Heterogeneity and clinical significance of ETV1 translocations in human prostate cancer

Abstract

A fluorescence in situ hybridisation (FISH) assay has been used to screen for ETV1 gene rearrangements in a cohort of 429 prostate cancers from patients who had been diagnosed by trans-urethral resection of the prostate. The presence of ETV1 gene alterations (found in 23 cases, 5.4%) was correlated with higher Gleason Score (P=0.001), PSA level at diagnosis (P=<0.0001) and clinical stage (P=0.017) but was not linked to poorer survival. We found that the six previously characterised translocation partners of ETV1 only accounted for 34% of ETV1 re-arrangements (eight out of 23) in this series, with fusion to the androgen-repressed gene C15orf21 representing the commonest event (four out of 23). In 5'-RACE experiments on RNA extracted from formalin-fixed tissue we identified the androgen-upregulated gene ACSL3 as a new 5'-translocation partner of ETV1. These studies report a novel fusion partner for ETV1 and highlight the considerable heterogeneity of ETV1 gene rearrangements in human prostate cancer.

Figure 1

FISH detection of ETV1 gene re-arrangements. Top: Interphase nuclei are hybridised to probes that detect sequences immediately 3′ to the ETV1 gene (probe I, red) and immediately 5′ to the ETV1 gene (probe II, green). The red and green signals are separated when an ETV1 gene rearrangement occurs. (A) Signals from normal un-rearranged ETV1 loci (class N). (B) Rearranged ETV1 gene with separate red (3′) and green (5′) probes (class ETV1 Esplit). Bottom: Map of the ETV1 gene showing the position of the BACs used as probes in FISH assays. Probe I: E1 (RP11-27B1), E2 (RP11-138H16), E3 (CTD-2008I15) labelled with Cy3. Probe II E4 (RP11-905H4), E5 (RP11-621E24), E6 (RP11-115D14) labelled with FITC. The direction of transcription of genes at this locus are indicated by arrows.

Figure 2

FISH detection of translocation of ETV1 to chromosome 14(q13.3–21.1). Top: Interphase nuclei are hybridised to probes that detect sequences immediately 3′ to the ETV1 gene on chromosome 7 (probe I, red) and a green probe (probe V) consisting of six BACS spanning the 14q13.3–21.1 region. (A). Red and green signals are normally separated. (B) Co-localisation of red and green probes indicate juxtaposition of chr 7 ETV1 sequences with chr 14 (q 13.3–21.1). The lower panel shows the position of the BACs used for probe V: C1 (RP11-945C4), C2 (RP11-381L10), C3 (RP11-666J24), C4 (RP11-796F21), C5 (RP11-588D7), C6 (RP11-107E23) labelled with FITC. The relative position and direction of transcription of genes are indicated by the arrows.

Figure 3

ACSL3:ETV1 fusion. (A) ACSL3 (red) and ETV1 (blue) transcripts with ORFs in dark colour. Exons are numbered. A fusion transcript of ACSL3 exon 3 fused to ETV1 exon 6 was detected by 5′-RACE from exon 6 ETV1 sequences in prostate cancer sample 23. The ORF shown was predicted using software at www.dnalc.org. (B) Sequence across the ACSL3:ETV1 fusion boundary. Underlined regions indicate the position of primers used in RT–PCR to confirm the fusion. The predicted fusion gene initiation codon is indicated in red. ACSL3 sequence is in lower case and ETV1 sequence in upper case. (C) RT–PCR detection of an ACSL3:ETV1 fusion transcript in RNA extracted from formalin-fixed paraffin-embedded prostate cancer samples: lanes 1–12 are ETV1-rearranged tumour samples, lane 12: tumour sample 23, lane 13 negative control. (D) FISH assays to confirm fusion of ACSL3 with ETV1. Panel i: The ETV1 break-apart assay utilises probes corresponding to 3′-ETV1 sequences (red) and 5′-ETV1 sequences (green) (see also Figure 1). A nucleus with separated red and green probes confirming rearrangement of ETV1 is shown. Panel ii: The ACSL3 break-apart assay hybridised the same TMA slice used in the ETV1 break-apart assay to 3′-ACSL3 sequences (red) and 5′-ACSL3 sequences (green). These signals are coincident in the wild type, but are split on translocation of ACSL3. Comparison of the images in panels i and ii indicates co-localisation of 3′-ETV1 with 5′-ACSL3 and co-localisation of 5′-ETV1 and 3′-ACSL3. This is confirmed by ETV1-ACSL3 co-localisation assays (panel iii) demonstrating co-localisation of 3′-ETV1 sequences (red) and 5′-ACSL3 sequences (green) and (panel iv) demonstrating co-localisation of 3′-ACSL3 sequences and 5′-ETV1 sequences (red) in the same cell. Superimposition of the images in panels iii and iv confirms co-localisation of wild-type 3′-ETV1 (panel iii) with 5′-ETV1 (panel iv) and of wild-type 3′-ACSL3 (panel iv) with 5′-ACSL3 (panel iii). The genes and their direction of transcription are indicated by the arrowheads. (E) Map of the ACSL3 gene showing the position of the BACs used as probes in FISH assays. Probe XV: A1 (RP11-157M20) labelled with FITC. Probe XIV: A2 (RP11-136M23) and A3 (RP11-749C15) labelled with Cy3. Probes XV and probes XIV correspond, respectively, to sequences immediately 5′ (green) and 3′ (red) to the ACSL3 gene. Direction of gene transcription indicated by arrowheads.