NEDD9 promotes lung cancer metastasis through epithelial–mesenchymal transition

Authors

  • Yujuan Jin,

    1. State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, People's Republic of China
    Search for more papers by this author
    • Y.J., F.L. and C.Z. contributed equally to this work

  • Fei Li,

    1. State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, People's Republic of China
    Search for more papers by this author
    • Y.J., F.L. and C.Z. contributed equally to this work

  • Chao Zheng,

    1. State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, People's Republic of China
    Search for more papers by this author
    • Y.J., F.L. and C.Z. contributed equally to this work

  • Ye Wang,

    1. State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, People's Republic of China
    Search for more papers by this author
  • Zhaoyuan Fang,

    1. State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, People's Republic of China
    Search for more papers by this author
  • Chenchen Guo,

    1. State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, People's Republic of China
    Search for more papers by this author
  • Xujun Wang,

    1. Department of Bioinformatics, School of Life Science and Technology, Tongji University, Shanghai, People's Republic of China
    Search for more papers by this author
  • Hongyan Liu,

    1. State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, People's Republic of China
    Search for more papers by this author
  • Lei Deng,

    1. Department of Bioinformatics, School of Life Science and Technology, Tongji University, Shanghai, People's Republic of China
    Search for more papers by this author
  • Cheng Li,

    1. Department of Bioinformatics, School of Life Science and Technology, Tongji University, Shanghai, People's Republic of China
    Search for more papers by this author
  • Hongda Wang,

    1. State Key Laboratory of Electroanalytical Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, Jilin, People's Republic of China
    Search for more papers by this author
  • Haiquan Chen,

    1. Department of Thoracic Surgery, Fudan University Shanghai Cancer Center, People's Republic of China
    2. Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, People's Republic of China
    Search for more papers by this author
  • Yan Feng,

    Corresponding author
    1. State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, People's Republic of China
    • Correspondence to: Hongbin Ji, State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 320 Yue Yang Road, 200031 Shanghai, People's Republic of China, Tel.: +86-21-54921108, Fax: +86-21-54921101, E-mail: hbji@sibcb.ac.cn or Yan Feng, State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 320 Yue Yang Road, 200031 Shanghai, People's Republic of China, Tel.: +86-21-54921107, Fax: +86-21-54921101, E-mail: fengyan01@sibs.ac.cn

    Search for more papers by this author
  • Hongbin Ji

    Corresponding author
    1. State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, People's Republic of China
    • Correspondence to: Hongbin Ji, State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 320 Yue Yang Road, 200031 Shanghai, People's Republic of China, Tel.: +86-21-54921108, Fax: +86-21-54921101, E-mail: hbji@sibcb.ac.cn or Yan Feng, State Key Laboratory of Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 320 Yue Yang Road, 200031 Shanghai, People's Republic of China, Tel.: +86-21-54921107, Fax: +86-21-54921101, E-mail: fengyan01@sibs.ac.cn

    Search for more papers by this author

Abstract

Metastasis is the major cause for high mortality of lung cancer with the underlying mechanisms poorly understood. The scaffolding protein neural precursor cell expressed, developmentally down-regulated 9 (NEDD9) has been identified as a pro-metastasis gene in several types of cancers including melanoma and breast cancer. However, the exact role and related mechanism of NEDD9 in regulating lung cancer metastasis still remain largely unknown. Here, we demonstrate that NEDD9 knockdown significantly inhibits migration, invasion and metastasis of lung cancer cells in vitro and in vivo. The pro-metastasis role of Nedd9 in lung cancer is further supported by studies in mice models of spontaneous cancer metastasis. Moreover, we find that NEDD9 promotes lung cancer cell migration and invasion through the induction of epithelial–mesenchymal transition (EMT) potentially via focal adhesion kinase activation. More importantly, NEDD9 expression inversely correlates with E-cadherin expression in human lung cancer specimens, consistent with the findings from in vitro studies. Taken together, this study highlights that NEDD9 is an important mediator promotes lung cancer metastasis via EMT.

