Aging Influence on Pulmonary and Systemic Inflammation and Neural Metabolomics Arising from Pulmonary Multi-walled Carbon Nanotube Exposure in Apolipoprotein E-Deficient and C57BL/6 Female Mice

Abstract

Objective:

Environmental exposures exacerbate age-related pathologies, such as cardiovascular and neurodegenerative diseases. Nanoparticulates, and specifically carbon nanomaterials, are a fast-growing contributor to the category of inhalable pollutants, whose risks to health are only now being unraveled. The current study assessed the exacerbating effect of age on multiwalled-carbon nanotube (MWCNT) exposure in young and old C57BL/6 and ApoE^−/− mice.

Materials and methods:

Female C57BL/6 and apolipoprotein E-deficient (ApoE^−/−) mice, aged 8 weeks and 15 months, were exposed to 0 or 40 μg MWCNT via oropharyngeal aspiration. Pulmonary inflammation, inflammatory bioactivity of serum, and neurometabolic changes were assessed at 24 h post-exposure.

Results:

Pulmonary neutrophil infiltration was induced by MWCNT in bronchoalveolar lavage fluid in both C57BL/6 and ApoE^−/−. Macrophage counts decreased with MWCNT exposure in ApoE^−/− mice, but were unaffected by exposure in C57BL/6 mice. Older mice appeared to have greater MWCNT-induced total protein in lavage fluid. BALF cytokines and chemokines were elevated with MWCNT exposure, but CCL2, CXCL1, and CXCL10 showed reduced responses to MWCNT in older mice. However, no significant serum inflammatory bioactivity was detected. Cerebellar metabolic changes in response to MWCNT were modest, but age and strain significantly influenced metabolite profiles assessed. ApoE^−/− mice and older mice exhibited less robust metabolite changes in response to exposure, suggesting a reduced health reserve.

Conclusions:

Age influences the pulmonary and neurological responses to short-term MWCNT exposure. However, with only the model of moderate aging (15 months) in this study, the responses appeared modest compared to inhaled toxicant impacts in more advanced aging models.

INTRODUCTION

Aging is an ongoing degenerative process resulting from complex interactions between genetics and various environmental factors. Aging produces physiological and pathological changes, including reduced energy production and utilization, impaired molecular/cellular/tissue repair mechanisms, neuropathology, and increased inflammation from the accumulation of senescent cells (1–4). As a result, aging is the strongest risk factor for many pathologies, including cardiovascular, metabolic, and neurological diseases. Epidemiological studies on the public health impact of air pollution provide strong evidence for negative effects on aging-related diseases and further indicate that individuals of advanced age may be vulnerable to toxic outcomes.

Particulate matter (PM) inhalation (environmental and engineered) as an environmental driver of aging-related diseases has broad global health implications. The World Health Organization estimates that 9 out of 10 persons globally are exposed daily to PM-polluted air that surpasses their set limits, resulting in 7 million people worldwide dying prematurely every year (5). The full impact of aging on responses to inhaled pollutants is currently incompletely characterized, although studies have shown that long term exposure to ambient air pollution significantly reduces life expectancy and exacerbates age related diseases in human populations (6, 7). Air pollution exposure is significantly associated with increased adverse cardiovascular outcomes, with sustained exposure to PM promoting increased mortality resulting from ischemic heart disease, arrhythmias, and heart failure (8). Exposure to PM decreased blood oxygen levels in older male subjects (>80 years old) and accelerated pulse rate in all subjects, suggesting that advancing age compromises our ability to adequately respond to environmental stressors (9). Furthermore, reducing PM in ambient air by 10 μg/m³ is associated with an increase in life expectancy of approximately 0.77 years (7).

Neurological impacts of PM have largely focused on the role these exposures play in the development of Alzheimer’s disease and related dementia. Grande et al. demonstrated that long-term exposure to air pollutants led to associations between heart failure, ischemic heart disease, and increased risk of dementia (10). Neuroinflammation is a hallmark of Alzheimer’s and associated dementia (11, 12) and so environmental stressors that promote either microglial activation or infiltration of peripheral leukocytes into the brain may exacerbate aging and related dementias. Tyler et. al. showed that ozone exposure induced neuroinflammation, which could be significantly augmented in aging models, possibly as a result of blood-brain barrier deficits (13). The consequences of engineered nanomaterials (ENM) on both pulmonary and systemic health are a more recent focus. We previously showed that short-term pulmonary exposure to multiwalled carbon nanotubes (MWCNT) led not only to pulmonary inflammation, but to disruption in blood brain barrier integrity and activation of astrocytes, as well as recruitment of microglia to areas of albumin leakage (14). Very little data exists to inform our knowledge of the impact of aging in these exposures. In our human health effect studies of workers exposed to carbon nanotubes and nanofibers, age followed a bimodal distribution with 49 % of the participants 45 years of age or older further supporting evaluations related to aging (15).

