Particulate Matter (PM10) Promotes Cell Invasion through Epithelial–Mesenchymal Transition (EMT) by TGF-β Activation in A549 Lung Cells

Air pollution presents a major environmental problem, inducing harmful effects on human health. Particulate matter of 10 μm or less in diameter (PM10) is considered an important risk factor in lung carcinogenesis. Epithelial–mesenchymal transition (EMT) is a regulatory program capable of inducing invasion and metastasis in cancer. In this study, we demonstrated that PM10 treatment induced phosphorylation of SMAD2/3 and upregulation of SMAD4. We also reported that PM10 increased the expression and protein levels of TGFB1 (TGF-β), as well as EMT markers SNAI1 (Snail), SNAI2 (Slug), ZEB1 (ZEB1), CDH2 (N-cadherin), ACTA2 (α-SMA), and VIM (vimentin) in the lung A549 cell line. Cell exposed to PM10 also showed a decrease in the expression of CDH1 (E-cadherin). We also demonstrated that expression levels of these EMT markers were reduced when cells are transfected with small interfering RNAs (siRNAs) against TGFB1. Interestingly, phosphorylation of SMAD2/3 and upregulation of SMAD induced by PM10 were not affected by transfection of TGFB1 siRNAs. Finally, cells treated with PM10 exhibited an increase in the capacity of invasiveness because of EMT induction. Our results provide new evidence regarding the effect of PM10 in EMT and the acquisition of an invasive phenotype, a hallmark necessary for lung cancer progression.


Introduction
Air pollution presents one of the major environmental problems, inducing harmful effects in human health. These effects have a major impact in highly populated cities, where industrial activity and vehicle traffic increase emissions of multiple pollutants. Epidemiological and toxicological research led to the classification of air pollution as carcinogenic in Group 1 by the International Agency for Research on Cancer (IARC) [1]. Multiple epidemiologic studies have demonstrated that particulate matter (PM) inhalation has a negative impact on human health, increasing the occurrence of respiratory diseases such as chronic obstructive pulmonary disease, asthma, and, over the long term, lung cancer [2][3][4][5]. According to the International Agency for Research in Cancer (IARC), lung cancer is still the leading cause of cancer death worldwide. In 2018, approximately 1.7 million deaths were attributed to lung cancer cases. Interestingly, from all lung cancer cases, at least 29% were linked with ambient air pollution [6,7]. To act against air pollution, the World Health Organization (WHO) developed the ambient air quality database, with the aim of maintaining particles with an aerodynamic diameter less or equal to 10 µm (PM 10 ) and an 2. Results 2.1. Particulate Matter PM 10 Increases the Phosphorylation of SMAD2/3 and the Levels of SMAD4 in A549 Cells Cells exposed to PM 10 for 48 h and transfected with non-targeting siRNA showed an increase in the phosphorylation levels of SMAD2/3 when compared with control group (fold change→FC of 1.24 vs. 1.00; p < 0.05) ( Figure 1A). In addition, cells treated with TGF-β showed an increase in the phosphorylation levels of SMAD2/3 when compared with control group (FC of 1.19 vs. 1.00; p < 0.05). Cells transfected with TGFB1 siRNAs showed no changes in the phosphorylation levels of SMAD2/3 in non-treated, PM 10 , and TGF-β groups (FC of 0.89, 0.92, and 0.91 vs. 1.00, respectively; p < 0.05), suggesting a link between the phosphorylation status of SMAD2/3 and PM 10 treatment independent of TGFB1 repression. Moreover, cells exposed to PM 10 for 48 h and transfected with nontargeting siRNA showed an increase in the protein levels of SMAD4 when compared with control group (FC of 1.19 vs. 1.00; p < 