Lung cancer is a deadly disease with high mortality and its 5-year survival rate is approximately 15% globally. Metastasis, frequently observed in patients initially diagnosed with lung cancer, remains as an important factor in contribution to high mortality of lung cancer.[1] Despite great efforts in last decades, the molecular mechanisms involved in lung cancer metastasis process still remain far from being fully understood.

Recent studies have shown that NEDD9 (Neural precursor cell expressed, developmentally down-regulated 9, also known as HEF1), a scaffolding protein without known catalytic activity, is important for pro-metastasis behavior in several types of solid tumors.[2, 3] In melanoma, NEDD9 is frequently up-regulated by gene allele amplification and activates FAK (focal adhesion kinase) to promote cell invasion and metastasis.[2] NEDD9 complexes with DOCK3 and regulates the activation of Rac which are essential for promoting the mesenchymal-type movement of melanoma cells.[4] In glioblastoma, NEDD9 works as an effector of FAK to promote tumor cell aggression.[5] Interestingly, Nedd9/ mice displayed much less mammary tumor formation as well as lung metastasis, which is potentially due to the reduction of Fak and Src activation.[6] Consistently, we have recently found that NEDD9 is transcriptionally regulated by CRTC1/CREB complex in LKB1-deficient lung cancer and significantly contributes to tumor progression via the promotion of tumor differentiation status.[7] However, the roles of NEDD9 in lung cancer metastasis and the related mechanisms are still unclear.

In this study, we demonstrated that NEDD9 was important for lung cancer metastasis. Mechanistically, NEDD9 promoted lung cancer metastasis through induction of epithelial–mesenchymal transition (EMT) potentially via FAK activation. We further found that NEDD9 expression was inversely correlated with E-cadherin expression in human lung cancer specimens. Taken together, our data have uncovered the important role of NEDD9 and related mechanism in lung cancer metastasis as well as identified a potential target for future metastatic lung cancer therapy.

Material and Methods

Animal models

KrasG12D and Lkb1L/L mice were originally provided by Dr. T. Jacks and Dr. R. Depinho, respectively.[8] Nude mice (6 weeks old, male) were purchased from Shanghai SLAC Laboratory Animal (Shanghai, China). All mice were housed in a specific pathogen-free environment at Shanghai Institute of Biochemistry and Cell Biology and treated in strict accordance with protocols approved by the Institutional Animal Use Committee of the Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences.

For de novo lung cancer mice model, KrasG12D, Lkb1L/L mice were treated with 2 × 106 plague-forming unites of adeno-Cre at 6–8 weeks old as previously described.[9] Twelve to sixteen weeks after adeno-Cre treatment, primary lung tumors were dissected from the KrasG12D, Lkb1L/L mice and prepared for primary cell culture as described previously[10] as well as for histopathology examination and genotyping identification.

For lung seeding assay, nude mice were injected with 1 × 106 cells via tail veins. After 10 weeks of inoculation, all mice were euthanized and the lungs were dissected for photographing, weighting, gross inspection and pathology analysis. Lung metastasis was quantified by counting the total numbers of metastatic lesions from series sections of H&E staining.

For dual fluorescence-based tumor competitive spontaneous metastasis mouse model, primary lung cancer cell line (KL, 1 × 106) derived from KrasG12D, Lkb1/ mouse lung tumor was virally infected with Lenti-Ctrl-copGFP (Ctrl-copGFP) or Lenti-shNedd9#B-DsRed2 (shNedd9#B-DsRed2), and sorted by flow cytometry and randomly grafted into either flank of the same nude mice subcutaneously. In classical spontaneous cancer metastasis assay, 1 × 106 KL cell lines virally infected with pLKO-Puromycin (Ctrl-Puro) or pLKO-shNedd9#B-Puromycin (shNedd9#B-Puro) were grafted subcutaneously into individual nude mice. Primary tumor volume was monitored every 2 days and calculated as described previously.[11] Three weeks later, all mice were euthanized for primary tumor weighing and lung metastasis analyses based on fluorescence signal at gross inspection and microscopic analyses. Genomic DNA was also extracted from the whole lung of each mouse for detection of micro-metastasis by PCR amplification using primers specific for copGFP or DsRed2. Wild-type P53 served as internal control for genomic DNA quality. The primers used for wild type P53 (P53-WT) were: 5′-CACAAAAACAGGTTAAACCCAG-3′ (forward) and 5′-AGCACATAGGAGGCAGAGAC-3′ (reverse).

