Targeting Src-Hic-5 Signal Cascade for Preventing Migration of Cholangiocarcinoma Cell HuCCT1

Cholangiocarcinoma (CCA) is the second most common primary liver cancer with poor prognosis. The deregulation of a lot of oncogenic signaling molecules, such as receptor tyrosine kinases (RTKs), has been found to be associated with CCA progression. However, RTKs-based target therapy showed limited improvement suggesting a need to search for alternative targets for preventing CCA progression. To address this issue, we screened the oncogenic signal molecules upregulated in surgical tissues of CCAs. Interestingly, over-expression of hydrogen peroxide inducible clone-5 (Hic-5) coupled with over-activation of Src, AKT, JNK were observed in 50% of the cholangiocarcinoma with metastatic potential. To investigate whether these molecules may work together to trigger metastatic signaling, their up-and-down relationship was examined in a well-established cholangiocarcinoma cell line, HuCCT1. Src inhibitors PP1 (IC50, 13.4 μM) and dasatinib (IC50, 0.1 μM) significantly decreased both phosphorylated AKT (phosphor-AKT Thr450) and Hic-5 in HuCCT1. In addition, a knockdown of Hic-5 effectively suppressed activation of Src, JNK, and AKT. These implicated a positive cross-talk occurred between Hic-5 and Src for triggering AKT activation. Further, depletion of Hic-5 and inhibition of Src suppressed HuccT1 cell migration in a dose-dependent manner. Remarkably, prior transfection of Hic-5 siRNA for 24 h followed by treatment with PP1 or dasatinib for 24 h resulted in additive suppression of HuCCT1 migration. This suggested that a promising combinatory efficacy can be achieved by depletion of Hic-5 coupled with inhibition of Src. In the future, target therapy against CCA progression by co-targeting Hic-5 and Src may be successfully developed in vivo.


Introduction
Cholangiocarcinoma (CCA) is the second most common primary liver cancer [1] and the overall incidence of CCA has increased progressively worldwide [2]. CCA originates from the epithelium lining the biliary tree and can be divided into two main classes in accordance with the anatomical origin: intrahepatic CCA (iCCA) and extrahepatic CCA

RNA Interference
Hic-5 expression was transiently knocked down by 12-25 nM Hic-5 siRNA (Dharmacon, Lafayette, CO, USA) for 48 h, according to the manufacturer's protocol. The depletion of Hic-5 was validated by Western blot and RT-PCR. The target sequence of Hic-5 siRNA is "GGAGCUGGAUAGACUGAUG".

Proliferation Assay
Cell proliferation assay was performed by utilizing 3-[4,5-dimethylthiazol-2-yl]-2,5diphenyl tetrazolium bromide (MTT) (Sigma-Aldrich, Poole, UK) as a water-soluble yellow dye that is readily taken up by viable cells and reduced by the action of mitochondrial dehydrogenases. After appropriate treatment of the cells, the dye was released and dissolved by DMSO and the absorbances were measured under 535 nm wavelength.

Wound Healing Migration Assay
A wound healing migration assay was performed using a culture insert (ibidi GmbH, Gräfelfing, MUC, Germany) with a septum located between two small chambers for cell seeding. After cells reached confluence in complete medium for 24 h, the culture insert was removed followed by monitoring the migration of cells toward the blanking area between two monolayers of cells within 48 h of appropriate treatments under serum-free condition. Quantitation of cell motility was performed by counting the cells that have migrated into the blanking area using Image J software (version 1.50 i).

Transwell Migration Assay
Cells were seeded on a 24-well trans well migration insert (Nalge Nunc International, Rochester, NY, USA) in a complete medium for 24 h. After appropriate treatments, cells that had migrated to the underside of the insert membrane were stained with 0.3% crystal violet. The cells on the topside of the insert membrane were rubbed with a cotton swab. The migrated cells on the underside were imaged using phase-contrast microscopy with ×200 magnification. Quantitation of the migrated cell was performed by measuring the intensity of crystal violet staining, using Image J software.

