Antitumorigenic Effects of Inhibiting Ephrin Receptor Kinase Signaling by GLPG1790 against Colorectal Cancer Cell Lines In Vitro and In Vivo

Erythropoietin-producing hepatocellular receptors (Eph) promote the onset and sustain the progression of cancers such as colorectal cancer (CRC), in which the A2 subtype of Eph receptor expression has been shown to correlate with a poor prognosis and has been identified as a promising therapeutic target. Herein, we investigated, in vitro and in vivo, the effects of treatment with GLPG1790, a potent pan-Eph inhibitor. The small molecule has selective activity against the EphA2 isoform in human HCT116 and HCT15 CRC cell lines expressing a constitutively active form of RAS concurrently with a wild-type or mutant form of p53, respectively. GLPG1790 reduced EPHA2 phosphorylation/activation and induced G1/S cell-cycle growth arrest by downregulating the expression of cyclin E and PCNA, while upregulating p21Waf1/Cip1 and p27Cip/Kip. The inhibition of ephrin signaling induced quiescence in HCT15 and senescence in HCT116 cells. While investigating the role of CRC-related, pro-oncogenic p53 and RAS pathways, we found that GLPG1790 upregulated p53 expression and that silencing p53 or inhibiting RAS (human rat sarcoma)/ERKs (extracellular signal-regulated kinase) signaling restrained the ability of GLPG1790 to induce senescence in HCT116 cells. On the other hand, HCT15 silencing of p53 predisposed cells to GLPG1790-induced senescence, whilst no effects of ERK inhibition were observed. Finally, GLPG1790 hindered the epithelial-mesenchymal transition, reduced the migratory capacities of CRC, and affected tumor formation in xenograft models in vivo more efficiently using HCT116 than HCT15 for xenografts. Taken together, our data suggest the therapeutic potential of GLPG1790 as a signal transduction-based therapeutic strategy in to treat CRC.


Introduction
Colorectal cancer (CRC) is the third and second most commonly occurring cancer in men and in women, respectively, with 1 million of new cases diagnosed every year [1]. CRC develops through different stages and is promoted by the accumulation of acquired or inherited genetic mutations that eventually promote malignant evolution [2]. In this regard, mutation of adenomatous polyposis coli (APC), Kirsten rat sarcoma viral oncogene homolog (KRAS), and tumor protein 53 (TP53 or p53) have been extensively investigated and have been shown to play a key role in the progression of CRC [3]. However, CRC cells comprised complex genetic traits, and the role of additional mutations or aberrant expression of additional genes remains unknown.
e erythropoietin-producing hepatocellular carcinoma receptors (Eph) constitute the largest family of receptor tyrosine kinases (RTKs). e family comprised sixteen total receptors divided into either A-or B-subclasses. e EphA group consists of EphA1 to EphA10 receptors, and the EphB group comprised six receptors named EphB1 to EphB6. Eph proteins bind membrane-bound ligands named ephrins generating a very complex signaling network. Eph receptors and ephrins expressed in different cells interact in trans to activate bidirectional signaling cascades from Eph and ephrin, which have been called "forward" and "reverse" signals, respectively. Eph and ephrins coexpressed in the same cell interact in cis, a phenomenon that inhibits trans signaling [4,5].
Eph regulates several biological processes through different kinase-mediated forward and reverse pathways that include small GTPases of Rho members, RAS, focal adhesion kinase (FAK), the PI3 kinase pathway (PI3K), the Jak/Stat pathway, and Src [4,5]. Because of their critical role in regulating several physiological processes, the aberrant expression of Ephs and ephrins has been reported in many human tumors including CRC [6,7]. Aberrant expression of Ephs and ephrins has also been shown to promote cancer cell growth, migration, and invasion through the disruption of multiple molecular mechanisms [5][6][7]. Several studies shown that significant upregulation of EphA1, EphA2, EphA3, EphA4, EphB2, and EphB4 occurs throughout CRC progression [7,8], and EphA2 has been shown to be a marker for advanced disease and a poor prognosis and is a critical therapeutic target [9].
We have recently developed a preclinical, orally bioavailable small molecule named GLPG1790 that is able to inhibit, at nanomolar concentrations, the activity of various Eph receptors, with a particularly strong efficiency versus EphA2 [10]. e molecule to efficiently diminishes the phenotypic transformation of several breast, rhabdomyosarcoma, and glioblastoma cancer cell lines, both in vitro and in vivo [10][11][12], and our recent preliminary data suggest a potential therapeutic role for GLPG1790 also in treating CRC [13]. e present study has investigated the therapeutic efficiency of GLPG1790 against HCT116 and HCT15 CRC cell lines, both expressing a mutated form of KRAS in addition to either a wild-type p53 protein (p53 WT ) or mutated p53 (p53 MT ), respectively [14]. Herein, GLPG1790 significantly inhibited tumor cell growth both in vitro and in vivo, where treatment-induced quiescence or senescence in a cell type-dependent manner and affected the epithelial-mesenchymal transition program. Interestingly, we found that the mutational status of p53, along with RAS (human rat sarcoma)/ERKs (extracellular signal-regulated kinase) signaling, affected GLPG1790-mediated anticancer activity that was higher in p53 WT expressing cells. us, data collected here support the idea that targeting Eph may offer a new means of developing novel anticancer strategies against CRC.

