Unfolded Protein Response Is Activated by Aurora Kinase A in Esophageal Adenocarcinoma

Simple Summary Esophageal cancer is the 6th most common cause of cancer-related deaths in 2018 worldwide, with a 5-year survival rate of around 20%. This study reported that esophageal adenocarcinoma cancer cells take advantage of unfolded protein response to survival. Our data using both in vitro cancer cell models and in vivo mice models discovered, for the first time, that Aurora kinase A hijacks pro-survival unfolded protein response in esophageal adenocarcinoma, promoting the survival of cancer cells under reflux-mediated stress conditions. Abstract Unfolded protein response (UPR) protects malignant cells from endoplasmic reticulum stress-induced apoptosis. We report that Aurora kinase A (AURKA) promotes cancer cell survival by activating UPR in esophageal adenocarcinoma (EAC). A strong positive correlation between AURKA and binding immunoglobulin protein (BIP) mRNA expression levels was found in EACs. The in vitro assays indicated that AURKA promoted IRE1α protein phosphorylation, activating prosurvival UPR in FLO-1 and OE33 cells. The use of acidic bile salts to mimic reflux conditions in patients induced high AURKA and IRE1α levels. This induction was abrogated by AURKA knockdown in EAC cells. AURKA and p-IRE1α protein colocalization was observed in neoplastic gastroesophageal lesions of the L2-IL1b mouse model of Barrett’s esophageal neoplasia. The combined treatment using AURKA inhibitor and tunicamycin synergistically induced cancer cell death. The use of alisertib for AURKA inhibition in the EAC xenograft model led to a decrease in IRE1α phosphorylation with a significant reduction in tumor growth. These results indicate that AURKA activates UPR, promoting cancer cell survival during ER stress in EAC. Targeting AURKA can significantly reverse prosurvival UPR signaling mechanisms and decrease cancer cell survival, providing a promising approach for the treatment of EAC patients.


TCGA and GEO Datasets Analysis
TCGA data for AURKA and BIP mRNA expression in EAC tissue samples were analyzed using the online tool UCSC Xena at https://xena.ucsc.edu/, accessed on 1 October 2020 [23]. The mRNA expression data of the esophageal adenocarcinoma (EAC) cohort were obtained from the Cancer Genome Atlas (TCGA) (https://portal.gdc.cancer.gov/, accessed on 1 February 2021) and Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/ gds, accessed on 1 May 2021) database. Samples diagnosed with EAC were enrolled for further analysis. The number of EAC samples included in each cohort was as follows: TCGA (n = 78), GSE26886 (n = 21), GSE37201 (n = 22), GSE37200 (n = 15), and GSE74553 (n = 52). We used R language (R 3.6.1; https://www.r-project.org/, accessed on 1 May 2021) for genetic annotation. When a gene contained multiple probes, these probes' mean was used for the gene expression value. Differential expression gene analysis was performed with the limma package. To analyze AURKA-related signaling pathways in EAC, we divided the samples of each data set into two groups. High-expression (top 25% AURKA expression samples) and low-expression (lowest 25% AURKA expression samples) groups were analyzed using mean ± SD of the AURKA mRNA value as the cutoff point. The ClusterProfiler v3.12.0 package (https://guangchuangyu.github.io/software/clusterProfiler, accessed on 1 May 2021) [24] was performed for GSEA analysis using DEG comparing AURKA high-and low-expression groups. All hallmark gene sets were obtained from MSigDB (https://www.gsea-msigdb.org/, accessed on 1 May 2021).

