RBFOX3 Promotes Gastric Cancer Growth and Progression through Activating HTERT Signaling

Tumor invasion, metastasis, and recrudesce remain a considerable challenge in the treatment of gastric cancer (GC). Herein, we first identified that RBFOX3 (RNA binding protein fox-1 homolog 3) was significantly up-regulated in GC tissues and negatively linked to the survival rate of GC patients. RBFOX3 promoted cell division and cell cycle progression in vitro as well as in vivo. Furthermore, RBFOX3 increased cell invasion and migration ability. Interestingly, both the suppression of GC cell multiplication and invasion moderated by the silencing of RBFOX3 was rescued by HTERT up-regulation. Additionally, RBFOX3 augmented the resistance of GC cells to 5-fluorouracil (5-Fu) by repressing RBFOX3. Mechanistically, exogenous upregulation of RBFOX3 triggered promoter activity and HTERT expression thereby enhancing the division and development of GC cells. Importantly, our findings revealed that RBFOX3 interacted with AP-2β to modulate the HTERT expression as demonstrated by co-immunoprecipitation analysis. In conclusion, our study indicates Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 16 April 2020 doi:10.20944/preprints202004.0264.v1 © 2020 by the author(s). Distributed under a Creative Commons CC BY license. that high expression of RBFOX3 promotes GC progression and development but predicts worse prognosis by stimulating HTERT signaling. Moreover, the results suggest that the RBFOX3/AP-2β/HTERT pathway is a novel target for the development of therapeutic agents for the prevention and treatment of GC reappearance and metastasis.

enhances RBFOX3 binding target UGCAUG motif [10] . However, recent studies suggest that RBFOX3 also modulates various physiological pathological processes. It has been reported that neuronal nuclei (NeuN) is a product of the Rbfox3 gene, revealing RBFOX3 as a marker exclusively expressed in post-mitotic neurons [11] . Another study showed that RBFOX3 could bind to DNA in vitro [12,13] . RBFOX3 has also been reported to control biogenesis of some miRNAs including primary-miRNAs (pri-miRNAs) that lack a UGCAUG motif [14] . We hypothesized that in addition to its splicing functions, RBFOX3 regulates various biochemical processes that are poorly understood. Here, we find that RBFOX3 regulates HTERT expression to promote GC. We find that RBFOX3 is significantly upregulated in GC, and that it correlates with poor survival. Functional analyses revealed that that RBFOX3 enhances GC growth, metastasis, and chemoresistance. Mechanistic studies reveal that RBFOX3 overexpression elevates HTERT expression, promoting GC progression. Moreover, our data indicate that RBFOX3 interacts with AP-2β to regulate HTERT expression. Taken together, our findings highlight the RBFOX3/HTERT signaling axis as a potential novel therapeutic target against GC.

Materials and Methods samples
Clinical samples 178 paired human GC tissues and adjacent non-cancer tissue were obtained from the tissue biobank at the department of pathology at the Second affiliated hospital of Nanchang University. The samples had been biobanked between August 2017 and December 2019 and 92 were from males and 86 from females. Adjacent, matched noncancer tissues were collected > 5 cm away from the edge of the cancerous foci. Ethical approval for this study was provided by the clinical research ethics committee of the second affiliated Hospital of Nanchang University.

Streptavidin-agarose pull down assay
HTERT promoter binding proteins were analyzed by streptavidin-agarose pulldown assays. Briefly, nuclear proteins were extracted from GC cells and 1mg incubated with 10µg of biotinylated double-stranded DNA probes (Sigma-Aldrich) corresponding to nucleotide -351 to -149 of the HTERT promoter region and 100µl of streptavidin-agarose beads (Sigma-Aldrich) at 4°C overnight. The DNA-protein complex was then pulled down by centrifuging at 500 g for 10 minutes, at 4°C.

Identification of HTERT promoter-binding proteins
HTERT promoter-bound proteins from section streptavidin-agarose pulldown assay were analyzed by mass spectrometry. Briefly, bound proteins were separated by 10% SDS-PAGE and visualized by sliver staining (Beyotime, P0015A). After reduction and alkylation, the bands of interest were candidate protein bands were digested MSgrade trypsin solution (Promega, CAS9002-07-7). The digested peptides were then identified by mass spectrometry. The identities of the proteins of interest were verified on available databases and software.

Chromatin immunoprecipitation (ChIP) assay
ChIP assay was sone using Carey's protocol. The PCR products were resolved on a 2% agarose gel and visualized by Gel-Red staining.