Cell culture

Human lung cancer cell lines A549, CRL-5810, CRL-5907 and CRL-5807 were purchased from ATCC (American type culture collection, the Global Bioresource Center). All cells were cultured in RPMI-1640 medium (Hyclone) supplemented with 10% fetal bovine serum (FBS) (Biochrom AG, Germany) and antibiotics (100 U/ml streptomycin and 100 µg/ml penicillin) (Invitrogen). The human cell line HEK-293T was maintained in DMEM (Hyclone) supplemented with 10% heat inactivated fetal calf serum and antibiotics. Lentivirus infection was used for gene overexpression or knockdown in lung cancer cells as previously described.[12] Stable cell lines were selected and maintained in 2 µg/ml puromycin or 500 µg/ml G418. The primary mouse cell line KL was maintained in DMEM with 10% FBS.

Tissue specimens and immunohistochemistry analysis

A total of 62 human lung cancer specimens were collected with patient consents in Fudan University Shanghai Cancer Center (Shanghai, China) from January 2008 to June 2009 with the approval from the Institute Research Ethics Committee. All the lung cancer specimens were immunostained for NEDD9 and E-cadherin, and analyzed as previously described.[7, 13] For assessment the correlation of NEDD9 and E-cadherin expression, p value was calculated by Pearson Chi-square test using SPSS 13.0 statistical software package. A value of p < 0.05 was considered as significant (two tailed).

Statistical analysis

Data were analyzed by Student's t test; p < 0.05 was considered to be significant (two tailed). Error bars represent standard error of mean (s.e.m.).

Results

NEDD9 is important for invasion, motility and metastasis of lung cancer cells

To investigate the role of NEDD9 in lung cancer metastasis, we first used shRNAs to knockdown NEDD9 expression in A549 cells (Figs. 1a and 1b) and found that decreased NEDD9 expression dramatically inhibited the invasiveness in matrigel culture, the motility in wound healing assay as well as migration and invasion in Boyden chamber assay in A549 cells (Figs. 1c1f). Consistently, NEDD9 knockdown inhibited cell growth (Supporting Information Fig. 1a) and resulted in much less and smaller colonies in soft agar assay (Supporting Information Figs. 1b and 1c). The inhibitory role of NEDD9 knockdown in cell invasion and motility was also confirmed in CRL-5810 cells (Supporting Information Figs. 2a–2f). However, NEDD9 knockdown seemed to have little effect upon the growth of CRL-5810 cells (Supporting Information Fig. 2g), which might reflect the variable effect of NEDD9 knockdown upon lung cancer cell proliferation. We further examined the impact of NEDD9 knockdown upon lung cancer metastasis using lung seeding assays in vivo. A549 cells with or without NEDD9 knockdown were injected intravenously into nude mice (five to six mice each group) via tail vein. All the mice were euthanized for gross inspection and pathology inspection 10 weeks afterwards. Many large lung tumor nodules were visible on lung surface of control group, whereas NEDD9 knockdown resulted in dramatic reduction in nodule number, nodule size and overall lung weight (Figs. 1g−1i), suggesting that NEDD9 knockdown significantly decreased the lung colonization ability of A549 cells as well as suppressed metastasis. Taken together, these data have demonstrated that knockdown of NEDD9 significantly decreased lung cancer metastasis, highlighting an essential role of NEDD9 in the metastatic program of lung cancer.

Figure 1.

