Targeting MiR-126 Blocks Progression of Myelodysplastic Syndrome and Transformation into Acute Myeloid Leukemia

Clonal-analysis studies have demonstrated that MDS is possibly derived from clonal hematopoietic stem cells (HSC) or early myeloid progenitors. These aberrant hematopoietic stem and progenitor cells (HSPCs) are resistant to therapy and expand during relapse. Understanding mechanisms contributing to their homeostasis is critical. MiR-126-3p (miR-126) is highly expressed in myeloid leukemia stem cell and restrains cell-cycle progression (Nature Medicine, 2018). Additionally, higher miR-126 levels reportedly correlate with poor outcome in MDS patients (pts; Korean J Intern Med, 2019). Thus, we postulated that miR-126 may play a relevant role in MDS pathogenesis. Accordingly, we found miR-126 levels were negatively correlated with overall survival [higher miR-126 (n=28) vs lower miR-126 (n=28) expression: 804 days vs undefined, p<0.05] and associated with leukemia transformation (transformation vs no transformation: 34.55±12.18 vs 13.37±2.89, p<0.05) in a cohort of consecutive MDS cases seen at our institution. Notably, miR-126 levels were higher in CD34+CD38- relative to CD34+CD38+ (0.32±0.03 vs 0.18±0.02, p<0.01) from same MDS pts. Similarly, when analyzed NUP19/HOXD13 (NHD13) mouse which recapitulates well human MDS (PANS, 2008), we found miR-126 levels were higher in the more primitive subset of Linage-c-kit+sca-1+ (LSK) cells relative to a more mature cells (Linage-c-kit+sca-1-, LK) (1.55±0.36 vs 0.13±0.04, p<0.01). Knockdown of miR-126 in a human MDS derived cell line (MDS-L) using short hairpin RNA (shmiR-126) increased cell cycling (G1: KD 52.4%±1.2% vs ctrl 63.9%±3.0%, p<0.05) and apoptosis (KD 18.8%±1.2% vs ctrl 2.0%±0.3%, p<0.001).

We have recently developed a novel miR-126 inhibitor (miRsten) by linking a miR-126-specific antisense oligodeoxynucleotide (ODN) to a cytosine-guanine dinucleotide (CpG) ODN(Nature Medicine, 2018. We therefore tested the combination of miRsten with decitabine (DAC) on primary MDS CD34+ cells; cells exposed to miRsten were more sensitive to DAC treatment than controls cells (miRsten +DAC [combination]: 46.3%±3.2% vs DAC 30.5%±1.4%, p<0.05, n=4). MDS CD34+cells treated exvivo with the combination (comb) of miRsten and DAC for 72 hrs had reduced engraftment ability in NSGS mice compared with cells treated with either drug alone or vehicle at 12 wks post-BMT (miRsten 2.7%±0.3% vs vehicle 5.6%±0.3%, p<0.0001; comb 0.8%±0.2% vs vehicle 5.6%±0.3%, p<0.0001; comb 0.8%±0.2% vs miRsten 2.7%±0.3%, p<0.001; comb 0.8%±0.2% vs DAC 1.5%±0.2%, p<0.05, n=5). Dysplastic CD45.2+ BM cells from donor NHD13+ mice were transplanted into preconditioned congenic recipients expressing CD45.1. The transplanted animals were treated with vehicle, miRsten (10mg/kg/day), DAC (0.5mg/kg, 3 times/ week) or combination of miRsten and DAC for 3 wks (n=6). Combination treatment resulted in partial reversal of MDS-like disease, with a significant increase in hemoglobin, platelet, red blood cell and white blood cell. The combination also significantly reduced NHD13+ chimerism (miRsten 73.6%±2.9% vs vehicle 87.2%±2.3%, p<0.01; comb 50.2%±4.4% vs vehicle 87.2%±2.3%, p<0.001; comb 50.2%±4.4% vs miRsten 73.6%±2.9%, p<0.01; comb 50.2%±4.4% vs DAC 67.8%±4.1%, p<0.05, n=6). In secondary transplantation experiments, the mice receiving BM cells from comb-treated donors showed reduced NHD13+ CD45.2+ cell long-term engraftment compared with recipients of cells from donors treated with either drug alone or vehicle (miRsten 28.6%±5.1% vs vehicle 36.7%±7.1%, p<0.06; comb 9.1%±2.1% vs vehicle 36.7%±7.1%, p<0.0001; comb 9.1%±2.1% vs miRsten 28.6%±5.1%, p<0.01; comb 9.1%±2.1% vs DAC 23.3%±5.3%, p<0.05, n=7). We also monitored the leukemia transformation in a cohort of tertiary transplants. After donor cells robust engrafted, we treated the transplants as above four groups. Notably, leukemia transformation was delayed, evidenced by reduced WBC and blast counts as well as extended survival in recipients of cells from donors treated with comb (median survival: miRsten 102d vs vehicle 34d, p<0.0001; comb 188d vs vehicle 34d, p<0.0001; comb 188d vs miRsten 102d, p<0.0001; comb 188d vs DAC 67d, p<0.0001, n=10).

Taken together, our results indicate that miR-126 targeting may disrupt MDS cell growth in vivo; mechanistic studies to explain how miR-126 contributes to MDS maintenance and optimize therapeutic targeting of this miR-126 with miRsten are ongoing.