Phosphorylated tubulin adaptor protein CRMP-2 as prognostic marker and candidate therapeutic target for NSCLC


  • Erik Oliemuller,

    1. Oncology Division, Center for Applied Medical Research (CIMA), University of Navarra, 55 Pamplona, Spain
    Search for more papers by this author
    • E.O. and R.P. contributed equally to this work

  • Rafael Peláez,

    1. Oncology Division, Center for Applied Medical Research (CIMA), University of Navarra, 55 Pamplona, Spain
    Search for more papers by this author
    • E.O. and R.P. contributed equally to this work

  • Saray Garasa,

    1. Oncology Division, Center for Applied Medical Research (CIMA), University of Navarra, 55 Pamplona, Spain
    Search for more papers by this author
  • María J. Pajares,

    1. Oncology Division, Center for Applied Medical Research (CIMA), University of Navarra, 55 Pamplona, Spain
    Search for more papers by this author
  • Jackeline Agorreta,

    1. Oncology Division, Center for Applied Medical Research (CIMA), University of Navarra, 55 Pamplona, Spain
    Search for more papers by this author
  • Rubén Pío,

    1. Oncology Division, Center for Applied Medical Research (CIMA), University of Navarra, 55 Pamplona, Spain
    2. Department of Biochemistry and Molecular Biology, University of Navarra, 55 Pamplona, Spain
    Search for more papers by this author
  • Luis M. Montuenga,

    1. Oncology Division, Center for Applied Medical Research (CIMA), University of Navarra, 55 Pamplona, Spain
    2. Department of Histology and Pathology, University of Navarra, 55 Pamplona, Spain
    Search for more papers by this author
  • Alvaro Teijeira,

    1. Oncology Division, Center for Applied Medical Research (CIMA), University of Navarra, 55 Pamplona, Spain
    Search for more papers by this author
  • Susana Llanos,

    1. Tumor suppression Group, Spanish National Cancer Research Center (CNIO), Madrid, Spain
    Search for more papers by this author
  • Ana Rouzaut

    Corresponding author
    1. Oncology Division, Center for Applied Medical Research (CIMA), University of Navarra, 55 Pamplona, Spain
    2. Department of Biochemistry and Molecular Biology, University of Navarra, 55 Pamplona, Spain
    • Oncology Division, Center for Applied Medical Research (CIMA), University of Navarra, Pío XII 55 Pamplona, Spain
    Search for more papers by this author


Collapsin response mediator protein-2 (CRMP-2) is the first described and most studied member of a family of proteins that mediate the addition of tubulin dimers to the growing microtubule. CRMPs have mainly been studied in the nervous system, but recently, they have been described in other tissues where they participate in vesicle transport, migration and mitosis. In this work, we aimed at studying the role of CRMP-2 in lung cancer cell division. We first explored the expression of CRMP-2 and phosphorylated (Thr 514) CRMP-2 in 91 samples obtained from patients with localized nonsmall cell lung cancer. We observed a significant correlation between high levels of nuclear phosphorylated CRMP-2 and poor prognosis in those patients. Interestingly, this association was only positive for untreated patients. To provide a mechanistic explanation to these findings, we used in vitro models to analyze the role of CRMP-2 and its phosphorylated forms in cell division. Thus, we observed by confocal microscopy and immunoprecipitation assays that CRMP-2 differentially colocalizes with the mitotic spindle during cell division. The use of phosphodefective or phosphomimetic mutants of CRMP-2 allowed us to prove that anomalies in the phosphorylation status of CRMP-2 result in changes in the mitotic tempo, and increments in the number of multinucleated cells. Finally, here we demonstrate that CRMP-2 phosphorylation impairment, or silencing induces p53 expression and promotes apoptosis through caspase 3 activation. These results pointed to CRMP-2 phosphorylation as a prognostic marker and potential new target to be explored in cancer therapy.

