Although NF-κB is known to play a pivotal role in lung cancer, contributing to tumor growth, microenvironmental changes, and metastasis, the epigenetic regulation of NF-κB in tumor context is largely unknown. Here we report that the IKK2/NF-κB signaling pathway modulates metastasis-associated protein 2 (MTA2), a component of the nucleosome remodeling and deacetylase complex (NuRD). In triple transgenic mice, downregulation of IKK2 (Sftpc-cRaf-IKK2DN) in cRaf-induced tumors in alveolar epithelial type II cells restricted tumor formation, whereas activation of IKK2 (Sftpc-cRaf-IKK2CA) supported tumor growth; both effects were accompanied by altered expression of MTA2. Further studies employing genetic inhibition of MTA2 suggested that in primary tumor growth, independent of IKK2, MTA2/NuRD corepressor complex negatively regulates NF-κB signaling and tumor growth, whereas later dissociation of MTA2/NuRD complex from the promoter of NF-κB target genes and IKK2-dependent positive regulation of MTA2 leads to activation of NF-κB signaling, epithelial–mesenchymal transition, and lung tumor metastasis. These findings reveal a previously unrecognized biphasic role of MTA2 in IKK2/NF-κB-driven primary-to-metastatic lung tumor progression. Addressing the interaction between MTA2 and NF-κB would provide potential targets for intervention of tumor growth and metastasis.
These findings strongly suggest a prominent role of MTA2 in primary tumor growth, lung metastasis, and NF-κB signaling modulatory functions.
Lung cancer is a multifaceted disease in which inflammation plays an important role in tumor progression. Mutation of KRas occurs in 30% lung adenocarcinoma cases. Blasco and colleagues using KRas+/G12V mice concluded that cRaf expression was found to be essential for mediating KRas signaling and tumor formation. Several subsequent studies revealed a potential role for cRaf in lung development and carcinogenesis (1, 2).
Previous studies have shown that cRaf activates not only the mitogenic cascade but also NF-κB signaling either via an autocrine loop or even more direct through MEKK1 (3). NF-κB transcription factor dependent signaling demonstrated NF-κB as a key player in diverse immune and inflammatory responses and regulates a wide range of genes (4, 5), many of which play roles in neoplastic transformation. Indeed, aberrant NF-κB expression and activity are observed in many cancers (6). As to the NF-κB signaling, in unstimulated cells, the NF-κB1/RelA (p50/p65)—heterodimers are sequestered in the cytoplasm by a family of inhibitors called IκBs. Although, upon stimulation, activation of the NF-κB is initiated by the signal-induced degradation of IκB proteins, which occurs primarily via activation of a kinase called the IKK. IKK is composed of a heterodimer of the catalytic IKKα and IKKβ (also known as IKK1 and IKK2, respectively) subunits and a regulatory protein NEMO (NF-κB essential modulator) or IKKγ (7). Several studies have attempted the use of selective IKK2 inhibitors to modulate NF-κB activity since targeting NF-κB sensitizes cancer cells to chemotherapy in non–small cell lung carcinoma (NSCLC; ref. 8) and breast cancer (9). In support of these data, Bay11-7085, a selective IKK2 inhibitor, increases the therapeutic effectiveness of bortezomib, a proteasome inhibitor, for ovarian cancer (10). However, these attempts failed to reach final clinical trials owing to their off-target effects.
These studies emphasize the importance of acquiring more insight into the regulation of NF-κB. Increasing evidence has shown that NF-κB is regulated epigenetically, especially in the context of tumorigenesis. For example, members of NF-κB family can, themselves, act as epigenetic regulators of NF-κB–dependent gene activation; the homodimer of p50 interacts with HDAC1 to inhibit proinflammatory genes (11). This inhibition has also been reported with respect to lipopolysaccharide (LPS) tolerance (12). A study concerning acute myeloid leukemia suggests that C/EBP displaces HDACs from p50 homodimers and thus drives the activation of anti-apoptotic genes (13). Another mechanism for epigenetic regulation is via diverse chromosome regulatory machinery such as the nucleosome and remodeling deacetylase complex (NuRD) complex. The NuRD complex was first discovered in 1998 in different species. NuRD is critical for chromatin assembly, transcription, and genomic stability (14). Its main structural components are metastasis-associated proteins 1/2 (MTA1/2), retinoblastoma protein associated protein 46/48 (RbAp46/48); its enzymatic components are chromodomain helicase DNA binding protein 3/4 (CHD3/4) and HDAC1/2. In lung cancer, Liu and colleagues showed that nuclear MTA2 correlates with lung cancer cell proliferation, tumor size, and lymph node metastasis, making it a potential candidate as a prognostic marker (15). Moreover, MTA2 overexpression was correlated with tumor progression and poor prognosis in a study conducted on patients with esophageal squamous cell carcinoma (16). However, how MTA2/NuRD complex regulates NF-κB signaling pathway and how it is contributing to tumorigenesis remains to be elucidated.
Hitherto, the epigenetic regulation of NF-κB has been an intriguing conundrum. We thus investigated how MTA2, an essential member of the NuRD complex, regulates NF-κB and its role in tumorigenesis and the components of the tumor microenvironment.