NEDD9 is important for invasion, motility and metastasis of A549 cells. (a) Real-time RT-PCR showed the efficiency of NEDD9 knockdown in A549 cells. Data were shown as mean ± s.e.m., n = 3. (b) Western blot showed the efficiency of NEDD9 knockdown in A549 cells. Actin served as control. (c) NEDD9 knockdown significantly inhibited invadopodia of A549 cells in matrigel. Photos were taken on the tenth day. Scale bar: 40 µm. (d) shNEDD9 inhibited cell motility of A549 cells in wound healing assay. Representative photos were taken at 0 hr and 36 hr post wound healing. Scale bar: 100 µm. (e, f) NEDD9 knockdown significantly inhibited migration and invasion of A549 cells assessed by Boyden chamber assay. Representative photos (e) and the migrated cell number per high-power field (HPF, f) were shown. Scale bar: 50 µm (e). Data were shown as mean ± s.e.m. ***p < 0.001 (f). (g) Gross inspection of lungs dissected from nude mice receiving A549 cells, A549 cells with shNEDD9#1 or shNEDD9#2 via tail vein injection. Scale bar: 1cm. (h) Significant difference of lung weights was found between the control (Ctrl) group and shNEDD9 groups in A549 lung seeding assay. Data were shown as mean ± s.e.m. * p < 0.05, ** p < 0.01. (i) Pathology analyses showed that NEDD9 knockdown significantly decreased lung metastasis of A549 cells in nude mice lung seeding assay. Scale bars: 500 µm (top panels) and 100 µm (middle and bottom panels).

Figure 2.

NEDD9 promotes migration, invasion and motility of lung cancer cells. (a, b) Ectopic expression of NEDD9 in CRL-5907 cells was assessed by real-time RT-PCR (a) and Western blot (b). Data were shown as mean ± s.e.m., n = 3 (a). Actin served as loading control (b). (c, d) Ectopic NEDD9 expression in CRL-5907 cells increased colony size in matrigel culture (c) and motility in wound healing assay (d). Scale bar: 40 µm (c) and 100 µm (d). (e, f) Ectopic NEDD9 expression significantly increased migration and invasion of CRL-5907 cells assessed by Boyden chamber assay. Representative photos (e) and the migrated cell number per high-power field (HPF, f) were shown. Scale bar: 50 µm (e). Data were shown as mean ± s.e.m. *** p < 0.001 (f). (g, h) Ectopic expression of NEDD9 in CRL-5807 cells was assessed by real-time RT-PCR (g) and Western blot (h). Data were shown as mean ± s.e.m., n = 3 (g). Actin served as loading control (h). (i, j) Ectopic expression of NEDD9 increased colony size in matrigel culture (i) and motility of CRL-5807 cells in wound healing assay (j). Scale bar: 50 µm (i) and 100 µm (j). (k, l) Ectopic NEDD9 expression significantly increased migration and invasion of CRL-5807 cells assessed by Boyden chamber assay. Representative photos (k) and the migrated cell number per high-power field (HPF, l) were shown. Scale bar: 50 µm (k). Data were shown as mean ± s.e.m. *** p < 0.001 (l).

Ectopic expression of NEDD9 enhances the abilities of invasion and motility of lung cancer cells

We next analyzed the biological function of ectopic NEDD9 expression in lung cancer cell lines. We found that NEDD9 overexpression in CRL-5907 cells (Figs. 2a and 2b), which have low endogenous NEDD9 expression, displayed more and larger colonies in matrigel culture (Fig. 2c), higher motility in wound healing process (Fig. 2d) and stronger migration and invasion in Boyden chamber assay (Figs. 2e and 2f). CRL-5907 cells with ectopic NEDD9 expression produced larger colonies in soft agar while showed no superiority on proliferation (Supporting Information Figs. 3a–3c), which might be due to the difference between 2D and 3D cell culture system. Similarly, ectopic expression of NEDD9 in another NEDD9 low-expressed lung cancer cell line CRL-5807 (Figs. 2g and 2h) had the promotive effects on cellular migration, invasion and motility (Figs. 2i−2l) without cell growth acceleration (Supporting Information Fig. 4).

Figure 3.

