Drugs and reagents

Syk non-enzymatic inhibitor C-13 was purchased from ChemBridge, Inc (San Diego, CA) (ID number 6197026) and was dissolved to a final concentration of 10 mM in dimethylformamide (DMF). Syk catalytic inhibitor R406 was from Santa Cruz Biotechnology and was dissolved to a final concentration of 10 mM in dimethyl sulfoxide (DMSO). Syk inhibitor Piceatannol was from Sigma-Aldrich and was dissolved to a final concentration of 40 mM in DMSO. Recombinant human EGF (1 mg/mL) was from Sigma-Aldrich. Cetuximab (ERBITUX^®) (5 mg/mL) was purchased from Merck (Darmstadt, Germany). All stock solutions were aliquoted and stored at -20C. All reagents unless otherwise mentioned were from Sigma (St Louis, Mo).

Cell culture

All cell lines were from the ATCC and were cultured in antibiotic- and antimycotic-free medium (Gibco). The human colorectal adenocarcinoma HCT-116, HT29, SW480 and SW620 were grown in RPMI Glutamax medium, and DLD-1 cell line was grown in DMEM/F12 (50/50) Glutamax medium. All culture media were supplemented with 10% (v/v) FBS. The DIFI cell line was grown in RPMI Glutamax medium supplemented with 20% (v/v) FBS.

For EGF induction experiments, cells were serum-starved for 12 hours, and they were subsequently treated with the recombinant human EGF (50 ng/mL) or with EGF (50 ng/mL) + Cetuximab (50 μg/mL) in serum free culture medium, for 10 minutes.

Cell propagation and passaging were as recommended by the American Type Culture Collection (ATCC). For cell enumeration, cells were detached with TrypLE Express Enzyme (Gibco), and they were counted using Z1 Particle Counter (Beckman Coulter). The cell lines were tested and authenticated by short-tandem repeat profiling (LGC Standards and Eurofins Genomics). All experiments were performed at least three times.

Syk silencing by small hairpin RNA

To design long isoform-specific and global Syk specific shRNA, we used siRNA sequences from Pinos et al. [6]. Four different shRNA expressing pSIREN vectors with the puromycin selection marker and specific to Syk were used. shRNA G1, G2 inhibit global gene expression and shRNA L1, L2 are alternative exon-specific. shRNA LUC with Luciferase sequence was used as negative control. Lentiviral production was performed by cotransfecting 1x10⁶ 293T cells (for lentiviral packaging) with 3 μg of pSiren vector in which anti-Syk shRNAs where cloned, 1 μg of the packaging vector gagpol, and 1 μg of envelope vector, using JetPRIME (Polyplus), according to the manufacturer’s instructions. Lentiviral particles were harvested, filtered and then used to infect 1x10⁶ cells for 48 hours. Following shRNA transduction, cells were selected with 1 μg/mL puromycin for one week and stable clones were pooled.

shRNA target sequences

shSykG1: 5’ GCAGATGGTTTGTTAAGAG 3’; shSykG2: 5’ GTCGAGCATTATTCTTATA 3’; shSykL1: 5’ GTTCCCATCCTGCGACTTG 3’; shSykL2: 5’ GGTCAGCGGGTGGAATAAT 3’; shLUC: 5’ TTACGCTGAGTACTTCGA 3’.

RNA extraction and quantitative PCR

Total RNA extractions were performed from 0.5 to 1.10⁶ cells using the ZR RNA MiniPrep Kit (Zymo Research) as recommended by the manufacturer. Reverse transcription was performed on 1 μg RNA using PrimeScript Reverse Transcriptase (Takara Bio.), SYBR Premix Ex Taq 2X master mix (Takara Bio.) and 5 pmol of both forward and reverse primers. Quantitative PCR reactions were performed on 1 ng cDNA. HPRT was amplified as internal control.

qPCR primer sequences

1. Syk L: for 5’ TCAGCGGGTGGAATAATCTC 3’; rev 5’ TGCAAGTTCTGGCTCATACG 3’
2. Syk S: for 5’ TGGCAGCTAGTCGAGCATTA 3’; rev 5’ CAGGGGAGGACGCAGGAT 3’
3. Syk L+ S: for 5’ GAAGCCATATCGAGGGATGA 3’; rev 5’ CCACATCGTATGTCCAGCAC 3’
4. HPRT: for 5’ CTGACCTGCTGGATTACA 3’; rev 5’ GCGACCTTGACCATCTTT 3’

Cell growth inhibition assay (2D assay)

Cell growth was evaluated using the CellTiter-Glo^® Luminescent Cell Viability Assay (Promega) according to manufacturer’s recommendations. Briefly, 3,000 cells/well were seeded in 96-well plates. After 24 hours, drugs were added in serial dilution. Cells were incubated for 96 hours, after which, a volume of CellTiter-Glo^® reagent equal to the volume of cell culture medium was added to each well. The plate was shaked on an orbital shaker for 15 minutes at room temperature to induce cell lysis. The luminescence was recorded using a Thermo Fisher Scientific Multiskan EX plate reader. The IC50 was determined graphically from the cell growth curves.

Migration

DLD-1 or HCT-116 cells were serum starved for 16 h in a 1%-FBS-medium. Cells were detached with TrypLE Express Enzyme (Gibco). 5x10⁴ cells were resuspended in a 1%-FBS-medium and they were seeded on Fluoroblok 24-well plate permeable inserts, with 8 μm pores (Corning). Following a 20-22h incubation with chemoattractant (10% FBS in RPMI medium), the inserts were transferred in a 2^nd 24-well plate containing a 4 μg/mL Calcein AM solution (Sigma Aldrich). The fluorescence was read on an inverted fluorescence microscopy and migrated cells were counted with ImageJ software.