Microtubules are the major cytoskeletal components in eukaryotic cells. The microtubule network is needed for the maintenance of cell shape and polarity. In addition, it participates in the intracellular transport of vesicles and organelles, and constitutes the primary cilia, the sensory organelle in eukaryotic cells.1 During cell division, the microtubules form the mitotic spindle, a highly specialized and dynamic structure that mediates the alignment of replicated chromosomes to the equatorial plane and their subsequent transmission to daughter cells.2 Microtubules are dynamic polymers formed by tubulin heterodimers, in continuous equilibrium between growth and destruction.3 To accomplish microtubule formation, the concurrent activity of other proteins called microtubule-associated proteins (MAPs) is needed. MAPs contribute to transport, localization and regulation of microtubule components.4 One of these MAPs is the family of collapsin response mediator proteins or CRMPs.5

CRMPs are a family of five phosphoproteins (CRMP1–5) highly conserved in mammals. CRMPs are involved in the regulation of microtubule polymerization and in axonal outgrowth.6 CRMP-2 was the first CRMP described in cells of neuronal origin.7 It is a 62 kDa phosphoprotein that mediates the addition of tubulin to the positive end of the growing microtubule.8 Specifically, CRMP-2 transports tubulin dimmers throughout the growing microtubule bound to other MAPs such as dynein,9 kinesin or numb.10

To accomplish its function, CRMP-2 must be dephosphorylated since its phosphorylation diminishes its affinity for tubulin. There are several protein kinases that phosphorylate CRMP-2: Cdk5 phosphorylates CRMP-2 on Ser 522 which serves as a priming residue for further phosphorylation by GSK3β on Thr 509, Thr 514 and Ser 518 residues.11 Less frequent phosphorylation events involve CRMP2 Thr 555 and Tyr 479 which are targets of ROCK1 and Yes kinases.12

Inspite of being initially described in neurons, CRMP-2 has recently been described in non-neuronal cells such as leukocytes,13 fibroblasts14 and neuroblastoma cells,15 and in fact, is currently considered a protein that is broadly expressed within tissues. Interestingly, CRMP-2 has been detected in the mitotic spindle of transformed mouse and human cells.14 Recently, it was demonstrated that CRMP-2 binds to tubulin during mitosis, whereas its depletion leaded to destabilized anaphase microtubules and altered spindle position.5

Cancer cells often present important defects in their cell cycle allowing for uncontrolled cell proliferation. Therefore, one of the most effective strategies in the treatment of cancer is to target cell division with antimitotic drugs directed against proteins or structures needed to complete the cell cycle.16 Here, we describe for the first time the sequential phosphorylation-dependent interaction between CRMP-2 and tubulin on the mitotic spindle of lung adenocarcinoma cells. We show that altered phosphorylation of CRMP-2 leads to aberrant mitosis and cell death. In addition, we observed a positive correlation between high expression of phosphorylated CRMP-2 and poor outcome in nonsmall cell lung cancer (NSCLC) patients, while the levels of total CRMP-2 were not related to survival.

Material and Methods


We analyzed a series of 91 patients with localized NSCLC that underwent surgery at the Clinica Universidad de Navarra between 2000 and 2007. Histological diagnosis was performed based on the 2004 World Health Organization classification of lung tumors.17 Pathologic staging of the tumors was performed according to the International System for Staging Lung Cancer.18 Patient inclusion criteria were: complete resection of the primary lung tumor, tumor histology (adenocarcinoma, squamous cell carcinoma, adenosquamous carcinoma and large cell carcinoma), no malignancy within the 5 years previous to surgery and no neo-adjuvant therapy. Postsurgical adjuvant therapy was administered in 51 NSCLC patients, while 40 individuals were treated exclusively with surgery. The study protocol was approved by the ethics committee from Clinica Universidad de Navarra. Written informed consent was obtained prior to participation. Clinical and pathological characteristics of the patients are shown in Supporting Information Table I.

Immunohistochemistry of clinical samples from NSCLC patients

Formalin-fixed, paraffin-embedded tissue sections were evaluated. Endogenous peroxidase activity was quenched, and antigen retrieval was carried out by citrate buffer pH 6, using Lab Vision PT Module (Thermo Scientific, Wilmington, DE, USA). Nonspecific binding was blocked using 5% normal goat serum in Tris-buffered saline (TBS) for 30 min. Sections were incubated with anti-CRMP-2 antibody (Immuno-Biological Laboratories, Minneapolis, USA) or anti-phosphorylated CRMP-2 (Thr514, Abcam, Cambridge, UK), overnight at 4°C. Afterward, the samples were incubated with Envision polymer (Dako, Glostrup, Denmark) for 30 min at room temperature. Negative controls were carried out by omission of the primary antibody or incubation with an isotype control antibody. Supporting Information Figure 1 provides a table that summarizes the different antibodies used for CRMP-2 detection, their specific epitopes and two experiments to prove P-CRMP-2 T514 antibody specificity.

