Ataxia-telangiectasia mutated activation mediates tumor necrosis factor-alpha induced MMP-13 up-regulation and metastasis in lung cancer cells

Despite that ataxia-telangiectasia mutated (ATM) is involved in IL-6 promoted lung cancer chemotherapeutic resistance and metastasis, the exact role of ATM in tumor necrosis factor-alpha (TNF-α) increasing tumor migration is still elusive. In the present study, we demonstrated that TNF-α promoted lung cancer cell migration by up-regulation of matrix metalloproteinase-13 (MMP-13). Notably, by gene silencing or kinase inhibition, we proposed for the first time that ATM is a key up-stream regulator of TNF-α activated ERK/p38-NF-κB pathway. The existence of TNF-α secreted in autocrine or paracrine manner by components of tumor microenvironment highlights the significance of TNF-α in inflammation-associated tumor metastasis. Importantly, in vivo lung cancer metastasis test showed that ATM depletion actually reduce the number of metastatic nodules and cancer nests in lung tissues, verifying the critical role of ATM in metastasis. In conclusion, our findings demonstrate that ATM, which could be activated by lung cancer-associated TNF-α, up-regulate MMP-13 expression and thereby augment tumor metastasis. Therefore, ATM might be a promising target for prevention of inflammation-associated lung cancer metastasis.

Here, we found that TNF-α promote lung cancer metastasis by MMP-13 up-regulation and ATM activation. ATM inhibition largely decrease TNF-α augmented lung Research Paper cancer cell migration. The existence of TNF-α secreted in autocrine or paracrine manner by component of tumor microenvironment highlights the significance of TNF-α in inflammation-associated tumor metastasis. Importantly, in vivo lung cancer metastasis test showed that ATM depletion actually reduce the number of metastatic nodules and cancer nests in lung tissues, verifying the critical role of ATM in lung cancer metastasis. Therefore, ATM might be a promising target for prevention of inflammationassociated lung cancer metastasis.

TNF-α level has a positive correlation with cell migration in lung cancer cells
A549, LTEP-a-2 and NCI-H520 cells possess stronger migration abilities than NCI-H446 and NCI-H1299 cells [11]. To explore the effect of TNF-α on cell migration, we firstly determined TNF-α level in a panel of lung cancer cells. As predicted, the higher TNF-α level was revealed in A549, LTEP-a-2 and NCI-H520 cells ( Figure 1a). When NCI-H446 or NCI-H1299 cells were replenished with TNF-α, the migration abilities increased accordingly (Figure 1b). A significant repression of cell migration of A549, LTEP-a-2 and NCI-H520 cells was achieved when TNF-α was inhibited by inhibitor ( Figure  1c) or gene silencing ( Figure 1d). As cell viabilities had not been affected by above treatments (Supplementary Figure S1), our data demonstrate that there is a positive correlation between TNF-α level and cell migration in lung cancer cells.

TNF-α up-regulate the expression and the activity of MMP-13
Since there is a difference of MMP-13 in expression and activity between NCI-H446 and A549 cells (Supplementary Figure S3a), we therefore explored TNFα's effect on MMP-13 expression. As shown in Figure 3, TNF-α potently increased MMP-13 expression in both protein ( Figure 3a) and mRNA (Figure 3b) level. MMP-13 activity was also increased upon TNF-α's stimulation ( Figure 3c). Moreover, TNF-α depletion achieved a significant MMP-13 reduction (Figure 3d-3e). Apart from MMP-13, the expression or activity of MMP-1, MMP-2, MMP-3 and MMP-9 were increased by TNF-α treatment as well (Supplementary Figure S3b-S3d). Collectively, above data support that TNF-α promote cell migration by up-regulation of MMPs expression and activity.

ATM is involved in TNF-α inducing ERK/p38-NF-κB pathway activation in lung cancer cells
Owing to TNF-α' effects on ATM, ERK, p38 and p65 activation (Supplementary Figure S4a-S4d), the inhibition of ERK or p38 decreasing TNF-α induced p65 activation (Supplementary Figure S4e) indicate that ATM might play a potential role in TNF-α induced ERK/p38-NF-κB activation. To address this issue, ATM inhibition was performed and TNF-α induced ERK/p38-NF-κB activation was evaluated. As showed in Figure

The components of lung cancer microenvironment could secret TNF-α in autocrine or paracrine manner upon LPS or chemotherapeutic drug stimulation
As main components of tumor microenvironment, both mouse and human immune cells efficiently produce TNF-α in low concentration of LPS (Figure 6a-6d) or chemotherapeutic agents (Figure 6e-6h) treated condition. TNF-α IHC determination of lung cancer metastasis test showed that TNF-α expression in cancer nests and surrounding tissues was decreased by the depletion of ATM (Figure 6i). Other groups also found that TNF-α was highly expressed in human non-small cell lung carcinoma [23][24]. These results suggest that TNF-α might be an important component in tumor microenvironment to trigger inflammation associated tumor metastasis.