Apolipoprotein E (ApoE), produced primarily in the liver, is also expressed in multiple lung cells (16, 17) and in the central nervous system in astrocytes and microglia (to a lesser extent) (18, 19). It has been shown to play an importatnt role in maintaining pulmonary homeostasis and modulating responses to various inhalation exposures including immunoresponses. Additionally, the role of ApoE as a genetic risk factor for pathogenesis of the aging-related neurodegenerative Alzheimer’s disease is well explored (20–22). Altered pulmonary inflammatory responses to inhaled toxicants may be an important adaptation of aging and may influence the systemic and neural responses. Acute organic dust exposure in aged male C57BL/6 mice produced reduced inflammatory responses in the lung compared with young adult controls (19). Repeated exposures, however, led to increased inflammatory cell recruitment, suggesting that aging alters endogenous responses to particulate exposures and may impair the homeostatic benefits of inflammation (23). Reduced pulmonary inflammation in older animals was also observed following inhaled ozone exposures; interestingly, however, neuroinflammation was increased (24), suggesting that acute pulmonary responses to air pollution exposure may be protective in reducing systemic toxicity.

In this study we assessed the impact of age on exposure to multiwalled carbon nanotubes (MWCNT), an engineered nanomaterial with well-studied pulmonary health effects, in young (2 months) and old (15 months) healthy and ApoE-deficient (ApoE^−/−) female mice. ApoE^−/− mice exhibit elevated levels of circulating cholesterol that leads to vascular pathology (25); humans exhibit an age-related increase in cholesterol that is also associated with vascular morbidity and mortality (26). MWCNT have been shown to induce consistent dose-dependent pulmonary inflammation and systemic toxicity following acute exposures (14, 27). Neuroinflammatory activation observed following MWCNT exposure, in conjunction with serum peptide changes led us to examine the impact of these exposures on the initial pulmonary responses in a model of vascular disease with aging as a key component.

METHODS

Animal model and sample collection

Specific pathogen-free female C57BL/6 and ApoE^−/− mice (Jackson Laboratory) aged 6–8 weeks and 15 months were used in this study. Animals were housed in an Association for Assessment and Accreditation of Lab Animal Care International-approved animal facility at the University of New Mexico, with all experimental procedures approved by Institutional Animal Care and Use Committee (IACUC) of the University of New Mexico. Animal care and use procedures were conducted in accordance with the US Public Health Service’s Policy on Humane Care and Use of Laboratory Animals (https://grants.nih.gov/grants/olaw/references/phspol.htm) and the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals (https://grants.nih.gov/grants/olaw/Guide-for-the-Care-andUse-of-Laboratory-Animals.pdf). Food and pre-packaged water were provided ad libitum in ventilated cages in a temperature- and humidity-controlled environment with a 12-h light/dark cycle. Mice were administered MWCNT via oropharyngeal aspiration at doses of 0 μg or 40 μg (n = 5–7/group). The MWCNTs were prepared in a dispersion media (DM) consisting of mouse serum albumin (0.6 mg/mL) and 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (10 μg/mL); DM was used as a control vehicle. The MWCNT material used in this study, Mitsui-7, has been extensively characterized (28–31), with a mean diameter of 49 nm and length of 3.86 μm (geometric standard deviation = 1.94). Mice were euthanized 24-h following MWCNT pulmonary dosing. At the time of euthanasia, blood and bronchoalveolar lavage fluid (BALF) were collected. Serum was spun from blood that was allowed to coagulate on ice for 30 min. Following transcardial perfusion with ice-cold 1X PBS, the cerebellum was collected for metabolomic assessment.

Pulmonary Inflammation Assessment – Mesoscale Assay

MWCNT-induced changes in BALF cytokines were determined using the Meso Scale Discovery MULTI-SPOT V-PLEX^® Cytokine Assay System Pro-inflammatory Panel 1 (mouse) Kits (K15048D - Meso Scale Diagnostics LLC, Rockville, MD) according to manufacturer’s instructions. Briefly, BALF was collected by lavaging lungs of euthanized MWCNT exposed mice with ice-cold 1X PBS. Cell fractions were removed by centrifugation and 50 μl/well of the supernatant were loaded onto sample plates pre-coated with capture antibodies for the following cytokines: interferon gamma (IFN-γ), interleukin-10 (IL-10), interleukin-12 (IL-12p70 active heterodimer), interleukin-1 beta (IL-1β), interleukin-2 (IL-2), interleukin-4 (IL-4), interleukin-5 (IL-5), interleukin-6 (IL-6), chemokine C-X-C motif ligand 1 (CXCL1, aka KC/GRO), and tumor necrosis factor alpha (TNF-α). Plates were incubated with gentle shaking for 2 h at room temperature. Plates were washed 3X with buffer containing 1X PBS and 0.05% Tween 20. Detection antibody was added to each well and allowed to react for 1 h at room temperature. Plates were washed as above, and Read Buffer was added to each well. Plates were analyzed on an MSD QuickPlex SQ 120 instrument (MSD, AI0AA-0); Discovery Workbench (v. 4.0) software calculated cytokine concentrations using a linear regression analysis of the standard curve. All concentrations were normalized to total BALF protein, determined using a standard Bradford protein assay.

