotated metabolites are indicated by pound sign (#). Selected annotated metabolites are individually plotted in (D-H): (D). S-Adenosylhomocysteine (m/z 385.1286, rt 76s); (E). Butyrylcholine (m/z 175.157, rt 76s); (F) CTP (m/z 481.9768, rt 87s); (G) 3-AMP (m/z 348.07019, rt 281s); and (H) uracil (m/z 157.0255, rt 281s). Significance for each condition vs. control is indicated by asterisks, with *p < 0.05, and **p < 0.01. nf V₂O₅: V₂O₅ nanofiber.

As with the nanoparticle network map, the network for nanofibers centered on AMP and was connected to many fatty acyl-CoAs (Fig. 6B), but subclusters were less well defined. One well defined subcluster was centered on glutamate and contained aspartate, ornithine and related metabolites (Fig. 6B, lower left). Niacin was present in a central subcluster with ATP; however, related metabolites as observed for V₂O₅ nanoparticles were not present. S-Adenosylhomocysteine was present but associated with histidine and histamine instead of niacin (Fig. 6B, upper left). Both purine and pyrimidine metabolites were present, but these were not clearly associated with specific subclusters. Specific examples of top metabolites associated with V₂O₅ nanofibers from the PLS-DA (Fig 6C) included S-adenosylhomocysteine (Fig. 6D), butyrylcholine (Fig. 6E), CTP (Fig. 6F), AMP (Fig. 6G) and uracil (Fig. 6H).

Organelle morphology and function with V nanomaterial exposure

To examine the effect of V₂O₅ nanomaterials on cellular organelles including lysosomes and phagolysosome, we performed TEM analysis with 1 ppm exposure. Fusion of lysosomes was commonly observed in V₂O₅ nanomaterial-exposed cells (Fig. 7B–7F), an event normally triggered by autophagic activities (Luzio, Pryor et al. 2007). Autophagic vacuoles (autophagosomes and autolysosomes) were apparent, and internalization of V₂O₅ nanoparticles in lysosomes (Fig. 7B–7F) were present. Use of a fluorescence-based lysosomal activity measurement showed that BEAS-2B cells exhibited increased lysosomal activity in response to 1 ppm V₂O₅ nanoparticles and nanofibers (Fig. 8). V₂O₅ nanoparticles and nanofibers had no apparent difference in effect on lysosomal activity at 1 ppm (Fig. 8B), however, and both V₂O₅ nanoparticles and nanofibers had less stimulating effect on the lysosomal activity at 0.5 ppm (Fig. S6A and S6B).

Fig. 7.

Representative TEM images of control BEAS-2B (A-B), 1 ppm V₂O₅ nanoparticles (C-D), and 1 ppm V₂O₅ nanofiber (E-F) after 24 h. Autophagic vacuoles including autophagosomes and autolysosomes were also noted in the V₂O₅ nanoparticle exposed cells (blue arrows). Fusion of lysosomes was commonly observed in both V₂O₅ nanoparticles or V₂O₅ nanofiber exposed cells (green arrows). Internalization of V₂O₅ nanoparticles or nanofibers was found in lysosome (red arrowheads). Note that it is technically challenging to capture a full picture of fibers in a cell from a monolayer, especially at such low concentration. Increased cell debris in lysosomes indicates stimulated breakdown of proteins and other cellular components (purple arrowheads). Swollen and dilated mitochondria with disrupted cristae structure were noted in both V₂O₅ nanoparticles or V₂O₅ nanofibers exposed cells (yellow arrows). Increased endoplasmic reticulum activity in both V₂O₅ nanoparticles or V₂O₅ nanofibers is indicated by black stars. Normal mitochondria are indicated by black arrows. Scale bars are indicated in μm.

Fig. 8.

Intracellular lysosomal activity in BEAS-2B cells following V₂O₅ nanofiber and nanoparticles exposure at 1 ppm. Representative images are shown in (A). Quantitative analyses are shown in (B). Bafilomycin A1 (1x) was used as a positive control that inhibits lysosomal function. nf V₂O₅: V₂O₅ nanofiber; np V₂O₅: V₂O₅ nanoparticle; SQS: self-quenched substrate.

Mitochondrial changes were also evident by TEM, with irregular shape, increased perimeter and enlarged size, and disorganized cristae (V₂O₅ nanofiber, Fig. 7E–7F; nanoparticle, Fig. 7D). JC-1 staining of control cells showed red fluorescence (Fig 9A,B), while an increased green fluorescence indicated a decrease in mitochondrial membrane potential in cells exposed to V₂O₅ nanofiber and nanoparticles at 1 ppm. Cells exposed to V₂O₅ nanofiber and nanoparticles at 0.5 ppm displayed a similar effect on mitochondrial membrane potential comparing to BEAS-2B cells exposed 1 ppm V₂O₅ nanofiber and nanoparticles (Fig. S6C and S6D).

Fig. 9.