NEDD9 positively regulates partial EMT of lung cancer cells. (a–d) Ectopic expression of NEDD9 increased Vimentin (VIM) expression and decrease E-cadherin (E-cad) expression in CRL-5907 cells assessed by real-time RT-PCR (a, b), Western blot (c) and immunofluorescence staining (d). Data were shown as mean ± s.e.m. ** p < 0.01, *** p < 0.001 (a, b). The expression level of Vimentin was quantified by densitometry. Actin served as control (c). Scale bar: 5 µm (d). (e–h) Knockdown of NEDD9 decreased expression of Vimentin but increased E-cadherin expression in A549 cells assessed by real-time RT-PCR (e, f), Western blot (g) and immunofluorescence staining (h). Data were shown as mean ± s.e.m. *** p < 0.001 (e, f). The expression level of Vimentin was quantified by densitometry. Actin served as control (g). Scale bar: 5 µm (h).

Figure 4.

FAK potentially contributes to NEDD9-induced EMT in lung cancer cells. (a) Western blot showed that knockdown of FAK partially reversed the modulation of Vimentin (Vim) or E-cadherin (E-cad) expression by ectopic NEDD9 expression in CRL-5907 cells. Actin served as control. (b) Ectopic expression of FAK partially reversed the modulation of Vimentin or E-cadherin expression by NEDD9 knockdown in A549 cells. Actin served as control. (c, d) Representative photos showed FAK knockdown partially inhibited colony size of CRL-5907 cells with NEDD9 overexpression in matrigel (c) as well as inhibited the cell motility in the process of wound healing (d). Photos were taken on the tenth day and scale bar was 50 µm (c). Photos were taken at 0 hr and 48 hr post wound healing and scale bar was 100 µm (d). (e, f) Invasion of CRL-5907 cells with/without ectopic NEDD9 expression and/or FAK knockdown were assessed by Boyden chamber assay. Representative photos (e) and the migrated cell number per high-power field (HPF, f) were shown. Scale bar: 50 µm (e). Data were shown as mean ± s.e.m. *** p < 0.001 (f). (g, h) Representative photos showed ectopic FAK expression partially promoted invadopodia of A549 cells with NEDD9 knockdown in matrigel (g) as well as promoted the cell motility in the process of wound healing (h). Photos were taken on the tenth day and scale bar was 50 µm (g). Photos were taken at 0 hr and 48 hr post wound healing and scale bar was 100 µm (h). (i, j) Invasion of A549 cells with/without NEDD9 knockdown and/or FAK overexpression were assessed by Boyden chamber assay. Representative photos (i) and the migrated cell number per high-power field (HPF, j) were shown. Scale bar: 50 µm (i). Data were shown as mean ± s.e.m. *** p < 0.001 (j).

NEDD9 promotes partial EMT process of lung cancer cells

To understand the biological basis involved in lung cancer metastasis promoted by NEDD9, we performed microarray analysis using lung cancer cell lines with either ectopic NEDD9 expression or NEDD9 knockdown. Interestingly, we found that the EMT and EMT-related pathways including Mesenchymal signaling and TGF-β1 pathway were among the most significantly signaling pathways regulated by NEDD9 expression perturbation (Supporting Information Figs. 5a–5c, and Supporting Information Tables 1–3). EMT is an important event during cancer progression and metastasis, which improves cell intrinsic capabilities for local invasion and distant organ metastasis.[14, 15] We found NEDD9 overexpression in CRL-5907 cells increased both the mRNA and protein levels of the mesenchymal marker Vimentin but decreased the epithelial marker E-cadherin by real-time RT-PCR, Western blot and immunofluorescent staining analyses (Figs. 3a3d). Reciprocally, NEDD9 knockdown decreased Vimentin expression and increased E-cadherin expression in A549 cells (Figs. 3e−3h). Consistent regulation of the EMT-related genes (such as SNAIL, MTA3, AKT1 and LOXL2)[16-21] by NEDD9 was validated in both lung cancer cell lines with either NEDD9 overexpression or NEDD9 knockdown (Supporting Information Figs. 6a–6d). Meanwhile, NEDD9 overexpression or knockdown did not dramatically change the morphology of lung cancer cells (data not shown). Together, these results indicated that NEDD9 overexpression has conferred the lung cancer cells with EMT-like biochemical features, which might promote lung cancer cell invasion and motility through regulation of the intrinsic metastatic capability.

