Identification of fusion genes in breast cancer by paired-end RNA-sequencing

Abstract

Background: Until recently, chromosomal translocations and fusion genes have been an underappreciated class of mutations in solid tumors. Next-generation sequencing technologies provide an opportunity for systematic characterization of cancer cell transcriptomes, including the discovery of expressed fusion genes resulting from underlying genomic rearrangements.

Results: We applied paired-end RNA-seq to identify 24 novel and 3 previously known fusion genes in breast cancer cells. Supported by an improved bioinformatic approach, we had a 95% success rate of validating gene fusions initially detected by RNA-seq. Fusion partner genes were found to contribute promoters (5' UTR), coding sequences and 3' UTRs. Most fusion genes were associated with copy number transitions and were particularly common in high-level DNA amplifications. This suggests that fusion events may contribute to the selective advantage provided by DNA amplifications and deletions. Some of the fusion partner genes, such as GSDMB in the TATDN1-GSDMB fusion and IKZF3 in the VAPB-IKZF3 fusion, were only detected as a fusion transcript, indicating activation of a dormant gene by the fusion event. A number of fusion gene partners have either been previously observed in oncogenic gene fusions, mostly in leukemias, or otherwise reported to be oncogenic. RNA interference-mediated knock-down of the VAPB-IKZF3 fusion gene indicated that it may be necessary for cancer cell growth and survival.

Conclusions: In summary, using RNA-sequencing and improved bioinformatic stratification, we have discovered a number of novel fusion genes in breast cancer, and identified VAPB-IKZF3 as a potential fusion gene with importance for the growth and survival of breast cancer cells.

Figure 1

Fusion gene identification by paired-end RNA-sequencing. (a) Identification of fusion gene candidates through selection of paired-end reads, the ends of which align to two different and non-adjacent genes. (b) Identification of the exact fusion junction by aligning non-mapped short reads against a computer generated database of all possible exon-exon junctions between the two partner genes. Separation of true fusions (left) from false positives (right) by examining the pattern of short read alignments across exon-exon junctions. Genuine fusion junctions are characterized by a stacked/ladder-like pattern of short reads across the fusion point. False positives lack this pattern; instead, all junction matching short reads align to the exact same position or are shifted by one to two base pairs. Furthermore, this alignment is mostly to one of the exons.

Figure 2

Experimental validation of identified breast cancer fusion transcripts. RT-PCR validation of fusions found in MCF-7 and KPL-4 (upper), SK-BR-3 (middle), and BT-474 (lower). Also shown is the marker and the negative control.

Figure 3

Genomic structure, validation and functional significance of VAPB-IKZF3. (a) Exonic expression of VAPB-IKZF3 is indicated by sequencing coverage (red). Copy number changes measured by array comparative genomic hybridization (aCGH; black dots) in reference to normal copy number (horizontal grey line) and fusion break points (vertical grey line) are indicated. Gene structures are shown below the aCGH data. Arrows below gene structures indicate which strand the genes lie on. Fusion transcript structure is pictured below wild-type (wt) gene structures. (b) Interphase FISH showing amplification of VAPB and IKZF3 and the VAPB-IKZF3 fusion in BT-474. White arrows indicate gene fusions. (c) Expression of the 5' and 3' partner genes and the fusion gene. RPKM denotes reads per kilobase per million sequenced short reads. (d) Quantitative RT-PCR validation of small interfering RNA (siRNA) knock-down efficiency of cells transfected either with a scramble siRNA or with gene-specific siRNAs. Error bars show standard deviation. (e) CTG cell viability analysis of cells transfected either with a scramble siRNA or with gene-specific siRNAs. Asterisks indicate the statistical significance of growth reduction: ***P < 0.001. Error bars show standard deviation.

Figure 4

Genomic rearrangements in SK-BR-3 and BT-474. (a) Circos plots representing chromosomal translocations in SK-BR-3 (upper right) and BT-474 (lower left). Chromosomes are drawn to scale around the rim of the circle and data are plotted on these coordinates. Selected chromosomes involved in the fusion events are shown in higher magnification. Each intrachromosomal (red) and interchromosomal (blue) fusion is indicated by an arc. Copy number measured by aCGH is plotted in the inner circle where amplifications are shown in red and deletions in green. N denotes the number of fusion genes per cell line. (b) Fusion gene formation in the ERBB2-amplicon region. Fusion partner genes within and near the amplicon region are connected with black lines (both partners on chromosome 17), or location of the other partner is indicated (partner gene on different chromosomes). Smoothed aCGH profiles (log2) for SK-BR-3 (blue) and BT-474 (red) indicate copy number changes in reference to normal copy number (horizontal grey line). ERBB2, which is not fused (arrow), and chromosomal positions (bottom) are indicated.

Figure 5

Exclusive expression of the exons of the 3' partner genes taking part in the fusions. Exonic expression of CEP250 in ZMYND8-CEP250 (upper), ENSG00000236127 in CSE1L-ENSG00000236127 (second from top), GSDMB in TATDN1-GSDMB (second from bottom) and BCAS4 in BCAS3-BCAS4 (lower) is indicated by sequencing coverage (red). Copy number changes measured by aCGH (black dots) in reference to normal copy number (horizontal grey line) and fusion break points (vertical grey line) are indicated. Chromosomal positions and transcript structures are shown below the aCGH data. Transcript structures above and below chromosome coordinates denote forward and reverse strand, respectively.