Assessment of immunohistochemical staining

Two observers (M.J.P and S.G.) evaluated the samples independently and unaware of the clinicopathological characteristics of the patients. After tissue examination, 90 and 88 samples were used for the study of the association between total and phosphorylated CRMP-2 expression and clinicopathological parameters respectively due to sample availability.

Staining scores were established by semiquantitative analysis. Both, extension and intensity of staining were considered to establish an “H score” for each sample as described.19 Discordant independent reading was resolved by simultaneous review by the two observers. REMARK criteria were followed throughout all the study.20

Cell culture

The human NSCLC cell line derived from lung adenocarcinoma NCI-A549 was purchased from American Type Culture Collection (ATCC, LGC-Promochem SL, Barcelona, Spain) and grown in RPMI 1640 medium (GIBCO, Barcelona, Spain) supplemented with 10% FetalClone III (HyClone, Thermo Scientific, Madrid, Spain) and 100 units/ml penicillin-streptomycin. Cells were maintained at 37°C in a humidified 5% CO2 incubator. To ensure cell line authentication, a specific Kras mutation in exon 2 (c.34 G>A) and a 23 base-pair deletion of SMARCA4 exon 15 were detected by PCR amplification of genomic DNA and subsequent sequencing of the PCR products.

Generation of CRMP-2 phosphomutants and cell transfection

The cDNA sequence encoding the open reading frame of human CRMP-2 was cloned into pcDNA3.2/V5/GW/D-TOPO (Invitrogen, Barcelona, Spain).21 Phosphomimetic (Ser 522 to Asp) and phosphodefective CRMP-2 (Ser522 to Ala) mutants were obtained by mutagenesis using the QuickChange Site Directed Mutagenesis kit (Agilent Technologies, Madrid, Spain) and the following oligonucleotides:





Wild type CRMP-2 and mutants were subcloned into pLEGFP-C1 vector (Mountain View, CA, USA) and transformed in E.coli competent cells. The recombinant plasmid was purified using Qiagen Plasmid MaxiPrep kit (Qiagen, Madrid, Spain). To transiently transfect A549 cells, FuGENE reagent (Roche, Barcelona, Spain) was used following manufacturer's instructions.

CRMP-2 silencing

A549 cells (3.5 × 105) were transfected with 2 μM siRNA nontargeting control sequence (siNT- 5′-UAAGGCUAUGAAGAGAUACUU-3′) and CRMP-2 siRNA (siCRMP2- 5′-GGAUCACGGGGUAAAUUCCUU-3′), using DharmaFECT-1 (Thermo Scientific) according to the manufacturer's instructions.

Western blot analysis

Total cell extracts were prepared using RIPA buffer as described.19 The antibodies used were: anti-CRMP-2 (C4G) (IBL International, Hamburg, Germany); anti-phosphorylated CRMP-2 (Ser522) and anti-phosphorylated CRMP-2 (Thr 509, Thr514) from Kinasource (Dundee, Scotland); anti-phosphorylated and non-phosphorylated GSK-3β (Calbiochem-Merck, Darmstadt, Germany) and anti-cleaved caspase-3 and p21 (Cell Signalling, MA, USA); anti-p53 (DakoCytomation, Glostrup, Denmark A/S); anti-β-Actin (Sigma-Aldrich, Madrid, Spain) and anti-α-tubulin antibody (Abcam, Cambridge, UK). Secondary antibodies were: anti-mouse IgG HRP-conjugated (GE Healthcare, Madrid, Spain); anti-rabbit IgG HRP (Santa Cruz Biotechnology, Inc., CA, USA) and anti-sheep IgG HRP-conjugated (DakoCytomation).

Coimmunoprecipitation assays

Five hundred micrograms of total protein were incubated with 2 μg of anti-CRMP-2 antibody for 1 hr at 4°C on an orbital shaker. Afterward, 20 μl of agarose conjugated immunoreactive protein A (Sigma–Aldrich) were added and incubated for 16 hr at 4°C. The protein-antibody-protein A complex was washed with RIPA buffer and centrifuged at 16,000 rpm at 4°C for 1 min. The immunoprecipitated protein was detected by anti-tubulin Western blot analysis.


