Aberrant Hedgehog Signaling in Myelodysplastic Syndrome Accelerates Leukemic Transformation Via Smoothened-Dependent and Smoothened-Independent Gli1 Activation

Introduction: Myelodysplastic syndromes (MDS) are malignant hematopoietic stem cell disorders with increased risk of transformation to secondary acute myeloid leukemia (sAML) that carries uniformly dismal prognosis. The exact mechanisms driving this process remain obscure. Aberrant Hedgehog (Hh) signaling has been demonstrated in a number of solid tumors and hematologic malignancies. An emerging body of literature suggests the importance of Hh signaling in leukemic transformation of MDS. Moreover, a recent Phase II clinical trial using Smoothened (SMO) inhibitor with low-dose chemotherapy resulted in significant prolongation of overall survival (Lancet et al. Blood 2016). Thus, we sought to examine the clinical impact of aberrant Hh activation in MDS and further explore the mechanism of Hh-driven leukemic transformation.

Methods/Results:

To determine the pathway activity, we quantified the level of GLI1 in CD34+ cells from patients with MDS and healthy controls in published databases (N=17 and N=159, respectively; GSE58831) and while absent in controls, GLI1 was significantly upregulated in MDS patients (p<0.05). Analysis of serial samples revealed Hh activation at disease progression in 67% of patients (4/6). In addition, GLI1 levels were significantly higher in poor-risk AML with complex cytogenetics comparing to favorable risk disease (p<0.001, TCGA).

To further evaluate the impact of of Hh activation in MDS progression, we crossed Nup98-HoxD13 (NHD13) mice with mice conditionally expressing the constitutively active mutant of Smoothened (SmoM2). Double transgenic NHD13/SmoM2 mice succumbed to aggressive myeloid leukemia (median survival of NHD13/SmoM2 and NHD13 controls, 3 months vs. 13 months, respectively, p<0.0001) characterized by leukocytosis (p<0.05), splenomegaly (p<0.0001) and significant expansion of immature myeloid progenitors in peripheral blood (p<0.0001) and bone marrow (p<0.0001). Analysis of hematopoietic stem/progenitor (HSPC) compartment revealed expansion of phenotypic granulocyte macrophage progenitors (GMP) (6244 vs. 131487 mean GMPs/femur , in NHD13 v. NHD13/SmoM2 respectively, p<0.05). To determine whether Hh activation resulted in aberrant self-renewal in GMP compartment, we carried out an in vivo repopulation assay. While NHD13 GMPs showed only transient engraftment, characteristic of non-self-renewing progenitors, NHD13/SmoM2 GMPs robustly engrafted and significantly expanded over time, and all recipient animals developed aggressive leukemia within 2-3 months. The engraftment was sustained through secondary transplantation. Our results suggest that Hh activation resulted in abnormal self-renewal in normally non-self-renewing progenitors.

To delineate the mechanism leading to aberrant activation of self-renewal we compared the transcriptome of GMPs isolated from NHD13/SmoM2 and NHD13 controls. In addition to expected Hh activation, Gene Set Enrichment Analysis (GSEA) showed enrichment in gene signatures associated with other developmental pathways such as Wnt or TGF-β signaling as well as EWS/FLI1-protooncogen signature. Interestingly, these pathways have been found to act via direct activation of GLI1 .

Using human AML (TF-1) and MDS (MDS-L) cell lines we confirmed that GLI1 may be activated in the absence of SMO. GLI1 knock-down and/or inhibition with GANT61 resulted in growth arrest (p<0.05, both) and significantly lower clonogenic potential in vitro (p<0.05, both) more so than upstream SMO inhibition.

Conclusions: Our data suggest that aberrant Hh/GLI1 signaling is frequent in MDS and marks disease progression and leukemic transformation. In some cases, GLI1 can be activated in a SMO-independent manner and therapeutic strategies focused on targeting downstream Gli transcription factors may be more effective than currently used Smo inhibitors.