In order to develop more targeted abatement strategies it is important to investigate whether particle-induced health effects can be attributed to specific sources or to compounds present in the complex PM mixture. The aim of this study was to compare cellular responses to WSPs from different phases of the combustion cycle and from different combustion conditions. In addition, the effects induced by the WSPs were compared to those induced by a reference sample from traffic. WSPs from different combustion phases did not differ in their potency to induce cytokine release, cell death or reductions in cell numbers. However, particles collected during medium-temperature combustion were more toxic than the reference particles collected from traffic or during a higher wood-combustion temperature. WSPs collected during medium-temperature combustion were also more potent inducers of IL-6 than WSPs collected during high-temperature incomplete combustion.
Differences in chemistry
Since the chemical composition of WSPs is known to depend on the combustion process and temperature [17, 19, 49], we expected the three WSPs from different combustion phases to differ in chemical composition. Indeed, the PAH-content of particles from the ‘mixed-smoke’ and ‘start-up’ sessions was approximately two times higher than in particles collected during the ‘burn-out’ session, but the levels of refractory elements were similar. In comparison to the differences between WSPs from medium- and high-temperature combustion, with the content of PAHs and refractory elements being five and 3300 times higher in Wood(high-temp), the differences between the medium-temperature WSPs must be considered as minor. A possible explanation for the high similarity between the WSPs from medium temperature combustion could be that the combustion conditions during collection of particles from the different phases were quite similar. The total emissions of hydrocarbons have been reported to be high in the start-up and flaming phases of the combustion cycle [17, 18], but to our knowledge the variation in PAH content with the combustion cycle has not yet been fully described. Although the PAH content in the samples from the different combustion phases was relatively similar, there could be larger differences in the content of other organic compounds, such as oxy-PAHs, ketones and quinones, that was not reflected in the total PAH content. In accordance with the similar chemical composition of the medium-temperature WSPs, these samples also induced similar biological effects.
The differences in chemical composition between the high-and medium-temperature wood smoke samples reflect differences in combustion conditions. The 3300 times higher level of refractory elements in the high-temperature compared to the medium-temperature wood smoke sample is in agreement with the potassium K-shell absorption edge observed only in the Wood(high-temp) NEXAFS spectrum. This difference is also in accordance with the higher levels of refractory elements reported for increasing combustion temperatures in the literature [19, 50, 51]. Similarly, the five times higher PAH content in the Wood(high-temp) sample is in accordance with the NEXAFS analysis showing a higher C = C peak in Wood(high-temp) than Wood(mixed smoke). In agreement with our findings, PAH formation has been reported to increase with increasing combustion temperatures, with optimal temperatures for PAH formation between 700 and 900°C [18, 51–55]. At even higher combustion temperatures, the PAHs are thermally decomposed, resulting in decreasing PAH levels during more complete combustion conditions [52, 56–58]. However, other factors also influence the PAH emissions, such as oxygen supply, moisture content and residence time in the combustion zone [52, 53, 55], and if available these factors should be considered in the interpretation of PAH data. Although a range of studies report lower PAH levels for poor combustion conditions in line with the present findings, two recent studies report very high levels of PAHs in emissions from poor combustion conditions [20, 59]. These contradicting data emphasize the importance of sufficient characterisation, not only of the emitted particulate matter applied in the toxicological studies, but also of factors determining the combustion conditions. An improved characterisation of factors determining the combustion conditions would have been feasible in the present study.