Western Blotting.
Western blot analyses were performed as previously described [17,18]  EphA2 α-Tubulin * * * * * *     showing the percentage of cell in each cell-cycle phase for HCT116 and HCT15 cells ± GLPG1790. Results represent the mean value of four independent experiments ( * P < 0.05 or * * P < 0.01 vs. control). (b) Cell lysates from HCT116 and HCT15 cells ± GLPG1790 (IC 50 ) at the indicated times were analyzed by immunoblotting with specific antibodies for indicated proteins. α-Tubulin expression acted as a loading control for samples. Densitometric analysis of three independent experiments has been reported below the blots ( * P < 0.05, * * P < 0.01 or * * * P < 0.001 vs. control).

Statistical Methods.
Continuous variables were summarized as the standard error of the mean (SEM). For continuous variables, statistical comparisons between control and treated groups were established by carrying out oneway ANOVA tests or the Student's t-test (in cases where there were two independent groups). Dichotomous variables were summarized by absolute and/or relative frequencies.
For dichotomous variables, statistical comparisons between control and treated groups were established by carrying out Fisher's exact test. For multiple comparisons, the level of significance was corrected by multiplying the P value by the number of comparisons performed (n) according to the Bonferroni correction. All tests were two-sided and were determined by Monte Carlo significance. P values less than 0.05 were considered statistically significant. All statistical analyses were performed using the SPSS® statistical analysis software package, version 10.0 [21].

