The different response pattern between lung epithelial cells and lung cancer cells after radiation revealed by RNA-sequencing (RNA-seq)
To determine whether ionizing radiation has different biological effects on normal lung cells and cancer cells, we treated the lung epithelial cell line BEAS-2B, non-small cell lung cancer cell line A549, and small cell lung cancer cell line H446 with 10 Gy X-ray radiation, and RNA-seq was performed. The results showed 120 differentially expressed genes in BEAS-2B cells, 3278 differentially expressed genes in H446 cells, and 5295 differentially expressed genes in A549 cells, with only 17 of these genes commonly expressed in all cell lines, revealing the potential different response patterns between lung epithelial cells and lung cancer cells after radiation (Fig. 1A, B). GSEA revealed that lung epithelial cells strongly expressed genes involved in the inflammatory response and apoptosis pathways, while lung cancer cells highly upregulated genes associated with DNA damage repair (Fig. 1C, D). We further analyzed the top 44 genes enriched in inflammatory response pathway in irradiated lung epithelial cells, and protein–protein interaction analysis showed that the NLRP3/IL-1β axis was at the center (Fig. 1E, F). To determine whether the NLRP3 inflammasome is associated with pulmonary fibrosis, the expression of NLRP3 was analyzed in serial sections of organizing pneumonia and normal lung tissue from the patients. We found that NLRP3 was overexpressed in organizing pneumonia tissues compared to normal lung tissues (Fig. 1G). These results indicate a correlation between NLRP3/IL-1β activation and inflammation post-radiation.
Radiation activates NLRP3 inflammasome canonical pathways in lung epithelial cells
To estimate the changes in the levels of NLRP3 inflammasome proteins after radiation, qRT-PCR and western blotting analyses of NLRP3 inflammasome canonical pathway proteins were performed. We found that the mRNA levels of NLRP3 inflammasome components, including NLRP3, CASP1 and IL-1β, as well as the pyroptosis marker gasdermin D (GSDMD), were significantly upregulated after radiation in lung epithelial cells, and that radiation increased the protein expression levels of NLRP3, ASC, cleaved CASP1, and cleaved GSDMD in a radiation dose-dependent manner, while NLRP3 expression was not upregulated in lung cancer cells (Fig. 2A, B, Additional file 1: Fig. S1). NLRP3 interacts with ASC to initiate inflammasome assembly. We found that radiation also induced the colocalization of NLRP3 with ASC (Fig. 2C). The IL-1β levels in the supernatants of irradiated BEAS-2B, H446, A549, and H460 cells at different time points were detected by ELISA, and the results showed that irradiated BEAS-2B cells secreted significantly higher amounts of IL-1β compared with irradiated lung cancer cells, which was most evident 48 h post radiation. The change in IL-1β levels correlated with the increase in the radiation dose (Fig. 2D). However, the change in IL-18 levels was very modest in lung epithelium compared with lung cancer cells, indicating a more profound role of IL-1β in lung epithelium-induced inflammation (Additional file 1: Fig. S2).
To determine whether the activation of CASP1 is dependent on NLRP3, BEAS-2B cells were transfected with NLRP3-silencing siRNA. Transfection efficiency was analyzed using western blotting and qRT-PCR. NLRP3 inhibited both the activation of CASP1 and secretion of IL-1β secretion (Fig. 2E, F, Additional file 1: Fig. S3). To further investigate the correlation between NLRP3 inflammasome, lung epithelial cells, and RILI, PHBE were isolated from human bronchial segments. As shown in Fig. 2G, PHBE cells had a typical cobblestone appearance and presented positive immunostaining for the epithelial marker cytokeratin and negative immunostaining for the fibroblast marker α-SMA. Consistent with the above findings, the NLRP3 inflammasome was activated in PHBE after radiation, and the secretion of IL-1β was significantly increased (Fig. 2H, I). Taken together, these results indicate that radiation induces NLRP3 inflammasome activation and pyroptosis in the lung epithelial cells.
Since NLRP3 may be the key molecule in RILI as well as in tumorigenesis and progression, we treated A549 and H446 cells with MCC950, a specific small-molecule inhibitor that selectively blocks NLRP3. The result revealed the inhibition of the proliferation ability of both H446 and A549 by MCC950 (Additional file 1: Fig. S4), indicating that the suppression of NLRP3 may inhibit tumor growth and alleviate the progression of RILI.