Influence of combustion conditions
Presently, particles from incomplete combustion with medium-temperature were more potent inducers of cytotoxicity and IL-6 release than particles originating from incomplete combustion with higher temperature. In line with this, particles from poor combustion conditions have been reported to be more toxic in macrophage and fibroblast cell lines than particles from more complete combustion conditions [20, 22, 60]. With respect to inflammatory potential, particles from smouldering combustion were more potent inducers of macrophage inflammatory protein (MIP)-2, the murine analogue of IL-8, than particles from normal flaming combustion, whereas the release of TNF-α was similar for the two combustion conditions, and IL-6 could not be detected . Similarly, intratracheal instillation of WSPs from poor combustion conditions increased the expression of MIP-2, monocyte chemotactic protein (MCP)-1 and heme oxygenase (HO)-1 mRNA in rat lung, whereas particles from more complete combustion conditions had no effect . These particles from poor combustion conditions also induced the highest expression of inflammatory markers in a parallel cell culture study in two human cell lines . Thus, the present finding of medium-temperature WSPs being more potent inducers of IL-6 than high-temperature particles appears to be in accordance with the previously published literature. It should however be kept in mind that the physicochemical properties vary with the combustion conditions, and based on the limited chemical analyses performed in the various toxicological studies it is difficult to compare the applied combustion conditions.
Although the present as well as previous findings suggest that both the cytotoxic and inflammatory effects of WSPs increase with decreasing combustion temperatures, it is necessary to perform more extensive toxicological studies to verify this hypothesis. Preferentially, the effects of well-characterised particles from a wider range of combustion conditions should be investigated simultaneously in the same biological model system. A recent study compared emissions from two different combustion conditions and related their in vitro toxicity data to the respective emission factors (mg PM1/MJ fuel energy) . The particles had a similar inflammatory potential if compared on a mass basis, but after adjustment for the emission factors the particles from low-temperature (smouldering combustion) induced 5–20 fold higher responses compared to particles from more complete combustion conditions. This illustrates the importance of considering the exposure levels in addition to the relative toxicity on a mass basis. However, although emissions from poor combustion conditions (smouldering) are known to be much higher than emissions from more complete combustion conditions (flaming), little is known about the fate of these particles in the atmosphere and the human exposure . Moreover, the relative pulmonary deposition of these two classes of particles is not known, and these gaps in knowledge preclude a full evaluation of the relative toxicity of WSPs from varying combustion conditions.
Importance of the organic fraction
Comparison of the effects induced by native WSPs and the corresponding organic extracts and washed particles, indicated that the organic fraction was of major importance for the biological effects induced by WSPs from the different combustion phases. The washed particles did however also increase the release of IL-6 and IL-8 significantly. This could be due to an incomplete removal of the organic compounds, effects induced by the insoluble carbon core, or possibly by adsorbed metals. The organic extracts of the reference samples Traffic and Wood(high-temp) were investigated in the same series of experiments, but the data were published previously . In line with the present results, the organic fraction of Wood(high-temp) accounted for the majority of the cytokine release, as well as the reduction in cell number. This may suggest that the organic fraction is of importance regardless of the combustion conditions used to generate the wood smoke. The toxicity of WSPs from varying combustion conditions has previously been linked to the condensable organic matter present in the smoke. In that study, the condensable organic matter had similar toxicity when compared on an equal mass basis, but the amount of condensable organic matter emitted was 200 times higher during poor combustion conditions as compared to complete combustion conditions . The amount of reactive oxygen species (ROS) present in wood smoke emissions has also been found to increase with decreasing combustion efficiency, and the amount of ROS was linked to the content of organic carbon in the particles . Although the ROS levels present in PM may not correlate with cellular effects , these data support that the amount of reactive chemicals in wood smoke emissions increase with decreasing combustion efficiency. Moreover, WSPs from low combustion temperature were recently found to induce higher levels of cellular ROS in monocytic and epithelial cell lines than particles from higher combustion temperatures .
