Unravelling the Complexities of Hedgehog Mediated Signal Transduction in Acute Myeloid Leukaemia and Normal Haematopoiesis

Primary cilia regulate hedgehog (Hh) signal transduction; these sensory organelles being present on most mammalian cells. Aberrant Hh activity has been implicated in malignancy with defective primary cilia expression linked to disease. The role the Hh pathway plays in AML has not been fully elucidated and it is unclear whether AML cells express primary cilia. Our aims were (1) to determine the presence of primary cilia and (2) further evaluate the role of Hh signalling in AML.

Primary cilia were identified in all AML (n=23), and 20% of normal (n=10) bone marrow trephines (BMTs) by immunocytochemistry (ICC). Primary cilia were not identified in AML cell lines, primitive (n=4) or mature hematopoietic cells (n=6) isolated from peripheral blood, suggesting they are lost once cells migrate from the bone marrow (BM) microenvironment.

Despite the heterogeneity of AML, analysis of 76 primary AML samples demonstrated clear evidence of Hh pathway activity with up regulation of SMO (p<0.001), confirmed by ICC (n=8), and Gli-1 (p<0.05) in 55% of cases. Sanger sequencing (n=36) did not identify a mutation in SMO to account for the increased activity, although genomic instability was apparent with a high level of SNPs (mean 4; range 3-8), and silent and missense mutations (mean 1; range 1-3 per sample) identified. Samples with a >5-fold increase in SMO (n=28) showed a >5-fold up regulation of genes associated with chemoresistance and poor survival, including MECOM and FOXM1 (both p=0.0001), ABCC1, HOXA3, HOXA9, TWIST1 and SNAIL1 (all p<0.001), KIT (p<0.01), CD44, MMP2 and DNMT3b (all p<0.05); with a 4-fold reduction in pro-apoptotic genes BAK and BAX. Next Gen sequencing (n=9) confirmed an aggressive phenotype with 67% of patients having a c-Kit mutation and 46% (n=13), stratifying within the high risk category.

The BM microenvironment is important in AML. Immunohistochemistry (IHC) on human AML BMTs (n=37) enabled analysis within this unique environment. Proteins associated with poor prognosis and a more aggressive phenotype were up regulated 2-30 fold, including BCL-2 (p<0.0001), CD44 (p<0.001) and Vimentin (p<0.05). SHH was significantly down regulated (p<0.001) within the blast population. Secreted SHH, measured by ELISA, was up regulated suggesting paracrine activity. Impaired post translational modification of SHH was demonstrated with protein located within the nuclei by IHC and ICC. Nuclear expression of SHH was limited to primitive (CD34+) cells and absent from mature (CD14+, CD15+) cells. This correlated with a 20-fold reduction in HHAT the acetyltransferase involved in Hh processing in primitive cells compared to MNCs (p=<0.01).

In vitro, SMO inhibition with cyclopamine reduced cell proliferation in myelomonocytic cell lines (Kasumi-1 p<0.05, MOLM-13 p<0.01, MV4-11 p<0.05, OCI-AML3 p<0.001 and THP-1 p<0.01). No change in early or late apoptosis was seen. Cell cycle arrest was demonstrated with an increase in G0-G1 within OCI-AML3 (p<0.05) and a 3-fold reduction in cell division. Cells demonstrated a striking morphological change with increased cytoplasm, granules and loss of nucleoli. Reduction in cell proliferation and morphological changes were also observed when SMO was knocked down using a SMO siRNA. OCI-AML3 cells demonstrated increased expression of CD11b (p<0.001) and CD11c (p<0.0001) by FACS in keeping with a more mature phenotype. Gene expression correlated with a 2-fold reduction in early lineage markers, GATA1, SCL, NAB2 (p<0.05) and 3-4 fold up regulation of genes involved in differentiation and maturation, PU1, GMCSF and GCSF (all p<0.05), and IRF8 (p<0.01). Primary AML samples (n=6) were treated with the clinical grade SMO antagonist, PF-04449913, alone and in combination with cytarabine. Combination treatment caused a significant reduction in proliferation (p<0.05) with an increase in apoptosis. LTCIC showed a trend towards differentiation with a greater number of CFU-M and CFU-G than CFU-GM compared to no drug control (n=4).

This is the first report to demonstrate primary cilia on hematopoietic cells, with an increased frequency observed in AML. Their absence when cells migrate from the BM fits with their function and suggests a 'switching off' of the Hh pathway occurs on maturation. The ability of SMO inhibition to cause differentiation, in genetically diverse AML cell lines and primary AML is promising. SMO inhibition should continue to be explored as a potential therapy in AML.