Figure 4: ATM mediate TNF-α inducing ERK/p38-NF-κB activation in lung cancer cells. Cells were conferred ATM
silencing a-c, e. or CGK733 treatment d, f. prior to TNF-α (2 ng/ml) stimulation and the phosphorylation of ERK, p38 or p65 was investigated by confocal fluorescence microscope (a-c) or western blot (d-f). One representative from three experiments is shown. CGK: CGK733. The immunoblots were cropped to improve the clarity and conciseness of the presentation. www.impactjournals.com/oncotarget

ATM inhibition decrease lung cancer metastasis in vivo
Tail-vein-injected mouse model was used to assess lung cancer metastasis in vivo [25]. As shown in Figure 7a, the number of white-tan nodules of metastatic tumor counted by naked eyes dropped when ATM was knocked down. Hematoxylin and eosin staining in Figure 7b showed that: a large number of normal alveolar structures were destroyed and replaced by tumor cells with prominent and irregular nuclei, more cancer nests were observed in control group. Conversely, more normal alveolar structures and a significant drop of the number of cancer nests were found in ATM depletion condition. These observations confirm that high level of ATM promote lung cancer metastasis in vivo.   (d, h). The concentration of LPS for splenocytes, PBMC, Raw264.7, DCs was 100, 50, 50, 20 ng/ml respectively. Data are presented as the mean±SEM, n=3. **p<0.01; ***p<0.001, Student t test or one-way ANOVA with post Newman-Keuls test. One representative from three experiments is shown. SP: Splenocytes; DCs: dendritic cells. i. 8×10 5 NCI-H520 cells conferred ATM silencing were transferred to BALB/c nude mice (5-6 weeks old) through tail vein (n=3 per group) and the lungs were performed TNF-α IHC staining. www.impactjournals.com/oncotarget DISCUSSION TNF-α and IL-6 are the major pro-inflammatory cytokines to promote metastasis [11,22,[26][27][28][29]. In the present study, we demonstrated that the activation of ATM-ERK/p38-NF-κB, which induced by TNF-α in autocrine or paracrine manner, up-regulate MMP-13 expression and thereby augment lung cancer metastasis. Most importantly, ATM inhibition potently repress the lung cancer metastasis in vivo. Therefore, ATM might be a promising target for prevention of inflammation-associated lung cancer metastasis.
TNF-α, which expressed by the components of tumor microenvironment, was functionally controlled by other pro-inflammatory cytokines in paracrine manner [30]. In our observation, TNF-α level has a positive correlation with metastasis of lung cancer, indicating that tumor cells itself might regulate metastasis in autocrine manner. As lung carcinoma epithelium cell with K-ras mutation have higher level of TNF-α, the higher level of pro-inflammatory cytokines might be due to the overall effects of oncogene activation. Apart from endogenous TNF-α, in our study, LPS and chemotherpeutic agents up-regulate TNF-α expression, indicating that TNF-α can act as a paracrine cytokine in malignancies.
Recently, we demonstrated that IL-6 is critical for lung cancer metastasis [11] and chemotherapeutic Figure 7: ATM inhibition decrease lung cancer metastasis. In vivo lung cancer metastasis model was established as described in Figure 6i. After sacrifice, the intact mouse lungs were dissociated to count the white-tan nodules of metastatic tumor by naked eyes a. and then embedded to perform H&E staining b. For H&E staining, the number of cancer nests in lungs was counted under microscope with 100×magnification; 24 slides per condition (b). www.impactjournals.com/oncotarget resistance [12]. Interestingly, our current data revealed that IL-6 could be well induced by TNF-α treatment (Supplementary Figure S7a-S7d). Moreover, IL-6 deficiency abrogated TNF-α's effects on cell migration (Supplementary Figure S7f-S7g) and MMPs expression (Supplementary Figure S7h), indicating that TNF-α regulate IL-6 expression [32][33]. Considering that IL-6 triggering ATM activation facilitate lung cancer metastasis via MMP-3/MMP-13 up-regulation, the above data suggest that TNF-α promoting lung cancer metastasis mostly, if not all, is depend on the upregulation of IL-6.
ATM, an upstream regulator of NF-κB, play a critical role in IL-6 increasing lung cancer metastasis and chemotherapeutic resistance [11][12]. Erk1/2 and p38 were also documented to be up-stream regulators of NF-κB signaling in response to TNF-α treatment [34][35][36][37][38]. In this study, ATM inhibition abolish the effect of TNF-α on ERK/p38-NF-κB activation, unveiling that ATM is a key up-stream signal molecule in NF-κB activation. Further exploration is needed to focus on the cross-talk among inflammatory mediators and cellular effectors, which will insight the mechanism of initiation and development of inflammation associated cancers.
We therefore propose a novel mechanism for TNF-α augmented lung cancer cell migration. In this model, ATM is the key up-steam regulator of TNF-α activated ERK/ p38-NF-κB pathway, which play a vital role in promoting lung cancer cell migration by up-regulating MMP-13 expression (Figure 8). Inactivation of TNF-α-ATM-ERK/ p38-NF-κB decrease lung cancer metastasis, providing advanced evidences for ATM inhibitor usage in lung cancer treatment.

