TGFβ-Induced Lung Cancer Cell Migration Is NR4A1-Dependent

Lung cancer is the leading cause of cancer-related deaths in the United States and it is estimated that in 2017, 222,500 new cases of lung cancer will be diagnosed in this country and 155,780 patient deaths will be observed (1). More than 85% of lung cancers are classified as non–small cell lung cancer (NSCLC) and despite significant advances in treatment regimens, the overall survival rate of patients with NSCLC is 15.9% and this rate has not significantly improved over the past decade (2). Smoking is the major risk factor for lung cancer and exposure to secondary smoke, various occupational exposures, air pollution, and genetic factors also contribute to the high incidence of this disease. NSCLC is a highly complex disease with multiple subtypes and histologies that are accompanied by mutations of oncogenes (i.e., EGFR, KRAS, EML4-ALK) and tumor suppressor genes (i.e., p53; refs. 3–6) and lung cancer therapy is driven, in part, by the tumor type and its pathologic and molecular characteristics and traditional surgery, radiation, and combinations of cytotoxic and mechanism-based drugs are extensively used (3–7). Targeted therapies for treating lung cancer have had limited success, and the more recent development and applications of immunotherapeutics that target programmed cell death ligand 1 (PD-L1) and programmed cell death 1 (PD-1) are promising new approaches (8, 9). Despite the advances in lung cancer chemotherapy, the improvement in patient survival remains low and most therapies are accompanied by unwanted side-effects and drug resistance. Thus, it is critical to develop new therapeutics that target multiple prooncogenic pathways and can be used in combination therapies.

The TGFβ family of ligands and receptors play an important and somewhat paradoxical role in cancer in which TGFβ acts as an inhibitor of early-stage cancers, but acts as a tumor promoter for later stage cancers. Several studies report that TGFβ induces lung cancer cell migration/invasion and EMT, and this involves multiple kinases and downstream targets (10–19). A recent study showed that TGFβ-induced migration/invasion of triple-negative breast cancer cells was also NR4A1-dependent where NR4A1 interacts with Arkadia, AXIN2, and RNF12 to induce proteasome-dependent degradation of SMAD7, resulting in TGFβR1/TGFβR2 homodimerization and activation (20). We have also confirmed that NR4A1 plays a key role in breast cancer invasion where TGFβ induces nuclear export of NR4A1 that interacts with E3 ligase complex proteins to induce SMAD7 ubiquitination and degradation (20, 21). We previously reported that NR4A1 was a negative prognostic factor for lung cancer patient survival and NR4A1 was a prooncogenic factor regulating lung cancer cell proliferation and survival (22), and this has also been observed in cell lines derived from other solid tumors (23–30). Structure–activity studies among a series of 1,1-bis(3′-indolyl)-1-(substituted phenyl)methane compounds showed that some of these analogues bound NR4A1 and in cancer cell lines, acted as NR4A1 antagonists (22–30). The most active compound 1,1-bis(3′-indolyl)-1-)p-hydroxyhenyl)methane (CDIM8; DIM-C-pPhOH), which acts as a nuclear NR4A1 antagonist (29) in lung and other cancer cell lines, inhibited NR4A1-dependent prooncogenic genes/pathways (22–30). We hypothesized that DIM-C-pPhOH would also inhibit TGFβ-induced lung cancer cell migration/invasion, and our results show for the first time that TGFβ-induced invasion of lung cancer cells is due to JNK1-dependent phosphorylation and nuclear export of NR4A1 that is inhibited by NR4A1 antagonists.