GLPG1790 Induces G1/S Cell-Cycle Growth Arrest by
Promoting Quiescence of HCT15 and Senescence Status HCT116 Cells. e dose of GLPG1790 able to affect 50% of CRC cell viability (IC 50 ) was investigated. After two days of  Journal of Oncology treatment of HCT116 and HCT15 using 0.1-10 μM GLPG1790 produced IC 50 values of 1.87 ± 0.32 μM and 6.98 ± 0.67 μM, respectively (Figure 1(a)). 24 hours after treating cells with a dosage equal to the IC 50 of each cell type, GLPG1790 significantly reduced EphA2 phosphorylation/ activation by 71 ± 9% in HCT116 and 90.4 ± 6.2% in HCT15 cells (Figure 1(b)). e effect of GLPG1790 on the proliferation rate of cells (Figure 1(c)) and the ability of cells to form colonies (Figure 1(d)) were also investigated. Four days of treatment with GLPG1790 reduced the proliferation rate of CRC cells by 96.5 ± 13.3% in HCT116 and 98.2 ± 18.4% in HCT15 cells. After 14 days of treatment, colonies-forming capacity of HCT116 and HCT15 cells was reduced 66 ± 3.1% and 66 ± 5.2%, respectively. Flow cytometry showed that treating cells with GLPG1790 (IC 50 ) for 24 hours significantly reduced DNA replication (S phase) by primarily arresting cells in the G 1 phase of the cell cycle (Figure 2(a)). Consistent with the G 1 /S arrest, GLPG1790 decreased expression levels of cyclin E (CycE) and PCNA whilst it increased p21 Waf1/Cip1 and p27 Cip/Kip protein expression in both cell lines (Figure 2(b)). Unexpectedly, treatment upregulated the cyclin D1 (CycD1) and CDK4 protein   Journal of Oncology expression both in HCT116 and HCT15 cell lines (Figure 2(b)). erefore, investigating whether the antiproliferative effects of GLPG1790 were reversible or not, CRC cell lines were transiently treated for 4 days with GLPG1790 (IC 50 ). As shown in Figure 3(a), removing GLPG1790 restored the ability of HCT15 to proliferate. is was contrary to effects observed after removing GLPG1790 from HCT116 cells, which remained persistently arrested (Figure 3(a)). Notably, after four days of treatment, β-galactosidase activity (Figure 3(b)) and large morphology (Figure 3(c)), widely used as markers of senescent cells, were significantly increased in HCT116 but not in HCT15 cells. GLPG1790 did not induce apoptotic cell death as suggested by the inability of treatment to increase the active (cleaved) form of caspase 3 and PARP (Figure 3(d)). Taken together, these results suggest that the inhibition of EPH signaling is able to affect CRC-transformed phenotypes by inducing either quiescence or senescence, but not cell death, in a cell type-dependent manner.
e roles of p53 and RAS, both key players in the regulation of senescence of CRC cells [22,23], were investigated. HCT116 and HCT15 cell lines were transfected with specific small interfering RNAs (siRNAs) directed against p53 mRNA (p53-siRNA), while a sequence against a C. Elegans gene (CTR-siRNA) was used as a negative control (Figure 4(a)). Western blotting analysis, performed 36 hours after transfection, revealed that p53 protein levels were specifically reduced in p53-siRNA-transfected cells (Figure 4(a)). 36 hours after transfection, cells were treated with GLPG1790 (IC 50 ), and β-galactosidase activity (Figure 4(b)) was assessed. As shown in Figures 4(b) and 4(c), silencing p53 diminished the ability of GLPG1790 to promote β-galactosidase activity 54.2 ± 7.1% in HCT116 (Figure 4(b)). However, β-galactosidase activity was enhanced in HCT15 cells by 70.23 ± 12.6% (Figure 4(b)). 36 hours of GLPG1790 treatment increased p53 expression levels in HCT116 but not in HCT15 cells (Figure 4(c)). e role of KRAS was investigated by assessing the effects of GLPG1790 on the activation status of ERK, which is known to be its principal pro-oncogenic downstream target. As shown in Figure 5(a), GLPG1790 (IC 50 ) treatment persistently increased the ERK activation in HCT116 cells whilst no effects were observed in HCT15 cells (Figure 5(a)). Cotreating HCT116 cells with the MEK/ERK inhibitor U0126 (10 μM) diminished the activity of β-galactosidase induced by GLPG1790 ( Figure 5(b)) and similarly diminished the expression of the senescence marker proteins p16 INK4 , p21 Waf1/Cip1 , and p27 Cip/Kip (Figure 5(c)). Taken together, this evidence suggests that both p53 and KRAS-ERKs-mediated signaling mediate the antiproliferative, senescence-promoting effects of GLPG1790 differently depending on CRC cell type.  Figure 5: GLPG1790 induces senescence in HCT116 cells via the activation of ERK signaling. (a) Cell lysates from HCT116 and HCT15 cells treated with 2 μM and 7 μM GLPG1790, respectively, for 24 or 48 h were analyzed using immunoblotting with specific antibodies for indicated proteins. α-Tubulin expression functions as a loading control. (b) β-Galactosidase activity was assessed by ELISA using HCT116 and HCT15 cells that had been cotreated with U0126 (10 μM) and GLPG1790, 2 μM or 7 μM, respectively, for 36 h ( * * * P < 0.001 vs. untreated, III P < 0.001 vs. GLPG1790). (c) Cell lysates from HCT116 and HCT15 cells cotreated with U0126 (10 μM) and GLPG1790, 2 μM or 7 μM, respectively, for 36 h were analyzed by immunoblotting with specific antibodies for indicated proteins. α-Tubulin expression functions as a loading control.