RNA Sequencing and Gene Set Enrichment Analysis
A total of 18 RNA samples were extracted from FLO-1, OE33, and SK-GT4 control (triplicates) or alisertib 400 nM 48 h treated (triplicates) cells. RNA seq was performed on a total of 1 µg RNA from each sample as described before [25]. Gene set enrichment analysis (GSEA) [26] was performed based on the RNA sequencing data obtained from all three EAC cell lines. The gene set used for UPR pathway analysis was Gene Set: HALLMARK_UNFOLDED_PROTEIN_RESPONSE (https://www.gsea-msigdb.org/gsea/ msigdb/cards/HALLMARK_UNFOLDED_PROTEIN_RESPONSE, accessed on 1 May 2021), which contains 22 genes upregulated during unfolded protein response.

qRT-PCR and Human Esophageal Tissue Samples
Deidentified human tissue samples from 32 esophageal cancer and 32 normal esophageal tissue samples were collected from the National Cancer Institute Cooperative Human Tissue Network (CHTN) and the pathology archives at Vanderbilt University Medical Center (Nashville, TN, USA). All tissue samples were collected, coded, and deidentified in accordance with the Institutional Review Board-approved protocols. The histology and age information are included in Table S1. Total RNA was purified using the miRNeasy mini kit (Qiagen, Redwood City, CA, USA). miRNA cDNA was reverse-transcripted as described before [27]. Quantitative real-time PCR (qRT-PCR) was performed using a Bio-Rad CFX Connect Real-time System with the threshold cycle number determined by Bio-Rad CFX manager software version 3.0. Primers that detect human genes were ordered from Integrated DNA Technologies (Coralville, IA, USA). The genes and sequences of qRT-PCR primers are given in Table S2. Results of target genes were normalized to human HPRT1.

Cell Viability ATP-Glo and Clonogenic Cell Survival Assay
Control siRNA or AURKA siRNA transfected FLO-1, and OE33 cells were seeded at 1000 cells per well in 96-well plates and treated with tunicamycin (range: 0.3-300 ng/mL) or AURKA inhibitor AK-01 (LY3295668) (range: 0.20-200 nM) or AURKA inhibitor alisertib (rang: 7.8-500 nM) or IRE1α inhibitor 3 6 -DMAD (range: 15.6-1000 nM) or PBS (control) for 5 days. Cell viability was measured using the CellTiter-Glo Cell Viability Assay (Promega, Madison, WI, USA). Changes in absorbance were recorded in a FluolarStar luminescence microplate reader (BMG Labtech, Ortenberg, Germany). FLO-1 or OE33 cells were seeded 500 cells/well in 6-well plates treated with tunicamycin (range: 0.0-400 ng/mL) or alisertib (range: 0-250 nM) or a combination for 48 h. Following treatments, cells were washed with PBS following incubation in a drug-free DMEM cell culture medium for 10 days. Subsequently, the media were removed, and cells were fixed with 4% paraformaldehyde solution for 10 min at room temperature. The cells were then gently washed with PBS and stained overnight with crystal violet (0.05% crystal violet in 50% methanol). Following overnight staining, the excess dye was gently washed off with PBS. The plates were photographed. Colony formation and cell survival were evaluated by quantifying the dye signal in each well with ImageJ image analysis software (https://imagej.nih.gov/ij/, accessed on 1 May 2021).

Western Blotting
Cells were collected after 5% trypsinization of the culture plate, followed by centrifuging at 12,000 rpm at 4 • C for 10 min. Cell pellets were resuspended in RIPA buffer containing protease inhibitor cocktail and phosphatase inhibitor sodium orthovanadate (Santa Cruz Biotechnology Inc., Dallas, TX, USA) on ice for 30 min with vortex every 10 min. Protein concentrations were measured using a Bio-Rad Protein Assay (Bio-Rad Laboratories, Hercules, CA, USA). An amount of 20 µg proteins from each sample were subjected to SDS-PAGE and transferred onto nitrocellulose membranes (PerkinElmer, Waltham, MA, USA). Membranes were blocked with 5% bovine serum albumin (BSA, Sigma-Aldrich). To detect target proteins, membranes were probed with specific primary antibodies overnight. The next day, membranes were washed for 10 min with TBS-T 3 times, followed by incubation with anti-rabbit or anti-mouse secondary antibodies. Protein bands were detected using chemiluminescence reagents (Millipore, Billerica, MA, USA).

