Phosphorylation of LCRMP-1 by GSK3β Promotes Filopoda Formation, Migration and Invasion Abilities in Lung Cancer Cells

LCRMP-1, a novel isoform of CRMP-1, can promote cancer cell migration, invasion and associate with poor clinical outcome in patients with non-small-cell lung cancer (NSCLC). However, the underlying regulatory mechanisms of LCRMP-1 in cancer cell invasiveness still remain obscure. Here, we report that GSK3β can phosphorylate LCRMP-1 at Thr-628 in consensus sequences and this phosphorylation is crucial for function of LCRMP-1 to promote filopodia formation, migration and invasion in cancer cells. Impediment of Thr-628 phosphorylation attenuates the stimulatory effects of LCRMP-1 on filopodia forming, migration and invasion abilities in cancer cells; simultaneously, kinase-dead GSK3β diminishes regulation of LCRMP-1 on cancer cell invasion. Furthermore, we also found that patients with low-level Ser-9-phosphorylated GSK3β expression and high-level LCRMP-1 expression have worse overall survival than those with high-level inactive GSK3β expressions and low-level LCRMP-1 expressions (P<0.0001). Collectively, these results demonstrate that GSK3β-dependent phosphorylation of LCRMP-1 provides an important mechanism for regulation of LCRMP-1 on cancer cell invasiveness and clinical outcome.


Introduction
Metastasis contributes to treatment failure and death in the majority of cancer patients [1]. The capacity of cancer cells to progressive metastasis is controlled by complicated cellular processes, involving microenvironmental changes, increasing ability of cell migration or invasion, multiple genetic events and regulatory factors. Until now, many master inducers and suppressors of cancer metastasis has been identified to be involved in these processes, and thus unraveling upstream regulatory pathways of these proteins may facilitate depicting detailed molecular mechanisms for cancer metastasis [2].
Glycogen synthase kinase-3b (GSK3b) is known as a multitasking serine/threonine kinase that control numerous cellular processes including glycogen metabolism, cell differentiation, apoptosis, cytoskeleton rearrangement, cell cycle regulation, and cell proliferation [3,4]. GSK3b regulates a broad range of substrates through phosphorylation at optimal consensus motifs (Ser/Thr-X-X-X-Ser/Thr, where X is representative of any amino acid) [3,5]. Usually, most common substrates of GSK3b need a specific priming kinase to increase the efficiency of first phosphorylation at serine or threonine residues that near to the four residues of GSK3b phosphorylation site in the carboxyl terminus. For example, casein kinase 1 prior primes b-catenin to GSK3b phosphorylation [6], and casein kinase 2 is a priming kinase of glycogen synthase [7].
Since LCRMP-1 and CRMP-1 have opposite function on cancer migration and invasion, whether the function of LCRMP-1 may be regulated by GSK3b should be further studied. In the present report, we investigate possible regulation of GSK3b on LCRMP-1. Here, we demonstrated that GSK3b can phosphorylate LCRMP-1 and modulate filopodia formation, cancer cell migration and invasion. We further confirm the GSK3bphosphorylated site in LCRMP-1, investigate its function for cell invasiveness and evaluate its clinical significant in NSCLC patients.