Animals
Pathogen-free C57BL/6 or BALB/c nude mice (female, 6-8 weeks old) were bought from Shanghai Laboratory Animal Center of Chinese Academy of Sciences (China) and kept at the Animal Center of Xiamen University. The protocol was approved by the Committee on the Ethics of Animal Experiments of the Xiamen University.

Cell culture and cell lines
Human small cell lung cancer (SCLC) NCI-H446 cells, non-small cell lung cancer (NSCLC) NCI-H1299 cells, lung adenocarcinoma LTEP-a-2 cells and squamous cell carcinoma NCI-H520 cells and Raw264.7 cells were obtained from Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). Human lung carcinoma A549 cells were kindly provided by Professor GH. Jin (Xiamen University). All the cells were grown in RPMI-1640 or DMEM medium containing 10% FBS.

Bone marrow-derived murine DCs
Bone marrow-derived immature dendritic cells (imDCs) were prepared as previously described [39]. Briefly, bone marrow mononuclear cells were prepared from bone marrow suspensions of C57BL/6 mice by depletion of red cells and then were cultured at a density of 1×10 6 cells/ml in RPMI 1640 complete medium with 10 ng/ml of GM-CSF and 1 ng/ml of IL-4. Non-adherent cells were gently washed out on day 4 of culture; the remaining loosely adherent clusters were used as im-DCs and were further stimulated with LPS, DDP or CPT.

Murine splenocytes preparation
Spleens isolated from C57BL/6 mice were washed in PBS and ground into single cells. Then the cell suspension was re-suspended followed by depletion of red cells and cultured at a density of 1×10 6 cells/ml in RPMI 1640 medium for further treatment.

Preparation of human peripheral mononuclear cells (PBMCs)
Briefly, human PBMC was prepared from the donor's blood by gradient density centrifugation using PBMC isolation reagents and cultured at a density of 1×10 6 cells/ml in RPMI 1640 medium for further treatment. Healthy subjects were selected according to the Declaration of Helsinki principles and signed an informed consent.

RT-qPCR
Endogenous TNF-α expression of lung cancer cells and MMPs expression were investigated by RT-qPCR analysis as previously described [11] The PCR primer sequences were described in Supplementary Table S1.

ELISA
To detect the release of TNF-α or IL-6 in corresponding lung cancer cell lines and determine MMP-3/MMP-13 activity, enzyme double-antibody indirect immunoassays with respective ELISA kits was performed in accordance with manufacturer protocol [40].

Western blot
Cells were treated with TNF-α (2 ng/ml) for indicated periods or pretreated with inhibitors or siRNA prior to TNF-α stimulation. The expressions of indicated proteins were determined via western blot analysis. ß-actin was used as a loading control.

Transwell migration assay
Cell migration ability was determined via transwell migration assay as previous description [11,22].

Confocal immunofluorescence assays
The effects of TNF-α or ATM siRNA on ATM, p65, ERK or p38 phosphorylation levels was investigated using immunofluorescence assays as previous description [12].

Flow cytometric measurements
The effects of LPS, DDP or CPT on TNF-α expression were assayed by flow cytometric assays [12].

In vivo lung cancer metastasis model
Cell migration ability was determined via murine lung cancer metastasis model [25]. Briefly, NCI-H520 cells transfected with control or ATM siRNA were inoculated into BALB/c nude mice through the tail vein. 27 days after the injection, the mice were sacrificed and the lungs were dissociated and preserved for further studies. Each experiment had three mice per condition.

H&E and immunohistochemistry staining
In vivo cell migration ability and the expression of TNF-α in lung tissues were determined via H&E and Immunohistochemistry staining as described before [40].

Statistical analysis
All experiments were repeated at least three times to confirm the similar results. Data were presented as the mean ± SEM. Student's t test or one-way ANOVA with the post Newman-Keuls test was applied. Statistical differences were considered to be significant at p<0.05.