Different expressions of ATF6, EMC6, XBP1, CHOP, and APAF1 in PC and normal tissues were analyzed in Gene Expression Profiling Interactive Analysis 2 (GEPIA2; http://gepia2.cancer-pku.cn/#index). GEPIA2 is an interactive web server for analyzing the expression data of RNA from 9,736 tumors and 8,587 normal samples from the Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) datasets (Tang et al., 2019). The p-value cutoff of the expression of gene for analysis was 0.01. The |Log2FC| cutoff of the expression of gene for analysis was 1, and we used log2 (TPM + 1) for log scale.

Survival analysis for overall survival (OS) and disease-free survival (DFS) in GEPIA2 was also used to estimate the relationship between PC prognostic value and the expression of genes ATF6, EMC6, XBP1, CHOP, and APAF1. The meaning of DFS was similar to recurrence-free survival (RFS) for this study. Hazards ratio (HR) was calculated based on Cox PH Model. The 95% confidence interval (CI) was shown by dotted line. The expression median value of gene for analyzing was identified as group cutoff to distinguish the high-expression group and the low-expression group.

PC and adjacent normal pancreatic tissue samples were retrospectively collected from 69 PC patients, including 39 males and 30 females, with a mean age of about 57 years ranging from 34 to 79 years. These patients underwent surgical resections in Nanfang Hospital, Southern Medical University, between October 2010 and April 2019. A more detailed information about the patients is shown in Supplementary Table S1. Before the experiments, each patient provided a written informed consent. After resection, each sample was frozen at −80°C until analysis and all patients were followed up until May 2019. Complete clinical and pathological data and follow-up documentations were recorded and analyzed for all patients in the study, who had never received preoperative chemotherapy or radiotherapy. The tumor stages in the study were classified according to the Union for International Cancer Control (UICC). RFS was defined as the period from the date of pancreatic resection until the date of recurrence diagnosis. OS was defined as the period from the date of surgical resection until the date of death or last follow-up. Disease-specific survival (DSS) was defined as the period from the date of surgical resection until the date of death due to PC. The collection and analysis of tissue and data were approved by the Ethics Committee of the Southern Medical University.

Pancreatic tissues were fixed in 4% neutral phosphate-buffered formalin, embedded in paraffin, and cut into 5-μm thick sections. Hematoxylin and eosin (H&E) staining were performed by experienced pathologists and followed by double-blinded histological evaluations. For the immunohistochemical detection, the sections were sequentially incubated overnight at 4°C with anti-ATF6 (Bioss, diluted 1:100), anti-XBP1 (Bioss, diluted 1:200), anti-CHOP (Bioss, diluted 1:100), anti-EMC6 (Proteintech, diluted 1:100) and anti-APAF1 (Abcam, diluted 1:100) antibodies. After 30 min of incubation with secondary antibodies at room temperature, the sections were counterstained with DAB solution and hematoxylin. Positively stained cells were evaluated by two experienced pathologists according to a previously described protocol (Zhou et al., 2020). Five high magnification areas were evaluated from each sample.

The immunohistochemical scores were assessed based on the intensity of staining and the proportion of stained cells. The scores of 0, 1, 2, and 3, respectively, corresponded to negative, weak, moderate, and strong staining intensities. The proportion of positively stained cells for each intensity was scored as follows: 0 (0%–5% positive cells), 1 (5%–25% positive cells), 2 (26%–50% positive cells), 3 (51%–75% positive cells), and 4 (76%–100% positive cells). The IHC scores given by each pathologist were calculated by multiplying the proportion of positively stained cells by the staining intensity scores. The final IHC scores were the mean value of scores from two pathologists and divided into low expression (0–7) and high expression (8–12) groups.

The human PC cell lines, SW1990, HUPT4, PATU8988, PANC1, and ASPC1, were obtained from American Type Culture Collection (ATCC, Rockville, MD, USA). All the cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco). All mediums were supplemented with 10% fetal bovine serum (FBS) and maintained in a 37°C and 5% CO₂ atmosphere.

ATF6, EMC6, and APAF1 expression and functions were investigated by Western blotting, qRT-PCR, CCK8 assay, and Transwell assay. Si-ATF6, Si-EMC6, and Si-APAF1 were, respectively, used to inhibit the expression of ATF6, EMC6, and APAF1. The constructs OE-ATF6, OE-EMC6, and OE-APAF1 were used to, respectively, overexpress ATF6, EMC6, and APAF1. The negative control (NC) and the vector were designed and synthesized by RiboBio Co., Ltd. (Guangzhou, China). PC cells were separately seeded in 24-well plates at a density of 5 × 10⁵ cells and transfected with Si-ATF6, Si-EMC6, Si-APAF1, OE-ATF6, OE-EMC6, OE-APAF1, NC, and the vector using riboFECT mRNA Transfection Reagent (RiboBio Co., Ltd. Guangzhou, China) according to the recommendations of the manufacturer. After 48 h of transfection and incubation in a 37°C and 5% CO₂ atmosphere, the cells were harvested for subsequent experiments.

