Characterisation of FLT3 alterations in childhood acute lymphoblastic leukaemia

Abstract

Background: Alterations of FLT3 are among the most common driver events in acute leukaemia with important clinical implications, since it allows patient classification into prognostic groups and the possibility of personalising therapy thanks to the availability of FLT3 inhibitors. Most of the knowledge on FLT3 implications comes from the study of acute myeloid leukaemia and so far, few studies have been performed in other leukaemias.

Methods: A comprehensive genomic (DNA-seq in 267 patients) and transcriptomic (RNA-seq in 160 patients) analysis of FLT3 in 342 childhood acute lymphoblastic leukaemia (ALL) patients was performed. Mutations were functionally characterised by in vitro experiments.

Results: Point mutations (PM) and internal tandem duplications (ITD) were detected in 4.3% and 2.7% of the patients, respectively. A new activating mutation of the TKD, G846D, conferred oncogenic properties and sorafenib resistance. Moreover, a novel alteration involving the circularisation of read-through transcripts (rt-circRNAs) was observed in 10% of the cases. Patients presenting FLT3 alterations exhibited higher levels of the receptor. In addition, patients with ZNF384- and MLL/KMT2A-rearranged ALL, as well as hyperdiploid subtype, overexpressed FLT3.

Discussion: Our results suggest that specific ALL subgroups may also benefit from a deeper understanding of the biology of FLT3 alterations and their clinical implications.

Fig. 1. Overview of FLT3 study.

A total of 342 patients with DNA-seq (n = 267) and RNA-seq (n = 160) data were included in the study. Sixteen PM in 15 patients were identified using both DNA-seq and RNA-seq data. An ITD in 4 patients and 8 rt-circRNAs in 16 patients were identified using RNA-seq data.

Fig. 2. Diagram showing mutations reported in FLT3 in childhood ALL.

In the upper part of the figure, FLT3 mutations reported in the current study. In red, mutations selected for functional validation. In the lower part, FLT3 mutations reported in literature. Fully displayed mutations (text) have been functionally characterised. Detail information about all previously reported mutations is available in Supplementary Tables 3 and 4.

Fig. 3. In vitro validation of FLT3 mutations.

a Ba/F3 cells stably transduced with FLT3 WT, D835Y, Y589D, G846D and the empty vector were seeded at a density of 1 × 10⁵ in the presence or absence of IL3. Viable cells were assessed with CellTiter Glo assay after 72 h. Cell growth in the presence of IL3 was set to 100% for each cell line. Data represent values ± SD of triplicates. b Ba/F3 cells expressing the indicated constructs were starved overnight in media without IL3. STAT5 and ERK activation was analysed by Western blot from crude cell extract using 100 µg of proteins and phospho-specific antibodies, membranes were then stripped and reprobed with antibodies against total STAT5, ERK and GAPDH. For the visualisation of FLT3 receptor, 250 µg of proteins were used. Ba/F3 cells expressing FLT3 D835Y and G846D were incubated with increasing concentrations of c midostaurin and d sorafenib in the absence of IL3. Cell viability was determined after 48 h. Data are presented as percentage of untreated cells. Data represent mean values ± SD of triplicates.

Fig. 4. Characterisation of rt-circURAD-FLT3.

a Fusion transcript URAD::FLT3 containing URAD-exon 1 and exons 16–24 of FLT3 as detected by FusionCatcher and Arriba from RNA-seq data. b Schematic showing the rt-circURAD-FLT3 generated by the fusion transcript. Position of primers used for RT-PCR and Sanger sequencing are indicated by arrows. c Validation of the existence of rt-circURAD-FLT3 in patient TC0133 by RT-PCR after RNAse R treatment, labelled R; cDNA generated using random priming, labelled S; cDNA generated using gene-specific priming, labelled B; blank. d Sanger sequencing confirms the fusion between 3´ exon 1 of URAD and 5´ exon 16 of FLT3 and the backsplicing junction between 3´exon 23 of FLT3 and 5´exon 1 of URAD.

Fig. 5. FLT3 expression.

Box plots showing FLT3 gene expression levels measured by FPKM values extracted from transcriptome data and log-transformed by a ALL subtype, b FLT3 alterations, c white blood cell counts (considering high levels WBC ≥ 50×10⁹/L), d involvement of CNS, e MRD status at day 32 of induction, f relapse. The boxes extend from the 25th to 75th percentiles and the middle line represents median values. Comparisons assessed by ANOVA and t test, two sides. *P value < 0.01; **P value < 0.001; ***P value < 0.0001. Only significant P values (<0.05) are indicated.