Immunohistochemistry (IHC)
IHC was performed by the EnVision+ Dual Link System-HRP (DAKO, Carpinteria, CA, USA), a two-step staining technique using an HRP labeled polymer conjugated with secondary antibodies. Briefly, the tissue section was incubated with Dual Endogenous Enzyme Block to remove any endogenous peroxidase activity. Subsequently, the sample was incubated with primary Ab for 30 min, followed by the second Ab-HRP labeled polymer for 30 min. Staining is completed by a 5-10 min incubation with 3,3 -diaminobenzidine (DAB+) substrate chromogen which results in a brown-colored precipitate at the antigen site.

Statistical Analysis
ANOVA test was conducted to analyze the intensity differences between samples on the Western blot and the differences in cell motility between the indicated CCAs. Quantitative data were expressed as mean ± coefficient variation (CV). Correlation analyses assessing the expression and/or activation of the signal molecules associated with the extent of metastatic potentials of various CCAs were analyzed by a Chi-square test (SPSS 16.0 software, Chicago, IL, USA).

Screening the Enhanced Signal Molecules Associated with CCA Metastasis
To identify the signal molecules as potential therapeutic targets for preventing CCA progression, about 15 surgical tumor tissues of CCA were obtained from Hualein Tzu Chi hospital (IRB 109-148-A). The samples were processed for Western blot of a lot of critical signal molecules known to be involved in tumor metastasis of CCA, including the oncogenic RTKs (EGFR, Her2, Her3, and c-Met) [10][11][12]; Src, one of the nonreceptor tyrosine kinase [13,14]; JNK, one of the MAPK [15,16]; and AKT, a downstream molecule of PI3K [17][18][19][20]. In addition, Hic-5, one of the focal adhesion molecules known to be associated with progression of hepatocellular carcinoma (HCC) [21][22][23], was included. More than a two-fold increase of the expression and/or activation of the aforementioned signal molecules in tumor tissues compared with those in the normal counterparts were regarded as over-express or over-activated. Conversely, for the association of the signal molecules elevated in tumor tissues with metastasis, clinical-pathological data of CCA patients were employed. The criteria for evaluation of the metastatic potentials of each CCA tumor were set as 0, +, ++, +++ according to whether tumor cells were present in the lymph-vascular system and the number of lymph nodes exhibiting tumor invasion (Table 1). Among the CCA tested, 90% of them exhibited minor (+), moderate (++) or high (+++) metastatic potentials (Chi-square test, p < 0.05, n = 15). After Western blot analysis of the signaling molecules in CCA tissues, correlation analyses were performed to assess whether the over-expression and/or over-activation of the aforementioned signal molecules were associated with the extent of metastatic potentials of various CCAs. Accordingly, Hic-5 proteins were significantly increased in about 65% of metastatic CCA tissues (Chi-square test, p < 0.05, n = 13) similar to that observed in HCC [10]. Remarkably, over-expression of Hic-5 coupled with over-activation of Src, AKT, JNK (but not multiple RTKs such as EGFR, c-Met, Her2, and Her3) were observed in 50% of the metastatic CCAs (Chi-square test, p < 0.05, n = 13). In contrast, only 10-20% of them exhibited overexpression of multiple RTKs such as EGFR, c-Met, and Her3 but not elevated Hic-5, p-Src, p-AKT, and p-JNK (data not shown). In addition, 15% of CCA did not show any increase of these signal molecules. Representative results of four of the metastatic CCA samples SD37, SD114, SD143, and SD144 are demonstrated in Figure 1. Case SD37, SD114, and SD144 ( Figure 1a) exhibited simultaneous over-expression of Hic-5 coupled with over-activation of Src, AKT, JNK, compared with that in the normal counterpart. In contrast, EGFR and Her3 but not Hic-5, p-Src, p-AKT, and p-JNK, were elevated in SD143 (Figure 1b). We further validated the expression of Hic-5 on CCA tissue by immunohistochemistry (IHC) analysis. As demonstrated in Figure 2, much higher Hic-5 staining was observed in tumor tissues of SD37, SD114, and SD144 (but not SD143) than that in the normal counterpart, consistent with those observed in Western blot (Figure 1a,b). Interestingly, most of the enhanced Hic-5 staining locates within the peripheral regions which can be suspected of the migration front of the primary tumor. within the peripheral regions which can be suspected of the migration front of the prima tumor.