Trizol Reagent (Merck, Germany) was used to extract RNA from cultured cells according to the instructions of the manufacturer. QRT-PCR was performed using the SYBR Premix Ex Tag kit (Takara Biotechnology Co., Ltd.) and the Applied Biosystems 7500 Real-Time PCR system (Thermo Fisher Scientific Inc., UK). The primers were designed and synthesized by RiboBio Co., Ltd. (Guangzhou, China). ATF6 forward, 5′-CGC CTT TTA GTC CGG TTC TT-3′ and reverse, 5′-CCA GTT GGT AAC AAT GCC ATG T-3′; EMC6 forward, 5′-GTC GCC AAG ATT TGC TCC CT-3′ and reverse, 5′-AAA CAC ACA ATG CCG GTA CAC-3′; APAF1 forward, 5′-GAT CCA CAC AGG CCA TCA CA-3′ and reverse, 5′-GGC GGG AGT CTA TGT TCC AC-3′. GAPDH forward, 5′-ATC ATC AGC AAT GCC TCC TG-3′ and reverse, 5′-ATG GAC TGT GGT CAT GAG TC-3′. ATF6, EMC6, and APAF1 expressions were normalized by GAPDH. The 2−^△△Ct method was used to calculate the relative expression levels. The expression levels of the genes were measured by qRT-PCR.

RiPA buffer (GenStar, China) was used to extract the proteins from the transfected cell lines. The proteins were loaded and separated on SDS-PAGE, and transferred onto PVDF membranes (Millipore, USA). After blocking with 5% non-fat milk, the membranes were incubated with anti-ATF6 (Bioss, China), anti-EMC6 (Proteintech, USA), anti-APAF1 (Abcam, UK), and anti-GAPDH (Fude Biological Technology Co., Ltd. China) antibodies overnight at 4°C. Following this step, the membranes were incubated with horseradish peroxidase-coupled secondary antibodies for 1 h, and finally, the expression of the proteins was revealed using the enhanced chemiluminescence solution (ECL, PerkinElmer, USA).

The cell counting kit-8 assay was performed to assess the viability of the transfected cell lines. The transfected cells were seeded in a 96-well plate and cultured for 72 h. Then, 10 μl of CCK8 solution (Dojindo, Japan) was added to each well and incubated for 2 h. The absorbance at 450 nm was measured using a microplate reader.

The Transwell invasion assay was performed to measure the invasion ability of the transfected cells using the Transwell chambers. The transfected cells were suspended in serum-free medium and seeded in the upper chambers that were precoated with Matrigel. The medium containing 10% FBS was placed in the lower chamber. After 24 h of incubation at 37°C, the cells on the bottom chamber were stained with 0.1% crystal violet for 10 min at 37°C. The number of stained cells in the lower chambers was calculated using a microscope.

The SPSS software version 26.0 (SPSS, Chicago, IL) and the GraphPad Prism software version 8.2 (San Diego, CA, USA) were used for statistical analyses. The data were reported as mean ± standard deviation. The paired Student’s t-test or the chi-square test were used to analyze the expression differences of genes among normal, high, and low expression groups. The Kaplan–Meier method was performed for survival curves between low-expression and high-expression groups using the log-rank test. To identify the factors involved in PC, the Cox proportional hazards regression method was applied using univariate and multivariate analyses. Differences were considered significant if p < 0.05.

GEPIA2 was used to evaluate the relationship between the expression of ATF6, EMC6, XBP1, CHOP, APAF1, and prognosis value of PC. There was an upregulated trend of ATF6, EMC6, APAF1, and CHOP expression in PC compared with that in normal pancreatic tissues, while the result of XBP1 was opposite in this event (Figure 1C). Noticeably, the elevated expression of ATF6 and reduced expression of EMC6 and APAF1 showed a statistically significant association with poor OS for PC, but not with DFS, while the expression level of CHOP and XBP1 had no correlation with OS and DFS of PC (Figure 2).

The clinicopathological features of the 69 patients are described in Table 1. Briefly, the median age of the patients at diagnosis was 58 years (age range: 34–79 years), and 39.1% of patients were >60 years old and 63.8% had tumors with sizes>3 cm. Patients were male in 56.5%, and the proportion of smokers was 23.2%. Adenocarcinoma was the most common histological type that was observed in 95.7% of patients. Stage I (42.0%) and II (44.9%) were common, with involvement of lymph nodes, and vascular and neural invasions observed in 31.9%, 13.0%, and 26.1%, respectively.

To determine the expression levels of ER stress-related and apoptosis-related proteins in human PC, we collected PC and normal pancreatic tissues for immunohistochemistry analysis. The expression of the ER stress-related proteins, ATF6, XBP1, CHOP, and EMC6, and the apoptosis-related protein, APAF1, were significantly higher in PC tissues compared with those in normal pancreatic tissues (Figures 1A,B). Compared with these results to GEPIA2, the reason for the different results in XBP1 may be the differences of sources of data analyzed. The RNA-seq datasets were used by GEPIA2, instead of the IHC data. These results demonstrate that the expressions of ATF6, XBP1, CHOP, EMC6, and APAF1 were upregulated in PC.

