Reversine inhibits Colon Carcinoma Cell Migration by Targeting JNK1

Colorectal cancer is one of the most commonly diagnosed cancers and the third most common cause of cancer-related death. Metastasis is the leading reason for the resultant mortality of these patients. Accordingly, development and characterization of novel anti-cancer drugs limiting colorectal tumor cell dissemination and metastasis are needed. In this study, we found that the small molecule Reversine reduces the migration potential of human colon carcinoma cells in vitro. A coupled kinase assay with bio-informatics approach identified the c-Jun N-terminal kinase (JNK) cascade as the main pathway inhibited by Reversine. Knockdown experiments and pharmacological inhibition identified JNK1 but not JNK2, as a downstream effector target in cancer cell migration. Xenograft experiments confirm the effect of JNK inhibition in the metastatic potential of colon cancer cells. These results highlight the impact of individual JNK isoforms in cancer cell metastasis and propose Reversine as a novel anti-cancer molecule for treatment of colon cancer patients.

In this study, we identified Reversine as a potent inhibitor of colon cancer cells migration and metastasis. The substance is effective by interference with JNK1-signaling.

Results
Reversine and SP600125 inhibit colon carcinoma cell migration. In a previous study, we developed a screening assay using the two dimensional Oris TM cell migration to target invasive sarcoma cell lines by treating with several mitosis and cytoskeleton inhibitors at 3 different doses (0.1, 1 and 10 μM) for 24 hours 13 . The aim of this test was to discover potent anti-migratory agent/s (≥50% as a cut off) without any major cell toxicity (≤30% of toxicity as limit) (Fig. S1). By calculating both the ratio of the field area and cells count of the migration zone it was found that Reversine and SP600125 at 10 μM prevent the migration of sarcoma cells 13 . Thus, we decided to investigate the effect of these molecules on the migration of the human colon carcinoma cell line RKO which are considered as one of the most invasive colorectal carcinoma cell lines 35 .
In a first experimental approach, we performed a wound-healing test and found that the decrease of cell free area was significantly delayed in the presence of SP600125 (55 +/− 0.7%) or Reversine (48 +/− 0.1%) (Fig. 1A). The two dimensional Oris TM cell migration assay confirmed the anti-migratory effect of Reversine and SP600125 compared to solvent treated RKO cells (Fig. 1B), Reversine inhibited migration by 40 +/− 0.1% and SP600125 by 37 +/− 0.1%. In addition, SP600125 and Reversine reduced individual cell migration, tracked by time laps microscopy ( Fig. 1C). Moreover, using Boyden chamber assay, cell invasion was completely abolished in cells treated with Reversine and SP600125 for 24 h (Fig. 1D) To exclude toxic effects of Reversine and SP600125, the RKO cells were treated with Reversine and SP600125 for 24 hours and subsequently used for Flow Cytometry (FACS) analysis to assess cell death associated parameters. The cells were co-stained with the propidium iodide to check plasma membrane integrity, and Annexin V-FITC to measure cell membrane scrambling and phosphatidylserine translocation, which are used as a hallmark of apoptosis. Cell membrane scrambling and phosphatidylserine outer membrane exposure was similar in Reversine and SP600125 treated cells and DMSO-treated cells ( Fig. 2A). Further, we used DiOC 6 (3), to measure the mitochondrial transmembrane potential (Δψm). Mitochondrial potential and cell membrane permeability were not significantly different between Reversine or SP600125 treated cells and DMSO treated cells (Fig. 2B). Moreover, cell cycle FACS analysis showed that subG1 fraction relative to DNA degradation was not significantly different between SP600125 or Reversine treated cells and DMSO-treated cells (Fig. 2C). Together, these results suggest that 24 h treatment of colon cancer cells with Reversine and SP600125 does not affect cell death parameters. In parallel, we decided to investigate the cytotoxicity effect of Reversine and SP600125 treatment in non-cancerous cells including non-transformed fibroblast IMR90 and normal human colon mucosal epithelial cell NCM460 using the propidium iodide and DiOC 6 staining. It was found that mitochondrial potential and cell membrane permeability were not different between treated and control cells (Fig. 2D). The evaluation of the subG1 fraction further confirmed this finding (Fig. 2E). Together, these results suggest that 24 h treatment of non-cancerous cells with Reversine and SP600125 does not affect cell death parameters.