Cross Talk between Hic-5 and Src for Activating AKT in HuccT1
It is worthy of noting that not only the roles of Hic-5, Src, AKT, and JNK in metastatic signaling have been well established as mentioned above, but also the interaction and regulation between them were frequently mentioned. For example, Src activity was required for Hic-5 to promote TGF-β induced EMT, AKT-mediated EMT [24], and migration of gastric cancer [25]. In addition, JNK positively cross-talks with Hic-5 in the signaling pathway leading to HCC progression [22]. Since we found a large proportion of the metastatic CCAs exhibited simultaneous over-expression of Hic-5 and over-activation of Src, AKT, and JNK, it is very probable that they work together to trigger metastatic signaling for CCA progression. To address this issue, the up-and-down relationship of them were examined in a well-established CCA cell line, HuCCT1. HuCCT1 is derived from ascites of intrahepatic CCA (NCIt:C35417) within the metastatic site (ORDO: Orphanet 70567). Specifically, over-expression of Hic-5 coupled with activation of Src, AKT, and JNK were observed in HuCCT1 (data not shown). To begin with, we found one of the Src inhibitors PP1 (an inhibitor of the Src family Lck/Fyn) at 40 μM concentration dramatically suppressed protein expression of Hic-5 (by 90%) and significantly inhibited the activation of AKT (represented by phosphorylated AKT at Thr450) by 55% ( Figure 3a). Moreover, a highly potent Src inhibitor (dasatinib) at 0.1 μM greatly inhibited the expression of Hic-5 and activation AKT by 77 and 58%, respectively ( Figure 3c). However, phosphorylation of JNK (p-JNK) was not significantly inhibited by both Src inhibitors (data not shown). Conversely, knockdown of Hic-5 greatly suppressed the activation of Src (represented by phosphorylated Src at Tyr416, p-Src) and phosphorylation of JNK by 50-69% in a dosedependent manner using 12.5, 25, and 30 nM Hic-5 siRNA (Figure 3b), whereas AKT phosphorylation was significantly inhibited by Hic-5 siRNA at 25 and 30 nM (by 25-65%) but not at 12.5 nM (Figure 3b). Thus, Hic-5 expression was required for activation of Src, and on the contrary, Src activity was also required for Hic-5 expression indicating Hic-5 locates both upstream and downstream of Src. Whereas both Hic-5 and Src were required for AKT phosphorylation, Hic-5 but not Src was required for JNK phosphorylation. Taken together, these results implicated signal cross-talk occurred between Hic-5 and Src to trigger AKT activation in HuCCT1. However, JNK seems located downstream of Hic-5 but

Cross Talk between Hic-5 and Src for Activating AKT in HuccT1
It is worthy of noting that not only the roles of Hic-5, Src, AKT, and JNK in metastatic signaling have been well established as mentioned above, but also the interaction and regulation between them were frequently mentioned. For example, Src activity was required for Hic-5 to promote TGF-β induced EMT, AKT-mediated EMT [24], and migration of gastric cancer [25]. In addition, JNK positively cross-talks with Hic-5 in the signaling pathway leading to HCC progression [22]. Since we found a large proportion of the metastatic CCAs exhibited simultaneous over-expression of Hic-5 and over-activation of Src, AKT, and JNK, it is very probable that they work together to trigger metastatic signaling for CCA progression. To address this issue, the up-and-down relationship of them were examined in a well-established CCA cell line, HuCCT1. HuCCT1 is derived from ascites of intrahepatic CCA (NCIt:C35417) within the metastatic site (ORDO: Orphanet 70567). Specifically, over-expression of Hic-5 coupled with activation of Src, AKT, and JNK were observed in HuCCT1 (data not shown). To begin with, we found one of the Src inhibitors PP1 (an inhibitor of the Src family Lck/Fyn) at 40 µM concentration dramatically suppressed protein expression of Hic-5 (by 90%) and significantly inhibited the activation of AKT (represented by phosphorylated AKT at Thr450) by 55% (Figure 3a). Moreover, a highly potent Src inhibitor (dasatinib) at 0.1 µM greatly inhibited the expression of Hic-5 and activation AKT by 77 and 58%, respectively ( Figure 3c). However, phosphorylation of JNK (p-JNK) was not significantly inhibited by both Src inhibitors (data not shown). Conversely, knockdown of Hic-5 greatly suppressed the activation of Src (represented by phosphorylated Src at Tyr416, p-Src) and phosphorylation of JNK by 50-69% in a dose-dependent manner using 12.5, 25, and 30 nM Hic-5 siRNA (Figure 3b), whereas AKT phosphorylation was significantly inhibited by Hic-5 siRNA at 25 and 30 nM (by 25-65%) but not at 12.5 nM (Figure 3b). Thus, Hic-5 expression was required for activation of Src, and on the contrary, Src activity was also required for Hic-5 expression indicating Hic-5 locates both upstream and downstream of Src. Whereas both Hic-5 and Src were required for AKT phosphorylation, Hic-5 but not Src was required for JNK phosphorylation. Taken together, these results implicated signal cross-talk occurred between Hic-5 and Src to trigger AKT activation in HuCCT1. However, JNK seems located downstream of Hic-5 but not Src and probably not involved in the signaling pathway regulated by Src-Hic-5 cascade.
not Src and probably not involved in the signaling pathway regulated by Src-Hic-5 cascade.