0.05) ( Figure 1B). Cells treated with TGF-β also showed an increase in protein levels of SMAD4 when compared with control group (FC of 1.27 vs. 1.00; p < 0.05). Cells transfected with TGFB1 siRNAs showed no change in the protein levels of SMAD4 in non-treated, PM 10 , and TGF-β groups (FC of 1.13, 1.07, and 1.12 vs. 1.00, respectively; p < 0.05), suggesting a link between the abundance of SMAD4 and PM 10 treatment independent of TGFB1 repression. control group (FC of 1.19 vs. 1.00; p < 0.05). Cells transfected with TGFB1 siRNAs showed no changes in the phosphorylation levels of SMAD2/3 in non-treated, PM10, and TGF-β groups (FC of 0.89, 0.92, and 0.91 vs. 1.00, respectively; p < 0.05), suggesting a link between the phosphorylation status of SMAD2/3 and PM10 treatment independent of TGFB1 repression. Moreover, cells exposed to PM10 for 48 h and transfected with non-targeting siRNA showed an increase in the protein levels of SMAD4 when compared with control group (FC of 1.19 vs. 1.00; p < 0.05) ( Figure 1B). Cells treated with TGF-β also showed an increase in protein levels of SMAD4 when compared with control group (FC of 1.27 vs. 1.00; p < 0.05). Cells transfected with TGFB1 siRNAs showed no change in the protein levels of SMAD4 in non-treated, PM10, and TGF-β groups (FC of 1.13, 1.07, and 1.12 vs. 1.00, respectively; p < 0.05), suggesting a link between the abundance of SMAD4 and PM10 treatment independent of TGFB1 repression.

Figure 1.
Particulate matter PM10 increases the phosphorylation of SMAD2/3 and the levels of SMAD4 in A549 cells after 48 h of treatment: (A) Particulate matter exposure increases the phosphorylation levels of SMAD2/3 in A549 cells transfected with non-targeting siRNA, with a similar effect in cells treated with TGF-β for 48 h. (B) PM10 exposure increases the protein levels of SMAD4 in A549 cells transfected with non-targeting siRNA, with a similar effect in cells treated with TGF-β for 48 h. These effects are not prevented in A549 cells transfected with TGFB1 siRNAs. Cells were exposed to PM10 (10 µ g/cm 2 ) and TGF-β (5 ng/mL) for 48 h. The quantitative results were expressed after normalization using GAPDH as a loading control. Data are reported as the means ± SD of three independent experiments. The images are representative of the data obtained. (*) indicates statistically differences between treatments. Densitometry was performed using arbitrary units p < 0.05.

Particulate Matter PM10 Induce the Expression of TGFB1 and EMT Components in A549 Cells
Cells exposed to PM10 for 48 h and transfected with non-targeting siRNA showed an increase in the expression levels of TGFB1 when compared with control group (FC of 1.36 vs. 1.00; p < 0.05) ( Figure 2). In addition, cells treated with TGF-β showed an increase in the expression levels of TGFB1 when compared with control group (FC of 3.06 vs. 1.00; p < 0.05). Moreover, cells transfected with TGFB1 siRNAs showed a decrease in TGFB1 expression levels in non-treated, PM10, and TGF-β groups (FC of 0.13, 0.06, and 0.28 vs. 1.00, respectively; p < 0.05), maintaining the effect in TGFB1 knockdown. Cells exposed to PM10 Figure 1. Particulate matter PM 10 increases the phosphorylation of SMAD2/3 and the levels of SMAD4 in A549 cells after 48 h of treatment: (A) Particulate matter exposure increases the phosphorylation levels of SMAD2/3 in A549 cells transfected with non-targeting siRNA, with a similar effect in cells treated with TGF-β for 48 h. (B) PM 10 exposure increases the protein levels of SMAD4 in A549 cells transfected with non-targeting siRNA, with a similar effect in cells treated with TGF-β for 48 h. These effects are not prevented in A549 cells transfected with TGFB1 siRNAs. Cells were exposed to PM 10 (10 µg/cm 2 ) and TGF-β (5 ng/mL) for 48 h. The quantitative results were expressed after normalization using GAPDH as a loading control. Data are reported as the means ± SD of three independent experiments. The images are representative of the data obtained. (*) indicates statistically differences between treatments. Densitometry was performed using arbitrary units p < 0.05. 