To exclude any anti-proliferative effect of Reversine or SP600125 that may overlap with the anti-migration effect, several assays were performed. Crystal violet cell proliferation assay showed that Reversine and SP600125 has no influence in cell proliferation after 24 h of treatment ( Fig. 2F and Fig. S2). A contrario, treatment of cells with 200 ng/ml Nocodazole, which we used here as a potent anti-mitotic drug, showed a complete proliferation arrest (Fig. 2F). Next, we analyzed and compared the status of cell cycle in Reversine and SP600125 treated cells compared to DMSO. Treatment of cells Reversine and SP600125 continued to cycle and undergo polyploidy as well as exhibited histone H3 phosphorylation, which is a marker of ongoing cell mitosis (Fig. 2G). Altogether, this data together with our previous findings demonstrate that SP600125 and Reversine do not influence cancer cells proliferation at 24 h but instead induces cells polyploidization 17,23 . Reversine and SP600125 target the JNK pathway. Reversine and SP600125 are reported as broadspectrum inhibitors of serine/threonine kinases 17,23 . To identify the common target kinases of these small molecules, we decided to perform a kinase profiler screening 36 (Fig. 3). Treatment of RKO cells with 10 µM of Reversine or SP600125 inhibited several kinases with varying efficiency. We selected kinases that were inhibited more than 90%. As a result, in a panel of 222 kinases, SP600125 and Reversine inhibited 36 and 93 kinases, respectively. However, 22 kinases were inhibited by both drugs (Fig. 3B-C). Focusing on these 22 kinases, and to select the pathway involved in the anti-migration effect of SP600125 and Reversine, we choose a bio-informatics approach. Genes relative to the kinases list were used in enrichment analysis to identify over-represented gene ontology (GO) categories (Fig. S3).
Regulation of the JNK cascade exhibited the highest enrichment score (Term p-value after correction with Bonferroni = 1,42 × 10 −6 ) for the common SP600125 and Reversine-inhibited kinase (Fig. 4A). The other top gene-sets were mitotic spindle organization, protein auto-phosphorylation, peptidyl-serine phosphorylation, and vascular endothelial growth factor (VEGF) signaling pathways ( Fig. 4A-C). To confirm the role JNK in the migration process, we decided to investigate the expression and activation of JNK in the RKO cells compared to the normal human colon mucosal epithelial cell NCM460. It was found that JNK is highly expressed and Scientific REPORts | (2018) 8:11821 | DOI:10.1038/s41598-018-30251-w activated in the human colon cancer cells compared to the non-cancerous cells (Fig. 4D). Moreover, by checking the level of total JNK and phospho JNK in control vs. treated cells with Reversine or SP600125, we found that Reversine affect slightly the expression of JNK protein, particularly the levels of JNK1 (The lower band in the western corresponding to 46 KDa, Fig. 4E). Furthermore, Reversine or SP600125 treatment completely abolishes  37 . Interestingly, Reversine also inhibited activation of p38 but not ERK. We also evaluated the level of Akt activation (another player of cell migration) 38 and confirmed the Reversine selectivity, a contrario of SP600125, for Akt 39 (Fig. 4F).  40 . In our study, we excluded JNK3 since it is mainly expressed in neurons, heart and testis. To distinguish between JNK1 and JNK2-mediated cell migration, we treated cells with a specific inhibitor of JNK2 (JNK inhibitor IX, Santa Cruz Biotechnology, Dallas, USA) 41,42 and performed a wound-healing (Fig. 5A,B) as well as a two dimensional Oris TM cell migration assay (Fig. 5C,D). Wound recovering and migration zone invasion were similar after 24h-treatment with the JNK2 inhibitor and control treated cells. To confirm this data, we treated the cells with small interfering RNAs specific for JNK1 or JNK2. As depicted in Fig. 6A, RKO cells were transfected with unrelated siRNA (siUNR), siJNK1/2, or siJNK2 in a 6-well plate for 24 h and condensed further into a 96-well plate. The efficacy of siRNA JNK1/2 and JNK2 were studies using Western blot analysis (Fig. S4). After the cells formed a monolayer, a scratch assay was performed, and cells migration was monitored by microscopy. It was evident that the decline of cell free area was significantly delayed following knockdown of JNK1/2 compared to siRNA UNR (74 +/− 0.4% of inhibition) or siJNK2-treated cells (Fig. 6B,C). To exclude toxic effect of JNK1 and JNK2 knockdown, FACS analyses of the death related parameters were investigated. It was found that siRNA transfection of RKO cells with JNK1 or JNK2 did not induce any significant changes in the overall cell survival (Fig. 6D,E). Moreover, we analyzed the cell cycle of RKO after the knockdown of JNK1/2 and JNK2 and found that the siRNA transfection was not influencing cell cycle or cell proliferation (Fig. 6F).