Depleting Hic-5 Expression and Inhibiting Src Activity Suppress HuccT1 Migration in A Concerted Manner
We further investigated whether the Hic-5-Src positive feedback signal cascade is responsible for the tumor progression of HuCCT1. Initially, we examined whether Hic-5-Src signaling is required for cell growth of HuCCT1 using an MTT assay. The results showed that inhibition of Src activity by dasatinib (at 0.01-0.2 μM) but not knockdown of

Depleting Hic-5 Expression and Inhibiting Src Activity Suppress HuccT1 Migration in a Concerted Manner
We further investigated whether the Hic-5-Src positive feedback signal cascade is responsible for the tumor progression of HuCCT1. Initially, we examined whether Hic-5-Src signaling is required for cell growth of HuCCT1 using an MTT assay. The results showed that inhibition of Src activity by dasatinib (at 0.01-0.2 µM) but not knockdown of Hic-5 (using Hic-5 siRNA at 25 nM) marginally inhibit cell proliferation (by about 8-10%) of HuCCT1 (data not shown). Thus this signal cascade is not majorly involved in the proliferation of HuCCT1. Further, we examined whether Src activity is required for HuCCT1 cell migration by wound healing assay using Src inhibitor PP1. The cells were pre-treated with various doses of PP1 followed by re-plating the inhibitor-treated cells into a wound healing culture insert. As shown in Figure 4a, PP1 (6.7~26.8 µM) suppressed HuCCT1 cell migration by 30-70%. Further, using the trans well assay, HuCCT1 cell migration was effectively suppressed by the more potent Src inhibitor, dasatinib (at 0.01 to 5 µM), by 14-90% in a dose-dependent manner (Figure 4b). Thus Src activity is essential for HuCCT1 cell migration. To assess whether Hic-5 expression is also required for the migratory ability of HuCCT1, we transfected the cell with different doses (12,30, and 50 nM) of Hic-5 siRNA, followed by observing the extent of decrease in cell motility using a wound-healing assay. As demonstrated in Figure 4c, cell motility was decreased by 30-45% in cells transfected with 30 and 50 nM Hic-5 siRNA and marginally suppressed by 12.5 nM Hic-5 siRNA. This indicated depletion of Hic-5 decreased HuCCT1 cell migration in a dosedependent manner. Since we found positive cross-talk between Hic-5 and Src (Figure 3a Hic-5 (using Hic-5 siRNA at 25 nM) marginally inhibit cell proliferation (by about 8-10%) of HuCCT1 (data not shown). Thus this signal cascade is not majorly involved in the proliferation of HuCCT1. Further, we examined whether Src activity is required for HuCCT1 cell migration by wound healing assay using Src inhibitor PP1. The cells were pre-treated with various doses of PP1 followed by re-plating the inhibitor-treated cells into a wound healing culture insert. As shown in Figure 4a, PP1 (6.7~26.8 μM) suppressed HuCCT1 cell migration by 30-70%. Further, using the trans well assay, HuCCT1 cell migration was effectively suppressed by the more potent Src inhibitor, dasatinib (at 0.01 to 5 μM), by 14-90% in a dose-dependent manner (Figure 4b). Thus Src activity is essential for HuCCT1 cell migration. To assess whether Hic-5 expression is also required for the migratory ability of HuCCT1, we transfected the cell with different doses (12,30, and 50 nM) of Hic-5 siRNA, followed by observing the extent of decrease in cell motility using a wound-healing assay. As demonstrated in Figure 4c

