Comprehensive analysis identifies AMIGO2 as a potential prognosis biomarker of pancreatic adenocarcinoma

Introduction

Pancreatic cancer is known for its high mortality rate, with a mere 8% 5-year survival rate (1). Pancreatic ductal adenocarcinoma (PDAC) is the most common pathological type of the pancreatic cancer (1). While surgical resection remains the primary choice of PDAC therapy, only a small percentage (10–15%) of newly diagnosed individuals meet the criteria. The surgical criteria for pancreatic cancer include localized tumor, no distant metastasis, good overall health, normal liver function, patient willingness, and typically involved patients under 75 years old with clinical stage II or lower (2,3). The majority of patients will eventually succumb to metastasis due to absence of alternative treatment options that could enhance the prognosis of PDAC patients (4,5). The reasons behind the high fatality rate of PDAC are numerous, primarily stemming from the difficulty in early detection (the diagnostic criteria for pancreatic cancer involve a combination of clinical symptoms, imaging studies, tumor marker assessments, and histopathological examinations), often occurring after distant metastasis has already taken place (6).

Genomic and transcriptomic investigation have provided insights into the underlying mechanisms of PDAC pathogenesis, highlighting the significance of gene mutations and aberrant signaling pathways. Notably, the KRAS driver mutation, present in over 85% of cases, along with frequent inactivation of tumor suppressors TP53, SMAD4, and CDKN2A (over 50%), have been identified as critical contributors (7,8). In terms of clinical management, systemic chemotherapy is the standard approach for most patients, irrespective of their eligibility for surgical intervention. Over the past decade, two novel combination regimens have emerged as the preferred first-line therapy for advanced PDAC. One such regimen, known as FOLFIRINOX (9), combines 5-fluorouracil (5-FU), leucovorin, irinotecan, and oxaliplatin. The combination of gemcitabine and albumin nanoparticle conjugate of palitaxel (nab-paclitaxel) represents the second treatment option for PDAC (9-11). Recently, a clinical trial demonstrated promising antitumour activity with liposomal irinotecan in combination with fluorouracil, leucovorin, and oxaliplatin (NALIRIFOX) in treatment-naive patients with metastatic PDAC (12). Therefore, it is imperative to identify a biomarker that can accurately predict the prognosis of PDAC.

The adhesion molecule with Ig like domain family 2 (AMIGO2), a cell adhesion molecule containing leucine-rich repeats (LRRs), is initially identified as a factor involved in neurite outgrowth (13,14). Recent studies have revealed that AMIGO2 may have a significant role in the development and progression of cancer. AMIGO2, which is regulated by BRD2/4, is upregulated in melanoma tissues and has been identified as a critical gene for melanoma survival (15). The knockdown of AMIGO2 results in alterations to the sphere-forming potential of ovarian cancer cells as well as a reduction in adhesion and invasion in vitro. Additionally, there is a significant decrease in intra-peritoneal metastasis (16). Furthermore, clinical samples have indicated that AMIGO2 serves as a prognostic factor for colorectal and gastric cancer (17,18). Moreover, AMIGO2 has been found to diminish the innate sensitivity of non-small cell lung cancer cells to cisplatin by inhibiting GSDME-mediated pyroptosis (19). Some researches havs shown that AMIGO2 may play a role in PDAC (20-24). Despite emerging studies providing evidence of the important role of AMIGO2 in various cancers, there is a lack of comprehensive analysis regarding its prognostic value and significant role in pancreatic adenocarcinoma (PAAD).

Our study evaluated the expression level and prognostic significance of AMIGO2 in pan-cancer and demonstrated that it was a potential tumor promoter in PAAD. We present this article in accordance with the TRIPOD reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-7/rc).

Methods

Data sources

The gene expression profile across all tumor samples and paired normal tissues to be analyzed by GEPIA (http://gepia.cancer-pku.cn/, accessed on February 2025) (25). The relative expression of AMIGO2 in Gene Expression Omnibus (GEO) datasets GSE71729 (26) (n=252, accessed on June 2024) was download from http://www.ncbi.nlm.nih.gov/geo. The Kaplan-Meier survival analysis for AMIGO2 messager RNA (mRNA) expression in The Cancer Genome Atlas (TCGA) and GEO [GSE62452 (27) (n=65) and GSE57495 (28) (n=63), accessed on December 2023] was analysed by PanCanSurvPlot (https://github.com/lanagarmire/PanCanSurvPlot, accessed on December 2023) (29). The clinical information of PAAD in the TCGA database was obtained from https://www.cancer.gov/ccg/research/genome-sequencing/tcga (accessed on December 2023). This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Genetic alteration of AMIGO2 in pancreatic cancer

Genetic alteration of AMIGO2 in pancreatic cancer from different cohorts was analyzed from the cBioPortal Database (https://www.cbioportal.org/, accessed on December 2023).

PAAD nomogram development and evaluation

A multivariate analysis incorporating factors significantly related to overall survival (OS) was performed using the R package “rms” 6.8.1 (https://cran.r-project.org/web/packages/rms/index.html, accessed on December 2023) (30). The calibration plot was used to determine discrepancies between actual and nomogram-predicted OS probabilities.