A549 cells were seeded at 150 cells/mm2 on microscopy glass slides, previously covered with collagen at a final concentration of 50 μg/ml (Menzel-Glaser, Braunschweig, Germany). Cells were fixed in 4% paraformaldehyde in TBS for 20 min and permeabilized by incubation with 0.5% Triton X-100 at room temperature for 5 min. Nonspecific binding was blocked by incubation with 1/10 goat serum (Sigma–Aldrich) for 30 min at room temperature. Incubation with a specific anti-CRMP-2 antibody and anti-tubulin antibody was carried out overnight. Samples were incubated 30 min at room temperature with secondary Alexa fluor 488 goat anti-mouse or Alexa fluor 594 goat anti-rabbit IgG (Invitrogen) and visualized in a Zeiss fluorescence microscope Axio Imager M1(Oberkochen, Germany). All images were captured and processed using the Zeiss Axiovision program.

Cell synchronization and in vivo confocal microscopy

A549 cells (1 × 105) were seeded in 35 cm2 confocal microscopy glass plates (MatTek Corporation, Ashland, MA, USA). When needed, cells were transiently transfected with each of CRMP-2-GFP constructs and 36 hr after transfection, nocodazole was added at a concentration of 50 nM for 16 hr. Afterward, cell culture medium was renewed, and 500 μl of 2.5 mM Draq5 agent (Biostatus Limited, Leicestershire, UK) was added for 10 min in complete RPMI medium. Cells were visualized using an in vivo confocal microscopy PerkinElmer Ultraview ERS model (Waltham, MA, USA). Images were taken every 2 min during 2 hr covering 1 mm Z stacks, and subsequently mounted using Ultraview software from PerkinElmer. Cells in prometaphase were distinguished by cell rounding and chromatin condensation. Metaphase was determined as the time point of chromosome alignment in the equatorial plane, and anaphase as the time point when the sister chromatids started to migrate towards the cellular poles. Total duration of anaphase was determined as the time between metaphase and the moment in which the chromatids were located at the most distant cellular poles. Finally, telophase was determined by the clear establishment of the cleavage furrow between the two daughter cells by indentation of the cell membrane.

Determination of cell viability and apoptosis

For determination of cell viability by the neutral red assay, 5 × 103 A549 transfected cells were plated in 96-well plates (TPP, MO, USA) in complete RPMI culture media. After 24, 48, 72 and 96 hr of culture, the medium was removed, and the cell viability was tested as previously described.22 Cell apoptosis was determined by annexin V- PI staining (Serotec, Oxford, UK) following manufacturer's instructions. Apoptotic cells were quantified using a Becton Dickinson FACScan (San Jose, CA, USA).

Cell cycle study by flow cytometry

NSCLC cells (106) were fixed in 70% ethanol in PBS overnight at 4°C. Afterward, the cells were centrifuged, washed and resuspended in PBS at a concentration of 2 × 106 cells/ml. RNase A was added at 0.5 mg/ml final concentration, and samples were incubated 1 hr at 37°C. Subsequently, the cells were labeled with 50 μg/ml PI (Fluka, Sigma-Aldrich) in darkness at 4°C and analyzed using a FACScan, Becton-Dickinson System. Flow cytometry data were processed using CellQuest software (BD-PharMingen).

Statistical analysis

The association between the expression of CRMP-2, or phosphorylated-CRMP-2 expression, and clinicopathological parameters was analyzed by Pearson's chi-square test. Overall survival (OS) of patients was represented using Kaplan-Meier curves, and significant differences between high-CRMP-2 and low-CRMP-2 groups were tested using the log-rank test. Recurrence free survival time (RFS) was calculated from the date of surgery to the date of recurrence or death. Survival time was calculated from the date of surgery to the date of death.

In the remaining analyses, normal distribution was assured with the Kolmogorov–Smirnov test, and data were compared with Student's t test for two independent samples. A p-value less than 0.05 was considered statistically significant.