Serum Cumulative Inflammatory Potential Assay

Mouse cerebrovascular endothelial cells (mCEC) were obtained from a commercial vendor (Cell Biologics, Chicago, IL) and maintained according to manufacturer’s recommendations at 37ºC and 5% CO₂ with complete endothelial cell medium supplemented with 5% fetal bovine serum. All experiments were conducted with cells between passages 3 and 8. To determine the serum cumulative inflammatory potential of MWCNT exposure, MBECs were treated with serum isolated from MWCNT-exposed and control mice as previously described (32, 33). Briefly, MBECs were serum starved overnight then incubated in FBS-free culture media supplemented at a final concentration of 5% v/v serum from control (0 μg) or 40 μg MWCNT exposed mice for 4 h and harvested. RNA was isolated using the RNeasy Mini Kit (QIAGEN, Germantown, MD), and reverse transcribed prior to gene expression analyses via quantitative real-time PCR (qPCR). Expression of mouse Ccl2 (Mm00441242_m1), Icam1 (Mm00516023_m1), Il6 (Mm00446190-m1), Tnfa (Mm00443258_m1), Tgfb (Mm01178820_m1), and Vcam1 (Mm01320970_m1) (Applied Biosystems, Foster City, CA) was measured using the TaqmanR Gene Expression protocol (ThermoScientific, Waltham, MA) following the manufacturer’s instructions. Relative gene expression normalized to the endogenous gene expression control TATA-Box Binding Protein (TBP) (Mm00446973_m1) gene was determined using the 2^−ΔΔCT method for all samples with threshold cycle values (CT) under 35. Results are expressed as fold change.

Cerebellar Metabolomics

Reagents

Acetonitrile (ACN), methanol (MeOH), ammonium acetate, and acetic acid, all liquid chromatography – tandem mass spectrometry (LC-MS/MS) grade, were purchased from Fisher Scientific (Pittsburgh, PA). Ammonium hydroxide was bought from Sigma-Aldrich (Saint Louis, MO). DI water was provided in-house by a Water Purification System from EMD Millipore (Billerica, MA). PBS was bought from GE Healthcare Life Sciences (Logan, UT). The standard compounds corresponding to the measured metabolites were purchased from Sigma-Aldrich (Saint Louis, MO) and Fisher Scientific (Pittsburgh, PA).

Tissue preparation

Briefly, each tissue sample (~20 mg; N=5–6 per group) was homogenized in 200 μL MeOH:PBS (4:1, v:v, containing 1,810.5 μM ¹³C₃-lactate and 142 μM ¹³C₅-glutamic Acid) in an Eppendorf tube using a Bullet Blender homogenizer (Next Advance, Averill Park, NY). Then 800 μL MeOH:PBS (4:1, v:v, containing 1,810.5 μM ¹³C₃-lactate and 142 μM ¹³C₅-glutamic Acid) was added, and after vortexing for 10 s, the samples were stored at −20°C for 30 min. The samples were then sonicated in an ice bath for 30 min, centrifuged at 14,000 RPM for 10 min (4°C), and 800 μL supernatant was transferred to a new Eppendorf tube. The samples were then dried under vacuum using a CentriVap Concentrator (Labconco, Fort Scott, KS). Prior to MS analysis, the obtained residue was reconstituted in 150 μL 40% PBS/60% ACN. A quality control (QC) sample was pooled from all the study samples.

LC-MS/MS

The targeted LC-MS/MS method used here was modeled after that developed and used in a growing number of studies with detailed lists of metabolites (34–39). Briefly, we targeted ~300 metabolites selected from >35 metabolic pathways of biological significance, including glycolysis, TCA cycle, purine metabolism, amino acid metabolism, etc. All LC-MS/MS experiments were performed on an Agilent 1290 UPLC-6490 QQQ-MS (Santa Clara, CA) system. Each sample was injected twice, 10 μL for analysis using negative ionization mode and 4 μL for analysis using positive ionization mode. Both chromatographic separations were performed using hydrophilic interaction chromatography (HILIC) on a Waters XBridge BEH Amide column (150 × 2.1 mm, 2.5 μm particle size, Waters Corporation, Milford, MA). The flow rate was 0.3 mL/min, auto-sampler temperature was kept at 4°C, and the column compartment was set at 40°C. The mobile phase was composed of Solvents A (10 mM ammonium acetate, 10 mM ammonium hydroxide in 95% H₂O/5% ACN) and B (10 mM ammonium acetate, 10 mM ammonium hydroxide in 95% ACN/5% H₂O). After the initial 1 min isocratic elution of 90% B, the percentage of Solvent B decreased to 40% at t=11 min. The composition of Solvent B maintained at 40% for 4 min (t=15 min), and then the percentage of B gradually went back to 90%, to prepare for the next injection.

The mass spectrometer is equipped with an electrospray ionization (ESI) source. Targeted data acquisition was performed in multiple-reaction-monitoring (MRM) mode. All LC-MS/MS parameters, including MRM transitions, collision energy (CE), retention time, etc., were confirmed and validated using metabolite standards. The whole LC-MS system was controlled by Agilent Masshunter Workstation software (Santa Clara, CA). The extracted MRM peaks were integrated using Agilent MassHunter Quantitative Data Analysis (Santa Clara, CA).

Statistical analyses of Metabolomics Data RStudio

Packages used for dataset analyses and presentation: qqman for the Manhattan plot (https://github.com/drveera/ggman), dplyr for data transformation (https://cran.r-project.org/web/packages/dplyr/index.html), ggplot2 for graphing resultant plots (https://cran.r-project.org/web/packages/ggplot2/index.html), ggpubr for additional plotting of data (https://cran.r-project.org/web/packages/ggpubr/index.html), car for some statistical analysis and multi-way ANOVAs (https://cran.r-project.org/web/packages/car/index.html), foreign for read-write capabilities (https://cran.r-project.org/web/packages/foreign/index.html), rcompanion for graph alterations (https://cran.r-project.org/web/packages/rcompanion/index.html), and tidyverse for aesthetic changes (https://cran.r-project.org/web/packages/tidyverse/index.html).