Enrichment analysis

The AMIGO2 expression related genes were identified using Pearson correlation analysis. The Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was performed using ClusterProfiler 4.10.0 (https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html, accessed on December 2023) (31) in R, and gene set enrichment analysis was conducted using GSEA v4.0 (https://www.gsea-msigdb.org/gsea/index.jsp, accessed on December 2023) at a significance threshold of false discovery rate (FDR) <0.25 and P<0.05. For gene set enrichment analysis (GSEA), the top genes with the strongest positive and negative correlations with AMIGO2 expression were used. Based on the median expression level of AMIGO2, PDAC samples from TCGA were stratified into two distinct cohorts: an AMIGO2-high expression group and an AMIGO2-low expression group.

Analysis of immune infiltration

The “Gene Set Variation Analysis (GSVA)” 1.50.0 package (https://bioconductor.org/packages/release/bioc/html/GSVA.html, accessed on December 2023) was utilized for single sample GSEA (ssGSEA) (32). Based on CIBERSORT, EPIC, MCPCOUNTER, QUANTISEQ, and XCELL algorithms in TIMER2.0 (https://cistrome.shinyapps.io/timer/, accessed on December 2023) (33), the association between AMIGO2 expression and immune cell infiltration levels were estimated.

Single-cell RNA (scRNA) sequencing analysis

A dataset of PAAD scRNA-seqs was acquired from the GEO database (GSE212966, accessed on December 2023) with a total of six of PDAC and six adjacent noncancerous resection specimens (34). Using the merge function in the Seurat R packages 5.1.0 (https://satijalab.org/seurat/, accessed on December 2023) (35), the datasets were integrated and clustered according to Uniform Manifold Approximation and Projection (UMAP) plots generated using the UMAP algorithm (https://arxiv.org/abs/1802.03426, accessed on December 2023) (36). The gene expression in specific cells was mapped using dot plots, which were annotated with metadata from the datasets.

Analyzing the drug sensitivity

The half maximal inhibitory concentration (IC50) was defined as an inhibitory concentration of 50%. Based on the TCGA dataset, IC50s for PAAD patients were calculated using “oncoPredict” 1.2 R package (https://github.com/pvreda/oncoPredict, accessed on December 2023) (37). The correlation between AMIGO2 gene expression and IC50s of drugs were analyzed by Pearson correlation analysis.

Cell lines

The Mia-Paca2, PATU8988, CFPAC-1, and SW1990 cell lines of human PAAD, along with HEK293T-17, were acquired from the American Type Culture Collection (ATCC). Mia-Paca2, PATU8988, SW1990 cells were cultured in DMEM (Gibco, California, USA) with 10% fetal bovine serum (FBS) (Lonsera, Suzhou, China) and 100 U/mL penicillin/streptomycin (Beyotime Biotechnology, Nanjing, China) at 37 ℃ and 5% CO₂. CFPAC-1 cells were cultured in IMDM (Gibco) with 10% FBS and 100 U/mL penicillin/streptomycin at 37 ℃ and 5% CO₂.

Plasmids

The oligo-containing AMIGO2-shRNA#1 and AMIGO2-shRNA#2 (sequence information from sigma) were sub-cloned into the EcoR I and Age I-digested pLKO.1 vector. AMIGO2-shRNA#1 sequence: GTCTTGACTTATCGTCCAATA, AMIGO2-shRNA#2 sequence: CGATGGATTTGTATGTTGGAA. DNA sequencing (Qingke, Hangzhou, China) verified the sequences of the plasmids generated.

Producing and infecting lentiviruses

To package lentiviruses, recombinant lentiviral plasmids harboring AMIGO2 shRNA were with co-transfected with packaging auxiliary vectors (VSVG, delta 8.9) into 293T cells. Puromycin screening was carried out to determine which cell lines were stably knocking down AMIGO2.

Immunoblot

SDS-PAGE was used to separate proteins, PVDF membranes (Millipore, Darmstadt, Germany) were used to transfer proteins, primary antibodies and horseradish peroxidase-conjugated secondary antibodies (Jackson Immunoresearch, Pennsylvania, USA) were used to immnuoblot the proteins. Finally, signal was detected by Enhanced Chemiluminescence. Rabbit antibodies against AMIGO2 (AF2080, 1:1,000) was purchased from R&D. Mouse antibodies against β-actin (3700, 1:5,000) were from Cell Signaling Technology (Danvers, MA, USA). Full and uncropped western blots are uploaded as the “supplement file” ( Figure S1).

Colony formation assay

Incubating CFPAC-1 cells for one week in a six-well tissue culture plate was followed by fixing them with 10% neutral formalin and staining them with 0.5% crystal violet solution. Staining dye was eluted from cells using 10% acetic acid solution. A Varioskan flash microplate reader was then used to measure the absorbance at 540 nm.

Statistical analysis

The Chi-squared test and Fisher’s exact test were used to determine the differences between categorical variables. Wilcoxon-signed-rank test was used to detect differences in AMIGO2 expression between normal and tumor tissues. Statistical differences between three groups were analyzed by one-way analysis of variance (ANOVA). Based on the Kaplan-Meier method and log-rank test, the survival distribution of the patients was determined. The Cox proportional hazard model was used for both univariate and multivariate analyses. P<0.05 was set for statistical significance.

