Loss of Heterozygosity for KrasG12D Promotes Malignant Phenotype of Pancreatic Ductal Adenocarcinoma by Activating HIF-2α-c-Myc-Regulated Glutamine Metabolism

Loss of heterozygosity (LOH) for KRAS, in which a wild-type KRAS allele is progressively lost, promotes invasive and migratory abilities of pancreatic ductal adenocarcinoma (PDAC) cells and tissues. Moreover, the occurrence of KrasG12D-LOH activates nonclassical glutamine metabolism, which is related to the malignant behavior of PDAC cells. Herein, we aim to demonstrate the regulatory link between hypoxia-inducible factor-2α (HIF-2α) and glutamine metabolism that mediates malignant phenotypes in KrasG12D-LOH PDAC cells. HIF-2α-shRNA knockdown lentivirus transfection and metabolite analysis were performed in KrasG12D-LOH and KrasG12D cell lines, respectively. Cell proliferation, migration, and invasion were examined using Cell Counting Kit-8, colony formation, and Transwell assays. Cell cycle phase and apoptosis were determined using flow cytometry. Western blotting and real-time quantitative PCR were also performed. Additionally, a subcutaneous xenograft mouse model was established. LOH stimulated HIF-2α activity and transactivated c-Myc, which has a central regulatory effect on glutamine metabolism independent of hypoxia. Meanwhile, HIF-2α silencing repressed KrasG12D-LOH PDAC cell proliferation, invasion, and migration. HIF-2α knockdown inhibited glutamine uptake and GOT1 expression via a c-Myc-dependent pathway. Collectively, KrasG12D-LOH can activate HIF-2α to regulate c-Myc-mediated glutamine metabolism and promote malignant phenotypes. Moreover, targeting HIF-2α-c-Myc regulated nonclassical glutamine metabolism, providing a new therapeutic perspective for KrasG12D-LOH PDAC.


