Traffic-related ultrafine particles impair mitochondrial functions in human olfactory mucosa cells - Implications for Alzheimer's disease

Abstract

Constituents of air pollution, the ultrafine particles (UFP) with a diameter of ≤0.1 μm, are considerably related to traffic emissions. Several studies link air pollution to Alzheimer's disease (AD), yet the exact relationship between the two remains poorly understood. Mitochondria are known targets of environmental toxicants, and their dysfunction is associated with neurodegenerative diseases. The olfactory mucosa (OM), located at the rooftop of the nasal cavity, is directly exposed to the environment and in contact with the brain. Mounting evidence suggests that the UFPs can impact the brain directly through the olfactory tract. By using primary human OM cultures established from nasal biopsies of cognitively healthy controls and individuals diagnosed with AD, we aimed to decipher the effects of traffic-related UFPs on mitochondria. The UFP samples were collected from the exhausts of a modern heavy-duty diesel engine (HDE) without aftertreatment systems, run with renewable diesel (A0) and petroleum diesel (A20), and from an engine of a 2019 model diesel passenger car (DI-E6d) equipped with state-of-the-art aftertreatment devices and run with renewable diesel (Euro6). OM cells were exposed to three different UFPs for 24-h and 72-h, after which cellular processes were assessed on the functional and transcriptomic levels. Our results show that UFPs impair mitochondrial functions in primary human OM cells by hampering oxidative phosphorylation (OXPHOS) and redox balance, and the responses of AD cells differ from cognitively healthy controls. RNA-Seq and IPA® revealed inhibition of OXPHOS and mitochondrial dysfunction in response to UFPs A0 and A20. Functional validation confirmed that A0 and A20 impair cellular respiration, decrease ATP levels, and disturb redox balance by altering NAD and glutathione metabolism, leading to increased ROS and oxidative stress. RNA-Seq and functional assessment revealed the presence of AD-related alterations in human OM cells and that different fuels and engine technologies elicit differential effects.

Keywords: Mitochondrial dysfunction; Olfactory mucosa (OM); Oxidative phosphorylation (OXPHOS); RNA sequencing (RNA-Seq); Redox balance; Ultrafine particles (UFP).

Graphical abstract

Fig. 1

Assessment of cellular viability. (A) Illustration of experimental workflow from receiving the OM biopsy, culturing and exposing cells to UFPs, and assessing cellular viability by MTT and LDH assays. (B, C) MTT reduction assay. OM cells from n = 4 individuals with diagnosed AD were exposed to different UFPs with a concentration of 20 l/ml. Results are presented as a fold change (FC) to vehicle-treated cells, with each cell line compared to its corresponding vehicle. (D, E) LDH release assay from collected cell culture medium after 24-h and 72-h exposure to different UFPs at a concentration of 20 l/ml, AD OM cells (n = 4). Results were calculated as a percentage from lysed cell control (TXT) = absolute death, and presented as an FC to vehicle-treated cells, each cell line compared to its corresponding vehicle. One-way ANOVA, Dunnett's multiple comparisons test * = p < 0.05; *** = p < 0.001.

Fig. 2

Transcriptomic assessment of 24-h UFP-exposed AD OM cells. (A) Venn diagram presents the overlap and individual DEGs in OM cells exposed to A0 and A20. (B) Illustration of the workflow from exposing OM cells to traffic-derived UFPs, extracting RNA, generation of cDNA-libraries, and performing RNA-Seq and data analysis. (C) Barplot showing the z-scores of the most perturbed Ingenuity Pathways enriched upon exposure to A0 and A20. (D) Stacked violin plot presenting the expression values distribution of a subset of DEGs from OXPHOS-pathway in AD and cognitively healthy control samples exposed to A0, A20, and vehicle. In each violin plot the black dot represents the mean expression value and the asterisk indicates significant (p < 0.05) difference compared to the vehicle.