Invasion

Same as migration, but the Fluoroblok insert was coated with 100 μL of 300 μg/μL BD Matrigel Matrix (BD Biosciences).

Apoptosis assay

A total of 2.5x10⁵ cells were plated, and after 24 hours, they were incubated with the indicated concentrations of compound C-13 for 2 hours. ShRNA-tranduced cells were plated following one week puromycin selection, and grown for 48 hours. Cells were stained with APC-labeled Annexin V and 7-aminoactinomycin D (Biolegend). For apoptosis determination, Annexin V- and 7-ADD-positive cells were quantified using a Gallios Cytometer (Beckman Coulter) and the Kaluza Software (Beckman Coulter).

Cell-cycle analysis

To determine the cell-cycle distribution, 2.5x10⁵ cells were plated in 25 cm2 flasks. After 48 hours, cells were synchronized by serum starvation during 12 hours. Quiescent cells were induced to re-enter the cell cycle by serum refeeding. After 15 hours, cells were washed in ice-cold PBS, fixed in 75% ethanol, and labeled with 40 mg/mL propidium iodide (Sigma-Aldrich) containing 100 mg/mL RNase A (Sigma). Cell-cycle progression was analyzed by flow cytometry with a Gallios Cytometer (Beckman Coulter) and quantified using the Kaluza Software (Beckman Coulter).

Immunofluorescence

5x10⁴ cells per well were seeded on Lab-Tek II glass chambers (Thermo Fischer Scientific). After 24 hours, cells were washed with ice-cold PBS, and they were fixed and permeabilized with ice-cold methanol (Sigma Aldrich) for 15 minutes at -20°C. Cells were incubated in PBS containing 2% BSA for 30 minutes at room temperature. Cells were incubated with the primary and the fluorescent antibodies for 1h at room temperature. The primary antibodies used were: rabbit polyclonal Syk (1:100) (sc-1077, Santa Cruz Biotechnology); mouse monoclonal anti-beta Tubulin (1:200) (T4026, Sigma). Primary antibodies were detected by Alexa Fluor 488-conjugated anti-mouse IgG (1:400) (715-546-151, Jackson ImmunoResearch laboratories) and APC-conjugated anti-rabbit IgG (1:200) (711-136-152, Jackson ImmunoResearch laboratories) antibodies, respectively. Nuclei were stained with Hoechst 33342 dye. After washing, the slides were mounted in VECTASHIELD Mounting medium (H-1400, Vector laboratories) and were analyzed visually using Zeiss Imager M2 microscope.

Patients and ethics statement

Colorectal cancer patients with synchronous and unresectable liver metastases were enrolled in a prospective study at the ICM Cancer center from January 2000 to June 2004 [24]. Normal colon, colon cancer and hepatic metastasis samples were collected at the time of surgery, prior to chemotherapy. The study was approved by ICM (Institut du Cancer de Montpellier) CORT (Comité de Recherche Translationnelle) ethical committee and all participating patients were informed of the study and had to provide signed written informed consent before enrollment. The ethics approval to use those cancer samples was provided by the ICM/CORT IRB committee, which collected the consent of all patients. All experimental protocols using these human samples were carried out in accordance with the French Guidelines and Regulations for Human Samples.

Tissue samples

From 13 CRC patients mentioned above, we collected pairs of colon primary tumors (CT) and normal colon mucosas (CN). These 13 pairs of samples were used to study the protein expression pattern of Syk expressed in CT versus the corresponding CN. Prior to western blot analysis, tumor tissue samples from patients were directly grinded in lysis buffer (150mM NaCl, 10mM Tris pH 7.4, 1mM EDTA, 1mM EGTA, 1% SDS, 1% Triton X-100, 0.5% NP-40, 2mM PMSF, 100mM NaF, 10mM sodium orthovanadate, one cocktail protease inhibitor tablet for 10 mL) using a Mixer Mill^® MM 300 unit (Qiagen, Valencia, CA). Protein concentration was determined with the Bradford assay (Pierce Coomassie Plus Protein Assay). Then, 50 μg of total proteins solubilized in 1x Laemmli sample buffer (10% glycerol, 5% β-mercaptoethanol, 2.3% SDS, 62.5 mM Tris-HCl (pH 6.8), and 0.1% bromophenol blue) were resolved by 10% SDS-PAGE.

Cell culture samples

Cells were solubilized in DOC modified lysis buffer (1% Igepal, 0.25% sodium deoxycholate, 0.1% SDS in PBS buffer supplemented with protease and phosphatase inhibitors) and the protein concentration was determined (PIERCE BCA Protein Assay, Thermo Fisher Scientific). Equivalent total protein amounts (50 μg per sample) solubilized in 1x Laemmli sample buffer were subjected to 8–10% SDS-PAGE.

