S153: DECRYPTING THE ROLE OF HSP90Α AND Β ISOFORMS TO OVERCOME RESISTANCE IN BCR-ABL1 LEUKEMIA

Background: The BCR-ABL1 fusion protein, which results from the chromosomal translocation t(9;22)(q34;q11) is detected in over 95% of CML, 25% of adult BCP-ALL and 3-5% of childhood BCP-ALL patients. The development of tyrosine kinase inhibitors (TKi) has enabled targeted treatment, however development of resistance is common and the BCR-ABL1 leukemia subtype is still associated with poor prognosis. Among the chaperone proteins involved in maintaining the homeostasis of cancerous cells, heat shock protein 90 (HSP90) have been widely studied, due to its crucial involvement in stabilizing a variety of oncogenic proteins, including BCR-ABL1 and STAT3/5. Therefore, inhibition of HSP90 can be an effective alternative approach to target TKi resistant and BCR-ABL1T315I mutant expressing cells. In the past, over 17 HSP90 inhibitors (HSP90i) have been evaluated in clinical trials, however associated resistances (e.g. HSR) and dose limited toxicity have thus far precluded their clinical approval. Aims: In mammalian cells, there are two predominant cytosolic isoforms of HSP90, a stress-inducible HSP90α and a constitutively expressed HSP90β isoform. In the past, the distinct roles of the isoforms were studied by applying pan- and isoform-targeted HSP90i. We aimed to decrypt the role of cytosolic HSP90 isoforms using genetically edited models and to identify novel therapeutic vulnerabilities to target therapy refractory BCR-ABL1+ leukemia. Methods: To study the kinetic implications of the resistance mechanism associated with pharmacological targeting of HSP90, we generated a transient knockdown (Kd) and a stable CRISPR-Cas9 based knockout (KO) model of HSP90α/β in BCR-ABL1+ cells. The edited cells were analyzed in ex vivo functional assays (e.g. IF staining and WB) and their transplantation efficiency was determined in NSG mice. Global transcriptomic and proteomic profiling was performed in conjunction with high throughput drug screening (HTDS) to identify therapeutic vulnerabilities upon loss of HSP90α/β. Results: HSP90α/β-KO or -Kd did not affect the expression of other HSP90 paralogues (TRAP1 and GRP94). Only HSP90β-KO cells displayed hyperactive pro-survival HSR. Other reported HSP90 isoform-dependent clients, such as Survivin, CDK4 and CDK6 were also not affected upon KO or Kd of HSP90α/β. As HSP90 is involved in stabilizing, proper folding and subcellular localization of BCR-ABL1 protein, immunofluorescence imaging demonstrated two fold higher abundance of BCR-ABL1 foci (cytoplasmic/nucleocytoplasmic region) in HSP90α KO cells, with hyperactive BCR-ABL1 and downstream signaling (pBCR-ABL and pSTAT5a). When KO cells were transplanted into NSG mice (n=5 mice per arm) the engraftment of the HSP90α-KO cells was significantly reduced with a prolonged overall survival of the animals (19 days, p=0.0083) as compared to HSP90β-KO and control group. To find possible target (s) in order to circumvent resistance associated with the use HSP90i, we performed HTDS with HSP90α/β KO cells, which revealed distinct sensitivities toward certain classes of inhibitors. One of these classes was CDK7i, which was differentially active in HSP90α KO cells. These results can be explained by our global transcriptomic results, which revealed hyperactive androgen receptor signaling in HSP90α KO cells (GSEA), which is known to be a prognostic marker for CDK7 inhibition. Summary/Conclusion: In vivo targeting of HSP90α is a promising approach to target BCR-ABL1+ leukemia cells. Moreover, the combination with CDK7i may help to overcome the resistance associated with clinical use of pan HSP90i.