Flow Cytometry Analysis of Cell Apoptosis/Death
Flow cytometry analysis of Annexin V and PI was performed using a FITC Annexin V apoptosis detection kit (BD Biosciences, San Jose, CA, USA) to quantitate ABS-induced apoptosis. EAC cells were treated with 200 nM alisertib or control. After 24 h, cells were treated with pH 4.0, 200 µM ABS, for 20 min, then recovered in regular medium for 3 h. According to the manufacturer's instruction, cells were then collected for FITC Annexin V and PI staining and subjected to flow cytometry analysis at the Flow Cytometry Shared Resource at the Sylvester Comprehensive Cancer Center.
The pL2-IL1β transgenic mice were a kind gift from Dr. Timothy Wang (Columbia University), a model of chronic esophageal inflammation that develops BE and EAC, as previously described [30]. The mice received drinking water containing 0.3% deoxycholic acid (DCA) at neutral pH for 7 months. Then the mice were sacrificed and subjected to histological analysis of the squamocolumnar junctions at the gastroesophageal junctions. All animal work was approved by the Institutional Animal Care and Use Committee.

AURKA Overexpression Correlates with UPR Activation in EAC
We and others have shown that AURKA plays vital biological functions to promote cancer cell survival [9,14]. To further identify novel roles of AURKA in EAC, AURKA mRNA expression levels at TCGA and four GEO databases containing 188 human esophageal adenocarcinomas tissue samples were analyzed. We performed signal pathway enrichment analysis using the upper quartile of AURKA high expressing compared with the lower quartile of AURKA low expressing EAC samples. The results demonstrated that the unfolded protein response (UPR) pathway was highly activated in AURKA high EAC samples compared with AURKA low tissues ( Figure 1A, left and middle panels, p < 0.01). To further test the correlation between AURKA and UPR in EAC, next-generation sequencing (NGS) analysis was performed between AURKA inhibitor (alisertib) treatment groups and control groups in FLO-1, OE33, and SK-GT4 cells. Gene set enrichment analysis using NGS data from the three cell lines demonstrated that unfolded protein response gene set expression was significantly downregulated in alisertib-treated EAC cells compared with control EAC cells ( Figure 1A, right panel, p < 0.001). UPR is known to be activated under ER stress in both normal and cancer cells [31,32]. Since few reports are published on the relationship of AURKA and ER stress, we further looked at our esophageal tissue samples and the TCGA database for BIP mRNA, a marker of ER stress [33] AURKA mRNA expression in EAC. Using RT-PCR in 30 normal esophageal samples and 32 EAC tissues, we detected a significant overexpression of both AURKA and BIP mRNA in EAC samples compared with the normal esophagus (NE) ( Figure 1B, p < 0.01). The results showed a strong positive correlation between AURKA and BIP mRNA in esophageal tissue samples ( Figure 1C, p < 0.0001). TCGA data analysis also demonstrated that AURKA mRNA was significantly overexpressed in EAC samples compared with normal esophageal tissues as expected ( Figure 1D, left panel, p < 0.001). In the meantime, BIP mRNA was significantly more expressed in EAC than normal esophageal samples ( Figure 1D, right panel, p < 0.01). The analysis demonstrated a significant positive correlation of AURKA mRNA and BIP mRNA expression in esophageal tissue samples ( Figure 1E, p < 0.0001, 11 normal esophagus samples, 79 esophageal adenocarcinoma samples). These findings suggest that UPR activation is positively correlated with AURKA expression in EAC samples. Based on these findings, we hypothesized that AURKA plays an essential role in activating UPR in EAC.