Materials and Methods
Mice and animal experiments
The Sftpc-cRaf mouse was originally generated by Ulf Rapp (17). The Tet-O-IKK2CA and Tet-O-IKK2DN mice were obtained from Thomas Wirth (University of Ulm; ref. 18) and Sftpc-rtTA mice were from Jackson Laboratory. Tet-O-IKK2CA and Tet-O-IKK2DN mice were crossed with the Sftpc-rtTA mice to obtain cell-specific expression of modified IKK2 in alveolar type II cells. The double transgenic offspring mice were crossed with Sftpc-cRaf mice to alter IKK2 in adenoma-forming ATII cells. Genotypes were verified by PCR from genomic tail DNA as described previously (Supplementary Table S1 for genotyping primers; ref. 19). Transgenes were induced with doxycyclin (500 mg/kg body weight; Ssniff) food at the age of 5 weeks and until euthanization.
For the syngenic model, C57BL/6 mice were injected subcutaneously with 3 × 106 LLC1 cells, intravenously with 1 × 106 LLC1 cells, and tumor growth was monitored for 20 to 21 days. Tumor volume based on caliper measurements was calculated by the modified ellipsoidal formula (L × W2)/2 as described previously (20, 21). For flow cytometry, lung homogenates were digested with collagenase D and DNase for 30 minutes. After preparing single cell suspension by passing through 100 and 40 μm filters, cells were stained with different antibodies for immune cells (20). All mouse studies were carried out according to the protocols approved by German federal authorities for animal research (Regierungspräsidium Giessen and Darmstadt).
For isolation of RNA from tissues, 50 to 100 mg tissue was homogenized in 1 mL TRIzol (Qiagen) twice at 6,500 rpm for 25 seconds using Precellys Ceramic Kit 1.4 (PEQLAB) and Precellys homogenizer. Total RNA isolated from homogenized tumor samples or cell lines was reverse transcribed using the ImProm-II Reverse Transcription System (Promega) with Oligo-dT primers according to the suppliers' recommendation. qPCR was carried out in triplicates using iQTM SYBR Green Supermix (Bio-Rad) according to the manufacturer's recommendations. Results were analyzed for the relative expression of mRNAs normalized against hypoxanthine-guanine phosphoribosyl transferase and 18S. A complete list of primers used in the study is shown in Supplementary Table S1.
Cell culture and activity assays
All cell lines were obtained from the ATCC and grown in DMEM (A427, A549, HEK) or RPMI 1640 medium (H2122, H1650, Colo699, LLC1) supplemented with 10% FCS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37°C with 5% CO2. The cell line was authenticated by the manufacturer and checked for Mycoplasma using LookOut Mycoplasma PCR Detection Kit. The cells were verified to be mycoplasma negative and were used between passages 10 to 25. Transduced cell lines were maintained in culture medium containing 6 to 8 μg/mL puromycin. To stimulate NF-κB signaling, cells were treated with TNFα (10 ng/mL) for 1 hour unless otherwise stated. Cells were transiently transfected with different plasmids using Lipofectamine 2000 at a ratio of 1:3. The NF-Gluc reporter plasmid for NF-κB activity was a kind gift of Bakhos A. Tannous, and experiments using this plasmid were performed as described previously (22). Briefly, cells were transfected with NF-Gluc, and after 48-hour samples from the supernatant were collected and incubated with 20 μmol/L coelenterazine and chemiluminescence was measured. The Epigenase HDAC Activity/Inhibition Direct Colorimetric Assay Kit (Epigentek) was used to determine HDAC activity in nuclear extracts of cells or whole-cell lysates. An amount of 10 μg protein was added per well, which are coated with acetylated histone HDAC substrate. The deacetylated products were detected with a specific antibody conjugated with horseradish peroxidase (HRP), the activity of which was measured spectrophotometrically at 450 nm.
Stable cell generation
Stable cells were generated via lentiviral transfection for both MTA2 OE and MTA2 KD. First, HEK293T cells were cotransfected with pLKO-1 empty vector or pLKO-1 carrying shMTA2 (for MTA2 KD) or pCDH -CMV-MCS-EF1-copGFP-T2A-Puro empty vector or pCDH-MTA2 (for MTA2 OE) together with the viral packaging vector psPAX2 and envelope vector pMD2G at a ratio of 3:2:1 using TurboFect transfection reagent (Thermo Fisher Scientific Inc.) and serum-free medium. After 48 hours of transfection, cell medium containing the viral particles was collected to transduce target cells (LLC1 cells with mouse shMTA2 and A549 cells with human shMTA2 and pCDH-MTA2) with the lentiviral vector of interest using polybrene (Fermentas) at a final concentration of 0.8 μg/mL. Two days later, positive cells were selected with puromycin at a concentration of 6 to 8 μg/mL according to the cell line.
Cell proliferation by BrdU incorporation
Cell proliferation was quantified using colorimetric cell proliferation ELISA BrdU Kit (Roche) according to the manufacturer's protocol. After 24 hours of serum starvation, the medium was replaced with medium to be tested and incubated for as long as required for the stimulation or inhibition. BrdU was then added to the cells for 2 hours in a serum-free medium. Cells were then fixed, and BrdUrd incorporated in proliferating cells was detected with an antibody conjugated with HRP. After the addition of the substrate, the plate was measured at 370 nm with a reference measurement at 492 nm.