Lung cancer cell lines (A549, H460, and H1299) were purchased from ATCC. A549 cells were maintained 37°C in the presence of 5% CO₂ in DMEM/Ham F-12 medium with 10% FBS with antibiotic H460, and H1299 lung cancer cells were maintained in RPMI1640 medium with 10% FBS and antibiotics. Alexa Fluor 488 and 455, Hoechst 33342, leptomycin B, SP600125, SB202190, LY294002, and PD98059 were obtained from Cell Signaling Technology, and TGFβ was purchased from BD Biosystems. DMEM, 14–22 Amide PKA inhibitor, ALK5i inhibitor (LY-364947), and 36% formaldehyde were purchased from Sigma-Aldrich and hematoxylin was purchased from Vector Laboratories. The antibodies and their sources are summarized in Supplementary Table S1. FLAG-NR4A1, FLAG-NR4A1-(A-B), and FLAG-NR4A1-(C-F) were synthesized in the laboratory using site-directed mutagenesis (31); pcDNA3-FLAG-MKK4WT, pcDNA3-FLAG-MKK7-JNK1A1WT [MKK7(CA)], and pcDNA3-FLAG-MKK7-JNK1A1APF [(MKK7(DN)] were purchased from Origene Technologies. pCMV5-FLAG-SMAD7 was a gift from Lin and colleagues (Department of Biochemistry, Hong Kong University of Science and Technology, Kowloon, Hong Kong, China).

A549, H460, and H1299 lung cancer cells (3.0 × 10⁵ per well) were seeded in DMEM/Ham F-12 medium supplemented with 2.5% charcoal-stripped FBS and were allowed to attach for 24 hours. After various treatments including knockdown of various genes (48 hours), cells were allowed to migrate for 24 hours, fixed with formaldehyde, and then stained with hematoxylin, and cells migrating through the pores were then counted as described previously (31).

RNA was isolated using Zymo Research Quick-RNA MiniPrep Kit. Quantification of mRNA (Slug, Snail, and NR4A1) was performed using Bio-Rad iTaq Universal SYBER Green 1-Step Kit using the manufacturer's protocol with real-time PCR. TATA binding protein (TBP) mRNA was used as a control to determine relative mRNA expression.

A549 cells were transfected with various constructs and, 6 hours after transfection, cells were treated with DMSO or various agents and immunoprecipitation experiments and subsequent analysis were carried out described previously (31).

The chromatin immunoprecipitation (ChIP) assay was performed using the ChIP-IT Express Magnetic Chromatin Immunoprecipitation Kit (Active Motif) according to the manufacturer's protocol. The treatment conditions and analysis were performed as described previously (31). The primers for detection of the NR4A1 promoter region were 5′-CCTGCCCTCGGGAAGG-3′ (forward) and 5′-CAGGCCGCGGGCTGAGG-3′ (reverse). PCR products were resolved on a 2% agarose gel in the presence of RGB-4103 GelRed Nucleic Acid Stain.

Lung cancer cells were treated with various agents/constructs, and nuclear and cytosolic fractions were isolated using Thermo Scientific NE-PER Nuclear and Cytoplasmic Extraction Kit according to the manufacturer's protocol. Fractions were analyzed by Western blots as described previously (31). GAPDH and p84 were used as cytoplasmic- and nuclear-positive controls, respectively.

A549 cells (1.0 × 10⁵ per well) were treated with either DMSO or TGFβ (5 ng/mL) was added for 4 hours after ± pretreatment with various agents or transfection for 48 hours. Cells were then fixed with 37% formalin, blocked, treated with fluorescent NR4A1 primary antibody [Nur77 (D63C5) XP] for 24 hours. Cells were then washed with PBS and treated with anti-rabbit IgG Fab2 Alexa Fluor 488 secondary antibody for 3 hours. Cells were then treated with Hoechst (Hoechst 33342) stain and phalloidin (Alexa Fluor 555 Phalloidin) for 15 minutes in the dark following manufacturer's protocol and visualized by confocal microscopy (Zeiss LSM 780 confocal microscope) as described previously (31) Cells were analyzed by Western blot analysis as described previously (24–28).