Results
GSK3b can phosphorylate LCRMP-1 at Thr-628 To predict whether the classic GSK3b phosphorylation consensus sequences are existed in LCRMP-1, we first aligned the protein sequences among CRMP-2, CRMP-1, and LCRMP-1 (Fig. 1A). Previous study showed that Cdk5 is a priming kinase that phosphorylates CRMP-2 at Ser-522, following with phosphorylation of CRMP-2 at Thr-514 by GSK3b and resulting in functional regulation of neuronal polarization [13]. Therefore, we found that protein sequences of LCRMP-1 contained highly consistent with Cdk5 and GSK3b phosphorylation motif, thus we speculated that a major potential phosphorylation site of LCRMP-1 is located at Thr-628 (Fig. 1A). To examine whether LCRMP-1 can be phosphorylated by GSK3b, HEK293T cells were cotransfected with wild-type Flag-tagged LCRMP-1 (WT) in the presence of empty vector, wild-type GSK3b (WT), constitutively active GSK3b (CA), or kinase-dead GSK3b (KD). Expression of GSK3b (CA) was more obviously detected mobility shifts (arrowheads) of LCRMP-1 (WT) than GSK3b (WT) (Fig. 1B, lane 2 and 3). However, slow-migration upper bands were completely disappeared in cells expressing kinase-deficient form of GSK3b (KD) (Fig. 1B, lane 4). These results suggest that LCRMP-1 is a substrate of GSK3b and slowly migrating bands were caused by its phosphorylation in vivo.
Thr-628 phosphorylation of LCRMP-1 is required for cancer cell invasion, migration and filopodia formation In our current reports, we have found that wild-type LCRMP-1 enhances filopodia formation, and promotes cell migration and invasion in noninvasive human lung cancer cell lines [10]. Since LCRMP-1 can be phosphorylated at Thr-628 by GSK3b, we next questioned whether the function of LCRMP-1 could be regulated by this phosphorylation. To address this, we generated a series of lentivirus that express GFP (control), non-tagged LCRMP-1 (WT), LCRMP-1 (T628A), or LCRMP-1 (T628D), and transduced them into low-invasive CL1-0 lung cancer cells which express low level of endogenous LCRMP-1 [11]. Protein expression of wild-type or mutant LCRMP-1 was confirmed by immunoblotting analysis using anti-LCRMP-1 antibodies ( Fig. 2A). These cells were then used to examine the ability of cell invasion. As expected, LCRMP-1 (WT) overexpression contributed to an increased cell invasiveness compared with GFP control (Fig. 2B). However, T628A nonphosphorylated mutant of LCRMP-1 was greatly diminished cell invasiveness (Fig. 2B). Conversely, phospho-mimic LCRMP-1 (T628D), which was expected to mimic the phosphorylated form, displayed enhanced invasion ability similar to wild-type LCRMP-1 (WT). To further explore the effect of Thr-628 phosphorylation of LCRMP-1 expression on CL1-0 cell motility, we performed video time-lapse microscopy assay to monitor moving tracks of at least 10 individual cells over a 20-hour period. Lentivirus-transduced CL1-0 cells expressing GFP-LCRMP-1 (WT) or GFP-LCRMP-1 (T628D) increased both migration distance and migration velocity compared to GFP vector (Fig. 2C, D, and E). However, GFP-LCRMP-1 (T628A) showed marked compression of distance and velocity of migration ( Fig. 2C, D, and E). These results suggested that phosphorylation of LCRMP-1 at Thr-628 is a prerequisite for cell invasiveness and cell migration.

GSK3b phosphorylates LCRMP-1 and modulates cancer cell invasion
To further detect the effects of GSK3b on LCRMP-1(WT)induced cancer cell invasion, lentivirus expressing GFP control, GSK3b (WT), GSK3b (CA), or GSK3b (KD) were infected in CL1-0/LCRMP-1 overexpression cells (lines 1015 and 1003) which have been previously shown to strongly induce cell invasiveness [10]. Consistent with our previous findings ( Taken together, these results demonstrated that GSK3b could modulate LCRMP-1 activity through a phosphorylation-dependent manner to control cancer cell invasion.

Low expression of inactive GSK3b and high expression of LCRMP-1 correlate with poor overall survival in NSCLC patients
Although our results consistently suggested that function of LCRMP-1 could be regulated by GSK3b phosphorylation, such studies do not fully reflect clinical malignancy. Accordingly, we extended our analysis by examining inactive form GSK3b and LCRMP-1 protein expression levels in tumor specimens from 142 NSCLC patients. The clinical characteristics of these patients are summarized in Table 1. Serial sections of each specimen were stained with antibodies against LCRMP-1 and Ser-9-phosphorylated GSK3b that indicated the status of inactive form GSK3b due to the Akt-mediated activation which results in suppression of GSK3b activity through phosphorylation at Ser-9 [15]. Our results showed typical staining of LCRMP-1 and Ser-9-phosphorylated GSK3b in patient's specimen (Fig. 5A). Consistent with our previous reports, high-level LCRMP-1 had significantly poor overall survival compared with low-level LCRMP-1 in patients with NSCLC [10,11]. Notably, analysis of the combined effect of both proteins on patients' prognoses revealed that patients with low-level expression of inactive form GSK3b and high-level expression of LCRMP-1 had poorer overall survival than those with high-level inactive form GSK3b expression and low-level LCRMP-1 expression (Fig. 5B, p,0.00001). Multivariable Cox proportional-hazards regression analyses, with a stepwise selection model, were present to evaluate the associations of various independent prognostic factors with patient survival ( Table 2). These results suggest that high activity GSK3b and high-level LCRMP-1, possibly mimicking the phosphorylated status of LCRMP-1, are associated with increasing cancer invasiveness and poorer overall survival.