Figure 5.

NEDD9 expression is negatively correlated with E-cadherin expression in human lung cancer specimens. A panel of 62 human lung cancer specimens was immunostained with NEDD9 and E-cadherin, respectively. The correlation between NEDD9 and E-cadherin was analyzed using Pearson Chi-square test using SPSS 13.0 statistical software package. A value of p < 0.05 was considered as significant (two tailed). (a) Representative NEDD9 and E-cadherin immunohistochemical images of human primary lung cancer specimens. Scale bar: 50 µm. (b) A significant negative correlation between NEDD9 and E-cadherin expression was found in human lung cancer specimens (r = −0.323, p = 0.011).

Figure 6.

Nedd9 is an essential downstream mediator for Lkb1-dificient lung cancer metastasis. (a) A scheme showing the strategy for spontaneous lung cancer metastasis model using the KrasG12D, Lkb1/ (KL) cells labeled with different fluorescence as indicated. KL cells virally infected with Ctrl-copGFP (green) or shNedd9#B-DsRed2 (red) were sorted by flow cytometry and then randomly grafted into either flank of the same nude mice subcutaneously (n = 6). (b) The tumor growth kinetics were shown and no significant difference was observed in nude mice grafted by KL cells virally infected with either Ctrl-copGFP (green) or shNedd9#B-DsRed2 (red). Data were shown as mean ± s.e.m., one-way ANOVA, compared with the Ctrl-copGFP group, p > 0.05 (n = 6). (c) No significant difference of primary tumor weights was observed between these two groups. Data were shown as mean ± s.e.m. p > 0.05 (n = 6). (d) Representative photo showing mouse lungs with spontaneous cancer metastasis indicated by arrow. (e) Representative pathology photo for lung lesions derived from spontaneous metastasis of subcutaneous KL cell xenografts. Scale bar: 500 µm (left) and 50 µm (right). (f) Detection of macro- or micro-fluorescence signals in primary subcutaneous tumors and spontaneous tumor metastasis in mouse lungs. The lung metastases were indicated by arrows. Representative phase images and H&E staining for primary tumors and mouse lungs with spontaneous metastasis were also shown. Scale bar: 50 µm. (g) PCR detection of spontaneous metastasis into lungs using whole lung genomic DNA with primers specific for either copGFP or DsRed2. Genomic DNAs from primary tumors were used as positive controls. Wild-type P53 served as internal control for genomic DNA quality.

NEDD9 promotes the intrinsic metastatic capability of lung cancer cells via EMT through FAK regulation

Considering that FAK is proposed to mediate the function of NEDD9 in melanoma and breast cancer metastasis,[2, 6] we then asked whether FAK was involved in the EMT process induced by NEDD9 in lung cancer cells. Our results showed that FAK knockdown inhibited FAK phosphorylation and Vimentin expression but increase E-cadherin expression regulated by NEDD9 overexpression in CRL-5907 cells (Fig. 4a); the converse phenomena were observed when FAK is ectopically expressed in A549 cells with NEDD9 knockdown (Fig. 4b). Moreover, we further detected the biological function of FAK in NEDD9-promoted lung cancer metastasis. The results showed that FAK knockdown could partially suppress invasion and motility of NEDD9 overexpressed CRL-5907 cells (Figs. 4c4f) or CRL-5807 cells (Supporting Information Figs. 7a–7d), while ectopic expression of FAK partially rescued the inhibition of invasion and migration caused by shNEDD9 in A549 cells (Figs. 4g−4j) and CRL-5810 cells (Supporting Information Figs. 7e−7h). Taken together, these data demonstrate that FAK is partially involved in NEDD9-induced EMT process and contributes to NEDD9-promoted lung cancer metastasis.

