Ku86 is important for TrkA overexpression‐induced breast cancer cell invasion

Purpose: We have recently shown that breast tumors express high levels of TrkA compared with normal breast tissues, with TrkA overexpression enhancing breast cancer cell invasion in vitro and metastasis in animal models. In this study, we tried to identify molecules involved in TrkA overexpression‐mediated biological effects in breast cancer cells.


Introduction
Several sets of growth factors and their cognate receptors are known to be involved in the regulation of cancer development [1][2][3][4]. Nerve growth factor (NGF) is the prototypic member of the neurotrophin family of proteins, well known for promoting survival and differentiation of neuronal cells during nervous system development. However, accumulating data indicate that NGF is also involved in cancer development [5][6][7]. NGF exerts its effects through two membrane receptors: the tyrosine kinase receptor TrkA and the receptor p75 NTR , a common receptor for all neurotrophins and pro-neurotropins. NGF binding to TrkA induces TrkA receptor dimerization and autophosphorylation of cytoplasmic tyrosines, leading to the activation of various signaling pathways, including the Ras/ MAPK pathway, the PLCg pathway, and the PI3K/Akt pathway [7]. The biological consequences of TrkA activation vary according to cell types. Hence, TrkA activation induces differentiation of neuronal precursors and neuroblastoma cells [8], whereas it induces proliferation in breast adenocarcinoma cells [9,10] and apoptosis in medulloblastoma cells [11].
Recently, we have shown that NGF and its tyrosine kinase receptor TrkA are overexpressed in breast cancers compared with normal breast tissues [12,13]. Inhibition of NGF with neutralizing antibodies or small interfering RNA strongly reduces tumor growth and metastasis of breast cancer cells xenografted in immunodeficient mice [12]. Moreover, TrkA overexpression in breast cancer cells leads to a constitutive activation of its tyrosine kinase, resulting in an increased cell growth and tumorigenicity [13]. Together, these findings point out the importance of the NGF/TrkA axis in breast cancer development. In addition, upregulation of TrkA has also been shown in other cancers including thyroid [14], lung [15], pancreatic [16,17], prostatic [18,19] and ovarian carcinomas [20,21]. Thus, identification of molecules involved in the enhancement of aggressiveness of TrkA overexpressing cancer cells would be important for both better understanding of oncogenesis and research of new molecular targets. Here, we have used a functional proteomic approach to identify molecules involved in TrkAmediated biological effects in breast cancer cells. We first separated proteins of total lysate using a cup-loading 2-DE system and then identified a series of putative modified proteins by MALDI and LC-MS/MS analysis. We found that Ku86, initially described to form a heterodimer with Ku70 to regulate DNA-dependent protein kinase (DNA-PK) that is crucial to DNA repair, was upregulated in TrkA overexpressing cells. Moreover, Ku86 was required for TrkAstimulated invasion of breast cancer cells.

Cell culture
The MDA-MB-231 human breast cancer cell line was from the American Type Culture Collection. MDA-MB-231 TrkA overexpressing cells were stably transfected and characterized in our laboratory [13]. Cells were routinely maintained in EMEM supplemented with 2 mM L-glutamine, 1% nonessential amino acids, 10% FCS, 40 U/mL penicillinstreptomycin, 40 mg/mL gentamycin. All cells were cultured at 371C in a humidified atmosphere of 5% CO 2 .

Sample preparation and 2-DE
Protein extraction was performed as previously described [22]. IEF was carried out using 18 cm Immobiline DryStrips pH 3-10 (GE Healthcare Bio-Sciences). IPG strips were reswollen overnight in 345 mL DeStreak (hydroxyethyl disulfide, GE Healthcare Bio-Sciences) rehydratation solution and 0.2% v/v carrier ampholytes 3-10 (Bio-Lyte s Bio-Rad), under 2 mL mineral oil. Prior to IEF, protein samples (150 mg) were first reduced (1 h) by adding tributylphosphine to a final concentration of 5 mM and secondly alkylated in the dark with 15 mM iodoacetamide for 90 min at room temperature. The samples were then cup-loaded near the anode of the IPG strips and focused in a Protean IEF cell (Bio-Rad) at a temperature of 201C. The IPG strips were initially conditioned for 30 min at 250 V (rapid voltage ramping), linearly ramped to 1000 V (1 h) and maintained at 1000 V for 1 h more. Then the electric voltage was slowly increased to reach 10 000 V in 1 h and focused at this voltage to give a total of 60 kVh. After focusing, the strips were equilibrated for 2 Â 15 min in 6 M Urea, 30% w/v glycerol, 2% w/v SDS, 0.125 M Tris, 0.1 M HCl, containing either 50 mM DTT (first equilibration step) or 150 mM iodoacetamide (second equilibration step) [23]. SDS-PAGE was performed as previously reported [24]. The gel patterns were visualized by silver nitrate staining [25] for analytical purposes or by colloidal CBB G-250 [26] in the case of micropreparative separations.