Immunoblotting

Following electrotransfer to Amersham nitrocellulose membrane (GE Healthcare, pore size 0.45μm) using a standard protocol, the membranes were first stained with Ponceau S (Sigma), to confirm protein loading equivalence, then blocked 1 hour at room temperature with gentle shaking in TBS containing 0.1% Tween-20 (TBST) and 5% skimmed milk. Primary antibodies were diluted in blocking solution and hybridized overnight at 4°C with gentle shaking. The antibodies used were: mouse monoclonal anti-Syk (1:500) (4D10, #sc1240 Santa Cruz Biotechnology); rabbit polyclonal anti-Phospho-Akt (1:2,000) (#4060, Cell Signaling Technology); mouse monoclonal anti-Phospho-p44/42 MAP Kinase (1:2,000) (#9106, Cell Signaling Technology); rabbit polyclonal anti-p44/42 MAP Kinase (1:1,000) (#9102, Cell Signaling Technology); mouse monoclonal anti-beta Tubulin (1:5,000) (SAB4200715, Sigma); rabbit polyclonal anti-EGF Receptor (1:1,000) (#2232, Cell Signaling Technology); goat polyclonal anti-Akt (1:200) (sc-1618, Santa Cruz Biotechnology); HRP-linked anti-goat antibody (1:5,000) (sc-2033, Santa Cruz Biotechnology); HRP-linked anti-rabbit antibody (1:5,000) (sc-2004, Santa Cruz Biotechnology); HRP-linked anti-mouse antibody (1:5,000) (sc-2314, Santa Cruz Biotechnology).

Membranes were washed 5 times (10 min. each) in TBST and appropriate HRP-conjugated secondary antibodies were diluted in blocking solution and hybridized 1 hour at room temperature with gentle shaking. After 5 washes in TBST, the proteins were detected by chemiluminescence generated with Western Lightning Chemiluminescence Reagent Plus (GE Healthcare) and detected on PXi Syngene Chemiluminscence system. For reprobing, membranes were incubated at room temperature for 10 min in stripping solution (100 mM Glycine pH 2.8; 0.1% Igepal and 1% SDS) with gentle shaking. Membranes were then washed (3 times, 5 min. each) with TBST and blocked as described previously, before rehybridization.

Mass spectrometry

In vivo studies

Syk isoforms expression in CRC cell lines

We assessed the expression levels of total Syk (L+S) as well as Syk (L) and Syk (S) isoforms in six CRC cell lines by RT-qPCR (Fig 1A). This analysis showed that these cell lines express different levels and splicing isoform ratios of Syk: the expression levels of global Syk (L+S) and that of Syk long isoform transcript are 5 to 4000 fold higher in HT29 and DIFI cell lines compared to the four other cell lines. Among the six cell lines, DIFI expresses the higher level of Syk short isoform transcript, which is 8 to 1800 fold higher than the other cell lines. The mRNA level is well correlated with the protein expression level detected by western blot analysis (Fig 1B), except for SW620 that expressed a high level of protein despite a low level of mRNA, suggesting that unknown post-translational mechanisms may affect Syk expression. However, we could not visualize at the protein level, the two Syk isoforms due to their co-migration by SDS-PAGE and the absence of suitable isoform-specific antibodies.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Targeting of Syk alternative spliced isoforms.

(A) The relative expression of Syk (L), Syk (S) and total Syk (L+S) in colorectal carcinoma cell lines was detected by qPCR; HPRT was used as an internal control. The Y-axis of the plot is expressed in a base-2 log scale. The figure is representative of three independent experiments. (B) Top panel- The expression of total Syk was evaluated by western blot. Bottom panel- Densitometric quantification of immunoblot analyses of Syk expression over control. The impact of shRNAs targeting the pre-mRNAs encoding Syk proteins evaluated by: (C) RNA expression monitored by qPCR, (D) RT-PCR; HPRT was used as an internal control, and (E) western blot, and compared to mock transduced (shLUC) HCT-116 (panel Ea), DLD-1 (panel Eb) and DIFI (panel Ec) CRC cells. Experiments were performed in three biological and three technical replicates, and an average for each cellular phenotype was calculated for each shRNA. (E) Bottom panel- Densitometric quantification of Syk expression over control, for immunoblots shown in panels Ea, Eb and Ec. shRNA G1, G2 inhibit global gene expression & shRNA L1, L2 are alternative exon-specific. shRNA LUC with Luciferase sequence was used as negative control. The qPCR relative expression values for the overall gene expression Syk (L+S), long splice isoform (Syk L) and short splice isoform (Syk S) were normalized against HPRT used as control. Error bars represent the mean ± SD of three independent experiments (*, P<0.05; **, P<0.01; ***, P<0.001, compared with control). For the western blots, total proteins were extracted one week post-transfection and equal amount of proteins were loaded on an 8% SDS-PAGE. The anti-Syk and anti-Tubulin antibodies used are described in materials & methods. Tubulin served as a loading control.

https://doi.org/10.1371/journal.pone.0274390.g001

Syk long isoform is implicated in CRC cell lines survival and mitosis

Our goal was to evaluate the role of Syk alternative spliced isoforms in cancer biology of colorectal carcinomas. Short hairpin RNA (shRNA) and RNA interference (RNAi) have become preferred tools for evaluating the impact of gene expression on the viability of cancer cells, and for understanding individual isoform functions. Compared to RNAi, shRNA can be integrated into genomic DNA for long-term or stable expression, leading to a longer knockdown of the target mRNA and a durable down-regulation of protein expression, thus allowing the evaluation of protein knockdown effects on cellular functions.

For our purpose, we used shRNA-based isoform-specific silencing to investigate the effects of Syk splicing modulation on cell growth and cell survival of HCT-116, DLD-1 and DIFI colorectal cell lines. We compared the effects of shRNA on global and long-isoform specific expression by RT-qPCR and RT-PCR (Fig 1C and 1D). Global shRNAs (shG1, shG2) equally inhibited the expression of long and short splice isoforms of Syk. Long Isoform specific shRNAs (shL1, shL2) inhibited the accumulation of the long splice isoform while proportionally reducing the overall gene expression level. Interestingly, Syk (L) depletion increased the expression of the short isoform to 1.6 fold in DLD-1 and to 2.8 fold in DIFI cell line. These observations were confirmed by western blot analysis using antibodies against total Syk protein to evaluate the impact of the shRNA-induced mRNA downregulation at the protein level (Fig 1E). Our results showed that among the global shRNAs, shG1 and among the isoform specific shRNAS, shL2 were the most efficient ones for reducing Syk gene expression both at the mRNA and at the protein levels.

