Deciphering the Prognostic and Therapeutic Significance of Cell Cycle Regulator CENPF: A Potential Biomarker of Prognosis and Immune Microenvironment for Patients with Liposarcoma

Abstract

Liposarcoma (LPS) is one of the most common subtypes of sarcoma with a high recurrence rate. CENPF is a regulator of cell cycle, differential expression of which has been shown to be related with various cancers. However, the prognostic value of CENPF in LPS has not been deciphered yet. Using data from TCGA and GEO datasets, the expression difference of CENPF and its effects on the prognosis or immune infiltration of LPS patients were analyzed. As results show, CENPF was significantly upregulated in LPS compared to normal tissues. Survival curves illustrated that high CENPF expression was significantly associated with adverse prognosis. Univariate and multivariate analysis suggested that CENPF expression could be an independent risk factor for LPS. CENPF was closely related to chromosome segregation, microtubule binding and cell cycle. Immune infiltration analysis elucidated a negative correlation between CENPF expression and immune score. In conclusion, CENPF not only could be considered as a potential prognostic biomarker but also a potential malignant indicator of immune infiltration-related survival for LPS. The elevated expression of CENPF reveals an unfavorable prognostic outcome and worse immune score. Thus, therapeutically targeting CENPF combined with immunotherapy might be an attractive strategy for the treatment of LPS.

Keywords: CENPF; cell cycle; immune infiltration; liposarcoma; prognostic biomarker.

Figure 1

Schematic diagram of candidate genes screening for LPS prognostic biomarkers.

Figure 2

Expression features of CENPF in LPS. (A–D) Compared to normal tissues, CENPF expression was upregulated in the LPS of different histological types, including WDLPS, DDLPS, MRCLPS and PLPS. (E) Comparison of CENPF expression between WDLPS and DDLPS. (F) Comparison of CENPF expression in LPS patients with different survival status. (G,H) Sensitivity and specificity of ROC curves of CENPF expression in discriminating LPS or DDLPS from normal adipose tissue. (I) The discrimination ability of CENPF expression between WDLPS and DDLPS. LPS: liposarcoma; DDLPS: dedifferentiated liposarcoma; MRCLPS: myxoid (round cell) liposarcoma; PLPS: pleomorphic liposarcoma; WDLPS: well-differentiated liposarcoma.

Figure 3

Validation of predictive performances of CENPF for LPS by plotting Kaplan–Meier (KM) curves. (A) KM curves of OS for DDLPS patients classified by CENPF expression (data derived from TCGA dataset). (B) Impacts of CENPF expression on DRFS. Survival curves based on CENPF expression of patients suffering from LPS (data derived from GSE30929 dataset). Verification of prognostic ability of CENPF for DDLPS patients classified by different clinicopathological features: (C) age: ≥60; (D) gender: male; (E,F) new tumor event after initial treatment; (G) radiation therapy. LPS: liposarcoma; DDLPS: dedifferentiated liposarcoma; OS: overall survival; DRFS: distant recurrence-free survival.

Figure 4

Correlation analysis between CENPF and previously reported DEGs or biomarkers of LPS. (A) Correlation matrix was plotted by Pearson’s rank correlation test with circles representative of the significance of the correlation. Negative and positive correlation were indicated by blue and red boxes, respectively. Color intensity and absolute value of correlation coefficient were directly proportional to the correlation intensity. (B) Validation of the prognostic values of known DEGs or biomarkers of LPS relevant to CENPF (|correlation coefficient| > 0.4) by survival curves. OS: overall survival; DDLPS: dedifferentiated liposarcoma; DEGs: differentially expressed genes.

Figure 5

Gene ontology (GO) enrichment analysis of co-expressed genes of CENPF. (A) Co-expressed gene network of CENPF. (B) Protein–protein interaction (PPI) network of CENPF. The colors of lines displayed on the networks corresponded to the different interactions with CENPF. Enrichment map and gene-concept network of (C) biological process (BP), (D) cell component (CC) and (E) molecular function (MF). Each dot on the graph represents a group of genes, and the dot size denotes the number of these genes. A redder color indicates a more significant enrichment.

Figure 6

Gene set enrichment analysis (GSEA) of CENPF. (A) A summary of GSEA analysis. Patients were classified into two subgroups according to CENPF expression (upper 50% vs. lower 50%). (B) Heat map of top 50 DEGs enriched in CENPF high-expression and low-expression subgroup. (C,D) Gene set enriched in CENPF high- and low-expression subgroup, respectively.

Figure 7

CENPF-related immune infiltration analysis of LPS tumor microenvironment (TME). Composition analysis of immune cells infiltrated in LPS tissues: (A,B) the proportion of various infiltrating immune cells in TME of LPS. (C,D) Comparison of the abundance of infiltrating immune cells between CENPF high-expression and low-expression subgroup. Analysis of immune score in TME of LPS tissues: (E) the correlation between CENPF expression and immune score. (F) KM curve of OS for DDLPS patients categorized by the immune score (high vs. low) with the number of censored patients at each timepoint (numbers at risk) listed under it. The optimal cut-off of survival curves was determined by X-tile [34]. (G) Correlation analysis between CENPF expression and scores of immune cells with correlation coefficient marked on the matrix plot. Positive and negative correlations were indicated by red and blue boxes, respectively. Color intensity and absolute value of correlation coefficient were directly proportional to the correlation strength. OS: overall survival; DDLPS: dedifferentiated liposarcoma.

Figure 8

Correlation analysis between CENPF and previously identified therapeutic targets for LPS. (A) Correlation matrix diagram. (B–H) CENPF expression was significantly positively correlated with several previously identified therapeutic targets for LPS clinical treatment research.

Figure 9

Farnesylation of CENPF affects cell cycle progression. (A) Localization of CENPF at different phases of cell cycle. Farnesylation of CENPF occurs from S phase to prophase of mitosis, involving connection of farnesyl isoprene-like compounds to the cysteine thiol group of CAAX peptide motif, catalyzed by farnesyltransferase (FTase). (B) SCH66336 (lonafarnib) competitively inhibits the binding of FTase to CAAX peptide of CENPF, thereby repressing its farnesylation. This competitive inhibition leads to a significant reduction of active CENPF in nucleus at prophase of mitosis and early centromere. As a result, G2/M transition is delayed due to the dysfunction of CENPF, thus cell cycle progression is restrained. F: farnesyl isoprenoid; FTase: farnesyl transferase; KTs: kinetochores; CAAX: CAAX box peptide (C: cysteine; A: an aliphatic amino acid; X: methionine, threonine, serine or glutamine); SCH66336: also known as Lonafarnib, a FTase inhibitor.