Invasive ability of the cells was measured using a transwell chamber (Falcon Cell Culture Insert; Corning Inc.) containing membranes with pores of 8 μm, which were initially coated with 100 μL/chamber Matrigel (VWR International). Cells were suspended in serum-free medium and 5 × 104 cells/membrane were seeded on top of the Matrigel. DMEM containing 10% FCS was added to the lower compartment as a chemo-attractant. Following incubation at 37°C in 5% CO2 for 20 hours, the noninvasive cells on the upper side of the membrane were removed using a cotton swab. The invasive cells on the lower surface were fixed in 100% methanol for 3 minutes and stained in 10% crystal violet for 10 minutes. After a final wash in distilled water, membranes were dried and mounted on slides with pertex. Slides were scanned using a NanoZoomer digital slide scanner C9600 (Hamamatsu Photonics) and invasive cells were quantified by ImageJ software.
Chromatin immunoprecipitation (ChIP) was performed as described previously (23) with slight modifications. Briefly, cells subjected to various treatments were fixed with formaldehyde at a final concentration of 1% for 10 minutes at room temperature. Glycine was then added at a final concentration of 125 mmol/L for 5 minutes to quench formaldehyde. Cross-linked cells were washed three times with cold 1× PBS and collected by scraping. Cell nuclei were separated and sonicated using Bioruptor Next gen (Diagenode) for 16 cycles at 90% power. The chromatin solution was collected, diluted with DB-dilution buffer supplemented with protease and phosphatase inhibitors, and then incubated with 80 μL protein A or G agarose beads conjugated with salmon sperm DNA for 1 hour at 4°C. Aliquots of DNA-protein samples were collected after centrifugation at 1,000 × g for 3 minutes and incubated overnight at 4°C with 2 μg of antibody or IgG in a final volume of 1.8 mL. From the DNA-protein sample, 10% to 20% were stored as input control. After incubation agarose beads were added and bound DNA-protein were washed with 800 μL of low salt washing buffer, then high salt washing buffer, then LiCl washing buffer and finally twice with TE buffer. DNA was cross-linked and eluted the DNA with 0.5 μL RNase A (1 mg/mL, Fermentas, Germany) at 37°C on thermo-shaker (900 rpm) for 2 hours followed by 5 μl of 10 mg/mL proteinase K (Sigma-Aldrich) at 56°C for 2 hours with shaking. Finally, NaCl was added to a final concentration of 0.2 mol/L and incubated overnight at 65°C. Finally, eluted DNA was diluted 1:4 with water and used for qPCR. (primers used for ChIP assay and their sequences and antibodies are listed in Supplementary Table S1).
Lungs or cells were lysed in ice-cold RIPA buffer (50 mmol/L Tris-HCl, pH 8.0, 150 mmol/L NaCl, 0.1% SDS, 0.5% deoxycholate, and 1% Nonidet P-40) with protease and phosphatase inhibitor cocktails (Roche) using a homogenizer in a cold environment. After centrifuging for 10 minutes at full speed at 4°C, the supernatant containing proteins was quantified using the BioRad Protein Quantification Kit. Protein samples were mixed with 5× SDS sample application buffer and boiled for 3 to 5 minutes, and then 20 μg protein was loaded on 10% SDS PAGE gels and transferred to nitrocellulose membranes. After blocking, membranes were incubated with primary antibodies and secondary antibodies. The antibodies used in the study are listed in Supplementary Table S1.
Cells grown on chamber slides were treated and fixed with acetone-methanol (1:1). After blocking with 5% BSA, cells were incubated with primary antibodies. Indirect immunofluorescence was conducted by incubation with Alexa Fluor 488/555-conjugated secondary antibodies (Supplementary Table S1). After incubation, slides were counterstained with 4,6-diamidino-2-phenylindole (DAPI, for nuclear staining) and mounted with fluorescent mounting medium (Dako Cytomation). At the end of the procedure, fluorescence images were acquired.
Fractionation of nuclear lysate by sucrose gradient
H1650 cells were cultured in 10-cm culture plates, two for each condition. After treatment with and TNFα for 6 hours, cells were washed twice with PBS and collected by scraping for subsequent centrifugation. Using 2 mL of hypotonic lysis buffer complemented with protease and phosphatase inhibitors, cell pellets were lysed by vortexing and incubation on ice for 10 to 15 minutes followed by centrifugation. Nuclear fractions were lysed by 450 μL native nuclear lysis buffer. A sucrose gradient was prepared by mixing 2 mL of 5%, 13.75%, 22.5%, 31.25%, and 40% of sucrose. Tubes were incubated overnight at 4°C. The nuclear samples were overlayed on top of the sucrose gradient and centrifuged at 37,5000 rpm for 16.5 hours. After centrifugation, the plastic tubes were gently pierced from the bottom to collect 500 μL fractions. The proteins collected were then examined by immunoblotting.
Histology, immunofluorescence staining, and IHC
Mouse lung tissues were collected at different times and fixed with 10% formalin, paraffin-embedded and sectioned for hematoxylin and eosin staining, IHC, and immunofluorescence staining. Antibodies are listed in Supplementary Table S1. For immunofluorescence, tissue sections were hydrated and incubated with trypsin for 10 minutes for antigen retrieval or with citrate for 30 minutes. Then, tissues were blocked in 5% BSA with 0.1% Triton-X in PBS for 1 hour at room temperature, to avoid unspecific binding of the antibodies. The following primary antibodies were used: RELA, C-Raf, PCNA, MTA2, and IKK2, all diluted 1:200. After incubation with the primary antibodies at 4°C overnight, slides were incubated with the corresponding Alexa Fluor-labeled secondary antibodies (Supplementary Table S1) at a dilution of 1:1,000, counterstained with DAPI (Life Technologies) and mounted with Dako fluorescent mounting media (Dako, North America, Inc.). Quantification of total fluorescent intensity was carried out by computer-aided image analysis using ImageJ software.
