Blood from COPD and non-COPD subjects were collected at the Hospital “Monaldi-Azienda Ospedaliera (AORN)-Ospedale dei Colli” in Naples, Italy, after signed informed consent. The experimental protocol was performed in accordance with the guidelines and regulations provided and accepted by the Ethical Committee of the “Monaldi-AORN-Ospedale dei Colli” Hospital (approval number 1254/2014). COPD and non-COPD subjects were 60 ± 10 (mean ± SEM) years of age. COPD subjects were smokers or former smokers; non-COPD subjects were divided in smokers and non-smokers. Blood was collected and used within 24 h.

Mononuclear cells were isolated according to Ficoll’s protocol as already reported (11). Briefly, blood (5 ml) was mixed with cell medium (5 ml) supplemented with sole antibiotics and Ficoll medium (Life Sciences, Italy). PBMC layer was collected, and platelets were separated by centrifugation at 150 g for 10 min. PBMCs were then collected in cell medium, plated (10⁵ cells/well) and treated for 1, 3, or 5 h, accordingly. Treatment was performed in duplicate for each patient. Experimental time points were repeated for each patient.

Ultrafine particles were collected from laboratory premixed flames, which were run in fuel-rich conditions feeding ethylene/air and ethylene-2,5-dimethylfuran/air mixtures with an equivalence ratio of 2.0 and divided in two parts according to their sizes: sub-10 nm particles and larger soot particles (16).

Sub-10 nm particles were mostly constituted of organic carbon (17). They were stacks of few aromatic molecules, which had high-molecular mass and were constituted by four to six fused benzene rings with a dimension of about 1.2 nm connected by chain-like bridge (van der Waals interactions).

Larger soot particles were more graphitic carbon structures (17); they had sizes ranging from 20–40 nm, typical of the primary soot particles, to 100–200 nm typical of the chain-like aggregates of the primary particles: ultraviolet–visible and infrared spectroscopy and Raman spectroscopy were used for structural analysis of the carbon-network constituting the particles (18, 19). Soot particles appeared as a network of aromatic structures with few peripheral H atoms. Elemental analysis confirmed the low presence of H atoms in the soot particles, as the amount of C was approximately 95–98%, in mass, of the total material, and H was approximately 1–2%, with the rest being trace compounds, possibly oxygen.

Sub-10 nm particles and primary soot particles and aggregates were dispersed in bidistilled water to obtain a suspension with a concentration of 5 ppm (5 µg/ml). The mass concentrations were determined assuming a density of 1.2 and 1.8 g/cm³ for sub-10 nm and soot particles, respectively. Here, we define NOC-E and NOC-ED the samples of sub-10 nm particles collected burning ethylene and the ethylene/dimethylfuran, whereas Soot-E and Soot-E/DMF the larger sizes soot particles collected burning ethylene and the ethylene/dimethylfuran, respectively. A summary of the particle characteristics is reported in Table 1.

IL-18 and IL-33 were measured in cell-free supernatants (75 × 10⁴ cells/well) using commercially available ELISA kits (eBioscience, CA, USA). No differences in cytokine release were observed according to stage of COPD patients. 8-OH-dG was measured following manufacturer’s instructions after 3 h of treatment (5 × 10⁶ cells/well) (Elabscience, USA). The release of caspase-4 was analyzed by an ELISA kit patented by ImmunePharma s.r.l. (RM2014A000080 and PCT/IB2015/051262) (Department of Pharmacy, University of Salerno, Italy). According to the patent policy, and because it is still not commercially available, it is not at the moment possible to describe the technical approach to determine the release of caspase-4 by using the ImmunePharma’s ELISA kit.

Intracellular Ca²⁺ concentrations ([Ca²⁺]i) were measured as previously reported after 1 h of treatment (5 × 10³ cells/well) (20). Data were expressed as percentage of delta increase of fluorescence ratio (F340/F380 nm) induced by ionomycin (1 µM) or carbonyl cyanide p-trifluoromethoxy-phenylhydrazone (FCCP, 0.05 µM)—basal fluorescence/basal fluorescence ratio (F340/F380 nm).

