Unraveling the Unfolded Protein Response Signature: Implications for Tumor Immune Microenvironment Heterogeneity and Clinical Prognosis in Stomach Cancer

doi:10.21203/rs.3.rs-3115880/v1

Download PDF

Research Article

Unraveling the Unfolded Protein Response Signature: Implications for Tumor Immune Microenvironment Heterogeneity and Clinical Prognosis in Stomach Cancer

https://doi.org/10.21203/rs.3.rs-3115880/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 02 May, 2024

Read the published version in Aging →

Version 1

posted

You are reading this latest preprint version

Background

Stomach cancer is a leading cause of cancer-related deaths globally due to its high grade and poor response to treatment. Understanding the molecular network driving the rapid progression of stomach cancer is crucial for improving patient outcomes.

Methods

This study aimed to investigate the role of unfolded protein response (UPR) related genes in stomach cancer and their potential as prognostic biomarkers. RNA expression data and clinical follow-up information were obtained from the TCGA and GEO databases. An unsupervised clustering algorithm was used to identify UPR genomic subtypes in stomach cancer. Functional enrichment analysis, immune landscape analysis, and chemotherapy benefit prediction were conducted for each subtype. A prognostic model based on UPR-related genes was developed and validated using LASSO-Cox regression, and a multivariate nomogram was created. Key gene expression analyses in pan-cancer and in vitro experiments were performed to further investigate the role of the identified genes in cancer progression.

Results

A total of 375 stomach cancer patients were included in this study. Analysis of 113 UPR-related genes revealed their close functional correlation and significant enrichment in protein modification, transport, and RNA degradation pathways. Unsupervised clustering identified two molecular subtypes with significant differences in prognosis and gene expression profiles. Immune landscape analysis showed that UPR may influence the composition of the tumor immune microenvironment. Chemotherapy sensitivity analysis indicated that patients in the C2 molecular subtype were more responsive to chemotherapy compared to those in the C1 molecular subtype. A prognostic signature consisting of seven UPR-related genes was constructed and validated, and an independent prognostic nomogram was developed. The gene IGFBP1, which had the highest weight coefficient in the prognostic signature, was found to promote the malignant phenotype of stomach cancer cells, suggesting its potential as a therapeutic target.

Conclusion

This study identified distinct molecular subtypes in stomach cancer patients and developed a prognostic signature to guide personalized treatment. Targeting IGFBP1, which promotes stomach cancer malignancy, could be a promising therapeutic strategy.

Stomach cancer

Unfolded protein response

Prognostic signature

IGFBP1

Tumor immune microenvironment

Stomach cancer poses a significant health burden, with low survival rates, especially among patients with metastatic disease [1]. The efficacy of current treatment options, including surgery, chemotherapy, and targeted therapies, is hindered by the intricate nature of cancer and the dynamic immune landscape [2, 3]. Enhancing patient prognosis requires the development of innovative prognostic models and exploration of new therapeutic approaches to aid clinical decision-making and improve outcomes.

The unfolded protein response (UPR) is a cellular stress response activated by unfolded or misfolded proteins accumulated in the endoplasmic reticulum (ER) that overwhelm ER’s coping ability[4, 5]. The unfolded protein response (UPR) is known to induce cell cycle arrest, activate transcription of UPR-related stress genes, inhibit protein translation, and facilitate the degradation of misfolded proteins through the activation of specific kinases. In cases where unresolved endoplasmic reticulum (ER) stress persists, the UPR may shift towards promoting apoptosis [6–8]. As the UPR maintains ER homeostasis and regulates autophagy, glucose metabolism, radox balance and so on, the impact of UPR on malignant tumor is complicated. Accumulating studies indicated the prognostic value of UPR [9–10]. In X-box binding protein 1 (XBP1)-luciferase transgenic mice, variations in UPR reflected different metabolic and hypoxic microenvironment and predicted tumor growth [11]. In cutaneous melanoma, hepatocellular carcinoma, bladder cancer, breast cancer and glioma, UPR related signature has been constructed to predict survival or therapeutic effect[12–16]. As tumor cells often grow under stressful conditions, UPR activation affects cell fate decisions and is reasonably believed to be a prognostic marker.

In this study, UPR-related gene expression was utilized to stratify patients with stomach cancer into two molecular subtypes, which demonstrated contrasting survival rates, immune characteristics, chemotherapy responses, and stemness scores. A prognostic signature based on UPR-associated genes was developed to predict patient outcomes, and a nomogram was created to assess the risk level. These findings underscore the potential of UPR-related genes in identifying high-risk patients, facilitating personalized treatment approaches, and pinpoint IGFBP1 as a promising target for therapeutic intervention.

The acquisition of clinical and gene expression data

The latest RNA expression data and clinical follow-up information were downloaded from the Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) database. To eliminate batch effects between datasets, data correction was performed through the combat algorithm. Then, the HALLMARK UNFOLDED PROTEIN RESPONSE related genes were acquired from the Molecular Signatures Database (https://www.gsea-msigdb.org/gsea/msigdb/index.jsp). These genes have been found to have an important role during the unfolded protein response process in previous studies.