NEDD9 expression is inversely correlated with E-cadherin expression in human lung cancer specimens

We then asked whether NEDD9 contributes to EMT process of lung cancer cells in vivo. To this end, we detected the expression levels of NEDD9 and E-cadherin of 62 human primary lung cancer specimens (Fig. 5a). Interestingly, an inverse correlation between NEDD9 and E-cadherin expression was observed (r = −0.323, p = 0.011) (Fig. 5b). Our result showed that 13 out of 20 specimens (65%) with NEDD9 low expression were found to have high E-cadherin expression, while only 31% (13/42) had high E-cadherin expression among those 42 specimens with high NEDD9 expression by immunostaining analysis (Fig. 5b). These data indicated the negative correlation of NEDD9 and E-cadherin in human lung cancer tissues, which may hints reprogramming of epithelial cell identity could be important to the metastatic phenotype of human lung cancer cells with NEDD9 high expression in vivo.

Nedd9 is a critical downstream mediator of lung cancer metastasis evoked by Lkb1 deficiency

We have recently found that Nedd9 is a critical downstream mediator in Lkb1-deficient lung cancer progression, whereas its role in lung cancer metastasis is unclear yet.[7] Our in vitro study here has indicated a pro-metastasis role of NEDD9 in lung cancer. To dig deep into this further, we used a dual fluorescence-based tumor competitive spontaneous metastasis mouse model which was based on fluorescence detection of spontaneous metastasis of mouse primary lung cancer cell line (KL) derived from KrasG12D, Lkb1/ lung tumor. KL cells with Nedd9-knockdown expressed DsRed2 while control KL cells expressed copGFP. For shNedd9#B had higher knockdown efficiency than shNedd9#A in KL cells (Supporting Information Figs. 8a and 8b), we chose to graft shNedd9#B-DsRed2 cells and the Ctrl-copGFP cells randomly into either flank of nude mice subcutaneously (Fig. 6a). To avoid the inhibition effect of shNedd9#B on primary tumor growth, we optimized the conditions of inoculation with more cell numbers for injection and shorter time course for tumor growth. No significant difference of primary tumor volumes and weights was observed in 3 weeks after xenografts (Figs. 6b and 6c). Spontaneous lung metastases were then analyzed by gross and pathology inspection (Figs. 6d and 6e). Interestingly, analyses of macro-fluorescence signal on lung surfaces showed that most if not all the spontaneous lung metastasis were positive for copGFP (Fig. 6f). Further detailed micro-fluorescence signal analysis on pathology sections also showed similar results (Fig. 6f). It was further confirmed by PCR detection of micro-metastasis in mouse lungs (Fig. 6g). More importantly, this was consistent with the findings from another set of experiment using classical spontaneous cancer metastasis mouse model, in which different groups of KL cells (with or without Nedd9#B knockdown) were subcutaneously grafted into individual nude mice and subjected to lung metastasis analyses 3 weeks later (Supporting Information Figs. 9a–9c). These data clearly show that Nedd9 acts as an important downstream mediator in Lkb1-deficient lung cancer metastasis.

Discussion

Metastasis is the most important factor in contribution to the high mortality of lung cancer. However, the mechanisms involved in this process remain to be fully understood. In this study, we demonstrate that NEDD9 is an important player in promoting lung cancer metastasis via regulation of the EMT process potentially through FAK activation. Downregulation of NEDD9 expression dramatically decreases the invasiveness, motility and metastasis of lung cancer cells and alleviated spontaneous lung cancer metastasis in vivo. The potential role of NEDD9 in EMT is further supported by the observation of an inverse correlation of NEDD9 and E-cadherin in human lung cancer specimens. Thus, our study deciphered the mechanism involved in NEDD9-promoted lung cancer metastasis and identified NEDD9 as a potential target for metastatic lung cancer treatment.

Several studies have previously linked NEDD9 to metastasis in multiple solid tumors including breast cancer,[3, 6] melanoma[2] and glioblastoma.[5] Minn et al. showed that decreased NEDD9 expression is associated with lung metastasis of breast cancer.[22] Contradictorily, Izumchenko et al. showed that MMTV-PyVmT, Nedd9/ mice had a reduction in number of primary breast tumors, as well as produced a trend toward fewer lung metastases,[6] which hints the opposite function of NEDD9 in suppressing breast cancer metastasis reported as before. Consistent with the potent pro-metastatic role of NEDD9 in melanoma[2] and glioblastoma,[5] we have previously shown a promotive role of NEDD9 in lung cancer progression.[7] Here, we have further extended our previous findings and shown that NEDD9 directly promotes lung cancer metastasis in mouse models. We reason that NEDD9 might have cell type specific function upon tumor progression and metastasis, which awaits further investigation.

