The c-KIT Exon 11 Deletions Cause Resistant to Imatinib Treatment of GISTs through Induction of ULK1-SRC related Autophagy


 Background: Exon 11 deletion of c-KIT is the common mutation site with the worst prognosis of gastrointestinal stromal tumors (GISTs). Autophagy has shown to have both protumor and antitumor functions and protect GIST cells from imatinib (IM)-induced apoptosis. In this study, we aimed to explore whether c-KIT exon 11 deletions of GISTs drive refractory to IM treatment through regulation of autophagy.Methods: 251 GISTs were included in this study to determine the association of c-KIT exon 11 deletions with prognosis and IM resistance. Immunohistochemistry, immunofluorescence, Western blotting were used to assess the correlation between c-KIT exon 11 deletions and autophagy and the underlying mechanism. We also studied the antitumor effect of IM in combination with the autophagy inhibitor chloroquine (CQ) in GIST62 cell derived xenograft mice model.Results: GISTs with c-KIT exon 11 deletions had worse survival and IM-resistant phenotype and exhibited higher autophagy levels as compared with GISTs harboring wild type or other c-KIT mutations. In addition, LC3-II expression strongly increased in GIST cell lines carrying c-KIT exon 11 deletions. Treated with autophagy inhibitors or knockdown of the autophagy-related gene ATG5 effectively inhibit GIST cells viability. Co-treatment with IM and CQ significantly suppressed tumor growth as compared with either agent alone, suggesting that autophagy protect GIST cells from IM-induced cell death both in vitro and in vivo. Furthermore, c-KIT exon 11 deletions can induce autophagy by activating SRC and ULK1. SRC inhibitor can block this action.Conclusion: Our study clearly demonstrates that KIT exon 11 deletions induce ULK1-SRC related autophagy is one of reasons for imatinib resistance clinically. Combination with autophagy inhibitor shows significant promise in treatment for imatinib refractory GISTs.

effectively inhibit GIST cells viability. Co-treatment with IM and CQ significantly suppressed tumor growth as compared with either agent alone, suggesting that autophagy protect GIST cells from IMinduced cell death both in vitro and in vivo. Furthermore, c-KIT exon 11 deletions can induce autophagy by activating SRC and ULK1. SRC inhibitor can block this action.

Conclusion:
Our study clearly demonstrates that KIT exon 11 deletions induce ULK1-SRC related autophagy is one of reasons for imatinib resistance clinically. Combination with autophagy inhibitor shows significant promise in treatment for imatinib refractory GISTs.

Background
Gastrointestinal stromal tumor (GIST), originating from the lineage of interstitial cells of Cajal (ICC) [1], is the most common mesenchymal neoplasm of the gastrointestinal tract with the annual incidence rate of approximately 10-15 cases per million population [2]. Currently, surgery is the 3 mainstay of treatment for patients with resectable GIST; however, approximately 50% of GIST patients develop recurrence or metastasis after complete resection of the primary tumor [3]. Imatinib (IM), a selective tyrosine kinase inhibitor is the standard of care for patients with unresectable or metastatic GISTs. It has led to significant improvements in survival of GIST patients. However, the majority of patients inevitably acquire drug resistance, highlighting the urgent need for identification and targeting of the mechanisms underlying resistance to IM.
GISTs are typically immunoreactive for KIT, also known as CD117. This protein is observed in over 95% of GISTs, and is the most specific and sensitive marker in differentiating GISTs from other mesenchymal tumors in the GI tract [4]. KIT is a 145 kD transmembrane glycoprotein and serves as a receptor for stem cell factor (SCF). Binding of SCF activates KIT downstream pathways such as MAPK, AKT, SRC, and STAT3, consequently regulating a variety of cell functions, including proliferation, survival, apoptosis, and motility [5,6]. About 75-80% of sporadic GISTs have c-KIT mutations, which have been considered a major driving force behind tumor development [7]. The most common mutation in KIT affects the juxtamembrane region encoded by exon 11. This region, just inside the cell membrane, is a helical domain that functions to inhibit KIT kinase activity in the absence of ligands [8]. Mutagenesis studies on c-KIT have demonstrated that mutations of this domain disrupt this function, allowing ligand-independent receptor dimerization and constitutional activation. Several reports have shown that c-KIT exon 11 deletions confer worse prognosis and predispose GISTs to metastasis [9]. Codons 557 and 558 are the most common sites for deletion in c-KIT exon 11 and c-KIT exon 11 deletions involving codons 557-558 are highly associated with liver metastasis of GISTs [10].
Autophagy is an evolutionarily conserved pathway of cellular catabolic degradation in response to starvation or stress. It can degrade and recycle long-lived proteins and organelles, serving as protein quality controller [11]. Based on its functions to keep cellular homeostasis, autophagy can act as a crucial adaption pathway that can promote tumorigenesis and survival of cancer cells under nutrient defeat. Accumulating evidence has indicated that autophagy has protumor functions. For examples, autophagy can enhance the tolerance of stress, which maintains tumor cell survival [12]. Cancer cells 4 have high metabolic demands and exposure to metabolic stress is shown to impair survival in autophagy-deficient cells in vitro. Hypoxia and nutrient deprivation can activate the autophagy for recycling of ATP and to maintain cellular biosynthesis and survival [13]. In genetically engineered mouse models driven by Kras or Braf mutation, depletion of ATG5 and ATG7 resulted in diminished tumor burden [14]. In addition, autophagy can also maintain activity of pancreatic cancer stem cells that contribute to chemoresistance of cancer cells, and blockade of autophagy potentiates antitumor effect of gemcitabine [15], suggesting that autophagy can function as a tumor promoter. Previous research has shown that inhibition of autophagy increases drug sensitivity in GIST cells with c-KIT exon 11 deletions [16], raising that autophagy acts as a survival mechanism. Therefore, in this study, we were interested in determining whether c-KIT exon 11 deletions cause resistant to imatinib treatment in GISTs by up-regulating autophagy.