Introduction
Pancreatic cancer is one of the most common malignant gastrointestinal cancers and is characterized by late diagnosis, early metastasis, lack of a specific targeted treatment, and poor prognosis [1]. Moreover, the prognosis of pancreatic ductal adenocarcinoma (PDAC) has not significantly improved over the past few decades; it has a five-year survival rate of only 6%. Moreover, the specific pathogenesis of pancreatic cancer remains unclear [2,3]. Patients with PDAC usually have no obvious symptoms and are diagnosed at an advanced stage of the disease, which is the primary cause of the associated low survival rate [4,5]. Diagnosed patients are eligible for surgery, chemotherapy, radiotherapy, and immunotherapy; however, there is currently a lack of effective targeted therapies [6]. As such, research has focused on the molecular mechanisms of pancreatic cancer, which may be of great significance for accurate diagnosis, efficient treatment, and improved prognosis.
Current studies indicate that pancreatic cancer is mainly regulated by tumor-suppressor gene inactivation and oncogene activation [7]. The KRAS gene is one of the most frequently mutated oncogenes in many cancers, and can be found in nearly all PDAC cases [8][9][10].
Current studies indicate that pancreatic cancer is mainly regulated by tum pressor gene inactivation and oncogene activation [7]. The KRAS gene is one of th frequently mutated oncogenes in many cancers, and can be found in nearly all cases [8][9][10]. KRAS belongs to a class of genes that encodes guanosine triphosphat regulates downstream signaling pathways through growth factor receptors [11]. I tion to KRAS, TP53, CDKN2A, and SMAD4 tumor suppressor genes are significan tated in pancreatic cancer [12][13][14]. Moreover, loss of heterozygosity (LOH), whic to the transformation of a heterozygous allele to a homozygous state, is a commo genic mutation associated with the occurrence and development of various [15,16]. In our previous experiment, we found that Kras G12D -LOH exists in transgen and human pancreatic cancer cells, wherein heterozygous PDAC cells lose their wi KRAS allele through mutation (Figure 1a). Furthermore, Kras G12D -LOH PDAC cells stronger ability to proliferate and invade than PDAC cells without LOH [17][18][19]. T cific downstream-signaling mechanism of Kras G12D -LOH responsible for its high p ation and invasion remains unclear and requires further study.  Warburg defined tumors as metabolic diseases in the 1920s. It was suggested that even under aerobic conditions, glucose does not participate in the tricarboxylic acid cycle but instead obtains adenosine triphosphate through lactic acid formation via aerobic glycolysis [20]. Glutamine (Gln) is a nonessential amino acid abundant in plasma and contributes to cancer cell genesis, proliferation, and metastasis [21,22]. Gln metabolism can provide carbon and nitrogen sources for tumor cells and participates in the synthesis of biological macromolecules, including lipids, proteins, and other amino acids, which play an important role in the maintenance of cell redox homeostasis, activation of cell-signaling pathways, and regulation of glucose and lipid metabolism [23]. Previous studies have shown that PDAC cells are heavily dependent on Gln metabolism for their growth needs and highly sensitive to Gln deprivation [24,25]. Moreover, unlike other tumor cells entering the tricarboxylic acid cycle by glutamate dehydrogenase-mediated deamination, glutamate is converted to aspartate by aspartate aminotransferase (GOT1) in PDAC cells. Glutamatederived aspartate is transferred to the cytoplasm to generate oxaloacetate, which in turn produces malate by malate dehydrogenase (MDH1) and pyruvate by malic enzyme (ME1). The reduced nicotinamide adenine dinucleotide phosphate (NADPH) that is generated to restore oxidized glutathione during this pathway is vital for maintaining the intracellular redox balance. Pancreatic tumor cells metabolize Gln via a nonclassical pathway, which is mediated by KRAS [26]. In our previous study, LOH for Kras G12D stimulated glutaminolysis in PDAC cells [19]. Kras G12D -LOH can increase protein and mRNA levels of related molecular markers, including GLS1, GOT1, GOT2, MDH1, and ME1, under both normoxic and hypoxic conditions [19]. However, the specific mechanism underlying Kras G12D -LOH and Gln metabolism remains unclear.
Hypoxia-inducible factors (HIFs) serve as the main regulators of a series of molecular biological changes in tumor cells, including angiogenesis, erythropoiesis, glycolysis, and subsequent tumor progression [27]. HIF-α can be regulated by oxygen and consists of three subunits: HIF-1α, HIF-2α, and HIF-3α [28]. Current studies have shown that HIF-2α correlates with normal cell proliferation, differentiation, energy metabolism, erythropoiesis, stem cell maintenance, inflammation, tumor proliferation, metastasis, and angiogenesis [29]. HIF-2α is hydroxylated by prolyl hydroxylase and degrades rapidly under aerobic conditions after binding to von Hippel-Lindau tumor suppressor protein (pVHL) [30]. The degradation of HIF-α is inhibited in response to hypoxia, and HIF-2α can initiate transcription of the target gene by binding to the hypoxia response elements of target genes to form a transcription-initiation complex [31,32]. Meanwhile, intracellular lactate signaling, which imitates a response to hypoxia, reportedly promotes Gln metabolism through the HIF-2α-signaling pathway in oxidative cancer cells in a manner dependent on c-Myc activation [33]. Similarly, studies have shown that HIF-2α affects PDAC cell growth by regulating Gln metabolism under prolonged hypoxia [34].
Our aim was to determine the role of HIF-2α in the Kras G12D -LOH-mediated malignant phenotype and examine the relationship between Kras G12D -LOH and HIF-2α-regulated Gln metabolism that occurs independently of hypoxia.

Kras G12D -LOH Elevated HIF-2α Expression in PDAC Cells at Both Protein and mRNA Levels
To confirm the regulatory role of HIF-2α in Gln metabolism, we investigated HIF-2α protein and mRNA expression under normoxic conditions. Notably, both HIF-2α protein and mRNA expression were significantly elevated in Kras G12D -LOH PDAC cells compared with Kras G12D PDAC cells without LOH under normoxic conditions (p < 0.05, p < 0.01; Figure 1b,c). We used the Gene Expression Profiling Interactive analysis (GEPIA) database (http://gepia.cancer-pku.cn/ accessed on 16 April 2021) to analyze the HIF-2α mRNA level in pancreatic cancer tissues and normal pancreatic tissues and found that HIF-2α mRNA levels were significantly higher in pancreatic cancer tissues than in normal pancreatic tissues (p < 0.05; Figure 1d).