The administration of SP600125 in vivo affects the metastasis of human colon carcinoma RKO cells.
To evaluate the therapeutic and anti-metastatic potential of JNK inhibition in vivo, animals were injected intravenously with 2 × 10 6 RKO cells/mouse. The mice were treated intraperitoneally every 3 days with a control vehicle or 10 mg/kg SP600125 (Fig. 7A). The SP600125-treated mice showed fewer tumor metastases in the liver (Fig. 7B,C), lung (Fig. 7F,G), spleen, and kidney (Data non shown) compared with control-treated animals. Moreover, the average and ratio of liver weight compared to the body weight were higher in vehicle-compared to SP600125-treated mice (Fig. 7D,E). No differences could be observed in the average body weight comparing vehicle and SP600125-treated mice (Fig. 7H). To further confirm the implication of JNK inhibition in vivo, we extracted the proteins from tumors isolated from lung and liver for western blotting. Densitometic analysis demonstrated that SP600125 reduces phospho JNK level in treated lung and liver comparatively to control. (Fig. S5).

Discussion
Metastasis is the leading cause of cancer mortality including colon cancer, which is one of the most invasive and metastable tumors 43 . Thus, it is very important and urgent to develop anti-cancer drugs to reduce tumor cells dissemination. Here, we report the identification of the small molecules Reversine and SP600125, which can interfere with JNK signaling, as inhibitors of migration and metastasis of colon carcinoma cells. In our study, we decided to use the primary colon cancer RKO cell line which harbors the microsatellite instability (MSI) phenotype and represents 15% of all CRCs. RKO cells harbor mutations in KRAS, BRAF, PIK3CA and PTEN and are considered as one of the most invasive colon cancer cell lines 44 . We took advantage of a novel screen assay, using the two dimensional Oris TM cell migration developed in a previous study 13 . The previous test revealed Reversine and SP600125 as potent anti-migratory drugs of invasive sarcoma cells. In the present study, we explored the effect of Reversine and SP600125 on the human colon carcinoma RKO cells. We exposed cells for 24 h to both small molecules and found that this treatment significantly delayed the migration and invasion of RKO cells using different assays, namely wound-healing, two dimensional Oris TM assay, individual cell tracking assay, and Boyden chamber assay. The anti-migratory effect of these drugs was not due to toxicity or cell apoptosis. Moreover, the treatment did not affect cell proliferation at 24 h, which is in line with our previous findings 17,23 . In this work we could also show that 24 h treatment with Reversine or SP600125 is non-toxic to the non-cancerous cells including non-transformed fibroblast and human colon mucosal epithelial cells. The kinase profiler screening identified several kinases that were inhibited by both Reversine and SP600125. This result in combination with bio-informatics (GO: gene ontology) approaches identified the JNK family activation as the most significant pathway inhibited by Reversine and SP600125. To distinguish between JNK1 and JNK2 activation, RKO cells were treated with specific siRNAs targeting JNK1/2 or JNK2, as well as treatment with a specific JNK2 inhibitor. The results demonstrated that only JNK1 inhibition delays the motility of these cells. Western blot analysis showed that JNK is overexpressed and activated in RKO cancer cells compared to normal colon mucosal epithelial cells. We could also show that Reversine, similar to SP600125, reduces the activation and expression of JNK in treated cells. Additionally, we demonstrate that activation of JNK related signaling pathway proteins including c-Fos, c-Jun, and p38 but not ERK were significantly decreased with Reversine treatment. The specificity of Reversine but not SP600125 could be observed against the activation of AKT, which is in line with previous finding 39 .
Reversine is a specific inhibitor of the human mitotic kinase Mps1 30 , whereas SP600125 is a well-known inhibitor for JNK1/2 14   and JNK have a high degree of sequence similarity in their ATP-binding pocket 16,46 . Mutation of methionine M108 to glutamine in JNK1 renders it insensitive to SP600125 inhibition 47 . Moreover, the corresponding mutation in Mps1 (M602Q) also proved significantly less sensitive to SP600125 16,17 . These results in combination with our findings suggest that Reversine and SP600125 share the common target kinase JNK.
It is becoming increasingly clear that JNK1 is essential for cell migration. In the present study, we found that treatment with SP600125 of intravenously injected tumor cells reduces liver and lung tumor metastases. This effect is in line with previous findings that showing the same effect upon treatment of animals with Reversine 48 . Several studies investigated the effect of JNK1 inhibition by chemicals, small interfering RNA, or dominant-negative mutant, and could confirm the role of JNK-signaling in migration of several different cell types. In these studies, the overexpression of a dominant-negative mutant of JNK1 reduced the migration of rat urinary bladder NBT-II cells 49 , cortical neurons 50 , human umbilical vein endothelial cells 51 , and bovine aortic endothelial cells 52 . Moreover, JNK1 siRNA transfection delays the migration of gastric cancer cells 53 , mouse hepatocellular carcinoma cells 54 , and human melanoma cells 55 . Furthermore, JNK inhibitors were vastly employed to inhibit migration in several cell lines including human dermal fibroblasts 56 , smooth muscle cells 57 , murine cortical neuronal cells 50 , schwann cells 58 , murine epidermal keratinocytes, murine dermal fibroblasts 59 , thyroid cancer cells 60 uterine luminal epithelial cells 61 and human lung adenocarcinoma cells 62 .
The JNK signaling is implicated in cell scattering during epithelial-to-mesenchymal transition (EMT) in pulmonary epithelial cells 63 . Moreover, JNK1 is an important regulator of cell-cell adhesion. Indeed, the epithelial tight junctions (TJ) and adherent junctions (AJ) reassembly is accelerated by JNK1 but not by JNK2 inhibition, leading to changes in epithelial morphology and migration delay 64 . In addition, cytoskeletal proteins are important JNK substrates, including proteins associated with the microtubules (DCX, MAP1B, MAP2, Tau and WDR62), proteins associated with the actin cytoskeleton (MARCKSL1 and SMTL2), and focal adhesion proteins (paxillin and β-catenin) 65 . Moreover, recent studies have connected additional JNK target proteins to the cytoskeleton control, including actin filament assembly proteins (Formin1, FHOD3), microtubule-associated proteins (CCSER1), and molecular motors (Myosin-9B). However, these important processes remain largely limited and further studies are needed to explore the mechanism 65 .
In cancer, JNK is widely documented as implicated in several tumors including melanoma, head and neck, breast, ovarian and gastric cancers, suggesting that JNK may be an attractive target for cancer therapy 66 . However, it has also been proposed that JNK1 and JNK2 have different specific cellular targets in cancer and thus the mechanisms involving individual JNK proteins is generally unknown. In the present study we could provide evidence