siRNA experiments were conducted as described previously (21). The siRNA complexes used in the study that were purchased from Sigma-Aldrich are as follows: siGL2-5′: CGU ACG CGG AAU ACU UCG A; siNR4A1(1): SASI_Hs02_00333289; siNR4A1(2): SASI_Hs02_00333290; siAxin2(1): SASI_Hs01_ 00110148; siAxin2(2): SASI_Hs01_00110149; siArkadia(1): SASI_Hs01_ 00064840; siArkadia(2): SASI_Hs01_00064841; siRNF12 (1): SASI_Hs01_00238255; siRNF12(2): SASI_Hs02_ 00348888; siTAK1(1): SASI_Hs02_00335227; siTAK1(2): SASI_ Hs01_00234777; siTAB1(1): SASI_Hs01_00094398; siTAB1(2): SASI_Hs02_00340933; siTRAF6(1): SASI_Hs01_00116391; siTRAF6(2): SASI_Hs01_00116390; siMKK4(1): SASI_Hs02_ 00334897; siMKK4(2): SASI_Hs02_00334898; siMKK7(1): SASI_Hs01_00059905; siMKK7(2): SASI_Hs01_00059906; siJNK1(1): SASI_Hs01_00010441; siJNK1(2): SASI_Hs01_ 00010442; sic-fos (1): SASI_Hs01_00184572; sic-fos(2): SASI_ Hs01_00184573; siATF2(1): SASI_Hs01_00147372; siATF2(2): SASI_Hs01_00147373; siElk1(1): SASI_Hs02_00326325; siElk1(2): SASI_Hs02_00326324. The following siRNA complexes that were used in this study were purchased from Santa Cruz Biotechnology: c-jun siRNA(h): sc29223; c-jun siRNA(h2): sc-44201 SRF siRNA: sc-36563; PKACα siRNA: sc-36240.

A549 lung cancer cells (3.0 × 10⁵ per well) were seeded in DMEM/Ham F-12 medium supplemented with 2.5% charcoal-stripped FBS and were allowed to attach for 24 hours. Cells were then treated with above described treatments as used in other assays, then lysed with PKA lysis buffer (made in the laboratory using the manufacturer's recipe). PKA activity assay (Promega) was performed following manufacturer's protocol, and then lysates were resolved on a 2% agarose gel.

Statistical significance of differences between the treatment groups was determined as described previously (31).

In this study, we initially used three NSCLC cell lines (A549, H460, and H1299) to investigate the role of NR4A1 in TGFβ-induced migration/invasion using a Boyden Chamber assay. TGFβ-induced migration of the three cell lines (Fig. 1A) and cotreatment with the NR4A1 antagonist CDIM8, the nuclear export inhibitor leptomycin B (LMB), the TGFβ receptor inhibitor ALK5i, or knockdown of NR4A1 by RNA interference (RNAi; siNR4A1) significantly inhibited the TGFβ-induced cell migration. The results also showed that CDIM8 and siNR4A1 also inhibited basal migration of the lung cancer cell lines. TGFβ also induced nuclear export of NR4A1 in A549, H460, and H1299 lung cancer cells (Fig. 1B–D, respectively), which was inhibited by LMB and CDIM8. We also observed that TGFβ induced both expression and phosphorylation (S351) of NR4A1 and this was inhibited by cotreatment with CDIM8 and LMB (Fig. 1E). The intracellular location of NR4A1 in these experiments was determined by Western blots of nuclear and cytosolic extracts using GAPDH (cytosolic) and P84 (nuclear) as subcellular controls.

We also examined the effects of kinase inhibitors on TGFβ-induced cell migration and the JNK inhibitor SP600125, but not p38MAPK (SP202190), p42/44MAPK (PD98059), or PI3K (LY294002) inhibitors blocked TGFβ-induced migration of A549, H460, and H1299 cells (Fig. 2A). SP600125 also inhibited TGFβ-mediated nuclear export of NR4A1 in A549 (Fig. 2B), H460 (Fig. 2C), and H1299 (Fig. 2D) cells, whereas cotreatment with SB202190, LY294002, or PD98059 did not inhibit nuclear export of NR4A1 in cells treated with TGFβ, indicating that TGFβ-induced nuclear export of NR4A1 was JNK-dependent in lung cancer cells. TGFβ also induced phosphorylation of (S351) NR4A1, JNK1, c-jun, and c-fos, which was also inhibited by SP600125, demonstrating that TGFβ induces JNK and genes downstream from JNK.