GLPG1790 In Vivo Reduces the Growth Capacity of CRC Cells.
erapeutic efficiency of GLPG1790 was then investigated using a CRC xenograft tumor model in nude mice. For any cell line, twenty mice were subcutaneously injected with 1 × 10 6 cells/per mouse; after tumors reached a volume of ∼0.2-0.3 cm 3 , animals were divided in two groups of 10 mice each. Animals were treated with GLPG1790 (30 mg/kg) 5 days/week for 4 weeks. As shown in Figure 6(a), at the end of the experiment, GLPG1790 reduced tumor volume by 74.6 ± 11.6% and 52.3 ± 11.2% in mice receiving HCT116 and HCT15 xenografts, respectively. Tumor weights were significantly decreased in mice treated with GLPG1790 in comparison to controls, with 70-90% inhibition of HCT116-derived tumors and 50-90% inhibition for HCT15-derived tumors (Figure 6(b)). e number of mice experiencing TP significantly differed across the groups, and this was confirmed by the median values of TP (Figure 6(c)). In the vehicle group, tumor progression was completed within three weeks of receiving HCT116 xenografts and within one week of receiving xenografts of HCT15 ( Figure 6(c)). With respect to GLPG1790-treated mice, in HCT116-grafted mice, TP never occurred, while for HCT15grafted mice, TP began in the first week and was completed two weeks later (Figure 6(c)). Immunoblotting on excised tumors showed that GLPG1790 treatment downregulated EPHA2 phosphorylation compared with tumors from vehicle-treated mice (Figure 6(d)).

GLPG1790 Affects the Expression of Epithelial-to-Mesenchymal Transition Markers and the Capacity of CRC Cells to
Migrate. e ability of GLPG1790 to delay the epithelial-tomesenchymal transition (EMT) process and the related migration of CRC cells were investigated. GLPG1790 treatment (IC 50 ) rapidly and persistently increased the expression of E-cadherin and ZO-1 and decreased levels of Survivin and Snail (Figure 7(a)). Further, the molecule efficiently reduced the capacity of both CRC cell lines to migrate (Figure 7(b)). Taken together, these data indicated that GLPG1790 restrains the pro-oncogenic EMT process and diminished the capacity of cancer cells to migrate.