Targeted NAD+ pathway analysis

After receiving the metabolomics data, we removed a single outlier mouse with cerebellum weight of 7.8 grams after performing a Grubb’s test (Z=4.460, Zcrit=3.10324). The remaining metabolites were normalized per mouse weight. Linear discriminate analyses were performed for each mouse strain followed by multiway ANOVA (p<0.10). To the statistically significant metabolites, Tukey’s Honest Significant Differences tests were performed (confidence level = 0.95) to determine groupings.

Overall metabolite examination

The entire metabolite dataset was run recursively through three-way ANOVAs and a secondary database was generated from metabolites that showed statistical significance in at least one category of comparison. That dataset was used for the generation of the Manhattan plot (cutoff line at 1.3 = p<0.05, cutoff line at 2 = p<0.01.

Statistics

All statistics were conducted in GraphPad Prism (v9.0) or RStudio. Data were tested for Guassian normality using a Shapiro-Wilk test; most data sets were observed to be normal. The infrequent nature of non-Guassian distribution of data led to the choice of standard parametric statistical tests for all assays. Data were assessed as a two-way ANOVA, considering MWCNT and age as the 2 factors. Tukey’s post-hoc multiple comparison test was used when appropriate. A three-way ANOVA was used for metabolomic data, considering MWCNT, age, and strain.

RESULTS

Pulmonary Inflammation

BALF Inflammatory Cell Profile: C57BL/6 Mice

Post-exposure assessment of BALF cell counts and total protein from MWCNT- and DM- exposed young and old mice revealed an age-independent inflammatory activation in C57BL/6 female mice (Figure 1). Total cell count and total protein were unchanged across age groups by MWCNT exposure (Figure 1A, 1D). Macrophage density was not significantly changed by MWCNT exposure (Figure 1B), although old animals display a lower trend in DM and MWCNT groups. The same MWCNT exposure led to a statistically significant increase in polymorphonuclear neutrophil (PMN) infiltration in both young and old animals (p<0.05) (Figure 1C). The unchanged total cell and protein values may represent a shift in the inflammatory response resulting from macrophage mediated increase in PMN infiltration in C57BL/6 animals.

Figure 1.

Airway inflammation in C57BL/6 female mice exposed to MWCNT. A. Total cells in bronchoalveolar lavage (BALF); B Macrophage differential cell counts; C. Polymorphonuclear Leukocytes (PMN) differential cell counts; and D. Total BALF protein. N=5–6 per group. Data presented are means ± SEM. Asterisks (*) indicate significant effect of MWCNT by a 2-way ANOVA. No significant effect of age was noted.

BALF Inflammatory Cell Profile: ApoE^−/− Mice

MWCNT exposure in the ApoE^−/− model of impaired cholesterol clearance exhibited a greater overall inflammatory response as compared to C57BL/6 mice (Figure 2). ApoE^−/− female mice showed no significant differences in the total cell count in response to MWCNT (Figure 2A) but exhibited a significant age-independent decrease in macrophage differential cell counts following MWCNT exposure (p<0.001) (Figure 2B). The decreased macrophage cell count was accompanied by a statistically significant and robust BALF PMN increase following MWCNT exposure in both young and old animals (p<0.0001; Figure 2C). The corresponding total protein assessment indicated an age-dependent BALF protein shift with older animals showing a significant increase in total protein in MWCNT-exposed groups compared with their DM controls ((p<0.001) and the MWCNT-exposed young counterparts (p<0.05; Figure 2D).

Figure 2. Airway inflammation in ApoE^−/− female mice exposed to MWCNT.

A. Total cells in bronchoalveolar lavage (BALF); B. Macrophage differential cell counts. Asterisks (***) indicate significant (p<0.001) effect of MWCNT by a 2- way ANOVA. No significant effect of age was noted; C. Polymorphonuclear Leukocytes (PMN) differential cell counts. Asterisks (****) indicate significant (p<0.0001) effect of MWCNT by a 2-way ANOVA. No significant effect of age was noted; and D. Total BALF protein. 2-way ANOVA revealed an interactive effect of age and MWCNT exposure, with old ApoE−/− mice displaying a significant increase in total protein compared to all other groups by Tukey’s multiple comparison test (*p<0.05; ***p<0.001). N=5–6 per group. Data presented are means ± SEM.

The influence of age on total protein content after MWCNT treatment was statistically significant by two-way ANOVA analysis (p<0.05), indicating an augmented response in older animals. Young ApoE^−/− females did not show changes in total protein after MWCNT compared to young DM-treated mice, and the overall levels of protein were consistent with that measured in C57BL/6 mice. The responses to MWCNT exposure between C57BL/6 and ApoE^−/− groups showed more robust responses in ApoE^−/− mice, suggesting there may be an increased susceptibility to acute MWCNT exposure in this model compared to C57BL/6.