Results

Pan-cancer analysis reveals AMIGO2 overexpression and prognostic significance

To systematically investigate the role of AMIGO2, we designed a multi-step analytical framework (Figure 1). Initial pan-cancer RNA-seq analysis using GEPIA-TCGA data revealed significant AMIGO2 upregulation in various types of cancers, including lymphoid neoplasm diffuse large B cell lymphoma (DLBC), bladder urothelial cancer (BLCA), lung adenocarcinoma (LUAD), PAAD, and other malignancies (Figure 2A). Subsequently, the impact of AMIGO2 expression on patient prognosis across multiple cancer types was investigated through survival analysis. Univariate Cox (UniCox) regression analysis of overall survival identified AMIGO2 as a high-risk factor (HR>1, P<0.05) in several cancer types. The findings from this analysis indicated that AMIGO2 served as a significant risk factor for glioma (GBMLGG) (HR =1.48, P=8.8e−24), brain lower grade glioma (LGG) (HR =1.42, P=7.9e−10), PAAD (HR =1.39, P=4.6e−4), and other cancer types (Figure 2B). The analysis of disease-free interval (DFI) revealed that AMIGO2 as a risk factor in various cancers, including PAAD (HR =1.72, P=2.5e−3), kidney renal papillary cell carcinoma (KIRP) (HR =1.26, P=0.01), stomach and esophageal carcinoma (STES) (HR =1.28, P=9.2e−3), and others (Figure 2C). These pan-cancer findings positioned PAAD as the cancer type with the most consistent AMIGO2-assoicated risk across survival endpoints.

Figure 1 Schematic design of the study. GEO, Gene Expression Omnibus; HPA, Human Protein Atlas; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Figure 2 Pan-cancer expression profile of AMIGO2. (A) The AMIGO2 expression profile across all tumor samples and paired normal tissues. (B) OS analysis of AMIGO2 OS by univariate Cox regression in TCGA pan-cancer. (C) DFI analysis of AMIGO2 in TCGA pan-cancer based on univariate Cox regression. AMIGO2, adhesion molecule with Ig like domain family 2; CI, confidence interval; DFI, disease free interval; OS, overall survival; TCGA, The Cancer Genome Atlas.

AMIGO2 overexpression predicts poor clinical outcomes in PAAD

Consistent with a previous report of elevated AMIGO2 expression in PAAD tissues (20), our analysis of GEO datasets demonstrated a significant elevation of AMIGO2 expression in tumor tissues compared to normal pancreas. Moreover, the expression of AMIGO2 in metastatic tissues exhibited a higher level compared to primary tissues (Figure 3A). To obtain additional confirmation, the protein level of AMIGO2 in PAAD was evaluated through immunohistochemical (IHC) utilizing the Human Protein Atlas (HPA). The results of the IHC staining revealed a robust expression of AMIGO2 in PAAD tissues, while its expression in normal pancreas was either absent or minimal (Figure 3B). Subsequently, we investigated the potential association between increased AMIGO2 expression and prognosis. We divided patients into high/low expression groups using the median value of AMIGO2 expression or the optimal cutoff calculated by the “surviminer” package. Notably, PAAD patients from high expression group demonstrated a significantly shorter overall survival in TCGA dataset (Figure 3C) and GEO validation datasets (Figure 3D,3E), respectively. In this study, it was observed that high expression of AMIGO2 was associated with shorter progression free interval (PFI) time (Figure 3F), disease specific survival (DSS) time (Figure 3G), and disease-free interval (DFI) time (Figure 3H). These findings collectively suggested that AMIGO2 expression level was elevated in patients with PAAD and was significantly correlated with poor prognosis.

Figure 3 AMIGO2 expression patterns and Kaplan-Meier survival analysis in PAAD. (A) Expression of AMIGO2 mRNA in normal pancreas, primary patients, and metastatic patients in the GSE71729 dataset. One-way ANOVA. *, P<0.05; ***, P<0.001. (B) Based on the HPA database (https://www.proteinatlas.org/), the IHC staining of AMIGO2 in normal pancreas and PAAD tissues. Image credit goes to the HPA. The links to the individual normal and tumor tissues of each protein are provided for AMIGO2 (https://www.proteinatlas.org/ENSG00000139211-AMIGO2/tissue/pancreas#img; https://www.proteinatlas.org/ENSG00000139211-AMIGO2/cancer/pancreatic+cancer#img). (C-E) The Kaplan-Meier survival analysis for AMIGO2 mRNA expression following optimal cutoff values in the TCGA (C), GSE62452 (D) and GSE57495 (E) datasets. (F-H) Correlation between AMIGO2 expression and PFI (F), DSS (G), DFI (H) times. AMIGO2, adhesion molecule with Ig like domain family 2; ANOVA, analysis of variance; DFI, disease free interval; DSS, disease specific survival; HPA, Human Protein Atlas; mRNA, messenger RNA; IHC, immunohistochemical; OS, overall survival; PAAD, pancreatic adenocarcinoma; PFI, progression-free interval.

Analysis of AMIGO2, nomogram construction and evaluation

To assess the independent relationship between AMIGO2 and other clinical characteristics in PAAD patients, both univariate and multivariate Cox regression analyses were conducted. The examination of TCGA data through univariate Cox regression analysis yielded significant findings, indicating that AMIGO2 expression (HR =1.363, P<0.001) and age (HR =1.023, P=0.04) were independent prognostic variables for patients with PAAD (Figure 4A).

Figure 4 An analysis of the prognosis of AMIGO2 in PAAD. (A) Using univariate and multivariate Cox regression, AMIGO2 in PAAD was evaluated as an independent prognostic factor. (B) The nomogram for predicting the OS of PAAD patients. (C,D) Calibration curves for OS of PAAD at 1-, 3-year intervals. AMIGO2, adhesion molecule with Ig like domain family 2; CI, confidence interval; Exp, expression; N, node; T, tumor; PAAD, pancreatic adenocarcinoma.

