Development of RNA-FISH Assay for Detection of Oncogenic FGFR3-TACC3 Fusion Genes in FFPE Samples

Abstract

Introduction and objectives: Oncogenic FGFR3-TACC3 fusions and FGFR3 mutations are target candidates for small molecule inhibitors in bladder cancer (BC). Because FGFR3 and TACC3 genes are located very closely on chromosome 4p16.3, detection of the fusion by DNA-FISH (fluorescent in situ hybridization) is not a feasible option. In this study, we developed a novel RNA-FISH assay using branched DNA probe to detect FGFR3-TACC3 fusions in formaldehyde-fixed paraffin-embedded (FFPE) human BC samples.

Materials and methods: The RNA-FISH assay was developed and validated using a mouse xenograft model with human BC cell lines. Next, we assessed the consistency of the RNA-FISH assay using 104 human BC samples. In this study, primary BC tissues were stored as frozen and FFPE tissues. FGFR3-TACC3 fusions were independently detected in FFPE sections by the RNA-FISH assay and in frozen tissues by RT-PCR. We also analyzed the presence of FGFR3 mutations by targeted sequencing of genomic DNA extracted from deparaffinized FFPE sections.

Results: FGFR3-TACC3 fusion transcripts were identified by RNA-FISH and RT-PCR in mouse xenograft FFPE tissues using the human BC cell lines RT112 and RT4. These cell lines have been reported to be fusion-positive. Signals for FGFR3-TACC3 fusions by RNA-FISH were positive in 2/60 (3%) of non-muscle-invasive BC (NMIBC) and 2/44 (5%) muscle-invasive BC (MIBC) patients. The results of RT-PCR of all 104 patients were identical to those of RNA-FISH. FGFR3 mutations were detected in 27/60 (45%) NMIBC and 8/44 (18%) MIBC patients. Except for one NMIBC patient, FGFR3 mutation and FGFR3-TACC3 fusion were mutually exclusive.

Conclusions: We developed an RNA-FISH assay for detection of the FGFR3-TACC3 fusion in FFPE samples of human BC tissues. Screening for not only FGFR3 mutations, but also for FGFR3-TACC3 fusion transcripts has the potential to identify additional patients that can be treated with FGFR inhibitors.

Fig 1. Schematic representation of how FGFR3-specific probes and TACC3-specific probes were designed.

Fig 2. A schematic figure explaining how bDNA-FISH works.

Fig 3. FGFR3-TACC3 fusion transcript detection by RT-PCR.

(A) Schematic representation of FGFR3-TACC3 fusion mRNA and PCR primers position. (B) Agarose gel separation of the FGFR3-TACC3 fusion specific RT-PCR amplicons. (C) Sanger sequencing chromatogram of FGFR3-TACC3 fusion specific RT-PCR products. The arrowhead and solid bar indicate breakdown point or region of the 2 genes.

Fig 4. FGFR3-TACC3 fusion transcript detection by RNA-FISH.

RNA-FISH image of RT112 and RT4 (fusion-positive controls) and HSC-39 (negative control) xenograft FFPE tissue. mRNAs of FGFR3 and TACC3 were detected by RNA-ISH using fluorescent probes (Alexa 647 for FGFR3 and Alexa 546 for TACC3, respectively), and signals from FGFR3 and TACC3 were shown as red and green, respectively, in the figure. The small boxed areas are enlarged in the adjacent large boxes. Fusion mRNAs appeared in microscope images as yellow, which are merged signals from red and green colors. Cell nuclei was stained with DAPI and shown as blue. Scale bars in the figure are 10 μm. For the detection of fusion signals for each sample, image data from 2 fluorescent probes were analyzed with IN Cell Analyzer 2000 and the number of overlapped/co-localized signals was counted and divided by the total number of FGFR3 signals and TACC3 signals and plotted in a scatter graph.

Fig 5. Detection of FGFR3-TACC3 fusion genes in FFPE clinical samples by RNA-FISH.

Scatter diagram of FGFR3-TACC3 co-localization ratios of 10 non-overlapping fields for each sample. The number of co-localized signals was divided by the number of FGFR3 signals and TACC3 signals, and the quotients were plotted in Y- and X-axis, respectively. Four samples in the right upper quadrant were thought to be fusion positive by RNA-FISH, and were confirmed as fusion positive by RT-PCR analysis. Small gray dots represent cases that were thought to be negative for the fusion gene.

Fig 6. Representative RNA-FISH images of fusion-positive case TKB014.

The small boxed areas are enlarged in the adjacent large boxes. Scale bars in the figure are 10 μm.

Fig 7. Detection of FGFR3-TACC3 fusion transcripts in clinical samples by RT-PCR.

(A) Agarose gel separation of the FGFR3-TACC3 fusion-specific RT-PCR amplicons. (B) Sanger sequencing chromatogram of FGFR3-TACC3 fusion-specific RT-PCR products. Arrowheads indicate breakdown points of the 2 genes.

Fig 8. FGFR3 mutation and FGFR3-TACC3 fusion status.

(A) The heatmap shows the distribution of FGFR3 mutations and FGFR3-TACC3 fusions with respect to T stage and pathological grade. (B) Subgroup analysis of NMIBC by T stage.