Because the TGFβ–JNK–NR4A1 (nuclear export) pathway is critical for enhanced migration of lung cancer cells, we used A549 cells as a model to further investigate the role of upstream kinases in this pathway. MKK4 and MKK7 are upstream from JNK1, and overexpression of FLAG-MKK4 (wild-type) enhanced invasion of A549 cells and this was inhibited by LMB, CDIM8, and SP600125 (Fig. 3A). Overexpression of MKK4 also induced nuclear export of NR4A1 and this was inhibited by LMB, CDIM8, and SP600125 (Fig. 3B). Overexpression of FLAG-MKK7(CA) also induced A549 cell migration, which was inhibited by LMB, CDIM8, and SP60012, and TGFβ-induced migration was inhibited by a dominant negative FLAG-MKK7(DN) (Fig. 3C). MKK7 overexpression also induced nuclear export of NR4A1 which was inhibited by LMB, CDIM8, SP600125, and dominant-negative MKK7 (Fig. 3D). We also observed that TGFβ-induced migration in A549 cells was inhibited by transfecting a construct expressing MKK7(DN) and also by knockdown of JNK1 (siJNK1) and upstream kinases including MKK4 (siMKK4), and MKK7 (siMKK7), TRAF6 (siTRAF6), TAK1 (siTAK1), and TAB1 (siTAB1) (Fig. 3E). TGFβ-induced nuclear export of NR4A1 was also inhibited in A549 cells transfected with siJNK, siMKK4, and siMKK7, confirming that the intact MKK4/7-JNK pathway is required for NR4A1 nuclear export (Fig. 3F). We also observed that knockdown of the upstream kinases TAB1, TAK1, and TRAF6 inhibited TGFβ-induced nuclear export of NR4A1 and the loss of TAB1 increased levels of nuclear NR4A1 (Fig. 3G). TRAF6 potentially plays a role in activation of PKA, such as recruitment of PKA to the plasma membrane or enhance dissociation of regulatory subunits of PKA. TRAF6 is K63 polyubiquitinated and forms signaling cascades that activate MAPK (like JNK1). Knockdown efficiencies are indicated in Fig. 3H. Knockdown studies were performed using at least two different oligonucleotides (see Materials and Methods).

We also investigated the effects of TGFβ, MKK4, and MKK7(CA) alone and in various combinations with TGFβ, LMB, SP600125, and CDIM8, and also MKK7(DN) alone and in combination with TGFβ by immunostaining and confocal microscopy. In DMSO-treated cells, NR4A1 was primarily nuclear and this was significantly decreased after treatment with TGFβ, whereas TGFβ-mediated nuclear export of NR4A1 was inhibited after cotreatment with LMB, CDIM8, and SP600125 or transfected with MKK7(DN) (Supplementary Figs. S1 and S2). Overexpression of FLAG-MKK4 and FLAG-MKK7(CA) also induced nuclear export of NR4A1 as determined by confocal microscopy and this response was inhibited after cotreatment with LMB, CDIM8, and SP600125 (Supplementary Fig. S2).

