FER promotes cell migration via regulating JNK activity

Abstract Objectives Cell migration has a key role in cancer metastasis, which contributes to drug resistance and tumour recurrence. Better understanding of the mechanisms involved in this process will potentially reveal new drug targets for cancer therapy. Fer is a non‐receptor protein tyrosine kinase aberrantly expressed in various human cancers, whereas its role in tumour progression remains elusive. Materials and Methods Transgenic flies and epigenetic analysis were employed to investigate the role of Drosophila Fer (FER) in cell migration and underlying mechanisms. Co‐immunoprecipitation assay was used to monitor the interaction between FER and Drosophila JNK (Bsk). The conservation of Fer in regulating JNK signalling was explored in mammalian cancer and non‐cancer cells. Results Overexpression of FER triggered cell migration and activated JNK signalling in the Drosophila wing disc. Upregulation and downregulation in the basal activity of Bsk exacerbated and eliminated FER‐mediated migration, respectively. In addition, loss of FER blocked signal transduction of the JNK pathway. Specifically, FER interacted with and promoted the activity of Bsk, which required both the kinase domain and the C‐terminal of Bsk. Lastly, Fer regulated JNK activities in mammalian cells. Conclusions Our study reveals FER as a positive regulator of JNK‐mediated cell migration and suggests its potential role as a therapeutic target for cancer metastasis.

pathways involved in the occurrence and development of cancer are evolutionarily conserved between human and flies. 15,16 This conservation enables Drosophila to model most hallmarks of mammalian cancer. Besides, a diversity of cancer models has been established in Drosophila. 17,18 Fer is a non-receptor protein tyrosine kinase expressed ubiquitously. 19 Early studies in Fer mainly focus on its role in regulating cell adhesion. Recent studies show that Fer is also implicated in numerous cellular processes such as cell cycle, cell proliferation, and migration. 19 In addition, several lines of evidence support a role of Fer in cancer promotion. Firstly, it has been reported that Fer is aberrantly upregulated or activated in various cancers such as breast, prostate, hepatic and lung cancers. [20][21][22][23] Secondly, the high expression of Fer is significantly correlated with poor prognosis in renal cell carcinoma. 24 Besides, a recent research reports that Fer overexpression activates NF-κB pathway and confers quinacrine resistance to several cancer cell lines. 25 Nevertheless, due to the complexity of cancer development and the limitations of our understanding of Fer, the molecular mechanism by which Fer promotes tumour progression remains elusive.
In this report, we investigate the function of FER, the Drosophila homolog of Fer, in a Drosophila model of cancer metastasis. Our results demonstrate that FER promotes cell migration in the wing disc, and the JNK signalling pathway mediates FER-induced cell migration. Elevation or reduction in the activity of JNK signalling via genetic manipulations strengthens or suppresses FER-induced cell migration, respectively. Besides, our work also reveals that FER is a positive regulator of the JNK signalling pathway and acts through Bsk. FER interacts with and phosphorylates Bsk, which require both the kinase domain and the C-terminal of Bsk. Lastly, our data show that the regulatory role of FER in JNK signalling is conserved in mammalian cells.

| Immunostaining
Staining was carried out according to standard protocol. 26 Briefly, third-instar larva wing discs were fixed with 4% formaldehyde in PBS-T for 20 minutes and blocked with 5% bovine serum albumin in PBS-T for 30 minutes. Discs were then incubated with primary antibodies at 4°C overnight followed by incubating with secondary antibodies at room temperature for 2 hours. Mounted discs were analysed with a Zeiss LSM 880 confocal microscope and a Nikon DS-Ri1 fluorescence microscope. Images were processed with Zeiss Zen and Adobe Photoshop software.

| Co-immunoprecipitation
Co-immunoprecipitation was performed as previously described. 27 Briefly, the transfected cells were lysed with lysis buffer at 4°C for 30 minutes. Insoluble fractions were removed, and supernatants were incubated with anti-Myc agarose conjugates at 4°C for 4 hours.
Then, the beads were washed with washing buffer and dissolved with SDS loading buffer.