To assess the prognostic capacity of various factors on individual overall survival, a nomogram was developed, incorporating AMIGO2 and other clinical variables including age, T stage, and N stage. Subsequently, the nomogram was employed to determine the predictive impact on 1-, 3-, and 5-year overall survival outcomes in patients with PAAD (Figure 4B). The calibration chart provided evidence of the nomogram’s accuracy and effectiveness (Figure 4C,4D). Our investigation unveiled that AMIGO2 was a potential prognostic marker.

Structural and genomic characterization of AMIGO2 in PAAD

To elucidate the characteristics of AMIGO2, we initially examined the predicted three-dimensional structure of the AMIGO2 protein using AlphaFold (38) (Figure 5A). Then we utilized the STRING website (39) to identify potential protein interactions with AMIGO2 (Figure 5B). The data obtained indicated that TMEM178A, RGS14, SLC38A4, and other proteins may interact with AMIGO2. Subsequently, we conducted an analysis of the genomic alterations of AMIGO2 in pancreatic cancer. These alterations included amplification and deep deletion, occurring at a certain frequency in pancreatic cancer. Specifically, approximately 4% of acinar cell carcinoma of the pancreas exhibited mutations in AMIGO2 (Figure 5C), with several mutation sites identified, including E356G (Figure 5D). To elucidate the correlation between AMIGO2 and gene mutations in PAAD, we constructed a mutation landscape. The findings of our study illuminated that a substantial proportion of PAAD patients, specifically 72.6%, carried KRAS mutations, and an additional 60.3% possessed TP53 mutations (Figure 5E). Notably, our analysis further uncovered a compelling association between the expression levels of AMIGO2 and both KRAS missense mutations and TP53 mutations. PAAD patients with high AMIGO2 expression exhibited higher proportions of the KRAS and TP53 mutations (Figure 5E). These results provided insights into the characteristics of AMIGO2.

Figure 5 An overview of AMIGO2’s characteristics. (A) AlphaFold’ prediction of the structure of AMIGO2. (B) Proteins associated with AMIGO2 analyzed by the STRING website. (C) An analysis of the alteration frequency of AMIGO2 gene for the cases in several studies was conducted through cBioportal Database. (D) Map of AMIGO2 mutations. (E) Genomic alterations in the top genes in the two groups with high AMIGO2 expression and low expression analyzed by Sangerbox 3.0. Chi-square test. AMIGO2, adhesion molecule with Ig like domain family 2.

Functional enrichment analysis of AMIGO2 in PAAD

To further investigate the molecular characteristics of AMIGO2, we identified its expression-related genes and generated a volcano plot (Figure 6A). The results showed that high AMIGO2 expression was positively correlated with elevated expression levels of GRP87 (hypoxia-inducible chaperone linked to chemotherapy resistance), CYP27C1 (vitamin D hydroxylase with reported oncogenic functions), and PADI1, while a negative association was observed with GKAP1 (a GTPase regulator with tumor suppressor activity), SAT2, and SSR4 (Figure 6B,6C). The GSEA analyses conducted on the ranked co-expression genes revealed a predominant enrichment in various pathways, including “MAPK signaling pathway”, “p53 signaling pathway”, “T cell receptor signaling pathway”, “adherens junction”, “focal adhesion”, “cell cycle”, and “TGF-β signaling” (Figure 6D). Gene ontology analyses highlighted AMIGO2 was primarily associated with metabolic process, membrane functions, and protein binding (Figure 6E). These findings provided evidence for the significance of AMIGO2 and shed light on the potential molecular mechanisms underlying PAAD.

Figure 6 The enriched biological processes were identified through functional enrichment analysis. (A) AMIGO2 expression-related genes were identified on a volcano map by Pearson test (logFC>1, adj. P value <0.05). (B,C) The different expression gene based on AMIGO2 expression. The gene expression was positively correlated with AMIGO2 expression (C). The gene expression was negatively correlated with AMIGO2 expression. (D) KEGG functional enrichment analysis based on AMIGO2 at high and low expression levels. (E) The GO enrichment analyses based on AMIGO2 expression. All data were from the LinkedOmics Database, except for (D). AMIGO2, adhesion molecule with Ig like domain family 2; FC, fold change; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

The tumor microenvironment of PAAD and the correlation with AMIGO2

Given the critical role of tumor-immune interactions, an examination was conducted to explore the relationship between the expression of AMIGO2 and immune microenvironment. The analysis of immunofunctional data revealed that the low expression group exhibited heightened cytolytic activity, human leukocyte antigen (HLA) expression, and type II interferon (IFN) response in comparison to the high expression group (Figure 7A). Additionally, by using four algorithms (CIBERSORT, EPIC, MCPCOUNTER, and TIMER), diverse immune cells profiles were constructed from PAAD samples in order to investigate the correlation between the immune microenvironment and AMIGO2 expression. By integrating these algorithms, a substantial infiltration of cancer-associated fibroblast (CAF), macrophage, and a limited infiltration of B cells were observed in patients exhibiting high expression of AMIGO2 (Figure 7B). Chen et al. also revealed that AMIGO2 was upregulated in M2 macrophages, which also was verified by in vitro experiments (20). Conversely, high expression of AMIGO2 was found to be positively correlated with elevated expression of immune checkpoints, particularly CD274 (PD-L1), SIGLEC15, PDCDILG2 (PD-L2), and LAG3 (Figure 7C).