Discussion
Our study primarily investigated the regulatory mechanism of post-translation modification associated with the cancer cell migration and invasiveness of LCRMP-1. Here, we showed that GSK3b-dependent phosphorylation of LCRMP-1 positively regulates filopodia formation, migration and cancer cell invasion. On the basis of GSK3b-phosphorylated consensus motifs, Thr-628 amino acid residue of LCRMP-1 is the master phosphorylation site for GSK3b (Fig. 1). A substitution of Thr-628 for Ala in LCRMP-1 led to impair filopodia formation, migration and cancer cell invasion, whereas a replacement of Thr-628 to Asp greatly restored its function ( Fig. 2 and 3). Consistent with these observations, ectopic expression of kinase-dead GSK3b diminished LCRMP-1-induced invasive ability (Fig. 4). Moreover, clinical NSCLC patients with low-level of inactive GSK3b and high-level LCRMP-1 protein expression is associated with poor overall survival than those with high-level inactive form GSK3b expression and low-level LCRMP-1 expression (Fig. 5B). Thus, our results provide evidence to support the crucial mechanism of GSK3b-dependent phosphorylation to control the LCRMP-1mediated filopodia formation, migration and invasive abilities in cancer cells.
Sequence analysis indicated that there has a Cdk5 (priming kinase) phosphorylation site in both LCRMP-1 and CRMP-1 (Fig. 1A). After phosphorylation at Ser-636 by Cdk5, GSK3b in turn phosphorylates Ser-632 and Thr-628 sequentially. Therefore, GSK3b may induce slower migrating bands in LCRMP-1 including both Ser-632 and Thr-628 phosphorylation in cells, and the bands induced by constitutively active GSK3b were more active than that of wild-type GSK3b (Fig. 1B). In detailed analysis, we found that LCRMP-1 mutant, T628A, could block most of GSK3b phosphorylation induced migrating bands (Fig. 1C). This may indicate that Thr-628 of LCRMP-1 may be the dominant and important phosphorylation site for GSK3b phosphorylation. However, we could not exclude the possibility that constitutively active GSK3b can phosphorylate other sites of LCRMP-1. The process of tumor invasion is primarily through alternations of the extracellular matrix including actin polymerization and filopodia formation [16]. Our findings reveal that blockage of GSK3bmediated phosphorylation of LCRMP-1 at Thr-628 leads to a regression of filopodia formation. A previous report was described that GSK-3 phosphorylation for Paxillin is necessary for cytoskeletal rearrangement [17]. Thus, we speculate that GSK3b may positively promote regulation of actin cytoskeleton and tumor invasion. In contrast to positively role, GSK3b is also reported to perform its suppressive roles for its substrates [18]. GSK3b phosphorylation for some nuclear transcription factors, such as b-catenin and Snail, trigger proteasomal degradation, following with suppression on epithelial-mesenchymal transition (EMT) and tumor invasion [6,[19][20][21]. GSK3b simultaneously localizes in cytoplasm and nucleus [22], as consistent with our results of immunohistochemical staining (Fig. 5A), and LCRMP-1 is a stable cytosolic protein [10]. Therefore, GSK3b govern cell invasion is possibly dependent on the characterization of its protein substrates. High-level LCRMP-1 expression is associated with poor overall and disease-free survival compared to low expression group in NSCLC patients [10,11]. Thus, LCRMP-1 potentially serves as a best candidate to identify high-risk patients. In this study, we showed a very interesting finding that patients with low-level phosphorylated GSK3b and high-level LCRMP-1 expressions had worse overall survival than the other catalog groups. Focusing on the low-level LCRMP-1 expression or the high-level LCRMP-1 expression group, we could also found that high-level of phosphorylated GSK3b expression may discriminate better outcome from low-level phosphorylated GSK3b expression, respectively. The combined effects of inactive form GSK3b and LCRMP-1 protein expression may have important clinical implications to indicate the high-risk subset of NSCLC patients as candidates for additional effective adjuvant therapy. The results suggest that GSK3b phosphorylation of LCRMP-1 is associated with poor clinical outcome. However, there still have some limitations in our experiments. Although the reference indicated that Ser-9phosphorylated GSK3b can indicate the status of inactive form GSK3b due to the Akt-mediated activation which results in suppression of GSK3b activity through phosphorylation at Ser-9 [15], whether Ser-9 phosphorylation of GSK3b inhibits its ability to phosphorylate LCRMP-1 is still not clear. To solve this problem, generating a specific anti-phospho-Thr-628 LCRMP-1 antibody for immunohistochemistry should be the most important issues in the future. This may confirm the clinical significance of GSK3b induced phosphorylation of LCRMP-1 in NSCLC patients.
In addition, we also found that patients with low-levels (n = 73) or high-levels (n = 69) of Ser-9-phosphorylated GSK3b expression were almost equal in distribution. Cells exist at different actual levels of active GSK3b unless distinct signaling pathways for inhibiting GSK3b are triggered concurrently, such as MAPK and phosphatidylinositol-3-OH kinase/Akt pathways [20]. Thus, we speculate that distinct activated extent of signaling pathways to induce an inhibition of GSK3b activity may lead to different outcome for patients with LCRMP-1 expression. Therefore, further investigating upstream signaling pathway for regulation of GSK3b may provide a better diagnosis for NSCLC patients with low-levels or high-levels of LCRMP-1 expression.
In conclusion, we show a new regulatory mechanism for GSK3b to phosphorylate an invasion enhancer LCRMP-1 and thus could further fine-tune cancer cell invasion abilities. Additionally, LCRMP-1 expression and Ser-9-phosphorylated GSK3b levels may have clinical implications in the outcome prediction of patients with NSCLC.