Phosphorylated CRMP-2 is expressed in primary tumors from NSCLC patients

We began our study by analyzing the expression of total CRMP-2 and phosphorylated CRMP-2 (phospho-Thr 514) in lung carcinoma samples from a cohort of 91 NSCLC patients. Immunohistochemistry analysis showed that CRMP-2 was expressed in the cytoplasm of both normal and tumor tissues, although its expression was more intense in tumor cells (Supporting Information Figure 2A). Interestingly, phosphorylated CRMP-2 was predominantly nuclear and was detected only in tumor tissues (Fig. 1).

Figure 1.

Expression of phosphorylated CRMP-2 in NSCLC tissues is associated with poor overall survival. (a) Immunohistochemical staining for phosphorylated CRMP-2 protein (Thr 514) in normal lung (NL), adenocarcinoma (AC) and squamous cell carcinoma (SCC) samples. (b) Kaplan Meier curves of overall survival (OS) and recurrence-free survival (RFS) for patients with high and low expression of phosphorylated CRMP-2 (median of CRMP-2 H-score was used as cut-off). (c) Kaplan Meier curves of OS in patients divided according to the use of adjuvant treatment.

The relationship between the expression of CRMP-2 or phosphorylated CRMP-2 and clinicopathological parameters was evaluated by the chi-squared test. Total CRMP-2 expression was significantly higher in adenocarcinoma than in squamous cell carcinoma (p < 0.001; Supporting Information Table II). We did not find association of the expression of total CRMP-2 or phosphorylated CRMP-2 with other clinicopathological characteristics (gender, age, pT, stage, nodal infiltration, histology and smoking history).

Phosphorylated CRMP-2 is associated with a poor outcome in patients with NSCLC

We determined whether expression of total CRMP-2 or phosphorylated CRMP-2 was associated with clinical outcome in NSCLC patients. No association was found between total CRMP-2 expression and OS or RFS. On the other hand, a significant association between high levels of phosphorylated CRMP-2 (Thr 514) and shorter OS was found (p = 0.034; Fig. 1b). Interestingly, the association between high phosphorylated CRMP-2 expression and worse OS was only found in patients no treated with adjuvant therapy (p = 0.011; Fig. 1c). We did not find significant differences in OS when adenocarcinoma and squamous cell carcinoma histologies were analyzed separately, probably due to the limited number of patients in each group.

To further evaluate whether CRMP-2 phosphorylation is a feature of transformed cells, we examined the levels CRMP-2 phosphorylation on residues Thr 509, Ser 518 and Ser 522 (detected with 3F4 antibody), in protein extracts obtained from primary bronchial epithelial cells (NHBE), immortalized and nontransformed bronquial epithelial cells (BEAS), and two NSCLC cell lines: A549 and H1299. Remarkably, only NSCLC cell lines showed simultaneous phosphorylation on residues Thr 509, Ser 518 and Ser 522 (Supporting Information Figure 2B).

CRMP-2 colocalizes with tubulin during mitosis

The finding of phosphorylated CRMP-2 in the nucleus of NSCLC cells, and its correlation with a poor outcome in lung cancer patients, led us to consider the role of CRMP-2 phosphorylation during mitosis. To investigate this, we first used immunofluorescence to analyze CRMP-2 and tubulin colocalization in dividing and nondividing cells. CRMP-2 could be detected in the cytoplasm of nondividing cells, and it colocalized with tubulin more clearly in the perinuclear region known as the microtubule organizing center (Fig. 2a). During mitosis, clear CRMP-2/tubulin colocalization could be observed in prometaphasic microtubules (Fig. 2b). This colocalization grew in intensity in metaphase spindles, and CRMP-2 images mirrored tubulin structure when separately analyzed. In contrast, only mild staining of CRMP-2 was observed in the spindle poles and in the spindle mid-zone in cells in anaphase. Finally, CRMP-2/tubulin costaining could be detected in cells in late cytokinesis in the structure known as the midbody. To further explore whether there existed a physical interaction between CRMP-2 and tubulin in cycling cells, we carried out coimmunoprecipitation assays in A549 cells synchronized with nocodazole. These assays demonstrated increased tubulin/CRMP-2 coprecipitation 1 hr after cell cycle release (Fig. 2c), when the majority of the cells (86%) were in the pro-metaphase or metaphase stages of the cell cycle (Supporting Information Figure 3).

Figure 2.

