Int J Clin Oncol. 2023 Jan 6. doi: 10.1007/s10147-022-02285-8. Online ahead of print.
ABSTRACT
PURPOSE: Here, we investigated expression modules reflecting the reciprocal expression of the cancer microenvironment and immune response-related genes associated with poor prognosis in primary central nervous system lymphoma (PCNSL).
METHODS: Weighted gene coexpression network analysis revealed representative modules, including neurogenesis, immune response, anti-virus, microenvironment, gene expression and translation, extracellular matrix, morphogenesis, and cell adhesion in the transcriptome data of 31 PCNSL samples. RESULTS : Gene expression networks were also reflected by protein-protein interaction networks. In particular, some of the hub genes were highly expressed in patients with PCNSL with prognoses as follows: AQP4, SLC1A3, GFAP, CXCL9, CXCL10, GBP2, IFI6, OAS2, IFIT3, DCN, LRP1, and LUM with good prognosis; and STAT1, IFITM3, GZMB, ISG15, LY6E, TGFB1, PLAUR, MMP4, FTH1, PLAU, CSF3R, FGR, POSTN, CCR7, TAS1R3, small ribosomal subunit genes, and collagen type 1/3/4/6 genes with poor prognosis. Furthermore, prognosis prediction formulae were constructed using the Cox proportional-hazards regression model, which demonstrated that the IP-10 receptor gene CXCR3 and type I interferon-induced protein gene IFI44L could predict patient survival in PCNSL.
CONCLUSION: These results indicate that the differential expression and balance of immune response and microenvironment genes may be required for PCNSL tumor growth or prognosis prediction, which would help understanding the mechanism of tumorigenesis and potential therapeutic targets in PCNSL.
PMID:36607476 | DOI:10.1007/s10147-022-02285-8