Statistical analyses were performed with GraphPad Prism 5 and 6 Software. Student t test (two-tailed) was used to compare two groups. When more than two groups were compared, differences among the groups were determined by one-way ANOVA with Tukey posttest for unpaired nonparametric variables. Data are expressed as mean ± SEM, and statistical significance was set at P ≤ 0.05.
IKK2 is required for NF-κB activation and subsequent tumor growth
KRAS, which is at the top of the mitogenic cascade, plays an essential role in many cancers. Activation of its downstream target, cRaf, is indispensable for the formation of lung adenomas (24). Hence, we used a mouse model of cRaf to dissect its function and possibly its regulation. In the Surfactant Protein C (Sftpc)-cRaf mouse model, alveolar type II (ATII) cells grow uncontrollably, leading to the formation of lung adenomas or hyperplasia (Fig. 1A). Lung homogenates of Sftpc-cRaf mice exhibited increased expression of genes involved in NF-κB signaling (Rela, Nfkb1, Tnf), regulator kinases Ikbka and Ikbkb (IKK1 and IKK2, respectively) as well as the NF-κB target genes (Mmp2, Mmp9, Fn1, Ccnd1), suggesting the activation of NF-κB (Fig. 1B). These results support that the canonical IKK2–NF-κB axis is one of the downstream signaling pathways activated by cRaf, which is thought to be crucial for the activation of KRAS signaling as well (25). To further explore this axis, a triple transgenic mouse model was established to alter the Inhibitor of NF-κB kinase subunit beta (IKK2; constitutive activation, CA, and dominant-negative inhibition, DN) in ATII cells harboring the Sftpc-cRaf transgene (Fig. 1C). Mice with dominant-negative IKK2 (Sftpc-cRaf-IKK2DN) had fewer tumors as quantified by hematoxylin and eosin staining image analysis and by magnetic resonance imaging (MRI), whereas IKK2 activation in ATII cells (Sftpc-cRaf-IKK2CA) promoted tumor formation owing to increased cell proliferation as indicated by cRaf and PCNA expression (Fig. 1D and F). In accordance, compliance of the respiratory system (Crs), tissue damping, and elastance also improved upon downregulation of IKK2 (Fig. 1E). Furthermore, the effects of IKK2 modulation also caused changes in the lung immune microenvironment. IKK2 downregulation largely suppressed the infiltration of macrophages, T cells, neutrophils, and dendritic cells into the lung, whereas IKK2 activation enhanced the accumulation of these cells (Supplementary Fig. S1).
MTA2 play a potential role in IKK2-driven lung metastasis
As tumorigenesis was markedly inhibited in Sftpc-cRaf-IKK2DN lungs, we next investigated the impact of IKK2 upregulation and downregulation in epithelial cells. The double transgenic mouse lines with altered IKK2 in ATII cells (Sftpc-IKK2CA and Sftpc-IKK2DN; Supplementary Fig. S2A) were subcutaneously injected with LLC1 lung cancer cells. Notably, constitutive activation of IKK2 created a favorable niche for LLC1 metastasis to the lung; more metastatic nodules were observed in Sfptc-IKK2CA lungs as compared with wild-type (WT) and Sfptc-IKK2DN lungs (Fig. 2A and B). Further, ATII cells were isolated from Sfptc-IKK2DN lungs and screened for differential expression of major genes known to play a role in cancer using RT2 profiler PCR array. Both NF-κB and non-NF-κB target genes were regulated in Sfptc-IKK2DN ATII cells as compared with WT ATII cells, which includes both NF-κB (e.g., Rela, Nfkb1, and Mmp9) and non-NF-κB target genes (e.g., Mta2 and Tek; Fig. 2C; Supplementary Table S2). In addition, we performed mRNA expression profiling of the above-mentioned genes from laser microdissected (LMD) tumors from WT, Sftpc-cRaf, Sftpc-cRaf-IKK2CA, and Sftpc-cRaf-IKK2DN mice. The expression of Mta2, Mmp9, and Tek increased in Sftpc-cRaf and Sftpc-cRaf-IKK2CA as compared with WT, whereas only Mta2 showed a significant reduction in Sftpc-cRaf-IKK2DN as compared with Sftpc-cRaf-IKK2CA (Fig. 2D; Supplementary Fig. S2B). Similar to known NF-κB target genes (Rela, Nfkb1, and Mmp9), expression of Mta2 in lung homogenates (Fig. 2E; Supplementary Fig. S2C) and immunofluorescence staining in lungs (Fig. 2F) confirmed upregulation of Mta2 in Sftpc-cRaf and Sftpc-cRaf-IKK2CA and downregulation in Sftpc-cRaf-IKK2DN mice. To understand the role of MTA2 in IKK2-driven lung metastasis, LLC1 MTA2 KD cells were generated (Supplementary Fig. S2D) and injected in Sfptc-IKK2CA and Sfptc-IKK2DN mice. The injection of LLC1 MTA2 KD cells showed a significant reduction in metastatic tumor nodules in both Sfptc-IKK2CA and Sfptc-IKK2DN mice compared with LLC1 control cells (Fig. 2G and H), indicating that MTA2 plays a potential role in IKK2-driven lung metastasis.
