The heat shock protein DNAJB2 as a novel biomarker for essential thrombocythemia diagnosis associated with immune infiltration.

BACKGROUND AND OBJECTIVE

Essential thrombocythemia (ET) is a rare myeloproliferative malignancy which may lead to severe thrombohemorrhagic complications. The diagnosis of ET is primarily based on bone marrow morphology and exclusion of other possibilities of myeloproliferative neoplastic diseases; the lack of gene biomarkers fails to provide a prompt diagnosis of ET. Therefore, this study was designed to identify biomarkers for early ET diagnosis, especially that associated with immune cell infiltration, by using the Gene Expression Omnibus (GEO) database and machine-learning algorithms.

METHODS

Two publicly available gene expression profiles (GSE9827 and GSE123732) from the GEO database were used to identify the differentially expressed genes (DEGs) between bone marrow samples of ET patients and healthy individuals, and functional enrichment analyses were conducted. The least absolute shrinkage and selection operator (LASSO) regression model and Support Vector Machine-Recursive Feature Elimination (SVM-RFE) machine-learning algorithm were performed to select the candidate gene biomarker. The expression level and diagnostic effectiveness of the identified gene biomarker were further validated using GSE567 and GSE2006 datasets. The involvement of infiltrating immune cells and their correlations with the gene biomarker were examined using cell-type identification by estimating relative subsets of RNA transcripts (CIBERSORT) algorithm.

RESULTS

There were 105 DEGs identified between ET and healthy control samples. Disease Ontology (DO) analysis showed that the diseases enriched by those DEGs were mainly human cancers, neurological diseases and inflammation while Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis demonstrated that pathways related to immune responses were primarily involved. The heat shock protein, DNAJB2, was identified as the potential biomarker for ET diagnosis with high effectiveness, with the area under the receiver operating characteristic (ROC) curve (AUC) equals to 0.905 in the validation cohort. The expression level of DNAJB2 in ET samples was indeed significantly higher than that in healthy control ones. The immune cell infiltration analysis showed that DNAJB2 was positively correlated with CD8+ T cells in ET with the proportion significantly higher than that in normal controls.

CONCLUSION

The present study identified DNAJB2 as a novel diagnostic biomarker for ET with high effectiveness based on ET and normal samples from the GEO database, which provides new insights into predicting ET with accuracy and promptness in clinical practice.