Figure 7 An analysis of the association between AMIGO2 expression and the tumor microenvironment. (A) ssGSEA scores of immune functions in two groups expressing high levels of AMIGO2 and those expressing low levels. Wilcoxon-signed-rank test. (B) The comparation of immune cell types between patients with high AMIGO2 expression and low AMIGO2 expression in different immune subtypes of PDAC. (C) mRNA levels of immune checkpoints CD274 (PD-L1), PDCDILG2 (PD-L2), CTLA4, LAG3, SIGLEC15, HAVACR2, TIGIT, and PDCD1 in those patients with high versus low AMIGO2 expression. Wilcoxon-signed-rank test. *, P<0.05; **, P<0.01; ***, P<0.001. PDAC, pancreatic ductal adenocarcinoma; AMIGO2, adhesion molecule with Ig like domain family 2; PD-L1, programmed death-ligand 1; ssGSEA, single sample gene set enrichment analysis.

Apart from infiltrating immune cells, the PAAD tumor microenvironment comprises various other cells and extracellular matrix (ECM) proteins. Through our analysis, we had identified 11 primary cell types present in this microenvironment, namely alpha cells, B cells, ductal cells, endothelial cells, fibroblasts, intermediate monocytes, macrophages, mast cells, peri-islet schwann cells, pancreatic stellate cells, and T cells (Figure 8A). The expression patterns of AMIGO2 in different cell types have been visually represented in Figure 8B. Additionally, we have conducted thorough investigation into the expression of AMIGO2 across various cell types (Figure 8C). In summary, our findings suggest that AMIGO2 was correlated with immune infiltration and may have the potential to be utilized as a novel biomarker for clinical interventions.

Figure 8 Single-cell type specificity of AMIGO2 mRNA. (A) The UMAP plot demonstrated main cell types in PDAC. (B) The UMAP plot showed AMIGO2 expression in main cell types. (C) Violin plots displaying the expression of AMIGO2 across the cell types identified in PDAC. AMIGO2, adhesion molecule with Ig like domain family 2; PDAC, pancreatic ductal adenocarcinoma; UMAP, Uniform Manifold Approximation and Projection.

AMIGO2 expression predicts therapeutic vulnerability to targeted agents

Given the observed correlation between AMIGO2 expression and unfavorable outcome, our study aimed to explore the association between AMIGO2 and anti-cancer drug sensitivity. Specifically, we found that patients with elevated AMIGO2 expression exhibited heightened sensitivity to BI-2536, a BRD4 inhibitor, thereby confirming AMIGO2 as a target of BRD4 (Figure 9A). Conversely, patients with high AMIGO2 demonstrated resistance to Cisplatin (Figure 9B), while the expression levels of AMIGO2 did not exhibit any significant correlation with other commonly utilized drugs in the treatment of PAAD, including 5-fluorouracil, olaparib, oxaliplatin, gemcitabine, paclitaxel, and docetaxel (Figure 9C-9H). These pharmacogenomic insights support AMIGO2-guided precision therapy.

Figure 9 The association between expression of AMIGO2 and drug sensitivity. (A-H) An association was found between the expression of AMIGO2 and drug sensitivity (IC50) of BI-2536, cisplatin, 5-fluorouracil, olaparib, oxaliplatin, gemcitabine, paclitaxel, docetaxel in the TCGA-PAAD. AMIGO2, adhesion molecule with Ig like domain family 2; IC50, half maximal inhibitory concentration; PAAD, pancreatic adenocarcinoma; TCGA, The Cancer Genome Atlas.

Biological significance of AMIGO2 in PAAD cells

In order to confirm the expression of AMIGO2, its levels were assessed in various PAAD cell lines. Initially, we conducted an analysis of the mRNA level of AMIGO2 in various PAAD cell lines obtained from GSE71729. Figure 10A illustrates that CFPAC-1 exhibited the highest AMIGO2 mRNA level among these cell lines. Two distinct short hairpin RNAs (shRNAs) were utilized to specifically reduce AMIGO2 expression in CFPAC-1 cells (Figure 10B). The results unambiguously demonstrated that inhibiting AMIGO2 expression significantly hindered the growth (Figure 10C) and colony formation (Figure 10D,10E) of CFPAC-1 cells. These outcomes offer compelling evidence for the role of AMIGO2 in the proliferation of PAAD cells.

Figure 10 The knockdown of AMIGO2 inhibited the growth of pancreatic cancer cells. (A) The AMIGO2 mRNA levels in GSE71729 datasets for different PAAD cell lines. (B) AMIGO2 expression in pancreatic cancer cell lines CFPAC-1 was knocked down with PLKO.1 lentivirus expressing control shRNA (csh) and two different AMIGO2 shRNAs (sh#1 and sh#2). (C) The knockdown of AMIGO2 inhibited the proliferation of CFPAC-1 cells. Cell proliferation was measured using the crystal violet assay. One-way ANOVA. (D,E) AMIGO2 knockdown decreased the number of colonies formed by CFPAC-1 cells. The representative image was shown on the left (D), and the quantitative chart was shown on the right (E). One-way ANOVA. Data are shown as mean ± SEM. *, P<0.05; ***, P<0.001. AMIGO2, adhesion molecule with Ig like domain family 2; ANOVA, analysis of variance; PAAD, pancreatic cancer; SEM, standard error of the mean.