To validate the associations between clinicopathological and molecular variables, and prognostic values in the PC patients, survival analysis was performed using the Kaplan–Meier method and the significance was tested with the log rank test. Univariate Cox regression analysis revealed that the TNM stage, lymph node involvement, and the expression levels of ATF6, EMC6, and APAF1, significantly correlate with RFS, OS, and DSS, while no significant differences were observed between the two groups based on age, sex, smoking, tumor size, neural and vascular invasions, and XBP1 and CHOP expression (Tables 2–4). Notably, the multivariate Cox regression analysis revealed that the TNM stage [HR = 3.578; p = 0.027], ATF6 expression [HR = 0.220; p < 0.001], EMC6 expression [HR = 2.571; p = 0.020], and APAF1 expression [HR = 2.426; p = 0.026] were also independent prognostic factors for RFS (Table 2). Simultaneously, the TNM stage [HR = 4.064; p = 0.014], lymph node involvement [HR = 0.380; p = 0.034], ATF6 expression [HR = 0.229; p = 0.001], EMC6 expression [HR = 2.956; p = 0.010], and APAF1 expression [HR = 2.369; p = 0.034] were also found to be significant prognostic factors for OS (Table 3). Moreover, the TNM stage [HR = 4.073; p = 0.017], lymph node involvement [HR = 0.396; p = 0.046], ATF6 expression [HR = 0.183; p < 0.001], EMC6 expression [HR = 3.275; p = 0.015], and APAF1 expression [HR = 2.887; p = 0.029] were also prognostic factors for DSS (Table 4). Kaplan–Meier survival plots indicated significantly higher survival rates at each time point for the ATF6^low, EMC6^high, and APAF1^high groups compared with those in the ATF6^high, EMC6^low and APAF1^low groups (p < 0.05, Figures 3A,D,E). There was no statistical significance for XBP1 and CHOP (p > 0.05, Figures 3B,C). Discrepancy of results in ATF6, EMC6, and APAF1 from GEPIA2 survival analysis and our experiment may be due to different sources of data analyzed. Another reason probably lies in the different group cutoff in distinguishing the high-expression group and the low-expression group. The samples were divided into two groups by expression median of genes in GEPIA2 database, instead of defining “high expression” as the IHC scores ≥8 in our study. Nevertheless, both GEPIA2 survival analysis and our study identified that the expression of ATF6, EMC6, and APAF1 was related to PC patients’ survival. The results demonstrate that among the genes that are related to ER stress and apoptosis in this research, only ATF6, EMC6, and APAF1 were associated with PC patients’ survival, which would be used in further studies.

To evaluate the expression of ATF6, EMC6, and APAF1 in pancreatic carcinoma cells by qRT-PCR, the PC cell lines, SW1990, HUPT4, PATU8988, PANC1, and ASPC1, were used. Noticeably, the highest and the lowest level of expression of ATF6 was observed in the cell lines, ASPC1 and SW1990 (Figure 4A). Therefore, SW1990 and ASPC1 were used as representative PC cell lines for ATF6 subsequent studies. Similarly, PATU8988 and SW1990 were selected for EMC6, PANC1, and ASPC1 was selected for APAF1 experiments (Figures 4B,C).

To further explore the function of ATF6, EMC6, and APAF1 in PC, we determined the effect of ATF6, EMC6, APAF1 on the proliferation and invasion abilities of PC cells using the CCK8 and Transwell assays on ATF6 transfected PC cell lines, SW1990 and ASPC1, EMC6 transfected PC cell lines, PATU8988 and SW 1990, APAF1 transfected cell lines, PANC1 and ASPC1. For this, qRT-PCR and Western blotting were used and revealed that the expression of ATF6, EMC6, and APAF1 were markedly increased in PC cells that were transfected with OE-ATF6, EMC6, and APAF1, and markedly inhibited in PC cells transfected with Si-ATF6, EMC6, and APAF1 when compared with the control (Figures 4D–L).

The result of the CCK8 assay showed that ATF6 overexpression and EMC6 or APAF1 knockdown enhances the growth of PC cells (Figures 5B,D,G), whereas ATF6 knockdown and EMC6 or APAF1 overexpression decreases the proliferation of PC cells compared with the control (Figures 5A,E,H). These results indicate that ATF6 promotes the viability of PC cells, while EMC6 and APAF1 have an inhibitory role.

The Transwell assay showed that the invasiveness of PC cells was markedly increased when ATF6 expression level is elevated and when EMC6 and APAF1 expression levels are reduced. However, ATF6 knockdown and EMC6 or APAF1 overexpression had the opposite effect (Figures 5C,F,I). Therefore, ATF6 promotes the invasion ability of PC cells, while EMC6 and APAF1 impair it.