Materials and Methods
Cell lines, culture conditions and reagents. Human colon carcinoma RKO cells were grown in McCoy's 5 A medium supplemented 10% fetal calf serum (FCS), 10 mM HEPES buffer, 100 units/mL penicillin G sodium and 100 μg/mL streptomycin sulfate. IMR90 human fibroblast cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS) and 100 units/mL penicillin G sodium. NCM460 cells were maintained in INCELL's enriched M3:10 medium (M310A; which is M3 Base medium plus supplements with 10% FBS and contains antibiotics). Cell lines were routinely maintained at 37 °C under 5% CO 2 . Figure 6. Inhibition of JNK1 and not JNK2 reduces the migratory potential of the human colon carcinoma RKO cells. (A-C) Cells were seeded into 6-well plate and were transfected with an unrelated small interfering (si) RNA (siUNR) or specific siRNAs directed against JNK1/2 or JNK2 (siJNK1/2 and siJNK2) (two well per condition). Upon 24 h, cells were harvested and seeded in a 96-well plate to ensure a monolayer confluence. After overnight culture a wound-healing test was performed and migration was evaluated after 24 h. Representative photomicrographs are shown in panel (B), the yellow broken lines delimit the cell-free area. Quantitative data of anti-migration % comparatively to control are presented in panel (C). (D-E) The toxicity of siRNAs JNK1/2 or JNK2 was evaluated by flow cytometry upon staining with the cell death-associated parameter dyes. RKO cells were transfected with siUNR, siJNK1/2 or siJNK2 for 48 h and then co-stained with the vital dye propidium iodide (PI) and the FITC-conjugated Annexin V for the detection of phosphatidylserine exposure (D) or co-stained with the vital dye propidium iodide (PI) and the mitochondrial membrane potential (Δψm)-sensing dye DiOC 6 (3) (F). Cells were transfected with siUNR, siJNK1/2 or siJNK2 for 48 h and then fixed with ethanol and labeled with the DNA dye PI, for the quantification of cell cycle. Representative histograms are shown. Data are reported in SEM; n = 3. ***(p < 0.001) indicates significant difference from the siUNR transfection (ANOVA).

Migration and Invasion assays.
i For wound-healing assay, scratches were performed on confluent cell monolayers with sterile 200 μl tips and monitored every 12 h by microscopy. Migration distances were expressed as percentages over control values. Indeed, as the scratches are not equal at time 0, we normalized the values by dividing every value by the time 0 value in every condition (Remaining free area/free area time 0). ii The two dimensional Oris TM cell migration assays were performed according to the manufacturer's instructions (Platypus Technologies, Madison, USA). Briefly, cells were seeded (4 × 10 4 cells per well) into 96-well plates with "silicone stopper" and grown overnight. Then stoppers were removed, and cells were incubated for an GraphPad Prism using ANOVA with Tukey's test as post-hoc and T-test (for the animal experiments).