CRMP-2 is associated with tubulin in the mitotic spindle. (a) Confocal imaging of CRMP-2 (green) and tubulin (red) staining by immunofluorescence in nondividing A549 NSCLC cells. (b) CRMP-2/tubulin immunodetection and confocal imaging of mitotic A549 cells. (c) CRMP-2/ tubulin immunoprecipitation assays. (d) Detection of GSK3β phosphorylation in A549 cells. (e) Detection of CRMP-2 phosphorylation on Ser 522, Thr 509 and Thr 514 in A549 cells.

CRMP-2 phosphorylation modifies its binding to the mitotic spindle in dividing cells

CRMP-2 is phosphorylated on Thr 509, Thr 514 and Ser 518 residues by the protein kinase GSK3β, that is inhibited by phosphorylation. CRMP-2 phosphorylation impairs its binding to tubulin. As shown in Figure 2d, the expression of phosphorylated and, therefore, inactive GSK3β, increased 1 hr after nocodazole release, when maximal tubulin/CRMP-2 interaction had been detected.

Next, we performed Western blotting to detect CRMP-2 phosphorylation in the same samples, and as depicted in Figure 2e, CRMP-2 phosphorylation on Ser 522, a Cdk5 target residue, did not change throughout the experiment. This result is a direct consequence of Cdk5 constitutive activation in immortalized cells.23 However, CRMP-2 phosphorylation on the GSK3β targeted residues Thr 509 and Thr 514 reached maximal phosphorylation when the cells were synchronized in prometaphase (Post Noc 0h), decreased 1 hr later, when the culture was enriched in metaphase, and returned to base-line levels 2 hr later, when the majority of the cells proceeded through telophase.

CRMP-2 phosphodefective and phosphomimetic mutants lead to altered duration of mitosis

To fully dissect the importance of CRMP-2 phosphorylation during the cell cycle, we transiently transfected A549 NSCLC cells with CRMP-2-GFP-fusion constructs of phosphodefective (CRMP-2 S522A) and phosphomimetic (CRMP-2 S522D) CRMP2 mutants, as this priming residue allows further CRMP-2 phosphorylation. Immunocytochemistry and immunoprecipitation experiments showed that CRMP-2 phosphodefective mutants interacted more actively with the mitotic spindle (Supporting Information Figure 4A and 4B).

To determine if deregulated CRMP-2 phosphorylation impacts cell division, we transiently transfected A549 cells with the aforementioned constructs, and analyzed their mitosis by time-lapse microscopy. Zero time was established as the moment when cells rounded and chromatin condensation was evident. As shown in Figure 3, cells transfected with wild-type CRMP-2 had the shortest stage of mitosis (33 min vs. 50 min in control A459 NSCLC cells), whereas those cells transfected with CRMP-2 phosphomimetic or phosphodefective expressing mutants did not show any variation in mitosis length compared to cells transfected with empty vector. Separate analysis of the duration of each phase of the cell cycle showed that cells transfected with CRMP-2 or its phosphodefective forms reached anaphase more rapidly (9–10 min) than empty-vector-transfected A549 cells (16 min), presented shorter anaphase and significantly longer cytokinesis. Conversely, cells transfected with phosphomimetic CRMP-2 presented delayed entry into anaphase (25 min). These effects might result from increased microtubule stabilization in those cells that expressed the phosphodefective form of CRMP-2, thereby making it difficult for the cell to progress through anaphase. In contrast, those cells that expressed the phosphomimetic form of CRMP-2 experienced delayed spindle assembly because of decreased CRMP-2 affinity for tubulin.

Figure 3.

Transfection of A549 NSCLC cells with CRMP-2 phosphomutants alters mitosis duration. (a) Duration of mitosis in A549 cells transfected with CRMP-2 phosphodefective (S522A), phosphomimetic (S522D) or wild-type (S522) constructs, and with GFP empty vector (mock). (b) Total mitosis duration in the same cell populations as in A. At least ten cell divisions where analyzed per transfectant. Data obtained were compared against cells transfected with empty vector. *p < 0.05. (c) Quantification of the percentages of cells in each phase of the cell cycle at different time-points after nocodazole removal. At least, 40 cells were counted per each time-point.

