PM2.5 Compositions analysis
As shown in Table1, the dominant components of metal ions and part non-metallic ions in PM2.5 were Al (1255.33 ± 5.51 ng/mg), Ca (9351.68 ± 69.97 ng/mg), Fe (2691.15 ± 20.84 ng/mg), Na (1567.68 ± 13.40 ng/mg), K (1567.43 ± 2.98 ng/mg), Mg (1074.41 ± 6.25 ng/mg) and S (12688.06 ± 66.99 ng/mg). The average concentrations of OC and EC in the PM2.5 were 527.11 ± 144.62 µg/mg and 119.71 ± 50.11 µg/mg, and the OC/EC ratio was about 4.61 ± 1.16. Among the total of 16 PAHs elements measured, Naphthalene, Acenaphthylene and Pyrene were the most abundant elements in the PM2.5.
Table 1
Metals, Carbon and organic compositions detected in PM2.5
Metals
|
Concentration (ng/mg)
|
OC\EC
|
Concentration (µg/mg)(µg/mg)
|
Al
|
1255.33 ± 5.51
|
OC
|
527.11 ± 144.62
|
As
|
11.53 ± 0.46
|
EC
|
119.71 ± 50.11
|
B
|
28.59 ± 0.39
|
OC/EC
|
4.61 ± 1.16
|
Ba
|
95.20 ± 1.00
0.11 ± 0.00
|
PAHs
|
Concentration (ng/mg)
|
Be
|
Bi
|
1.69 ± 0.97
|
Naphthalene
|
32.38 ± 1.34
|
Ca
|
9351.68 ± 69.97
|
Acenaphthylene
|
1.16 ± 0.49
|
Cd
|
2.96 ± 0.01
|
Acenaphthene
|
0.56 ± 0.32
|
Co
|
2.02 ± 0.15
|
Fluorene
|
0.42 ± 0.30
|
Cu
|
94.05 ± 1.18
|
Phenanthrene
|
0.14 ± 0.06
|
Fe
|
2691.15 ± 20.84
|
Anthracene
|
0.01 ± 0.01
|
Hg
|
1.16 ± 0.07
|
Fluoranthene
|
0.21 ± 0.15
|
K
|
1567.43 ± 2.98
|
Pyrene
|
0.66 ± 0.14
|
Li
|
4.28 ± 0.21
|
Benzo [A] Anthracene
|
0.07 ± 0.01
|
Mg
|
1074.41 ± 6.25
|
Chrysene
|
0.16 ± 0.12
|
Mn
|
181.71 ± 1.58
|
Benzo [B] Fluoranthene
|
0.39 ± 0.16
|
Mo
|
8.43 ± 0.04
|
Benzo [K] Fluoranthene
|
0.15 ± 0.11
|
Na
|
1567.68 ± 13.40
|
Benzo [A] Pyrene
|
0.15 ± 0.06
|
Ni
|
27.37 ± 0.16
|
Indene (1, 2, 3-cd) Pyene
|
0.39 ± 0.06
|
P
|
291.58 ± 0.15
|
Diphenyl Anthracene (A, H)
|
0.06 ± 0.08
|
Pb
|
106.62 ± 0.70
|
Benzo [G, H, I] Pyrene
|
0.17 ± 0.13
|
S
|
12688.06 ± 66.99
|
|
|
Sb
|
14.08 ± 0.58
|
|
|
Se
|
12.61 ± 1.52
|
|
|
Si
|
1019.26 ± 7.59
|
|
|
Sn
|
13.56 ± 0.49
|
|
|
Sr
|
35.02 ± 0.33
|
|
|
Ti
|
62.17 ± 0.22
|
|
|
V
|
73.25 ± 0.36
|
|
|
W
|
18.56 ± 1.70
|
|
|
Zn
|
493.88 ± 0.89
|
|
|
Zr
|
4.23 ± 0.39
|
|
|
OC: organic carbon; EC: elemental carbon; PAHs: Polycyclic Aromatic Hydrocarbons. |
PM 2.5 induces lipid droplet formation in and promotes the aggressive phenotypes of lung cancer cells.
To study the effect of PM2.5 on lung cancer, we treated A549 and H1975 LUAD cells with PM2.5. After the treatment, we observed clear increases in lipid droplets in both cell lines evaluated with electron microscope or Oil Red O staining compared to those untreated cells (Fig. 1a). Cell migration assays showed that A549 and H1975 treated with PM2.5 migrated faster than those untreated cells, although A549 showed large variation in migration at 24 h (Fig. 1b). Colony formation assays showed increased colony formation in both A549 and H1975 after the cells were treated with PM2.5 for 14 days (Fig. 1c). Transwell assays demonstrated that A549 and H1975 cells treated with PM2.5 became more aggressive, and the number of cells penetrating through the Matrigel membrane increased by 2 ~ 4 times (Fig. 1d).