Knockdown of MTA2 has an opposite influence on primary and metastatic tumor growth
To study the impact of MTA2 on tumor growth, human (A549, H1650) and mouse (LLC1) lung cancer cells with stable overexpression (MTA2 OE) and knockdown (MTA2 KD) of MTA2 were established by transfection with plasmids encoding MTA2 cDNA and lentivirus-mediated short hairpin RNA, respectively. A549 and H1650 MTA2 OE cells increased mRNA expression of MTA2, no effect was observed MTA1 and MTA2 (Fig. 3A; Supplementary Figs. S3A and S3B). Overexpression of MTA2 in A549 decreased their cell growth and proliferation (Fig. 3B) by decreasing proliferation markers at mRNA level (PCNA, MKI67) and protein level (PCNA, pERK, CCNB1, and CCND1; Fig. 3C; Supplementary Fig. S3C). Further overexpression of MTA2 in H2122, H1650, and Colo699 lung cancer cell lines also showed decreased proliferation (Supplementary Fig. S3D). On the other hand, MTA2 overexpression increased invasion of cells as compared with controls (Fig. 3D) by decreasing epithelial markers (CDH1 and KRT18), while increasing expression of epithelial markers (FN1 and VIM) at mRNA and protein level in both A549 and H1650 cells (Fig. 3E and F; Supplementary Figs. S3E–S3G) obviously suggesting MTA2 involvement in epithelial–mesenchymal transition (EMT) processes. Opposite effects of proliferation and EMT phenotype were observed in A549 and H1650 MTA2 KD cells (Fig. 3G–L; Supplementary Figs. S3H–S3L). Similarly, increased proliferation and decreased invasion were observed upon knockdown of MTA2 in mouse lung cancer syngeneic LLC1 cells (Supplementary Fig. S3M). These findings support a differential role of MTA2 on primary tumor growth and metastasis in vitro.
Further, LLC1 MTA2 KD cells were subcutaneously and intravenously injected into syngeneic C57BL/6 (WT) mice to understand the role of MTA2 in vivo primary and metastatic tumor growth, respectively. Similar to Sfptc-IKK2CA (Fig. 2G and H), intravenous injection of LLC1 MTA2 KD cells in WT mice reduced metastatic tumor growth (Fig. 4A; Supplementary Fig. S4A).
Interestingly, subcutaneous injection of LLC1 MTA2 KD cells showed increased tumor growth compared with LLC1 cells injected to in WT mice (Fig. 4B; Supplementary Fig. S4B), suggesting that in the absence of constitutive activation of IKK2, MTA2 restricts primary tumor growth either by alteration in the tumor microenvironment or due to direct effect on tumor cell specific NF-κB signaling. The analysis of immune cells by FACS in LLC1 control and LLC1 MTA2 KD subcutaneous tumor revealed that MTA2 KD tumors showed increased infiltration of macrophages, but T cells, neutrophils, and dendritic cells did not show any significant alteration in MTA2 KD tumors compared with control tumors (Supplementary Fig. S4C). Interestingly, MTA2 KD tumors showed upregulation of NF-κB signaling (Nfkb1, Rela), NF-κB target genes (Nfkbia, Mmp9, Tnf), and proliferation genes (Ccnd1, Myc), indicating negative regulation of NF-κB signaling by MTA2 (Fig. 4C and D).
As evidences suggests that, in the course of tumor progression, increased inflammation and reactive oxygen species (ROS) can regulate NF-κB signaling as well as trigger an oxidative stress-induced response, we explored whether increased inflammatory and/or oxidative stress conditions (stimulating with TNFα and H2O2) can explain opposite roles exerted by MTA2 in primary tumor growth and metastasis (26, 27). In corroboration, the short stimulation (2 hours) of A549 cells with two different concentrations of TNFα (10 and 100 ng) and H2O2 (25 and 250 μmol/L) altered the expression of NF-κB signaling genes as well as epithelial–mesenchymal markers. A549 cells treated with a higher concentration of TNFα (100 ng) and H2O2 (250 μmol/L) increased the expression of IKK2, RELA, MTA2, NF-κB target genes (MMP9, CCND1) and mesenchymal marker (FN1), whereas decreased the expression of the epithelial marker (CDH1) compared with control cells and A549 cells treated with a low concentration of TNFα and H2O2 (Supplementary Fig. 4D). Altogether, these results suggest that, in the progression of IKK2-driven primary to metastatic tumor growth, MTA2 may play a biphasic role in IKK2–NF-κB signaling.
MTA2 plays a biphasic role in IKK2–NF-κB signaling
To understand the contribution of MTA2 to IKK2–NF-κB signaling, we stimulated A549 cells with TNFα for different time durations. Upon TNFα stimulation, (i) the expression of IKK2 and RELA increased starting from 1 hour (Fig. 5A and B), (ii) the expression of MTA2 increased from 6 hours (Fig. 5A and B), and (iii) the expression of NF-κB target genes (MMP9, FN1, NFKB1, CCND1) drastically increased from 6 hours (Fig. 5C; Supplementary Fig. S5A). Notably, MTA2 was identified as one of the overexpressed genes in the metastatic lungs of Sfptc-cRaf-IKKCA and Sfptc-IKKCA lungs. Therefore, to understand IKK2-mediated MTA2 modulation and vice versa, we performed siRNA-mediated knockdown of IKK2 or MTA2 in unstimulated (0 hours), 1-hour TNFα-stimulated (time point at which expression of IKK2, RELA increased but not the MTA2 and NF-κB target genes, 1 hour) and 6-hour TNFα-stimulated A549 cells (time point at which expression of all the genes are increased, 6 hours). In si_IKK2 cells as compared with si_NS (control), the expression of IKK2 and RELA was downregulated at all time points of TNFα stimulation, but the expression of MTA2 and the target genes of NF-κB (MMP9, CCND1, FN1) were downregulated only after 6 hours stimulation with TNFα (Fig. 5D–F; Supplementary Fig. S5B). Wherein, in si_MTA2 cells as compared with si_NS (control), expression NF-κB target genes showed an increase in the absence (0 hours) and during short-term (1 hour) stimulation with TNFα and decrease after 6 hours stimulation with TNFα (Fig. 5G–I; Supplementary Fig. S5C). Altogether these results suggest that—(i) in the absence (0 hours) and during short-term (1 hour) stimulation with TNFα, MTA2 negatively regulates target genes, without regulating the expression of IKK2, whereas (ii) during the long-term (6 hours) stimulation with TNFα, IKK2–RELA–MTA2 axis positively regulates the expression of target genes, indicating that MTA2 plays a biphasic role in IKK2–NF-κB activation.
