Identification of therapeutic targets in lung adenocarcinoma using Mendelian randomization and multi-omics

Abstract

Background: Lung adenocarcinoma (LUAD) remains associated with limited effective pharmacological treatment options. This study aimed to identify potential therapeutic targets for LUAD through the integration and analysis of multi-omics datasets.

Methods: A meta-analysis was conducted using two extensive proteomics datasets, the UK Biobank Proteomics Project (UKB-PPP) and the Fenland study, to identify disease-associated targets for LUAD through the Summary-Data-Based Mendelian Randomization method. Sensitivity analysis, including heterogeneity tests for dependent instruments, were conducted to validate the findings. The prognostic relevance of the identified candidate targets was assessed using transcriptomic data. Functional interactions were explored via protein-protein interaction network analysis, while single-cell analyses were employed to determine cell-specific expression patterns and differentiation trajectories. Potential side effects and therapeutic indications of these targets were evaluated using phenome-wide association studies and pharmacological data mining.

Results: Following meta-analysis, a primary significant target, intercellular adhesion molecule 5 (ICAM5), along with potential targets FUT8 and KLK13, were identified as therapeutic candidates for LUAD. FUT8 demonstrated a positive association with LUAD risk (OR = 1.02, p = 0.049), while ICAM5 (OR = 0.88, p = 0.002) and KLK13 (OR = 0.85, p = 0.021) exhibited negative associations. ICAM5 was further identified as an independent prognostic factor for patient survival (HR: 0.788, 95% CI: 0.663-0.936, p = 0.007) and revealed significant diagnostic and prognostic utility in LUAD. ICAM5 expression correlated with various immune infiltration patterns, suggesting potential modulation of the tumor immune microenvironment. Single-cell analysis revealed that ICAM5 did not directly impact LUAD cell differentiation, though its downstream target, MUC1, may contribute to differentiation processes, particularly in KRAS-mutated LUAD. Furthermore, phenome-wide association studies did not reveal substantial evidence of adverse phenotypes linked to ICAM5, supporting its safety profile for drug development.

Conclusion: ICAM5 emerges as a promising biological marker with significant prognostic and therapeutic potential in LUAD.

Keywords: Drug target; ICAM5; LUAD; Mendelian randomization; Prognosis.

Fig. 1

Expression of Candidate Targets. A SMR analysis depicting the pathogenic significance of candidate targets. B Differential expression of candidate targets in LUAD tissues compared to normal tissues. C Prognostic evaluation of candidate targets for LUAD. D Diagnostic performance of candidate targets, including Kaplan–Meier survival curve analysis. D–K The prognostic impact of significant candidate targets on LUAD. LUAD lung adenocarcinoma, SMR summary data-based Mendelian randomization, K-M Kaplan–Meier

Fig. 2

Predictive models of candidate targets and clinical evaluation. A Mulberry plot depicting the relationship between candidate targets and prognosis. B Nomogram integrating candidate targets and clinical features. C Calibration plot demonstrating the accuracy of the nomogram. D–F Restricted cubic spline analysis exploring the nonlinear associations of candidate targets with prognosis

Fig. 3

Clinical Association of ICAM5. A Differences in T stage expression of ICAM5. B Differences in N stage expression of ICAM5. C Differences in M stage expression of ICAM5. D Differences in TNM stage expression of ICAM5. E Variations in tumor location when compared to ICAM5 expression. F Relationship between ICAM5 expression and patient age. G Multivariate analysis evaluating ICAM5 as an independent prognostic factor. H Survival analysis of the predictive model. I Nomogram constructed using ICAM5 and clinicopathological features. J Calibration plot of the nomogram. K Time-dependent ROC curve analysis evaluating the predictive performance of the model

Fig. 4

Functional evaluation of ICAM5. A PPI network analysis depicting the interactions of ICAM5 with other proteins. B Friend analysis ranking candidate targets based on network topology parameters. C, D Structural domain analysis highlighting the protein domains of ICAM5. E GO and KEGG pathway analysis depicting the biological functions and pathways associated with ICAM5. F GSEA analysis revealing the involvement of ICAM5 in cell cycle regulation and immune responses. G DO analysis linking ICAM5 to specific diseases, including benign and malignant conditions

Fig. 5

Relationship between ICAM5 and immune infiltration. A xCell analysis depicting the relationship of ICAM5 with various immune cell types. B MCP-counter analysis showing immune infiltration patterns associated with ICAM5. C QuanTIseq analysis evaluating immune cell compositions correlated with ICAM5. D ESTIMATE analysis indicating the relationship between ICAM5 and Immune, Stromal, and ESTIMATE Scores. E CIBERSORT analysis detailing the correlation of ICAM5 with specific immune cell subtypes. F Correlation analysis revealing the associations of ICAM5 with multiple immune infiltration targets, including CD274

Fig. 6

Single-Cell analysis of candidate targets. A Classification of cell types based on single-cell RNA-seq data. B, C Expression patterns of candidate targets, including FUT8, ICAM5, KLK13, and MUC1, across different cell types. D, E Pseudotime analysis depicting differentiation trajectories of cells and the dynamic expression of candidate targets, particularly MUC1, during the progression from KACs to LUAD-KRAS mutant cells

Fig. 7

In vitro validation of the tumor-suppressive effect of ICAM5. A Relative mRNA levels of ICAM5 in HCC827 cells with knockdown of ICAM5 (shICAM5-1 and shICAM5-2) compared to the control (shCtrl). B Relative mRNA levels of ICAM5 in NCI-H1975 cells with knockdown of ICAM5 (shICAM5-1 and shICAM5-2) compared to the control (shCtrl). C, D The proliferation of cells with ICAM5 knockdown (shICAM5-1 and shICAM5-2) was assessed by CCK-8 assay over 5 days (left to right: HCC827, NCI-H1975, and two other cell lines). E, F The colony formation assay was used to assess the impact of ICAM5 knockdown on the colony formation ability of LUAD cells in vitro (left to right: HCC827, NCI-H1975, and two other cell lines). G, H Transwell assays were employed to evaluate the impact of ICAM5 knockdown on the in vitro migration ability of LUAD cells (G HCC827; H NCI-H1975)