SH3BP2 Silencing Increases miRNAs Targeting ETV1 and Microphthalmia-Associated Transcription Factor, Decreasing the Proliferation of Gastrointestinal Stromal Tumors

Simple Summary A previous study showed that silencing the adaptor molecule SH3 Binding Protein 2 (SH3BP2) reduced oncogenic KIT and PDGFRA receptor levels and impaired gastrointestinal stromal tumor (GIST) growth. This study tries to get insights into the molecular mechanism underlying this effect. The silencing of SH3BP2 induces miRNAs (miR-1246 and miR-5100), which target microphthalmia-associated transcription factor (MITF) and ETV1, a linage survival factor involved in GIST tumorigenesis. Altogether, this results in decreased tumor cell viability and enhanced apoptosis. Abstract Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumors of the gastrointestinal tract. Gain of function in receptor tyrosine kinases type III, KIT, or PDGFRA drives the majority of GIST. Previously, our group reported that silencing of the adaptor molecule SH3 Binding Protein 2 (SH3BP2) downregulated KIT and PDGFRA and microphthalmia-associated transcription factor (MITF) levels and reduced tumor growth. This study shows that SH3BP2 silencing also decreases levels of ETV1, a required factor for GIST growth. To dissect the SH3BP2 pathway in GIST cells, we performed a miRNA array in SH3BP2-silenced GIST cell lines. Among the most up-regulated miRNAs, we found miR-1246 and miR-5100 to be predicted to target MITF and ETV1. Overexpression of these miRNAs led to a decrease in MITF and ETV1 levels. In this context, cell viability and cell cycle progression were affected, and a reduction in BCL2 and CDK2 was observed. Interestingly, overexpression of MITF enhanced cell proliferation and significantly rescued the viability of miRNA-transduced cells. Altogether, the KIT-SH3BP2-MITF/ETV1 pathway deserves to be considered in GIST cell survival and proliferation.


Introduction
Gastrointestinal stromal tumors (GISTs) are the most common type of soft tissue sarcoma in the intestinal tract [1]. They are derived from the interstitial cells of Cajal (ICCs), located in the submucosa and myenteric plexus of the gastrointestinal tract [2]. The pathogenesis of GISTs is defined by mutually exclusive mutations in KIT (75-80%) and platelet-derived growth factor receptor α (PDGFRA) genes (5-10%). Additionally, 10-15% of GISTs lack KIT mutations, the so-called "wild type." They are classified as deficient in succinate dehydrogenase (SDH)-deficient and non-SHD-deficient. Non-SDHdeficient include NF type 1 neurofibromatosis and GISTs with BRAF, KRAS, and PIK3CA mutations [3].

Cell Culture
Human GIST cell lines GIST882, GIST48, and GIST-T1 were kindly provided by Dr. S. Bauer. GIST cell lines were-cultured as described elsewhere [11]. Transient transfections were carried out using Opti-MEM (Gibco, Carlsbad, CA, USA). The mycoplasma test was performed routinely in all cell lines used.

RNA Extraction, Retrotranscription, and PCR Assays
Total RNA was extracted with a miRCURY RNA Isolation Kit (Exiqon, Vedbaek, Denmark) from NT control and SH3BP2 knockdown GIST cells. cDNA was generated by reverse transcription using the miRCURY LNA RT Kit. Quantitative, Real-Time PCR for miRNA PCR assay was performed using the miRCURY SYBR Green PCR Kit, and following miRCURY LNA miRNA PCR assay protocol on a LightCycler ® 480 Instrument II (LifeScience Roche). miR-30c-5p and miR-335 were used as housekeeping miRNA genes.