PM 2.5 upregulates the expression of lipogenic genes and induces lipid metabolism disorder.
To understand the mechanisms by which PM2.5 induces lipid droplet formation and promotes LUAD progression, we analyzed the transcriptomes of A549 cells with RNA-seq after PM2.5 treatment. The transcriptomic analysis revealed 263 differentially expressed genes (DEGs) between cells treated with and without PM2.5. Of these DEGs, 148 were upregulated, and 115 were downregulated (Fig. 2a). Using the Kyoto Encyclopaedia of Genes and Genomes (KEGG) algorithm to interrogate the DEG profiles, we found significant signal enrichment in fatty acid metabolism (Fig. 2b). The top two genes involved were SREBFs (Fig. 2c). We confirmed these findings with RT-qPCR analysis on SREBF1 and its downstream targets, FASN and ACACA, all of which were upregulated significantly after PM2.5 treatment (Fig. 2d).
PM 2.5 promotes N-SREBP1 translocation into the nucleus, inducing lipid droplet formation and LUAD progression.
Following the above findings, we speculate that PM2.5 may upregulate the expression of SREBP1 and its downstream genes, which leads to changes in lipid metabolism, resulting in increases in lipid droplets and aggressive cell behaviors. However, the level of total SREBP1 protein was not elevated after PM2.5 exposure in either lung cancer cell lines based on western blot analysis, while its downstream proteins FASN and ACACA were clearly upregulated (Fig. 3a). Using fluorescence immunohistochemical staining for SREBP1, we found that the protein was enriched in the cell nucleus (Fig. 3b). Western blot analysis showed that N-SREBP1 and SREBP cleavage activating protein (SCAP) were upregulated after PM2.5 exposure (Fig. 3a). Next, we used siRNA to knock down SREBF1 and the main downstream gene FASN. Oil Red O staining showed that the number of intracellular lipid droplets in the knockdown cells after PM2.5 exposure was reduced (Fig. 3c). Additionally, knocking down the two genes inhibited the PM2.5-induced increases in cell invasion, migration and colony formation (Fig. 3d-3f), suggesting that SREBP1 plays an important role in the accumulation of lipid droplets and the aggression of lung cancer cells induced by PM2.5 exposure.
D-limonene inhibits PM 2.5 -induced lipid metabolism disorder in lung cancer cells by upregulating miR-195.
To investigate the molecular targets and mechanisms of D-limonene in lung cancer, we analysed the transcriptomes of A549 cells treated with or without D-limonene after PM2.5 exposure. Compared to those without D-limonene treatment, cells with D-limonene treatment showed downregulation of the genes involved lipid biogenesis, such as SREBF1, FASN and ACACA (Fig. 4a). These observations were confirmed by our RT-qPCR analyses (Fig. 4b). Interestingly, we found that the expression of miR-195, a known tumor suppressor, was substantially upregulated after D-limonene treatment (Fig. 4c). Using miRWalk [33] and miRSystem [34], we found that miR-195 was predicted to interact with SREBF1 in the CDS region when energy < -20, accessibility < 0.0001, me < -8. The in silico prediction also indicated that the miRNA could interact with ACACA and FASN mRNAs at their 3' UTRs. The prediction that the seed sequence of miR-195 can directly bind to the CDS region of SREBF1 and the 3' UTR of FASN and ACACA was verified by our dual luciferase assays. Our experiments showed that the relative luciferase activity was reduced in the WT groups (SREBF1, FASN, ACACA) compared with the mutant groups (Fig. 4d). After miR-195 was overexpressed in A549 cells, western blot analysis showed that SREBP1, FASN and ACACA were significantly inhibited (Fig. 4e).
To verify the effects of D-limonene and miR-195 on inhibiting lipid metabolism disorder caused by PM2.5 exposure, A549 cells were treated with or without miR-195 after PM2.5 exposure. We found that miR-195 treatment significantly inhibited the upregulated expression of the SREBF1, FASN and ACACA genes induced by PM2.5 exposure (Fig. 4f). After treatment with D-limonene or miR-195, lipid droplet formation induced by PM2.5 exposure was decreased dramatically in the cells (Fig. 4g). Western blot analysis showed that D-limonene or miR-195 decreased the protein levels of SREBP1, FASN and ACACA, which were upregulated in A549 and H1975 cells after PM2.5 exposure (Fig. 4h). Colony formation and Transwell experiments showed that both D-limonene and miR-195 inhibited the PM2.5 exposure-induced proliferation and invasion of lung cancer cells (Fig. 4i-4j).
