Outcomes for patients with childhood acute lymphoblastic leukaemia (ALL) are increasingly successful. Nevertheless, 20% of cases relapse due to failure to eradicate the disease, emphasising the need for more effective treatment. We and others have shown that multiple ALL subpopulations contain leukaemia initiating cells (LIC) that have the capacity to engraft and generate leukaemias in immune deficient mice. In addition, we have shown that some of these populations are resistant to current therapies and these cells may be responsible for disease relapse. Therefore, future therapies should be designed to target and eradicate all LIC populations. Preliminary results from our whole genome microarray analyses on bulk B-ALL cells and LIC subpopulations have indicated that certain key proteasome pathway and Hsp90 genes are similarly expressed across the LIC subpopulations. Therefore, inhibitors of these genes may be effective on all LIC subpopulations. Proteasome and heat shock protein (Hsp) 90 inhibitors are promising agents in cancer therapy. Bortezomib, a specific proteasome inhibitor, currently used in multiple myeloma, disrupts various signalling pathways leading to cell cycle arrest and apoptosis. Hsp90 is involved in the maturation and stabilisation of a wide range of oncogenic client proteins involved in survival and tumour progression. Hsp90 inhibitors, such as 17-DMAG and Celastrol have been shown to suppress B- and T-ALL cells, while sparing normal haemopoietic stem cells (HSC). 17-DMAG targets the ATP-binding domain of Hsp90 and Celastrol disrupts the Cdc37-Hsp90 complex, so when used in combination their potential to target leukaemia cells might be improved. The aim of this study was to investigate the effects of Bortezomib and the Hsp90 inhibitor combination 17-DMAG and Celastrol (Hsp90i) in childhood ALL. The viability and functional capacity of specific LIC subpopulations from 4 B-ALL cases and 5 CD34+/CD38- HSC samples following drug treatment were assessed by flow cytometry and functional assays. Cells were stained with Annexin V, 7AAD and antibodies against CD34 and CD19 (B-ALL) or CD34 and CD38 (normal cells). Following treatment with 1nM Bortezomib, the proportions of surviving bulk ALL cells, CD34+/CD19+, CD34+/CD19- and CD34-/CD19+ LIC subpopulations were significantly reduced to ≤23.6±21% compared to untreated cells (P≤0.025). The CD34-/CD19- subpopulation demonstrated resistance to Bortezomib with 61.6±28% remaining viable (P=0.144). Encouragingly, all LIC subpopulations responded to Hsp90i treatment (0.1nM Celastrol + 10nM 17-DMAG) with the proportions surviving ranging from 3.0±4% to 10.5±5% compared to untreated cells (P≤0.002). Normal cells were largely unaffected by Bortezomib and Hsp90i (75.9±15% and 90.8±29% surviving, respectively P=0.22). When ALL subpopulations were sorted, treated with Bortezomib then set up in long-term culture, viable cell numbers decreased significantly compared to untreated samples, even in the CD34-/CD19- subpopulation (up to 725-fold from 11.6±16x105 cells to 2.4±1x103, P=0.013). Treatment with Hsp90i also resulted in a significant reduction in viable cells after 4 weeks in culture (up to 161-fold to 8.0±9x103, P=0.015). The ex vivo effects of these drugs on LIC proliferation have also been assessed using NOD/LtSz-scid IL-2Rγc null (NSG) mice. High levels of leukaemia engraftment were observed in NSG with untreated bulk T-ALL cells and LIC subpopulations (60.3%-86.8% leukaemia). Results to date indicate that treatment with Bortezomib, prior to inoculation, prevented engraftment of bulk and all LIC in most cases or suppressed it to only 1.3±2.3% leukaemia cells in 1 NSG inoculated with bulk ALL (P=0.02). Similarly, treatment with Hsp90i prevented engraftment in bulk and LIC subpopulations in 2 of 3 cases. In the third case, a relapse sample, engraftment of some LIC was prevented and significantly reduced in NSG injected with bulk (68.73%), CD34+/CD7+ (24.06%) and CD34-/CD7+ (48.36%) cells (P=0.02). None of the treatment conditions have significantly affected the engrafting capacity of normal HSC (16.21±13.7% with untreated cells, 5.3±9% with Bortezomib P=0.11 and 5.8±7% with Hsp90i P=0.12). In conclusion, proteasome and Hsp90i show promising results in the killing of most LIC in B-ALL and T-ALL and further in vivo studies are warranted to explore their full potential.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.