BALF Cytokine Expression: C57BL/6 mice

To further determine the level of inflammatory activation induced by MWCNT exposure, we assessed cytokine protein expression via Mesoscale ELISA assay. Differential exposure- and age-related cytokine profiles were revealed in both C57BL/6 and ApoE^−/− groups. In C57BL/6 mice, MWCNT treatment and aging interactions were seen with major inflammatory cytokines IL-β, IL-6, and CCL2, with older mice exhibiting potentiated responses to MWCNT (Figure 3A–C). TNFα and MIP-1α were both significantly elevated by MWCNT treatment, but unaffected by age (Figure 3D,E). CXCL1 and CXCL10 both displayed significant induction due to MWCNT treatment that was attenuated in 15-mo mice (Figure 3,F.G). Several cytokines remained unchanged by MWCNT and age (IFNγ, IL-10, IL-12p70; Figure 3H–J).

Figure 3.

Airway cytokine expression in C57BL/6 female mice exposed to MWCNT. Cytokines are grouped according to pattern of response across dose and age. Some cytokines increased with MWCNT treatment and were augmented by advanced age, including IL-1β, IL-6, and MCP-1 (A-C, respectively). Some cytokines responded to MWCNT treatment and were unaffected by age, including TNFα and MIP-1α (D, E). Cytokines that were elevated by MWCNT in 2-mo mice but significantly reduced in 15-mo old mice included KC-GRO and IP-10 (F, G). IFNγ, IL-10, and IL-12p70 were unchanged by MWCNT treatment or age. Asterisks indicate significant (*p<0.05; **p<0.01; ***p<0.001; ****p<0.0001) effect by a 2-way ANOVA with Tukey’s multiple comparison test. N=5–6 per group. Data presented are means ± SEM.

BALF Cytokine Expression: ApoE^−/− mice

Unlike C57BL/6 mice, aging did not impact the magnitude of major inflammatory cytokines IL-β, IL-6, CCL2, and TNFα response to MWCNT in ApoE^−/− (Figure 4A–D). 15-mo ApoE^−/− mice did not exhibit a significant MIP-1α response to MWCNT (p=0.09), although younger ApoE^−/− mice did (Figure 4E). Similar to findings in C57BL/6 mice, MWCNT treatment induced CXCL1 and CXCL10 BALF increases that were significantly greater in young mice compared to 15-mo mice (Figure 4F,G). Interestingly, MWCNT exposure produced a significant decrease in IFNγ BALF protein in 15-mo ApoE^−/− mice (Figure 4H). IL-10 and IL-12p70 cytokines remained unchanged by either MWCNT and age (Figure 4I,J). In general, ApoE−/− mouse responses appeared dampened compared to C57BL/6. This was possibly a factor in limiting the lack of aging-associated augmentation of IL-β, IL-6, MCP-1 and TNFα responses to MWCNT.

Figure 4.

Airway cytokine expression in ApoE^−/− female mice exposed to MWCNT. Cytokines are grouped according to pattern of response across dose and age. Some cytokines increased with MWCNT treatment but were unaffected by age, including IL-1 p, IL-6, MCP-1, TNFa, and MIP-1a (A-E, respectively). Cytokines that were elevated by MWCNT but significantly reduced in 15-mo old mice included KC- GRO and IP-10 (F,G). 15-mo old mice treated with MWCNT displayed a significant reduction in IFNy, and effect not seen in 2-mo old mice. IL-10 and IL-12p70 were unchanged by MWCNT treatment or age. Asterisks indicate significant (*p<0.05; **p<0.01; ***p<0.001; ****p<0.0001) effect by a 2-way ANOVA with Tukey’s multiple comparison test. N=5–6 per group. Data presented are means ± SEM.

A 3-way ANOVA was applied to compare the differences between C57BL/6 and ApoE^−/− mice, along with age and MWCNT treatment. The 3-way ANOVA identified strain-dependent differences for MIP1a, CXCL10, IL1b, CXCL1, IL12p70, IL-10, and IFNγ. In all cases, the ApoE^−/− mice appeared to have lower values at baseline or after response to MWCNT. Strain x MWCNT interactions were observed only for CXCL10 and IL-10.

Serum Cumulative Inflammatory Potential Assessment

To begin to define the systemic consequences of MWCNT exposure, we performed an in vitro assay using mCECs. mCECs were incubated with 5% serum from C57BL/6 and ApoE^−/− animals to assess the degree of systemic bioactivity generated by lung MWCNT exposure and any age-related consequences (Figure 5 and 6). Using two-way ANOVAs, no effects were noted for MWCNT or age in each strain (C57BL/6 and ApoE^−/−). However, given that some of the trends were similar across the strains, we considered that our data may be underpowered and applied a three-way ANOVA to better incorporate all the data. Several significant trends emerged. Vcam1, for instance, was significantly upregulated in ApoE−/− mice (p=0.0084) and we also saw an Age x Strain interaction (p=0.049). A similar Strain related effect was seen for Tgfb (p=0.0106) along with an interaction between Age x MWCNT treatment (p=0.0212). A Strain effect was also seen for Icam1 (p=0.046).

Figure 5.

Serum Cumulative Inflammatory Potential (SCIP) assessment of serum from C57BL/6 female mice exposed to MWCNT. Mouse cerebrovascular endothelial cells (mCEC) were treated with serum from old and young dispersion media (DM) and MWCNT. Inflammatory gene expression changes were assessed via qPCR. No statistically significant changes were induced in endothelial cells in culture following exposure to serum from mice treated with MWCNT. A non-significant trend towards a decrease in cytokine gene expression in both young and old groups was observed for IL-6. Data assessed using 2-way ANOVA. N=5–6 per group. Data presented are means ± SEM.