TGFβ induces NR4A1 in breast cancer cells, (31) and this was also observed in lung cancer cells (Fig. 1E) and the mechanism of this response was further investigated. Treatment of A549 cells with TGFβ for 5 hours induced a >10-fold increase in NR4A1 mRNA levels and these effects were inhibited after cotreatment with CDIM8, SP600125, or ALK5i (TGFβ receptor inhibitor), but not LMB (Fig. 4A) or after transfection with siJNK1, siMKK4, and siMMK7 (Fig. 4B). The effects of TGFβ alone on induction of NR4A1 protein were minimal (Fig. 3B and F) as were the effects of LMB on this response, and this was in contrast to the induction of NR4A1 gene expression by TGFβ (Fig. 4A). Downstream targets of JNK, such as a c-jun, c-fos, ATF2, Elk-1, and SRF, bind AP1 (c-jun, c-fos), CRE (c-jun, ATF2), and SRE (Elk-1, SRF) promoter elements, and CRE and SRE sites were identified within the NR4A1 promoter (Fig. 4C). Therefore, we used a ChIP assay to investigate association of c-Jun/ATF2 and Elk-1 and SRF with the CRE/SRE motifs using primers that cover the −807 to −703 region of the NR4A1 promoter. A549 cells were treated with TGFβ, transfected with FLAG-MKK4 or FLAK-MKK7(CA) alone and this resulted in recruitment of c-jun, ATF2, and SRF and also Pol II to the promoter and Elk-1 was constitutively bound in control (DMSO) cells (Fig. 4D). CDIM8 and SP600125 but not LMB blocked recruitment of c-jun, ATF2, and SRF to the NR4A1 promoter, and the result for LMB correlated with its effect (or lack thereof) on TGFβ-induced levels of NR4A1 mRNA (Fig. 4A). ChIP assay results in cells transfected with FLAG-MKK7(DN) showed that TGFβ-induced recruitment of c-Jun, ATF2 and SRF to the promoter was blocked by the DN plasmid (Fig. 4D). We did not detect any c-fos bound to the NR4A1 promoter, which is consistent with the fact that no putative AP1 promoter elements were identified within the promoter. The time-dependent activation of JNK phosphorylation by TGFβ was also accompanied by activation and/or induction of phosphorylated ATF2, SRF, c-jun, and Elk-1 (Fig. 4E) and these results are consistent with recruitment of these factors to the NR4A1 promoter as results of the ChIP assay (Fig. 4D). TGFβ-induced NR4A1 mRNA (Fig. 4F) and protein (Fig. 4G) expression was inhibited in A549 cells after knockdown of c-jun, ATF2, SRF and Elk-1 but not c-fos (Fig. 4F) and this complemented results of the ChIP assay. There was some off-target variability in this experiment; for example, knockdown of c-jun also resulted in decreased c-fos expression and, loss of Elk-1 and SRF increased levels of c-jun and this may indicate some interactions and crosstalk between of these transcription factors. Because c-jun, ATF2, and SRF are recruited to the NR4A1 promoter and regulate expression of NR4A1, we also observed that their loss (by RNAi) also resulted in decreased TGFβ-induced migration (Fig. 4H).

TGFβ-induced activation of CRE and CRE binding factors suggests that protein kinase A (PKA) may also be activated by TGFβ, and we therefore used a fluorescent peptide (kempTide) with multiple PKA phosphorylation sites and show that TGFβ alone or in combination with LMB, CDIM8, and SP600125 induced phosphorylation (activity), whereas this response was not observed in cells cotreated with TGFβ plus ALK5i, the TGFβ receptor inhibitor (Fig. 5A). In cells transfected with FLAG-MKK4-WT or FLAG-MKK7(CA) alone or in combination with CDIM8, LMB and SP600125 or FLAG-MKK7(DN) ± TGFβ, phosphorylation of PKA was not observed. As a positive control, we also observed increased phosphorylation of PKA in cells overexpressing the PKA catalytic subunit (Fig. 5A). Treatment of cells with the PKA inhibitor 14–22 Amide or knockdown of PKA-Cα (siPKA-Cα) by RNAi inhibited TGFβ-induced A549 cell migration, but did not affect MKK4/7-induced migration, which are downstream from PKA (Fig. 5B and C). We also investigated the effects of 14–22 Amide and siPKA-Cα (Fig. 5D and E) on TGFβ/MKK4/7-induced nuclear export of NR4A; only the TGFβ-induced effect was inhibited and MKK4/7 differentially enhanced nuclear export of NR4A1 independent of PKA inhibition. Results obtained for MKK4 ± 14-22 Amide (Fig. 5D) were somewhat inconsistent; however, results in Supplementary Fig. S3 confirm these observations using confocal microscopic analysis. We also examined the effects of 14-22 Amide and siPKA-Cα (Fig. 5F and G) on NR4A1 and the JNK1 pathway in A549 cells treated with TGFβ and show that phosphorylation of NR4A1, JNK1 and jun were inhibited and SMAD7 levels were increased compared with that observed in cells treated with TGFβ alone.