MicroRNA Array Profiling
All experiments were conducted at Exiqon Services, Denmark. The quality of all the total RNA was verified by an Agilent 2100 Bioanalyzer profile. 750 ng total RNA from both sample and reference was labeled with Hy3™ and Hy5™ fluorescent labels, respectively, using the miRCURY LNA™ microRNA Hi-Power Labeling Kit, Hy3™/Hy5™ (Exiqon, Vedbaek, Denmark), following the procedure described by the manufacturer. The Hy3™-labeled samples and a Hy5™-labeled reference RNA sample were mixed pair-wise and hybridized to the miRCURY LNA™ microRNA Array 7th Gen (Exiqon, Denmark), which contains capture probes targeting all microRNAs for humans, mice, or rats registered in the miRBASE 18.0. The hybridization was performed according to the miRCURY LNA™ microRNA Array Instruction manual using a Tecan HS4800™ hybridization station (Tecan, Austria). After hybridization, the microarray slides were scanned and stored in an ozone-free environment (ozone level below 2.0 ppb) to prevent potential bleaching of the fluorescent dyes. The miRCURY LNA™ microRNA Array slides were scanned using the Agilent G2565BA Microarray Scanner System (Agilent Technologies, Inc., Santa Clara, CA, USA), and the image analysis was carried out using the ImaGene ® 9 (miRCURY LNA™ microRNA Array Analysis Software, Exiqon, Denmark). The quantified signals were background corrected (Normexp with offset value 10, see [15]) and normalized using the global Lowess (Locally Weighted Scatterplot Smoothing) regression algorithm. Among the 502 human miRNAs detected by the array, we discarded any miRNA with an Average Hy3 signal under 7.5.

Lentiviral Transduction
Lentiviral particles to silence the SH3BP2 gene expression were previously described [5]. Lentiviral transduction for NT (non-target) was performed as described in [4] with slight modifications. PLenti-III-mir-GFP-blank was the plasmid used as a control. Plenti-III-miR-GFP miRNAs (miR-1246 and miR-5100) were obtained from Applied Biological Materials Inc (Richmond, BC, Canada). GIST cells were transduced in the presence of 8 µL/mL of Polybrene (Santa Cruz, CA, USA), and puromycin selection (1 µg/mL) was carried out after one day from transduction.

Cell Cycle Analysis by Flow Cytometry
GIST cells were collected on the 5th and 7th days after transduction. The cells were fixed with 70% ethanol at 4 • C overnight and stained with propidium iodide buffer, as described elsewhere [18]. Data were acquired in FACS Calibur and analyzed using model Dean/Jet/Fox FlowJo 7.6 software.

Statistical Data Analysis
After determining the normal distribution of the samples and variance analysis, an unpaired student's t-test was used to determine significant differences (p-value) between the two experimental groups. A one-way ANOVA test was used to determine significant differences (p-value) between several experimental groups. All results are expressed as mean ± standard error of the mean (SEM).

SH3BP2 Silencing Reduces ETV1 Levels in GIST Cell Lines
In previous work, we showed that silencing of SH3BP2 diminished KIT, PDGFRA, and MITF levels lead to a reduction in tumor growth in vitro and in vivo [5]. To better understand the role of SH3BP2 in GIST survival, we checked whether silencing of SH3BP2 was also affecting ETV1, a master regulator of the normal linage of interstitial Cajal cells, which cooperates with KIT in GIST [14,19]. As shown in Figure 1 and Supplementary Material S1, silencing of SH3BP2 reduces MITF, as previously reported, and ETV1 protein levels in imatinib-sensitive and imatinib-resistant GIST cell lines.

miRNA Profiling of SH3BP2-Silenced GIST Cells
SH3BP2 silencing reduces MITF at the protein level but not at the mRNA level in GIST [5], suggesting a post-transcriptional regulatory mechanism. KIT can regulate MITF through selective miRNA expression in mast cells [12]. Thus, we next performed a miRNA microarray to identify miRNAs regulated by SH3BP2 in GIST882 and GIST48-silenced cells to get insights into the signaling pathway that leads to apoptosis. Figure 2 shows the heat map representation of the two-way hierarchical clustering of miRNAs and samples. Interestingly, the samples cluster according to their biological group, meaning a very different miRNA profile exists between Non-Target and SH3BP2-silenced cells independently of the cell type.
A p-value < 0.05 was used to define significantly deregulated miRNAs between the different groups. This criterion identified 162 and 130 miRNAs in GIST882 and GIST48, respectively, with 107 in common ( Figure 2A). In Figure 2B, a four-way Venn diagram shows that 32 miRNAs are downregulated in both cell lines, and 56 are up-regulated among the significantly changed miRNAs. Among them, a threshold of 1.5-fold change defined the 21 most up-regulated and the 12 most downregulated miRNAs in both cell lines ( Figure 2C). Several databases were used to predict miRNA-target interactions with these miRNAs (Tables S1 and S2).

