Hh/Gli antagonist in acute myeloid leukemia with CBFA2T3-GLIS2 fusion gene

Pediatric acute myeloid leukemia (AML) carrying CBFA2T3-GLIS2 fusion gene deserves particular interest, being associated with a grim prognosis in all the reports published so far [1‐3]. The incidence of this aberration is 17 and 8% in pediatric non-Down syndrome acute megakaryoblastic and in pediatric cytogenetically normal AML, respectively [1‐3]. The expression profile of CBFA2T3-GLIS2 is associated with upregulation of both Hedgehog (HH) and bone morphogenic protein (BMP) signaling [1, 4].

The protein GLIS2 shares a highly homologous zinc finger domain with members of the GLI proteins, the final effectors of classic Hedgehog pathway. GANT61 is a GLI inhibitor showing a potent effect on the inhibition of transcription activity of GLI proteins, blocking their binding to DNA [5‐8]. Considering the high homology of the DNA-binding domain between GLIS2 and GLI family proteins, we hypothesized that GANT61 might be used to specifically target the CBFA2T3-GLIS2 fusion gene in pediatric AML.

In the present study, we investigated the in vitro effects of GANT61 on AML cell lines and primary cells from AML patients harboring the CBFA2T3-GLIS2 fusion gene.

The materials and methods are detailed in Additional file 1. Molecular analysis of CBFA2T3-GLIS2 fusion gene is reported in Additional file 2: Figure S1. Genetic features of control AML cell lines without GLIS2 fusion are reported in Additional file 3: Table S1.

Our results showed that AML cell lines with CBFA2T3-GLIS2 fusion gene have a higher sensitivity to GANT61 than other AML cell lines without this genetic aberration (Fig. 1a). Similar results were obtained on primary leukemia cells isolated from AML patients, being the IC₅₀ of the GLIS2-positive leukemia and negative primary cells 13.6 and 41.6 μM, respectively (Fig. 1b).

Treatment with GANT61 induced an increase of about 30% of apoptotic cells (Fig. 1c) and block of cell cycle in G0/G1 phase only in M07e and WSU-AML lines positive to CBFA2T3-GLIS2 (Fig. 1d and Additional file 4: Figure S2).

We further analyzed the expression profile of cell lines and primary cells following GANT61 treatment. Through qPCR, we demonstrated that GANT61 treatment led to a significant reduction of the expression of BMP2 and GLIS2 (Fig. 2a, b). In order to fully characterize the effect of GANT61 treatment on whole transcriptome profile of CBFA2T3-GLIS2-positive cells, gene expression was assessed by microarray analysis. After GANT61 treatment, the expression of target genes of CBFA2T3-GLIS2, such as CRISP3, GATA3, H2AFY, or NCAM1(CD56) was significantly downregulated (p < 0.05) (Fig. 2c). Moreover additional genes were downregulated by the treatment with GANT61 (p < 0.05), as for example, genes involved in cell cycle control, such as CDC25A, CDC7, CDCA2, and CCNA2. The expression of these genes is required for progression through cell cycle, and their expression is aberrant in AML, as well as in other malignancies [9, 10]. In our model, the expression of these genes is downregulated following GANT61 treatment, this partly explaining the cell cycle arrest observed in AML cell lines with GLIS2 fusion after GANT61 treatment (Fig. 1d). Genes involved in cell proliferation (KIF14, MELK, MCM10, NUF2) and epigenetic regulators namely DNMT1 and DNMT3B were also present. Considering the particular interest of these DNA methyltransferase genes, we performed chromatin immunoprecipitation (ChIP) analysis using a CBFA2T3-specific antibody on WSU-AML and M07e cell lines. Our findings showed that CBFA2T3-GLIS2 fusion protein directly binds to the proximal promoter of DNMT1 and DNMT3B, positively regulating their expression (Fig. 2d). Overexpression of DNMT genes could lead to DNA hypermethylation and could be involved in the leukemogenesis process.

Since in the classic Hedgehog signaling pathway, several target genes involved in feedback mechanisms (HHIP, PTCH1, GLI1) have been described; we hypothesized that CBFA2T3-GLIS2 chimeric protein could regulate wild-type GLIS2 protein with the same feedback mechanism. Western blotting analysis confirmed the presence of chimeric protein and showed also the presence of wild-type GLIS2. These results may point to the role of the fusion protein in regulating wild-type GLIS2. Following GANT61 treatment, the expression of both proteins was decreased with respect to untreated samples, suggesting that GANT61 treatment targeted both CBFA2T3-GLIS2 fusion protein and also wild-type GLIS2 expression (Fig. 2e). Moreover, other AML cell lines not carrying the CBFA2T3-GLIS2 fusion gene did not show expression of GLIS2 (data not shown).

On the other hand, western blotting analysis showed that expression of GLI1 and GLI2 did not decrease following GANT61 treatment (Additional file 5: Figure S3).

Another study demonstrated a high sensitivity of this subgroup of AML with GLIS2 fusion to Aurora A kinase (AURKA) inhibitors [4]. We therefore investigated the effect of GANT61 and AURKA inhibitor MK-0457 in M07e and WSU-AML cell lines carrying the CBFA2T3-GLIS2 fusion gene. Cell lines were incubated for 48 h with either single drugs or a combination of the two drugs at a constant ratio of 1:10 (GANT61:MK-0457). The combined treatment showed a higher cytotoxic effect when compared to each single drug, and the two inhibitors displayed a synergistic effect on cell growth, as indicated by the CI value (Additional file 6: Figure S4).

This work provides preliminary data from preclinical in vitro and ex vivo studies targeting pediatric AML with CBFA2T3-GLIS2 fusion gene. Although further investigation will be required to confirm these results, our experience with GANT61 represents a preliminary background for further evaluating in vivo the inhibition of GLIS2-mediated transcription in AML harboring the CBFA2T3-GLIS2 fusion gene.