Lentiviral Construction and Cell Transfection
To generate clones stably overexpressing RBFOX3, we infected SGC-7901 and

Colony formation and 5-Ethynyl-20-deoxyuridine (Edu) assay
200-400 GC cells/well were onto 6-well plates and cultured for 2 weeks. Next, colonies were stained with 1% crystal violet and counted. Colony formation assays were performed as we previously described [15] . All experiments were performed in independent triplicates.
1×10 5 SGC-7901 and MGC-803 cells/well were seeded onto 6-well plates and cultured to 75%-85% confluence. 100μL of 10uM prewarmed Edu was added into each well and the cells put in culture 2 hours. They were then fixed with 4% PFA for 15 minutes. 50μL of 2mg/mL glycine solution was then decolorized for 5 min. The cells were then permeabilized with 0.3% TriX-100 PBS for 10 minutes, followed by Apollo and Hoechst staining. The cells were then examined by confocal microscopy and the Hoechst nuclear staining imaged.

Wound healing and transwell invasion assays
To evaluate the migration and invasion capacity of the GC cells after BRFOX3 silencing or overexpression, we performed wound healing assay and trans-well invasion assays. GC cells stably overexpressing RBFOX3 or RBFOX3 silenced, and the respective controls were cultured on 6-well plates to confluence. The monolayers were then scratched with a 10μl pipette tip. The cells were then imaged 0 and 24 hours later to monitor migration. These experiments were performed in independent triplicates. Cell invasion assays were done using BD BioCoat matrigel invasion chambers (BD, 354480) by following manufacturer instructions. Invading cells in 5 randomly selected fields of view were counted under a light microscope.

Western blot
Whole cell and nuclear protein extracts extracts were prepared using Complete lysis-M reagent (Roche, 4719956001) and RIPA lysis buffer (Beyotime, P0013B).
Protein concentration was determined using BCA assay (ThermoFisher Scientific, 23221-23230). Proteins were then separated by 8%-10% SDS-PAGE and transferred onto 0.45μm PVDF membranes for antibody staining and detection.

Co-immunoprecipitation (CoIP) assay
Equal amounts of nuclei protein extracts from different cell lines were incubated with the indicated antibodies. Next, 50μL agarose-conjugated protein-A/G beads (Merck Millipore, YB36403ES03) were added and the mixture incubated at 4°C overnight. After extensive washing with ice cold PBS 1X, the beads were mixed with loading buffer and heated at 4°C for 5 minutes. The supernatant was then subjected to western blot analysis. Tumor growth and metastases were monitored weekly using an IVIS system (Caliper Life Sciences). The mice were sacrificed after six weeks and the tumors surgically removed for analyses.

Statistical analysis
Data were presented as mean ± SD of at least three independent experiments.
Statistical analyses were done using SPSS version 11.0. * indicates statistical significance. p value <0.05 was considered statistically significant.

RBFOX3 is overexpressed in human GC and positively correlates with tumor progression
To determine the expression pattern of RBFOX3 in human GC tissues, RBFOX3 DNA digestion products were first identified by agarose gel electrophoresis. The DNA digestion products were then stably expressed in SGC-7901 and MGC-803 cells ( Figure   1A). RT-qPCR analysis of RBFOX3 expression revealed that it was consistently elevated in GC tissues (p<0.001, Figure 1B). IHC analyses of 89 GC tissues and 89 matched adjacent non-cancer gastric tissue revealed elevated RBFOX3 staining in the GC tissues relative to controls (p<0.01, Figure 1C). Similar observations were made upon western blot analysis of RBFOX3 in 37 frozen GC tissues and paired non-cancer control tissues (p<0.05, Figure 1D). Western blot analysis of cell lines revealed that RBFOX3 was poorly expressed in the normal gastric cell line, GES-1 relative to the GC cell lines, MKN45, AGS, SGC-7901, MGC-803, BGC-823 ( Figure 1E). Next, we evaluated the association between RBFOX3 protein expression and clinicopathological features of GC (Table 1). This analysis indicated that RBFOX3 expression correlates with tumor differentiation (p=0.017), AJCC clinical stage (p=0.015) and TNM stage (p=0.004). Additionally, Kaplan-Meier survival analysis revealed significantly lower overall survival (OS) and disease-free survival (DFS) rates in patients exhibiting high tumor RBFOX3 levels relative to those expressing low RBFOX3 levels ( Figure 1F, 1G).
Univariate and multivariate Cox regression analysis showed that advanced TNM stage and high RBFOX3 expression significantly corelated with unfavorable OS and DFS.
Taken together, these results indicate that High RBFOX3 protein expression is associated with poor prognosis, highlighting the potential of RBFXO3 as an independent prognostic marker in GC (HR=2.670; 95% CI: 1.471-3.917; p=0.001; Table 2).

