The seventh pathogenic fusion gene FIP1L1-RARA was isolated from a t(4;17)-positive acute promyelocytic leukemia

The majority of acute promyelocytic leukemia (APL) cases are characterized by the expression of the chimeric fusion gene PML-RARA. Although the PML-RARA fusion gene is detected in more than 95% of APL cases, RARA has also been found to fuse with other partner genes in some APL variants. To date, five such partner genes have been reported: PLZF, NPM, NuMA, Stat5b and PRKAR1A.1,2 These fusion gene products however, must meet a number of common prerequisites for APL pathogenesis to ensue. The RARA gene portion of the fusion gene products ought to be from exon 3, and the fusion gene products must form homodimers as well as repress retinoic acid-responsive transcriptional activity.3,4 We hereby report the cloning of a seventh fusion gene from an APL variant and the functional characterization of its product.

A 90 year-old woman was clinically diagnosed for APL. The karyotype was 47, XX, t(4;17)(q12;q21), +8. FISH analysis showed that 94% of the bone marrow cells had the RARA split signal without the PML-RARA fusion signal (Figure 1A).

To identify the 5’-fusion partner of RARA, we adopted the 5’-RACE method (5’-Full RACE Core Set, Takara Bio) according to the manufacturer’s instructions. Briefly, the reverse primer 5’-GCGCTTTGCGCACCT-3’ was designed, which was complimentary for exon 3 of the RARA gene. Following reverse transcription using total mRNA from the patient’s bone marrow cells, the cDNA obtained was ligated by T4 RNA ligase. The ligated product was amplified by the nested polymerase chain reaction (PCR). PCR primer sequences were as follows: 1st PCR primers (5’-CTGCAGAAGTGCTTTGAAGT-3’, 5’-CACCTTGTTGATGATGCAGT-3’) and 2 PCR primers (5’-GAGTGCTCTGAGAGCTACAC-3’, 5’-CGGTGA-CACGTGTACACCAT-3’). The products obtained were cloned and sequenced directly. As a result, FIP1L1 was identified as the fusion partner of RARA (Figure 1B). The RARA portion in this case starts, as expected, from exon 3 and is fused to exon 15 of FIP1L1. While cloning the full length FIP1L1-RARA, we isolated three isoforms of FIP1L1-RARA; all of these isoforms are inframe (Figure 1C). We also confirmed the mRNA expression of RARA-FIP1L1 by means of RT-PCR analysis (data not shown). FIP1L1 is known to form a fusion gene with PDGFRA that causes chronic eosinophilic leukemia.5 In a similar fashion to FIP1L1-PDGFRA, which produces several isoforms caused by alternative splicing, the isoforms of FIP1L1-RARA also seemed to be generated by alternative splicing of the FIP1L1 portion.6 FIP1L1-RARA was recently isolated from a case of juvenile myelomonocytic leukemia (JMML).7 In the JMML case, as in our case, the fusion gene was generated between exon 15 of FIP1L1 and exon 3 of RARA. At the moment, the reason why FIP1L1-RARA causes two different phenotypes of leukemia is unknown, nevertheless we propose two hypotheses. One possibility is that the difference in cell lineage derived from the identical fusion gene may be due to some additional mutation, allowing for the progression of a particular disease and not another. Alternatively, the fusion gene may target different progenitor populations and influence the phenotype.

Figure 1.FIP1L1-RARA was identified from t(4;17)-positive APL cells. (A) Morphology of the leukemia cells shows hypergranular promyelocytes with Auer rods (upper panel). FISH, using a PML probe (red signal) and a RARA probe (green signal), was performed for nuclei of a leukemia cell in interphase. Split FISH signals of RARA (arrow) indicate rearrangement of RARA (lower panel). (B) The sequence analysis of the identified fusion gene from the reverse sequence of RARA exon 3 identified FIP1L1 as the fusion partner gene. The fusion gene between FIP1L1 and RARA is in frame and the translated amino acid sequence is shown. (C) Schematic representation of the estimated organization of FIP1L1-RARA rearrangement at the genomic level and the isolated isoforms. Isoform 1 lacks exons 2 and 11 and gains an additional exon 13a. Isoform 2 lacks exon 11. Isoform 3 lacks exon 11 and gains exon 13a.