10 Induce the Expression of TGFB1 and EMT Components in A549 Cells Cells exposed to PM 10 for 48 h and transfected with non-targeting siRNA showed an increase in the expression levels of TGFB1 when compared with control group (FC of 1.36 vs. 1.00; p < 0.05) ( Figure 2). In addition, cells treated with TGF-β showed an increase in the expression levels of TGFB1 when compared with control group (FC of 3.06 vs. 1.00; p < 0.05). Moreover, cells transfected with TGFB1 siRNAs showed a decrease in TGFB1 expression levels in non-treated, PM 10 , and TGF-β groups (FC of 0.13, 0.06, and 0.28 vs. 1.00, respectively; p < 0.05), maintaining the effect in TGFB1 knockdown. Cells exposed to PM 10 and transfected with non-targeting siRNA also exhibited an increase in expression levels of SNAI1 (Snail), SNAI2 (Slug), ZEB1 (ZEB1), CDH2 (N-cadherin), ACTA2 (α-SMA), and VIM (vimentin) (FC of 1.21, 2.79, 1.30, 1.39, 1.41, and 3.02 vs. 1.00, respectively; p < 0.05), and a decrease in expression levels of CDH1 (E-cadherin) when compared with control group (FC of 0.29 vs. 1.00, respectively; p < 0.05). A similar effect in EMT genes SNAI1 (Snail), SNAI2 (Slug), ZEB1 (ZEB1), CDH2 (N-cadherin), ACTA2 (α-SMA), VIM (vimentin), and CDH1 (E-cadherin) was observed in cells treated with TGF-β (FC of 2.98, 4.10, 1.86, 3.09, 1.84, 3.12, and 0.12 vs. 1.00, respectively; p < 0.05). When cells were transfected with TGFB1 siRNAs, EMT gene SNAI1 (Snail) exhibited a decrease in expression levels in non-treated and PM 10 , but an increase in expression levels in TGF-β group (FC of 0.91, 0.64, and 3.60, respectively; p < 0.05) when compared with their non-targeting siRNA pair. A similar effect was observed in CDH1 (E-cadherin) (FC of 0.75, 0.27, and 0.07, respectively; p < 0.05) and CDH2 (N-cadherin) (FC of 0.75, 0.78, and 2.39, respectively; p < 0.05). Cells transfected with TGFB1 siRNAs exhibited a decrease in expression levels in non-treated group, but an increase in expression levels in PM 10 and TGF-β-treated groups in SNAI2 (Slug) (FC of 0.80, 3.22, and 5.11, respectively; p < 0.05) and VIM (vimentin) (FC of 0.69, 1.94, and 3.54, respectively; p < 0.05). On the other hand, cells transfected with TGFB1 siRNAs exhibited no changes in expression levels in non-treated and PM 10 groups, but an increase in expression levels in PM10 and TGF-β-treated groups in ZEB1 (ZEB1) (FC of 1.06, 1.36, and 2.40, respectively; p < 0.05) and ACTA2 (α-SMA) (FC of 1.12, 1.29, and 1.92, respectively; p < 0.05).  . Cells were exposed to PM10 (10 µ g/cm 2 ) and TGF-β (5 ng/mL) for 48 h. The quantitative results were expressed after normalization using GAPDH as a control. Data are reported as the means ± SD of three independent experiments. The images are representative of the data obtained. (*) indicates statistically differences between treatments. Relative quantification or fold change (FC) was performed using the 2 −ΔΔCt ; p < 0.05. (N-cadherin), and VIM (vimentin). Cells were exposed to PM 10 (10 µg/cm 2 ) and TGF-β (5 ng/mL) for 48 h. The quantitative results were expressed after normalization using GAPDH as a control. Data are reported as the means ± SD of three independent experiments. The images are representative of the data obtained. (*) indicates statistically differences between treatments. Relative quantification or fold change (FC) was performed using the 2 −∆∆Ct ; p < 0.05.