Although the organic fraction was found to influence the release of inflammatory mediators, toxicity and cell numbers, the total PAH content was not associated with any of the biological effects in the linear regression analysis. This may point towards a limited importance of the PAHs in the inflammatory and toxic effects. Cellular studies of individual PAHs report effects on cytokine release, cytotoxicity and DNA damage [64–66]. This may suggest that these compounds contribute to the observed cellular effects after all, although the concentrations used in those studies were 10- to 1000-fold higher than the PAH concentrations during PM exposure in the present study. When linear regression was performed for the individual PAHs, these were however associated with negative slopes for IL-6 and IL-8, suggesting a decreased mediator release for increased levels of PAHs. Interestingly, PAHs have been negatively associated with release of pro-inflammatory mediators both in macrophages and mouse lung in studies using urban air particles, and the authors suggested that PAHs were associated with an immunosuppressive effect [67, 68]. Overall, the R2 values were relatively low in the regression analysis, and it is necessary to do further experiments to investigate the role of these single PAHs in wood smoke induced effects. Generally, the present data and data from the literature [20–22] show that the in vitro toxicity and inflammatory potential of the particles increases with decreasing combustion temperatures, and none of these studies show a positive correlation between biological effects and PAH content. Thus, other organic compounds than PAHs seem to also be involved in the wood smoke-induced effects. Future studies of the toxicological effects of wood smoke should therefore include analysis of other organic compounds than the traditionally analysed unsubstituted PAHs like EPA PAH16.
To further characterise the organic fraction of the WSPs applied in the present study, two samples were analysed by NEXAFS. The results suggested a higher content of quinone-like compounds in the medium-temperature than the high-temperature wood smoke sample. Since the medium-temperature particles also induced the highest effects on release of IL-6 and toxicity, the content of quinone-like compounds appears to co-vary with the inflammatory and toxic effects. Interestingly, a study that applied fractionation of organic extracts of WSPs into different polarity extracts suggested that oxy-PAHs and quinones contributed to cellular oxidative stress to a larger extent that the unsubstituted PAHs . Moreover, different quinones, including naphthoquinone and phenanthraquinone have been reported to induce inflammatory effects in vivo and in vitro, including recruitment of inflammatory cells to the lung and expression of pro-inflammatory mediators [70–72]. Based on the present findings and the literature, quinones may emerge as a group of organic compounds potentially involved in the biological effects of WSPs. However, further studies are necessary to confirm this hypothesis including fractionation studies for a wider range of WSPs and biological endpoints. These methods should also be combined with further chemical characterisation studies in order to search for other groups of organic compounds that might be involved in WSP toxicity.
The reference WSPs from high-temperature incomplete combustion have previously been associated with reduced proliferation in this co-culture . Actually, the previous series of experiments showed that the cell numbers did not increase at all over time, and in combination with a lack of necrosis and apoptosis, these data suggested a complete stop in the proliferation. It could therefore be speculated that the further reduction in cell numbers induced by the medium-temperature combustion particles in the present study was due to cytotoxicity. This hypothesis is also in line with the increased release of LDH induced by all the medium-temperature WSPs, but not the high-temperature sample. Also, ambient PM and cigarette smoke have been reported to reduce cell proliferation [73, 74]. In the present study, we observed an accumulation in cells in the S/G2 phase in THP-1 monocytes that could contribute to the reduced proliferation. The most severe effects on the cell cycle were observed in the A549 pneumocytes, but these histograms could not be analysed due to the severe artefact effects introduced by the organic extracts. In line with the present results, Wood(high-temp) was recently found to induce accumulation of bronchial epithelial cells in the S-phase , and other types of PM have been reported to induce cell cycle effects like G1 arrest or a delay in G2 or mitosis [73, 75]. In a previous study, the DNA damage induced by Wood(high-temp) in mono-cultures of THP-1 and A549 was determined by the comet assay. In comparison to PM from other sources, both Wood(high-temp) and its organic extract were potent inducers of DNA damage, and WSPs also induced FPG sites, suggesting oxidative DNA damage . DNA damage may cause cell cycle delays or arrests, since damaged DNA may stop the cells from passing through various checkpoints in the cell cycle . Thus, oxidative DNA damage induced by organic compounds in the WSPs is a possible mechanism for the cell cycle effects observed in the present study.