FIP1L1-RARA has already been isolated; however, the function of the gene product has not yet been analyzed. Thus, we examined the potential of FIP1L1-RARA to form a homodimer. The full length cDNAs of FIP1L1, RARA and FIP1L1-RARA were cloned into the T7-tagged or FLAG-tagged expression vectors, and these vectors were transiently transfected and used for immunoprecipitation and immunoblotting analysis. The result revealed that all three identified isoforms of FIP1L1-RARA (Figure 2A, lanes 9–11), as well as RARA-FIP1L1 (Figure 2A, lane 16), form homodimers. Also, FIP1L1-RARA associated with either FIP1L1-RARA or FIP1L1, but not with RARA (Figure 2B), confirming that the homodimerization of FIP1L1-RARA is dependent on the FIP1L1 portion. Interestingly, the FIP1L1 portion of FIP1L1-PDGFRA does not have the ability to form a homodimer.8 The breakpoint in FIP1L1-RARA is in intron 15 of FIP1L1. On the other hand, the breakpoints in FIP1L1-PDGFRA were distributed among introns 10 to 13 of FIP1L1,9 which leads us to believe that the different breakpoints of FIP1L1 in FIP1L1-RARA and FIP1L1-PDGFRA may confer different attributes upon the fusion products at the point of homodimerization.

Figure 2.FIP1L1-RARA forms a homodimer and suppresses retinoic acid-dependent transcriptional activity. (A) HEK293 cells, in a 6 cm-dish, were transfected with both FLAG-tagged and T7-tagged expression vectors. T7-FIP1L1-RARA was co-expressed with either FLAG-FIP1L1-RARA or FLAG-RARA-FIP1L1 (lanes 1–4 and 9–12). T7-RARA-FIP1L1 was co-expressed with FLAG-FIP1L1-RARA or FLAG-RARA-FIP1L1 (lanes 5–8 and 13–16). The cell lysates were subjected to immunoprecipitation and immunoblotting analyses. Immunoblotting of the whole cell lysates (WCL) with anti-T7 polyclonal antibody (MBL) confirmed the expression of T7-FIP1L1-RARA and T7-RARA-FIP1L1 (lanes 1–8). Immuno-precipitation (IP) with anti-FLAG M2 antibody (SIGMA) was subsequently subjected to the immunoblotting analysis with either anti-T7 antibody or anti-FLAG M2 antibody (lanes 9–16). (B) T7-FIP1L1-RARA was co-expressed with mock (lane 1), FLAG-FIP1L1-RARA (lane 2), FLAG-FIP1L1 (lane 3) or FLAG-RARA (lane 4). Immunoblotting of the WCL with anti-T7 antibody confirmed the expression of T7-FIP1L1-RARA (upper panel). IP with anti-FLAG M2 antibody was subsequently subjected to immunoblotting analysis with either anti-T7 antibody (middle panel) or anti-FLAG M2 antibody (lower panel). (C) HEK293 cells, in 35 mm-dishes, were transfected with 0.25 μg of the retinoic acid responsive-luciferase vector, which contains seven repeats of the retinoic acid-response element (RARE) in the RARβ2 gene, combined with 2 μg of empty vector or the expression vectors FLAG-PML-RARA, FLAG-FIP1L1-RARA or FLAG-RARA-FIP1L1 respectively. One day after transfection, the culture media were exchanged with fresh culture media supplemented with the indicated concentration of ATRA. Following two days of ATRA treatment, the cells were harvested and luciferase activities were analyzed. The luciferase activity without ATRA treatment was arbitrarily assigned as 1.0 and the results are shown as the mean ± SD. The analysis was performed in triplicate assays and the results were reproducible.

Subsequently, we examined the effect of FIP1L1-RARA on the retinoic acid response. FIP1L1-RARA represses retinoic acid-dependent transcriptional activity at the lowest concentration of all-trans retinoic acid (ATRA), as determined by the luciferase assay (Figure 2C). At higher concentrations of ATRA, luciferase activity is induced. As shown in Figure 2C, FIP1L1-RARA has an ATRA response similar to that of PML-RARA. Consistent with this result, our patient achieved a complete remission by oral administration of ATRA alone (50 mg daily) (Mori et al., manuscript in preparation).

In our examination, the expression of the reciprocal RARA-FIP1L1 fusion gene was also detected. The analysis of PML-RARA transgenic mice suggested that the reciprocal product, a particular isoform of RARA-PML transcripts, plays a role in disease progression.10,11 At this moment we are not suggesting that RARA-FIP1L1 has any functional role in leukemogenesis, but the homodimerization of RARA-FIP1L1 implies that it may possess a pathological function.

To summarize, FIP1L1-RARA forms a homodimer and represses the retinoic acid response. We therefore propose that FIP1L1-RARA is the seventh pathogenic fusion gene of APL.