To further corroborate these observations, we took a different approach and quantified the number of cells in each phase of mitosis 30, 60 and 90 min after nocodazole treatment in fixed cellular preparations. As summarized in Figure 3c, there existed a delayed metaphase/anaphase transition due to accumulation of cells in metaphase in cells expressing phosphodefective CRMP-2 constructs (55% from total mitotic cells 60 min after nocodazole retrieval, vs. 30% from control cells). This population also presented the smaller counts of cells in anaphase (5% from total cell division) and increased numbers of cells in telophase/cytokinesis (10% vs. 2–5% in the remaining cell populations) most likely due to tubulin stabilization. On the contrary, those A549 NSCLC cells that overexpressed the phosphomimetic form of CRMP-2 had delayed their entry into metaphase (only 10% of total dividing cells 30 min after nocodazole withdrawal), but upon reaching this stage, they proceeded normally.

Cells expressing phosphodefective forms of CRMP-2 have incomplete cytokinesis and p53 activation

Interestingly, we found that 33% of A549 NSCLC cells transfected with phosphodefective forms of CRMP-2 did not complete cytokinesis but reverted to double-nucleated cells (Fig. 4 and Supporting Information Video 1). Because aberrant cell division is associated with p53 activity,24 we determined the expression of p53 and of its downstream effector p21 in A549 cells stably expressing each CRMP-2 construct. As expected, increased levels of p53 and p21 (Fig. 4c) were observed in cells stably expressing CRMP-2 phosphodefective mutants. We, thus, analyzed cell viability under normal and stress conditions. To this end, we serum-starved A549 cells transfected with the CRMP-2 constructs for 48 hr, and analyzed p53 expression and cell apoptosis by annexin/PI staining. The results obtained, summarized in Figures 5a5c, demonstrate an increased p53 expression and apoptotosis and necrosis in cells that expressed the CRMP-2 phosphodefective mutant.

Figure 4.

A549 NSCLC cells transfected with phosphodefective CRMP-2 mutants present aberrant cytokinesis. (a) Representative photograms of time-lapse video microscopy of aberrant mitosis in A549 cells transfected with a CRMP-2 phosphodefective mutant that leads to a binucleated cell. (b) Quantification of the number of aberrant nuclei in A549 cells transfected with different CRMP-2 phosphomutants. One hundred nuclei were counted per condition and compared to cells transfected with empty vector. *p < 0.01. (c) Western blot analysis of p53 and p21 expression in A549 NSCLC cells transfected with CRMP-2 expression vectors and grown under normal serum conditions.

Figure 5.

Cells expressing phosphodefective mutants of CRMP-2 have higher p53 expression and are prone to p53-induced cell death. (a) p53 expression, as detected by Western blotting in A549 cells cultured under serum-deprived conditions for 48 hr. (b) Quantification of the percentage of apoptotic (annexin +/ PI −) and necrotic (annexin +/ PI+) A549 cells transfected with different CRMP-2 mutants and grown under serum free conditions for 48 hr. Statistics show significance compared to cells transfected with empty vector. **p < 0.01. (c) Representative dot plot of annexin/PI staining in serum-starved A549 NSCLC cells.

CRMP-2 silencing induces apoptosis in A549 cells

Next, we aimed to ascertain whether CRMP-2 silencing would also be a good therapeutic approach. First, we tried to stably silence CRMP-2 expression by transfection with shRNA-expressing plasmids, but CRMP-2 shRNA-transfected cells died several days after transfection (data not shown). We then, transiently transfected A549 cells with CRMP-2 siRNA oligonucleotides, and as shown in Figures 6a and 6b, CRMP-2 silencing significantly decreased cell viability in correlation with p53 induction. From these results, we can presume that CRMP-2 expression is crucial for cell survival, as its silencing induced cell death most likely through p53 activation during cell division, which occurred when cells were transiently transfected with the phosphodefective forms of CRMP-2.

Figure 6.

CRMP-2 silencing induces cell death. (a) Cell viability assays performed on A549 cells transfected with CRMP-2 specific siRNA or nonspecific control siRNA. (b) Western- blot analysis for the detection of total and phosphorylated (Thr 514) CRMP-2, cleaved caspase 3, and p53 expression in A549 cells transfected with CRMP2 siRNA or control oligonucleotides.