Genotyping of unfolded protein response in stomach cancer

Clustering unsupervised analysis was used to identify genotyping from the expression of UPR-related genes. Then, patients in each genotyping were performed to survival analysis. The number of clusters and stability of clusters were determined by consensus clustering algorithm. Using R package “ConsensuClusterPlus”, 100 iterations were performed to ensure classification stability with the above steps.

Analysis of functional enrichment

The UPR genomic subtype of stomach cancer was identified through unsupervised clustering analysis. Differential gene analysis was subsequently conducted using the "limma" package, comparing cluster 1 molecular subtype and cluster 2 molecular subtype with the criteria of log₂ fold change (FC) > 1 or < -1, and a false discovery rate (FDR) p-value below 0.05. Concurrently, Gene Ontology (GO) functional analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were performed based on the differentially expressed genes (DEGs). The "clusterProfiler" R package (v3.0.0) was employed for KEGG pathway analysis and Gene Ontology (GO) analysis.

The immune landscape analysis

The Extended Polydimensional Immunome Characterization (EPIC), quanTIseq (a method for quantifying tumor immune contexture), xCell, and MCPcount were employed to evaluate cellular components and immune responses between clusters based on UPR-related genes. The purpose of this comparison was to thoroughly assess the stomach cancer immune microenvironment, as a comprehensive understanding of the immune landscape can provide valuable insights into tumor progression, prognosis, and response to therapy. By examining the immune cell composition and interactions, researchers can identify potential therapeutic targets and develop more effective strategies to improve patient outcomes.

Chemotherapy benefit predicted by clusters

The chemotherapeutic response for each sample was predicted using the largest publicly accessible pharmacogenomics database, the Genomics of Drug Sensitivity in Cancer (GDSC, http://www.cancerrxgene.org/). The R package "pRRophetic" was employed to facilitate the prediction process. Ridge regression was applied to estimate each sample's half-maximal inhibitory concentration (IC₅₀).

Developing and validating prognostic models of UPR-related signature

The number of UPR-related genes was reduced using LASSO-Cox regression. Individual normalized gene expression levels, weighted by their corresponding multiple-cox coefficients, were utilized to develop a UPR-related score formula. The external validation cohort was then employed to verify the stability of this prognostic formula.The log-rank test was used to assess survival differences between two groups. The "timeROC" package was used to compare the predictive accuracy of prognostic formula.

Develop a multivariate nomogram

Univariate cox regression analysis was performed to screen the overall survival (OS) associated UPR-related genes involved in score formula. 6 genes met the threshold of p-value > 0.05. Next, the multivariate cox regression analysis of the same genes with clinical characteristics identified independent factors associated with stomach cancer prognosis. Following a multivariate cox analysis, genes or clinical characteristics with p-values less than 0.05 were considered as independent prognostic factors. Using the nomogram function provided in the ‘rms’ library, a nomogram was created in R software.

Key gene expression analyses in pan-cancer

Pan-cancer analyses was performed on the gene with the highest correlation coefficient which was considered as the key gene. The differential expression of this gene in other pan-cancer tissues and the adjacent normal tissue, and the association of expression of this gene with ESTIMATE score and Stemness genes in tumors were analyzed for investigation of this gene's role in cancer.

Cells culture

The immortalized gastric epithelial cell line (GES-1) and two GC cell lines (MKN45 and AGS) were obtained from Sun Yat-sen memorial hospital, Sun Yat-sen University. Both GES-1 and MKN45 cells were cultured in RPMI-1640 medium with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin, while AGS cells were cultured in F12K medium with 10% FBS and 1% penicillin/streptomycin. The culture atmosphere is 37°C, 90% humidity, and 5% CO2.

Cell transfection

Before transfection, two stomach cancer cell lines, MKN45 and AGS, were transferred into 6-well plates. A combination of small interfering RNA (siRNA) (Ribobio, Guangzhou, China) mixed with Lipofectamine 2000 transfection reagent (Invitrogen, CA, USA) was used to transfect the tumor cells after they had reached a sufficient confluency. In the experiment, siRNAs that are validated were transfected into the cells, while nontarget control siRNAs were transfected into the control cells. The complete medium was added to the transfected cells after 6h of incubation with the siRNAs, and the cells were cultured for 48 h before being used for further experiments.

RNA isolation and RT-qPCR analysis

To extract total RNA, TRIzol reagent (Vazyme Biotech Co.,Ltd, Nanjing,China) was utilized. According to the manufacturer’s instructions, 1 ug cDNA was generated using the HiScript III first Strand cDNA Synthesis Kit (Vazyme Biotech Co.,Ltd, Nanjing,China). Then, we used the Quant Studio TMDx from Applied Biosystems as well as the ChamQ Universal SYBR qPCR Master Mix kit from Vazyme Biotech (Vazyme Biotech Co.,Ltd, Nanjing,China). Activation of the polymerase was performed at 95°C for 30 seconds, followed by 35 cycles at 95°C for 5 seconds and 60°C for 30 seconds.The 2^–ΔΔCt method was used to determine the relative expression of mRNA. Gene expression levels were normalized using β-actin.