Cell lines
The human GIST cell lines, GIST62 with KIT-negative and GIST48 with a homozygous c-KIT exon 11 Transfection GIST62 cells in 6-well plates were transfected with 2 µg of plasmids, including pcDNA3.1b-KIT exon 11 △557-558, pcDNA3.1b-KIT wild type, pcDNA3.1b-KIT exon11 V560D, and empty pcDNA3.1b vector as control using lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. After transfection for 48 hours, we harvested cells for further study. We also used geneticin (800 µg/mL) for selection to establish stable cell lines.

Protein Extraction
Cells in culture dishes were washed with ice-cold PBS and were collected to microcentrifuge tube using cell scraper. The cells were lysed with ice-cold RIPA cell lysis buffer (Millipore Corporation) containing protease inhibitor cocktail (Millipore Corporation) and phosphatase inhibitor cocktail (Sigma-Aldrich). The cell lysate was centrifuged at 13,000 rpm at 4 °C for 30 minutes, and then the supernatant was collected to a new 1.5 mL microcentrifuge tube. The protein concentration of the samples was measured by BCA protein assay (Thermo Fisher Scientific), and all samples were stored at -80 °C.

Western Blotting
Equal amounts of protein in 1 × SDS loading buffer were heated at 95 °C for 10 minutes. Samples were separated by 8 or 10% SDS-PAGE and transferred to Immobilon-P membranes (Millipore Corporation) at 100V for 100 minutes. The membrane was then blocked with 5% milk diluted in 0.05% TBST for 1 hour at room temperature. After blocking, membrane was probed by primary antibodies diluted in 0.05% TBST overnight at 4 °C. Next day, the membrane was washed 10 minutes for three times with 0.05% TBST followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 hour at room temperature. The membrane was washed 10 minutes for three times with 0.05% TBST. The blots signal was developed by using Immobilon Western Chemiluminescent HRP Substrate (Millipore Corporation) according to the manufacturer's instruction and captured by Biospectrum Imaging System (UVP).

Immunohistochemical (ihc) Analysis
The formalin-fixed paraffin embedded tissues were cut into 5 µm-thick sections. Antigen retrieval was performed by autoclaving the sections at 121 °C for 10 minutes in sodium citrate buffer (10 mM, pH 6.0). After blocking with the blocking buffer (Thermo Fisher Scientific) at room temperature for 30 minutes, the sections were incubated with specific primary antibodies at 4 °C overnight and then with biotinylated secondary antibodies at room temperature for 30 minutes. The immunoreaction was 6 visualized using a DAB chromogen system (DAKO). The cell nuclei were stained with hematoxylin.