Syk has been previously described as implicated in cell-survival pathways and apoptosis. We evaluated the impact of shRNA-based global and long-isoform specific silencing of Syk on the viability of CRC cells. Inhibiting the expression of the long isoform of Syk resulted in an increase of apoptotic marker-positive cells in the three cell lines (Fig 2). Accordingly, cell proliferation and the ratio of viable cells were markedly decreased by shRNA-mediated silencing of the long isoform of Syk (Fig 2). On the opposite, Syk general knock down with global shRNA shG1 did not significantly affect cell apoptosis and viability. However, it increased cell proliferation of HCT-116 and DLD-1 cells, and had the opposite effect in DIFI cells where Syk general knock down reduced cell proliferation (Fig 2). Transfection of an unrelated shRNA (shLUC) did not affect any of these cellular phenotypes. We concluded that changing the splicing pattern of Syk influences CRC cell viability and apoptosis, in a manner that is independent of the overall level of gene expression.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Changing the splicing pattern of Syk affects cell viability and induces apoptosis.

The impact of shRNAs targeting global Syk (shG1) & long-isoform-specific Syk (shL2) pre-mRNAs on cellular phenotype evaluated by: cell proliferation, viability and apoptosis, and normalized to unrelated-transduced shRNA (shLUC) HCT-116, DLD-1 and DIFI CRC cells. Cell numbers were monitored using cell counts (Countess II, Promega), apoptosis was monitored using Annexin V-FITC marker and analyzed by flow cytometry (Gallios Beckman Coulter) and viability was monitored using CellTiter-Glo Luminescent cell viability assay (Promega). The relative changes were calculated and used to generate the bar graphs. The error bars represent the variation observed in three biological and three technical replicate (*, P<0.05; **, P<0.01; ***, P<0.001, compared with control).

https://doi.org/10.1371/journal.pone.0274390.g002

K-Ras mutations occur as an early event in about 50% of CRC cases, resulting in constitutive activation of the RAS-RAF-MEK-ERK pathway, and a subsequent resistance to anti-EGFR therapy by monoclonal antibodies Cetuximab and Panitumumab. Thus, we focused our studies on the role of Syk in HCT-116 and DLD-1 cell lines, which both bear the activating K-Ras (G13D) mutation, and which express low and high levels of Syk, respectively.

Previous works reported the implication of Syk in the control of cell cycle and mitosis [9,10,28]. To further investigate the effect of Syk alternative splicing on cell proliferation, we monitored the impact of shRNA-based Syk targeting on cell-cycle progression of HCT-116 and DLD-1 cell lines. Our data showed that changing the alternative splicing of Syk led to the accumulation of cells in the G2-M phase, consistent with a defective mitosis, and to the emergence of cells with a >4N DNA content due to polyploidization through mitotic skipping (Fig 3A). This was particularly true for DLD-1 cells, which were more affected than HCT-116 in their proliferation and survival (Fig 2).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Syk (L) isoform regulates cell-cycle progression and cytokinesis.

(A) DLD-1 cells were transduced with Syk shG1 and shL2 shRNAs or shLUC (control shRNA), and the cell cycle distribution of cells one week post-infection was analyzed using flow cytometry. The accumulation of transduced cells in G0-G1, S, G2-M and > 4N phases of each cell line is normalized to its control and is shown as a bar graph. The data are an average of three independent transduction experiments and error bars represent the mean ±SD (*, P<0.05; **, P<0.01). (B) Representative fluorescence microscopy images of DLD-1 cells in mitosis. DLD-1 cells were transduced with shLUC (control shRNA) (panel B1), Syk shG1 shRNA (panels B2-B3) and shL2 shRNA (panels B4-B5). One week post-infection, cells were cultured for 24 hours in Labtek glass chambers and stained with anti-Syk antibody (APC, red), anti-beta Tubulin (Alexa 488, green). Nuclei were stained with Hoechst 33342 dye (blue). Slides were analyzed visually using Zeiss Imager M2 microscope (system magnification: 60x). Two-color fluorescence merge images of representative fields show normal size nuclei and normal bipolar spindles in control cells (panel B1), while cells tranduced with Syk shG1 (panels B2) and shL2 shRNAs (panel B4) showed abnormal metaphases with aberrant accumulation of unaligned chromosomes in midzone (red arrows) and abnormal multipolar spindles, enlarged nuclei and abundant Syk expression (panels B3 and B5). Similar results were observed in three independent experiments. Scale bar: 10 μm.

https://doi.org/10.1371/journal.pone.0274390.g003

Upon the visual examination of DLD-1 cells expressing shRNA that target Syk global and isoform specific expression, we noted an increasing amount of mitotic cells with abnormal metaphases and chromosome alignment aberrations in comparison to cells expressing unrelated shRNA (Fig 3B). We also noted an increasing amount of “giant cells” displaying large and aberrant nuclei with supernumerary centrosomes and multipolar anaphase spindles confirming the role of Syk in mitotic exit and cytokinesis (Fig 3B). This mitotic arrest was particularly visible following Syk long-isoform knockdown, but was also present following Syk general knock down with global shRNA, indicating that Syk (L) controls mitotic exit and cytokinesis. Taken together, our data suggest that Syk (L) plays an important role in the cell survival and the control of the cell cycle in CRC cell lines.