Peripheral blood mononuclear cells were stained for flow cytometry analysis (BD FACSCalibur Milan, Italy) using the following antibodies: CD14-PE and NLRP3-PeCy5.5. Healthy non-smoker-, smoker-, and COPD-derived PBMCs (2 × 10⁵ cells/well) were stained for MitoSOX Mitochondrial Superoxide Indicator as indicated in the manufacturer’s guide (Life Technologies, USA).

Total RNA was isolated from PBMCs (10⁷ cells/well) by using the RNA extraction kit (Qiagen, Milan, Italy). Reverse Transcription was performed by using first-strand cDNA synthesis kit (Qiagen, Milan, Italy) followed by PCR. Thermal cycling conditions were as follow:

5 min at 95°C, followed by 40 cycles of 30 s at 95°C, 60 s at 54°C, 30 s at 60°C for OGG1.

Primer pairs were as follow:

OGG1: Forward 5′-GACAAGACCCCATCGAATGC-3′

Data are reported as median ± interquartile range or as mean ± SEM. Statistical differences were assessed with one-way analysis of variance followed by Bonferroni’s multiple comparison post test, and p values less than 0.05 were considered significant.

To mimic environmental pollution, PBMCs were isolated from smokers, non-smokers, and COPD patients, and then treated with two classes of combustion-generated UFPs for 5 h. Figure 1 reports the release of IL-18 after the addition of Soot-E (A), Soot-E/DMF (B), NOC-E (C), and NOC-ED (D) at concentrations of 50 and 100 pg/ml. Clearly, UFPs induced the release of IL-18 from PBMCs obtained from unstable exacerbated COPD patients (black bars, Figure 1). To note, the release of IL-18 from PBMCs of unstable/exacerbated COPD patients was significantly higher than that observed from PBMCs of non-smoker (Figure 1, white bars) and smoker (Figure 1, dotted bars) volunteers. Noticeably, Soot-E (Figure 1A) and Soot-E/DMF (Figure 1B) did not induce IL-18 release from PBMCs obtained from non-smoker (Figures 1A,B, white bars) and smokers (Figures 1A,B, dotted bars), compared with the control (CTR)/basal levels.

Similarly, the same mass concentrations of NOC-E (Figures 1C) and NOC-ED (Figure 1D) increased the release of IL-18 from PBMCs obtained from unstable COPD (Figures 1C,D, black bars) and were almost non-active at inducing IL-18 release from non-smoker- derived (white bars) and smoker-derived (dotted bars) PBMCs (Figures 1C,D).

The same mass concentrations of NOC and soot particles correspond to different surface area of the particles exposed to cells due to the different sizes of the particles. In particular, 100 pg/ml of soot have a surface area of 2.2E−4 cm²/ml whereas the same amount of NOC has a surface area of 2E−3 cm²/ml, implying that NOC particles expose higher surface area than soot particles, due to their lower size. To verify, whether the size of these sub-10 nm particles was a limitation, we treated cells with coronene and pyrene, polycyclic aromatic hydrocarbons with sizes comparable to those measured for the aromatic compounds that constitute the nanoparticles (1 pg/ml up to 1 ng/ml). As shown in Figures S1A–C in Supplementary Material, both coronene and pyrene (10 pg/ml) significantly increased IL-18 release from PBMCs of non-smokers (Figure S1A in Supplementary Material), smokers (Figure S1B in Supplementary Material), and unstable COPD (Figure S1C in Supplementary Material) at 10 pg/ml corresponding to 6E−4 cm², implying that the size of the particles was not as relevant as the nature of the particles. Moreover, similarly to NOC-E and NOC-ED particles, we observed that cells were more responsive to the lower concentrations than the higher (data not shown), reaching a plateau at 10 pg/ml. Interestingly, PBMCs from non-smokers (Figure S1A in Supplementary Material) and smokers (Figure S1B in Supplementary Material) were less responsive in terms of IL-18 release compared with PBMCs obtained from unstable COPD patients (Figure S1C in Supplementary Material), which instead showed a significant increase. These data underlie that the organic nature of UFPs, as well as soot and NOC, is of relevant importance for cell responsiveness in terms of IL-18 release.