Particulate Matter PM10 Activates the Invasive Phenotype in A549 Cells
A549 cells exposed to PM10 for 48 h and transfected with non-targeting siRNA showed an increase in the invasiveness levels when compared with control group (FC of 1.39 vs. 1.00; p < 0.05) (Figure 4). In addition, cells treated with TGF-β showed an increase in invasiveness level when compared with control group (FC of 1.74 vs. 1.00; p < 0.05). Cells transfected with TGFB1 siRNAs showed no changes in invasiveness levels in the non-treated, PM10, and TGF-β groups (FC of 1.00, 0.97, and 1.00 vs. 1.00, respectively; p < 0.05), suggesting a role of TGFB1 expression in the induction of the invasive phenotype. . Particulate matter PM10 activates the invasive phenotype in A549 cells. Particulate matter exposure increases the invasiveness of A549 cells exposed to PM10 and transfected with non-targeting siRNA, with a similar effect in cells treated with TGF-β for 48 h. These effects are prevented in A549 cells transfected with TGFB1 siRNAs. Cells were exposed to PM10 (10 µ g/cm 2 ) and TGF-β (5 ng/mL) for 48 h. The quantitative results were expressed after obtaining the relative fluorescence units of each condition. Data are reported as the means ± SD of three independent experiments. The images are representative of data obtained. (*) indicates statistical differences between treatments p < 0.05.

Discussion
In the present study, we evaluated the effect of PM10 in the induction of EMT through the TGF-β pathway, considering the phosphorylation of EMT cascade initiators such as SMAD2/3, as well as the upregulation of SMAD4, in order to follow the transcription ac-  10 activates the invasive phenotype in A549 cells. Particulate matter exposure increases the invasiveness of A549 cells exposed to PM 10 and transfected with non-targeting siRNA, with a similar effect in cells treated with TGF-β for 48 h. These effects are prevented in A549 cells transfected with TGFB1 siRNAs. Cells were exposed to PM 10 (10 µg/cm 2 ) and TGF-β (5 ng/mL) for 48 h. The quantitative results were expressed after obtaining the relative fluorescence units of each condition. Data are reported as the means ± SD of three independent experiments. The images are representative of data obtained. (*) indicates statistical differences between treatments p < 0.05.

Discussion
In the present study, we evaluated the effect of PM 10 in the induction of EMT through the TGF-β pathway, considering the phosphorylation of EMT cascade initiators such as SMAD2/3, as well as the upregulation of SMAD4, in order to follow the transcription activation of multiple EMT markers. We found that PM 10 is capable of inducing the expression of TGFB1 and EMT markers SNAI1, SNAI2, ZEB1, CDH2, ACTA2, and VIM as well as their respective products TGF-β, Snail, Slug, ZEB1, N-cadherin, α-SMA, and vimentin in the lung A549 cell line. Cell exposed to PM 10 also shows a decrease in the expression of CDH1 and its product, E-cadherin. Interestingly, expression levels of these EMT markers were reduced when cells are transfected with small interfering RNAs (siRNAs) against TGFB1. However, phosphorylation of SMAD2/3 and upregulation of SMAD induced by PM 10 was not affected by transfection of TGFB1 siRNAs. As a response to these molecular changes in EMT components on cells treated with PM 10 , the biological consequence is the increase in the capacity of invasiveness.
EMT is involved in multiple pulmonary fibrotic diseases as well as in lung cancer [12,[33][34][35][36]. According to the physiological and pathological processes where is involved, EMT has been classified into three types. While type 1 and type 2 are related to embryonic development and tissue regeneration/cicatrization, respectively, type 3 EMT has been widely recognized as necessary for cancer cells in order to confer an invasive and metastatic phenotype, where cancer cells not only show conversion from epithelial to mesenchymal cells, but also undergo mesenchymal-epithelial transition (MET) in sites where they are established after metastasis [37]. Some studies have detected the induction of type 2 and 3 EMT in lung diseases, providing some clues regarding the relevance of PM in EMT [38,39].