| X-gal staining
Third-instar wing discs were dissected in PBS and fixed with 1% glutaraldehyde in PBS-T for 20 minutes. The samples were then stained for β-galactosidase as previously described. 27

| FER promotes cell migration in the wing disc
It has been reported that Fer is able to modulate cell migration in numerous cell lines. 29,30 To explore the precise function of Fer in an intact organism, we employed the Drosophila wing disc epithelium paradigm, a well-established system to study cell migratory behaviour in Drosophila. 31 We ectopically expressed Drosophila Fer (FER) at the A/P compartment boundary of the wing disc by patched-GAL4 (ptc-Gal4). Compared to the control, many FER-expressing cells migrated from the A/P compartment boundary to the posterior part of the wing disc ( Figure 1A-B, and Figure S1). In addition, the x-z imaging of these discs showed that FER-express- in Drosophila cell invasive migration and tumour metastasis. 34,35 Consistently, we found that overexpression of FER markedly improved the expression of MMP1 ( Figure 1C-D, and Figure S3).
Together, these results confirm that overexpression of FER promotes cell migration ( Figure 1E).

| FER activates JNK signalling in the wing discs
Next, we investigated how FER induced cell migration. Activation of the JNK signalling pathway has been demonstrated to trigger cell migration and activate MMP1 expression. 6 Notably, overexpression of As overexpression of FER activated the JNK signalling pathway, we continued to assess whether JNK signalling was required for FER-induced cell migration. To test this notion, we altered the activity of JNK signalling in the ptc-Gal4 > FER background. To slightly elevate the JNK activity, we removed one copy of JNK negative regulator puc (puc E69/+ ). puc E69/+ mutant alone could not affect the expression pattern of ptc-Gal4 in the wing discs ( Figure 2E). However, this genetic alteration significantly aggravated the spreading of ptc-Gal4 > FER cells ( Figure 2F-2G, and Figure S4). We also reduced the JNK signalling by expressing bsk DN , a dominant negative bsk construct. Co-expression of bsk DN strongly suppressed the migration of ptc-Gal4 > FER cells ( Figure 2H and Figure S4). These results indicate that FER-induced cell migration in the wing discs is mainly attributed to the activation of the JNK signalling pathway.

| FER is required for JNK signalling
The above results suggested that FER was sufficient to activate JNK signalling. We further asked whether FER was necessary for the signal transduction of JNK signalling. Previous studies showed that the downregulation of cell polarity genes such as dlg or scrib resulted in JNK-dependent cell migration. 36,37 Consistently, we observed that in ptc-Gal4 > scrib-IR discs, numerous scrib knockdown cells migrated away from A/P boundary, along with increased expression of MMP1 ( Figure 3A-B). Importantly, we found that the knockdown of FER no- in adults. Several studies have reported that this phenotype can be rescued by reducing JNK activity. 38,39 We found that removing one copy of FER or knocking down FER largely alleviated egr-induced small eye phenotype ( Figure 3D-G, and Figure S5). Together, these results suggest that FER is required for the outputs of the JNK signalling pathway.

| FER regulates JNK signalling through Bsk
The above results indicated that FER was a positive regulator of the JNK signalling pathway. To genetically map the location of FER in this pathway, we disrupted JNK signal transduction by knocking down its key components in the ptc-Gal4 > FER background.
The JNK signalling pathway is transmitted by a series of precisely regulated kinases including Egr, dTRAF2, dTAK1, Hep and Bsk.
Ectopic FER expression with ptc-Gal4 (ptc-Gal4 > FER) induced cell migration, accompanied by increased MMP1 and decreased E-cad expression. We found that all of these phenotypes remained unaffected by knocking down dTRAF2, dTAK1 or hep, but were strongly suppressed by knocking down bsk (Figure 4), indicating that FER might act downstream of or in parallel with hep, but upstream of bsk. Consistently, increased phosphorylation of Bsk was observed by co-expressing FER in S2 cells ( Figure 5A). In contrast, the p-Bsk levels remained unaffected by the co-expression of the kinase-dead FER K1092M (FER KD ) mutant ( Figure 5A), suggesting that the kinase activity of FER is required for its function in Bsk.
We performed the co-immunoprecipitation (coIP) assay and found that FER interacted with Bsk ( Figure 5B). Bsk consists of an N-terminal protein kinase domain (amino acids [aa] 1-292) and a short C-terminal extension (aa 293-372), and the anti-p-JNK antibody specifically detect the phosphorylation events in the TPY motif of the kinase domain. 40 To determine which region of Bsk mediates its interaction with FER, we constructed Bsk ΔC (aa 1-292), which lacked the C-terminal region. Unexpectedly, the interaction between FER and Bsk ΔC was much stronger than that between FER and the full-length Bsk ( Figure 5B). Typically, phosphorylation of a substrate significantly reduces its affinity to its kinase. 41,42 The above data thus suggest that FER binds to Bsk at the kinase domain but phosphorylates Bsk at its C-terminus. The lack of the phosphorylation site in Bsk ΔC stabilized FER-Bsk interaction. In consistent with this hypothesis, Bsk ΔC was not activated by FER ( Figure 5C), indicating that the phosphorylation at the C-terminus by FER primes for Bsk activation. Taken together, we propose a molecular mechanism by which FER activates Bsk ( Figure 5D). FER binds to the kinase domain of Bsk and phosphorylates its C-terminus, which consequently triggers Bsk activation in the TPY motif.