Lung cancer remains the leading cause of cancer-related death in Europe and the United States. Despite advances in treatment, the prognosis remains poor, with only 15% of patients surviving more than 5 years from the time of diagnosis.25 Drugs that depolymerize microtubules, such as vinca alkaloids, or stabilize them such as taxanes, have been developed as potent anticancer agents26 since they detain aberrant cell division, the primary hallmark of transformed cells.27

In light of this, we were interested in determining the role of CRMP-2, a tubulin adaptor protein, in NSCLC. CRMP-2 expression was recently described in lung,28 but, to our knowledge, there are no reports in the literature examining its role in lung neoplasm. In contrast, other members of the family have been associated with lung cancer. For example, the ratio between the expression of CRMP-1 and its large isoform LCRMP-1 is associated with patient outcome,29 and CRMP-5 has been established as a specific marker for lung neuroendocrine tumors.30

In this study, we analyzed CRMP-2 expression in 91 primary tumors obtained from lung carcinoma patients. The results showed a significant correlation between the expression of the phosphorylated form of CRMP-2 and poor outcome in NSCLC patients, while the expression of total CRMP-2 protein was not associated with survival. Interestingly, we did not observe staining for phosphorylated CRMP-2 in the nucleus of normal epithelium adjacent to the tumor. Therefore, the presence of phosphorylated forms of this protein appears to be a distinctive feature of highly proliferative cells. In accordance to these, we described how dynamics of CRMP-2 interaction with tubulin during mitosis is a result of its function as a tubulin adaptor protein. Strong CRMP-2/ tubulin colocalization with microtubules in pro-metaphase and metaphase occurs, when active polymerization of tubulin takes place to duplicate centrosomes31 and establish the mitotic spindle.32 On the contrary, during anaphase, the CRMP-2 interaction with tubulin is restricted to the mitotic spindle poles, and to some extension, at the spindle midzone. These results are congruent with studies showing intense tubulin polymerization at the cellular poles during anaphase, simultaneous with microtubule depolymerization in the leading kinetochore fibers, which occurs to maintain spindle tension.33 During telophase, we observed CRMP-2 colocalization with tubulin in the intercellular bridge, which is a structure rich in microtubules.34 Prekeris and coworkers demonstrated the need of extensive microvesicle trafficking to allow cytokinesis.35 In this process, tubulin forms intracellular rails where motor proteins cargo membrane vesicles and contractile ring proteins that form the cleavage furrow36 and fuse in cell abscission. In this vein, a number of reports have demonstrated the interaction of CRMP-2 with motor proteins such as dynein9 or kinesin-1,10 both of which are important for the transport of vesicles towards the intercellular bridge. Therefore, the presence of CRMP-2 in the cleavage furrow structure chiefly reflects the dynamic nature of tubulin at these structures.

The spatial and temporal polymerization of tubulin must be finely choreographed with the intervention of MAPs and other microtubule-binding proteins to allow sister chromatid migration towards the cellular poles. Our results from mitosis timing showed that the presence of mutant forms of CRMP-2 alters the duration of mitosis. More interestingly, the expression of CRMP-2 phosphodefective forms in A549 NSCLC cells, deferred them from anaphase and induced p53 expression and incremental cell death. These findings may well be a consequence of the triggering of an “anaphase wait” signal due to alterations in tubulin/chromosome dynamics.37 In addition, we observed an incremental increase in the number of binucleated cells in NSCLC cells that overexpress phosphodefective mutants of CRMP-2. Interestingly, it has been demonstrated that interfering with vesicle recycling to the intercellular bridge leads to binucleated cells.38

In summary, the results described herein present the phosphorylated form of CRMP-2 protein as a putative prognostic marker in NSCLC. Furthermore, we provide a mechanistic explanation for these findings because CRMP-2 participates in cell division in a manner that depends upon its phosphorylation state. Therefore, detecting increased expression of phosphorylated CRMP-2 in transformed cells and its association with mitotic spindle is a reflection of elevated cell proliferation, which makes this protein a druggable candidate for causing cell death in p53-expressing tumors.


The authors are grateful to Drs Y. Ihara and Y. Morishima, (University of Tokyo, Tokyo, Japan) for providing with anti-phosphorylated CRMP2 (3F4) antibody and Manolo Serrano (CNIO, Spain) for providing pcDNA-CRMP-2 expression vector. A Teijeira had a scholarship from the Spanish Ministry of Science and Innovation (MICINN) and R. Peláez a Young Investigator Grant from the Torres-Quevedo Program, Spanish Ministry of Science and Innovation (MICINN).