Ethics statement
This investigation was approved by the Institutional Review Board of the National Taiwan University Hospital and obtained informed written consent statement from all participant patients involved in our study.

Patients and Tumor Specimens
Lung tumor tissue specimens were obtained from patients (n = 142) with histologically confirmed NSCLC who had undergone complete surgical resections at the National Taiwan University Hospital (Taipei, Taiwan) between December 28, 1995, and December 26, 2005. This investigation was approved by the Institutional Review Board of the National Taiwan University Hospital. The enrolled patients had not been treated with neoadjuvant chemotherapy or irradiation therapy. All specimens were formalin fixed, sectioned, stained with hematoxylin and eosin, and examined by microscopy. Pathological staging was performed by Dr. Yih-Leong Chang (Department of Pathology and Graduate Institute of Pathology, National Taiwan University) according to the international staging system for lung cancer [23].

Immunohistochemical analysis
Immunohistochemical staining of tumor tissue samples from patients with NSCLC was carried out as previously described [11]. In brief, the sections for analysis of LCRMP-1 or phosphorylated GSK3b protein expression were first autoclaved in Trilogy Solution (Cell Marque Corp., Rocklin, CA.) or Antigen Retrieval Citra Solution (Biogenex, San Ramon, CA) at 121uC for 10 minutes. The samples were subsequently made a treatment of 3% H 2 O 2 -methanol, incubation with DakoCytomation Dual EndogenousEnzyme Block (DakoCytomation, Inc., Carpinteria, CA) for 10 minutes, Ultra V Block (Lab Vision Corporation, Fremont, CA) for 10 minutes, antibody-dilution buffer (Ventana Medical Systems, Inc., Tucson, AZ) for 10 minutes, and finally with a phosphorylated GSK3b (Cell signaling, Danvers, MA) antibody for 6 hours at room temperature or a polyclonal anti-LCRMP-1 antibody (C2; 1:300 dilution) overnight at 4uC. Detection of the immunostaining was determined using Super Sensitive Non-Biotin Polymer HRP Detection System (BioGenex, San Ramon, CA) according to the manufacturer's protocol.

