In the present study we showed that 7.5 μg/cm2 of a well-characterized urban fine PM (Milan winter PM2.5) caused alterations in different phases of the cell cycle, resulting in apoptotic cell death, tetraploid G1 cells (binucleated) and cells with MN.
PM exposure has previously been reported to result in an accumulation of cells at various cell cycle phases [20, 22, 23]. Besides PM characteristics and dose, time of analysis and the specific cell line used may also influence the results obtained [23, 33]. We have previously reported that 25 μg/cm2 of Milan winter PM2.5 induced mitotic arrest in BEAS-2B cells after 20 h of exposure which later resulted in mitotic cell death . Here we investigated the in vitro effects of a PM-dose which is among the lowest reported in literature to give biological effects, in an effort to approach environmental human exposure levels. Using this dose, the various phases of the cell cycle were differently affected and little mitotic apoptosis was observed. As results on cell cycle distribution are highly dependent on the time of the analysis, the cell cycle progression has been followed at different time points. A significant increase of cells in G2/M phases already occurred after 3 h of exposure. The G2/M increase was sustained up to 24 h, but it consisted of alterations at three different phases of the cell cycle progression. The combined use of flow cytometry and fluorescence microscopy revealed an early (after 3 h) delay in the G2 phase. This was followed by an increased number of cells in mitosis (after 10 h). Finally, cytokinesis was affected, because an increased number of non-mitotic tetraploid (4 N) G1 cells was seen after 24 h. The increase of cells in the subG1 region suggests that part of the cells affected by PM treatment die through apoptosis at 40 h.
The cell cycle delay has often been linked to DNA damage and the DNA damage response [20, 23, 34]. The G2/M transition checkpoint is a non-genomic and rapid-response system activated by DNA damage response . The rapid G2 block is primarily induced in a transient mode and requires p53 transcriptional activity to ultimately produce a sustained block [24, 27]. Transient or sustained by p53, the checkpoint protein kinase Chk2 is a pivotal messenger of this system. In the present study we observed a significant increase in the level of the active phosphorylated form of Chk2 (pChk2) in cells treated with winter PM2.5 for 3 h, which is in line with the accumulation in G2 phase reported. The levels of pChk2 decrease to control values after 10 h of exposure, suggesting that the cells have overcome the G2 arrest and have entered mitosis. Accordingly, the levels of p53 and pp53 appear not to be affected by PM treatment at 3 and 10 h; these data confirm that cells exposed to PM were arrested transiently in G2 by a p53-independent pathway at 3 h of exposure and then escape from G2 into mitosis after 10 h.
When studying DNA damage and DNA damage responses in vitro it is essential to avoid cell lines with TP53 mutations, as the loss of p53 activity is linked to defects in cell cycle control and apoptosis after DNA damage . Here we used BEAS-2B cells, which are reported to have normal p53 activity, and for this reason have been widely used to study cell cycle alterations  and mechanisms involved in PM-induced toxicity [37, 38]. Nevertheless, it should be noted that this cell line is SV 40-transformed, thus these effects should be further explored in primary human lung epithelial cells and/or in vivo.
The alterations of the cell cycle may not only depend on DNA damage but also on damages to other macromolecules, as well as on changes in protein phosphorylation and ion concentrations . As shown in the present study, the various cell cycle steps affected in PM2.5-exposed cells suggest that several types of initial damage might be involved. The mitotic arrest was characterized by disequilibrium in the different mitotic phases (higher incidence of pro- and metaphase cells versus ana- and telophase ones) suggesting possible structural dysfunctions of microtubules (MT) and of mitotic spindle assembly. Furthermore, mitotic cells presented various aberrations of the mitotic apparatus, including tripolar, multipolar and incomplete spindles. Moreover, γ-tubulin staining showed centrosomes amplification. Similar spindle aberrations have been reported in Chinese hamster fibroblasts after exposure to PM10  and in our previous study, where preliminary results showed the presence of tripolar cells . These findings indicate that PM may act as spindle poison, directly perturbing microtubules dynamics, and suggest the activation of the spindle assembly checkpoint (SAC) as a mechanism for the M/A delay. Indeed, centrosomes amplification and increased number of spindle poles are known to cause a delay in the anaphase onset through SAC activation . Further, SAC can also be activated by the presence of incomplete bipolar spindles with lagging chromosomes, similar to the ones we found. Pole-associated chromosomes are a regular transient feature of astral spindle assembly, when an initial monotelic attachment brings the chromosomes towards the centrosomes. Under normal conditions this feature should be rapidly corrected by an Aurora-B kinase-based mechanism . The presence of a high percentage of cells with pole-associated chromosomes (10% in PM-treated samples) suggests a delay in the rearrangement of this attachment.