RBFOX3 promotes GC cell proliferation and cell cycle progression in vitro
Next, we evaluated the function of RBFOX3 in GC cell growth in vitro by stably overexpressing RBFOX3 in SGC-7901 cells and stably silencing it in MGC-803 cells.

RBFOX3 regulates GC cell migration and invasion in vitro
The HTERT signaling pathway has been reported as a modulator of cancer cell migration and invasion [16][17][18] . We therefore investigated the role of RBFOX3 on GC cell migration and invasion using wound healing and transwell invasion assays. These  Figure 3H).

RBFOX3 inhibition enhabces 5-Fu sensitivity in GC cells
Although 5-Fu (5-fluorouracil) is commonly used to treat GC, it's dosage is limited by its negative side effects [19] . We hypothesized that given the effect of RBFOX3 on GC growth and proliferation, its knockdown might enhance sensitivity GC sensitivity to 5-Fu. We observed that 5-Fu inhibits GC cell viability and colony formation in a dose dependent manner. Additionally, these effects were significantly stronger in the background RBFOX3 silencing (p<0.05, Figure 4A). Next, we first generated MGC-803 stably expressing RBFOX3 shRNA (shRB-1 and shRB-2) or the negative control (sh-NC). This analysis revealed that cell viability in RBFOX3 knockdown cells treated with 5-Fu was significantly lower relative to mock knockdown cells treated with 5-Fu (p<0.05, Figure 4B). Moreover, western blot analysis revealed that RBFOX3 expression was significantly reduced upon 5-Fu treatment relative the controls (p<0.05, Figure 4C). Taken together, this suggested that RBFOX3 regulates GC cells sensitivity to 5-Fu. However, further studies are needed to fully understand the mechanisms of RBFOX3's regulation of drug resistance.

RBFOX3 binds to the HTERT promoter in GC cells
Previous studies have uncovered novel regulators of the HTERT promoter in lung and liver cancer using streptavidin-agarose bead pull down assays [17,20] .  Figure 6A).
Similar results were obtained in MGC-803 stably silenced for RBFOX3 expression (p<0.05, Figure 6B). Additionally, cell proliferation, viability and invasion capacity were restored (p<0.05, Figure6 C-H). Together, these findings further confirm the regulation of HTERT expression by RBFOX3.

RBFOX3 interacts with AP-2β to regulate HTERT expression
Next, we investigated whether RBFOX3 bound to the HTERT promoter interacts with other transcription factors. Previous studies have found that HTERT promoterbinding proteins include KLF4 [21] , RFPL3 [22] , CPSF4 [23] and AP-2β [17] . To test this possibility, we performed a CoIP pull-down experiment and using a RBFOX3-specific antibody observed that RBFOX3 interacts with AP-2β ( Figure 7A-B). Moreover, we found that AP-2β binds to the HTERT promoter ( Figure 7C). A co-immunofluorescence analysis of SGC-7901 and MGC-803 cells revealed that RBFOX3 primarily localizes in the cytoplasm while AP-2β was primarily found in the nucleus ( Figure 7D). The colocalization between AP-2β and RBFOX3 was observed in the cytoplasm ( Figure 7D).
Additionally, we found that AP-2β overexpression enhances RBFOX3 to the HTERT promoter ( Figure 7E). Conversely, AP-2β knockdown weakened the binding of RBFOX3 to the HTERT promoter even in the context of RBFOX3 overexpression (SGC-7901 cells) ( Figure 7E). A luciferase reporter analysis showed that AP-2β overexpression enhanced the HTERT promoter activity while AP-2β knockdown suppressd HTERT promoter activity ( Figure 7G). To test the effect of AP-2β on HTERT expression, we performed western blot analysis and found that AP-2β overexpression enhanced HTERT expression. Connversely, AP-2β knockdown suppressed HTERT expression ( Figure 7F). Moreover, AP-2β knockdown suppressed RBFOX3 overexpression driven GC cell invasion. Conversely, AP-2β overexpression partially rescued the inhibition of GC invasion caused by RBFOX3 knockdown ( Figure 7H).
These findings suggest that RBFOX3 interacts with AP-2β to regulate the HTERT expression and GC metastasis.

Deregulation of RBFOX3 suppresses tumor growth in GC orthotopic xenografts
To evaluate whether RBFOX3 has oncogenic functions, we created GC orthotopic mouse xenografts using SGC-7901 cells carrying a luciferase reporter and stably overexpressing RBFOX3, as well as MGC-803 cells carrying as luciferase reported and stably knocked down for RBFOX3. As controls, we used vector groups and sh-NC groups, respectively. The transfected cells and the control cells were inoculated into the right and left renal capsule of the same mouse, respectively (8 mice per group). Tumor growth was monitored using IVIS. 6 weeks into the experiment, we observed that RBFOX3 knockdown tumors were significantly smaller than the controls while RBFOX3 overexpression ones were significantly bigger than the controls ( Figure 8A and 8D). Analysis of tumore size revealed that RBFOX3 up-regulation significantly promoted tumor volume and weight ( Figure 8B and 8C). However, knocking down RBFOX3 observed the opposite result ( Figure 8E and 8F). Taken together, these results indicate that deregulation of RBFOX3 may suppresses tumor progression in vivo.