Syk inhibition increases chemotaxis and Matrigel engagement

To investigate any correlation between Syk expression and cell motility, HCT-116 and DLD-1 cell lines were tested in a Fluoblok chemotaxis assay. We monitored the impact of Syk knockdown on cell motility, using shRNA-based Syk global and isoform specific expression. Chemomigratory ability was significantly increased in cells expressing shRNA that target Syk global expression relative to cells expressing unrelated shRNA, while Syk long-isoform knockdown did not affect cell migration ability (Fig 4A).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Syk inhibition increases chemotaxis and Matrigel engagement.

(A) HCT-116 and DLD-1 cells (5x10⁴) transduced with shRNAs targeting global Syk (shG1) & long-isoform-specific Syk (shL2) or unrelated-transduced shRNA (shLUC), and (B) cells pretreated with Piceatannol or vehicle (DMSO) were serum starved, labelled with CellTracker dye, and migrated toward FCS in Fluoroblock chambers. (C) HCT-116 and DLD-1 cells (5x10⁴) transduced with Syk shRNAs or control shRNA were serum starved, labelled with CellTracker dye and seeded on Matrigel-coated Transwell inserts of 8 μm pore size (BD Falcon). FCS was used as chemoattractant in the Transwell lower part. After 24 hours at 37°C, migration was observed and counted microscopically. Images were obtained and total number of cells per field of view was measured using ImageJ software. Error bars represent the mean number of migrated cells in triplicate wells, representative of three independent experiments (*, P<0.05).

https://doi.org/10.1371/journal.pone.0274390.g004

Next, we used the pharmacologic inhibition of Syk to further study its role in cell motility. Piceatannol, a hydroxystilbene derivative of resveratrol, preferentially inhibits the kinase activity of Syk in in vitro assays and is widely used as a Syk-selective inhibitor. We used a non-toxic concentration range of Piceatannol (up to 12.5 μM) commonly used in studies using human cells. As shown in Fig 4B, Piceatannol increased FCS-induced chemomigration of both HCT-116 and DLD-1 cell lines to the same manner than shRNA-based Syk global knockdown.

Cell attachment and spreading on matrix proteins is a key step in invasion. When cultured on Matrigel, many tumor cells attach and elongate, and this, together with motility, results in formation of cordlike structures and ultimately invasion. We used this experimental metastasis model to evaluate the role of Syk in colorectal carcinoma invasion and metastasis. As shown in Fig 4C, DLD-1 cell lines expressing shRNA that target Syk global expression, as well as long-isoform specific shRNA, displayed increased cell invasion compared to cells expressing control shRNA. This tendency was also observed in Matrigel invasion of HCT-116 cell lines expressing both shRNA, however, it did not reach a significant statistical value because of the high variability of this experiment.

Together, our data indicated that global Syk knock down increases cell proliferation and cell motility, while the absence of Syk (L) isoform affects cell viability and induces cell apoptosis.

Syk expression is required for the activation of PI3K-Akt survival pathway

The epidermal growth factor receptor (EGFR) signaling pathway is commonly activated in colorectal cancer. Activation of this pathway occurs after ligand binding to EGFR, which leads to EGFR phosphorylation and oligodimerization at the plasma membrane. This in turn triggers a chain of downstream signaling events that include activation of the Ras/mitogen-activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K)-Akt pathways that are crucial mediators of growth factor-induced proliferation and cell survival, respectively. Previous studies have shown that the presence of activating PIK3CA and Ras mutations in CRC cell lines result in constitutive activation of these pathways downstream of EGFR signaling.

Considering the presence of activating K-Ras (G13D) mutation in both HCT-116 and DLD-1 cell lines, as well as PIK3CA (H1047R) mutation in HCT-116 cell line, we investigated the effects of Syk targeting on EGFR ligand-induced activation of Erk and Akt. For this purpose, HCT-116 and DLD-1 cells transduced with shRNAs that target Syk global and long-isoform specific expression or firefly luciferase (shLUC; control) were serum starved, then treated with EGFR ligand, EGF. Cell lysates were subsequently analyzed by immunoblotting using activation state-specific antibodies for Erk and Akt (S1A Fig). Our data showed that HCT-116 cells display constitutively active Erk1/2 and Akt proteins, shown by their phosphorylation state, which remained comparable between untreated and EGF-treated cells. However, EGF treatment induced the phosphorylation of Erk1/2 and Akt in DLD-1 cell line (p-Erk; p-Akt; shLUC: -/+ EGF).

In both cells lines, shRNA-mediated Syk depletion did not affect the phosphorylation levels of Erk1/2 which remained comparable between EGF-treated and untreated cells (p-Erk; shG1, shL2: -/+ EGF). However, targeting Syk global and Syk (L) isoform affected the level of constitutive phosphorylation of Akt in HCT-116 cells, as well as the level of EGF-induced phosphorylation of Akt in both cell lines (p-Akt; shG1, shL2: -/+ EGF) (S1A Fig).

Cetuximab is a chimeric IgG1 monoclonal antibody that targets the extracellular domain of EGFR, blocking ligand binding to the receptor and leading to the inhibition of EGFR signaling. The pretreatment of the cells with Cetuximab attenuated the mitogenic effect of EGF and inhibited the phosphorylation of Akt and Erk1/2 in both cell lines; however, the phosphorylation of Erk1/2 in HCT-116 cell lines was mildly affected by Cetuximab. In all cases, the expression level of EGFR in the different cell lines, analyzed by western blot (S1A Fig) and by FACS (data not shown) was not affected. Together, these results showed that Syk (L) expression is required for the sustained and/or growth factor-induced activation of the PI3K/Akt pathway.

