Stroma-Mediated Chemotherapy Resistance in Acute Myeloid Leukemia Cells

AML is a devastating malignancy with a relapse rate near 50% in children, despite very toxic chemotherapy. Once a child relapses, the chance of survival is very low. Therefore new, rational therapies for AML are desperately needed. Accumulating evidence shows that the bone marrow stromal environment protects a subset of leukemia cells and allows them to survive chemotherapy, eventually leading to recurrence. Our goal is to delineate the mechanisms underlying stroma-mediated chemotherapy resistance in AML cells, which could potentially lead to new therapies for AML.

We used two human bone marrow stromal cell lines, HS-5 and HS-27 for our studies. Both provide physical contact with AML cells, while HS-5 cells secrete many more cytokines and growth factors than HS-27 stromal cells. To verify the difference between HS-5 and HS-27 in their secreted soluble factors, both stroma-conditioned media were harvested and soluble factors were quantified by multiplex cytokine assay for 42 individual soluble factors. We detected 23 factors in HS-5 conditioned medium, including G-CSF, IL-6, and MCP-3 at very high levels. HS-27-conditioned medium contained only a few cytokines at similar levels as HS-5, e.g., VEGF and Fractalkine.

Next, we performed co-culture experiments to determine the ability of each stromal cell line to confer resistance to chemotherapy. Human AML cell lines (NB-4, THP-1 and Kasumi-1) were cultured alone or co-cultured with HS-5 or HS-27 cells, and treated with etoposide, mitoxantrone or cytarabine for 48 hours. Cells were then harvested and labeled with annexin V-FITC. Stromal cells were identifiable by stable mOrange expression, and the percentage of apoptotic AML cells (FITC positive and mOrange negative) was determined by FACS. Both HS-5 (p<0.001) and HS-27 (p<0.05) cells protected NB-4 and THP-1 cells from etoposide-induced apoptosis (apoptosis rate at the 3 uM dose: 86.6±1.4% NB-4 alone vs. 33.9±2.9% with HS-5 vs 60.7±2.5% with HS-27). The results with THP-1 were similar to NB-4. Using the same method, we demonstrated that both stromal cells protected NB-4 and THP-1 from the toxic effects of all three chemotherapy agents; Kasumi-1 were resistant to all three agents, even when cultured alone.

To delineate if the protection induced by stromal cells against chemotherapy was dependent on adhesion pathways and/or soluble factors, we performed Transwell co-culture assays. Different from regular co-culture, there is no physical contact between AML and stromal cells, while soluble factors secreted by stromal cells can reach AML cells. In the absence of physical contact, both stromal cells provided little protection for NB-4 and THP-1 against etoposide and cytarabine; while both NB-4 and THP-1 were still protected against mitoxantrone. Those results suggest that the protection provided by both stromal cells against etoposide and cytarabine mostly relies on cell-cell contact; as for mitoxantrone, soluble factors secreted by both stromal cells seem more important. Surprisingly, HS-5 and HS-27 provided similar degrees of protection against all three chemotherapies.

To discover genes in AML cells that are induced by interaction with stromal cells and may contribute to chemotherapy resistance, oligonucleotide microarray analysis was done using total RNA extracted from NB-4 and THP-1 cells cultured alone or co-cultured with stromal cells. We found that 43 genes were upregulated by HS-5, and over 1000 genes were either up- or down-regulated by HS-27. Among them, eighteen genes were upregulated by both stromal cell lines. Since HS-5 and HS-27 provided similar degrees of protection against chemotherapy, those eighteen commonly upregulated genes are likely to be important for stroma-induced chemotherapy resistance. Excitingly, seven out of those eighteen genes, e.g., including CYR61, CAV1, TM4SF1, have been reported to contribute to chemotherapy resistance in various cancer types. Further studies are underway to determine if those genes are responsible for stroma-induced chemotherapy resistance.

This study suggests that distinct pathways in the microenvironment mediate resistance to different chemotherapy drugs. Elucidating the precise drug-specific mechanisms involved is likely to result in promising combination therapies to reduce chemotherapy resistance and relapse, and thereby improve survival for children with AML.

No relevant conflicts of interest to declare.