Integrated pan-cancer analysis of ADM's role in prognosis, immune modulation and resistance

Abstract

Introduction: Adrenomedullin (ADM), a multifunctional peptide, has been implicated in various inflammatory and autoimmune diseases. However, its role in cancer, particularly in NSCLC, remained under-explored. This called for a pan-cancer analysis of ADM, investigating its expression, genomic alterations, prognostic value, immune associations, and relations with drug sensitivity to provide insights into its potential as a therapeutic target and biomarker.

Methods: ADM expression data from normal and tumor tissues was retrieved and analyzed through HPA and Timer 2.0 online platforms. Genetic alterations, copy number variations (CNVs), and methylation patterns were analyzed using cBioPortal and GSCA platforms. The data for survival analysis was extracted from TCGA and GEO database and analyzed through GEPIA and PrognoScan online platforms. ADM's correlations with immune checkpoint genes, immune cell infiltration, MSI, and TMB were evaluated using data from Timer and TCGA via R. Drug sensitivity analysis was performed with GDSC and CTRP databases, supported by network visualizations. IHC staining was conducted on LUAD patients' samples to assess ADM's relationship with EGFR-TKI resistance and immune microenvironment.

Results: ADM was widely expressed across normal tissues, with high levels in adipose tissue, endocrine organs, digestive and reproductive systems. Pan-cancer analysis revealed that ADM expression was upregulated in multiple cancer types, including CESC, ESCA, GBM, HNSC, KICH, KIRC, LUSC, PCPG, THCA, and UCEC, and correlated with advanced pathological stages in THCA, KIRP, and HNSC. Furthermore, high ADM expression was significantly linked to poor prognosis in patients with LGG, LUAD, MESO, THYM, LIHC, HNSC, GBM, KICH, KIRP, CESC, PAAD, and STAD, while its negative influence on OS and RFS was validated in LUAD. In addition, ADM exhibited genetic alterations, including amplification and deep deletion across multiple cancer types. Strong and consistent positive correlations were witnessed between ADM and several immune checkpoint genes, including CD274 (PD-L1), CD276, TNFRSF18, TNFSF9, and PVR in pan-cancer analysis, indicating its role in the development of suppressive immune microenvironment and T cell exhaustion. Besides, ADM showed significant correlations with immune cell infiltration, and TMB/MSI, highlighting its role in immune regulation and its potential as a predictive biomarker for immunotherapy. Significantly, ADM expression was correlated with multiple drug sensitivity, particularly chemotherapy and tyrosine kinase inhibitors (TKIs) therapy. Moreover, positive correlations between its expression and EGFR-TKI resistance, CD8⁺ T cell infiltration and tumor proportion score (TPS) in LUAD were validated in patients' samples, emphasizing its potential in guiding personalized therapy.

Discussion: This pan-cancer analysis revealed ADM's pivotal role in progression, immune modulation, and therapeutic response, especially in LUAD. ADM held promise as a prognostic biomarker and a potential therapeutic target in immune modulation and resistance management. Future research should focus on experimental validation and elucidation of ADM-mediated pathways, which might provide novel insights into cancer biology and improve clinical outcomes.

Keywords: adrenomedullin (ADM); immune modulation; lung adenocarcinoma (LUAD); pan-cancer analysis; resistance.

Figure 1

Expression profile of ADM in normal human tissues. (A) mRNA expression of ADM across various normal human tissues. (B) Protein expression of ADM in normal human tissues. (C–F) Representative IHC staining showing ADM expression in normal cerebellum, lung, stomach, and liver. The ADM expression regions were marked with arrows.

Figure 2

(A) ADM mRNA expression in tumor and normal tissues based on data from TCGA and GTEx (n = 17,672). (B) ADM expression in tumors and their paired adjacent normal tissues from TCGA (n = 1,394). (C) Analysis of the association between ADM expression and cancer pathological stages in THCA, PAAD, LUAD, LIHC, KIRP, and HNSC was performed using the GEPIA database. *p<0.05, **p<0.01, ***p<0.001; ns, not significant.