EGF exerts opposite effects on Syk alternative splicing

Growth factors can affect cell survival by regulating the activity of splicing modulators and the splicing of kinases such Syk. We investigated the effect of short-term EGF treatment on the alternative splicing of Syk in HCT-116 and DLD-1 cell lines in vitro. Our data showed that EGF exposure increases exon 9 inclusion and the expression levels of Syk (L) isoform in DLD-1 cell line, while it exerts the opposite effect in HCT-116 cell line by inducing exon skipping and the increase in Syk (S) isoform expression (S1B Fig). Therefore, we concluded that mitogenic signaling regulates both the amount and the ratio of Syk splicing isoforms but in a manner dependent on cell background.

Considering the accumulating evidence supporting the role of Syk (L) isoform in CRC cell survival and cell cycle regulation, we further investigated the role of Syk downstream of EGFR signaling. For this, we performed Syk “pull-down assays” on lysates of serum-starved and EGF-stimulated DLD-1 cells, and we analyzed the components of the captured complexes by high-resolution mass spectrometry to identify Syk interactomes. Our data showed that upon EGF treatment, there is an increased association of Syk with ligands that are implicated in the control of apoptosis, and in the regulation of mitosis and cell cycle G2/M checkpoint, such as Nuclear death domain protein p84N5, which activates a G2/M cell cycle checkpoint prior to the onset of apoptosis [29]; RAD50, which regulates mitotic progression independent of DNA repair functions [30]; and cell division cycle protein 23 homolog (CDC23), a protein component of anaphase-promoting complex (APC) [31] (S1C Fig). These data confirm the implication of Syk in the control of the cell cycle and the apoptosis of CRC cell lines, which is enhanced under the influence of oncogenic growth factors such as EGF.

Syk Long isoform expression is associated with tumorigenesis in human CRC tissues

In a previous study, we reported a gene expression signature associated with treatment response in advanced CRC patients [24]. Thirteen pairs of colon primary tumors (CT) and normal colon mucosas (CN) were collected at the time of surgery, before chemotherapy. We used these samples to determine the profile of Syk expression in primary human colorectal tumors. For this purpose, we performed western blot analysis of protein extracts from tumor tissues of these 13 untreated patients and we compared the expression of Syk protein to that of normal adjacent tissues of the same patient. Our data showed that Syk protein levels were higher in 50% of the analyzed human tumors compared to the normal tissues (Fig 5A). However, we were not able to analyze the ratio of Syk (L) versus Syk (S) isoforms due to the absence of long isoform-specific antibodies.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Syk (L) isoform expression is associated with tumorigenesis in human CRC tissues.

(A) Top panel- Western blot analysis of the protein extracts of primary colorectal tumors of 13 untreated patients. T: Tumor tissue and N: normal adjacent tissue of the same patient. Bottom panel- Densitometric quantification of immunoblot analyses of Syk expression in tumor tissues over the corresponding normal tissues. Red arrows indicate the tumor samples in which the ratio of Syk expression is at least two fold higher compared to the normal tissues. (B) Analysis of Syk mRNA differential gene expression in tumor and normal tissue of N = 50 patients from TCGA database COAD and READ cohorts. (C) Correlation between Syk mRNA and Syk protein expression in tumor tissues of N = 52 patients. (D) Analysis of Syk (L) isoform expression over the total Syk transcripts in tumor tissues of N = 50 patients. The horizontal dashed line indicates the median of PSI value in tumor samples. (E) High proportion of Syk (S) isoform are only present in tumors. The PSI values were compared in 50 patients presenting both tumor and normal tissue value. The above gray curve corresponds to the 1.5 limit of Syk (L) expression in tumor sample over normal tissue ratio, whereas the bottom gray curve is the reverse one. Dashed line figures out the 0.25 PSI limit defining low PSI samples. (F) Differentially expressed genes between low and high PSI tumors. Heatmap represents the z-score transformed gene expression values. Clinical data on TCGA patient were recovered from [26]. CMS groups were recovered from [32]. K-Ras mutations were downloaded from cbioportal. KRAS means all mutations except G13D, G12V and G12D, which are the most prevalent ones in TCGA patients. The relationship between PSI expression and MSI score is according to [27]. The REG value is calculated as the mean of z-score of gene expression in the low-PSI cluster of 40 samples. (G) GO terms enrichment between positive REG (583 genes) and negative REG genes (1070 genes) as defined in panel F. Red colors correspond the upregulated genes and blue colors to down regulated genes in the PSI low associated cluster of 40 patients. The score is the log10 of the corrected p-value of enriched terms in each group. The sign of the score is the sign of the REG value (i.e. mean of z-score in the low-PSI group).

https://doi.org/10.1371/journal.pone.0274390.g005

Next, we conducted an analysis integrating Syk global and isoform-specific expression data, as well as clinical meta-data of 622 CRC patients (COAD and READ cohorts) from The Cancer Genome Atlas (TCGA). This analysis showed a high Syk mRNA expression score in CRC tumors (S2A Fig). Accordingly, the comparison of mRNA expression levels of global Syk in tumor and normal tissue of same patients (N = 50) revealed an upregulation of Syk gene expression in two thirds of the primary tumors (Fig 5B). Furthermore, high mRNA expression was associated with high protein level in TCGA tumors but the opposite was not always true (N = 52) (Fig 5C). This is possibly due to Syk protein turnover differences.