Discussion
Limitless self-renewal is a hallmark of cancer [24] . Telomere maintenance and telomerase activation has been reported to promote cancer cell proliferation [25] . The transcriptional regulation of HTERT is believed to modulate telomerase activation in human cancers. It is thought that a contributing factor to cancer development is their acquired ability to overcome senescence by maintaining telomere length [26,27] . HTERT has been found to be upregulated in various tumors via genetic and epigenetic means including HTERT amplifications, HTERT structural variation, HTERT promoter mutations and HTERT promoter methylation [28,29] . As the catalytic subunit of telomerase, HTERT plays a decisive role in cell unlimited replication [29][30][31] . Recent findings have implicated HTERT in various human diseases, including cancer [17,30,32] .
Currently, HTERT deregulation is considered a hallmark of cancer and a potential therapeutic target [30,31,33] . Mounting evidence indicates that tumor-specific cellular factors may be differentially expressed and specifically bind to the HTERT promoter, thereby modulating HTERT expression and tumor development [32,34] . Here, we found that RBFOX3 highly expressed in GC tissues and cell lines, and that RBFOX3 dunctions as an oncogene. Our data showed that patients expressing high RBFOX3 levels exhibit significantly shorter OS and DFS. Furthermore, univariate and multivariate analyses showed that high RBFOX3 expression may independently predict poor GC prognosis.
Next, we investigated significance of high RBFOX3 expression in GC. Functional analysis revealed that RBFOX3 overexpression promotes proliferation, invasion and migration GC cells. However, in the background of stably silenced RBFOX3, the proliferation, invasion and migration of GC cells were suppressed. Additionally, we also observed that RBFOX3 influences GC cell sensitivity to 5-Fu. It has been previously reported that RBFOX3 binds to the promoter region of HTERT, thereby regulating HTERT signaling and tumor growth. Interestingly [17,33] , we find that the RBFOX3 knockdown-mediated suppression of both GC proliferation and invasion is rescued by HTERT overexpression. Additionally, HTERT suppression did not affect RBFOX3 levels. In addition, cell proliferative and invasion were restored. Taken together, these experiments further indicated that HTERT expression is regulated by RBFOX3.
Finally, we observed that RBFOX3 interacts with AP-2β to regulate HTERT expression. AP-2β binding to the HTERT promoter was confirmed by ChIP. AP-2β Overexpression enhanced RBFOX3 binding to HTERT promoter. Conversely, AP-2β knockdown weakened RBFOX3 binding to the HTERT promoter even in RBFOX3 overexpressing SGC-7901 cells. A luciferase reporter assay revealed that AP-2β overexpression enhances HTERT promoter activity, which is onhibited by AP-2β knockdown. To evaluate the effect of AP-2β on HTERT expression, we perfomed western blot analysis and observed AP-2β upregulates HTERT while HTERT downregulation is achieved by AP-2β knockdown. Analysis of the effect of AP-2β on cell viability by MTS assay, revealed that AP-2β knockdown inhibits the RBFOX3 overexpression driven GC cell growth. On the other hand, AP-2β overexpression partially rescued RBFOX3 knockdown-mediated growth inhibition. Together, these data show that RBFOX3 interacts with AP-2β to regulate HTERT expression and GC cell growth.
RBFOX3 is an antigen of the neuronal marker antibody NeuN [35] , which makes its involvement in GC surprising. Recent studies indicate that RBFOX3 has a wide range of physiological functions on top of its traditional role as a an alternative splicing factor [17,36] . RBFOX3 has been found to bind DNA in vitro and to control transcription of a subset of microRNAs [12,14] . However, its role in mammalian development and homeostasis has remained elusive. RBFOX3 possesses an RNA recognition motif (RRM)-type RNA binding domain (RBD) which enables it to regulate splicing events on various transcripts by binding the (U)GCAUG sequence on RNA [14,37] . Here, have presented strong evidence that RBFOX3 binds to the HTERT promoter to regulate its expression and GC cell growth. However, it is not clear whether the RNA-binding function of RBFOX3 in these processes. Even so, our findings reveals a novel role for RBFOX3 and further examination of its roles in tumorigenesis is meritted.
In summary, our study reveals that RBFOX3 functions as an oncogene, driving GC cell proliferation, migration and invasion. Furthermore, we have find that RBFOX3