Cancer pathways where EMT is involved are well known. In this context, the induction of EMT is mainly associated with factors such as TGF-β, SMAD2/3, MAPK, IL-6, Wnt, SHH, and Notch [40,41]. The target of these pathways involves multiple transcription factors but show affinity to SNAI1, SNAI2, and ZEB. This step in the EMT signaling is crucial because these transcription factors have a direct impact on E-cadherin expression. Thus, EMT signaling is oriented to cause an inhibition of epithelial markers and, in the same way, cause an induction of mesenchymal markers.
EMT markers that are well established and usually detected in vitro or in vivo include a decrease in E-cadherin expression, accompanied by an increase in the expression levels of N-cadherin, α-SMA, and vimentin [41,42]. Additionally, an increase in the expression levels of MMPs and type 1 collagen is also induced [43,44]. Together, these EMT components create a microenvironment in which growth factors, degrade extracellular matrix enzymes and extracellular matrix proteins, and orchestrate the activation of autocrine and paracrine signals oriented to sustain EMT in cancer [45,46].
The mechanisms by which air pollution is capable of inducing cell transformation depend on a complex interaction between environmental and individual factors. PM is a mixture of organic, inorganic, and biological entities interacting not only in the air, but also in their site of deposition. The composition of this mixture is dependent of the location where the PM is collected, the sources of emission near to the collection site, and the season in which PM is sampled [47][48][49][50]. While epidemiological studies have reported a sustained increase in risk for lung cancer incidence of approximately 8% per each increase of 10 µg/m 3 in PM 10 concentrations [51], other evidence points to the fact that PM 10 as well as PM 2.5 are associated with poorer lung cancer survival, emphasizing their effect in early stage non-small cell lung cancer [52]. Because air pollution has been catalogued as a Group I carcinogen by the IARC, the underlying process of malignant cell transformation is expected to follow the multistep sequence described for these group of diseases: initiation, promotion, and progression [53]. As part of the cancer progression step, a reversible process called epithelial-mesenchymal transition (EMT) takes place. The goal of EMT in cancer is the colonization of distant sites, starting with the signaling necessary for the induction of a phenotype suitable for cell migration (mesenchymal), followed by the invasion of adjacent tissues, and finally, colonizing distant tissues trough metastasis and promoting the establishment by switching again the phenotype through the epithelial condition [54]. Thus, the EMT process is the program that converts a benign tumor into an aggressive and highly invasive neoplasm.
Several cancer progression programs are subject to change by PM 10 , and these effects have been linked to deregulated expression of genes caused by PAHs contained in PM [24]. Further, PM exposure activates inducers of EMT in lung cancer cells, impacting positively the downstream signaling of multiple mesenchymal markers [23]. Chronic exposure to PM also induces a well-defined EMT morphology in A549 cells [25]. Furthermore, chronic exposure has described the carcinogenic mechanisms of PM in the induction of EMT, emphasizing the role of the organic fraction in the induction of different degrees of EMT progression and increased invasion ability in BEAS-2B cells in a concentration-dependent manner [26]. In fact, BEAS2B exposed to PM exhibits a decrease in E-cadherin, followed by an increase in mesenchymal markers such as α-SMA, Snail, Slug, ZEB1, and extracellular matrix degradation related to an increase in TGFB1, MMP3, and MMP9 [55].
