Objective: This work aimed to verify the candidate biomarkers for keloid disorder (KD), and analyze the role of immune cell infiltration (ICI) in the pathology of keloid disorder.Methods: The keloid-related datasets (GSE44270 and GSE145725) were retrieved from the Gene Expression Omnibus (GEO). Then, differential expressed genes (DEGs) were identified by using the “limma” R package. Support vector machine-recursive feature elimination (SVM-RFE) and LASSO logistic regression were utilized for screening candidate biomarkers of KD. The receiver operating characteristic (ROC) curve was used to evaluate the diagnostic power of candidate biomarkers. The candidate biomarkers were further verified through qRT-PCR of keloid lesions and the matched healthy skin tissue collected from eight cases. In addition, ICI in keloid lesions was estimated through single-sample gene-set enrichment analysis (ssGSEA). Finally, the potential drugs to the treatment of KD were predicted in the Connectivity Map Database (CMAP).Results: A total of 406 DEGs were identified between keloid lesion and healthy skin samples. Among them, STC2 (AUC = 0.919), SDC4 (AUC = 0.970), DAAM1 (AUC = 0.966), and NOX4 (AUC = 0.949) were identified as potential biomarkers through the SVM-RFE, LASSO analysis and ROC analysis. The differential expressions of SDC4, DAAM1, and NOX4 were further verified in collected eight samples by qRT-PCR experiment. ICI analysis result showed a positive correlation of DAAM1 expression with monocytes and mast cells, SDC4 with effector memory CD4+ T cells, STC2 with T follicular helper cells, and NOX4 with central memory CD8+ T cells. Finally, a total of 13 candidate small molecule drugs were predicted for keloids treatment in CMAP drug database.Conclusion: We identified four genes that may serve as potential biomarkers for KD development and revealed that ICI might play a critical role in the pathogenesis of KD.

BackgroundMelanoma is among the most malignant immunologic tumor types and is associated with high mortality. However, a considerable number of melanoma patients cannot benefit from immunotherapy owing to individual differences. This study attempts to build a novel prediction model of melanoma that fully considers individual differences in the tumor microenvironment.MethodsAn immune-related risk score (IRRS) was constructed based on cutaneous melanoma data from The Cancer Genome Atlas (TCGA). Single-sample gene set enrichment analysis (ssGSEA) was used to calculate immune enrichment scores of 28 immune cell signatures. We performed pairwise comparisons to obtain scores for cell pairs based on the difference in the abundance of immune cells within each sample. The resulting cell pair scores, in the form of a matrix of relative values of immune cells, formed the core of the IRRS.ResultsThe area under the curve (AUC) for the IRRS was over 0.700, and when the IRRS was combined with clinical information, the AUC reached 0.785, 0.817, and 0.801 for the 1-, 3-, and 5-year survival, respectively. Differentially expressed genes between the two groups were enriched in staphylococcal infection and estrogen metabolism pathway. The low IRRS group showed a better immunotherapeutic response and exhibited more neoantigens, richer T-cell receptor and B-cell receptor diversity, and higher tumor mutation burden.ConclusionThe IRRS enables a good prediction of prognosis and immunotherapy effect, based on the difference in the relative abundance of different types of infiltrating immune cells, and could provide support for further research in melanoma.