Cell proliferation assay

96-well culture plates with 1.0 × 10³ cells per well were seeded with cells and cultured for 3 days. CCK8 reagent (Dojindo Laboratories, Kumamoto,Japan) was added to each well every 24 hours and incubated for 1 hour at 37°C to determine the viable cells. The microplate reader (Molecular Devices) was used to measure the absorbance at 450 nm. In addition, approximately 700 cancer cells were plated into six-well plates, and the cells were allowed to grow for approximately two weeks. The mixture of 4% paraformaldehyde and 0.5% crystal violet staining was performed for 15 minutes after two washes with phosphate-buffered saline (PBS). Cell clones were calculated for different groups.

Flow cytometry assay

The Annexin V-FITC/PI Cell Apoptosis Detection Kit (Beyotime Biotechnology, Shanghai, China) was used for flow cytometry. (FITC)-Annexin V and propidium iodide (PI) were used to stain the collected cells.BD FACSCalibur (BD, USA) was used to measure and analyze the apoptosis rate of cells.

Transwell migration and invasion assay

The invasion and migration experiment was conducted using a Boyden chamber containing transwell membrane filter inserts (Corning).To plant AGS cells, add 100 ml of serum-free F12K medium to the upper cavity, and to plant MKN45 cells, add 100 ml of serum-free RPMI-1640 medium.5×104 cells were injected into 24-well Transwell chambers (pore size 8 *m) containing complete medium containing 10% fetal bovine serum for migration assays or cells were injected into Matrigel-coated chambers for invasion assays. Invasion and migration assays were conducted after 24 and 48 hours of cell culture, respectively. The cells that penetrated the membrane were fixed with paraformaldehyde for 10 minutes, stained with 0.5% crystal violet, and then removed from the upper chamber. Three random 100 × fields of view were used to count the number of cells that penetrated the membrane under the microscope.

Statistical analysis

Comparing gene expression between different clusters was done using Wilcoxon rank-sum test. Kaplan-Meier analysis was used to compare survival rates among groups using the log-rank test. Cox regression analyses of univariate and multivariate models were conducted to identify independent predictor. A statistically significant P value was less than 0.05. All statistical analyses were performed with R software (Version 4.0.3).

Clinical characteristics of stomach cancer patients

A total of 375 stomach cancer patients from the TCGA cohort, 60 patients from the GSE13861 cohort, and 40 patients from the GSE28541 cohort were included in our study. The clinical characteristics of these patients are detailed in Supplementary Table 1. Additionally, we gathered 113 unfolded protein response genes. The study's flow chart is illustrated in Fig. 1.

Expression characteristics of UPR genes in stomach cancer

The expression and functional connections of 113 UPR related genes were displayed with protein-protein interaction (PPI) network, and according to the correlation expression network, there is a positive correlation between the expression of UPR-related genes, demonstrating the close correlation between their functions (Fig. 2A-B).The distribution of mutations of 133 UPR related genes in stomach cancer were exhibited in waterfall plot, it is estimated that EIF4G1, DDX10, and EIF2AK3, which have the highest mutation rate, have a mutation rate of 4%, and the mutation type is predominantly missense mutations (Fig. 2C). Among all types of mutations, the missense mutation was the one with the highest frenquency. Meanwhile, C > T substitution obtained the highest proportion in single nucleotide polymorphism (SNP) in stomach cancer patients. Copy number variations (CNV) analysis of 113 UPR related genes were also performed (Fig. 2D). Among them, Vascular endothelial growth factor A (VEGFA) had the highest CNV gain mutation, whereas C-C motif chemokine receptor 4-NOT transcription complex subunit 4 (CNOT4) presented the most obvious deletion mutation. Gene Ontology (GO) and KEGG pathways analysis revealed that 113 UPR related genes were significantly enriched in protein modification and transport and RNA degradation (Fig. 2E-F).

Identification of molecular subtypes of UPR genes in stomach cancer

In an unsupervised Cophenetic, Silhouette Indicator analysis, two subtypes of clusters were determined to be the optimal number based on UPR-related genes in TCGA cohort (Fig. 3A–C).

Based on the principal component analysis (PCA), it was concluded that the distribution of patients clustered according to UPR-related genes was significantly different between the two types of patients (Fig. 3D). Then, we showed that the expression of UPR-related genes in the two clusters were different by the heatmap (Fig. 3E). Furthermore, we also analyzed the prognostic relationship between the two molecular subtypes, with the results showing that Cluster 1 (C1) group had a less favorable over-all survival (OS) probability (log rank test P = 0.017), disease-free survival probability (DFS) (log rank test P = 0.018), progression-free survival probability (PFS) (log rank test P = 0.0063) and disease specific survival (DSS) probability (log rank test P = 0.028) probability than the Cluster 2 (C2) group (Fig. 3F-I).In light of the above results, it is evident that, by clustering patients based on UPR signatures, we are able to correctly classify stomach cancer patients into molecular types, and there is indeed a difference in prognosis between them.