Immunofluorescence (if) Staining
Cells (3 × 10 4 cells/well) were seeded in an 8-well chamber slide (Millipore Corporation). After 24 hours of growth, cells were fixed in 4% paraformaldehyde for 10 minutes and permeabilized in 0.05% Triton X-100 for 15 minutes. Cells were then blocked in Super Blocking Buffer (Pierce) for 1 hour and incubated with primary antibodies overnight. Next day, cells were incubated with fluorochromeconjugated secondary antibodies for 1 hour at room temperature. Nuclei were counter-stained with DAPI. All the images were visualized using a fluorescence microscope (BX51, OLYMPUS).

Animal Model
Seven weeks old male NOD/SCID mice were subcutaneously injected with GIST62 V cells or GIST62 Δ cells (4 × 10 6 cells in 50 µL serum free medium with 50µL matrix gel). The tumor volume was calculated using the formula of (length × width 2 )/2. When tumor volume was > 100 mm 3  Values are expressed as means ± SEM. Differences between groups were determined by Student's t test and ANOVA. The survival probability was calculated with the Kaplan-Meier method. P value less than 0.05 was considered to be statistically significant (*p < 0.05, **p < 0.01, ***p < 0.001). with c-KIT exon 11 deletions had more metastasis rate, recurrence rate (Table 1). Furthermore, those patients also had the worst disease-free survival (Fig. 1A). We further compared the responses to IM of the four groups. GIST patients with c-KIT exon 11 deletions had the lower response rate than those in the other three groups (Fig. 1B). These results indicate that c-KIT exon 11 deletions are associated with poor prognosis in GISTs.

c-KIT
GISTs with c-KIT exon 11 deletions exhibit high levels of autophagy Autophagy has been widely implicated in mechanisms governing tumor development and chemoresistance [18]. In GIST, autophagy can protect tumor cells from IM-induced apoptosis [16]. To elucidate whether autophagy is involved in c-KIT exon 11 deletions-induced malignant properties and drug resistance of GIST cells, the correlation between c-KIT exon 11 deletions and autophagy was evaluated. We compared the autophagosomal marker LC3-II expression in human GIST specimens with wild-type c-KIT, c-KIT exon 11 deletion, and c-KIT exon 11 V560D mutation. We found GISTs harboring c-KIT exon 11 deletion had the highest expression of LC3-II ( Fig. 2A). IHC staining for LC3-II also showed that GISTs harboring c-KIT exon 11 deletion expressed a higher level of LC3 puncta than that with wild-type c-KIT (Fig. 2B). Quantification of LC3 puncta levels in GIST specimens revealed a positive correlation between c-KIT exon 11 deletion and autophagy (Fig. 2C). To further confirm the positive correlation between autophagy and c-KIT exon 11 deletions, we established stable 8 transfected cell lines from KIT-negative GIST62 cells: vector control GIST62V cells and GIST62D cells that carry c-KIT exon 11 deletion. We observed that LC3-II expression was higher in GIST62D cells and GIST-T1 cells but lower in GIST62V cells and GIST48 cells (Fig. 2D). IF staining for LC3 showed a higher level LC3 puncta in GIST62D cells than in GIST62V cells (Fig. 2E). Taking together, we prove that c-KIT exon 11 deletions upregulate autophagy in GIST cells.
Autophagy mediates c-KIT exon 11 deletions-induced GIST cell growth We next determined the effect of autophagy upregulation in GIST cells with c-KIT exon 11 deletions.
High expression of LC3-II and reduced expression of the autophagy marker SQSTM1/p62 indicating autophagy activation [19] was accompanied by increased expression of the cell proliferation marker Ki-67 in GIST62D cells but not in GIST62V cells (Fig. 3A), while there was no difference in apoptosis between GIST62D cells and GIST62V cells (supplemental Fig. 1). As expected, GIST62D cells also showed greater cell growth potential and mitotic activity than GIST62V cells did (Fig. 3B, 3C, and supplemental Fig 2). When autophagy was blocked by various pharmacological inhibitors including Bafilomycin A1 (Baf A1), chloroquine (CQ), and 3-methyladenine (3-MA), the viability of GIST62D cells was decreased in a dose-dependent fashion (Fig. 3D-F). A similar result also obtained after the core autophagy gene ATG5 inactivated by RNA interference in GIST62D cells (Fig. 3G). These data indicate that autophagy is required for c-KIT exon 11 deletions-mediated cell growth in GISTs.