Early IKK2–NF-κB activation via TNFα led to the release of MTA2/NuRD-repressing complex from NF-κB target genes
As MTA family members, together with the other components of the NuRD complex, are known to interact and modulate the expression of several genes, it was interesting to investigate whether MTA2 binds to NF-κB members through the NuRD complex. At first, we compared the response of A549 and H1650 to TNFα, which demonstrated that H1650 shows a more robust NF-κB activation compared with A549 (Fig. 6A). Therefore, to obtain more precise results, we performed nuclear fractionation in H1650 cells to assess how the composition of the multiprotein MTA2/NuRD complex and NF-κB members are altered by TNFα stimulation. Nuclear fractionation using sucrose gradient identified members of the NuRD complex (HDAC1, HDAC2, CHD4, and RbAp46) and NF-κB in the same fractions, and importantly an additional HDAC1 band appears in fractions 7 to 15 after stimulation with TNFα (Fig. 6B), suggesting an acetylated form of HDAC1. Acetylation of HDAC1, which is known to be inactivating (28), was confirmed in samples treated with TNFα as seen after pull-down with an antibody that recognizes acetylated proteins (Fig. 6C). HDAC1 activity was significantly reduced in cells treated with TNFα (fractions 7–15, Fig. 6D). Moreover, higher NF-κB activity was detected in fractions 7 to 15 after TNFα treatment (blue bars in Fig. 6E). A proximity ligation assay in A549 cells confirmed the direct interaction between MTA2 and RELA. This interaction was stronger upon MTA2 overexpression, whereas MTA2 knockdown abolished the interaction (Fig. 6F). Correspondingly, upon shorter TNFα stimulation to A549 cells, stronger MTA2-RELA interaction was observed in nuclei (Fig. 6G).
Finally, to determine the dynamics of MTA2 binding at the consensus NF-κB binding sites to promoters of NF-κB target genes (NFKB1 proximal, NFKB1 upstream, NFKBIA, MMP9, CCND1) during the short- and long-term TNFα stimulation, we performed ChIP at the time points mentioned above (0, 1, and 6 hours). Interestingly, both short- and long- term stimulation with TNFα leads to enrichment of RELA and NFKB1 (Fig. 7A; Supplementary Fig. S6A) and the release of MTA2 and other members of the NuRD complex, like RbAp46 and HDAC1 from the promoters of target genes (Fig. 7B; Supplementary Figs. S6B and S6C), suggesting a positive regulation of IKK2–NF-κB signaling by MTA2 occurs in the long term TNFα stimulation. At this point, it is worth mentioning that, in the absence of TNFα stimulation, NF-κB reporter activity in A549 cells was reduced with MTA2 overexpression, whereas knockdown of MTA2 increased NF-κB reporter activity (Supplementary Figs. S7A and S7B).
Further, to confirm the causative role of MTA2 in mediating IKK2–NF-κB signaling during short- and long-term TNFα stimulation, we performed overexpression or knockdown of MTA2 in the presence of TNFα at different time points mentioned above, followed by ChIP analysis. In A549 MTA2 KD, RELA was enriched at the promoters of NF-κB target genes (NFKB1 proximal, NFKB1 upstream, NFKBIA, MMP9, CCND1) only in the absence (0 hours) and during short-term (1 hour) stimulation with TNFα compared with A549 control cells (Supplementary Fig. S7C). On the contrary, in A549 MTA2 OE, occupancy of RELA was decreased and occupancy of MTA2 was increased at the promoters in the absence (0 hours) and during short-term (1 hour) stimulation with TNFα (Supplementary Figs. S7D and S7E). These results confirmed that the positive regulation of IKK2–NF-κB signaling by MTA2 occurs only in the long-term TNFα stimulation.