Immunofluorescence staining for observation of filopodia formation
Transfected or lentivirus-infected cells were fixed with 3.7% cold paraformaldehyde, washed with PBS, following by permeabilizing with 0.1% Triton X-100. The cells were then stained with rhodamine-conjugated phalloidin (red, Molecular Probes, Eugene, OR). The cells were mounted onto microscope slides with ProLongH Gold antifade reagent with DAPI (Molecular Probes) and then examined and photographed using LSM 700 laser scanning confocal microscope from Carl Zeiss.

Cell migration analysis
Moving tracks of migrating cells were performed by video timelapse microscopy as previously described [24]. In brief, cells were maintained in growth medium at 37uC/5% CO 2 and time-lapse images were observed under a AF 6000 LX microscope (Meyer Instruments,Inc.) for the time period of 20 hours. Images were taken with a CoolSNAP HQ CCD camera (Roper Scientific, NJ) at 5-minute intervals and processed by MetaMorph 5.0 software (Universal Imaging, Downingtown, PA).

Cell culture and transfection
The human lung adenocarcinoma cell lines (CL1-0 cells) were isolated from a 64-year-old male patient with a poorly differen-  tiated adenocarcinoma and selected in our laboratory by in vitro Transwell invasion to get 5 sublines with progressive invasiveness, with similar genotypic background (designated CL1-1, CL1-2, CL1-3, CL1-4, and CL1-5) as previously described [25]. HEK293T cell lines were purchased from American Type Culture Collection (ATCC, USA). The CL1-0 and HEK293T cells were grown in RPMI and DMEM medium containing 10% FBS and 2 mM L-glutamine (all from Invitrogen, Eugene, OR) at 37uC in a humidified atmosphere of 5% CO 2 -95% air, respectively. All cell lines in this study were tested with mycoplasma-free condition. All transfection experiments were carried out using Lipofectamine or Lipofectamine 2000 reagents (Invitrogen) according to the manufacturer's instructions.

Lentivirus production and transduction
GFP-tagged, untagged LCRMP-1 (WT, T628A and T628D), and myc-tagged GSK3b (WT, CA and KD form) were constructed by cloning their cDNA into pTYEF lentiviral vector. Briefly, HEK293T cells were contransfected with the indicated lentiviral vector and three helper plasmids pHP-dl-N/A, pHEF-VSVG, and pCEP4-Tat by using Lipofectamine 2000 reagents according to manufacturer's protocols. Virus-containing medium was collected at 24, 48, 72 hours post-transfection, centrifuged, and filtered through 0.45 mm-pore-size filters. The percentage of pTYEF-GFP-infected cells by flow cytometry were determined the relative lentivirus titers. Cells were infected with GFP or the indicated lentivirus in media containing polybrene (8 mg/ml). After twenty-four hours post-infection, cells were treated with fresh medium for 24-48 hours and then used for all experiments.

Cell lysate preparation and immunoblotting
All experiments were performed according to standard protocols. Briefly, preparation of whole-cell lysates for immunoblotting and immunoprecipitation were using IP lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 0.5% Nonidet P-40, 100 mM Na 3 VO 4 , 50 mM NaF, 30 mM sodium pyrophosphate) containing protease inhibitors (protease inhibitor cocktail; Roche Diagnostics, Basel, Switzerland). After brief sonication and centrifugation, protein samples were resolved by SDS-PAGE gels, transferred into PVDF membranes (Millipore), blotted with the indicated antibodies and finally detected chemiluminescent signals using X-ray films.

Statistical analysis
Data are shown as Mean 6 SEM. and statistical analyses were performed by Student's t-test or Pearson's x2 test. The overall survivals for patient groups with different expression signatures were determined using SPSS software (v10.0; SPSS, Inc., Chicago, IL) by the Kaplan-Meier method and two-sided log-rank tests. Immunoreactivity in more than 50% and 70% of the tumor specimens was defined to high level of LCRMP-1 and phosphorylated GSK3b expression, respectively. P values,0.05 were considered to be statistically significant.