miRNA Profiling of SH3BP2-Silenced GIST Cells
SH3BP2 silencing reduces MITF at the protein level but not at th GIST [5], suggesting a post-transcriptional regulatory mechanism. KIT through selective miRNA expression in mast cells [12]. Thus, we next pe microarray to identify miRNAs regulated by SH3BP2 in GIST882 an cells to get insights into the signaling pathway that leads to apoptosis. F heat map representation of the two-way hierarchical clustering of miR Interestingly, the samples cluster according to their biological group, m ferent miRNA profile exists between Non-Target and SH3BP2-sil pendently of the cell type.

miRNA Profiling of SH3BP2-Silenced GIST Cells
SH3BP2 silencing reduces MITF at the protein level but not at the mRNA level in GIST [5], suggesting a post-transcriptional regulatory mechanism. KIT can regulate MITF through selective miRNA expression in mast cells [12]. Thus, we next performed a miRNA microarray to identify miRNAs regulated by SH3BP2 in GIST882 and GIST48-silenced cells to get insights into the signaling pathway that leads to apoptosis. Figure 2 shows the heat map representation of the two-way hierarchical clustering of miRNAs and samples. Interestingly, the samples cluster according to their biological group, meaning a very different miRNA profile exists between Non-Target and SH3BP2-silenced cells independently of the cell type. A p-value < 0.05 was used to define significantly deregulated miRNAs between the different groups. This criterion identified 162 and 130 miRNAs in GIST882 and GIST48, respectively, with 107 in common ( Figure 2A). In Figure 2B, a four-way Venn diagram shows that 32 miRNAs are downregulated in both cell lines, and 56 are up-regulated among the significantly changed miRNAs. Among them, a threshold of 1.5-fold change defined the 21 most up-regulated and the 12 most downregulated miRNAs in both cell lines ( Figure 2C). Several databases were used to predict miRNA-target interactions with these miRNAs (Tables S1 and S2).
Next, we validated these five miRNAs in various GIST cell lines. Quantitative realtime PCR was carried out in SH3BP2 silenced GIST-T1 ( Figure 3A), GIST882 ( Figure 3B), and GIST48 cells ( Figure 3C). Only two of the five putative miRNAs (miR-1246 and miR-5100) exhibited significant differences between SH3BP2 shRNA and scramble transfection in all GIST cells. In parallel, only miR-1246 and 5100 overexpression in GIST cell lines show a reduction of MITF level by western blot (Supplementary Figure S1). We restricted
Next, we validated these five miRNAs in various GIST cell lines. Quantitative real-time PCR was carried out in SH3BP2 silenced GIST-T1 ( Figure 3A), GIST882 ( Figure 3B), and GIST48 cells ( Figure 3C). Only two of the five putative miRNAs (miR-1246 and miR-5100) exhibited significant differences between SH3BP2 shRNA and scramble transfection in all GIST cells. In parallel, only miR-1246 and 5100 overexpression in GIST cell lines show a reduction of MITF level by western blot (Supplementary Figure S1). We restricted further studies to these two miRNAs. The miRNAs sequence location on the MITF-A, the highest isoform expressed, and ETV1 genes, are shown in Supplementary Figure S2. further studies to these two miRNAs. The miRNAs sequence location on the MITF-A, the highest isoform expressed, and ETV1 genes, , is shown in Supplementary Figure S2.

MiR-1246 and miR-5100 Target ETV1 and MITF, and Overexpression Significantly Affects Cell Proliferation
As mentioned above, these miRNAs putatively bind to MITF or ETV1 mRNA, so we overexpressed them in the imatinib-sensitive GIST-T1 and imatinib-resistant GIST-48 cell lines to check ETV1 and MITF protein levels. The overexpression of GFP-miR-1246 and GFP-miR-5100 efficiently causes the downregulation of MITF and ETV1 protein levels ( Figure 4A,B and Supplementary Material S1). Consistently with this, we reported diminished cell proliferation in GIST cells, Figure 4C. The levels of transfection were similar in all cases (Supplementary Figure S3).