Differentially analysis of GO and KEGG pathways

The DEGs between C1 and C2 molecular subtypes can be calculated via the "limma" package, in which 84 genes in C1 that are up-regulated and 57 genes in C1 molecular subtype that are down-regulated. Then, GO and KEGG analysis results showed that 84 up-regulated genes in the C1 molecular subtype were significantly enriched in some cancer progressed biological processes and pathways, such as TGF-β signaling pathway, NF-κB signaling pathway (Fig. 4A-B), in contrast, 57 down-regulated genes are enriched in the ferroptosis, adenosine 5‘-monophosphate (AMP)-activated protein kinase (AMPK) signaling pathway and so on (Fig. 4C-D).

Differentially analysis of immune landscape

We utilized multiple algorithms for comprehensive immune cell abundance assessment in order to clarify differences between the molecular subtypes of immune cell infiltration, including CIBERSORT, EPIC, MCPcounter, and xCell. Types of tumor-infiltrating immune cells were profiled by the heatmap (Fig. 5). The proportion of B cells, CD8-positive T cells, and macrophages were significantly higher in C1 molecular subtype with lower UPR-related genes expression than C2 molecular subtype. This result suggested that the unfolded protein response may affect the composition of the tumor immune microenvironment.

Comparative analysis of drug sensitivity

For the purpose of understanding the sensitivity of patients with C1 and C2 molecular subtypes to chemotherapy, we examined the sensitivity of chemotherapy drugs to two different groups of patients. The IC₅₀ for 5-fluorouracil, doxorubicin, docetaxel, and camptothecin was lower in patients with C2 molecular subtype, indicating an increased sensitivity to chemotherapy in these patients. (Fig. 6A-D). In conclusion, the patients in C2 molecular subtype are indeed more sensitive to chemotherapy than those in C1 molecular subtype. Previously, survival analyses have shown that patients in C2 molecular subtype have a better prognosis than patients in C1 molecular subtype, the UPR-related classifier distinguished different prognosis of patients with stomach cancer via the potential influence on the sensitivity of chemotherapy drugs. Consequently, UPR-related gene function could be correlated with treatment sensitivity in stomach cancer patients, and UPR-signature related typing could be used as a reference index for treatment selection in stomach cancer patients.

Prognostic signature construction and validation

In order to simplify the prognostic model, LASSO-Cox regression analysis was performed on UPR-related genes to establish UPR-related risk signature. Finally, 7 UPR-related genes were selected for prognostic model construction (Fig. 7A–B). The risk score was calculated as follows: score = 0.2082 × EIF4E + 0.0556 × EXOSC1–0.1725 × IARS1–0.4039 × IGFBP1 + 0.0814 × SPCS1 + 0.1412 × TSPYL2 + 0.0951 × TUBB2A. Patients were divided into two risk groups based on their median scores. Kaplan-Meier curves consistently showed the patients at high risk had a worse OS than those at low risk (log-rank P < 0.0001) (Fig. 7C). In the analysis of survival prediction using the prognostic model, time-dependent receiver operating characteristic (ROC) curves were produced, with 1-, 3-, 5-year area under curves (AUCs) reaching 0.709, 0.726, 0.781 (Fig. 7D). Based on the optimal cutoff value, patients in the GSE37023 cohort were also divided into high-risk and low-risk groups to assess the stability of the UPR-related signature. The high-risk group likewise demonstrated poorer overall survival (OS) compared to the low-risk group (log-rank P < 0.0001) (Fig. 7E). Likewise, the AUC of the predictive signature was 0.782 at 1 year, 0.730 at 3 years, and 0.709 at 5 years (Fig. 7F).

In addition, the practical utility of the UPR signature when validated using other survival data from TCGA cohort. We observed that the high-risk group had lower disease specific survival compared to the low-risk group (log-rank p < 0.0001) (Fig. 8A), with AUC values of 0.675, 0.747, and 0.767 at 1, 3, and 5 years, respectively (Fig. 8B). Similarly, the high-risk group exhibited poorer disease-free survival than the low-risk group (log-rank p < 0.0001) (Fig. 8C), with AUC values of 0.801, 0.743, and 0.784 at 1, 3, and 5 years (Fig. 8D). Lastly, the high-risk group displayed lower progression fee survival compared to the low-risk group (log-rank p < 0.0001) (Fig. 8E), with AUC values of 0.695, 0.660, and 0.692 at 1, 3, and 5 years (Fig. 8F). The clinical implications of these findings are significant. The TCGA stomach cancer prognostic UPR signature can be utilized as a valuable tool for risk stratification, enabling healthcare professionals to identify high-risk patients who may require closer monitoring, more aggressive treatment, or personalized therapeutic strategies. This risk assessment can lead to improved clinical decision-making and ultimately better patient outcomes. Furthermore, by understanding the underlying molecular mechanisms associated with different risk levels, researchers can potentially identify new therapeutic targets for the development of novel treatments for stomach cancer patients.