These findings reveal a previously unrecognized tumor-suppressive role as well as NF-κB signaling modulatory function of MTA2. This concept, summarized in the diagram in Fig. 7C, is based on the findings that IKK2 alteration in ATII cells (both constitutive activation and dominant negative inhibition) in the presence of Sftpc-cRaf–induced tumor burden regulate the lung immune microenvironment and tumor growth; MTA2 was identified as a non-NF-κB target gene regulated by IKK2 and MTA2 is regulated according to the level of NF-κB activation within cancer cells. Under unstimulated conditions, MTA2 negatively regulates NF-κB pathway by forming the MTA2/NuRD corepressor complex and interacting with RelA at the promoters of NF-κB target genes, leading to decreased primary tumor growth (panel 1). However, the IKK2 activation mediated by short-time exposure to inflammatory or growth factor stimuli, does not regulate the expression of MTA2, but releases the MTA2/NuRD complex from the promoters thereby increasing the expression of NF-κB target genes that subsequently promotes primary tumor growth, emphasizing the biphasic role of MTA2 in IKK2–NF-κB signaling-mediated lung cancer progression (panel 2). On the other hand, in the presence of long-term exposure to inflammatory or growth factor stimuli, IKK2 not only increases the expression of NF-κB target genes, but also the expression of MTA2. Importantly, the upregulation of MTA2 regulates the expression of epithelial and mesenchymal markers, thereby promoting EMT and lung metastasis (panel 3).
Inflammation and cancer were linked more than a century ago when Rudolf Virchow identified leukocyte infiltration in neoplastic tissues (29). Many years of research followed until the discovery of NF-κB, and this paved the way for more insight into the contribution of inflammatory signaling to cancer development. NF-κB is not only involved in innate and adaptive immune responses, but it governs the expression of a wide set of diverse genes regulating neoplastic transformation (30, 31). Many tumors, especially solid tumors, exhibit constitutive activation of NF-κB canonical signaling pathway. This fact encouraged several researchers to develop NF-κB inhibitors and IKK2 inhibitors as an adjunct therapy to existing cancer treatment modalities. However, success is presently modest at best.
Our results in transgenic mice where cRaf is activated in epithelial cells (ATII) reveal an upregulation of the NF-κB pathway and activation of NF-κB pathway via IKK2 is shown to control survival, inflammation, and proliferation in the context of tumorigenesis. This was observed in several other models; in a mouse model of precancerous pancreatic lesions, Daniluk and colleagues showed that oncogenic Ras initiates a positive feedback loop of NF-κB activation that further pathologically activates Ras (32). Furthermore, increased NF-κB activation is well documented in most cancer cell lines, mice, and human tumors (33). In concordance with our results, a study by Xia and colleagues reported that IKK2 depletion in mouse lungs leads to reduced tumor proliferation mediated by a reduction in Timp-1 expression (34).
In addition, we observed changes in MTA2 levels and nuclear translocation, a noncanonical NF-κB target gene, upon IKK2 modulation. This is in line with previous studies suggesting that IKK2 not only regulates NF-κB but also has NF-κB- independent effects. Chariot and colleagues reported that IKK subunits in an NF-κB independent manner could phosphorylate transcription factors such as FOXO3 and p53 (35). As p53 binding sites are present at the promoter region of MTA2, suggesting p53 as a potential pathway regulating MTA2 gene (26). Moreover, p53 regulates miR-146a and miR-34a, the microRNAs that can posttranscriptionally regulates MTA2 (36, 37). In addition, our in vitro experiments with H2O2 suggest that oxidative stress/inflammatory stimuli lead to upregulation and/or nuclear exclusion of MTA2. Although similar stimuli were shown to stimulate the expression of MTA1 and subsequently the regulation of EMT (38), no underlying mechanisms have been ascribed. Thus, future studies are warranted to dissect the MTA proteins regulation by IKK2.
Surprisingly, both in vitro and in vivo data showed that MTA2 exerts opposite functions in lung cancer development as it initially inhibited primary tumor growth but later favored metastasis. There are limited publications showing the effect of silencing of MTA2 on tumor growth, Lu and colleagues showed that shMTA2 reduced both primary and metastatic tumor growth in a human breast cancer model with MDA-MB231 cells (39). It is worth mentioning that MDA-MB231 cells are highly aggressive and poorly differentiated and are derived from a metastatic site. This divergence between our data and that of Lu and colleagues might also be attributed to different mouse models that are used (nude mice versus syngeneic mice) and the influence of IKK2– NF-κB–MTA2 axis on tumor microenvironment, which is lacking in the nude mouse model used in the study of Lu and colleagues
These contrasting and novel roles of MTA2 in primary tumor growth and lung metastasis can partly explained by sustained inflammatory response at the later phases of tumor growth. ROS increase by growth factors or ongoing inflammation as well as oxidative stress-induced response, at the later phases of tumor growth, may result in IKK2-mediated upregulation of MTA2 and release of MTA2 from the promoters of NF-κB target genes, resulting in increased expression of NF-κB signaling associated inflammation and proliferative genes, EMT, and metastasis formation (27).
Importantly, we identified MTA2 as a novel regulator of NF-κB pathway and as well identified a biphasic role of MTA2 in NF-κB pathway modulation: (i) inhibiting the NF-κB pathway by forming the MTA2/NuRD corepressor complex at the promoters of NF-κB target genes in unstimulated conditions; and (ii) a subsequent function as an activator of NF-κB pathway by releasing the MTA2/NuRD corepressor complex from the promoters of NF-κB target genes in the presence of growth factor or inflammatory stimuli. Notably, stimulation experiments with inflammatory stimuli such as TNFα suggest that the release of MTA2/NuRD corepressor complex is regulated by IKK2 and is dependent on the duration and concentration of the inflammatory stimuli. These findings provide not only new mechanistic and functional links between NF-κB and the MTA2/NuRD complex but also establish new essential roles for the components of MTA2/NuRD complex in primary tumor growth. Interestingly, long-term stimulation with TNFα via IKK2 upregulates the expression of MTA2 that subsequently leads to EMT and lung metastasis. In accordance, several studies correlated the overexpression of MTA2 with poor prognosis in different cancer types, like lung cancer (40), colorectal cancer (41), and estrogen receptor-negative breast cancer (42). In our model, MTA2 was identified as one of the overexpressed genes in SpC-cRaf lungs. It is noteworthy that in animal mouse models, mutations of both Kras and Tp53 in the lung lead to the formation of aggressive metastatic lesions though Kras mutations alone merely formed adenocarcinomas (43, 44), whereas SpC-cRaf mutation leads to the formation of benign adenomas. These suggest that MTA2 overexpression is a later event during tumorigenesis. Along similar lines, previous studies demonstrate that the MTA2/NuRD complex is essential to the mediation of TWIST-controlled cancer cell invasion and metastasis (39, 45).