Fine vs. coarse particle size fractions
Presently, the PM0.1–2.5 and PM2.5–10 fractions induced very similar biological effects, and few statistically significant differences were detected. This is in contrast to in vitro studies of other particle samples in which the PM2.5–10 fractions seem to be more potent than the PM2.5 fractions in inducing a pro-inflammatory response [35, 36]. Possible explanations include a higher content of microbial components like endotoxin in the coarse fraction, and also that different sources contribute to the two size fraction causing differences in physicochemical properties [78, 79]. A possible explanation for the similar cellular responses induced by the fine and coarse particle fractions presently might be that these two fractions of wood smoke have similar physicochemical properties. During collection of the applied WSP samples the PM0.1–2.5 fraction accounted for 79–86% of the total PM10 emissions (data not shown), which is in agreement with the PM1 to PM10 ratio of 0.8 reported for the PM mass emissions from both smouldering and flaming combustion conditions . Since emissions from residential wood combustion are dominated by the PM2.5 fraction both with respect to number and mass concentrations , characterisation of the toxicity of the PM2.5 fraction rather than the PM2.5–10 fraction appears to be feasible in future toxicological studies.
Wood smoke vs. traffic
We have previously compared in vitro effects of particles collected from wood smoke and traffic and concluded that traffic-derived particles were more potent in inducing inflammatory responses, while WSPs were more potent in reducing cellular proliferation . In the present study, the pro-inflammatory potential of traffic-derived particles was not consistently higher, since wood smoke and traffic particles induced a similar release of inflammatory mediators. Overall, our present and previous findings suggest that wood smoke and traffic particles have a similar pro-inflammatory potential in vitro, but that WSPs seem more cytotoxic and more potent in reducing proliferation than traffic derived particles. However, the pulmonary effects depend on the deposited dose, and WSPs from poor combustion conditions were recently found to deposit to a lower extent in the lungs than diesel exhaust particles, providing different pulmonary doses . To compare the effects of particles from the two sources it is therefore necessary to perform inhalation studies accounting for possible differences in pulmonary deposition. In the published inhalation studies in animals and humans, the authors report that PM from diesel exhaust and wood smoke induce effects of similar magnitude [10, 11, 81, 82], whereas a more recent study suggests that WSPs may be less potent . It should however be kept in mind that the large variation in the physicochemical properties of particles emitted from each source precludes the comparison of the biological effects, since the differences reported will depend strongly on the composition of the particles included in the study.
The applied particle concentration of 40 μg/cm2 is likely to be higher than the average deposition on the lung surface during normal ambient concentrations. Particle deposition in the human airway is highly uneven, but the retention for fine particles has been suggested to be highest in the proximal alveolar region [30, 84, 85]. The slow particle clearance in the alveolar region, with biological half-lives up to 120 days, also contributes to increased particle exposure . The applied particle concentrations might be relevant for specific regions of the lung during long time exposure to high concentrations of air pollution. The single particle concentration applied presently was chosen based on results obtained in a previous study in which the effects of lower concentrations of particles with similar physicochemical properties did not cause a significant increase in cytokine release in the same co-culture model. That study showed a monotonous increase in cytokine release with increasing particle concentrations. Thus, although inclusion of a concentration-dependent response to the present particle samples would have provided a more robust data set, inclusion of more, lower, particle concentrations would most likely have given the same overall picture of the inflammatory potential of the analysed samples. Previously, the cytokine release induced by Wood(high-temp) was found to be attenuated for increasing exposure times, therefore inclusion of two time-points was considered more important in the present study than inclusion of several different particle concentrations. For risk assessment purposes a more complete evaluation of the relative toxicity of WSPs is necessary, both due to the limitation of applying only one particle concentration, but also due to the limitations inherent in in vitro studies, including altered cellular characteristics of cell lines and limited communication with other cells compared to cells in tissue.