Kras G12D -LOH Triggered HIF-2α-Dependent c-Myc Activation
Mitochondrial glutaminase and Gln metabolism can be enhanced by c-Myc via suppression of miR-23 [35]. However, a specific link between LOH and c-Myc has not yet been reported. Notably, c-Myc protein and mRNA levels were significantly higher in Kras G12D -LOH cells than in Kras G12D cells without LOH (p < 0.01; Figure 2a,b). Kras G12D -LOH promoted c-Myc activity, and HIF-2α knockdown repressed c-Myc expression both at the protein and mRNA levels (p < 0.01; Figure 2c,d). The LOH-induced elevation of c-Myc expression was blocked by HIF-2α silencing. According to these findings, LOH activated HIF-2α, which further stabilized c-Myc activity via a signaling pathway independent of hypoxia.
cancer tissues and normal pancreatic tissues and found that HIF-2α mRNA levels wer nificantly higher in pancreatic cancer tissues than in normal pancreatic tissues (p < 0.05 ure 1d).

Kras G12D -LOH Triggered HIF-2α-Dependent c-Myc Activation
Mitochondrial glutaminase and Gln metabolism can be enhanced by c-Myc via pression of miR-23 [35]. However, a specific link between LOH and c-Myc has not yet reported. Notably, c-Myc protein and mRNA levels were significantly higher in Kra LOH cells than in Kras G12D cells without LOH (p < 0.01; Figure 2a,b). Kras G12D -LOH prom c-Myc activity, and HIF-2α knockdown repressed c-Myc expression both at the protein mRNA levels (p < 0.01; Figure 2c,d). The LOH-induced elevation of c-Myc expression blocked by HIF-2α silencing. According to these findings, LOH activated HIF-2α, w further stabilized c-Myc activity via a signaling pathway independent of hypoxia.

LOH-Activated HIF-2α-c-Myc Signaling Enhanced Noncanonical Gln Metabolism
To determine the role of HIF-2α and c-Myc in Gln metabolism, we transfected shRNA lentivirus targeted with c-Myc and shRNA targeted with HIF-2α into Kras G12D and Kras G12D -LOH cells, respectively. HIF-2α or c-Myc silencing repressed GOT1 protein and mRNA levels in Kras G12D -LOH cells (p < 0.05 and p < 0.05, respectively; Figure 3a,b). LOH-activated GOT1 expression in Kras G12D -LOH PDAC cells was independently repressed by sh-HIF-2α or sh-c-Myc compared with that in the shRNA-NC and control groups. HIF-2α or c-Myc knockdown significantly decreased Gln consumption in Kras G12D -LOH PDAC cells (p < 0.01 and p < 0.05; Figure 3c). Moreover, we observed a reduction in NADPH levels in sh-HIF-2αand sh-c-Myc Kras G12D -LOH PDAC cells; specifically, improved NADP + ratios and reactive oxygen species (ROS) levels were observed (p < 0.05, and p < 0.05, respectively; Figure 3d,e). These findings suggest that inhibition of HIF-2α attenuated nonclassical Gln metabolism in Kras G12D -LOH pancreatic cancer cells, possibly in a c-Myc-dependent manner.

HIF-2α Knockdown Suppressed Invasion and Migration of Kras G12D and Kras G12D -LOH PDAC Cells
According to the results of Transwell assay, HIF-2α silencing reduced the invasion of Kras G12D and Kras G12D -LOH cells (p < 0.01; Figure 5a). Consistently, the sh-HIF-2α group had a lower number of invaded Kras G12D and Kras G12D -LOH cells than the shRNA-NC and control groups (p < 0.01; Figure 5b). These findings demonstrate that HIF-2α silencing can inhibit the migration and invasion of Kras G12D and Kras G12D -LOH PDAC cells.

HIF-2α Knockdown Increased Cell Cycle Arrest and Apoptosis of Kras G12D and Kras G12D -LOH PDAC Cells
The cell cycle assay results showed that the proportion of cells in the S phase was significantly lower in sh-HIF-2α Kras G12D -LOH and sh-HIF-2α Kras G12D groups than in the control and shRNA-NC groups, whereas HIF-2α knockdown induced G1 phase arrest (p < 0.01; Figure 6a). In addition, the sh-HIF-2α groups of Kras G12D -LOH and Kras G12D cells had an increased apoptosis rate (p < 0.01; Figure 6b). These results confirmed that the proliferative capacity of Kras G12D and Kras G12D -LOH PDAC cells was inhibited following HIF-2α knockdown.