Tumor metastasis is a complicate process involving cell survival, proliferation as well as migration and invasion. In this study, we have concluded that NEDD9 is important for lung cancer metastasis. Our data showed that Nedd9 knockdown significantly impaired the spontaneous lung metastasis of KL cells in vivo, and NEDD9 promoted lung cancer cell migration and invasion in vitro. Moreover, we demonstrate that lung cancer cells with NEDD9 knockdown formed smaller colonies in soft agar, as well as suppressed lung cancer cell growth, which indicated the potent role of NEDD9 in lung cancer cell proliferation. We reason that Nedd9 is similar to the subset of genes previously reported to promoting both cell proliferation and metastasis in breast tumorigenesis.[22]

Our study also highlights the important functional link between NEDD9 and EMT in lung cancer metastasis. Through bioinformatic analysis and ensued cellular biology analysis, we demonstrate that NEDD9 is involved in EMT process of lung cancer cells, consistent with two recent studies showing NEDD9 induced EMT in breast cancer[23, 24] as well as the positive correlation between NEDD9 and EMT marker expression in lung cancer.[25] EMT is an important event involved in early stage of cancer metastasis.[15] NEDD9 is reported to enhance FAK activation through the direct interaction, which results in an extensive tyrosine phosphorylation of itself sufficient to recruit effector proteins Crk and Crk-L essential for switching on the migration machinery.[26] Recent data have suggested that SRC mediates NEDD9-induced EMT through degradation of E-cadherin in breast cancer at post-translational level.[24] Interestingly, our study in lung cancer shows that FAK is involved in EMT process of lung cancer cells induced by NEDD9 at transcriptional level. Although we do not know how this occurs, we reason that the different mechanisms in breast cancer and lung cancer might reflect a tissue-specific regulatory mechanism for NEDD9 in promoting EMT. As a scaffold protein in cytoplasm, NEDD9 transcriptionally regulates EMT only by indirect ways, from which we consider FAK is not the direct downstream mediating NEDD9-promoted EMT process of lung cancer cells. Future work about the detailed mechanisms will be interesting to biochemically dissect the NEDD9 signaling involved in EMT process in lung cancer.

Furthermore, we have previously shown that NEDD9 works as an important regulator of lung cancer progression downstream of LKB1.7 LKB1 somatic mutations are frequently observed in lung cancer.[8, 27, 28] Recent co-clinical trial study has shown that LKB1-deficient lung tumors are intrinsically resistant to multiple drug treatments,[29] highlighting that this subset of lung cancer may represent as a unique subpopulation. Our studies in de novo mouse models have previously shown that LKB1-deficient metastatic lung tumors display an EMT pattern.[30] However, the detailed mechanism remains unknown. Our data here show that NEDD9, which is up-regulated in lung tumors by LKB1 deficiency,[7] partially promotes EMT via FAK activation and thus provides a potential explanation for our previous findings. Based on the findings from this study, it remains very attractive to test if FAK inhibition suppresses tumor progression and metastasis in KrasG12D, Lkb1L/L mouse model.

In summary, we have uncovered an important role of NEDD9 and related mechanisms involved in lung cancer metastasis. Although NEDD9 is important for cell mitosis[31] and cell survival,[32-34] the homozygous deletion of NEDD9 in mice does not seem to affect the health and fertility,[35] indicating that NEDD9 targeted therapy may produce a low side effect. In light of the importance of NEDD9 in lung cancer progression and metastasis, NEDD9 might serve as a promising target for metastatic lung cancer treatment.

Acknowledgements

The authors thank Drs. Kwok-kin Wong, Lynda Chin, Ronald A. DePinho and Tyler Jacks for material contribution.

Ancillary