A Potential Prognostic Gene Signature Associated with p53-Dependent NTRK1 Activation and Increased Survival of Neuroblastoma Patients

Abstract

Neuroblastoma is the most common extracranial solid tumour in children, comprising close to 10% of childhood cancer-related deaths. We have demonstrated that activation of NTRK1 by TP53 repression of PTPN6 expression is significantly associated with favourable survival in neuroblastoma. The molecular mechanisms by which this activation elicits cell molecular changes need to be determined. This is critical to identify dependable biomarkers for the early detection and prognosis of tumours, and for the development of personalised treatment. In this investigation we have identified and validated a gene signature for the prognosis of neuroblastoma using genes differentially expressed upon activation of the NTRK1-PTPN6-TP53 module. A random survival forest model was used to construct a gene signature, which was then assessed across validation datasets using Kaplan-Meier analysis and ROC curves. The analysis demonstrated that high BASP1, CD9, DLG2, FNBP1, FRMD3, IL11RA, ISGF10, IQCE, KCNQ3, and TOX2, and low BSG/CD147, CCDC125, GABRB3, GNB2L1/RACK1 HAPLN4, HEBP2, and HSD17B12 expression was significantly associated with favourable patient event-free survival (EFS). The gene signature was associated with favourable tumour histology and NTRK1-PTPN6-TP53 module activation. Importantly, all genes were significantly associated with favourable EFS in an independent manner. Six of the signature genes, BSG/CD147, GNB2L1/RACK1, TXNDC5, FNPB1, B3GAT1, and IGSF10, play a role in cell differentiation. Our findings strongly suggest that the identified gene signature is a potential prognostic biomarker and therapeutic target for neuroblastoma patients and that it is associated with neuroblastoma cell differentiation through the activation of the NTRK1-PTPN6-TP53 module.

Keywords: NTRK1; biomarkers; gene signature; neuroblastoma; prognosis.

Figure 1

The prognostic score differentiates survival outcomes. The prognostic score was evaluated on the two validation datasets (GSE85047 and TARGET) using Kaplan–Meier plots and time-dependent ROC curves. The Kaplan–Meier plots for (A) GSE85047 and (B) TARGET compares five-year event-free survival curves for low-risk and high-risk cohorts stratified by the median risk score of GSE49711. Time-dependent ROC curve analysis for (C) GSE85047 and (D) TARGET illustrates the prognostic score’s ability to predict event-free survival in neuroblastoma patients for cohorts at one, three, and five years.

Figure 2

The prognostic score is significantly associated with prognostic markers. The prognostic score was measured for its association with four prognostic markers as follows: (A,B) MYCN amplification (C,D) diagnosis under the age of 18 months, (E) tumour histology, and (F) NTRK1-PTPN6-TP53 activation. The association was statistically significant for all results, using the Mann–Whitney test. GSE85047 is missing the data required to measure the association for tumour histology and NTRK1-PTPN6-TP53 activation.

Figure 3

Kaplan–Meier analyses of BSG/CD147, FNBP1, GNB2L1/RACK1, and HAPLN4 genes. Kaplan–Meier analysis showed that (A) BSG/CD147, (B) FNBP1, (C) GNB2L1/RACK1, and (D) HAPLN4 are the genes in the signature that show the best differentiation between low- and high-risk patients. For BSG/CD147, GNB2L1, and HAPLN4, low expression is associated with favourable EFS. Conversely, for FNBP1, high expression is associated with favourable EFS.

Figure 4

Gene signature expression validation by RTqPCR. RT-qPCR analysis of the sixteen-gene signature in human neuroblastoma (A) SH-SY5Y and (B) SK-N-AS cell lines, both expressing the NTRK1-PTPN6-TP53 module at 32 °C and lacking the expression of the module at 37 °C. The key genes from the sixteen-gene signature were also analysed in neuroblastoma murine xenografts (C) 9464D and 262226. Relative mRNA abundance was determined by the 2−∆∆CT method. Statistical significance was denoted as * = p < 0.05 (paired t-test). Error bars represent the standard error of the mean calculated from 2 or 3 independent experimental replicates.

Figure 4

Gene signature expression validation by RTqPCR. RT-qPCR analysis of the sixteen-gene signature in human neuroblastoma (A) SH-SY5Y and (B) SK-N-AS cell lines, both expressing the NTRK1-PTPN6-TP53 module at 32 °C and lacking the expression of the module at 37 °C. The key genes from the sixteen-gene signature were also analysed in neuroblastoma murine xenografts (C) 9464D and 262226. Relative mRNA abundance was determined by the 2−∆∆CT method. Statistical significance was denoted as * = p < 0.05 (paired t-test). Error bars represent the standard error of the mean calculated from 2 or 3 independent experimental replicates.