Fig. 3

UFPs reach mitochondria and impair OXPHOS. (A)–(G) Seahorse Mitostress Test assay to assess mitochondrial functions in OM cells from cognitively healthy controls and ADs exposed to vehicle, Euro6, A0, and A20 for 24-h with a concentration of 20 l/ml and quantified in pmol/minute/μg protein. (A) Line plot with one cognitively healthy control and one AD OM cell line displaying oxygen consumption rate (OCR) at baseline followed by the administration of an ATP synthase inhibitor oligomycin at 20 min, the uncoupler FCCP at 50 min, and a mix of complex I and complex III inhibitors (rotenone and antimycin A at 80 min. All following respiratory parameters (B)–(G) present OM cells from cognitively healthy controls and ADs (n = 3/group) (B) Basal respiration: the last OCR measurement before injection of oligomycin-non-mitochondrial oxygen consumption (C) Maximal Respiration: the maximum OCR that can be achieved after the addition of FCCP (D) Proton Leak: OCR after the addition of oligomycin, indicating remaining basal respiration not coupled to ATP production (E) Spare respiratory capacity as %: maximal OCR – baseline OCR * 100 (F) Coupling Efficiency%: (baseline OCR - OCR after the addition of oligomycin)/baseline OCR * 100 %. (G) Non-mitochondrial oxygen consumption: Minimum rate of OCR after adding rotenone & antimycin A. All exposures are compared to the vehicle-treated cells inside the health status group with One-way ANOVA, Dunnett's multiple comparisons test * = p < 0.05; ** = p < 0.01; *** = p < 0.001. (H) Illustration of how UFPs reach even the mitochondrion and cause multiple effects. (I) ATP concentrations in OM cells after 24-h exposure to the vehicle, Euro6, A0, and A20, presented as μM. Each dot represents cell line n = 6 controls/ADs, calculated as mean from 4 technical replicates. One way ANOVA, Dunnett's multiple comparisons test ** = p < 0.01; *** = p < 0.001.

Fig. 4

(A) Illustration of the ETC with NADH-molecules as electron donors and releasing protons into the intermembrane space, finally resulting in pumping them back to the mitochondrial matrix while generating ATP (B) Amount of oxidized form (NAD+) and (C) reduced form (NADH) of nicotinamide adenine dinucleotide in cognitively healthy control/AD cells exposed to vehicle and A20 presented as pmol/mg of protein, Student's t-test, * = p < 0.05; ** = p < 0.01. (D) NAD+/NADH-ratio in cognitively healthy control/AD cells exposed to vehicle, Student's t-test # = p < 0.05. (E) NAD+/NADH-ratio presented as a grouped line plot illustrating the statistically significant effect of the exposure (# = p < 0.05), and the interaction effect of exposure and health status (* = p < 0.05) two-way ANOVA. In addition, there is a significant difference between AD veh – AD A20 (p < 0.05), two-way ANOVA followed by Sidak's multiple comparison test. (F) Total cellular ROS measured from the OM cells from cognitively healthy controls and ADs (n = 3/group) exposed to veh, Euro6, A0, and A20 for 24-h with a concentration of 20 l/ml. Results were calculated as AUC from fluorescence intensity and presented as a percentage from positive cell control (500 μM H₂O₂ – treated) for absolute ROS formation. Asterisks indicate the difference compared to vehicle. One-way ANOVA, Dunnett's multiple comparisons test ** = p < 0.01; *** = p < 0.001. The group effect assessed with two-way ANOVA did not quite reach statistical significance (p < 0.0646). (G) Violin plot presenting DEGs found in the transcriptomic analysis involved in glutathione metabolism, an asterisk indicates a significant (p < 0.05) difference compared to vehicle. (H) GSH in cognitively healthy controls and ADs exposed to vehicle and A20, presented as nmol/mg protein. Two-way ANOVA, significant exposure effect * = p < 0.05. (I) GSSG in cognitively healthy control and AD cells exposed to vehicle and A20, presented as nmol/mg protein. (J) GSH/GSSG ratio of cognitively healthy control and AD cells exposed to vehicle and A20. (K) GSH pool in cognitively healthy controls and ADs exposed to vehicle and A20, presented as nmol/mg protein. Significant health status effect, but exposure effect did not reach statistical significance (p < 0.0941), two-way ANOVA ## = p < 0.01.