Inhibition of autophagy reduces tumor growth and enhances IM cytotoxicity
To investigate the in vivo relevance of the in vitro observations, we used a xenograft model derived from GIST62 cells. We first compared the tumor growth rate between GIST62V cells and GIST62D cells by subqutaneous injection of the cells into NOD/SCID mice. The growth rate of GIST62Δ tumors was significantly higher than that of GIST62V tumors (Fig. 4A). Six weeks after injection, mice were sacrified and tumors were collected. The size and the weight were greater for GIST62Δ tumors as compared with GIST62V tumors (Fig. 4B and 4C), indicating that c-KIT exon 11 deletion promotes GIST tumor growth. Western blotting using collected tumors confirmed that LC3-II was highly expressed in GIST62Δ tumors but not in GIST62V tumors (Fig. 4D). We next assessed the effect of autophagy blockade on tumor growth and the therapeutic efficacy of IM. GIST62D cells were subqutaneously 9 inoculated into NOD/SCID mice to form tumors. When the volume reached around 100mm 3 , mice were randomized to receive a 2-week course of twice-weekly treatment with vehicle, CQ, IM, or in combination. Two weeks after cell inoculation, the mean volume of tumors treated with vehicle was about 450 mm 3 . Compared to CQ or IM alone treatment that decreased the mean tumor volume to about 300 mm 3 , CQ combined with IM treatment could more effectively reduce the tumor volume to about 120 mm 3 (Fig. 4E). Similar results were also seen in tumor size and tumor weight measured at the end point of the experiment (Fig. 4F and 4G). These in vivo results indicate that inhibiting autophagy can not only retard tumor growth but also enhance the susceptibility of GIST cells to IM.

ULK1 and SRC are involved in c-KIT exon 11 deletions-induced autophagy
To decipher the mechanism by which c-KIT exon 11 deletions induce autophagy in GIST cells, we first identified the autophagy-related proteins whose expression altered. We found that c-KIT exon 11 deletion mutation did not affect the expression of ATG5, ATG7, and Beclin 1 (Fig. 5A). However, ULK1 Serine-556 phosphorylation that contributes to autophagy activation increased, while ULK1 Serine-758 phosphorylation that deactivates ULK1 decreased in GIST62D cells as compared with GIST62V cells (Fig. 5A). The nonreceptor tyrosine kinase SRC is one of the major signaling molecule located downstream of KIT and has been reported to be involved in autophagy activation [20][21][22]. A previous study revealed that phosphorylation of mATG9 at Tyr8 by SRC functionally cooperates with Ser14 phosphorylation by ULK1 to regulate mATG9 trafficking and redistribution for autophagy initiation [23]. We therefore observed the activation of SCR in GIST cells with different c-KIT mutations by analyzing its phosphorylation and evaluated the correlation between SRC phosphorylation and autophagy. Enhanced expression of phosphorylated SRC (pSRC) and LC3-II observed in GIST62D cells and GIST T1 cells (Fig. 5B). GIST62D cells expressed the highest levels of pSCR and LC3-II than the other GIST62 transfectants, including GIST62V cells, GIST62 WT cells that overexpress wile-type KIT, and GIST62 V560D cells that carry c-KIT exon 11 V560D mutation (Fig. 5C). Treatment with a specific SRC kinase inhibitor bosutinib in GIST62D cells caused decreased SRC phosphorylation with a concomitant suppression in LC3-II expression. These data suggest that c-KIT exon 11 deletions 10 upregulate autophagy by activating ULK1 and SRC.