Differences in immune cell infiltration levels between the URPS groups

In order to better investigate the differences in immune cell infiltration between high and low-risk groups, we employed multiple immune assessment algorithms. Figure 9A illustrates the difference in immune cell infiltration levels between high-URPs and low-URPs groups. The majority of immune cells exhibit higher expression levels in the low-URPs group, particularly M1 Macrophages (P < 0.05), CD8 + T cells (P < 0.05), CD4 + memory T cells (P < 0.001), and CD8 + Tcm (P < 0.01). Furthermore, the heat map in Fig. 9B also reveals that the low-URPs group has increased immune cell expression. This finding suggests that patients in the low-risk group have stronger immune activity, which may contribute to a more favorable prognosis.

Independent Prognostic Value of Risk Models and construction of Nomogram

First, we analyzed the differences in various clinical factors (including age, gender, clinicopathological grade, clinicopathological stage, T stage, M stage, and N stage) between the low-URPs group and the high-URPs group in the TCGA cohort. We discovered that both stage and N stage were statistically different between the low-URPs and high-URPs groups (P < 0.05) (Fig. 10A). The composition of stage and N stage is illustrated in Figs. 10B and 10C. Next, to broaden the risk model's applicability, we performed validation for each stage of clinical factors, and the survival curve is presented in Fig. 11. Furthermore, univariate and multivariate Cox regression analyses were conducted on the TCGA cohort to evaluate the risk model's accuracy and determine if the risk score could serve as an independent prognostic factor for patient survival. Univariate Cox regression analysis revealed that age, stage, N stage, and URPs were significantly associated with patient prognosis (Fig. 12A). After adjusting for other confounding factors, multivariate analysis indicated that age, grade, and URPs were independent prognostic factors (Fig. 12B). Based on the TCGA cohort, we integrated clinical factors such as age, stage, N stage, and URPs to construct a nomogram to enhance survival prediction for patients with stomach cancer (Fig. 12C). Calibration plots for 1-, 3-, and 5-year overall survival (OS) demonstrated good agreement between nomogram predictions and actual observations (Fig. 12D). Clinically, this research offers valuable insights into the prognostic factors of stomach cancer, which can help clinicians make more informed decisions regarding treatment strategies. By presenting the model in the form of a nomogram and calibration plots, we provide a more intuitive and user-friendly tool that can be easily implemented in clinical practice to improve patient outcomes.

Pan-cancer analysis of IGFBP1

We have constructed a prognostic scoring model comprising seven genes, with IGFBP1 having the highest absolute coefficient value. Consequently, we consider IGFBP1 to be the key gene in our model. Furthermore, we aim to investigate the expression of this gene in other types of tumors and its relationship with the immune microenvironment. Then, pan-cancer analysis revealed that the expression of IGFBP1 in tumor tissues was significantly higher than that in normal tissues in TCGA-GBM, TCGA-LUAD, TCGA-COAD, TCGA-STES, TCGA-HMSC, and TCGA-THCA (Fig. 13A). Additionally, IGFBP1 expression was positively correlated with ImmuneScore, StromalScore, and ESTIMATEScore in TCGA-KIPAN, TCGA-MESO, and TCGA-PCPG (Fig. 13B). These findings demonstrate the potential role of IGFBP1 as a biomarker for tumor progression and the tumor microenvironment. Understanding the involvement of IGFBP1 in various cancers and its association with immune and stromal components may provide valuable insights for the development of novel therapeutic strategies, leading to improved patient outcomes.

IGFBP1 promotes the malignant phenotype of stomach cancer cells

We used three cell lines GES-1, AGS, and MKN45 to verify the expression of genes from UPR-related prognostic model. The expression of EIF4E, EXOSC1, IARS1, IGFBP1, and SPCS1 was found to be upregulated in stomach cancer cell lines, while TSPYL2 and TUBB2A showed downregulated expression (Fig. 14A-G). Supplementary Table 2 contains a list of all primer sequences. Notably, IGFBP1 demonstrated significant differences in expression among the stomach cancer cell lines. As a result, we further investigated the functionality of IGFBP1.

Then, using the stomach cancer cell lines MKN45 and AGS, we performed loss-of-function experiments and silenced expression of IGFBP1 to investigate its biological function. The efficiency of two siRNAs (siIGFBP1-1 and siIGFBP1-2) knockdown of IGFBP1 (Fig. 15A) were confirmed by RT-qPCR, and the sequences of both siRNAs are presented in Supplementary Table 3. In vitro, the knockdown of IGFBP1 resulted in a decrease of the proliferation of MKN45 and AGS cells as measured by the CCK8 assay (Fig. 15B) and clonal formation assay (Fig. 15C). The proportion of apoptotic cells increased significantly (MKN45, P < 0.01; AGS, P < 0.01) by knockdown of IGFBP1, indicating that IGFBP1 significantly decreased the antiapoptotic ability of the MKN45 and AGS cells (Fig. 15D).The knockdown of IGFBP1 significantly decreased migration and invasion of both MKN45 and AGS cells in the transwell migration and invasion assay (MKN45, P < 0.001; AGS, P < 0.001) (Fig. 15E). It has been demonstrated in vitro experiment that the IGFBP1 promotes the malignant phenotype of stomach cancer, including proliferation, invasion, and migration. In view of this, IGFBP1 is expected to become a new target for the treatment of stomach cancer.