Discussion

PAAD is widely recognized as a highly fatal malignancy. Despite notable progress in diagnostic and therapeutic approaches over the last ten years, the 5-year survival rate of pancreatic cancer remains alarmingly low, estimated at approximately 9% (1). A significant contributing factor to this grim prognosis is the absence of dependable biomarkers for accurate diagnosis. In our investigation, we successfully showcased the upregulation of AMIGO2 in PAAD, and further established a correlation between its heightened expression and shortened overall survival.

Genetically, PAAD is characterized by a limited number of recurring mutations in oncogenes and tumor suppressor genes, which have been linked to disease progression (7,8). These include four canonical mutations, namely KRAS (~85%), TP53 (60–70%), CDKN2A (>50%), and SMAD4 (~50%) (40-42). The remarkably high frequency of KRAS mutation in PAAD is comparable to that of BRAF mutation in melanoma. This implies that the Ras signaling pathway, known as a crucial oncogenic driver in the development of PDAC, represents a significant focus for the development of inhibitors. KRAS, in its canonical role, stimulates the MAPK signaling pathway through RAF-MEK-ERK. Intriguingly, our analysis revealed a correlation between high AMIGO2 expression and KRAS mutation in patients, with a notable enrichment in the “MAPK signaling pathway”. MAPKs contains JNK, ERK, so on. Giulia Benedetti et al. showed that AMIGO2 induction was ERK-dependent and on the contrary, inhibited by JNK (43). This suggests that KRAS activation may be involved in AMIGO2 expression on some extent.

On the other hand, our findings indicated a correlation between elevated AMIGO2 expression and TP53 mutation in patients (Figure 5E). Furthermore, GSEA analysis revealed an enrichment of “p53 signaling pathway” and “cell cycle” in these patients (Figure 6D). Published paper also shown that AMIGO2 silencing in melanoma cells and bladder cancer cells induces G1/S arrest (15,44). This suggests that AMIGO2 may play a role in regulating the cell cycle through the p53 pathway, thereby facilitating the proliferation of PAAD cells. In addition, our analysis demonstrated that the enrichment of “adherens junction” was associated with AMIGO2 expression upregulation (Figure 6D). Izutsu et al. also revealed that AMIGO2, encapsulated within extracellular vesicles derived from cancer cells, potentiates the adherence of liver endothelial cells to cancer cells (45). Future investigations should also consider the impact of AMIGO2 on KRAS-MAPK signaling and the p53 pathway.

Additionally, it is worth noting that higher cytolytic activity is indicative of enhanced anti-tumor immune function, as it is associated with immune infiltration. Our analysis revealed a correlation between high expression of AMIGO2 and decreased cytolytic activity and type II IFN response, suggesting that AMIGO2 over-expression may hinder immune function. Tumors in PADC often contain fibroblasts, which have the potential to regulate the progression of the disease. These fibroblasts are collectively referred to as CAFs and are believed to play a role in the development and advancement of PDAC, as well as the response to treatment (46-49). Pancreatic stellate cells are likely activated by TGF-β and other mediators to transform into CAFs in pancreatic cancer (46,49,50). Interestingly, a positive correlation was observed between high AMIGO2 expression and infiltration of CAFs in our study. Additionally, KEGG analysis indicated that patients with elevated AMIGO2 expression exhibited an enrichment in the “TGF-β signaling pathway”. CAFs possess the ability to secrete various cytokines and metabolites. We hypothesize that the up-regulation of AMIGO2 in PAAD activates the TGF-β signaling pathway, leading to the transformation of pancreatic stellate cells into CAFs. This process subsequently inhibits immune function and promotes the progression and metastasis of PAAD. Moreover, a positive correlation was observed between the expression levels of AMIGO2 and CD274, PDCDILG2, SIGLEC2, and HAVCAR2. Programmed death-ligand 1 (PD-L1), the ligand that pairs with programmed cell death protein 1 (PD-1), is found on the surface of some immune cells and certain non-immune cells. When PD-1 and PD-L1 bind, they send a signal to the immune system to reduce its activity, allowing cancer cells to evade immune surveillance and proliferate. Our analysis indicated that patients with PAAD who exhibit high AMIGO2 expression may potentially benefit from anti-PD-1 therapy.

The current treatment approach for unresectable pancreatic cancer primarily relies on chemotherapeutic agents such as gemcitabine and 5-FU. However, the median survival rate with these therapies remains low, typically not exceeding 12–18 months. These findings of our study indicated that PAAD patients with elevated AMIGO2 expression exhibited sensitivity to BI-2536, a BRD4 inhibitor, which is in line with the regulatory role of BRD4 in AMIGO2 expression (15). This suggests that BI-2536 may hold therapeutic potential for PAAD patients with high AMIGO2 expression. Conversely, our results demonstrated that PAAD patients with high AMIGO2 expression displayed resistance to Cisplatin, suggesting that Cisplatin may not be an appropriate treatment option for these individuals.

Despite a comprehensive analysis was conducted to underscore the prognostic importance and potential functional roles of AMIGO2 in PAAD, several critical issues still necessitate attention. Our examination of AMIGO2’s role in pancreatic cancer cells has laid a foundation, yet the functional implications of AMIGO2 in vivo in pancreatic cancer require further exploration. Furthermore, although several studies have documented a connection between various signaling pathways and AMIGO2 expression (15,44,45), aligning with our findings, additional pathways such as NOTCH, TGF-β, and T-cell receptor signaling require thorough verification. Moreover, the relationship between AMIGO2 expression and the response to BRD4 inhibitors in pancreatic cancer cells should be rigorously validated. To fully elucidate the underlying mechanisms, further investigation is urgently warranted in future studies.