D-limonene prevents pulmonary fibrosis caused by long-term exposure to PM 2.5 in normal lung epithelial cells and mice by repairing lipid metabolism disorder.
Following our observations that PM2.5 could promote tumor cell aggressive behaviors and that D-limonene could inhibit this process by reducing lipid metabolism disorder caused by PM2.5 exposure, we further explored whether PM2.5 also had a detrimental effect on normal lung tissue in vitro and in vivo and whether D-limonene can block or inhibit this effect. Normal lung epithelial cells (Beas-2b) were cultured under PM2.5 exposure for 30 passages. The PM2.5 treatment led to significant lipid droplets accumulation in the cells, and this phenomenon did not appear after the cells were treated with both PM2.5 and D-limonene (Fig. 5a). We also found a large amount of lipid droplets in the lung tissue of C57BL/6J mice after the animals were treated with PM2.5 inhalation for 60 days, but no lipid droplets were observed in the animals when D-limonene was orally administered simultaneously with PM2.5 inhalation (Fig. 5a). RNA-Seq analysis was performed on the lung tissues of mice treated with PM2.5 or PM2.5 plus D-limonene. The expression of SREBF1, FASN and ACACA were upregulated by PM2.5 exposure, and the upregulations were suppressed by D-limonene treatment. Similar results were also seen in normal lung epithelial cells, and the findings were verified by RT-qPCR (Fig. 5b, 5d). Our animal models also showed that long-term PM2.5 exposure significantly increased serum levels of TGs in the mice, and D-limonene treatment attenuated this effect (Fig. 5e).
To assess time-dependent variations, we collected Beas-2b cell samples every 5 passages, and the samples were analysed with western blotting. As shown in Fig. 5c, the expression of SREBP1, FASN and ACACA were significantly higher in PM2.5-exposed than in control groups, and the increases appeared as early as in the fifth generation of cells. These proteins were also upregulated in the mouse lung tissue after a long period of exposure to PM2.5. Similar to the observations in LUAD cells, the expression of SREBP1, FASN and ACACA, which were promoted by PM2.5 exposure, were inhibited by D-limonene treatment to the mice (Fig. 5c).
The RNA-Seq results of miR-195 KO and WT mice treated with PM2.5 with or without D-limonene showed that D-limonene inhibited the increases in expression of SREBF1, FASN and ACACA induced by PM2.5 exposure in the KO mice (Fig. 5f), but the inhibitory effect was weaker than that in the WT mice (Fig. 5g). Furthermore, Masson’s trichrome staining and western blotting analysis revealed that pulmonary fibrosis occurred in both WT mice and miR-195 KO mice after long-term exposure to PM2.5. This phenotype could be rescued by D-limonene treatment. The level of pulmonary fibrosis in miR-195 KO mice treated with D-limonene was significantly higher than that in WT mice (Fig. 5h-5k). The above results suggest that the presence of miR-195 in WT mice facilitates the D-limonene’s suppression and attenuation on PM2.5-induced lipid-related gene expression and of pulmonary fibrosis development, respectively.
PM 2.5 exposure increased TG levels in human plasma and lipid droplets in the lung tissue of cancer patients, while D-limonene intervention upregulated the expression of miR-195 in human plasma.
We assembled a cohort of 11,712 lung cancer patients with information on serum TG levels and their residences where air monitoring data were available on ambient PM2.5. Based on their residence, 7,663 individuals were classified as living in high exposure areas, and 4,049 were in low exposure regions. More patients living in high exposure regions had higher TG levels than those living low exposure regions, 60% vs. 52% (Fig. 6a). Multivariate logistic regression analysis showed that high PM2.5 exposure were associated with elevated serum TG levels after adjusting for sex and age (HR = 1.39, 95% CI: 1.29ཞ1.50) (Fig. 6b). Compared to those in Shanghai City, lung cancer patients from Hebei Province where PM2.5 pollution was higher than Shanghai, had more lipid droplets in their lung tissues according to the Oil Red O staining (Fig. 6c). A single-arm D-limonene intervention trial was conducted in ten people. Over a 4-week oral administration of D-limonene, individuals in the trial had substantial increases in serum levels of miR-195 (Fig. 6d).