Discussion
CRC remains an uncontrollable disease and is the second leading cause of cancer-related deaths [1]. As such, there is an urgent need for new therapeutic strategies to treat CRC. Eph, from the largest family of RTKs that by binding membrane bound ligands named ephrins, regulates many cellular functions through the modulation of several signal transduction pathways including the RAS/ERKs/MAPKs pathway [4,5] known to sustain CRC [24]. CRC cells express high levels of different Eph proteins including EphA2, a known marker of poor prognosis in advanced CRC and a potentially critical therapeutic target for the treatment of CRC [7][8][9]. We have recently shown the anticancer proprieties of GLPG1790, a new pan inhibitor of the Eph receptors with a strong efficiency versus EphA2, versus several breast, rhabdomyosarcoma, and glioblastoma cell lines, both in vitro and in vivo [10][11][12]. Since our preliminary in vitro investigation showed the therapeutic potential of GLPG1790 also versus CRC [13], we decided to better characterize the pharmacological action of this drug on CRC by using in vitro and in vivo approaches on p53 wild-type or mutated CRC cell lines.
GLPG1790 decreased the phosphorylation/activation levels of EphA2 and affected CRC cell viability and proliferation, inducing the promotion of cell-cycle arrest in the G 1 /S phase of the cell cycle. According to the cell-cycle distribution, GLPG1790 downregulated cyclin E and PCNA, thus affecting strategic molecular targets for the onset and progression of CRC [25,26]. Moreover, GLPG1790 upregulated the expression of cell-cycle inhibitors and tumor suppressors p21 Waf1/Cip1 and p27 Cip/Kip , which have previously been shown to revert the chemo-resistant phenotype of several cancer cell types [27]. Unexpectedly, GLPG1790 enhanced expression of cyclin D1, known to be a positive regulator of the G 1 /S cell-cycle transition [28]. We have recently shown that GLPG1790 upregulates the expression of cyclin D1 in RMS, inducing the cytoplasmatic/perinuclear accumulation of the protein [11]. Reports have shown that this accumulation correlates with a lower proliferative index in several cancer types [29,30]. We suppose that a similar mechanism occurs in CRC cells, and future investigations will be to fully elucidate how the protein regulates proliferation. Notably, GLPG1790, at the concentration tested, did not induce apoptosis on CRC cell lines suggesting that it worked as a cytostatic pure treatment. However, GLPG1790 could sensitize cells to the cytotoxic action of other drugs, as some of our preliminary data [13] and ongoing experiments suggest.
Cytostatic drugs induce either reversible or irreversible cell-cycle arrest termed quiescence or senescence, respectively. During quiescence-related growth arrest, cells diminish metabolic activity, protein synthesis, and other cellular functions.
is results in a lack cellular growth; however, removing cytostatic stimulus reactivates cell growth. Contrary to quiescence, cellular senescence is characterized by the nonreversible loss of proliferative potential, the increases in β-galactosidase activity, and cellular expansion [31][32][33]. GLPG1790 induced senescence in HCT116 cells as indicated by enhanced β-galactosidase activity and the acquisition of the characteristic morphology, which includes large and flattened cells that are markers of senescence [33]. Quiescence, however, occurred in the HCT15 cell line. e data collected here support the suggestion that GLPG1790 is a potentially powerful therapeutic with respect to the treatment of CRC. In fact, a treatment that induces senescence is expected to be more efficient than a treatment that induces quiescence [31]. is is because senescence is irreversible and also because senescent cells can be more easily eliminated via phagocytosis, a process that can result in tumor regression [32]. However, these data suggest that the ability of GLPG1790 to induce the maximum therapeutic effects against CRC progression likely depends on the molecular features of targeted cells. e ability to induce quiescence or senescence depends on several signaling pathways, including p53 and RAS/    T u m o r  1  2  3  1  2  3   ---+  +  +  GLPG1790   EphA2 PO4   EphA2   T u m o r  1  2  3  1  2  3 -  Journal of Oncology ERKs/MAPKs [22,23], known to drive the onset and progression of CRC [2,3]. ough no relationship has previously been shown between the mutational status of either p53 or RAS/ERKs/MAPKs and Eph signaling dysregulation, we decided to investigate whether the proteins had roles in mediating GLPG1790-induced effects. Silencing p53 with siRNA specific for the protein enhanced the ability of GLPG1790 to induce senescence in HCT15 cells only, while having the opposite effect on HC116, suggesting that the role of p53 varies depending on cellular context. p53 WT and p53 MT , expressed, respectively, by HCT116 and HCT15 [12], are known to promote or diminish, respectively, the activation of senescence in CRC [22,34,35]. is evidence seems to suggest that the mutational status of p53 could affect the ability of GLPG1790 to induce senescence and that the presence of the mutated p53, known to have acquired oncogenic features due to a gain-of-function mutation [36], has a negative effect on the therapeutic efficacy of GLPG1790. Furthermore, the fact that GLPG1790 increased the expression levels of p53 in HCT116 and did not induce any modulation in HCT15 cells reinforces the idea that p53 WT mediates whilst p53 MT counteracts senescence induced by the GLPG170. However, further tests will be needed to confirm whether there is indeed a relationship between the mutational status of p53 and the therapeutic efficiency of Eph inhibitors.
