A systems biology-based gene expression classifier of glioblastoma predicts survival with solid tumors

Abstract

Accurate prediction of survival of cancer patients is still a key open problem in clinical research. Recently, many large-scale gene expression clusterings have identified sets of genes reportedly predictive of prognosis; however, those gene sets shared few genes in common and were poorly validated using independent data. We have developed a systems biology-based approach by using either combined gene sets and the protein interaction network (Method A) or the protein network alone (Method B) to identify common prognostic genes based on microarray gene expression data of glioblastoma multiforme and compared with differential gene expression clustering (Method C). Validations of prediction performance show that the 23-prognostic gene classifier identified by Method A outperforms other gene classifiers identified by Methods B and C or previously reported for gliomas on 17 of 20 independent sample cohorts across five tumor types. We also find that among the 23 genes are 21 related to cellular proliferation and two related to response to stress/immune response. We further find that the increased expression of the 21 genes and the decreased expression of the other two genes are associated with poorer survival, which is supportive with the notion that cellular proliferation and immune response contribute to a significant portion of predictive power of prognostic classifiers. Our results demonstrate that the systems biology-based approach enables to identify common survival-associated genes.

Figure 1. Flowchart of Methods A, B and C.

Schematic method overview of a systems biology-based approach using either combined gene sets and the protein interaction network (Method A) or the protein interaction network alone (Method B) and conventional gene expression clustering (the SAM-based analysis) (Method C) for prognostic gene identification based on microarray gene expression data of primary GBMs.

Figure 2. The null distributions of S values of CDK2 and IFNGR1.

The null distribution of S values of CDK2 and IFNGR1 are computed from MDA-GBM data. The left panel is the null distribution of S value of cell proliferation-related gene CDK2; the right panel is the null distribution of S value of immune response-related gene IFNGR1.

Figure 3. Kaplan-Meier plots of overall survival for advanced gliomas generated by the 23-gene classifier.

(A) Two GBM cohorts UCSF-2 and CMBC. (B) Three HGG cohorts MDA, UCLA, and CMBC. (C) Three GBM training cohorts MDA, UCLA, and UCSF-1. STS, short-term survival group; LTS, long-term survival group. N, the number of patients within STS or LTS group. P values are indicated within plots. P< = 0.05 is defined as significance.

Figure 4. Kaplan-Meier plots of overall survival for other solid tumor types generated by the 23-gene classifier.

(A) Five lung cancer cohorts DFCI, PCH, CAN/DF, MSK, and UM-HLM. (B) Five breast cancer cohorts GIS, CRCM, SUSM, NCI and EMC. For NCI and EMC, the overall survival times were unavailable and thus time to distant metastasis for prediction was used instead. (C) One bladder cancer cohort AUH. (D) One ovarian cancer cohort MNI. STS, short-term survival group; LTS, long-term survival group. N, the number of patients within STS or LTS group. P values are indicated within plots. P< = 0.05 is defined as significance.

Figure 5. The protein interaction sub-network based on 23 prognostic gene-encoded proteins and their interacting partners.

Nodes represent gene-encoded proteins; links represent physical interactions. Nodes in color indicate enriched biological functions of the proteins. Red nodes represent cell proliferation, green nodes represent immune response, and half red and half green nodes represent both cell proliferation and immune response. Proteins in a bigger circle represent the 23 gene-encoded proteins; their interactions with partners with enriched biological functions are highlighted in blue links, whereas grey links represent interactions of the 23 gene-encoded proteins with other partners (white circle).