IRE1α Promotes EAC Cell Survival through AURKA
To test our hypothesis and further investigate the role of AURKA in the UPR pathway, we examined the protein expression levels of AURKA and critical UPR signaling proteins in three normal esophageal tissues (NE1, NE2, and NE3), Barrett's esophagus (BART and CPB), and esophageal adenocarcinoma cell lines (FLO-1 OE33, SK-GT4, and ESO26). Western blot results demonstrated that AURKA, p-IRE1α (S724) IRE1α, BIP, XBP1s, p-PERK (T982) PERK, and p-eIF2α(S51) protein expression levels were remarkably higher in Barrett's esophagus and EAC cells compared with normal esophageal tissues ( Figure 2A). IRE1α, PERK, and ATF6 are three well-known sensors that activate UPR, controlling the balance between cell survival and apoptosis during ER stress in cancer [33,34]. To test which sensors are vital in EAC cell models, we knocked down these proteins respectively in EAC cell lines using siRNA. Western blot data clearly showed that knockdown of PERK or IRE1α was successful in FLO-1 and OE33 cell lines ( Figure 2B). Interestingly, in both cell lines, IRE1α knockdown remarkably induced cleaved PARP expression, while the PERK or ATF6 knockdown had little effect on cleaved PARP induction ( Figure 2B,C). These results suggest that IRE1α is essential for EAC cell survival. To test whether AURKA plays a vital role in regulating IRE1α, we used AURKA siRNA knockdown in FLO-1 or OE33 cells. Western blot data demonstrated that AURKA knockdown decreased IRE1a phosphorylation and protein levels in both cell lines ( Figure 2D). To further test our hypothesis that AURKA promotes EAC cell survival during ER stress by inducing and activating prosurvival IRE1α expression, we used tunicamycin (TM) [35] to generate ER stress with or without AURKA siRNA knockdown in FLO-1 or OE33 cells. Western blot data indicated TM treatmentinduced BIP protein expression in both cell lines, as expected. In the meantime, AURKA knockdown decreased the IRE1α phosphorylation, expression, and protein expression level of cIAP2, a known downstream prosurvival target of IRE1α activation [36] ( Figure 2E). Interestingly, AURKA knockdown induced higher levels of cleaved PARP with TM-treated cells than control siRNA ( Figure 2E). These data strongly support our hypothesis that AURKA promotes IRE1α activation and cell survival in EAC cells.

IRE1α Activation Is Dependent on AURKA in Response to Acidic Bile Salt Reflux Conditions in EAC Cells
Acidic bile salts (ABSs) in gastroesophageal reflux disease (GERD) cause esophageal irritation and inflammation, the leading risk factor for EAC. ABSs induce DNA damage and oxidative stress in EAC cells. Recent studies have shown that cells under oncogenic stress, such as DNA damage and oxidative stress, develop ER stress, which activates UPR [37,38]. In this study, we found that AURKA induced IRE1α activation in EAC cells. Therefore, we investigated whether ABSs induce UPR in EAC and whether this induction is dependent on AURKA.

AURKA Protects EAC Cells from ABS-Induced Apoptosis
To investigate the prosurvival role of AURKA in EAC cells during ABS treatment, we used Annexin V/PI staining assay with or without ABS treatment (20 treatment + 3 h recovery), AURKA inhibitor alisertib (400 nM overnight pretreatment), or combinations. The results indicated that ABS induced cancer cell death as measured by Annexin V staining, PI staining, and double staining ( Figure 4A,B and Supplementary Figure S1C, p < 0.01). In the meantime, cell death was significantly higher with AURKA siRNA knockdown with or without ABS treatment compared with control siRNA ( Figure 4A,B and Supplementary Figure S1C, p < 0.01). Using Annexin V/PI staining in SK-GT4, ABS treatment induced significantly more cell death in AURKA siRNA knockdown than control SK-GT4 cells ( Figure 4C).

AURKA Binds to IRE1α in EAC Cells
To investigate the mechanisms by which AURKA phosphorylates IRE1α, w performed proximity ligation assay (PLA) for AURKA and IRE1α proteins in both FLOand OE33 cells. Our data indicated that AURKA was closely localized with the IRE1α protein in both cell lines ( Figure 5A and Supplementary Figure S2A). Furthermore, AB treatment significantly increased PLA signals. In contrast, the alisertib treatmen