Previous studies show that TGFβ induces proteasome-dependent SMAD7 degradation via an NR4A1/RNF12/Arkadia/Axin2 complex (21, 31, 32) and treatment of A549 cells with TGFβ or transfection with FLAG-MKK4 and FLAG-MKK7(CA) followed by immunoprecipitation with NR4A1 antibodies showed that NR4A1 interacts with Axin2 and SMAD7 but not RNF12 or Arkadia (Supplementary Fig. S4A–S4C). Cotreatment with LMB, CDIM8 or SP600125 significantly decreased these interactions and transfection with FLAG-MKK7(DN) blocked TGFβ-induced interactions of NR4A1 with Axin2 and SMAD7 (Supplementary Fig. S4C). The same treatment groups were used and A549 cells were also transfected with FLAG-NR4A1-LBD (containing the LBD region of NR4A1) and immunoprecipitated with FLAG antibodies, and results showed that SMAD7 interacted with the ligand-binding domain of TGFβ/MKK4/MKK7(CA)-activated NR4A1 (Supplementary Fig. S4D–S4F). Using a SMAD7-FLAG construct in 549 cells treated with TGFβ, we also showed that SMAD7 interacts with Axin2, RNF12, Arkadia, and NR4A1, and these interactions are blocked by LMB, CDIM8, and SP600125. Treatment of A549 cells with TGFβ or transfection with MKK4 or MKK7(CA) followed by immunoprecipitation by SMAD7 antibodies showed that a broad band of ubiquitinated SMAD7 proteins were formed (Fig. 6A–C). Moreover, the intensity of the TGFβ-induced ubiquitinated SMAD7 was inhibited by LMB, CDIM8, and SP600125. In addition, TGFβ-induced ubiquitination was blocked after transfection with FLAG-MKK7(DN) and minimal ubiquitinated SMAD7 was observed in the control IgG lane (Fig. 6C). TGFβ-induced ubiquitination of SMAD7 was inhibited after knockdown of Axin2, Arkadia, and RNF12 (Fig. 6D) and TGFβ-induced ubiquitination of SMAD7 was also inhibited by 14-22 Amide or after transfection with siPKA-Cα (Fig. 6E). We also observed that MKK4/7 enhanced ubiquitination of SMAD7 (Fig. 6F) and these responses were not blocked by inhibition of PKA as both kinases are downstream from PKA. In contrast, both 14-22 Amide and siPKA-Cα inhibited TGFβ-induced interactions of NR4A1, Axin2, Arkadia, and RNF12 with SMAD7 (Fig. 6G). The importance of the ubiquitin ligase complex members in mediating TGFβ-induced migration of A549 cells is consistent with results in Fig. 6H showing that knockdown of Axin2, Arkadia, or RNF12 inhibits the TGFβ-induced response. Figure 6I illustrates the specificity of the ubiquitin ligase complex proteins after knockdown by RNA interference. These results demonstrate the critical role of NR4A1, Axin2, Arkadia, and RNF12 in mediating degradation of SMAD7 and TGFβ-induced cell migration and identify several inhibitors of this pathway including the NR4A1 antagonist CDIM8.

Because TGFβ and elements of the TGFβ signaling pathway and the ubiquitin ligase complex proteins (including NR4A1) play a role in SMAD7 expression and ubiquitination, we further examined their role in SMAD7 degradation. A549 cells were treated with TGFβ (Fig. 7A) or transfected with FLAG-MKK4 (Fig. 7B), FLAG-MKK7(CA) or FLAG-MKK7(DN) ± TGFβ (Fig. 7C) plus or minus the proteasome inhibitor MG132. Kinase activation alone decreased expression of SMAD7 which is consistent with activation of TGFβ signaling; however, cotreatment with LMB, CDIM8, or SP600125 ± MG132 prevented SMAD7 degradation and this was consistent with their resulting blockade of TGFβ-induced signaling and cell migration (Figs. 1B and 2B). The critical effects of SMAD7 degradation on TGFβ-induced migration are illustrated in Fig. 7D in which TGFβ-, MKK7(CA)-, and MKK4-induced migration of A549 cells is blocked by cotreatment with MG132, which increases SMAD7 levels due to inhibition of proteasome-dependent degradation of SMAD7 (Fig. 7E). We also confirmed the critical role of TGFβ-induced SMAD7 degradation by showing that TGFβ-induced invasion can be inhibited by overexpression of SMAD7 (Fig. 7F). Figure 7G illustrates the unique TGFβ–NR4A1–SMAD7 interactions in lung cancer cells and the role of PKA-MKK4/7-JNK in mediating the phosphorylation of NR4A1 and its nuclear export. Although TGFβ-induced nuclear export of NR4A1 and its role in degradation of SMAD7 are common in breast and lung cancer cells, there are significant cell context-dependent differences in TGFβ-induced kinase pathways and PKA-dependent induction of NR4A1 in lung versus β-catenin/TCF/LEF-mediated induction of NR4A1 in breast cancer cells (31).