Discussion
Autophagy is not only a critical catabolic process essential for cellular homeostasis but has also been linked to cancer, playing a dual role in tumor cell survival and death [24]. In GISTs, c-KIT exon 11 deletion associates with worse prognosis and confers a more aggressive phenotype [9, 10]. However, the correlation between c-KIT exon 11 deletion and autophagy in GIST progression is not clear. In this study, we observed that GISTs with c-KIT exon 11 deletion mutations had a higher rate of resistance Our results show that SRC was highly phosphorylated in GIST cells with c-KIT exon 11 deletion but not in GIST cells with c-KIT exon 11 V560D or wild type c-KIT. The phosphorylation of SRC might be independent of KIT autophosphorylation but affected by different mutations-induced receptor conformational changes that can selectively recruit different intracellular signaling pathways [25,26].
Furthermore, some studies have found that autophagy has an important role in SRC-mediated oncogenic activities of cancer cells. For example, autophagy promotes focal adhesion disassembly and cell motility of metastatic breast cancer cells through the paxillin-LC3 interaction which is regulated by SRC [22]. Activation of the SRC/STAT3 pathway can induce heme oxygenase-1 expression by promoting autophagy to protect breast cancer cells from doxorubicin-induced cytotoxicity [27]. Consistently, in this study, we found that SRC phosphorylation positively correlated with autophagy induction, and inhibition of SRC could reduce autophagy levels, suggesting that SRC can act as a positive regulator of autophagy induced by c-KIT exon 11 deletion mutations in GIST cells.
Many tyrosine kinase inhibitors have been developed to target KIT activity to treat GIST, such as IM, sunitinib, regorafenib, sorafenib, nilotinib, ponatinib, masitinib, dasatinib, and dovitinib [28]. IM is the first line drug with significant improvement of GIST patients' survival, but some GIST patients with kit exon 11 deletion mutations show refractory to IM treatment soon. These patients should receive second line drug treatment by sunitinib. Our study proved combination of IM and autophagy inhibitor can inhibit tumor growth. Furthermore, SRC inhibitor can reduce the autophagy activity induced by kit exon 11 deletion, which meant combination with IM and SRC inhibitor might be used for those patients refractory to IM treatment. Gupta A et al also had proposed similar strategy to enhance GIST cytotoxicity and to diminish both cellular quiescence and acquired resistance in GIST patients [16].
Recently, the Hsp90AA1 inhibitor 17-AAG is another drug to inhibit KIT pathway, which blocks heat shock protein and thus promotes autophagy to degrade KIT protein [29]. Since 17-AAG exert a different mechanism to inhibit KIT, whether the combination of autophagy inhibitors with this drug has synergistic effect in GIST treatment need to be studied.
Many studies show that knockdown of autophagy-associated genes can inhibit cancer cells growth or increase anticancer efficiency. In our results, we found that inhibition of autophagy could inhibit the growth of GIST62Δ cells. Inhibition of autophagy may be a potential strategy to treat GISTs.
Additionally, cotreament with autophagy inhibitors becomes a more common option in cancer treatment. Effects of the combination of chemotherapy and the autophagy inhibitor CQ in many cancers have been assessed in pre-clinical studies [14,30,31]. These studies prove that inhibition of autophagy can serve as reliable method to treat cancers.   assay. *** P < 0.001 versus GIST62 V cells, unpaired t-test. D, E, and F, GIST62 cells were treated with the autophagy inhibitors Baf A1, CQ, and 3-MA at indicated doses for 24 hours.

Conclusion
The cell viability was measured by MTT assay. G, after knockdown of ATG5, the viability of GIST62 cells was measured by MTT assay. *** P < 0.001 versus control cells, unpaired ttest.
24 Figure 4 Autophagy inhibition reduces tumor growth of GIST cells harboring c-KIT exon 11 deletion and potentiates antitumor effect of IM. A, GIST62 V and GIST62 Δ cells were subcutaneously 25 inoculated into SCID mice. Tumor volume was calculated weekly. The graph depicts the growth curves of tumors. *** P < 0.001 versus GIST62 V cells, unpaired t-test. B, representative image of tumors dissected at the end of experiment from GIST62 V and GIST62 Δ cells is shown. C, tumor weight was measured at the end of the experiment. *** P < 0.001 versus GIST62 V cells, unpaired t-test. D, levels of autophagy in GIST62 V and GIST62 Δ tumors were confirmed by measuring LC3 II and p62 expression using western bloting. E, GIST62 Δ tumors were treated with imatinib (40mg/kg), chloroquine (CQ, 60mg/kg), and their combination. The tumor volumes were measured once three days after treatments. F, representative image shows GIST62 Δ tumors receiving different treatments that were dissected at the end of experiment. G, bar graph depicts tumor weight of different treatment groups measured at the end of the experiment. P 0.05; P 0.01; P 0.001, significant differences between groups, one-way ANOVA.
26 Western blotting using the indicated antibodies. E, schematic representation of the mechanism by which c-KIT exon 11 deletion induces autophagy in GIST cells.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.