Conclusions

Through a comprehensive analysis of the transcriptome and prognostic significance of AMIGO2 in pan-cancer, we have demonstrated that there is an increase in AMIGO2 expression in PAAD, which is associated with an adverse prognosis. These findings suggest that AMIGO2 may serve as a promising biomarker for PAAD.

Acknowledgments

We would like to thank the participants who provided data in the TCGA and GEO databases.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-7/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-7/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-7/prf

Funding: This study was supported by the Natural Science Foundation of Zhejiang Province, China (LQ23H160012), and Wenzhou Science and Technology Bureau of China (Y2023937).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-7/coif). The authors have no conflicts of interest to declare

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Heller M, Mann DA, Katona BW. Current Approaches of Pancreatic Cancer Surveillance in High-Risk Individuals. J Gastrointest Cancer 2025;56:61. [Crossref] [PubMed]
3. Mohamed G, Munir M, Rai A, et al. Pancreatic Cancer: Screening and Early Detection. Gastroenterol Clin North Am 2025;54:205-21. [Crossref] [PubMed]
4. Makohon-Moore A, Iacobuzio-Donahue CA. Pancreatic cancer biology and genetics from an evolutionary perspective. Nat Rev Cancer 2016;16:553-65. [Crossref] [PubMed]
5. Castellanos E, Berlin J, Cardin DB. Current treatment options for pancreatic carcinoma. Curr Oncol Rep 2011;13:195-205. [Crossref] [PubMed]
6. Kleeff J, Korc M, Apte M, et al. Pancreatic cancer. Nat Rev Dis Primers 2016;2:16022. [Crossref] [PubMed]
7. Jones S, Zhang X, Parsons DW, et al. Core signaling pathways in human pancreatic cancers revealed by global genomic analyses. Science 2008;321:1801-6. [Crossref] [PubMed]
8. Ryan DP, Hong TS, Bardeesy N. Pancreatic adenocarcinoma. N Engl J Med 2014;371:1039-49. [Crossref] [PubMed]
9. Conroy T, Desseigne F, Ychou M, et al. FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med 2011;364:1817-25. [Crossref] [PubMed]
10. Von Hoff DD, Ervin T, Arena FP, et al. Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine. N Engl J Med 2013;369:1691-703. [Crossref] [PubMed]
11. Wang-Gillam A, Li CP, Bodoky G, et al. Nanoliposomal irinotecan with fluorouracil and folinic acid in metastatic pancreatic cancer after previous gemcitabine-based therapy (NAPOLI-1): a global, randomised, open-label, phase 3 trial. Lancet 2016;387:545-57. [Crossref] [PubMed]
12. Wainberg ZA, Melisi D, Macarulla T, et al. NALIRIFOX versus nab-paclitaxel and gemcitabine in treatment-naive patients with metastatic pancreatic ductal adenocarcinoma (NAPOLI 3): a randomised, open-label, phase 3 trial. Lancet 2023;402:1272-81. [Crossref] [PubMed]
13. Kuja-Panula J, Kiiltomäki M, Yamashiro T, et al. AMIGO, a transmembrane protein implicated in axon tract development, defines a novel protein family with leucine-rich repeats. J Cell Biol 2003;160:963-73. [Crossref] [PubMed]
14. Ono T, Sekino-Suzuki N, Kikkawa Y, et al. Alivin 1, a novel neuronal activity-dependent gene, inhibits apoptosis and promotes survival of cerebellar granule neurons. J Neurosci 2003;23:5887-96. [Crossref] [PubMed]
15. Fontanals-Cirera B, Hasson D, Vardabasso C, et al. Harnessing BET Inhibitor Sensitivity Reveals AMIGO2 as a Melanoma Survival Gene. Mol Cell 2017;68:731-744.e9. [Crossref] [PubMed]
16. Liu Y, Yang J, Shi Z, et al. In vivo selection of highly metastatic human ovarian cancer sublines reveals role for AMIGO2 in intra-peritoneal metastatic regulation. Cancer Lett 2021;503:163-73. [Crossref] [PubMed]
17. Tanio A, Saito H, Amisaki M, et al. AMIGO2 as a novel indicator of liver metastasis in patients with colorectal cancer. Oncol Lett 2021;21:278. [Crossref] [PubMed]
18. Goto K, Morimoto M, Osaki M, et al. The impact of AMIGO2 on prognosis and hepatic metastasis in gastric cancer patients. BMC Cancer 2022;22:280. [Crossref] [PubMed]
19. Chen LK, Lin SP, Xie YH, et al. AMIGO2 attenuates innate cisplatin sensitivity by suppression of GSDME-conferred pyroptosis in non-small cell lung cancer. J Cell Mol Med 2023;27:2412-23. [Crossref] [PubMed]
20. Chen S, Du W, Feng K, et al. AMIGO2 is a pivotal therapeutic target related to M2 polarization of macrophages in pancreatic ductal adenocarcinoma. Aging (Albany NY) 2024;16:1111-27. [Crossref] [PubMed]
21. Yamaguchi N, Wu YG, Ravetch E, et al. A Targetable Secreted Neural Protein Drives Pancreatic Cancer Metastatic Colonization and HIF1α Nuclear Retention. Cancer Discov 2024;14:2489-508. [Crossref] [PubMed]