Figure 3

Pan-cancer mutational landscape of ADM. (A) The frequency of genetic alterations in ADM across a range of cancer types, representing 0.8% of samples(altered/profiled = 0.8% of 10,967 samples). (B) The mutation spectrum grouped by ADM in LUAD with the top 20 most frequently mutated genes presented. (C) Stacked bar plot of ADM mutation frequency across various cancer types, with LGG, ACC, and LUAD exhibiting the highest mutation rates. (D) Analysis of the correlation between ADM putative copy-number alteration (CNA) and ADM expression in pan-cancer tissues.

Figure 4

Correlations of ADM expression with CNV and DNA methylation. (A) Correlation between CNV and ADM mRNA expression in multiple cancer types via GSCA database. (B) Correlation between DNA methylation and ADM mRNA expression in various cancer types via GSCA database. (C) Heatmap showing the correlations between ADM expression and three methyltransferases, including DNMT1, DNMT3A, and DNMT3B, across different cancer types.

Figure 5

Pan-cancer analysis revealed prognostic significance of ADM expression. (A) Forest plot illustrating the impact of high ADM expression on overall survival (OS) across 41 cancer types, with significant associations in LGG, LUAD, MESO, THYM, LIHC, HNSC, GBM, KICH, KIRP, CESC, PAAD, and STAD. (*p<0.05, **p<0.01, ***p<0.001) (B) Kaplan-Meier plots for OS, showing the significant prognostic value of elevating ADM expression in CESC, HNSC, LGG, LIHC, LUAD, and MESO. (C) Survival analysis from the PrognoScan database, including GSE31210, GSE3141 and GSE8894 datasets, confirming the negative prognostic impact of high ADM expression on overall survival and relapse free survival in LUAD.

Figure 6

Correlations between ADM expression and immune response markers. (A) The correlations between ADM expression and 33 immune-related genes across multiple cancer types (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001). (B) The relationship between ADM expression and immune cell infiltration across various cancer types for B cells, CD4⁺ T cells, CD8⁺ T cells, neutrophils, macrophages, and dendritic cells (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001). (C, D) The correlations of ADM expression with tumor mutational burden (TMB) and microsatellite instability (MSI) in various cancer types. (E) Representative IHC staining images showing PD-L1 (CD274) and CD8 expression in LUAD samples with low and high ADM expression. Scale bars, 100 µm. The PD-L1 and CD8 expression regions were marked with arrows. (F) Scatter plot and linear fitting of ADM expression and tumor proportion score (TPS) in LUAD tissues. (G) Scatter plot and linear fitting of ADM expression and CD8⁺ T cell infiltration in LUAD tissues.

Figure 7

Pan-cancer analysis of correlation between ADM and drug sensitivity. (A) Correlation between ADM/ADM2 expressions and drug sensitivity from the GDSC database. (B) Relationship between ADM/ADM2 expressions and drug sensitivity from CTRP database. (C) Network analysis visualizing the associations between ADM expression and sensitivity to various compounds in the GDSC database. (D) Network representation of ADM-associated drug sensitivity based on data from CTRP database. (E) Representative IHC staining images of ADM in LUAD tissues before and after EGFR-TKI resistance. Scale bars, 100 µm. The ADM expression regions were marked with arrows. (F) Semi-quantitative bar chart of ADM expression levels from the IHC images.

Figure 8

ADM-associated pathways and functions in LUAD. (A, B) Heatmaps showing the top 50 significant genes positively and negatively correlated with ADM expression in LUAD. (C) KEGG pathway enrichment analysis revealed that ADM-associated genes were enriched in pathways such as p53 signaling, HIF-1 signaling, TNF signaling, cell cycle, and DNA replication. (D) GO enrichment of ADM-associated genes highlighted processes relating to DNA repair, chromatin regulation, and cell cycle modulation. (E) Representative ADM-associated genes enrichment plots: p53 signaling, cell cycle, HIF-1 signaling and TNF signaling.