AURKA Binds to IRE1α in EAC Cells
To investigate the mechanisms by which AURKA phosphorylates IRE1α, we performed proximity ligation assay (PLA) for AURKA and IRE1α proteins in both FLO-1 and OE33 cells. Our data indicated that AURKA was closely localized with the IRE1α protein in both cell lines ( Figure 5A and Supplementary Figure S2A). Furthermore, ABS treatment significantly increased PLA signals. In contrast, the alisertib treatment diminished such induction in both FLO-1 and OE33 cells, consistent with our findings that ABS-induced IRE1α phosphorylation is dependent on AURKA in EAC cells ( Figure 5A,B and Supplementary Figure S2A,B, p < 0,05). To further validate the binding of AURKA with IRE1α, we performed immunoprecipitation (IP) in FLO-1 cells with or without ABS or alisertib treatment and OE33 cells with or without ABS or AURKA siRNA knockdown. Our Western blot data demonstrated that ABS induced IRE1α phosphorylation in both cell lines. The induction of IRE1α phosphorylation was abrogated by alisertib treatment or AURKA siRNA knockdown, as expected ( Figure 5C and Supplementary Figure S2C). In the meantime, ABS remarkably promoted the binding of AURKA and IRE1α, detected by Western blot in IRE1α IP samples. Alisertib treatment or AURKA knockdown diminished the induction of binding between AURKA and IRE1α ( Figure 5D and Supplementary Figure S2D). Our data also indicated no change of binding between BIP and IRE1α with ABS or alisertib treatment in FLO-1 cells ( Figure 5D). We also performed in vitro kinase assay to investigate whether AURKA phosphorylates IRE1α directly. However, we cannot conclude due to the autophosphorylation of IRE1α in the experimental condition. Autophosphorylation of IRE1α is well [39,40]. These results suggest that AURKA binds to IRE1α to promote its autophosphorylation in EAC cells. This binding is strongly induced with ABS treatment. diminished such induction in both FLO-1 and OE33 cells, consistent with our findings that ABS-induced IRE1α phosphorylation is dependent on AURKA in EAC cells ( Figure 5A,B and Supplementary Figure S2A,B, p < 0,05). To further validate the binding of AURKA with IRE1α, we performed immunoprecipitation (IP) in FLO-1 cells with or without ABS or alisertib treatment and OE33 cells with or without ABS or AURKA siRNA knockdown. Our Western blot data demonstrated that ABS induced IRE1α phosphorylation in both cell lines. The induction of IRE1α phosphorylation was abrogated by alisertib treatment or AURKA siRNA knockdown, as expected ( Figure 5C and Supplementary Figure S2C). In the meantime, ABS remarkably promoted the binding of AURKA and IRE1α, detected by Western blot in IRE1α IP samples. Alisertib treatment or AURKA knockdown diminished the induction of binding between AURKA and IRE1α ( Figure 5D and Supplementary Figure S2D). Our data also indicated no change of binding between BIP and IRE1α with ABS or alisertib treatment in FLO-1 cells ( Figure 5D). We also performed in vitro kinase assay to investigate whether AURKA phosphorylates IRE1α directly. However, we cannot conclude due to the autophosphorylation of IRE1α in the experimental condition. Autophosphorylation of IRE1α is well [39,40]. These results suggest that AURKA binds to IRE1α to promote its autophosphorylation in EAC cells. This binding is strongly induced with ABS treatment.