| Fer regulates the JNK signalling pathway in mammalian cells
Having demonstrated that Fer promotes cell migration through activating the JNK signalling pathway in Drosophila, we next checked whether Fer played a conserved role in mammals. In 293T human embryonic kidney cells, overexpression of Fer largely increased the endogenous p-JNK levels ( Figure 6A). Fer overexpression is also observed in the pathological context. For example, Oncomine analysis (https ://www.oncom ine.org) showed that Fer was highly amplified in metastasized bladder cancers ( Figure 6B). T24 is a well-characterized human urinary bladder carcinoma cell line with high expression of Fer. To test the idea whether upregulated Fer participated in the regulation of the JNK signalling pathway in cancer cells, we knocked down Fer in T24 cells with Fer siRNA (si-Fer; Figure S6) and observed that Fer RNAi significantly decreased the activity of endogenous JNK ( Figure 6C). Together, these results suggest that Fer regulates the JNK signalling pathway in mammals.

| D ISCUSS I ON
Although aberrant activation or overexpression of Fer is found in various cancers, 20,22 the pathological consequence and the underlying mechanism of this misregulation remain elusive. In this study, we showed that overexpression of FER in Drosophila wing disc resulted in cell migration, a key step in the progression of cancer ( Figure 1).
Moreover, we observed that in scrib-IR-induced cell migration model, the knockdown of FER strongly suppressed the migration phenotype ( Figure 3A-C), which further validated the role of FER in regulating cell migration.
Given the pivotal role of FER in regulating cell adhesion, 19 the migrated behaviour in FER overexpressed cells could either be a secondary effect of the defect in cell adhesion or the consequence of cell migration. In this study, we ascribed it to the later for two reasons. Cells with disrupted adhesion often disperse into the surrounding wild-type cells without a direction. 43 However, FER-induced cell migration is unidirectional. We noted that ptc-Gal4 > FER cells migrated from the A/P boundary to the posterior compartment of discs ( Figure 1A-B). In addition, ectopic expression of FER caused aberrant MMP1 expression ( Figure 1C-D), which is a characteristic for cell migration. 6 MMP1 is not only the hallmark of cell migration but also a target gene of JNK signalling. 6 Previous studies have also shown that JNK signalling is closely associated with cell migration. Besides, it has been reported that gain-of-function FER mutant Drosophila displays dorsal closure defect and axon misrouting. Co-expressing puc is able to substantially correcting both of these phenotypes. 44 We thus analysed the role of the JNK signalling pathway in FER-induced cell migration in the wing discs. We found that JNK signalling was remarkably activated by overexpressing FER (Figure 2A-2D). More importantly, elevating JNK activity by deleting one copy of puc aggravated FER-induced cell migration.
Conversely, reducing JNK activity by expressing dominant negative bsk suppressed this phenotype ( Figure 2E-H). Taken together, we demonstrate that JNK signalling is responsible for FER-induced cell migration. As Fer has been shown to be involved in tumour metastasis in human cancers, it will be interesting to investigate whether JNK signalling accounts for Fer-mediated metastasis in mammals.
F I G U R E 6 FER regulates JNK signalling in mammalian cells. A, The JNK phosphorylation levels in the absence and presences of Fer in 293T cells. B, Fer expression in Modlich Bladder. Dataset is grouped by cancer type. The reporter used in the study to detect Fer expression is J03358. C, The phosphorylation levels of JNK in T24 cells with or without Fer knockdown. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant. Data are presented as mean ± SEM (n = 3)