Influence of Bone Marrow Stromal and Leukemic Cells on Cytarabine Chemo-Toxicity in Acute Myeloid Leukemia (AML)

Mesenchymal stromal cells (MSC) are known to protect leukemic cells from drug-induced toxicity within the bone marrow (BM) niche, however, less is known about leukemic impact on supportive MSC. The nucleoside-analogue, cytarabine (Ara-C), is a front-line therapy for acute myeloid leukemia (AML), entering cells via the human equilibrative nucleoside transporter (hENT1). Over a third of AML patients do not show continued response to Ara-C-based regimens, with chemo-resistance linked with repressed hENT1 availability in some patients, while other mechanisms remain unknown. In addition to chemo-resistance, DNA damage caused by chemotherapeutics such as Ara-C can persist in BM-MSC, which remain of host origin following allogeneic stem cell transplantation. This genotoxicity hinders cellular functionality, and may be implicated in long-term hematopoietic failure and secondary malignancies.

This study aimed to further elucidate chemo-resistance mechanisms, with particular focus on the contribution of leukemic cells to stromal cell toxicity; aiming to uncover potentially targetable features of resistant AML and reduce treatment burden on the BM.

Primary MSC were isolated from BM aspirates from patients both at diagnosis and post-treatment; following ethical approval and informed consent. MSC cultures were confirmed by immunophenotype (flow cytometry) and differentiation capacity (cytological staining) and used in a similar manner to that of cell lines. AML (HL-60, K562) and stromal (HS-5) cell lines were mono- or co-cultured using trans-well inserts for 24h, prior to 1-24h treatment with 25µM Ara-C (equivalent to in vivo standard dose; 100-200mg/m²). Cytotoxicity was monitored by viability and proliferation (CFSE tracing) assays, and chemo-sensitivity assessed with a drug-efficacy screening tool (bacterial bioluminescent biosensor). Genotoxicity was determined by micronucleus and alkaline comet assays, assessing division abnormalities and DNA fragmentation respectively. Differential cytokine secretion utilised an array, with quantification by ELISA.

In co-culture, stromal cells were sensitised to drug-induced cytotoxicity, while leukemic cells were themselves protected from treatment. Genotoxicity was also significantly increased in stromal cells (p=0.0397), while being significantly decreased in leukemic cells when co-cultured (p=0.0089), conferring with cytotoxicity findings. Similarly, BM-MSC from previously treated patients had significantly higher genotoxicity than patients at diagnosis (p=0.0138). While stromal cell proliferation remained unchanged regardless of intervention, data suggest increased proliferation in co-cultured leukemic cells compared to cells cultured alone. Chemo-sensitivity also increased in stromal cells in co-culture, while the opposite was seen for leukemic cells. Seven of 32 cytokines were differentially secreted by cell lines in co-culture compared to combined values from mono-cultured cells; CCL2, CXCL1, G-CSF, GM-CSF, IL-6, MIF and Serpin E1. Of these, the inflammatory cytokine MIF, macrophage migration inhibitory factor, was decreased in co-culture (p<0.0001), and has been implicated in the progression of other malignancies. Separation of cells following co-culture and treatment uncovered opposing MIF secretion profiles in cells with high (HL-60) and low (K562) sensitivity to Ara-C. Despite differential secretion, neither MIF, nor the MIF inhibitor ISO-1, had significant effects on chemo-sensitivity when cells were cultured in each condition for 24 hours.

Chemo-resistance is evidently a network of complex, interlinking mechanisms which are not easily identified in vitro. MIF remains a likely candidate for studies into AML chemo-resistance, with research ongoing. This study shows for the first time that the co-culture of AML and MSC alters the genotoxic effect of chemotherapy. Future research utilising larger patient cohorts is required to fully understand how cells in the BM micro-environment can be targeted. This could potentially improve not only the overall outcome for AML patients, but reduce the cytotoxic and long-term genotoxic complications of current therapies.