AURKA Promotes Survival under Tunicamycin-Induced ER Stress in EAC Cells
As the results indicated that AURKA induced IRE1α phosphorylation, we investigated the potential prosurvival role of AURKA under cytotoxic ER stress in EAC cells. FLO-1 or OE33 cells were treated with the ER stress inducer tunicamycin (TM, 250 or 500 ng/mL, 24 h) with or without AURKA siRNA knockdown. ATP-Glo results demonstrated that TM alone significantly decreased cell viability on day 3 in both cell lines ( Figure 6A and Supplementary Figure S3, p < 0.01). The combination of AURKA knockdown and TM significantly reduced cell viability more than AURKA siRNA alone, TM alone, or the control group ( Figure 6A and Supplementary Figure S3). Cell viability was also determined with different concentrations of TM, AURKA inhibitor AK-01, or a combination. Our ATP-Glo data indicated that combination treatment decreased TM IC50s in FLO-1 (from 74.2 to 23.99 ng/mL) and OE33 cells (from 38.69 to 23.65 ng/mL) ( Figure 6B and Supplementary Figure S4A-C). The IC50s of AK-01 were decreased in combination groups compared with single treatment (FLO-1: from 25.96 to 15.99 nM; OE33: from 392.24 to 63.14 nM) in Figure 6B and Supplementary Figure S4A-C. Similar ATP-Glo results were found using alisertib, another AURKA inhibitor ( Figure 6C and Supplementary Figure S4D). Our results demonstrated that the alisertib and TM combination dramatically sensitized EAC cells to the treatment. The IC50s of alisertib was decreased from 59.36 to 27.65 nM in FLO-1 and from 64.96 to 27.15 nM in OE33 cells. Meanwhile, the IC50s of TM was decreased from 185.9 to 16.59 ng/mL in FLO-1 and from 53.68 to 16.29 ng/mL in OE33 ( Figure 6C and Supplementary Figure S4D).
The clonogenic cell survival assay was performed to test the long-term effect of AU-RKA inhibition during ER stress in FLO-1, OE33 cells with TM alone, alisertib alone, or combination treatment. The results indicated that alisertib and TM combination treatment significantly decreased cell viability more than a single treatment, quantified by relative cell staining intensity in a long-term colony formation assay ( Figure 6D and Supplementary Figure S4E,F). To investigate whether there is a synergistic effect of AURKA inhibition and IRE1α inhibition, we tested the IRE1α inhibitor 3 6 -DMAD in FLO-1, OE33, and SK-GT4 cells. The ATP-Glo data indicated similar IC50s for all three cell lines from 153.7 to 175.52 nM (Supplementary Figure S5A). However, there was no synergistic effect of alisertib and 3 6 -DMAD in all three cell lines (Supplementary Figure S5B; OE33 data are not shown). In conclusion, these novel results indicated that AURKA inhibition combined with TM-induced ER stress could be a possible therapeutic approach for EAC.

AURKA and ABS Promote IRE1α Phosphorylation In Vivo
Tissues from FLO-1 and OE33 tumor xenograft mouse models [29] were investigated to examine the regulation of IRE1α by AURKA in vivo. Immunofluorescence staining of the xenograft tumor tissue samples demonstrated that alisertib treatment not only decreased xenograft tumor size but also significantly reduced phospho-IRE1α (p-IRE1α) staining, as compared with control tumor samples ( Figure 7A,B and Supplementary Figure S6, p < 0.01). Immunofluorescence staining data also demonstrated colocalization of AURKA and IRE1α (yellow staining), higher in the control group than in the alisertib treatment group ( Figure 7A and Supplementary Figure S6). To further confirm the IRE1α activation after bile salts in vivo, we used the pL2-IL1β mouse, and a model of chronic esophageal inflammation that develops BE-like lesions was investigated [30]. Immunofluorescence staining was performed for p-IRE1α and AURKA in normal control esophagus and gastroesophageal junction high-grade dysplasia (HGD) samples. p-IRE1α staining was remarkably stronger in HGD samples compared with normal tissues. Furthermore, we observed p-IRE1α and AURKA colocalization in HGD samples.

AURKA and ABS Promote IRE1α Phosphorylation In Vivo
Tissues from FLO-1 and OE33 tumor xenograft mouse models [29] were investigated to examine the regulation of IRE1α by AURKA in vivo. Immunofluorescence staining of the xenograft tumor tissue samples demonstrated that alisertib treatment not only decreased xenograft tumor size but also significantly reduced phospho-IRE1α (p-IRE1α) staining, as compared with control tumor samples ( Figure 7A,B and Supplementary Figure S6, p < 0.01). Immunofluorescence staining data also demonstrated colocalization of AURKA and IRE1α (yellow staining), higher in the control group than in the alisertib treatment group ( Figure 7A and Supplementary Figure S6). To further confirm the IRE1α activation after bile salts in vivo, we used the pL2-IL1β mouse, and a model of chronic esophageal inflammation that develops BE-like lesions was investigated [30]. Immunofluorescence staining was performed for p-IRE1α and AURKA in normal control esophagus and gastroesophageal junction high-grade dysplasia (HGD) samples. p-IRE1α staining was remarkably stronger in HGD samples compared with normal tissues. Furthermore, we observed p-IRE1α and AURKA colocalization in HGD samples.