The effect of PM in migration and invasion pathways in cancer cells has also been described in A549 cells. Toxicological assays as well as global transcriptome profiling exhibit an increase in the activity of proteases, as well as an invasive phenotype (due to the up-regulation of metalloproteases) when cells are exposed to PM 10 and PM 2.5 [28][29][30][31]. PM exposure has been also associated with methylation of EMT genes [27]. PM components also increase cell permeability, according to experiments measuring transepithelial electrical resistance. Deregulation of metalloprotease activity, as well as other signaling pathways, enhance the ability of invasion and induce orthotopic lung carcinoma metastasis in mice [32]. This evidence suggests that both, PM 10 and PM 2.5 are potent signal transducers, acting as activators of transcription of key regulators in EMT, a condition strongly linked to tumor progression [56].
A variety of cell lines have been considered for the evaluation of the effect of PM in the induction of EMT. These models include the epithelial bronchial cell lines BEAS2B, 16HBE, HBEC, and HBEC3, as well as the alveolar carcinoma cell line A549 and the lung carcinoma NCI-H441 and H292 cell lines. From these, BEAS2B and A549 are the most frequently used models [57]. Although several studies report similar effects of PM in both cell lines, we considered the A549 cell line to be a more suitable model for the study of EMT, once that is a transformed cell line. This feature is crucial in the evaluation of the carcinogenic process because invasion and metastasis are cancer hallmarks involved in progression. We performed experiments in A549 cells treated with PM 10 for 24 h (data not shown). However, the effects of PM 10 in EMT are more evident after 48 h of exposure. Moreover, when we performed the invasion assay in BEAS2B, no changes were observed when cells were treated with PM 10 or TGF-β and compared with controls (data not shown). These results suggest that BEAS2B could lack of some cancer hallmarks necessary for the acquisition of an invasive phenotype. Further studies should be addressed in search of the invasion capacity between these models in order to consider if a non-transformed cell line is suitable for analysis of EMT.
To our knowledge, this is the first study that has demonstrated the effect of PM 10 on the induction of the growth factor TGF-β, and its influence on EMT regulation. PM 10 induces TGFB1 overexpression followed by overexpression of EMT genes SNAI1 (Snail), SNAI2 (Slug), ZEB1 (ZEB1), CDH2 (N-cadherin), ACTA2 (α-SMA), and VIM (vimentin), and a decrease in the expression of CDH1 (E-cadherin) ( Figure 5). These changes in gene expression result in TGF-β autoinduction, the activation of type 3 EMT, and the subsequent acquisition of an invasive phenotype, features observed in cancer cells. We also demonstrated that PM 10 is capable of inducing phosphorylation of SMAD2/3 and an increase in protein levels of SMAD4 in A549 cells treated with PM 10 . Our results suggest that phosphorylation of SMAD2/3, as well as an increase in levels of SMAD4, are not influenced by TGFB1 overexpression caused by PM 10 exposure in A549 cells. Moreover, this is the first in vitro study to document the invasive phenotype associated with PM 10 exposure, integrating gene expression, protein synthesis, and signal transduction 9 of 14 of multiple EMT markers. These findings offer new evidence regarding the effect of air pollution in tumor progression, by the acquisition of a key hallmark in carcinogenesis. The effect of air pollution in EMT signaling needs to be evaluated by other approaches, such as the analysis of EMT markers in vivo exposed to PM or of lung cancer patients exposed to air pollution, to determine the role of this risk factor during cancer continuum, and to associate the effect of PM 10 in lung cancer susceptibility, incidence, and survival.