The Role of Focal Adhesion Signaling Molecules Responsible for Cell Migration in Tumor Metastasis
It has been well established that cell migration is one of the critical steps in tumor metastasis [26][27][28][29]. Cell migration is triggered by signaling pathways not only from cellular receptors (such as RTK) but also from focal adhesion signaling molecules including integrin, focal adhesion kinase (FAK), Src, Rac [30], Hic-5 [23,24], and other integrin-associated proteins. In one early study, IHC demonstrated Ep-CAM, an integrin-associated focal adhesion molecule capable of regulating cancer cell adhesion [31], was overexpressed in 95% CCA but not HCC [32]. FAK activation mediated HGF-induced proliferation and invasion of the human cholangiocarcinoma cell line, HuCCA-1 [33]. This was essential for tumor necrosis factor-alpha-dependent matrix metalloproteinase-9 production in a CCA cell line, CCKS1 [34]. A recent study further demonstrated FAK activation contributes to the initiation and progression of iCCA by inducing the YAP proto-oncogene [35]. Regarding the focal adhesion molecules in our study, not only that Src [30] and Hic-5 [21,22] are key effectors responsible for cell migration and tumor progression but also they may interact with each other [36]. Thus, targeting critical signal complex in focal adhesion may prevent metastasis. For example, it is promising to target Rac [37] and integrin [38] for preventing the progression of solid tumor such as cutaneous melanoma. Herein we suggested Src and Hic-5 can be candidate targets for therapeutic approaches against the progression of CCA.

The Role of Focal Adhesion Signaling Molecules Responsible for Cell Migration in Tumor Metastasis
It has been well established that cell migration is one of the critical steps in tumor metastasis [26][27][28][29]. Cell migration is triggered by signaling pathways not only from cellular receptors (such as RTK) but also from focal adhesion signaling molecules including integrin, focal adhesion kinase (FAK), Src, Rac [30], Hic-5 [23,24], and other integrin-associated proteins. In one early study, IHC demonstrated Ep-CAM, an integrin-associated focal adhesion molecule capable of regulating cancer cell adhesion [31], was overexpressed in 95% CCA but not HCC [32]. FAK activation mediated HGF-induced proliferation and invasion of the human cholangiocarcinoma cell line, HuCCA-1 [33]. This was essential for tumor necrosis factor-alpha-dependent matrix metalloproteinase-9 production in a CCA cell line, CCKS1 [34]. A recent study further demonstrated FAK activation contributes to the initiation and progression of iCCA by inducing the YAP proto-oncogene [35]. Regarding the focal adhesion molecules in our study, not only that Src [30] and Hic-5 [21,22] are key effectors responsible for cell migration and tumor progression but also they may interact with each other [36]. Thus, targeting critical signal complex in focal adhesion may prevent metastasis. For example, it is promising to target Rac [37] and integrin [38] for preventing the progression of solid tumor such as cutaneous melanoma. Herein we suggested Src and Hic-5 can be candidate targets for therapeutic approaches against the progression of CCA.

