Gains or deletions. Interestingly, only the fusion gene PPP1R12A-SEPT10 in KPL-4 was not associated with either copy number transitions or changes at the location of either of the fusion counterparts as detected with the 1M probe aCGH.Structural properties of the novel fusion genesMCF-7KPL-M 300 bp 200 bp PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/28914615 100 bp36 02 00 B 1 D3 2 4 00 M DH ET FIP 70 H SG SD C -S R F F3 H X1 A KI -EN 1-G 1-P 5C -LR -ZN -EI -ITC RE -P L N D 8 F1 67 1 5 -P H RA E1 TD KH DC M DR TH X3 S1 PD 2O RA CS TA AN CC SU W CY DH NF GA dHSK-BR-6 50 B 2 F8 5 2L 7 40 P2 9 AC -SN F3 CE YO O19 OK CF TM I3 AA0 ST B1 KZ 8- A-M Y 3-D -M M -P -KI H D 2 -M D A- K -I 1 -C 1 1 AC S6 PB YN B2 A2 AR MP B1 NE DO PD 2O AC RP VA ZM RA SK ST LA GL CP DI GA dHM 300 bp 200 bp 100 bpBT-Figure 2 Experimental validation of identified breast cancer fusion transcripts. RT-PCR validation of fusions found in MCF-7 and KPL-4 (upper), SCH 530348 web SK-BR-3 (middle), and BT-474 (lower). Also shown is the marker and the negative control.overexpression [16]. We also observed complex rearrangements, where multiple breaks in a narrow genomic region led to the formation of more than one gene fusion in the same sample. For instance, altogether six genes in the ERBB2-amplicons in BT-474 and SK-BR-3 took part in gene fusions (Figure 4b). As seen with the FISH analysis (Figure 3b; Additional file 4), the fusions were only seen in two to five copies per cell on average, indicating that the multiple genomic breakpoints required for the formation of high-level amplifications were probably contributing to the formation of the fusions as secondary genetic events. Another important group of gene fusions was associated with breakpoints of low-level copy number changes, involving both gains and deletions. These are interesting in the sense that they represent the types of fusion events leading to gene activation with no association with gene amplifications. For example, this is the case for TMPRSS2-ERG and many leukemia-associated translocations [17]. Eight out of 27 fusion genes (BSG-NFIX, CCDC85C-SETD3, DHX35-ITCH, CMTM7-GLB1, LAMP1-MCF2L, NOTCH1-NUP214, PPP1R12A-SEPT10 and SUMF1-LRRFIP2) identified here were not associated with high-level gene amplifications, but typically had one of the fusion partners associated with a low-Several consistent patterns observed for the gene fusions suggest their potential importance. First, most of the fusions (23 of 27) were predicted to be in-frame (Table 1), assuming that the splicing pattern of the rest of the transcript is retained. Should the reading frame not be retained across the fusion junction, it would likely lead to appearance of a premature stop codon and the transcript would be degraded by nonsense-mediated mRNA decay. Therefore, it is possible that some of the highly expressed fusions that were predicted to be out-offrame, such as ZMYND8-CEP250, may retain an intact open reading frame through alternative splicing or mutations that place the gene back in frame. Second, we observed 19 intra- and 8 interchromosomal translocations (Figure 4a; Additional file 6), which is in line with the previously observed pattern of intrachromosomal rearrangements occurring more frequently based on data from genomic sequencing [14]. Several (9 of 27) fusion partner genes were located on opposite strands, implying inversion, which in some cases has been followed by amplification of the rearranged region (for example, ZMYND8-CEP250). Third, the rearranged genes were occasionally exclusively expre.