22. Nakamura T, Kuwai T, Kitadai Y, et al. Zonal heterogeneity for gene expression in human pancreatic carcinoma. Cancer Res 2007;67:7597-604. [Crossref] [PubMed]
23. ZhangJFuX.High expression of amigo2 is an independent predictor of poor prognosis in pancreatic cancer.Research Square (Preprint). 2021. DOI:.10.21203/rs.3.rs-557347/v1
24. S. M. Differential expression of AMIGO2 in pancreatic cancer.OSF Preprints. 2023.
25. Tang Z, Li C, Kang B, et al. GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res 2017;45:W98-W102. [Crossref] [PubMed]
26. Moffitt RA, Marayati R, Flate EL, et al. Virtual microdissection identifies distinct tumor- and stroma-specific subtypes of pancreatic ductal adenocarcinoma. Nat Genet 2015;47:1168-78. [Crossref] [PubMed]
27. Yang S, He P, Wang J, et al. A Novel MIF Signaling Pathway Drives the Malignant Character of Pancreatic Cancer by Targeting NR3C2. Cancer Res 2016;76:3838-50. [Crossref] [PubMed]
28. Chen DT, Davis-Yadley AH, Huang PY, et al. Prognostic Fifteen-Gene Signature for Early Stage Pancreatic Ductal Adenocarcinoma. PLoS One 2015;10:e0133562. [Crossref] [PubMed]
29. LinAYangHShiYPanCanSurvPlot: A Large-scale Pan-cancer Survival Analysis Web Application.BioRxiv. 2022. doi: 10.1101/2022.12.25.521884
30. Harrell FE Jr. Regression Modeling Strategies. Springer Series in Statistics; 2015.
31. Yu G, Wang LG, Han Y, et al. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012;16:284-7. [Crossref] [PubMed]
32. Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics 2013;14:7. [Crossref] [PubMed]
33. Li T, Fu J, Zeng Z, et al. TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res 2020;48:W509. [Crossref] [PubMed]
34. Chen K, Wang Q, Liu X, et al. Immune profiling and prognostic model of pancreatic cancer using quantitative pathology and single-cell RNA sequencing. J Transl Med 2023;21:210. [Crossref] [PubMed]
35. Hao Y, Stuart T, Kowalski MH, et al. Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nat Biotechnol 2024;42:293-304. [Crossref] [PubMed]
36. McInnes L, Healy J, Melville J. Umap: Uniform manifold approximation and projection for dimension reduction. arXiv preprint. 2018. arXiv:1802.03426.
37. Maeser D, Gruener RF, Huang RS. oncoPredict: an R package for predicting in vivo or cancer patient drug response and biomarkers from cell line screening data. Brief Bioinform 2021;22:bbab260. [Crossref] [PubMed]
38. Abramson J, Adler J, Dunger J, et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 2024;630:493-500. [Crossref] [PubMed]
39. Szklarczyk D, Kirsch R, Koutrouli M, et al. The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res 2023;51:D638-46. [Crossref] [PubMed]
40. Cancer Genome Atlas Research Network. Electronic address: andrew_aguirre@dfci.harvard; . Integrated Genomic Characterization of Pancreatic Ductal Adenocarcinoma. Cancer Cell 2017;32:185-203.e13. [Crossref]
41. Waddell N, Pajic M, Patch AM, et al. Whole genomes redefine the mutational landscape of pancreatic cancer. Nature 2015;518:495-501. [Crossref] [PubMed]
42. Bailey P, Chang DK, Nones K, et al. Genomic analyses identify molecular subtypes of pancreatic cancer. Nature 2016;531:47-52. [Crossref] [PubMed]
43. Benedetti G, Bonaventura P, Lavocat F, et al. IL-17A and TNF-α Increase the Expression of the Antiapoptotic Adhesion Molecule Amigo-2 in Arthritis Synoviocytes. Front Immunol 2016;7:254. [Crossref] [PubMed]
44. Han D, Xiong B, Zhang X, et al. Knockdown of AMIGO2 suppresses proliferation and migration through regulating PPAR-γ in bladder cancer. Hereditas 2024;161:21. [Crossref] [PubMed]
45. Izutsu R, Osaki M, Nemoto H, et al. AMIGO2 contained in cancer cell-derived extracellular vesicles enhances the adhesion of liver endothelial cells to cancer cells. Sci Rep 2022;12:792. [Crossref] [PubMed]
46. Kalluri R. The biology and function of fibroblasts in cancer. Nat Rev Cancer 2016;16:582-98. [Crossref] [PubMed]
47. Quail DF, Joyce JA. Microenvironmental regulation of tumor progression and metastasis. Nat Med 2013;19:1423-37. [Crossref] [PubMed]
48. Hosein AN, Brekken RA, Maitra A. Pancreatic cancer stroma: an update on therapeutic targeting strategies. Nat Rev Gastroenterol Hepatol 2020;17:487-505. [Crossref] [PubMed]
49. Sahai E, Astsaturov I, Cukierman E, et al. A framework for advancing our understanding of cancer-associated fibroblasts. Nat Rev Cancer 2020;20:174-86. [Crossref] [PubMed]
50. Sherman MH. Stellate Cells in Tissue Repair, Inflammation, and Cancer. Annu Rev Cell Dev Biol 2018;34:333-55. [Crossref] [PubMed]