Figure 6.

Serum Cumulative Inflammatory Potential (SCIP) assessment of serum from ApoE^−/− female mice exposed to MWCNT. Mouse cerebrovascular endothelial cells (mCEC) were treated with serum from old and young dispersion media (DM) and MWCNT. Inflammatory gene expression changes were assessed via qPCR. No statistically significant changes were induced in endothelial cells in culture following exposure to serum from mice treated with MWCNT. Data assessed using 2-way ANOVA. N=5–6 per group. Data presented are means ± SEM.

Neurometabolomics

To assess neurological consequences of MWCNT exposure in young and old mice, we assessed the metabolic changes in cerebellum of exposed C57BL/6 and ApoE^−/− animals (Figure 7–10). Complete heatmaps of metabolites for C57BL/6 and ApoE−/− mice are shown in Figure 7 (raw data provided in Supplemental Table 1). Linear discrimination analyses found that C57BL/6 accuracy was estimated at 80% by machine learning and 92% of variance was explained in the dimensions shown (LD1=62.09%, LD2=29.67%, LD3=8.24%; Figure 8A). ApoE^−/− data were subjected to the same analysis; however, the amount of colinear variables deemed the accuracy at 66% by machine learning, with 79% of variance explained in the first two dimensions (LD1=51.36%, LD2=28.55%, LD3=20.9%; Figure 8B). Analysis of C57BL/6 metabolomic responses produced more discrete grouping with MWCNT inducing a shift in young mice, but with little effect in old mice. APOE^−/− animals showed more overlapping groupings between age and exposure, suggesting a muted overall metabolomic response. Further analyses are required to determine functional implications of the groups identified and the role the play in modulating cerebellar metabolic changes following MWCNT exposure.

Figure 7.

Heat map representation of cerebellar metabolomics of MWCNT-exposed C57BL/6 and APOE^−/− female mice. Untargeted analysis of all metabolites assessed. N=5–7 per group. Data presented are means ± SEM.

Figure 10.

Cerebellar NAD+ and related metabolites MWCNT exposed C57BL/6 and APOE^−/− female mice. Targeted analysis revealed MWCNT exposure, age, and strain-related changes to NAD+ and several related metabolites. N=5–7 per group. Data presented are means ± SEM. Adenosine diphosphate (ADP); Adenosine triphosphate (ATP); Nicotinamide adenine dinucleotide (NAD); Nicotinamide riboside (NR); nicotinamide mononucleotide (NMN); nicotinamide (NAM); nicotinic acid mononucleotide (NaMN); nicotinic acid adenine dinucleotide (NaAD). NMN (A) and NaAD (B) are precursors to NAD+ (C). ADP Ribose (D) can be made from NAD+ during times of DNA damage.

Figure 8.

Cerebellar metabolomics of MWCNT exposed C57BL/6 and APOE^−/− female mice. C57BL/6 (left) and APOE^−/− (right) linear discriminate analyses. C57BL/6 (A) and APOE (B) linear discriminate analyses. C57BL/6 accuracy was estimated at 80% by machine learning and 92% of variance was explained in the dimensions shown (LD1 =62.09%, LD2=29.67%, LD3=8.24%). ApoE^−/− was subjected to the same analysis. However, the amount of colinear variables deemed the accuracy at 66% by machine learning, with 79% of variance explained in the first two dimensions (LD1=51.36%, LD2=28.55%, LD3=20.9%).

A three-way analysis was conducted to examine the influence of the major experimental factors (Age, Strain, MWCNT treatment) on cerebellar metabolites. A Manhattan plot (Figure 9) summarizes the outcomes, with the ApoE−/− strain having a major influence on metabolite concentrations in the cerebellum, with over 20 metabolites altered with p<0.01. Age had a lesser effect with 7 metabolites different at p<0.01 and 19 at p<0.05 (no false discovery rate correction applied). MWCNT exposure on its own induced minimal alterations (6 metabolites of p<0.05), but interactions between age, strain, and MWCNT exposure identified several additional factors that may be altered by MWCNT only in older mice. A complete table of significantly altered metabolites, corresponding with the Manhattan plot is provided in the data supplement (Supplemental Table 2).

Figure 9.

Manhattan plot of cerebellar metabolomics of MWCNT exposed C57BL/6 and APOE^−/− female mice. 3-way ANOVA compared the influence of Age, MWCNT exposure, and Strain on the cerebellar metabolome. N=5–7 per group. Data presented are means ± SEM.

Targeted nicotinamide adenine dinucleotide (NAD+) pathway analysis was performed to assess changes in cerebellar energetics and oxidative stress and the role played by age in mediating MWCNT effects (Figure 10). NAD⁺ levels were significantly elevated by MWCNT exposure in young C57BL/6 animals, but conversely decreased by exposure in young ApoE^−/− animals. Old animals of either animal strain did not show significant changes with exposure to MWCNT. The precursor nicotinamide mononucleotide (NMN) was observed as statistically unchanged by exposure and age. Adenosine diphosphate (ADP) ribose was also investigated as a marker of MWCNT-induced DNA damage. Apart from converse responses seen in young DM and young MWCNT-exposed C57BL/6 mice, no other indicators of effect were seen. However, the NAD⁺ precursor nicotinic acid adenine dinucleotide (NaAD) was strikingly similar to NAD⁺ levels for all conditions, implying NaAD and the tryptophan pathway as the preferred cerebellar pathway for NAD⁺ production. No statistical differences were observed in the NaAD precursor, nicotinic acid mononucleotide (NaMN), and the ATP consumption at this conversion step would appear to be rate-limiting for NaAD production. In turn, this would limit the production of NAD⁺, as the NMN precursor was unchanged by exposure and age.