Stomach cancer is characterized by a high incidence and mortality rate, posing a significant health challenge worldwide. Despite advances in medical research, stomach cancer prognosis remains difficult to predict. The development of reliable prognostic indicators for stomach cancer is crucial for improving patient outcomes and guiding treatment strategies. Research suggests that the unfolded protein response (UPR) may be one of the novel mechanisms involved in regulating stomach cancer progression[17].

Herein, in this study, we developed a classifier based on UPR-related genes, which effectively stratified stomach cancer patients into two distinct molecular subtypes. Further analysis revealed that these subtypes exhibited significant differences in prognosis, immune microenvironment infiltration, and chemosensitivity. Subsequently, we constructed a prognostic scoring formula using seven UPR-associated genes, which effectively predicted the prognosis of stomach cancer patients. Finally, we identified IGFBP1 as a crucial factor in the prognostic model and verified its role in promoting the malignant phenotype of stomach cancer cells through functional experiments.

The unfolded protein response (UPR) operates as an adaptive mechanism during cancer progression, modulating mechanisms that trigger cell transformation, enhance survival, and adjust metabolic status [18]. Previous research has proved that UPR-related IRE1/XBP1 signaling pathway plays a role in tumor progression, with high expression of XBP1s protein correlating with poor prognosis in glioblastoma, triple-negative breast cancer, and pre-B acute lymphoblastic leukemia [19–21]. Therefore, UPR is closely related to the progression of multiple tumors. In addition, studies have shown that GPCRs modulate UPR signaling via ERS sensors, IRE1α, PERK, and ATF6, to support cancer cell survival and inhibit cell death [22]. By regulating downstream signaling pathways such as NF-κB, PI3K/AKT, TGF-β and Wnt/β-catenin, GPCRs also upregulate mesenchymal transcription factors including Snail, ZEB, and Twist superfamilies which regulate cell polarity, cytoskeleton remodeling, migration, and invasion [22–23]. These pathways play a critical role in various pathological conditions in humans, including cancer, by regulating a number of key processes involved in tumor formation and progression [24–25]. Our study reveals that clustering stomach cancer patients based on UPR-related gene signatures effectively classifies them into distinct molecular subtypes with different prognoses. The analysis showed that patients in Cluster 1 molecular subtype had a less favorable overall survival, disease-free survival, progression-free survival, and disease-specific survival probability compared to those in Cluster 2 molecular subtype. Furthermore, differentially expressed genes between the two clusters were significantly enriched in cancer progression-related biological processes and pathways, such as the TGF-β and NF-κB signaling pathways as well. In accordance with previous research findings, our study demonstrated that the TGF-β and NF-κB pathways are indeed classical UPR pathways involved in tumor developmentIn accordance with previous research findings, our study confirms that the TGF-β and NF-κB pathways are indeed classical UPR pathways involved in stomach cancer development.

Prognostic signature constructed in this study consisted of 7 UPR-related genes, among which IGFBP1 showed the highest positive correlation coefficient. IGFBP1 as a protein prolonging the half-life of the insulin-like growth factors (IGFs) by combination, is induced during ER stress via activating transcription factor 4 (ATF4) expression[26].

Increased secretion of IGFBP1 helps cells modulate cell metabolism and maintain adaptive response under ER stress. IGFBP1 may inhibit IGFs activity by attenuating the binding of IGFs to their receptors[27], or potentiate IGFs function by facilitating the interaction of IGFs with receptors or binding to cell membranes through IGFBP1’s Arg-Gly-Asp sequence[28]. Post-translational modifications of IGFBP1 may affect the regulation of IGFBP1 on IGFs[29]. However, complex regulatory network among IGFs and IGFBPs is still not elucidated. Also, other IGF-independent effects of IGFBP1 were found to regulate cell migration and growth by interaction with extracellular or intracellular partners when faced with ER stress[30]. Taken together, this study we conducted in vitro found that IGFBP1 is capable of proliferating, invasive and migrating stomach cancer cells in a proliferation-like manner.

In summary, UPR-related gene classifier and risk signature were constructed for survival prediction of patients with stomach cancer. The patients in Cluster 1 molecular subtype tended to be characterized as short survival time, low chemotherapy sensitivities, and low tumor-killing immune cells infiltration. Next, we developed a robust prognostic prediction signature based on 7 UPR-related genes, and examined the relationship between this risk model and the prognosis of stomach cancer patients. As a key component of the UPR prognostic signature, IGFBP1 serves as an independent prognostic factor for poor stomach cancer outcomes. We have confirmed that IGFBP1 promotes the malignant phenotype of stomach cancer, including proliferation, anti-apoptosis, invasion, and migration. IGFBP1 is expected to become an important therapeutic target for stomach cancer.

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Availability of data and materials

The data of this study will be made available by the corresponding authors.

Competing interests

The authors declare no competing interests.