After exposure to PM for 24 h the number of cells was slightly reduced relative to controls, without significant levels of mitotic-apoptosis. However, an increased number of non-mitotic cells with double amount of DNA (4 N), large or double nuclei, and cells with micronuclei (MN) were present, suggesting that cells, when arrested in mitosis, did not always complete cytokinesis. It is well known that cells arrested by SAC at the M/A transition point can exit mitosis without proper segregation of chromosomes and cytokinesis, if the damages are not properly corrected within a certain period of time. This process (called mitotic slippage) gives rise to cells with large or double nuclei (4 N, G1) and with multiple micronuclei , as we found. In agreement with the literature , cells with amplified centrosomes, forming tripolar mitotic spindles, seemed to go through karyokinesis, as tripolar cells in anaphase and telophase were frequently observed. These cells might contribute to the increased subG1 peak reported after 40 h of exposure, which can be only partly explained by the increase of apoptosis observed at this time point. In contrast, cells with more than three poles were never found in anaphase and telophase, suggesting that they failed the cytokinesis, resulting in binucleated or micronucleated cells.
Cells exposed for 24 h to PM also presented high levels of cyclin B protein. This further supports the hypothesis of SAC activation, as SAC inhibits the anaphase-promoting complex (APC)-dependent degradation of cyclin B. Moreover it has been demonstrated that cyclin B degradation not only is required for the transition to anaphase, but also for the onset of cytokinesis in Drosophila
. Interestingly, Burns et al.  found high levels of cyclin B1 in 4 N cells treated with nocodazole and paclitaxel. On the other hand, Brito and Rieder  reported that cyclin B degradation is required for mitotic slippage; thus the role of cyclin B in this event is still a matter of debate.
The results obtained from the various PM fractions (organic fraction versus inorganic and carbonaceous particles) showed that the organic components of Milan winter PM2.5 are very important for the effects on the cell cycle, as particles deprived of these compounds were ineffective. This observation is in line with previous results showing that Milan summer PM2.5, with low quantity of PAHs, had no effect on the mitotic progression . Accordingly, other data in the literature [45–47] describe the role of PM organic compounds in inducing toxicity. In most of these studies [48, 49], the high PAHs content has been associated with high genotoxicity, oxidative stress, and mitochondrial and cytoskeletal dysfunctions. Möller and colleagues  reported effects on phagocytosis, phagosome transport mechanisms and cytoskeletal integrity. PAHs-rich PM0.2, produced by combustion of solid fuels, induced G2/M arrest in macrophages , while organic extracts from PM2.5 and PM10 arrested the cell cycle of different human cell lines in G0/G1 [22, 51]. Several PAHs are able to alter the cell cycle in various ways; dibenzo[a,l]pyrene induces G2/M arrest in human mammary carcinoma MCF-7 cells , while it delays HEL fibroblasts in the S phase . Similarly, exposure to BaP leads to S phase accumulation in human hepatocarcinoma HepG2 and MCF-7 cells . Moreover, recent results have shown that the cell cycle status can impact on BaP metabolism and DNA damage . Thus, how PAHs adsorbed on PM may affect the cell cycle depends on the specific compounds present and the cells’ metabolic capacity. The compounds’ bioavailability is also of importance, which was demonstrated in the present study by the higher potential of the PM organic fraction in comparison with the whole-PM to induce ROS formation. On the other hand, the whole-PM longer sustained the cellular arrest in G2/M when compared to the organic fraction, and induced oxidative DNA damage. Thus, the localization of PAHs on the particles is probably of importance for some of the PM-induced effects. However, a role for other components cannot be excluded. These could be some metals in the water soluble PM fractions, which have been shown to alter mitosis progression [56, 57].
