Spred1 Insufficiency in the Hematopoietic and/or Vascular Compartments of the Bone Marrow (BM) Niche Promotes Aggressive Leukemogenesis in Chronic Myelogenous Leukemia (CML)

Spred1, a member of the Sprouty family of proteins and a negative regulator of RAS-MAPK signaling, is highly expressed in normal hematopoietic stem cells (HSCs) where it negatively regulates self-renewal activity. Lack of Spred1 function has been associated with aberrant hematopoiesis (Tadokoro, 2018). Spred1 knocked-out (KO) mice fed with high-fat diet develop a myeloproliferative phenotype (Tadokoro, 2018), and lower SPRED1 expression in acute myeloid leukemia associates with a poor outcome (Li, 2015; Olsson, 2014; Pasmant, 2015), suggesting a potential role of this gene as a tumor suppressor in myeloid malignancies. In CML, however, the role of Spred1 has not been fully dissected. Thus, we generated Spred1 KO CML (i.e., Spred1^-/-SCLtTA/BCR-ABL) mice by crossing Spred1 KO (a gift from Dr. Yoshimura, Japan) with inducible SCLtTA/BCR-ABL CML mice. Spred1 KO mice showed increased cell cycling of BM long-term HSCs (LTHSCs; Lin^-Sca-1⁺c-kit⁺Flt3^-CD150⁺CD48^-; G0: 62% vs 76%), and increased white blood cell (WBC) counts [14 vs 5.9 k/ul at 12 weeks (w) old, n=15 per group, p<0.0001], as compared to wt mice. Upon B/A induction by tetracycline withdrawal, Spred1^-/-SCLtTA/BCR-ABL mice had higher WBC (102.5 vs 12 k/ul at 4 w, n=15 per group, p<0.0001), more pronounced splenomegaly (spleen weight: 0.28g vs 0.19g, n=4 per group, p=0.06) and a significantly shorter survival (median: 39 vs 83 days, n=23 per group, p<0.0001) than Spred1 wt CML mice. In Spred1^-/-SCLtTA/BCR-ABL mice, we observed a more rapid expansion of circulating mature myeloid cells (CD11b⁺Gr-1⁺ cells: 63% vs 25%, n=8 per group, p<0.01) and a deeper decrease of BM LTHSCs (1,385 vs 2,164 per femur, n=5 per group, p<0.01) and increase of spleen LTHSCs (27330 vs 18546, n=5 per group, p<0.01) at 4 w after B/A induction compared with Spred1 wt CML mice. Further, we found a higher fraction of Spred^-/-SCLtTA/BCR-ABL mice (33% vs 10%) developed lymph node enlargement, with infiltration with pro-B lymphoblastic cells (B220⁺CD43⁺CD19⁺IgM⁻) compared with Spred1 wt CML mice. Altogether these features suggested that Spred1 insufficiency accelerates CML development and evolution to more aggressive phases of the disease. Since upregulation of Spred1 reportedly disrupts vascular integrity (Fish, 2008; Wang 2008), a finding that we have also confirmed in the BM niche, in order to evaluate separately the leukemogenic effect of Spred1 expression on different compartments of the BM niche, we generated the following conditional Spred1 KO strains: Spred1^flox(f)/fMxl-cre+ (Spred1 KO in HSCs, hereafter called Spred1^HSCΔ/Δ), Spred1^f/fTie2-cre+ [Spred1 KO in endothelial cells (ECs), hereafter called Spred1^ECΔ/Δ], Spred1^HSCΔ/ΔSCLtTA/BCR-ABL and Spred1^ECΔ/ΔSCLtTA/BCR-ABL by crossing SCLtTA/BCR-ABL with the above Spred1 KO mice. LTHSCs from Spred1^HSCΔ/ΔSCLtTA/BCR-ABL mice showed an increase in cell cycling, RAS/MAPK/ERK activity and Bcl-2 levels, and higher engraftment in recipient mice (blood: 9.7% vs 26.5% at 6w, 14.8% vs 42% at 8w, 14.7% vs 48% at 12w, n=10 per group, p<0.01), compared to Spred1 wt CML LTHSCs. Spred1^HSCΔ/ΔSCLtTA/BCR-ABL mice (n=15) showed enhanced leukemia progression (WBC: 19 vs 12 k/ul, p=0.004; CD11b⁺Gr-1⁺ in blood: 36% vs 25%, p=0.04 at 4 w after B/A induction) and a significantly shorter survival (median: 49.5 vs 83 days, p=0.01) compared to Spred1 wt CML mice (n=20). However, the disease in these mice appeared to be overall less aggressive than global Spred1 KO CML (i.e., Spred1^-/-SCLtTA/BCR-ABL) mice (WBC: 19 vs 102 k/ul; CD11b⁺Gr-1⁺ in blood: 36 vs 63%; Survival: 49.5 vs 39 days), suggesting that Spred1 depletion in other non-hematopoietic cell compartments may also be important for leukemogenesis. In fact, Spred1^ECΔ/ΔSCLtTA/BCR-ABL mice (n=8) showed enhanced leukemia progression (WBC: 26 vs 9.8 k/ul at 4 w after B/A induction, p=0.02), a trend for a reduced survival (median: 56 vs 83 days, p=0.09), and increased arteriolar vascularization, compared to Spred1 wt CML mice (n=20). Mechanistic studies on how endothelial Spred1 insufficiency co-participates in leukemogenesis are ongoing. Altogether our results support a role of Spred1 insufficiency in distinct BM niche compartments to produce a more aggressive CML phenotype, likely through different, but complementary mechanisms. Spred1 may therefore emerge as a novel target for advanced CML.