Role of HIF-2α in Kras G12D and Kras G12D -LOH PDAC Cells In Vivo
We further demonstrated the exact role of HIF-2α in malignant behavior and nonclassical Gln metabolism in vivo. After subcutaneous injection of HIF-2α-targeted shRNAinfected cells into nude mice, we observed that the weight of the removed tumor in the sh-HIF-2α group was significantly lower than that in the shRNA-NC group (p < 0.05; Figure 7a,b). In addition, immunohistochemistry revealed that HIF-2α silencing decreased the number of Ki67-positive PDAC cells compared with that in the shRNA-NC group (p < 0.01; Figure 7c). Consistent with the in vitro results, we noted that HIF-2α silencing inhibited LOH-induced c-Myc and GOT1 expression at both mRNA and protein levels (p < 0.01; Figure 7d,e). Overall, these results demonstrate that HIF-2α silencing may repress the malignant phenotype of Kras G12D -LOH PDAC cells in vivo, and HIF-2α-c-Myc signaling may regulate nonclassical Gln metabolism.

Discussion
This study mainly explored the role of HIF-2α in the Kras G12D -LOH cell-mediated malignant phenotype of pancreatic cancer. In particular, we investigated the downstream targets of Kras G12D -LOH and found that the protein and mRNA levels of HIF-2α in Kras G12D -LOH PDAC cells were significantly higher than those in Kras G12D PDAC cells. These results suggest that HIF-2α plays a key role in regulating the malignant phenotype and Gln metabolism in Kras G12D -LOH pancreatic cancer.
Hypoxia is among the fundamental factors regulating the HIF-signaling pathway [36]. Some studies have found that malignant tumor cells can promote the activation of the HIFsignaling pathway by inducing HIF-2α mRNA transcription, maintaining protein stability and regulating upstream and downstream target genes in various pathways [37,38]. In addition, mutation of VHL or inactivation of prolyl hydroxylase can promote HIF-2α expression [39]. Prolyl hydroxylation can be inhibited by ROS, nitric oxide, and specific oncogenes, such as activated RAS and v-Src [40,41]. The PI3K/AKT/mammalian target of rapamycin (mTOR) pathway activates HIF signaling by increasing HIF-α subunit protein translation [42]. Similarly, in the present study, we found that HIF-2α expression was significantly higher in Kras G12D -LOH pancreatic cancer cells than in Kras G12D pancreatic cancer cells under normoxia. Furthermore, the occurrence of LOH in Kras G12D PDAC cells was found to promote HIF-2α expression. The LOH for Kras G12D stabilizes HIF-2α expression, which was confirmed in the experimental 897 and 907 cell lines without hypoxia signaling. We successfully constructed sh-HIF-2α and shRNA-NC lentiviruses to transfect Kras G12D -LOH and Kras G12D pancreatic cancer cells. The results showed that inhibition of HIF-2α can suppress the proliferation, migration, and invasion of Kras G12D -LOH and Kras G12D pancreatic cancer cells. Moreover, HIF-2α silencing can block the cell cycle and promote apoptosis. Therefore, HIF-2α plays a crucial role in the regulation of malignant phenotypes in Kras G12D -LOH and Kras G12D pancreatic cancer cells.
Pancreatic cancer cells metabolize Gln via a nonclassical pathway. Gln-derived glutamate yields aspartate through GOT1, which is then transferred to the cytoplasm to produce oxaloacetate, malic acid, and pyruvate by MDH1 and ME1 [26]. In contrast to those in Kras G12D pancreatic cancer cells, the Gln uptake rate and NADPH/NADP + ratio of Kras G12D -LOH pancreatic cancer cells were significantly decreased after HIF-2α downregulation, whereas ROS levels increased compared with those in the control and shRNA-NC groups. HIF-2α-mediated Gln metabolism was more apparent in Kras G12D -LOH PDAC cells. In addition, LOH activated c-Myc protein and mRNA levels, which had a regulatory effect on Gln metabolism; moreover, HIF-2α downregulation inhibited c-Myc expression. The suppression of HIF-2α or c-Myc significantly decreased GOT1 levels in the Gln metabolic pathway compared with those in the control group. The occurrence of LOH first increases HIF-2α expression; HIF-2α then enhances c-Myc activity, which regulates Gln metabolism in a pathway independent of hypoxia. These results suggest that HIF-2α knockdown can significantly repress Gln metabolism in Kras G12D -LOH pancreatic cancer cells in a manner dependent on c-Myc, without a significant change in Gln metabolism in Kras G12D pancreatic cancer cells. Kras G12D -LOH can promote Gln metabolism in Kras G12D -LOH pancreatic cancer cells by regulating the HIF-2α-c-Myc pathway, and HIF-2α plays an important role in regulating Gln metabolism in Kras G12D -LOH pancreatic cancer cells. Furthermore, the regulatory trend of nonclassical Gln metabolism is consistent with the malignant behavior of Kras G12D -LOH pancreatic cancer cells, suggesting that Kras G12D -LOH may regulate the malignant phenotype of pancreatic cancer through HIF-2α/c-Myc-mediated Gln metabolism. However, the precise mechanism by which the occurrence of LOH stimulates the HIF-2α pathway and hypoxia signaling requires further exploration in future research. In addition, clinical data must be collected and analyzed to confirm the regulatory role of HIF-2α in Kras G12D -LOH PDAC. Although HIF-2α knockdown results in a similar phenotype in Kras G12D and Kras G12D -LOH cells, this is not the case for Kras G12D -LOH cells. Because Kras G12D -LOH cell lines have a higher proliferation and metastasis capability, they were more effective at inhibiting the malignant phenotype than Kras G12D cells after HIF-2α knockdown. In addition, the regulation mechanism of HIF-2α in Kras G12D cell phenotypes requires further investigation. We hypothesize that additional pathways regulate the malignant phenotype of Kras G12D and Kras G12D -LOH cell lines simultaneously.
In conclusion, the findings of this study suggest that HIF-2α silencing in Kras G12D -LOH pancreatic cancer cells significantly decreases Gln metabolism and inhibits cell proliferation and invasion. Meanwhile, although downregulation of HIF-2α expression has minimal effects on Gln metabolism in Kras G12D pancreatic cancer cells, it inhibits cell proliferation and invasion, suggesting that Kras G12D -LOH regulates the malignant biological behavior of pancreatic cancer in HIF-2α-mediated Gln metabolism. Therefore, HIF-2α can be activated by Kras G12D -LOH to regulate Gln metabolism and participate in malignant phenotypes.