Radiation induced NLRP3 inflammasome activation relies on glycolysis related ROS accumulation
Radiation-induced normal tissue injury is a dynamic and progressive process that begins with ROS production and accumulation [1]. ROS can cause DNA damage, and protein and lipid oxidization, followed by a series of biological processes that may lead to cell death and inflammation [26]. We found that ROS levels were significantly increased in the lung epithelial cell lines BEAS-2B and HBE in a radiation dose-dependent manner (Fig. 3A, Additional file 1: Fig. S5). To investigate whether ROS accumulation in lung epithelial cells triggered NLRP3 inflammasome activation, we used N-acetyl-l-cysteine (NAC), to treat lung epithelial cells prior to radiation, and found that NAC largely scavenged intracellular ROS in lung epithelial cells (Fig. 3A, Additional file 1: Fig. S5A). NAC pretreatment abolished radiation-induced NLRP3 inflammasome activation, CASP1 cleavage, and IL-1β secretion in BEAS-2B and PHBE cells, suggesting that ROS accumulation may trigger NLRP3 inflammasome activation in lung epithelial cells (Fig. 3B, H, Additional file 1: Fig. S5B).
According to the RNA-seq results, several pathways involved in metabolic processes, were enriched in irradiated lung epithelial cells (Fig. 3C). To investigate the metabolic alterations in lung epithelial cells caused by radiation, we first examined the intracellular levels of ATP and NADP(H) redox couples, which are involved in cellular energy metabolism and the maintenance of redox balance. We found that ATP levels increased, and the NADPH/NADP(H)total ratio decreased after radiation in BEAS-2B cells (Fig. 3D). We then investigated the glycolysis level after radiation at different time points and found that the lactic acid level in supernatants was significantly increased after radiation, while the glucose concentration was decreased, suggesting upregulated glycolysis in irradiated epithelial cells (Fig. 3E).
Studies have shown that p53 plays a vital role in antioxidant response and cell metabolism [27]. GSEA revealed that the p53 pathway and glycolysis were enriched in irradiated BEAS-2B cells (Fig. 3C). Among the core enriched genes, the most significant differentially expressed gene was dihydropyrimidinase-like 4 (DPYSL4) (Fig. 3C), which has been reported to be p53-inducible and is associated with OXPHOS and cellular energy supply [28]. Therefore, we examined the expression levels of DPYSL4 after radiation. We found that DPYSL4 was significantly upregulated in BEAS-2B and PHBE cells post-irradiation (Fig. 3F). Next, we sought to determine the function of DPYSL4 in irradiated epithelial cells. DPYSL4 silencing resulted in a lower ATP level and a higher NADPH/NADP(H)total ratio than in the control group after radiation (Fig. 3F, G). Since NADP+/NADPH redox couples are essential for maintaining cellular redox homeostasis [29], we examined the cellular ROS levels in DPYSL4-silencing or small interfering control lung epithelial cells. We found that silencing DPYSL4 greatly increased cellular ROS levels either in the presence or absence of radiation (Fig. 3I), suggesting a protective effect of DPYSL4 post radiation by regulating cellular energy metabolism. Consistent with the above findings, silencing DPYSL4 increased IL-1β secretion in BEAS-2B and PHBE cells, possibly due to increased cellular ROS levels (Fig. 3H).
To explore the role of ROS in RILI, the ROS levels in irradiated lung tissue of mice were detected. The results showed that the ROS level was significantly increased at day 7 post radiation and remained at high level until 5 months after radiation compared with that of unirradiated mice, indicating that ROS was involved in the whole process of RILI (Additional file 1: Fig. S6A). Moreover, immunohistochemistry revealed that DPYSL4 was greatly upregulated in bronchus 7 days after radiation and continued until 5 months after radiation (Additional file 1: Fig. S6B).
NLRP3 inflammasome activation triggers fibroblast mature and migration through IL-1β
A broad spectrum of molecules, including extracellular matrix (ECM) components, collagens, and ECM modulators, matrix metalloproteinases (MMPs), and tissue inhibitors of metalloproteinases (TIMPs), have been reported to participate in the process of lung fibrosis. To determine the effect of IL-1β on RPF, the human fibroblast cell line, 2BS, was used. qRT-PCR and western blotting were performed to observe the expression of collagens, TIMPs, and MMPs, as well as the fibroblast activation marker α-SMA. The results showed that IL-1β induced the expression of type I collagens, TIMP-1, MMP-3, and α-SMA in fibroblasts (Fig. 4A, B), indicating the effect of IL-1β on fibroblast activation and lung tissue remodeling. To validate these results, primary human lung fibroblasts (HLF) and mouse embryonic fibroblasts (MEF) were extracted (Additional file 1: Fig. S7A). Consistent with the results obtained from cell line 2BS, we found that IL-1β promoted the synthesis of type I collagen, α-SMA, TIMP-1, and MMP-3 in HLF and MEF (Fig. 4B, Additional file 1: Fig. S7B). To ascertain that the epithelium-derived IL-1β was responsible for the observed phenotype change, the culture medium of BEAS-2B was collected after 48 h of incubation post radiation, then added to 2BS and co-cultured for 24 h. Similar to the treatment with IL-1β, the conditioned medium significantly induced the mRNA expression of COL1A1, COL1A2, TIMP-1, MMP-3, and α-SMA (Fig. 4C). To further confirm whether NLRP3 inflammasome activation in irradiated lung epithelial cells mediated the response of fibroblasts, BEAS-2B cells were transfected with NLRP3, IL-1β siRNA, or control siRNA and then treated with 10 Gy radiation. The conditioned medium was collected 48 h later and co-cultured with 2BS or HLF. Both silencing NLRP3 and IL-1β in BEAS-2B cells significantly suppressed the activation of fibroblasts, and silencing IL-1β eliminated the effect of co-culturing, indicating that other pathways that activate IL-1β may be involved in fibroblast activation. If IL-1β was added to the conditioned medium of irradiated IL-1β-silencing BEAS-2B, the activation effect was restored (Fig. 4D). Similarly, the conditioned medium of irradiated PHBE greatly promoted the activation of HLF, further confirming the above results (Fig. 4D).