We next analyzed Syk splicing variants expression in CRC tumors using the COAD and READ cohorts of TCGA SpliceSeq database. This database reports the PSI (percent-splice-in) values for the splice events in tumor and normal adjacent tissues [25] (S2B Fig). We extracted PSI values for exon 9 of Syk gene, which correspond to the ratio of Syk (L) isoform expression over the total Syk transcripts (S2C Fig). Since Syk (L) isoform results from the inclusion of exon 9 in Syk gene, the PSI value reflects the proportion of Syk (L) isoform expression in patient’s tissue samples.

The comparison of PSI values of normal and tumor tissues of the same patients (N = 50) revealed that both normal and tumor tissues present high PSI values reflecting the dominant expression of Syk (L) isoform (S2B Fig). Our analysis showed that most tumor samples express higher amounts of Syk (L) than Syk (S) mRNA (PSI values > 0.5, N = 43/50) (Fig 5D). Moreover, the PSI values of tumor tissues in which Syk expression is down-regulated remain high, suggesting a dependence of tumors to Syk (L) isoform expression. Interestingly, 10% of the analyzed tumor tissues (N = 5/50), and none of the normal tissues, have low PSI values (PSI < 0.25), reflecting the expression of a high proportion of Syk (S) isoform (Fig 5E). The level of expression of global Syk in these tumor samples is comparable to that of the adjacent normal tissue (Fig 5D).

To further characterize this specific subpopulation and the consequences of Syk PSI value on tumor biology, we classified 619 CRC tumor samples based on gene expression data. We first classified samples as low PSI (PSI<0.25; 20 samples), medium PSI (452 samples) and high PSI (PSI>0.75; 147 samples) tumors. Next, we selected 1653 genes differentially expressed between the low and high PSI samples and we analyzed tumor subpopulations by hierarchical clustering (Fig 5F). Interestingly, we identified a small cluster of 40 samples designated as low PSI-like tumors. This cluster includes 17 low PSI, 13 medium PSI and 10 high PSI tumors. This group of 40 samples forms a specific cluster associated with a large set of 583 strongly over-expressed genes and 1070 down-regulated genes, compared to other tumor samples (REG value). The REG positive and negative groups overlap the two clusters obtained from unsupervised analysis. Our analysis showed that 495 genes from the first main cluster associated with REG positive values were related to translation and mitochondria, whereas 1158 genes from the second main cluster related to REG negative values were associated with centriole and centriolar functions (Fig 5G). Finally, this group of PSI-low tumors was not related to any classical CRC classification such as MSI status, CMS molecular classification, K-Ras mutational status, or TNM stage, nor to Syk total expression level.

Taken together, our analysis suggests the dependence of CRC tumors to the pro-survival Syk (L) isoform, except in the new identified subgroup of low PSI tumors in which gene reprograming has taken place to overcome the effect of Syk (L) down-regulation.

Treatment of CRC cell lines with C-13, an original non-enzymatic inhibitor of Syk

In our previous works, we developed an original approach for the identification of protein-protein interaction inhibitors of Syk [33]. This led to the discovery of new classes of non-enzymatic inhibitors of Syk with improved selectivity profile and which bind to a cavity located at the interface between the SH2 domains and the linker domain of Syk (Patent No WO2009133294). Among the newly isolated scaffolds, compound C-13 showed a high potential for the inhibition of mast cell degranulation in vitro (Syk IC50 = 2 μM) and for the prevention of anaphylactic shock in mice using oral delivery route in vivo (Syk IC50 = 110 mg/kg) [23].

We studied the impact of compound C-13 on the biology of CRC cell lines. We exposed a wild type (DIFI), and two K-Ras mutant (HCT-116, DLD-1) CRC cell lines to various concentrations of C-13 for 2 hours and then measured cell proliferation, viability and apoptosis 72 hours post-treatment. Our data showed that C-13 inhibits cell proliferation and induces apoptosis of the cells (Fig 6A) in a manner that mimic shRNA Syk (L) isoform-specific targeting but not total Syk targeting by shRNA (Fig 2). Moreover, C-13 inhibits FCS-induced chemomigration (Fig 6B and 6C) and invasion (Fig 6D and 6E) of CRC cell lines in a concentration-dependent manner. Interestingly, the cytotoxic effect of C-13 on DLD-1 cell line was higher than on the two other cell lines, and was confirmed by western blot analysis which showed that EGF-induced phosphorylation of Erk1/2 and Akt were both markedly reduced in DLD-1 cells pre-treated with C-13 (Fig 6F). FACS analysis showed that the level of membrane expression of EGFR in the cell lines was not affected by C-13 (data not shown). To demonstrate that the effect of C-13 on cells viability is indeed due to a Syk-specific inhibition and not to an off-target interaction, we used Syk-positive CRC cell lines DLD-1, SW620 and HCT-116, and Syk-negative breast cancer cells lines MDAMB231 and BT549 as control. Cells were exposed to various concentrations of C-13 for 72 h and cell viability was measured. Our data showed that C-13 specifically exerts cytotoxic effects on Syk-positive cell lines, while Syk-negative cell lines viability was unaffected (Fig 6G).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Treatment of CRC cell lines with C-13, an original non-enzymatic inhibitor of Syk.

