Identification of tumor-associated cassette exons in human cancer through EST-based computational prediction and experimental validation

Abstract

Background: Many evidences report that alternative splicing, the mechanism which produces mRNAs and proteins with different structures and functions from the same gene, is altered in cancer cells. Thus, the identification and characterization of cancer-specific splice variants may give large impulse to the discovery of novel diagnostic and prognostic tumour biomarkers, as well as of new targets for more selective and effective therapies.

Results: We present here a genome-wide analysis of the alternative splicing pattern of human genes through a computational analysis of normal and cancer-specific ESTs from seventeen anatomical groups, using data available in AspicDB, a database resource for the analysis of alternative splicing in human. By using a statistical methodology, normal and cancer-specific genes, splice sites and cassette exons were predicted in silico. The condition association of some of the novel normal/tumoral cassette exons was experimentally verified by RT-qPCR assays in the same anatomical system where they were predicted. Remarkably, the presence in vivo of the predicted alternative transcripts, specific for the nervous system, was confirmed in patients affected by glioblastoma.

Conclusion: This study presents a novel computational methodology for the identification of tumor-associated transcript variants to be used as cancer molecular biomarkers, provides its experimental validation, and reports specific biomarkers for glioblastoma.

Figure 1

Genes and introns classification based on ESTs data analyses. (A) Number of available ESTs for each tissue, classified as normal (N) or tumoral (T), according to the CGAP classification. (B) Number of genes analyzed for each tissue, classified on the base of their expression in the normal (N and NN) and tumoral (T and TT) status and of comparable expression in both conditions (E). (C) Number of tumoral (T) and normal (N) specific introns for each tissue. BMA: Bone marrow, BRE: Breast, CNS: Central nervous system (e.g. brain); COL: Colon; DER: Skin; END: Endocrine system; EYE: Eye; INT: Gastrointestinal tract; KID: Kidney; LIV: Liver; LYM: Lymphoreticular system; MSK: Musculoskeletal system; PLA: Placenta; PRO: Prostate; RES: Respiratory System; STO: Stomach; TES: Testis

Figure 2

Cassette and mutually exclusive exons. Schematic representation of a tumoral cassette exon (A) and mutually exclusive exons (B). Tumoral cassette exon or mutually exclusive exons are defined by two cancer specific splice site pairs, alternative to one or two normal specific splice site pairs.

Figure 3

Expression analyses of some normal/tumoral predicted cassette exons. Experimental validation of normal (A), and tumoral (B) predicted cassette exons (CE) and a pair of mutually exclusive CEs (C) by RT-qPCR. w/o CE = transcripts lacking CE; with CE = transcripts containing CE. For each condition, results are expressed as relative expression ratio of transcripts containing CE respect to transcripts lacking CE, used as calibrator. For each gene, a schematic representation of alternative transcripts and primer pairs used are reported.

Figure 4

Expression analyses of some CNS predicted cassette exons in gliobastoma samples. Experimental validation of nervous system specific predicted cassette exons (CEs) of STRADA, PCNP, ATP6v0A1 and CS genes by RT-qPCR in normal brains and in 8 patients affected by glioblastoma. Statistical analysis was performed using an unpaired one-tail T-test. w/o CE = transcripts lacking CE; with CE = transcripts containing CE. For each condition, results are expressed as relative expression ratio of transcripts containing CE respect to transcripts lacking CE, used as calibrator.