MiR-1246 and miR-5100 Target ETV1 and MITF, and Overexpression Significantly Affects Cell Proliferation
As mentioned above, these miRNAs putatively bind to MITF or ETV1 mRNA, so we overexpressed them in the imatinib-sensitive GIST-T1 and imatinib-resistant GIST-48 cell lines to check ETV1 and MITF protein levels. The overexpression of GFP-miR-1246 and GFP-miR-5100 efficiently causes the downregulation of MITF and ETV1 protein levels ( Figure 4A, B and Supplementary Material 1). Consistently with this, we reported diminished cell proliferation in GIST cells, Figure 4C. The levels of transfection were similar in all cases (Supplementary Figure S3).

MiR-1246 and miR-5100 Promote Apoptosis by Caspases 3/7 in GIST Cells
To analyze how miRNAs affect cell proliferation, we performed a viability assay and measured caspase 3/7 activity on overexpressed miRNAs GIST cells. Our results show a decrease in cell viability that correlates with an increase in caspase 3/7 activity in both cell lines ( Figure 5B,C). Previous studies reported that miR-5100 induces apoptosis throughout caspase 3 protein activity [26], and miR-1246 increases apoptosis by promoting caspase 3 and caspase 7 activity [27]; these results are consistent with the anti-apoptotic protein BCL2 (MITF-dependent target) reduction after overexpression of miRNAs ( Figure 5A and Supplementary Material S1).

MiR-1246 and miR-5100 Promote Apoptosis by Caspases 3/7 in GIST Cells
To analyze how miRNAs affect cell proliferation, we performed a viability assay and measured caspase 3/7 activity on overexpressed miRNAs GIST cells. Our results show a decrease in cell viability that correlates with an increase in caspase 3/7 activity in both cell lines ( Figure 5B, C). Previous studies reported that miR-5100 induces apoptosis throughout caspase 3 protein activity [26], and miR-1246 increases apoptosis by promoting caspase 3 and caspase 7 activity [27]; these results are consistent with the anti-apoptotic protein BCL2 (MITF-dependent target) reduction after overexpression of miRNAs ( Figure  5A and Supplementary Material 1).

MiR-1246 and miR-5100 Affect Cell Cycle Progression
MITF regulates CDK2 in melanoma, which is critical for tumor cell growth [7,28]. We further analyzed whether CDK2 was altered after miRNA overexpression. MITF reduction was accompanied by decreased CDK2 levels in GIST-T1 and GIST 48 ( Figure 6A,B and Supplementary Material S1). The overexpression of both miRNAs had different consequences in the cell cycle in both cell lines. GIST-T1 ( Figure 6C) overexpression induced a substantial increase in the G2 phase, while in GIST 48 ( Figure 6D), there is an accumulation in the S phase. Altogether, these results indicate that these miRNAs may regulate MITF-dependent targets and cell cycle progression.

MITF Overexpression Significantly Restores Cell Proliferation
Next, we assessed the specificity of the effect of MITF on the proliferation of miRNAtreated GIST cells. For that purpose, after three days of miRNA transduction (when cells were still viable), cells were transfected with MITF-GFP or GFP. Seven days after miRNA transduction, MITF levels and cell proliferation were assessed. Our data show that MITF reconstitution is detected by western blot (Figure 7A and Supplementary Material S1) and significantly increases cell proliferation ( Figure 7B). were still viable), cells were transfected with MITF-GFP or GFP. Seven days after miRNA transduction, MITF levels and cell proliferation were assessed. Our data show that MITF reconstitution is detected by western blot (Figure 7A and Supplementary Material 1) and significantly increases cell proliferation ( Figure 7B).