2-D gel evaluation
Digitized images of 2-D gels were acquired by scanning with a GS-800 calibrated densitometer under control of PDQuest Advanced software version 8.0 (Bio-Rad), which was also used for image analysis and construction of a local 2-D database. Image alignment, spot detection, background removal and expression analysis were performed using PDQuest Advance software (Bio-Rad). Fold changes and all statistical analysis were calculated based on normalized spot volumes where the global spots volume was used to perform normalization. A total of three gels per protein extraction and three extractions from independent experiments were made (nine gels in total per condition) for the study.

In-gel trypsin digestion and protein identification by MS
Coomassie blue-stained protein bands were excised from SDS-PAGE gel and processed for trypsin digestion as previously described [27]. Protein identification was realized using both MALDI-TOF (Voyager DE STR instrument, Applied Biosystems) and NanoLC-NanoESI-MS/MS (LCQ Deca XP1, Thermo-electron, San Jose, CA) as previously described [27]. Database searching was done with SwissProt 56.8 (410 518 sequences; 148 080 998 residues). Carbamidomethylation of cysteine was set as fixed modification, oxidation of methionine was set as a variable modification for all MASCOT searches. To ascertain unambiguous identification, searches were performed in parallel with Phenyx software using the same parameters.

Western blot
After 2-DE or SDS-PAGE separation, proteins were electrotransferred onto nitrocellulose membrane using a semi-dry transfer system (Trans-Blot SD cell, BioRad). Non specific protein binding sites were saturated for 1 h 30 at room temperature in TBS-0.1% Tween-20 reagent (TBST) containing either 5% BSA for Ku70, Ku86 and actin immunodetection. Membranes were then incubated overnight at 41C with 1:5000 anti-Ku70, 1:500 anti-Ku86, 1:500 anti-TrkA or 1:5000 anti-actin antibodies. After washes in TBST, peroxidase-conjugated anti-rabbit or anti-mouse IgG diluted in saturated solution was added for 1 h at room temperature and the membranes were washed several times in TBST before detection of peroxidase activity using chemiluminescent system.

Flow cytometry
Cells detached by trypsin-EDTA solution were incubated for 1 h at 41C with 20mg/mL of the indicated antibodies or matched control isotypes at similar concentrations. After washing with PBS containing 0.5% BSA, cells were incubated for 30 min at 41C with secondary fluorescein-labelled IgG. Cells were then analyzed in Coulter Epics XL/XL-MCl cytometer (Beckman Coulter, Villepinte, France).

Immunocytochemistry and confocal microscopy
MDA-MB-231 cells were seeded on Lab-Tek chamber slides pre-coated with type I collagen. Cells were washed in PBS pH 7.5, fixed in 4% paraformaldehyde for 20 min. Non specific protein binding sites were then blocked in PBS pH 7.5 containing 2% BSA and cells were incubated in blocking solution containing 10 mg/mL rabbit anti-TrkA and 10 mg/mL mouse anti-Ku70 or anti-Ku86 antibodies overnight at 41C. After washes in PBS pH 7.5, 10 mg/mL Alexa Fluor 546 goat anti-rabbit IgG and 10 mg/mL Alexa Fluor 488 donkey anti-mouse IgG were added for 1 h at 371C. Cells were washed in PBS pH 7.5 and mounted. Scanning fluorescence images were acquired using a Zeiss Axiophot microscope.