Author contributions

Drs Ouyang, Liu and Huang are co–first authors.

Concept and design: Drs Yao and Yu.

Data acquisition and correction, figure visualization and manuscript writing: Drs Ouyang and Liu.

Data bioinformatic and statistical analysis: All authors.

Validation in vitro experiments: Dr Liu.

Critical revision of the manuscript for important intellectual content: All authors.

Obtained funding: Drs Yao and Yu.

All coauthors have approved the final version of the manuscript.

Funding

This study was supported by grants 82273204, 81972471, 82003209 and 82073408 from the National Natural Science Foundation of China, grant 2023A1515012412 and 2023A1515011214 GuangDong Basic and Applied Basic Research Foundation, grant 2023A03J0722, 202206010078, 202102080072 and 202201020574 from the Guangzhou Science and Technology Project, grant 2018007 from the Sun Yat-Sen University Clinical Research 5010 Program, grant SYS-C-201801 from the Sun Yat-Sen Clinical Research Cultivating Program, grant A2020558 from the Guangdong Medical Science and Technology Program, grant 7670020025 from Tencent Charity Foundation, grant YXQH202209 from the Scientific Research Launch Project of Sun Yat-Sen Memorial Hospital, grant HL2021012 from the nursing research project of Sun Yat-Sen Memorial Hospital.

Acknowledgements

Not applicable.

Sung H, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209–49.
Chen W, et al. Cancer statistics in China, 2015. CA Cancer J Clin. 2016;66(2):115–32.
Fuchs CS, et al. Ramucirumab monotherapy for previously treated advanced gastric or gastro-oesophageal junction adenocarcinoma (REGARD): an international, randomised, multicentre, placebo-controlled, phase 3 trial. Lancet. 2014;383(9911):31–9.
Hetz C, Papa FR. The Unfolded Protein Response and Cell Fate Control. Mol Cell. 2018;69(2):169–81.
Bang YJ, et al. Trastuzumab in combination with chemotherapy versus chemotherapy alone for treatment of HER2-positive advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase 3, open-label, randomised controlled trial. Lancet. 2010;376(9742):687–97.
Senft D, Ronai ZA. UPR, autophagy, and mitochondria crosstalk underlies the ER stress response. Trends Biochem Sci. 2015 Mar;40(3):141–8.
Park SW, Ozcan U. Potential for therapeutic manipulation of the UPR in disease. Semin Immunopathol. 2013 May;35(3):351–73.
Ajoolabady A, Lebeaupin C, Wu NN, Kaufman RJ, Ren J. ER stress and inflammation crosstalk in obesity. Med Res Rev. 2023 Jan;43(1):5–30.
Tian Y, Merkwirth C, Dillin A, Mitochondrial UPR. A Double-Edged Sword. Trends Cell Biol. 2016 Aug;26(8):563–5.
Ko DK, Brandizzi F. Transcriptional competition shapes proteotoxic ER stress resolution. Nat Plants. 2022 May;8(5):481–90.
Spiotto MT, et al. Imaging the unfolded protein response in primary tumors reveals microenvironments with metabolic variations that predict tumor growth. Cancer Res. 2010;70(1):78–88.
Wan Q, Jin L, Wang Z. Comprehensive analysis of cancer hallmarks in cutaneous melanoma and identification of a novel unfolded protein response as a prognostic signature. Aging. 2020;12(20):20684–701.
Su Z, et al. An Unfolded Protein Response Related mRNA Signature Predicting the Survival and Therapeutic Effect of Hepatocellular Carcinoma. Comb Chem High Throughput Screen; 2022.
Zhang F, et al. An Unfolded Protein Response Related Signature Could Robustly Predict Survival Outcomes and Closely Correlate With Response to Immunotherapy and Chemotherapy in Bladder Cancer. Front Mol Biosci. 2021;8:780329.
Andruska N, et al. Anticipatory estrogen activation of the unfolded protein response is linked to cell proliferation and poor survival in estrogen receptor alpha-positive breast cancer. Oncogene. 2015;34(29):3760–9.
Zhang Q, et al. Characterization of an endoplasmic reticulum stress-related signature to evaluate immune features and predict prognosis in glioma. J Cell Mol Med. 2021;25(8):3870–84.
Brettingham-Moore KH, et al. Using gene expression profiling to predict response and prognosis in gastrointestinal cancers-the promise and the perils. Ann Surg Oncol. 2011;18(5):1484–91.
Su S, Shi YT, Chu Y, Jiang MZ, Wu N, Xu B, Zhou H, Lin JC, Jin YR, Li XF, Liang J. Sec62 promotes stomach cancer metastasis through mediating UPR-induced autophagy activation. Cell Mol Life Sci 2022 Feb 15;79(2):133.
Lin J, Liu H, Fukumoto T, Zundell J, Yan Q, Tang CA, Wu S, Zhou W, Guo D, Karakashev S, Hu CA, Sarma K, Kossenkov AV, Zhang R. Targeting the IRE1α/XBP1s pathway suppresses CARM1-expressing ovarian cancer. Nat Commun. 2021 Sep 7;12(1):5321.
Harnoss JM, Le Thomas A, Reichelt M, Guttman O, Wu TD, Marsters SA, Shemorry A, Lawrence DA, Kan D, Segal E, Merchant M, Totpal K, Crocker LM, Mesh K, Dohse M, Solon M, Modrusan Z, Rudolph J, Koeppen H, Walter P, Ashkenazi A. IRE1α Disruption in Triple-Negative Breast Cancer Cooperates with Antiangiogenic Therapy by Reversing ER Stress Adaptation and Remodeling the Tumor Microenvironment. Cancer Res. 2020 Jun 1;80(11):2368–2379.
Leung-Hagesteijn C, Erdmann N, Cheung G, Keats JJ, Stewart AK, Reece DE, Chung KC, Tiedemann RE. Xbp1s-negative tumor B cells and pre-plasmablasts mediate therapeutic proteasome inhibitor resistance in multiple myeloma. Cancer Cell 2013 Sep 9;24(3):289–304.
Kumari N, Reabroi S, North BJ. Unraveling the Molecular Nexus between GPCRs, ERS, and EMT. Mediators Inflamm 2021 Mar 2;2021:6655417.
Vázquez-Prado J, Bracho-Valdés I, Cervantes-Villagrana RD, Reyes-Cruz G. Gβγ Pathways in Cell Polarity and Migration Linked to Oncogenic GPCR Signaling: Potential Relevance in Tumor Microenvironment. Mol Pharmacol. 2016 Nov;90(5):573–86.
Chusri P, Kumthip K, Hong J, Zhu C, Duan X, Jilg N, Fusco DN, Brisac C, Schaefer EA, Cai D, Peng LF, Maneekarn N, Lin W, Chung RT. HCV induces transforming growth factor β1 through activation of endoplasmic reticulum stress and the unfolded protein response. Sci Rep. 2016 Mar 1;6:22487. doi: 10.1038/srep22487. PMID: 26927933; PMCID: PMC4772380.
Chusri P, Kumthip K, Hong J, Zhu C, Duan X, Jilg N, Fusco DN, Brisac C, Schaefer EA, Cai D, Peng LF, Maneekarn N, Lin W, Chung RT. HCV induces transforming growth factor β1 through activation of endoplasmic reticulum stress and the unfolded protein response. Sci Rep 2016 Mar 1;6:22487. doi: 10.1038/srep22487.
Rufo N, Garg AD, Agostinis P. The Unfolded Protein Response in Immunogenic Cell Death and Cancer Immunotherapy. Trends Cancer. 2017;3(9):643–58.
Marchand A, et al. Endoplasmic reticulum stress induction of insulin-like growth factor-binding protein-1 involves ATF4. J Biol Chem. 2006;281(28):19124–33.
Kajimura S, Aida K, Duan C. Insulin-like growth factor-binding protein-1 (IGFBP-1) mediates hypoxia-induced embryonic growth and developmental retardation. Proc Natl Acad Sci U S A. 2005;102(4):1240–5.
Fleming SD, Fazleabas AT, Bell SC. Differential regulation of insulin-like growth factor binding protein-1 and – 2 by insulin in the baboon (Papio anubis) endometrium. Reprod Biol Endocrinol. 2008;6:6.
Gupta MB. The role and regulation of IGFBP-1 phosphorylation in fetal growth restriction. J Cell Commun Signal. 2015;9(2):111–23.
Jones JI, et al. Insulin-like growth factor binding protein 1 stimulates cell migration and binds to the alpha 5 beta 1 integrin by means of its Arg-Gly-Asp sequence. Proc Natl Acad Sci U S A. 1993;90(22):10553–7.