DISCUSSION

In the present study, we assessed the lung inflammatory and neurometabolomic consequences of nanomaterial exposure in a susceptible model. We evaluated the role of age as a mitigating factor in MWCNT exposure-induced pulmonary and systemic outcomes in the hyperlipidemic ApoE^−/− female mouse model. We demonstrated age-independent pulmonary inflammatory modulation in which the ApoE-null model of hypercholesterolemia played a significant role in exacerbating cellular responses to exposure. Additionally, we showed, as expected from previous studies, robust neutrophil influx was induced in both ApoE^−/− and C57BL/6 groups (assessed by BALF differential cell counts) by MWCNT exposure, although significantly more so in ApoE^−/− (40, 41). A corresponding decrease in airway macrophages was observed in the ApoE^−/− groups, but not in the C57BL/6, after MWCNT exposure. This differential expression between macrophages and neutrophils in ApoE^−/− has been previously reported with carbon black, carbon nanotubes, and quantum dots (42, 43). Decreased macrophage function has also been reported with age and likely reflects an acceleration of particle clearance and influences recruitment of inflammatory cells to sites of exposure. Additionally, persistent neutrophil presence in both local lung environment and systemic circulation is facilitated by age-related loss of clearance efficiency with concurrent increases in ROS production (44). The natural increase of BALF neutrophils with age is well documented (45), although not observed in our study, as both old and young animals of both strains exhibited MWCNT exposure-induced increases independent of age. The expected adjuvant role of MWCNT in increasing neutrophil infiltration in aging animals was not observed in either C57BL/6 or ApoE^−/− animals. In contrast to our findings in ozone-exposed animals (13), C57BL/6 mice showed no significant changes in lung macrophages, although a non-significant trend towards overall lower macrophage presence was observed in older animals.

ApoE deficiency also exacerbated MWCNT-induced lung injury, as indicated by BALF protein, potentially due to enhanced pulmonary endothelial permeability (46) in older mice, an effect not seen in C57BL/6 mice. Increased vascular permeability with age is a common phenomenon resulting from vascular wall remodeling, barrier integrity changes and general age-related functional decline in endothelial function (47). Previous studies of MWCNT exposure in ApoE^−/− mice also demonstrated increased BALF total protein with various particulate and nanomaterial exposures (40). Long et al. showed that oxidative stress played a significant role in altering human endothelial cell permeability in vitro with MWCNT exposure increased ROS (48) and lipid accumulation in THP-1 macrophages cell lines (49). Although not assessed here, we also previously reported a corresponding increase in the cellular damage marker, lactate dehydrogenase, as an accompanying feature of MWCNT and other particulate exposures (29, 50, 51). These findings are important from a clinical perspective, since increased vascular permeability is common in aging and associated with adverse cardiovascular and neurological outcomes.

We have previously demonstrated significant serum peptide compositional changes following MWCNT exposure, which appear to be pulmonary-derived and impair blood-brain barrier permeability with accompanying neuroinflammatory activation (14, 27). Such a transfer of bioactivity may be facilitated by cytotoxicity induced by exposure as a result of increased oxidative stress resulting in impaired alveolar air-blood barrier as has been shown in prior studies using PM and nanomaterials (52–57). The reduction in blood-brain barrier has been shown with other inhaled toxicants, and is a likely pathway by which pulmonary toxicity can contribute to acute and long-term neurological outcomes. (13, 14, 24) Blood-brain barrier deficits are apparent in numerous neurological diseases, initiate neuroinflammatory outcomes, and are conjectured to be an essential player in the pathogenesis of aging-related neurodegenerative diseases. (11, 58) Additionally, Stapleton et al. reported serum albumin in BALF following MWCNT exposure (59), which introduction into the airways might exacerbate inflammatory responses leading to further disregulation of the airway-blood barrier. ApoE^−/− mice have been shown to be susceptible to pulmonary endothelial damage directly related to increased oxidized lipids (46). Higher baseline cholesterol levels in ApoE^−/− (60), along with toxicity from MWCNT exposure compounds oxidative stress appeared to promote endothelial and epithelial dysfunction and subsequent systemic inflammatory activation by pulmonary derived bioactives.