Invasion assay
BD Falcon inserts with a PET membrane/6.5 mm pores (BD Biosciences) were used for invasion assay. The inserts were pre-coated with GFR Matrigel (1:10 dilution, BD Biosciences). Cells (5 Â 10 4 ) were seeded on polycarbonate membrane insert and maintained in EMEM containing 0.1% FCS. For antibody neutralization, cells were pre-treated with 20 mg/mL of neutralizing antibodies against Ku70 and Ku86 during 30 min before seeding. After 16 h of culture, the insert was washed with PBS, and cells on the top surface of the insert were removed by wiping with a cotton swab. Cells that invaded the Matrigel and migrated to the bottom surface of the insert were fixed with methanol, stained by Hoechst 33258 and then counted on 10 random fields at 200 magnification under a Nikon Eclipse Ti-U fluorescent microscope.

Apotosis analysis
Cells were treated with 5 ng/mL TNF-related apoptosis inducing ligand (TRAIL) for 6 h. Apoptosis was determined by morphological analysis after fixation with methanol (10 min, À201C) and staining with 1 mg/mL Hoechst 33258 (10 min, room temperature, in the dark). A minimum of 500-1000 cells was examined for each case under fluorescent microscope and the results represented the number of apoptotic cells over the total number of counted cells.

Statistical analysis
Statistical significances were determined with two-tailed Student's t tests. All p-values were two-sided. po0.01 was considered as statistically significant.

Ku86 is upregulated in breast cancer cells overexpressing TrkA
Proteins of mock and TrkA overexpressing MDA-MB-231 cells were separated by 2-DE before analysis of protein spots with PDQuest software. A representative example of proteins separated on 2-D gel is shown in Fig. 1A.
Nearly 1500 spots were obtained in the ranges of MW Spots indicated in Fig. 1 were cut out of the gel and digested with trypsin before MALDI-TOF analysis. MASCOT search program was used to identify proteins. The table lists all the 25 peptides identified by MALDI-TOF. The underlined peptides are also identified by nanoLC-nanoESI-MS/MS. 12-120 kDa and pI 3-10. In TrkA overexpressing cells, more than 20 spots were found to be down-or upregulated (at least two-fold changes with po0.05) after analysis with PDQuest software. Proteins were identified by MALDI-TOF and LC-MS/MS. Among a dozen of differentially expressed proteins which remain to be validated, a significant increase of Ku86 protein was observed in TrkA overexpressing cells. Interestingly, Ku86 was found as a set of four close spots in empty vector transfected cells (mock) and TrkA overexpressing cells (Fig. 1B). MALDI-TOF spectrum showed 25 experimental tryptic peptides that matched to theoretical masses, leading to 32.5% sequence coverage with an average error mass of 0.034 Da (Fig. 1C, Table 1). This identification was consolidated by independent MS/MS analysis of corresponding spot from 2-D gel, as sequencing of five peptides revealed 8% sequence coverage with an average error mass of 0.295 Da ( Fig. 1D and E, Table 2). Importantly, Western blotting analysis also showed a similar increase of Ku86 in TrkA overexpressing cells ( Fig. 2A  and B), thus validating the 2-DE and MS analysis. However, RT-PCR showed no modification of Ku86 mRNA level (Fig. 2C), indicating that Ku86 upregulation is post-translationally controlled.

Ku86 is co-immunoprecipitated with TrkA
Ku86 and Ku70 can be associated to form a heterodimeric regulatory subunit of the DNA-PK that is crucial to DNA repair [28]. Ku proteins are also reported to be involved in cell proliferation, migration and invasion [29]. More recently, we have shown that Ku70 interacts with TrkA in MCF-7 breast cancer cells to stimulate cell survival in TRAIL-induced apoptosis [27].  . 3A and B), with a MASCOT score of 208 and 4% coverage (blanket). On the other hand, two peptides corresponding to Ku70 were also sequenced (586.71 and 568.63), with a Mascot score of 48 and 3% of coverage (Fig. 3C). Moreover, Western blot confirmed the presence of both Ku86 and Ku70 in TrkA co-immunoprecipitated proteins but not in isotype IgG co-immunoprecipitated proteins (Fig. 3D), suggesting the specific association of Ku proteins to TrkA.