In conclusion, our data strongly suggest prominent role of MTA2 in primary tumor growth, lung metastasis, and NF-κB signaling modulatory functions. Thus, addressing the interaction between MTA2/NuRD complex and NF-κB would provide potential targets for intervention of tumor growth and metastasis. As the NuRD complex contains HDAC subunits, HDAC inhibitors may represent one potential therapeutic avenue for targeting the interaction between MTA2/NuRD complex and NF-κB. A recent study demonstrated that HDAC inhibitors have a selective preference for different types of HDAC complexes (46), suggesting that targeting specific HDAC complexes may be feasible with enzyme inhibitors. However, it remains unclear whether one could selectively target tumor-promoting activities while sparing tumor-suppressive functions with this class of drugs. The other therapeutic option is to screen for the drugs modulating the activity or interactions of MTA2/NuRD complex and NF-κB and may represent a more selective approach.
Disclosure of Potential Conflicts of Interest
I. Singh reports grants from Deutsche Forschungsgemeinschaft (DFG, Bonn, Germany) through Emmy Noether program (SI 2620/1-1) during the conduct of the study. R. Dammann reports grants from German Center for Lung Research during the conduct of the study, grants from German Center for Lung Research, and other item from Beckmann Research Institute (RASSF1 patent) outside the submitted work. G. Barreto reports personal fees from Université Paris-Est Créteil (UPEC, Créteil, France) and grants from Deutsche Forschungsgemeinschaft (DFG, Bonn, Germany; grant no. BA 4036/4-1) during the conduct of the study, personal fees from Hawkeye Bio, Inc. (scientific advisor) outside the submitted work; in addition, G. Barreto has a patent for US Patent number 10501803 issued and licensed to Hawkeye Bio, Inc. (principal inventor of the intellectual property protected by the US patent with the number 10501803). No potential conflicts of interest were disclosed by the other authors.
N. El-Nikhely: Conceptualization, data curation, validation, investigation, methodology and writing-original draft. A. Karger: Data curation, investigation and methodology. P. Sarode: Data curation, investigation, methodology, writing-original draft, writing-review and editing. I. Singh: Data curation, investigation, methodology, writing-review and editing. A. Weigert: Conceptualization, resources, data curation, methodology, writing-original draft, writing-review and editing. A. Wietelmann: Data curation, software, investigation, visualization and methodology. T. Stiewe: Resources, funding acquisition, writing-review and editing. R. Dammann: Resources, funding acquisition, methodology, writing-review and editing. L. Fink: Resources, investigation, visualization and methodology. F. Grimminger: Conceptualization, resources, funding acquisition, project administration, writing-review and editing. G. Barreto: Conceptualization, resources, supervision and methodology. W. Seeger: Conceptualization, resources, funding acquisition, writing-original draft, project administration, writing-review and editing. S.S. Pullamsetti: Conceptualization, resources, formal analysis, funding acquisition, writing-original draft, project administration, writing-review and editing. U.R. Rapp: Conceptualization, resources, methodology, project administration, writing-review and editing. R. Savai: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, validation, investigation, methodology, writing-original draft, project administration, writing-review and editing.
R. Savai, S.S. Pullamsetti, W. Seeger, and U.R. Rapp were supported by the Max Planck Society. R. Savai, S.S. Pullamsetti, T. Stiewe, F. Grimminger, R. Dammann, G. Barreto, W. Seeger, and U.R. Rapp were supported by the UGMLC and Loewe Program. U.R. Rapp received a grant from German Cancer Aid (Deutsche Krebshilfe; 109081). R. Savai and S.S. Pullamsetti received Von-Behring-Röntgen-Stiftung (projects 66-LV06 and 63-LV01). R. Savai, S.S. Pullamsetti, W. Seeger, and F. Grimminger were supported by Cardio-Pulmonary Institute (CPI). R. Savai, S.S. Pullamsetti, T. Stiewe, R. Dammann, W. Seeger, and F. Grimminger were supported by the German Center for Lung Research (DZL). R. Savai and S.S. Pullamsetti were supported by the German Research Foundation (DFG) by the CRC1213 (Collaborative Research Center 1213), projects A01 (to S.S. Pullamsetti), A05 (to S.S. Pullamsetti), A10* (to R. Savai). S.S. Pullamsetti received European Research Council (ERC) Consolidator Grant (866051). We thank Yanina Knepper, Jeanette Knepper, Marianne Hoeck, Vanessa Golchert, for excellent technical assistance. We thank Prof. Thomas Wirth for providing Tet-O-IKK2DN and Tet-O-IKK2CA mice. We thank Prof. Bakhos A. Tannous for the NF-Gluc Luciferase reporter plasmid for NF-κB activity. We thank Siavash Mansouri for the cell growth experimental help.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.