Pro-inflammatory mediators such as TNF-α, TGF-β and IFNγ are known to alter endothelial homeostasis through disrupted endothelial cell proliferation, increased permeability, and altered endothelial cell connections/communication with the extracellular matrix (61). We assessed the levels of chemokines and cytokines in BALF of ApoE^−/− and C57BL/6 mice exposed to MWCNT in this study. Similar patterns of expression were observed between C57BL/6 and ApoE^−/− animals, with exposure increasing most cytokines assessed. C57BL/6 showed the greatest age-related expression changes in cytokine expression with IL-1β, IL-6, MCP-1 being increased in old animals, while CXCL1 and CXCL10 were decreased. Only CXCL1, CXCL10 (mimicking the decrease in the controls), and IFNγshowed age-related effects in ApoE^-/−. Given the high levels of CCL2, a monocyte/macrophage attractant, one might expect a corresponding increase in BALF macrophages but this was not the case in our acute study. The role of CCL2 in neutrophil recruitment is much debated; however, studies have shown that neutrophil recruitment and function at sites of bacterial infection in the lung was impaired in CCL2^−/− animals (62, 63). Despite similar neutrophil counts in the BALF of both age groups and strains, CCL2 was more significantly elevated in old C57BL/6 mice compared to young animals, while levels were not different between ApoE^−/− age groups. Elevated CCL2 levels also persist in aged ApoE^−/− when larger doses and longer experimental time points are employed (64). Decreases were also observed in the expression levels of CXCL1 and CXCL10 in the aged animals in both the C57BL/6 and ApoE^−/− groups. CXCL1, which plays a key role in neutrophil recruitment and activation, were surprising given the the increase in neutrophils observed in these groups. CXCL1 exists as both monomers and dimers, with both the monomer and dimer activating the CXCR2 receptors in neutrophils (65). CXCL10 is secreted in response to IFNγ signaling from and so the observed decrease in ApoE^−/− is understandable given then significant decrease in IFNγ levels. The substantial decrease in CXCL10 response to MWCNT in C57BL/6 aged animals is harder to explain and the implications are unclear.

Surprisingly, the bioactivity of serum from exposed female mice did not induce an inflammatory response in endothelial cells grown in culture in any of the groups assessed. This is in stark contrast to several previous studies showing increased inflammatory activation induced by MWCNT-exposed serum, when compared to serum from unexposed animals (14, 27, 29, 66, 67). We conclude that batches of primary endothelial cells may exhibit different sensitivity, contributing uncertainty to such approaches. Alternatively, as most previous work was conducted in male mice, there is a possiblity that female mouse serum is more protective of inflammatory alteration. While we have previously considered using such approaches as more broadly applicable clinical assessments of vascular risk, the present findings represent an important caveat that standardization of this ex vivo / in vitro methods may be more complicated than previously thought.

Further assessment of the systemic consequences of a pulmonary exposure to MWCNT centered on neurometabolic changes. The brain has high energy demands that must be maintained for optimal function. As we age, the loss in metabolic capacity often foreshadows cognitive impairments (68). The aging neuron is vulnerable to reduced metabolic capacity, highlighted as a loss of nicotinamide adenine dinucleotide (NAD+) concentrations, increased AMP/ATP ratio and purine and pyrimidine accumulation (69). This reduction in bioenergetics adversely impacts DNA repair mechanisms which appears central to a downward spiral of cellular functionality (70–72). Mice in the present study were not truly of “advanced age”, but at 15 months (aproximately mid-40s in humans), there were still substantial metabolic differences compared to 2 month old mice (approximately 20 years old in humans). Cerebellar metabolomic responses to MWCNT were modest, but statistically significant effects still highlight the systemic ramifications of pulmonary particulate exposures. Holistically, the aged mice exhibited a less robust response to MWCNT than young mice, based on the principal component analysis. This was most notable for NAD+, interestingly, with young mice displaying an increase in NAD+ levels, but this response was dampened in older mice and in the ApoE^−/− mice. Several metabolites were altered by aging that have been previously identified as aging-related metabolites in the cerebellum, including pantothenic acid, citrulline, adenosine, and carnosine. However, MWCNT alone altered only a few (5) published factors that are altered in the cerebellum by aging, including pantothenic acid and betaine. (73) MWCNT interactions with aging, strain, and both were also apparent, suggesting that complex pathways and responses may be involved in the linkage between lung inflammation and nearometabolomic changes. Future studies should examine these metabolomic outcomes across a wider range of response times, as well as in mice in the 18–24 month range. The 15 monoth old mouse is more of a model of middle age in humans, but as such may model a perimenopausal period. The impact of MWCNT exposure in perimenopausal women, whose hormonal changes may leave them already susceptible to cognitive impairments (74–76), should be considered in future studies.

The most obvious caveat to the straightforward interpretation of these studies is the difference in body weight between the young (~20g) and old (~35g) mice, for both strains. This difference in body weight means that older mice effectively received a reduced relative dose of MWCNT. In terms of those cytokines or other measures that appeared reduced in comparing the old versus young mice, e.g., CXCL1 and CXCL10, this diminution of response may mostly relate to the lower relative dose of MWCNT in older (larger) mice. However, this consideration of body weight amplifies the impact of those markers clearly increased in older mice, such as the IL-6, IL1β, and MCP-1 responses in C57BL/6 mice, and overall total protein levels in lavage from both strains.

Here we reported on the inflammatory potential of short-term MWCNT exposure in an aging WT and ApoE−/− mouse model. ApoE deficiency resulted in more robust pulmonary inflammatory cell recruitment and increased lung damage. While the modestly advance age of 15 months did not seem to play a major role in the inflammatory activation, the extent of protein leakage resulting from exposure was significantly worse in older mice, indicating a risk of more adverse and possibly prolonged effects in older populations. Additionally, while cerebellar metabolic changes indicated a modest CNS response to the MWCNT exposure, aging seemed to diminish responsiveness.