Membrane Ku86 and Ku70 are increased in TrkA overexpressing cells
Increased levels of Ku86 in TrkA overexpressing cells as well as the association of Ku proteins with TrkA prompted us to determine the levels of membrane Ku proteins by flow cytometry analysis. As shown in Fig. 4A and B, an increase of membrane Ku86 and Ku70 was observed in TrkA overexpressing cells compared to empty vector transfected cells. We then analysed subcellular distribution of these proteins in TrkA overexpressing cells by confocal microscopy after immunochemical staining (Fig. 4C). TrkA (blue fluorescence) and Ku proteins (green fluorescence) seemed to be mainly located in the nucleus and cytoplasmic membrane, though a faint and punctuate staining was also observed in the cytoplasm.

Ku86 is involved in TrkA-induced cell invasion
We have previously shown that TrkA overexpression leads to an increase in invasion and survival of breast cancer cells [13]. To determine whether Ku proteins were implicated in  TrkA-overexpression-induced biological effects, we inhibited Ku proteins by siRNA or neutralizing antibodies. As shown in Fig. 5A, specific siRNAs strongly decreased the expression of Ku86 and Ku70. Inhibition of Ku86 with both siRNA and the neutralizing antibody efficiently reduced invasion of TrkA overexpressing cells, whereas inhibition of Ku70 had no effect ( Fig. 5B and C). Interestingly, simultaneous inhibition of both Ku86 and Ku70 reduced cell invasion at a similar degree to that observed upon Ku86 inhibition alone. We then evaluated if Ku proteins could regulate apoptosis induction in TrkA overexpressing cells. For this, we transfected cells with siKu and then treated them with TRAIL (TNF-related apoptosis inducing ligand), a cytokine well known for apoptosis induction in breast cancer cells. As shown in Fig. 5D, siKu86 had no effect on apoptosis of cells whatever TRAIL treatment. In contrast, siKu70 induced apoptosis even in the absence of TRAIL and could further increase apoptosis induction by TRAIL, confirming our previous finding in MCF-7 breast cancer cells [27]. When cells were co-transfected with siKu86 and siKu70, no significant difference was observed compared to siKu70 alone-transfected cells. All together, these results indicated that Ku proteins functioned independently of each other to mediate TrkA overexpression-induced biological effects: Ku86 was only involved in cell invasion whereas Ku70 was implicated in cell survival.