Similarly to what performed for IL-18, we treated PBMCs from the three cohorts with soot and NOC UFPs. The addition of Soot-E (Figure 2A), Soot-E/DMF (Figure 2B, black bars), NOC-E (Figure 2C, black bars), and NOC-ED (Figure 2D, black bars) significantly increased the release of IL-33 at the concentration of 50–100 pg/ml from PBMCs obtained from unstable/exacerbated COPD patients. Also in this case, we observed that IL-33 release reached a plateau at 50–100 pg/ml whereas at higher concentrations of these UFPs (>100 pg/ml up to 1 ng/ml, data not shown) no increasing effect was observed. However, it is to note that despite what happened for IL-18, no statistical increase of IL-33 release was observed comparing results among the groups represented by non-smokers (Figure 2, white bars) and smokers (Figure 2, dotted bars). Instead, the stimulation of PBMCs derived by unstable COPD patients with Soot-E (Figure 2A, black bars), Soot-ED (Figure 2B, black bars), NOC-E (Figure 2C, black bars), and NOC-ED (Figure 2D, black bars) increased the levels of IL-33. It is to note, though, that we observed a statistical difference in IL-33 basal levels that were higher in smokers than unstable COPD-derived PBMCs (Figure 2, dotted bars vs black bars).

In our previous study, we demonstrated that the release of IL-1-like cytokines after UFP exposure was caspase-1- and NLRP3 inflammasome dependent in PBMCs from healthy smokers (11). To understand the molecular mechanism underlying the release of IL-18 and IL-33 from unstable COPD patients after organic UFP exposure, we carried on evaluating the role of NLRP3 and mitochondrial-dependent oxidative stress. We observed that the expression of NLRP3 in CD14⁺ PBMCs in basal conditions was significantly higher in COPD patients (Figure 3A, black bars) than non-smokers (Figure 3A, white bars) and smokers (Figure 3A, dotted bars). Similarly, mitochondrial-derived reactive oxygen species (mtROS) (identified as % Mitosox cells) were robustly higher in PBMCs from unstable/exacerbated COPD patients than non-smokers and smokers (Figure 3B). Starting from these observations and based on the concept that NLRP3 activation is strictly dependent on mtROS production (21), we measured mitochondria homeostasis under UFP treatment. The exposure to UFPs of PBMCs from non-smokers (Figure 3C, white bars) and smokers (Figure 3C, dotted bars) increased the release of calcium (Ca⁺²) from mitochondria only after an external stimulus (carbonyl cyanide p-trifluoromethoxy-phenylhydrazone, FCCP) during the measurement/detection, implying that UFP treatment did not alter mitochondrial calcium stores. By contrast, the measurement of Ca⁺² stores in the mitochondria of unstable COPD-derived PBMCs were lower (Figure 3C, black bars) than those observed in non-smokers and smokers (Figure 3C, white and dotted bars) after Soot-E, Soot-E/DMF, NOC-E, and NOC-ED (100 pg/ml) treatment. The above data imply that the stimulation of PBMCs with these particles had already induced the release of Ca⁺² from the mitochondria in unstable COPD-derived PBMCs. Because the release of Ca⁺² from the mitochondria is strictly correlated to the release of mtROS (20) and protein/DNA damage, the levels of 8-hydroxy-deoxyguanosine (8-OH-dG), a well-known marker for DNA damage derived from oxidative stress (22), were measured. Very interestingly, the addition of NOC-E and NOC-ED significantly increased the release of 8-OH-dG in non-smoker (Figure 3D, white bars) and smokers (Figure 3D, dotted bars). However, 8-OH-dG release was more pronounced in PBMCs from unstable/exacerbated COPD patients after the addition of Soot-E/DMF, NOC-E, and NOC-ED (100 pg/ml) (Figure 3D, black bars).

To note, we did not observe a significant increase in mtROS from non-smokers after the addition of all particle samples (Soot-E, Soot-E/DMF, NOC-E, and NOC-ED, each 100 pg/ml) compared with smokers and COPD PBMCs (data not shown). Therefore, we went on to analyze the levels of a repairing enzyme, 8-oxoguanine glycosylase, OGG1, highly important to avoid DNA damage following oxidative stress (23). The administration of all UFPs (100 pg/ml) significantly increased the levels of mRNA in PBMCs obtained from non-smokers (Figure 3E, white bars). By contrast, PBMCs from smokers did not show any increase compared with the basal levels of mRNA for OGG1 after all UFP (100 pg/ml) exposure (Figure 3E, dotted bars), but rather, a significant decrease was observed. Interestingly, PBMCs from unstable COPD patients showed no increase of OGG1 mRNA levels after Soot-E, Soot-E/DMF, NOC-E, and NOC-ED (100 pg/ml) exposure (Figure 3E, black bars).

