The Role of the BMP Pathway in Sustaining CML Stem Cells in the Bone Marrow Niche

Despite improved patient outcome using tyrosine kinase inhibitors (TKIs), chronic myeloid leukemia (CML) patients require life-long treatment due to leukemic stem cell (LSC) persistence. LSCs reside together with mesenchymal stem cells (MSCs) and hematopoietic stem cells (HSCs) in the bone marrow (BM) niche, which they modify to their advantage whilst impairing normal hematopoiesis. To date it has proved difficult to understand how LSCs both dominate and alter the niche, and to effectively target LSCs with current therapies. Recent studies have shown intrinsic and extrinsic deregulation of the bone morphogenetic protein (BMP) pathway in CML. These cells show altered stem cell fate, persistence, and response to BMP receptor (BMPR) antagonists, which affect cell behavior including cell cycle, apoptosis, and expansion (Laperrousaz et al. 2013; Grockowiak et al. 2017; Zylbersztejn et al. 2018; Toofan et al. 2018).

We compared TKIs alone and in combination with a BMPR inhibitor to gain improved insights into BCR-ABL1 dependent and independent regulatory mechanisms within the niche environment using this therapeutic approach.

K562 CML cell line and CML CD34⁺ primary cells were used in this study in combination with HS5 stromal cell co-culture. CML cells were treated with single or combination treatments of imatinib, the dual SRC-ABL1 TKI saracatinib, and the BMPR inhibitor dorsomorphin for 4h or 72h with and without BMP4 stimulation/co-culture. HSC and LSC interactions were also investigated using our artificial 3D BM niche model which comprises of magnetically levitated MSC spheroids embedded in medical-grade collagen type I, mimicking the BM biological and mechanical microenvironment, along with a high throughput microfluidic MSC spheroid formation system for drug testing. Kinase screens were performed on CML CD34⁺ cells using a chip-based microarray assay (PamGene) following 4h drug treatments. Phosphorylation data were then utilized for upstream kinase and pathway analysis using the metacore platform. This analysis facilitated identification of proteins showing a change in expression of ≥0.5-fold across all (n=3) patient samples, which was validated by qPCR and immunoblotting. Fluidigm multiplex qPCR was utilized to assess changes in expression of early response, self-renewal and differentiation genes. Flow cytometry was performed to investigate apoptosis, cell cycle progression and proliferation, alongside colony assays of primary CD34⁺CP-CML samples (n=3) following treatment.

We demonstrated a synergistic mode of action upon inhibition of the BMP pathway in combination with TKI treatment, resulting in increased apoptosis (p< 0.001), altered cell cycle (G2-M, p< 0.01), fewer cell divisions, and a reduction in CD34⁺cells. Primary patient samples displayed differential gene expression in relation to response for genes involved in cell cycle (CDKN1A, CDKN2B, RB1; p< 0.05), self-renewal (PBX1; p< 0.01) and cell survival (GATA1, CKIT, p< 0.05). Kinase prediction analysis identified kinases involved in cell growth, development, differentiation, apoptosis, and cell-cell adhesion in treated sample lysates, with consequent pathway analysis highlighting transcription factors, ETS1, TP53 and C-MYC as main regulators across all common pathways. QPCR identified significant changes in expression of ETS1 (p< 0.001) and C-MYC (p< 0.01) following treatment. Within the top 10 deregulated pathways based on PamGene TK profiling we also identified GAB1 and GAB2 as common effectors, known to play an important role in growth and differentiation of myeloid cells, and which were demonstrated by immunoblotting to be considerably downregulated following 72h single and dual treatments in K562. Co-culture of CD34⁺ cells on stroma was chemo-protective, however dual treatment was still able to elicit strong anti-proliferative effects. We are currently investigating these promising findings in more detail using our 3D niche model and microfluidic spheroid platform.

Taken together, these results provide vital insights into the mechanisms by which CML cells respond to current treatments, which is critical for improving therapeutic approaches and avoiding patient resistance or relapse. A combinatorial approach targeting the BMP pathway with BMP antagonists or small molecule inhibitors together with second-generation TKIs could open up new therapeutic possibilities.