AUTHOR=Xu Zhu , Wang Qin , Zhang Yiyao , Li Xiaolan , Wang Mei , Zhang Yuhong , Pei Yaxin , Li Kezhen , Yang Man , Luo Liping , Wu Chuan , Wang Weidong TITLE=Exploiting tertiary lymphoid structures gene signature to evaluate tumor microenvironment infiltration and immunotherapy response in colorectal cancer JOURNAL=Frontiers in Oncology VOLUME=14 YEAR=2024 URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2024.1383096 DOI=10.3389/fonc.2024.1383096 ISSN=2234-943X ABSTRACT=Background

Tertiary lymphoid structures (TLS) is a particular component of tumor microenvironment (TME). However, its biological mechanisms in colorectal cancer (CRC) have not yet been understood. We desired to reveal the TLS gene signature in CRC and evaluate its role in prognosis and immunotherapy response.

Methods

The data was sourced from The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO) databases. Based on TLS-related genes (TRGs), the TLS related subclusters were identified through unsupervised clustering. The TME between subclusters were evaluated by CIBERSORT and xCell. Subsequently, developing a risk model and conducting external validation. Integrating risk score and clinical characteristics to create a comprehensive nomogram. Further analyses were conducted to screen TLS-related hub genes and explore the relationship between hub genes, TME, and biological processes, using random forest analysis, enrichment and variation analysis, and competing endogenous RNA (ceRNA) network analysis. Multiple immunofluorescence (mIF) and immunohistochemistry (IHC) were employed to characterize the existence of TLS and the expression of hub gene.

Results

Two subclusters that enriched or depleted in TLS were identified. The two subclusters had distinct prognoses, clinical characteristics, and tumor immune infiltration. We established a TLS-related prognostic risk model including 14 genes and validated its predictive power in two external datasets. The model’s AUC values for 1-, 3-, and 5-year overall survival (OS) were 0.704, 0.737, and 0.746. The low-risk group had a superior survival rate, more abundant infiltration of immune cells, lower tumor immune dysfunction and exclusion (TIDE) score, and exhibited better immunotherapy efficacy. In addition, we selected the top important features within the model: VSIG4, SELL and PRRX1. Enrichment analysis showed that the hub genes significantly affected signaling pathways related to TLS and tumor progression. The ceRNA network: PRRX1-miRNA (hsa-miR-20a-5p, hsa-miR-485–5p) -lncRNA has been discovered. Finally, IHC and mIF results confirmed that the expression level of PRRX1 was markedly elevated in the TLS- CRC group.

Conclusion

We conducted a study to thoroughly describe TLS gene signature in CRC. The TLS-related risk model was applicable for prognostic prediction and assessment of immunotherapy efficacy. The TLS-hub gene PRRX1, which had the potential to function as an immunomodulatory factor of TLS, could be a therapeutic target for CRC.