In lung cancer cells, several reports demonstrate that TGFβ induces cell migration, invasion, and EMT through modulation of multiple genes/pathways (10–19) and these prooncogenic functions of TGFβ have been observed in many other tumor types (32–36). Recent studies in breast cancer cells show that TGFβ-induced migration involves the orphan nuclear receptor NR4A1, which is part of an ubiquitin ligase complex required for proteasome-dependent degradation of SMAD7, an inhibitor of TGFβ-activated signaling (20, 31). Studies in this laboratory previously showed the pro-oncogenic functions of NR4A1 in lung cancer cells and NR4A1 was overexpressed in tumors from patients with lung cancer and inversely correlated with their survival (22). On the basis of these observations, we hypothesized that NR4A1 may also play a role in TGFβ-induced lung cancer migration/invasion and that this pathway can be inhibited by DIM-C-pPhOH/CDIM8, a compound that binds nuclear NR4A1 and acts as an NR4A1 antagonist in cancer cells (24).

We initially used three lung cancer cell lines as models and show that TGFβ-induced migration was blocked by knockdown of NR4A1 or treatment with CDIM8, LMB, or the TGFβ receptor inhibitor Alk5i (Fig. 1). These data confirm that TGFβ-dependent activation of the TGFβ receptor is important for cell migration, and Western blot analysis confirmed that the TGFβ-induced response requires nuclear export of NR4A1, which is blocked by LMB and CDIM8. Results of kinase inhibitor studies show that the JNK1 inhibitor SP600125 also blocked TGFβ-induced cell migration, nuclear export of NR4A1 (Fig. 2) and inhibitors of nuclear export (LMB and CDIM8), and JNK also inhibited phosphorylation of NR4A1 (Fig. 1B–D and Fig. 2). These results are consistent with previous studies showing that selected apoptosis-inducing agents also induce phosphorylation-dependent nuclear export of NR4A1 through activation of JNK1 or other kinases (37–40). In addition, we also investigated both MKK4 and MKK7 which are upstream from JNK and demonstrate that overexpression of MKK4 or MKK7 recapitulated the effects observed with TGFβ in terms of enhanced cell migration and nuclear export of NR4A1 and inhibition of these responses by LMB, CDIM8, and SP600125 (Fig. 3; Supplementary Figs. S1 and S2). Moreover, MKK7(DN) also inhibited MKK7 and TGFβ-induced responses and thus, linking the upstream effects of TGFβ with MKK4/7-mediated activation of JNK.

Our previous studies in breast cancer cells (31) demonstrated that TGFβ-dependent phosphorylation and nuclear export of NR4A1 was due to activation of MKK3/MKK6 and p38 but not MKK4/7 and JNK. Moreover in breast cancer cells, we observed that TGFβ induced expression of both β-catenin and NR4A1, and the mechanism of NR4A1 expression involved β-catenin/TGF/LEF interactions with the NR4A1 promoter (31). In contrast, we did not observe induction of β-catenin in lung cancer cells treated with TGFβ, whereas TGFβ-induced expression of NR4A1 protein (1.5- to 2-fold; Fig. 1B–D) and RNA (>10-fold; Fig. 4A). We identified upstream CRE and SRE sites in the NR4A1 promoter at -807 and -703 (Fig. 4C) that bind c-jun/ATF2 and Elk1/SRF which are among some of the genes induced by JNK1 and these were induced by TGFβ in A549 cells (Fig. 4E). Moreover, like the upstream kinases, knockdown of c-jun, ATF2, Elk1 and SRF inhibited induction and nuclear export of NR4A1 and migration of A549 cells treated with TGFβ (Fig. 4).