Next, we investigated whether IL-1β affected the migration and proliferation of fibroblasts. Transwell assays demonstrated that IL-1β facilitated the migration ability of 2BS, HLF, and MEF (Fig. 4E, Additional file 1: Fig. S7C). The EdU experiment showed that IL-1β stimulated the proliferation of 2BS, HLF, and MEF (Fig. 4F, Additional file 1: Fig. S7D). Taken together, our results suggest that lung epithelial cells activate the NLRP3/IL-1β pathway after irradiation and induce the activation, proliferation, and migration of fibroblasts.
NLRP3 inflammasome was activated and associated with RPF in vivo
To further analyze the effect of the NLRP3 inflammasome in RILI, 6-week female C57Bl/6 mice were administered a single thoracic 16 Gy radiation dose to construct RILI mouse models. We evaluated the expression of the NLRP3 inflammasome in lung tissue 7 days after radiation. Radiation triggered an increase in NLRP3 and GSDMD expression in irradiated lung tissue, as revealed by western blotting (Fig. 5A). BALF and lung tissues were obtained at different time points after radiation. The IL-1β levels in BALF were significantly increased on day 7, which persisted over 21 days, while the change in IL-1β levels in serum did not reach statistical significance (Fig. 5B). IHC analysis of the lung tissue showed that NLRP3 was significantly upregulated in mouse bronchial tissue on days 1, 3, 7, 14, and 28 after radiation (Fig. 5C). Two months after radiation, Masson’s staining revealed lung tissue remodeling and collagen deposition around the bronchus, which was more evident 5 months after radiation. Furthermore, NLRP3 immunohistology remained positive 5 months post-radiation, indicating that it may participate in the whole process of RILI (Fig. 5D).
Clodronate liposomes were used to deplete the lung macrophages of mice to rule out the effect of macrophages. The depletion efficiency was confirmed by flow cytometry (Fig. 5E), and the mice were administered 16 Gy thoracic radiation. IHC analysis of the lung tissue demonstrated that NLRP3 was still positive in the bronchus after macrophage depletion (Fig. 5F). We found that IL-1β levels in the BALF of macrophage-depleted mice were significantly higher than that of control mice. Furthermore, IL-1β levels in the serum were augmented in macrophage-depleted mice, which was not observed in the mice of the control group, suggesting that macrophages may play an immunosuppressive role in the early phase of RILI (Fig. 5G). To further validate the effect of IL-1β, we used IL-1R−/− mice that received 16 Gy of thoracic irradiation (Additional file 1: Fig. S8). Lung tissue was collected after 5 months. Masson’s staining revealed a significantly higher amount of collagen deposition and more severe fibrosis in wild-type mice than in IL-1R−/− mice (Fig. 5H), confirming that targeting IL-1β is a promising strategy to alleviate RILI.
MCC950 has been proposed as a specific small molecule inhibitor that can selectively block NLRP3 inflammasome activation [30]. Previous study has shown that MCC950 treatment ameliorates RP and decreases cytokine production including IL-6, IL-18, and IL-1β [10]. We further explored whether MCC950 could improve RPF. Lung of irradiated mice were collected after 5 months. Masson’s staining showed that MCC950 significantly alleviated lung fibrosis induced by radiation (Fig. 5I). These findings suggested that MCC950 was a promising drug for the prevention of RPF.
Here, we demonstrated that after radiation, the NLRP3 inflammasome is activated in lung epithelial cells, where DPYSL4 regulated glucose metabolism alteration is related to intracellular ROS accumulation and is thus associated with NLRP3 inflammasome activation. IL-1β secreted by lung epithelial cells promotes the activation, proliferation, and migration of fibroblasts as well as promotes the process of lung tissue remodeling, thereby promoting the progression of RILI (Fig. 6).