LNA Treatment (Anti-miR-1246 Anti-miR-5100) Was Not Adequate to Revert the Apoptotic Phenotype in GIST-T1 SH3BP2 Silenced Cell
To determine whether miR-1246 and miR-5100 are the main ones responsible for the SH3BP2 silencing apoptotic phenotype, we analyzed the effect of LNA (Locked nucleic acid) miRNA inhibitors treatment on SH3BP2 silenced cells. LNA miRNA inhibitors are antisense oligonucleotides with perfect sequences complementary to their target miRNA that prevent miRNA hybridization with its regular cellular interaction partners. LNAs are taken up naturally by cells by a process known as gymnosis. We checked if LNA treatment reverted the apoptotic phenotype of SH3BP2 silenced cells. For that purpose, GIST-T1 cells were treated with LNA against miR-1246 and miR-5100, and afterward, cells were transduced with lentiviral particles shRNA-SH3BP2. Effective gymnosis was measured by FAN fluorescence microscopy each 24 h in treated cells. Our data show that LNAs treat- To determine whether miR-1246 and miR-5100 are the main ones responsible for the SH3BP2 silencing apoptotic phenotype, we analyzed the effect of LNA (Locked nucleic acid) miRNA inhibitors treatment on SH3BP2 silenced cells. LNA miRNA inhibitors are antisense oligonucleotides with perfect sequences complementary to their target miRNA that prevent miRNA hybridization with its regular cellular interaction partners. LNAs are taken up naturally by cells by a process known as gymnosis. We checked if LNA treatment reverted the apoptotic phenotype of SH3BP2 silenced cells. For that purpose, GIST-T1 cells were treated with LNA against miR-1246 and miR-5100, and afterward, cells were transduced with lentiviral particles shRNA-SH3BP2. Effective gymnosis was measured by FAN fluorescence microscopy each 24 h in treated cells. Our data show that LNAs treatment cannot block the apoptotic phenotype promoted by SH3BP2 silencing (Supplementary Figure S4). These results suggest that SH3BP2 action on apoptotic phenotype goes beyond miR-1246 and miR-5100.