The organic fraction seemed to be responsible for the increase of ROS observed at 2 h of exposure. ROS formation after PM exposure is associated with significant cell effects such as mitochondrial damage, increased production of cytokines and chemokines, as well as DNA damage [2, 58, 59]. Moreover, high levels of oxidants determine perturbation of the mitochondrial permeability and a disruption of electron transfer chain resulting in cellular apoptosis or necrosis . Mitochondria have been indicated as the main source of ROS generation in rat alveolar type II and human lung adenocarcinoma A549 cells exposed to a high dose of PM2.5 (50 μg/cm2) . However in this study, after exposure to 7.5 μg/cm2, only 40-50% of total ROS were localized at the mitochondria, while the rest of ROS were located in the cytoplasm. Moreover, the absence of mitochondrial superoxide formation indicated that mitochondria are not significantly involved in ROS production at 2 h. Considering these results, it is likely that the organic fraction is responsible for PM-induced ROS through P450-mediated metabolic activation of various PAHs and oxo-PAHs. The co-localization of ROS signal and mitochondria might be due to CYP enzymes, which have been recently reported to have also mitochondrial localization . Still, the contribution of other pathways (such as AKR or NADPH oxidase) cannot be excluded  and should be further investigated.
As mitochondrial superoxide formation was found at 24 h, this effect is likely secondary to ROS formation, and may be caused by the observed mitochondrial damage.
The results in this study show that PM was able to induce DNA damage as determined by comet assay, measuring strand breaks and alkali-labile sites. The AhR-response has previously been found to be of major importance in explaining the toxicity of various PM [21, 62, 63] and of its organic fraction . In accordance with this, antioxidants NAC and Thio, and the AhR/CYP enzymes inhibitor α-NF reduced the PM-induced DNA damage, as well as the G2 increase occurring at 3 h of exposure. These findings suggest that these effects were related to ROS and/or other reactive metabolites formed by AhR/CYP enzymes.
ROS-induced DNA damage includes various oxidative DNA base modifications as well as single and double strand breaks (SSBs and DSBs) [65, 66], while the reactive PAHs intermediates might also induce bulky DNA adducts [62, 67]. A further characterization of PM-induced DNA damage by 32P-postlabelling showed that the PM organic fraction induced higher bulky DNA adduct levels after 24 h of exposure, while no difference was seen after 3 h. Similar results following PM exposure have been reported by others [15, 62]. PAHs which form DNA adducts often require a two-steps activation , which might undergo competitive inhibition by non-genotoxic PAHs present in the PM complex mixture . Thus, the primary DNA damage detected by the comet assay might be those induced by organics and PAHs needing only one-step activation, such as nitro- and oxo-PAH.
Although the comet assay with Fpg was negative, the levels of 8-oxodG and γH2AX measured by immunostaining increased after 3 h of PM exposure, suggesting the presence of oxidative DNA damage and DSBs. A similar lack of effect of comet assay with Fpg, despite positive immunostaining, have previously been reported  and is probably due to an artefact; various micro and nanoparticles have been reported to interact with Fpg, decreasing the sensitivity of the assay , and PM may have similar effects.
Interestingly, 8-oxodG was increased by whole-PM but not by its organic extract, suggesting a more direct interaction of some PM component (including both metals and various PAHs) with the DNA in the nucleus [70–72]. It is known that 8-oxodG is induced by singlet oxygen and hydroxyl radical which, due to their high reactivity, will only react with DNA when generated in direct proximity . Thus, our results suggest that ROS formed in the cytosol when exposed to the organic fraction will not interact with the cellular DNA. Previous data in our laboratory indicated that PM may be in close contact with the chromosomes , but the current data is not conclusive and this potential nuclear localization of PM would require further investigations.
In conclusion, the dose used in the present study is among the lowest reported to have biological effects in vitro
. Our study shows that this low dose of winter PM2.5 induces an early G2 arrest followed by an arrest in M/A with a subsequent inhibition of cytokinesis and an increased formation of cells with double nuclei and MN. These effects are associated with a rapid DNA damage response and the formation of mitotic spindle aberrations. The early DNA damage and G2/M accumulation have been related to the formation of reactive electrophilic/radical metabolites via a P450-depending reaction. However, PM2.5 apparently also has spindle poison properties which contribute to the induction of the M/A arrest. The characterization of the process leading to double nuclei and MN in PM-exposed cells is of great importance, giving a possible explanation for PM-induced chromosomal aberrations. Such events could be central when explaining the increased lung cancer incidence associated with PM2.5 and deserve further investigations.