Identification of Hub Genes and Pathways in Preinfusion Chimeric Antigen Receptor (CAR) T-cell Products Associated With Cytokine Release Syndrome

Abstract

Background: Chimeric antigen receptor (CAR) T-cell therapy has transformed cancer management over the past decades, offering new hope to many patients. However, its effectiveness is often limited due to cytokine release syndrome (CRS), a life-threatening inflammatory response. Despite its clinical relevance, the molecular mechanisms underlying CRS, specifically in CAR T-cell products, remain poorly understood. This study aims to identify hub genes and pathways in preinfusion CAR T-cell products associated with CRS development and evaluate their potential as therapeutic targets through drug-gene interaction analysis and immune cell correlation profiling.

Methods: We examined gene expression data from 43 preinfusion clusters of differentiation 22 of CAR T-cell samples (CD22+), sourced from the Gene Expression Omnibus dataset GSE200296. Using the linear models for microarray data package, we identified differences in gene expression and conducted enrichment analyses to explore relevant biological pathways, including Kyoto Encyclopedia of Genes and Genomes and Gene Ontology terms. We built protein-protein interaction networks using the Search Tool for Retrieval of Interacting Genes/Proteins database to understand how these genes interact and pinpointed central "hub genes" with Cytoscape and the cytoHubba plugin. Our findings were validated using the GeneCards database (Weizmann Institute of Science, Israel) and an independent CRS-related dataset (GSE164805). Additionally, we analyzed immune cell populations and explored potential drug-gene interactions.

Results: Our study identified 24 genes with changed expression levels: 16 were downregulated and eight were upregulated. We identified five hub genes, interleukin (IL)1B, IL15, CD276, NCR2, and CCL17, as key contributors in CRS, which were primarily implicated in immune-related pathways, including cytokine-cytokine receptor interactions, IL17 signaling, and TNF signaling. These genes were especially expressed in monocytes, macrophages, and dendritic cells, confirming that those immune cell types play a critical role in CRS development. Through drug-gene interaction analysis, we found prospective therapies, such as enoblituzumab (targeting CD276) and canakinumab (targeting IL1B), which might assist in reducing CRS severity.

Conclusion: The study highlights IL1B, IL15, CD276, NCR2, and CCL17 as key CRS genes in preinfusion CAR T-cell products. Their dysregulation activity may contribute to the increased inflammation noted in CRS, pointing to a loss of regulatory control. Bringing us closer to better patient outcomes, these findings not only suggest that these genes could serve as valuable biomarkers for predicting CRS but also open the way for the development of more precise treatments such as combining drugs such as enoblituzumab and canakinumab, which might assist in reducing CRS severity and making CAR T-cell therapy safer and more effective, ultimately improving patient lives.

Keywords: bioinformatics analysis; car t-cell therapy; cytokine release syndrome; enrichment pathways analysis; hub genes.

Figure 1. Enhanced volcano plot of DEGs from the GSE200296 dataset

Genes are classified based on specific thresholds (p < 0.05, |log₂ FC| > 0.585). Red dots indicate genes significant by both p value and log₂ FC, blue dots represent genes significant only by p value, green dots show genes significant only by log₂ FC, and gray dots represent nonsignificant genes DEG: differential expression gene; log₂ FC: log₂ fold change; NS: not significant

Figure 2. Enrichment analysis for downregulated genes. (A) The KEGG pathways for downregulated genes (red dot plot), with dot size representing the number of genes associated with each pathway and color intensity indicating the statistical significance (p value). Pathways with higher gene counts and greater significance are shown as larger and darker dots. (B) GO terms associated with downregulated genes. The bar graphs represent the counts of genes associated with each GO term, and the color gradient from red to blue indicates the significance level of the p value, with red representing the most significant terms and blue representing less significant terms

KEGG: Kyoto Encyclopedia of Genes and Genomes; IL: interleukin; NF kappa B: nuclear factor kappa B; TNF: tumor necrosis factor; MHC: major histocompatibility complex

Figure 3. Enrichment analysis for upregulated genes. (A) The KEGG pathways for upregulated genes (blue dot plot), with dot size representing the number of genes associated with each pathway and color intensity indicating the statistical significance (p value). Pathways with higher gene counts and greater significance are shown as larger and darker dots. (B) GO terms associated with upregulated genes. The bar graphs represent the counts of genes associated with each GO term, and the color gradient from red to blue indicates the significance level of the p value, with red representing the most significant terms and blue representing less significant terms

KEGG: Kyoto Encyclopedia of Genes and Genomes; GO: Gene Ontology; IgA: immunoglobulin A; CCR: chemokine receptor

Figure 4. The PPI network and identification of feature genes. (A) The PPI network was constructed using the STRING database. (B) The top 10 feature genes were ranked based on their centrality using the MCC algorithm in the cytoHubba plugin shortest path method. Node colors range from deep red (higher rank) to lighter shades, highlighting their relative importance within the network. (C) Overlap Venn diagram of the top-ranked genes identified using the Degree, MCC, and Betweenness centrality algorithms in cytoHubba

PPI: protein-protein interaction; MCC: maximal clique centrality

Figure 5. Validation of CRS hub genes. (A) Overlap between the six hub genes identified in CAR T-cell therapy and 9,810 CRS-associated genes from the GeneCards database, confirming their relevance to CRS pathogenesis. (B) Volcano plot of DEGs from the GSE164805 dataset, highlighting the expression patterns of the six hub genes (p < 0.05, |log2FC| > 0.585). Red dots indicate genes significant by both p value and log2 FC, blue represents genes significant only by p value, green indicates genes significant only by log2 FC, and gray represents nonsignificant genes

CAR: chimeric antigen receptor; CRS: cytokine release syndrome; DEGs: differential expression gene; log₂ FC: log fold change; NS: not significant

Figure 6. Functional enrichment analysis of hub genes. (A) Enriched pathways associated with the hub genes, including cytokine-cytokine receptor interaction, rheumatoid arthritis, and IL17 signaling pathway. (B) GO analysis of hub genes, highlighting molecular functions and biological processes. Bar graphs represent gene counts and -log10 (p value) for significant terms

IL17: interleukin 17; TNF: tumor necrosis factor; GO: Gene Ontology; CCR: chemokine receptor

Figure 7. Heatmap of hub gene expression across immune cell populations

The heatmap shows the expression levels of hub genes across different immune cell types. Expression levels are color-coded, with deep red indicating high gene-level expression and deep blue indicating low-level expression. The Y-axis represents the hub genes, while the X-axis represents the immune cell populations NK: natural killer; ILC: innate lymphoid cells; DC: dendritic cells; MF: macrophage; Mo: monocytes; BA: B cells; GN: granulocytes; Eo: eosinophils; MC: mast cells; Ep: erythroid progenitors

Figure 8. The drug-gene interaction network of hub genes

The network illustrates interactions between hub genes highlighted in yellow triangles and potential therapeutic drugs shown in green rectangles. Shared drugs that interact with multiple hub genes are highlighted in pink rectangles