(A) Cells were exposed for 2 hours to various concentrations of C-13 and the impact of C-13 on cellular phenotype evaluated 72 hours post-treatment by: cell proliferation, viability and apoptosis, and normalized against DMF (vehicle)-treated HCT-116, DLD-1 and DIFI CRC cells. Data are the mean ±SD (error bars) of at least three independent experiments (*, P<0.05; **, P<0.01; ***, P<0.001; compared with control). (B-C) Cells pretreated for 2 hours with C-13 or vehicle (DMF) were serum starved, labelled with CellTracker dye, and migrated toward FCS in Fluoroblock chambers. (D-E) Cells pretreated with C-13 or vehicle (DMF) were serum starved, labelled with CellTracker dye and seeded on Matrigel-coated Transwell inserts of 8 μm pore size (BD Falcon). FCS was used as chemoattractant in the Transwell lower part. After 24 hours at 37°C, migration was observed and counted microscopically. Images were obtained and total number of cells per field of view was measured using ImageJ software. Representative microscopy images are shown in Fig 6C and 6E. Error bars represent the mean number of migrated cells in triplicate wells, representative of three independent experiments (*, P<0.05; **, P<0.01; ***, P<0.001; compared with control). (F) DLD-1 cells pre-treated with the indicated concentrations of compound C-13 or DMF (vehicle) for 2 hours, were incubated with the recombinant human EGF (50 ng/mL) or a combination of EGF (50 ng/mL) + Cetuximab (50 μg/mL) for 10 minutes. Top panel- Protein extracts from cell lysates were analyzed by western blot using the indicated antibodies. Bottom panel- Densitometric quantification of immunoblot analyses of phospho-Erk (+EGF) versus phospho-Erk (-EGF) over control and phospho-Akt (+EGF) versus phospho-Akt (-EGF) over control. G) Syk-positive CRC cell lines DLD-1, SW620 and HCT-116; and Syk-negative breast cancer cells lines MDAMB231 and BT549 were exposed to various concentrations of C-13 for 72 h; (H) HCT-116, DLD-1 and HT29 cell lines were treated in parallel with C-13 (vehicle DMF) and R406 (vehicle DMSO) for 72 h; and cell viability was measured using CellTiter-Glo Luminescent cell viability assay (Promega). The relative changes were calculated and used to generate the bar graphs. The error bars represent the variation observed in three biological and three technical replicate. (*, P<0.05; **, P<0.01; ***, P<0.001; compared with Syk-negative MDAMB231 cell line for panel G; compared with vehicle for panel H).

https://doi.org/10.1371/journal.pone.0274390.g006

Finally, we compared C-13 with R406 (the active metabolite of Fostamatinib R788), a specific, ATP-competitive inhibitor of Syk (Syk IC50 = 41 nM) that has shown efficacy in the treatment of autoimmune diseases [34] and in preclinical leukaemia studies [35,36]. We compared the effects of increasing concentration of the two drugs on the survival of HCT-116, DLD-1 and HT29 CRC cell lines that express low to high levels of Syk, respectively (Fig 6H). Our data showed that despite a 300 fold lower in vitro affinity for Syk compared to R406 (4 μM for C-13 [23] versus 12 nM for R406 [37]), C-13 affects the viability of Syk-positive CRC cell lines in a more efficient manner than R406. Together with the effect of Piceatannol on cell migration (Fig 4B), our data demonstrate that compounds that target Syk catalytic activity are less efficient than compound C-13 on the inhibition of CRC cell lines proliferation and motility.

Compound C-13 attenuates tumor growth in CRC xenograft mice model

We established two in vivo mice models to examine the anti-tumor effect of C-13 on CRC tumor growth. For these experiments, xenografts in athymic nude mice were established by subcutaneous injection of 1x10⁶ DLD-1 cells and the drug administration was initiated when tumors reached the volume of 40–70 mm3. For the intraperitoneal (IP) and oral administrations, we referred to our previous in vivo studies with C-13 in order to choose nontoxic doses of this compound and to minimize adverse effects [23].

We first investigated the effect of IP administration of C-13 on tumor growth. For this purpose, two groups of 9 mice received an IP administration of compound C-13 at a dosage of 0.35 mg/kg or DMF alone (vehicle). The drug was administered three times a week over 3 weeks. The tumor sizes were measured and recorded every 2 days, and tumor growth curves analyzed for each group. At the given dose, there was a significant decrease in tumor growth in the group treated with C-13 compared to the control group treated with vehicle only (p = 0.04) suggesting C-13’s potently inhibitory effect against tumor growth (Fig 7A). Kaplan-Meier survival analysis confirmed this observation (Fig 7B). We did not evidence any apparent toxicity of the treatment and body weight growth and physical activity of mice were similar in both groups. In a second in vivo experiment, we investigated the effect of the oral administration of C-13 on the tumor growth of DLD-1 cells xenografts in athymic nude mice. For this purpose, 100 mg/kg of C-13 or DMF alone (vehicle) were administered by gavage to two groups of 7 mice xenografted with DLD-1 cells, three times per week for 3 weeks. As shown in Fig 7C and 7D, there is a decrease in tumor growth of the group of mice treated with C-13 compared to the control group, although we did not reach a significant statistical value (p = 0.11).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. C-13 inhibits tumor growth of xenografts in Athymic Nude mice.

(A) Effect of intraperitoneal treatment with C-13 (0.35 mg/kg) or DMF (vehicle) (3 times/week) on tumor growth in Athymic Nude mice xenografted with 1x10⁶ DLD-1 CRC cells (n = 10 mice/group). (B) Kaplan-Meier survival curves of the mice described in B. C) Effect of oral administration of C-13 (100 mg/kg) or DMF (vehicle) (3 times/week) on tumor growth of Athymic Nude mice xenografted with 1x10⁶ DLD-1 CRC cells (n = 7 mice/group). D) Kaplan-Meier survival curves of the mice described in C.

https://doi.org/10.1371/journal.pone.0274390.g007

Together, our in vivo data confirm that C-13 is a promising and original Syk-specific molecule, with potential applications in the treatment of K-Ras-mutated CRC.