Mitochondrial membrane potential measured by JC-1 dye in BEAS-2B cells following V₂O₅ nanofiber and nanoparticles exposure at 1 ppm. Representative images are shown in (A). Quantitative analyses are shown in (B). H₂O₂ (1 mM) was used as a positive control. nf V₂O₅: V₂O₅ nanofiber; np V₂O₅: V₂O₅ nanoparticle.

Discussion

V₂O₅ nanofibers and V₂O₅ nanoparticles possess unique electronic and mechanical properties, and the present results show that V₂O₅ nanomaterials lead to significant alterations to cellular metabolic networks in BEAS-2B cells at ppm and ppb levels. Earlier studies showed that V₂O₅ nanoparticles induce cell death in fibroblasts (Ivanković, Musić et al. 2006, Fukuto, Kim et al. 2021), tumor cells (Ivanković, Musić et al. 2006, Raj and Chatterjee 2016), and human corneal epithelial cells (Kim, Gates et al. 2020, Domanico, Fukuto et al. 2022). The present studies extend understanding of V nanotoxicity by showing extensive metabolic effects at doses that do not cause cell death within 24 h. Extension of toxico-metabolomics to in vivo research could be useful for low exposures where overt tissue injury can be difficult to detect. For instance, toxico-metabolomics may detect effects of V₂O₅ nanoparticles at doses below those increasing lactate dehydrogenase (LDH), alkaline phosphatase (ALKP) and γ-glutamyltransferase (GGT) in bronchoalveolar lavage (BAL) fluid, decreased glutathione (GSH) in plasma, or increased collagen deposition and inflammatory response in rat lungs (Kulkarni, Kumar et al. 2014). Such an approach may be especially important because V₂O₅ nanomaterial manufacture and use are increasing, and V₂O₅ nanomaterials are being developed for therapeutic applications.

Among the widespread effects of V₂O₅ nanomaterials, metabolic effects were centered on fatty acid metabolism and linked to carnitine shuttle, niacin, aminosugar and purine and pyrimidine metabolism (see Fig. 4A and Fig. 6A). Glycosphingolipid metabolism was downregulated in the V₂O₅ nanoparticle group (Fig. 4A and Fig. S4B), with significant increase in dolichyl phosphate and decrease in ethanolamine phosphate and CDP-ethanolamine. Earlier studies showed that lipid mediators derived from phospholipids and sphingolipids play an important role in the pathogenesis of pulmonary fibrosis (Summer and Mora 2019, Suryadevara, Ramchandran et al. 2020) and lung tumors (Moreno, Jiménez‐Jiménez et al. 2018). An extensive disruption in lipid metabolism occurred in all V₂O₅ exposure groups, suggesting that disruption of lipid metabolism could serve as an early sign for potential pulmonary fibrosis. Additionally, the carnitine shuttle was upregulated in both V₂O₅ nanoparticles (Fig. 4A, Fig. S4A) and V₂O₅ nanofiber group (Fig. 6A, Fig. S5A) indicating changes in fatty acid, phospholipid and sphingolipid metabolism, which are recognized as potential markers for pulmonary fibrosis (Geng, Liu et al. 2022) and lung cancer (Pamungkas, Medriano et al. 2017). In particular, increasing fatty acid oxidation in epithelial cells may promote pro-fibrotic cytokines, stimulate myofibroblasts and contribute to lung fibrosis (Geng, Liu et al. 2022). In addition to effects on niacin metabolism, major amino acids effects were seen for alanine, aspartate, methionine, cysteine, aspartate, asparagine, glycine, serine, alanine and threonine metabolism. An earlier study on the plasma samples following traffic pollution exposure identified vanadium as closely associated with arginine, histidine, and methionine, amino acids linked to inflammation and oxidative stress (Liang, Ladva et al. 2019). A previous study of BEAS-2B cells also showed that exposure to electronic cigarette vapors caused oxidative stress and effects on key pathways of energy and amino acids metabolism (Jarrell, Smith et al. 2021).

Among other changes, N-acetyl-hexosamine 6-phosphate (p < 0.001) was of interest because it was increased at all nanomaterial exposure levels. Previous studies showed that the increase of N-acetyl-hexosamine 6-phosphate was consistently linked to pulmonary hypertension (Rafikova, Meadows et al. 2016), pulmonary arterial hypertension (Liu, Qin et al. 2022) and idiopathic pulmonary fibrosis (Sun, Fernandez et al. 2018). Enhanced aminosugar and nucleotide sugar metabolism were also observed in bleomycin-induced fibrosis (Sun, Fernandez et al. 2018) and in lungs with tumors (Kim, Lee et al. 2020). Notably, in the present study, metabolite effects differed even though related pathways were altered. For instance, an increase in dTTP was associated with nucleotide sugar metabolism in the nanofiber group; whereas, elevation in UTP was associated with aminosugar and pyrimidine metabolism in the nanoparticles group.