These data imply that treatment of PBMCs obtained from unstable/exacerbated COPD patients with Soot-E, Soot-E/DMF, NOC-E, and NOC-ED leads to mitochondrial dysfunction in that oxidative stress occurs. In this context, however, this effect is not countered by repairing enzymes, such as OGG1, establishing a pro-inflammatory/prooxidative cytoplasmic molecular pattern that leads to the release of IL-18 and IL-33 from PBMCs obtained from unstable/exacerbated COPD patients.

Because the release of mtROS and the presence of 8-OH-dG are able to induce the activation of NLRP3 inflammasome and because we found that NLRP3 protein levels are higher in unstable COPD patients, we went on by analyzing the molecular mechanism underlying the release of IL-18 and IL-33 from unstable COPD-derived PBMCs after combustion-generated UFP exposure. Therefore, unstable COPD-derived PBMCs were treated with pharmacological inhibitors for caspase-1, ac-Y-Vad (Y-Vad, 1 µg/ml) (20), for NLRP3, glybenclamide (Gly, 1 µM) (11, 20) and for caspase-8, z-IETD-fmk (IE, 0.5 µg/ml) (24, 25) together with combustion-generated UFPs.

The pharmacological inhibition of caspase-1 did not alter the levels of IL-18 from unstable COPD-derived PBMCs after Soot-E (Figure 4A), Soot-E/DMF (Figure 4B), NOC-E (Figure 4C), or NOC-ED (Figure 4D) exposure. In support, the inhibition of NLRP3 by means of Gly did not affect the levels of IL-18 from unstable COPD-derived PBMCs after Soot-E (Figure 4A), Soot-E/DMF (Figure 4B), NOC-E (Figure 4C), or NOC-ED (Figure 4D) exposure. Moreover, the pharmacological inhibition of caspase-8, another enzyme involved in the non-canonical inflammasome-dependent pathway (26), did not alter IL-18 levels (Figures 4A–D).

The same results were obtained for IL-33, where we observed that neither the inhibition of caspase-1 (Figures 5A–D) nor the inhibition of NLRP3 and caspase-8 (Figures 5A–D) altered the levels of the cytokine released from PBMCs obtained from unstable COPD patients exposed to combustion-generated UFPs.

Non-canonical inflammasome was also described as caspase-4-dependent in humans (caspase-11 in mice) (26). However, caspase-11 was defined as inducible, not constitutive, in mice (27). Therefore, we first analyzed the levels of mRNA for caspase-4 in PBMCs obtained from unstable COPD patients.

We found that caspase-4 mRNA levels were detectable in CTR cells and, interestingly, the treatment of cells with NOC-ED significantly increased caspase-4 mRNA levels (Figure 6A). To evaluate the potential involvement of caspase-4 in UFP effects, we used a patented ELISA kit (as described in Section “Materials and Methods”). Very interestingly, the addition of NOC-ED significantly increased the release of caspase-4 from PBMCs obtained by unstable COPD patients (Figure 6B). However, we were not able to reach a statistical difference when cells were treated with the other three UPFs samples, Soot-E or Soot-E/DMF or NOC-E (Figure 6B). Nevertheless, a potential increase was observed: 0.61 ± 0.32 ng/ml for Soot-E (100 pg/ml) vs 0.35 ± 0.09 ng/ml for CTR; 1.36 ± 0.74 ng/ml for NOC-E (100 pg/ml) vs 0.35 ± 0.09 ng/ml for CTR. Similar data were observed when COPD-derived PBMCs were treated with coronene (10 pg/ml; 1.74 ± 0.7 ng/ml).

Taken together, these data imply the extracellular release of caspase-4 and its involvement in IL-18 and IL-33 increase when PBMCs obtained from unstable COPD patients are treated with UFPs.