Approximately 30% to 50% of CRCs express a mutated RAS gene that promotes the constitutive activation of the ERK/MAPK downstream signaling pathway [37]. is pathway is a key regulator of cell proliferation and transformation, promotes CRC onset, and sustains CRC progression [38]. Interestingly, the Ephs are among the few receptor tyrosine kinases known to negatively regulate RAS signaling and its downstream pathways in a variety of cell types [39] though, under some circumstances, Eph receptors activate, rather than inhibit MAPK signaling [40]. We also have shown in previous reports that GLPG1790 induces irreversible growth arrest and differentiation by downregulating ERK in RMS cells [11]. Herein, GLPG1790 strongly and persistently increased the phosphorylation/ activation of ERK signaling in HCT116 cells indicating that relationship between RAS and Eph signaling is similar to that of RMS cells. However, the fact that both RMS [40] and CRC [12] cell lines express constitutively activated RAS, which cannot be regulated, suggests that Eph could modulate the ERK activation downstream of Ras. is could also explain why GLPG1790 failed to modulate ERKs in HCT15 cells expressing constitutively active Ras. Interestingly, p53 WT has been shown to regulate ERK activation [41], and the presence of a mutated p53 in HCT15 could be the reason for the failure to activate ERKs after treatment with GLPG1790. Given the role of ERKs in promoting  tumorigenesis, the fact that they are GLPG1790 activated in HCT116 might seem disadvantageous. However, depending on the duration, magnitude and its subcellular localization of ERKs activation various advantageous cell responses could be affected, such as proliferation, migration, differentiation, and death [42]. ough literature most often emphasizes the oncogenic potential of enhanced ERK activation, a growing number of conflicting studies have suggested that, under certain conditions, aberrant ERK activation can promote anticancer effects including senescence and cell death [43]. e biological significance of GLPG1790-mediated activation of ERKs requires further study; however, the fact that the inhibition of ERK signaling diminished senescence suggests an antioncogenic role for the protein.
CRC cells progressively acquire a metastatic phenotype. While several molecular mechanisms have been shown to orchestrate this process, the EMT has been shown to play a key role [44,45]. EMT is a biologic process that allows a polarized epithelial cell to acquire a mesenchymal cell phenotype characterized by an enhanced migratory capacity and invasiveness [44,45]. Considering that the Eph signaling primarily regulates the cell-to-cell interactions, we supposed that GLPG1790 could disrupt the EMT process by reducing the ability of CRC cells to migrate. Our hypothesis was confirmed by data collected here. Further, GLPG1790 more strongly disrupted EMT in HCT116 than in HCT15, in line with the role of p53 in contributing to metastasis [46].
Because of the natural limitations of the in vitro study model, we also tested the efficacy of the GLPG1790 in xenograft models. Herein, we have shown that although GLPG1790 efficiently blocked EphA2 phosphorylation in tumors derived from both HCT116-or HCT15-xenografted mice, HCT116 tumors were more sensitive to GLPG1790. ese results were supported in vitro data and, taken together, suggest that the presence of p53 WT represents the best condition for use of Eph inhibition-based therapy for CRC. e rate of both the incidence and mortality of CRC has declined over the last two decades. is success is attributable to the early detection and treatment of CRC [47]; however, CRC is still an uncontrollable disease and is the second leading cause of cancer-related deaths. e results of this study demonstrate the preclinical in vitro and in vivo antitumor activity of GLPG1790, a new potent pan inhibitor of EPH receptors, against human CRC cells. Data collected here further suggest that the mutational status of p53 could affect the responsiveness of CRC cells to therapy targeting Eph. Considering the fact that the CRC cell lines used here were derived from different tumors with completely different genetic backgrounds, other molecules may certainly participate in regulating responsiveness to GLPG1790 in CRC cells. ese results indicating that targeting Eph signaling may be used in a signal transduction-based therapy for CRC warrant further testing in preclinical and clinical CRC trials.

Data Availability
All data generated or analyzed during this study are included in this published article.

Ethical Approval
Studies on animal models were performed according to the guidelines established by the University of L'Aquila, Medical School and Science and Technology School Board Regulations, in compliance with the Italian government regulation no. 116 January 27, 1992, and national/international guidelines for the care and use of animals. e present study was approved by the Ethical Committee of the Medical School of the University of L'Aquila (555/2017-PR). is article does not contain any studies with human participants.

Conflicts of Interest
Filip Beirinckx, Philippe Pujuguet, Laurent Saniere, and Ellen Van der Aar are employees of Galapagos NV. e other authors declare that they have no conflicts of interest.

Authors' Contributions
Colapietro Alessandro, Francesco Marampon, and Claudio Festuccia contributed equally. CA, GLGIP, FP, BS, ADF, and IF planned and performed the experiments; FB, PP, LS, and EVdA developed and synthesised GLPG1790; RM, AF, DM, EF, AP, and VT analyzed data; FM and CF wrote and critically revised the manuscript.