No competing interests reported.

SupplementaryTable.docx

Download PDF

Journal Publication

published 02 May, 2024

Read the published version in Aging →

Version 1

posted

You are reading this latest preprint version

Unraveling the Unfolded Protein Response Signature: Implications for Tumor Immune Microenvironment Heterogeneity and Clinical Prognosis in Stomach Cancer

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Materials and Methods

The acquisition of clinical and gene expression data

Genotyping of unfolded protein response in stomach cancer

Analysis of functional enrichment

The immune landscape analysis

Chemotherapy benefit predicted by clusters

Developing and validating prognostic models of UPR-related signature

Develop a multivariate nomogram

Key gene expression analyses in pan-cancer

Cells culture

Cell transfection

RNA isolation and RT-qPCR analysis

Cell proliferation assay

Flow cytometry assay

Transwell migration and invasion assay

Statistical analysis

Results

Clinical characteristics of stomach cancer patients

Expression characteristics of UPR genes in stomach cancer

Identification of molecular subtypes of UPR genes in stomach cancer

Differentially analysis of GO and KEGG pathways

Differentially analysis of immune landscape

Comparative analysis of drug sensitivity

Prognostic signature construction and validation

Differences in immune cell infiltration levels between the URPS groups

Independent Prognostic Value of Risk Models and construction of Nomogram

Pan-cancer analysis of IGFBP1

IGFBP1 promotes the malignant phenotype of stomach cancer cells

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1