Identification of disulfidptosis-associated genes and characterization of immune cell infiltration in thyroid carcinoma

Abstract

Objective: The primary objective of this study is to conduct a comprehensive screening and analysis of differentially expressed genes related to disulfidoptosis (DEDRGs) in thyroid carcinoma (THCA). This entails delving into the intricate characterization of immune cell infiltration within the THCA context and subsequently formulating and validating a novel prognostic model.

Method: To achieve our objectives, we first delineated two distinct subtypes of disulfidoptosis-related genes (DRGs) via consensus clustering methodology. Subsequently, employing the limma R package, we identified the DEDRGs critical for our investigation. These DEDRGs underwent meticulous validation across various databases, alongside an in-depth analysis of gene regulation. Employing functional enrichment techniques, we explored the potential molecular mechanisms underlying disulfidoptosis in THCA. Furthermore, we scrutinized the immune landscape within the two identified subtypes utilizing CIBERSORT and ESTIMATE algorithms. The construction of the prognostic model for THCA entailed intricate methodologies including univariate, multivariate Cox regression, and LASSO regression algorithms. The validity and efficacy of our prognostic model were corroborated through Kaplan-Meier survival curves and ROC curves. Additionally, a nomogram was meticulously formulated to facilitate the prediction of patient prognosis. To fortify our findings, we conducted a comprehensive Bayesian co-localization analysis coupled with rigorous in vitro experimentation, aimed at unequivocally establishing the validity of the identified DEDRGs.

Result: Our analyses unveiled Cluster C1, characterized by elevated expression levels of DEDRGs, as harboring a favorable prognosis accompanied by abundant immune cell infiltration. Correlation analyses underscored predominantly positive associations among the DEDRGs, further affirming their significance in THCA. Differential expression patterns of DEDRGs between tumor samples and normal tissues were evident across the GEPIA and HPA databases. Insights from the TIMER database underscored a robust correlation between DEDRGs and immune cell infiltration. KEGG analysis elucidated the enrichment of DEDRGs primarily in pivotal pathways including MAPK, PPAR signaling pathway, and Proteoglycans in cancer. Furthermore, analyses using CIBERSORT and ESTIMATE algorithms shed light on the crucial role played by DEDRGs in shaping the immune microenvironment. The prognostic model, anchored by five genes intricately associated with THCA prognosis, exhibited commendable predictive accuracy and was intricately linked to the tumor immune microenvironment. Notably, patients categorized with low-risk scores stood to potentially benefit more from immunotherapy. The validation of DEDRGs unequivocally underscores the protective role of INF2 in THCA.

Conclusion: In summary, our study delineates two discernible subtypes intricately associated with DRGs, revealing profound disparities in immune infiltration and survival prognosis within the THCA milieu. The implications of our findings extend to potential treatment strategies for THCA patients, which could entail targeted interventions directed towards DEDRGs and prognostic genes, thereby influencing disulfidptosis and the immune microenvironment. Moreover, the robust predictive capability demonstrated by our prognostic model, based on the five genes (ANGPTL7, FIRRE, ODAPH, PROKR1, SFRP5), underscores its potential clinical utility in guiding personalized therapeutic approaches for THCA patients.

Keywords: disulfidptosis; immune cell infiltration; prognostic model; thyroid carcinoma.

Figure 1

The protocol of our study procedure.

Figure 2

Landscape of DRGs. (A) PPI network of DRGs. (B) Heatmap of DRGs related expression profile.

Figure 3

Consensus clustering analysis. (A) Represents that when k = 2, the matrix heatmap was neatly classified. (B) Heatmap of DRGs-related expression profile in cluster C1 and C2 subtypes.

Figure 4

Survival analysis in the TCGA-THCA. (A) Kaplan-Meier of the two clusters. (B) Kaplan-Meier of THCA patients regrouped according to Gender. (C) Kaplan-Meier of THCA patients regrouped according to T. (D) Kaplan-Meier of THCA patients regrouped according to N. (E) Kaplan-Meier of THCA patients regrouped according to M. (F) Kaplan-Meier of THCA patients regrouped according to Stage.