Targeting Hic-5 and Src Is Promising in Preventing CCA Progression
In the past decades, the local and systemic therapies along with target therapy aiming at metastatic signaling cascade including various oncogenic RTKs and diverse intracellular signaling did not significantly improve the prognosis of CCA patients, and strongly suggests the need for novel therapeutic agents and strategies. One of the major obstacles in targeting RTK signaling is due to acquired or inducible drug resistances. Usually, targeting upstream signal molecules is often attenuated by downstream or parallel alterations in the pathway. This is exemplified by resistance to anti-EGFR therapies in colon cancer, which can be mediated by downstream KRAS-or NRAS-activating mutations. In addition, RTK target therapy may be useless due to co-expression of multiple growth factors that can raise compensatory secondary signaling after treatment with specific tyrosine kinase inhibitors (TKIs) [39,40]. For example, in the anti-EGFR target therapy using gefitinib for lung cancer, MET (receptor of HGF) amplification leads to resistance by activating ERBB3 [41] signaling. Recently, metastatic signaling downstream of oncogenic receptors including Hic-5 and Src are emerging. Previously, Hic-5 is increased upon activation of 20-HETE/GPR75 (G-protein coupled receptor 75) that triggering metastatic features of androgen-insensitive prostate cancer cells [42]. In addition, Hic-5 is one of the critical mediators for c-Met to trigger HCC progression [43] and regulated ESCC cell migration and invasion induced by TGFβ [44]. Conversely, Src kinase is one of the downstream signal molecules mediating HER3 (coupled with EGFR or HER2) triggered tumor progression [45]. In addition, Src mediated the reactivation of RTK signaling responsible for the induced resistance of BRAF and MEK inhibitors during the treatment of BRAF mutant melanoma [46]. In the past decades, Src has been found to be a promising target for suppressing tumor progression. For example, an innate immune sensor, NOD1, exerted its antitumor effect on HCC by directly inhibiting the Src-MAPK axis. [47]. Moreover, dasatinib prevented CCA progression in PDX models [48] and reduced the viability of sorafenib-resistant (SR) HCC cells by inhibiting Src [49]. In this report, inhibition of Src activity by dasatinib (at 0.1 µM) prevented cell migration of HuCCT1 by 80% (Figure 3), and suppressed HuCCT1 cell proliferation by 8-10% (data not shown), consistent with the inhibitory effect of dasatinib on CCA progression observed in vivo [48]. Conversely, whether Hic-5 can be a target for preventing metastasis of tumors are far less investigated in preclinical/clinical setting. Previously, siRNA silencing of Hic-5 suppressed migration and invasion of pancreatic tumor [50] and HCC [21,22]. Herein, we demonstrated knockdown of Hic-5 suppresses migration of HuCCT1 in a dose-dependent manner (Figure 4c). It is worth investigating whether depletion of Hic-5 may prevent CCA progression in vivo.

Combined Therapy Is More Effective in Blocking Metastatic Signaling
Recently, combined therapy is emerging as a more effective and safe therapeutic approach in cancer prevention. This is based on multiple targeting against the metastatic pathways may reduce the compensatory signaling encountered in single targeting approaches. Moreover, a lower concentration of each of the inhibitors used in combined treatment can avoid the side effects caused by drugs at higher concentrations used in single drug treatments. In fact, co-targeting PARP1 and Src have been found to improve the therapeutic strategies for HCC [51]. In the present study, we demonstrated co-targeting Hic-5 and Src, which positively crosstalk with each other, may enhance the inhibitory effect on migration of HuCCT1 under lower concentration of Hic-5 siRNA and Src inhibitor dasatinib (Figure 4e).

Conclusions and Perspective
To develop more precise and personalized therapeutic strategies for improving the prognosis of CCA, the potential molecular target associated with metastasis of CCA were identified in CCA tissues. Elevated expression of Hic-5 coupled with activation of Src and the downstream AKT and JNK were observed in 50% metastatic CCAs. Pathway analysis further proved that Hic-5 and Src, which are critical interactive proteins in focal adhesion, cross-talk with each other to trigger the downstream AKT signaling. Remarkably, simultaneous suppression of both Hic-5 expression and Src activation prevent HuCCT1 migration in a collaborated manner. Although cell migration is one of the essential steps in tumor metastasis, the therapeutic implication conveyed in this study needs to be validated in more advanced models, such as 3D cultures using animal [52] and human [53] tissues. These can be employed for validating the efficiency of the double Hic-5/Src targeting approach in preventing the progression of CCAs with enhanced Hic-5-Src signaling.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: Not applicable, there is not any data available for supporting the results in this study.