Measurement of Gln Consumption, the NADPH:NADP + Ratio, and Intracellular ROS Levels
Four experimental cell lines were seeded in 6-well (10 5 /well) plates for 48 h. A colorimetric assay kit (Biovision, Milpitas, CA, USA), NADPH/NADP + assay kit (Jiancheng Bioengineering Institute, Nanjing, China), and ROS detection kit (Jiancheng Bioengineering Institute) were used to measure Gln consumption, the NADPH:NADP + ratio, and intracellular ROS levels, respectively, according to the manufacturers' instructions. The specific methods were described previously [19].

Cell Viability, Colony Formation, Transwell, Cell Cycle, and Apoptosis Assays
The relevant methods are described in our previous studies [19] and the Supplementary Materials.

Xenograft Mouse Model
Female BALB/c nude mice (4 ± 1 week old, 14 ± 5 g) were injected with 200 mL (1 × 10 6 cells) of sh-HIF-2α-transfected Kras G12D and Kras G12D -LOH cells. Additionally, shRNA-NC cells (5 mice per group) were subcutaneously injected into the right axillary fossa. Subsequently, each animal was sacrificed by cervical dislocation 14 days after inoculation, and the tumors were retrieved. The animal experiments complied with the ARRIVE guidelines and were approved by the Ethics Committee on Animal Care and Use of Southeast University (No. 20200101003).

Immunohistochemistry
The tumors retrieved from BALB/c nude mice were fixed in formalin and embedded in paraffin; 4 µm-thick sections were prepared. The sections were first incubated with a primary antibody for Ki67 (1:200, 12202T, Cell Signaling Technology, Boston, MA, USA) overnight at 4 • C and subsequently incubated at 37 • C for 30 min with a corresponding secondary antibody (1:2000, 111-035-003, Jackson ImmunoResearch Inc, West Grove, PA, USA) at 37 • C. Finally, the sections were re-stained with hematoxylin and observed under a microscope (Nikon Eclipse, Tokyo, Japan).

GEPIA Database Analysis
Analysis of HIF-2α mRNA expression in pancreatic cancer tissues and normal pancreatic tissues was conducted using the GEPIA (http://gepia.cancer-pku.cn/ accessed on 16 April 2021) database.

Statistical Analysis
All experiments were independently conducted at least three times. The results are presented as the mean ± standard deviation (SD). The obtained experimental data were analyzed using one-way analysis of variance and Student's t-test. The SPSS software (IBM Corp., Armonk, NY, USA) was used for all experimental analyses. p-values < 0.05 were considered statistically significant.