Figure 5

Differential expression and correlation analysis. (A) Volcanic map of DEDRGs in TCGA-THCA. (B) Volcanic map of DEDRGs in GSE33630. (C) Venn diagram of co-DEDRGs. ^* represents P < 0.05, ^*** represents P < 0.001, ^**** represents P < 0.0001, - represents no significant difference.

Figure 6

IHC of the DEDRGs in the HPA database.

Figure 7

Specific spatial distribution of the DEDRGs in cells.

Figure 8

Gene regulatory network. (A, B) represents the gene regulatory network of the DEDRGs (FLNA, IQGAP1, MYH10, and ACTN1). (C, D) represents gene regulatory network of the DEDRGs (ZHX2, INF2, ME1, and PDLIM1).

Figure 9

GO and KEGG analysis. (A) Represents the biological process bubble diagram of the DEDRGs in TCGA-THCA. (B) Represents the cellular component bubble diagram of the DEDRGs in TCGA-THCA. (C) Represents the molecular function bubble diagram of the DEDRGs in TCGA-THCA. (D) Represents the KEGG bubble diagram of the DEDRGs in TCGA-THCA. The bubble size represents the number of hub genes enrichment, and the color depth represents the level of significance.

Figure 10

Immune cell infiltration analysis. (A) Stack diagram of immune cell infiltration. (B) Violin diagram of immune cell infiltration between cluster C1 and cluster C2. (C) Violin diagram of immune score between cluster C1 and cluster C2. (D) Heatmap of correlation analysis between immune cells. ^* represents P < 0.05, ^** represents P < 0.01, ^*** represents P < 0.001, ^**** represents P < 0.0001, - represents no significant difference.

Figure 11

Immune checkpoints analysis and mutation landscape. (A) Expression difference of CTLA4 in two subtypes. (B) Expression difference of LAG3 in two subtypes. (C) Expression difference of PD-L1 in two subtypes. (D) Expression difference of PD-1 in two subtypes. (E) Expression difference of TIGIT in two subtypes. (F) Expression difference of CD274 in two subtypes. (G, H) Heatmap of correlation analysis between immune checkpoints. (I) Waterfall map of the Mutation landscape. ^* represents P < 0.05, ^** represents P < 0.01, ^*** represents P < 0.001, ^**** represents P < 0.0001, - represents no significant difference.

Figure 12

Construction and verification of prognostic model. (A) LASSO coefficient map of the prognostic genes. (B) After ten cross-verifications, the LASSO model of parameter selection is adjusted with minimum absolute shrinkage and selection. (C) Forest plot of the prognostic genes. (D) Heatmap of risk score. (E) Kaplan-Meier of high and low-risk group. (F) Predictive value of Cox prognostic model in THCA patients evaluated by ROC curves. (G) Box diagram of the prognosis genes expression in the high and low-risk groups.

Figure 13

Establishment of a nomogram model. (A) Nomogram model for evaluating the prognosis of THCA patients. (B) Multivariate Cox regression. (C) Calibration curve of nomogram model. (D) Kaplan-Meier of the high and low-risk groups in the validation cohort. ^* represents P < 0.05, ^** represents P < 0.01, ^*** represents P < 0.001, ^**** represents P < 0.0001, - represents no significant difference.

Figure 14

Relationship between risk scores and immune cell infiltration. (A) CD8+ T cells; (B) Memory B cells; (C) Activated memory CD4+ T cells; (D) Activated NK cells; (E) M0 macrophages; (F) Neutrophils; (G) M1 macrophages; (H) M2 macrophages; (I) Naïve B cells.

Figure 15

Verification of DEDRGs by Bayesian co-localization and experiment. (A–G) Bayesian co-localization analysis. (H) mRNA relative expression of PDLIM1 and INF2 (^*P < 0.05). (I) Protein level of INF2 and PDLIM1 in the normal group and control group.