Discussion
GISTs can be successfully treated with imatinib or other TKIs [1,29]. However, the necessity for new therapeutical approaches arose due to clinical resistance. We previously reported that silencing of SH3BP2 leads to a reduction of KIT expression at both mRNA and protein levels, as well as MITF at the protein level, resulting in a decrease in GIST tumor growth in vitro and in vivo [5]. In the same study, overexpression of MITF significantly reverses the apoptotic phenotype produced by SH3BP2 silencing, suggesting the involvement of this transcription factor in the regulatory mechanism in which SH3BP2 levels are critical. SH3BP2 silencing did not alter MITF mRNA levels but protein levels, suggesting a post-transcriptional mechanism. A miRNA microarray was performed in SH3BP2-silenced GIST882 and GIST48 cell lines (imatinib-sensitive and resistant cells) compared to non-silenced cells to get insights into the KIT-SH3BP2-MITF pathway. This microarray showed a different miRNA pattern when SH3BP2 was silenced. In parallel, we found that SH3BP2 silencing also targets ETV1, a master of ICC-transcription factor whose regulation is dependent on KIT signaling and is directly involved in the tumorigenic phenotype [14,19]. In this study, from the top miRNAs that were up-regulated, we focused on those that putatively target MITF and ETV1. After database analysis and cell validation, the miRNAs: miR-1246 and miR-5100 were selected for further studies. Overexpression assays showed that these miRNAs targeted MITF and ETV1 in GIST48 and GIST-T1. Consequently, the decrease in the levels of these transcription factors leads to a reduction in cell survival.
In this context, miR-1246 has been described as a tumor suppressor miRNA in prostate cancer, as authors showed that miR-1246 overexpression led to the inhibition of xenograft tumor growth over time [30]. They propose the exosomal-mediated release of miR-1246 to serum from tumor cells to evade its tumor suppressor role. Moreover, they suggest exosomal miR-1246 as a good biomarker to discern between benign of aggressive prostate cancer. Interestingly, exosomal miR-1246 has been proposed as a biomarker in gastric cancer (GC), and bioinformatics analysis revealed it as a tumor suppressor in GC [31]. Moreover, miR-1246, which can be induced by tumor suppressor p53, has been described as a tumor suppressor due to its capacity to reduce DYRK1A (a Down syndrome-associate kinase) levels, leading to the nuclear retention of NFATc1 and the induction of apoptosis [32]. Additionally, miR-1246 was downregulated in thyroid cancer, and the overexpression of miR-1246 affects PI3K/AKT pathway by regulating phosphoinositide 3-kinase adapter protein1 (PIK3AP1), resulting in less cell proliferation, diminished migration, and increasing apoptosis [33]. Furthermore, miR-1246 mediates LPS-induced pulmonary endothelial cell apoptosis in vitro and acute lung injury (ALI) in mouse models, which are at least partly attributed to the suppression of angiotensin-converting enzyme 2 (ACE2) [27]. In addition, miRNA-1246 mediates ALI-induced lung inflammation and apoptosis via the NF-κB activation and Wnt/β-catenin suppression [34]. Additionally, miRNA-1246 attenuates renal cell carcinoma's proliferative and migratory abilities by downregulating CXCR4 [35]. Nonetheless, the oncogenic role of miR-1246 has been reported in melanoma by conferring resistance to BRAF inhibitors [36] or enhancing migration and invasion through the adhesion molecule CADM1 in hepatocellular cancer [37].
Regarding miR-5100 activity as a tumor suppressor, our results follow Chijiiwa et al. [38]. The authors show that miR-5100 decreases the aggressiveness of the pancreatic cancer tumor models through the inhibition of PODXL, which promotes anti-adhesive and migratory characteristics of various cancers, and high levels of PODXL correlates with poor prognosis in many of them. Moreover, miR-5100 can increase the apoptosis level of gastric cancer cells and inhibit autophagy by targeting CAAP1 (conserved anti-apoptotic protein 1 or caspase activity and apoptosis inhibitor 1) [26].
However, miR-1246 and miR-5100 have been reported as oncogenic miRNAs in lung cancer [39]. One explanation for these contradictory results could be that miRNAs may vary their affinity to target mRNA depending on the cell lines, the pool of miRNAs that they could be cooperating, and the secondary structures in the 3 UTR of the target mRNA, which can affect the binding of a miRNA [40]. In conclusion, many factors could interfere with the miRNA functional effect in other cell lines.
The proapoptotic role of miR-1246 and miR-5100 in GIST cell lines could result from their ability to affect the cell cycle and regulate cell apoptosis. These actions can be related to a MITF reduction since BCL2 and CDK2 are MITF-dependent targets [7,[41][42][43]. BCL2 is found in most GIST patients [19] and correlates with a poor prognosis before imatinib treatment [44]. These miRNAs also induce cell cycle arrest in a cell line-dependent manner. CDK2 has been reported to regulate both G1/S and G2/M transitions. [45]. As previously noted, high double-negative CDK2-expressing cells were arrested in the mid-S phase. In contrast, low double negative CDK2 expressing cells progressed through early and mid-S phases but were still arrested in the late S/G2 phase [45], suggesting that the active CDK2 can be critical in the different phases. Recent research has shown that CDK2 deficiency slows colorectal cancer's S/G2 progression [46]. It would deserve further consideration to analyze the role and regulation of CDK2 in the different GIST cell lines.
The blockage of miR-5100 and 1246 using LNA did not reduce apoptotic effects due to SH3BP2 silencing, indicating that other miRNAs contribute to this phenotype. However, overexpression of MITF significantly restores cell survival after miR-5100 and 1246 transduction, suggesting that MITF is a crucial target for cell viability.
The role of MITF is well-known in melanoma [47], and recent studies suggest that MITF overexpression in kidney angiomyolipoma cells [48] and clear cell renal cell carcinoma (ccRCC) improve cell growth, proliferation, and invasion in vitro and in vivo [49]. Lately, we have described that the silencing of MITF results in decreased gastrointestinal stromal tumor cell viability in vitro and tumor growth in vivo [11].

Conclusions
Our results highlight the KIT-SH3BP2-MITF/ETV1 pathway for GIST cell survival and proliferation. Targeting ETV1 and MITF together will help break the positive feedback loop and indirectly target KIT independently of the mutations in the tyrosine kinase receptor.