PM10 Sampling
Collection of PM10 was performed in Mexico City area using the GMW model 1200 VFC HVPM10 high-volume particle collector with a flux of 1.13 m 3 /min (Sierra Andersen, Atlanta, GA, USA). Particles were captured on cellulose nitrate filters with a pore size of 3.0 μm (Sartorius AG, Goettingen, Germany) for 3 d per week. Storage of the filters was made at 4 °C in dark and dry conditions. After that, particles were removed by gently scraping off membranes, using a surgical blade. Samples were placed for storage into endotoxin-free glass vials. All vials were maintained at 4 °C in dark and dry conditions. Prior

PM 10 Sampling
Collection of PM 10 was performed in Mexico City area using the GMW model 1200 VFC HVPM10 high-volume particle collector with a flux of 1.13 m 3 /min (Sierra Andersen, Atlanta, GA, USA). Particles were captured on cellulose nitrate filters with a pore size of 3.0 µm (Sartorius AG, Goettingen, Germany) for 3 d per week. Storage of the filters was made at 4 • C in dark and dry conditions. After that, particles were removed by gently scraping off membranes, using a surgical blade. Samples were placed for storage into endotoxin-free glass vials. All vials were maintained at 4 • C in dark and dry conditions. Prior to their use for treatment in cell cultures, particles were always sterilized for 15 min at 121 • C and 15 PSI. Characterization of collected PM 10 is available in previous reports [58][59][60].

Lung Epithelial A549 Cell Culture and Exposure
Cell line A549 (human adenocarcinoma alveolar basal epithelial cells) was obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Cell culture was performed using F-12 Kaighn's medium (Gibco, New York, NY, USA) at 37 • C under an atmosphere of 5% of CO 2 . Only cultures reaching a confluence of 80% of the surface were considered before propagation or experiment running. All experiments were performed in 6-well plates, considering an initial concentration of 250,000 cells per well. Cells were kept in F-12 Kaighn's medium considering a concentration of 10% of fetal bovine serum (FBS) (Gibco, Brooklyn, NY, USA) for 48 h. Before treatment (4 h prior to exposure), cells were transfected with the Silencer Select Negative Control small interfering RNA (siRNA) (Ambion-Thermo Fisher, Waltham, MA, USA), a control non-targeting siRNA, employing the Xfect RNA Transfection Reagent (Takara, Clontech, Beijing, China) according to manufacturer's protocol. A549 cells were exposed to non-targeting siRNA to a final concentration of 20 nM. After 4 h, medium was removed, plates were then washed with PBS, and cells were treated with PM 10 , considering a final concentration of 10 µg/cm 2 in F-12 Kaighn's medium supplemented with 10% of FBS. This dose is equivalent to a 5-day exposure in humans, according to previous reports [61,62]. Cultures were evaluated at 48 h of treatment, comparing results with control cells (cultures without PM 10 treatment) Additionally, experiments included cells treated for 48 h with TGF-β, 5 ng/mL (Sigma, St. Louis, MI, USA), serving as positive control of EMT induction.

Knockdown of TGF-β
To compare the effect of PM 10 in the induction of EMT markers and evaluate the endpoints proposed, cultures of A549 cells were also transfected using a mixture of small interference RNAs (siRNAs) against TGFβ1 (mRNA of TGF-β), targeting multiple regions of its mRNA to improve the likelihood of degradation. Assays were performed in 6-well plates using an initial concentration of 250,000 cells per well. Cells were kept in F-12 Kaighn's medium with 10% of fetal bovine serum (FBS) (Gibco, Brooklyn, NY, USA) for 48 h. Then, medium was removed, plates were washed in PBS and cells were synchronized by serum starvation for 24 h using free-FBS F-12 Kaighn's medium. Before treatment (4 h prior to exposure), cells were transfected with the TGFβ1 siRNA (h2) (sc-270322) (Santa Cruz Biotechnology, Paso Robles, CA, USA), employing the XfectTM RNA Transfection Reagent (Takara, Clontech, Beijing, China) according to manufacturer's protocol. A549 cells were exposed to TGFβ1 siRNA to a final concentration of 20 nM (Supplementary Figure S1). After 4 h, medium was removed, plates were then washed in PBS, and cells were treated with PM 10 (10 µg/cm 2 ), and TGF-β (100 ng/mL; Sigma, St. Louis, MI, USA), respectively. Comparisons in all experiments were performed considering two conditions (cells transfected with non-targeting siRNA and cells transfected with TGFβ1 siRNAs) with three experimental groups each of them (control/non-treated, PM 10 , and TGF-β).