Dysregulated Expression of HSP70 Isoforms in Acute Myelogenous Leukemia (AML)

Abstract 1370

Stress-inducible heat shock protein 70 (HSP70) is a major cytoprotective factor and a molecular chaperone that interacts with HSP90 to form a multi-chaperone complex. Cancer cells are highly dependent on this complex due to their increased demand for protein synthesis. HSP70 overexpression inhibits apoptosis and has been associated with drug resistance. However, the contribution to drug resistance in AML of specific HSP70 isoforms remains unknown. As there is growing interest in therapeutically targeting HSP70, we investigated the expression of 7 different HSP70 isoforms in AML primary cells and leukemia cell lines and their response to a novel HSP70-inhbitor, YK5. A panel of 12 leukemia cell lines and 11 primary samples was used to determine the expression of HSP70 and their response to YK5. We also evaluated the changes to the HSP70 isoforms when exposed to either heat shock or YK5. We found MV4-11, MOLM-13, and U937 sensitive to YK5 (LD50 = 1.18μM, 1.03μM, and 2.31μM at 24 hours, respectively). In contrast, OCI-AML3, TUR and THP-1 were more resistant to the inhibitor. (LD50 = 9.92μM, 9.74μM, and 8.84μM at 24 hours, respectively). Non-tumor cells, however, were significantly less affected by treatment with YK5 (72% viable cord blood mononuclear cells after 24 hour treatment with 5μM YK5).

We found that the cell surface expression of HSP70 was higher in both cell lines and primary samples when compared their normal counterparts. Furthermore, quantitative PCR revealed that cell lines with higher levels of HSPA1A and lower levels of HSPA6 demonstrated higher sensitivity to YK5. Interestingly, higher levels of HSPA1A and lower levels of HSPA6 were also found in primary AML samples when compared to CD34+ cord blood cells, consistent with the relative insensitivity of normal cells to YK5. We further discovered, mining publicly available databases, that high levels of HSPA1A were associated poorer prognosis (p = 0.004), suggesting that YK5 would be beneficial to patients presenting high HSP70 expression. We also evaluated the effect of YK5 on the gene expression of the various HSP70 isoforms. Quantitative PCR revealed the ability of YK5 to downregulate HSPA6 and HSC70 (HSPA8) in both cell lines and primary samples. Strikingly, all HSP70 isoforms exhibited similar fold changes upon heat shock in primary samples, CD34+ cord blood cells, and leukemia cell lines, indicating that the cellular stress response is not damaged in AML. However, the specificity of HSP70 inhibition to leukemia cells and not normal cells suggests a dysregulated set of client proteins and increased dependency on HSP70 to maintain leukemic homeostasis.

In summary, we have found dysregulated expression of the HSP70 isoforms HSPA1A and HSPA6 in leukemia cells and that the expression levels of these isoforms correlate to the sensitivity of YK5-mediated HSP70 inhibition (HSPA1A: p=0.0012 and r²=0.801, HSPA6: p=0.0011 and r²=0.847).

*KKS and JFR contributed equally to this project