Discussion
We have previously shown that TrkA overexpression increases breast cancer cell growth, invasion, as well as survival. In order to understand the underlying mechanisms, we have used a functional proteomic approach to identify molecules involved in the TrkA-mediated biological effects. We first separated proteins using 2-DE system and then identified a series of putative modified proteins in TrkA overexpressing cells by MALDI and LC-MS/MS analysis. The cup-loading technology reduces the inter-experimental variations, allowing rapid identification of protein changes between samples on reproducible 2-D gels [30]. Among a dozen of differentially expressed proteins, which remain to be validated, we observed a significant increase of Ku86 in TrkA overexpressing cells, as revealed by direct quantification of corresponding spots in 2-D gels and Western blot analysis. This is of particular interest, as we have recently identified another member of the Ku family of proteins, namely Ku70, as a partner of TrkA signaling in breast cancer cells [27]. Indeed, we have shown that NGF treatment induces tyrosine phosphorylation of Ku70 upon its association to TrkA. Moreover, Ku70 is involved in TrkA-enhanced cell survival [27]. Here, we only found that Ku86 was upregulated in TrkA overexpressing cells, although both Ku70 and Ku86 were co-immunoprecipitated with TrkA. Visualisation of Ku86 as four close spots in 2-D gels may be due to post-translational modifications, as Ku proteins have been described to be acetylated and phosphorylated [27,31,32]. In spite of the increase of protein levels, we were not able to observe any significant variation of Ku86 mRNA level by RT-PCR, suggesting a mechanism of regulation at posttranscriptional level. Reinforcing this hypothesis, it has been recently described that VEGF can activate AKT which in turn inhibits Ku70 proteolysis by phosphorylating Hdm2, the ubiquitin ligase of Ku70 [33]. Thus, further studies will be needed to examine the exact underlying mechanism of Ku86 upregulation in our model. The fact that Ku86 was upregulated in TrkA overexpressing breast cancer cells is in line with previous demonstration that Ku86 is expressed in abundant levels in tissues with a high proliferative index or in cells stimulated to proliferate [34,35]. Moreover, upregulation of Ku proteins has been associated with the progression of some types of tumors. For instance, the levels of Ku86 are positively correlated with that of anti-apoptotic Bcl-2 in B cell chronic lymphocytic leukemia [36]. Ku86 has also been reported to be upregulated in bladder, breast and primary hepatocellular carcinomas, compared to adjacent non tumorous tissues [37][38][39]. Ku86 and Ku70 are mainly localized in the nucleus, where they form heterodimers to recruit the catalytic subunit of DNA-PK, which is involved in multiple biological processes such as DNA double-strand break repair, telomere length maintenance, cell cycle progression and transcriptional regulation [32,40]. However, cytosolic and membrane Ku proteins are increasingly reported to exert different functions independently of each other. Cytosolic Ku70 has been shown to bind to the pro-apoptotic protein Bax and inhibit Bax-mediated apoptosis by preventing its translocation to mitochondria. This anti-apoptotic function is mediated by a domain in the carboxyterminal of Ku70 and does not require the cooperative effects of Ku86 [41]. Accordingly, here we showed that Ku70 but not Ku86 was involved in the increased survival of TrkA overexpressing breast cancer cells. On the other hand, membrane Ku proteins have been reported to be associated with cell adhesion and migration [42,43]. It has been shown that hypoxia-stimulated invasion of neuroblastoma and breast carcinoma cells involves upregulation of membrane Ku86 [44]. Similarly, Ku proteins are found to interact with matrix metalloproteinase 9 at the membrane of highly invasive normal and tumoral hematopoietic cells [45]. Translocation of Ku proteins from the nucleus to the plasma membrane can enhance migration of monocytes [46]. In this study, we found an increase of membrane Ku86 and Ku70 in TrkA overexpressing cells. Moreover, inhibition of Ku86 with siRNA or neutralizing antibodies strongly reduced TrkA-stimulated invasion, indicating that Ku86, especially membrane Ku86, was involved in this process. Our previous work shows that activation of signalling pathways including PI3/AKT and MAP kinases is required for TrkA overexpression-enhanced cell invasion and survival [13]; here we showed that Ku86 was involved in TrkA-overexpression-stimulated cell invasion while Ku70 was only implicated in TrkA-overexpression-enhanced cell survival. Thus, it will be interesting to determine the link between classical signaling pathways and the specific implication of Ku proteins in these processes. In conclusion, we showed by functional proteomic exploration that Ku86 is upregulated in TrkA overexpressing breast cancer cells and is involved in TrkA-induced cell invasion. Therefore, upregulation of Ku86 in tumor cells overexpressing TrkA might be a mechanism leading to an increase of metastasis. Although further in vitro and in vivo investigations will be required to test this hypothesis, these data reveal Ku86 as a new potential player in the intracellular signaling leading to breast cancer cell metastasis.
The authors have declared no conflict of interest.

Clinical Relevance
Increased receptor tyrosine kinase activity has been found in a number of different human cancers, motivating development of anti-cancer therapies designed to abrogate their functional contribution. Thus, Herceptin has been clinically used to target the tyrosine kinase receptor Erb-B2. However, less than 15% of patients with breast cancer are Erb-B2 positive, critically limiting the impact of Herceptin treatments. Better understanding of the involvement of other tyrosine kinase receptors in cancer development is therefore important for therapeutic improvements. The Trk family of neurotrophin receptors is emerging as an important player in the development of several types of cancers, including breast cancer. We have recently shown that breast tumors express high levels of TrkA compared to normal breast tissues, with TrkA overexpression enhancing breast cancer cell invasion in vitro and metastasis in animal models. In this study, we showed by functional proteomic exploration that Ku86 is upregulated in TrkA overexpressing breast cancer cells and is involved in TrkA-induced cell invasion. Therefore, upregulation of Ku86 in tumor cells overexpressing TrkA might be a mechanism participating to metastasis. Our findings suggest that TrkA and its down stream signalling pathways should be regarded as potential new targets for breast cancer therapy.