Previous studies show subcellular disorganization in patients with pulmonary diseases (Cloonan and Choi 2016, Weidner, Jarenbäck et al. 2018). Importantly, lipid and fatty acid metabolism play important roles in proper function of endoplasmic reticulum and are required for the export of folded proteins from endoplasmic reticulum lumen (Velázquez, Tatsuta et al. 2016). Disrupted lipid and fatty acid metabolism can lead to endoplasmic reticulum stress (Han and Kaufman 2016), which could further contribute to inflammatory and profibrotic signaling, and drive downstream fibrotic remodeling in IPF lungs (Katzen and Beers 2020). We observed an elevated endoplasmic reticulum activity in cytoplasm of BEAS-2B cells exposed to V₂O₅ nanoparticles (Fig. 7D) and nanofibers (Fig. 7F), indicating a potential response to the stress induced by nanosized V₂O₅. Together with the data showing changes in lysosome activity (Fig. 7 and 8) and mitochondrial function (Fig. 7 and 9), our data suggest that a systematic organelle disruption might cause diverse alterations in lipid, amino acid, and aminosugar metabolism (Fig. 4 and 6). Earlier research showed that V₂O₅ nanoparticles accumulated explicitly in the lysosomes and mitochondria and impaired lysosomal function, caused mitochondrial damage, and promoted autophagy (Suma, Padmanabhan et al. 2020).

Additional studies will be needed to establish a causal relationship between disruption of organelles and metabolism, however, because several studies indicate that metabolic dysfunction can cause organellar effects. Earlier studies also showed that the niacin pathway may play critical role in attenuating pulmonary diseases such as pulmonary hypertension (Jia, Bai et al. 2020) and pulmonary fibrosis (Nagai, Matsumiya et al. 1994). The disruption in the niacin pathway by both V₂O₅ nanoparticles and nanofibers could severely impair mitochondrial function in energy metabolism. Moreover, vitamin D deficiency is reportedly associated with pulmonary fibrosis and impaired lung function and may cause over-activation of the renin-angiotensin system, which aggravates extracellular matrix deposition and lung fibrosis (Finklea, Grossmann et al. 2011, Shi, Liu et al. 2017). We observed vitamin D₃ was significantly decreased in V₂O₅ nanofiber exposure group, suggesting a possible role of vitamin D deficiency in organellar dysfunction. It should be noted however, that no significant changes were observed in the V₂O₅ nanoparticles group.

In summary, the present results show that exposure to V₂O₅ nanoparticles and nanofibers at sub-ppm and ppb levels cause a widespread alteration in cellular metabolism in human lung airway epithelial BEAS-2B cells, including lipid metabolism, carnitine shuttle, niacin, amino acid, and aminosugar metabolism. Cells exposed to V₂O₅ nanoparticles or nanofibers showed a number of shared pathways, while distinct patterns in individual metabolites were also present. V₂O₅ nanoparticles exhibited a more significant impact on glycerophospholipid and glycosphingolipid metabolism (Fig. 4); whereas, V₂O₅ nanofibers had more effects on sialic acid, cholesterol, and carbon fixation pathway (Fig. 6). The results indicate that the physiochemical properties of V₂O₅ nanomaterials, especially the shape and the size, may play important roles in toxicity. The observed disruption in cellular metabolism aligns with observed damage to lysosomes, endoplasmic reticulum, and mitochondria (Fig. 7, 8, and 9). Taken together, these results show that V₂O₅ nanomaterials at concentrations which do not cause cell death in BEAS-2B cells cause widespread effects on metabolism and organellar functions which could contribute to disease-causing pulmonary inflammation and fibrosis.

Conclusion

V₂O₅ is a toxic, redox-active compound with commercially useful high reactivity at nanoscale. The present study reveals that activities of V₂O₅ nanomaterials at environmentally relevant low concentrations induced a wide range of metabolic responses in a human bronchial epithelial cell line. The results showed that both V₂O₅ nanoparticles and nanofibers had dose-response effects on aminosugar, amino acid, fatty acid, carnitine, niacin and nucleotide metabolism. Differential effects of the particles and fibers included dibasic acid, glycosphingolipid and glycerophospholipid pathway associations with V₂O₅ nanoparticles, and cholesterol and sialic acid metabolism associations with V₂O₅ nanofibers. Results also showed distinct responses induced by V₂O₅ nanoparticles and nanofibers in cellular organelles, including mitochondria, lysosomes and endoplasmic reticula. Altogether, this study shows that V₂O₅ nanoparticles and nanofibers cause metabolic and ultrastructural changes in an airway cell model which could contribute to cell dysfunction and disease. The results further suggest utility of high-resolution metabolomics to strengthen mechanistic knowledge of nanomaterial interactions to evaluate toxicity in settings that represent real-life exposures.

Highlights.

• Low level nanosize V₂O₅ materials exposure causes significant metabolic alterations in human airway epithelial cells

• V₂O₅ nanoparticles and nanofibers share similar dose response on aminosugar, fatty acid, and carnitine metabolism

• V₂O₅ nanoparticles exhibit differential effects on dibasic acid, glycosphingolipid and glycerophospholipid pathway

• V₂O₅ nanofibers indicate differential effects on cholesterol and sialic acid metabolism