Discussion
Esophageal adenocarcinoma is an aggressive malignancy with an estimated 16,000 deaths in the United States in 2020 [41]. Chronic GERD is considered the leading risk factor for developing a premalignant metaplastic condition known as BE and its progression to EAC [42]. However, the current understanding of the underlying biology and molecular mechanisms of EAC remains limited. Excessive protein production places cancer cells in endless ER stress. Therefore, cancer cells must adapt and tilt the balance of survival over death during ER stress by activating unfolded protein response (UPR) [43]. In this study, we demonstrated the role of AURKA in promoting prosurvival UPR under reflux conditions in EAC.
Although the role of AURKA in promoting cancer cell survival and drug resistance has been established [14,44,45], its role in regulating prosurvival UPR in cancer cells remains unknown. Analysis of TCGA, GEO databases, and human tissue samples demonstrated a robust positive correlation between AURKA and increased UPR/ER stress in EAC. Recent studies indicated that IRE1α, PERK, and ATF6 are essential ER stress sensors, regulating the balance between cell survival and death during ER stress through UPR [33]. Our data not only indicated a remarkable increase in UPR proteins in EAC cells but also showed the critical prosurvival role of IRE1α compared with the other two primary ER stress sensors, PERK and ATF6. As we showed the co-overexpression of AURKA and UPR key proteins, our data indicated that AURKA induced IRE1α levels and phosphorylation in EAC cells. IRE1α and its downstream sXBP1 promote cell survival during ER stress in several malignancies [33,46].
EAC is a unique malignancy that develops under continuous cellular stress mediated by chronic GERD, forcing cells to develop adaptive prosurvival cellular mechanisms to overcome reflux-induced stress [47]. Surprisingly, to the best of our knowledge, no studies in the literature have investigated the UPR pathway under reflux/ABS conditions in EAC.
Our results indicate that AURKA-induced IRE1α tilts the cellular homeostatic balance towards cell survival rather than death during an overwhelming ER stress under oxidative genotoxic reflux conditions. Of note, we detected colocalization and binding of AURKA to IRE1α during reflux/ABS conditions in vitro and in vivo. At the same time, neither ABS treatment nor AURKA affected BIP binding to IRE1α. Based on our data, we suggest that AURKA binding to IRE1α enhances IRE1α stability and autophosphorylation under ER stress conditions without affecting BIP binding to IRE1α. Autophosphorylation of IRE1α has been described as a critical mechanism for counteracting cell death and promoting survival under ER stress [40,48]. Our findings elucidate a previously unrecognized mechanism by which AURKA activates UPR in EAC cells to promote cell survival under reflux conditions.
To examine the efficacy of targeting AURKA and ER stress in EAC, we combined AURKA inhibition and ER stress inducer treatment. Our data indicated that both AURKA knockdown and pharmacological inhibition synergize with an ER stress inducer in short-or long-term experimental settings. Similar findings were discovered in different cancers where TM synergized with several antitumor therapeutics [49][50][51]. In the meantime, we also tested the efficacy of AURKA inhibition and IRE1α inhibition in EAC. The results demonstrated a lack of substantial synergistic effect when AURKA inhibition is combined with IRE1α inhibition. These findings are expected and provide additional proof that AURKA inhibition is sufficient to inhibit IRE1α, not necessitating the addition of IRE1 inhibitors. Future therapeutic strategies should focus on treating EAC cells with the combination of an AURKA inhibitor and ER stress inducer.

Conclusions
In summary, our findings demonstrate for the first time the function of AURKA in regulating prosurvival UPR through IRE1α. This AURKA-dependent UPR activation provides a new paradigm in esophageal tumorigenesis and treatment.

Institutional Review Board Statement:
The study design regarding deidentified human data or tissues was approved by the Institutional Research Ethics Committee. All animal studies were carried out following the protocols approved by the Institutional Animal Care and Use Committee of the University of Miami (UM-20-110).
Informed Consent Statement: Not applicable.