Previous studies report that induction of NR4A1 expression in multiple cell types is associated with PKA, cAMP, or cAMP inducers (41–46) and cadmium induction of NR4A1 in A549 cells is both PKA- and MAPK-dependent (41). These reports, coupled with the identification of PKA-activated transcription factors interacting with the NR4A1 promoter (Fig. 4D), suggested that PKA may be a potential kinase downstream from TGFβ/TGFβ receptor in lung cancer cells. Moreover, there is prior evidence demonstrating that TGFβ/TGFβ receptor induces PKA (47–49). Data illustrated in Fig. 5 confirm that TGFβ induces PKA activity and the PKA inhibitor 14-22 Amide or transfection with siPKA-Cα inhibits TGFβ-induced migration and NR4A1 expression, nuclear export of NR4A1 and transcription factors associated with induction of NR4A1. TGFβ also induces NR4A1 in fibroblasts through activation of a SMAD3/SMAD4/Sp1 complex bound to GC-rich sites in the NR4A1 promoter (50), thus illustrating cell context–dependent differences in regulating NR4A1 expression.

Thus, TGFβ activates the TGFβ receptor–PKA–MKK4/7–JNK1 pathway which in turn phosphorylates NR4A1 and subsequently undergoes nuclear export (Fig. 7G). Although TGFβ/kinase-dependent nuclear export of NR4A1 is necessary for A549 cell migration, this process also involves subsequent responses associated with extranuclear NR4A1 because TGFβ-induced A549 cell migration is also inhibited by LMB and CDIM8 (Fig. 1). Previous studies in breast cancer cells showed that phosphorylated NR4A1 was a necessary component of a RNF12/Arkadia/Axin2/SMAD7 complex that induced ubiquitination and proteasome-dependent degradation of SMAD7, which in turn activated TGFβ/TGFβ receptor signaling (31, 36). Results illustrated in Fig. 6 and Supplementary Fig. S4 confirm that this same complex is also functional in A549 cells and is necessary for ubiquitination and subsequent degradation of SMAD7. Thus, another major difference between breast and lung cancer cells is TGFβ-dependent activation of p38 (breast) versus PKA/JNK (lung), which is required for nuclear export of NR4A1 and subsequent activation of proteasome-dependent degradation of SMAD7. The importance of SMAD7 degradation in mediating TGFβ-induced A549 cell migration is also supported by results showing that overexpression of SMAD7 inhibits the TGFβ-induced effect (Fig. 7F). Figure 7G illustrates that the mechanism of TGFβ-induced migration is a cyclic rather than a linear process because inhibition is observed by TGFβ receptor inhibitors (Alk5i), kinase inhibitors, NR4A1 antagonists, nuclear export, and proteasome inhibitors. Thus, effective treatment of TGFβ-induced lung cancer progression could involve a number of agents including the CDIM/NR4A1 antagonists which block not only TGFβ-induced migration but several other NR4A1-regulated prooncogenic genes/pathways in lung cancer cell lines (22).

No potential conflicts of interest were disclosed.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Conception and design: E. Hedrick, S. Safe

Development of methodology: E. Hedrick, S. Safe

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E. Hedrick, K. Mohankumar, S. Safe

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E. Hedrick, K. Mohankumar

Writing, review, and/or revision of the manuscript: E. Hedrick, S. Safe

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): E. Hedrick, S. Safe

Study supervision: S. Safe

Other (carried out experiments): K. Mohankumar

The financial assistance of the NIH (P30-ES023512, to S. Safe), Texas AgriLife Research (to S. Safe), and Sid Kyle Chair Endowment (to S. Safe) is gratefully acknowledged.

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.