BMP2/SMAD pathway is upregulated in erythroleukemic blasts and contributes to proliferation and self-renewal. (A) Quantitative reverse transcription polymerase chain reaction for Bmp2 in MEPs from WT, J^VF, PEL J^VFP^−/−, and PEL J^VFP^172/– mice (n = 5-6 mice per genotype). J^VFP^−/− vs WT, P = .0064; J^VFP^−/− vs J^VF, P = .0062; J^VFP^172/– vs WT, P = .007; J^VFP^172/– vs J^VF, P = .007. (B) Western blot (top) for BMP2 in c-kit⁺ spleen cells (side-by side duplicates for each murine genotype are presented) and immunohistochemistry (IHC) (bottom) for BMP2 in BM and liver. Magnification, 400×. (C) The fraction of pSMAD1/5/9–positive cells within CD45-MEPs measured by flow cytometry in WT (n = 4), J^VF (n = 11), PEL J^VFP^−/− (n = 12), and PEL J^VFP^172/– mice (n = 10). J^VFP^−/− vs J^VF, P < .0001; J^VFP^172/– vs J^VF, P < .0001. (D) Top: total number of colony-forming units (CFUs) generated by J^VFP^−/− and J^VFP^172/– erythroleukemic blasts (Lin^–CD45^–c-Kit⁺ spleen cells) containing sh. Control (Control), sh.Bmp2_1 (sh_1), and sh. Bmp2_2 (sh_2) (n = 4 for each group). J^VFP^−/−: sh_1 vs Control, P < .0001; sh_2 vs Control, P < .0001. J^VFP^172/–: sh_1 vs Control, P < .0001; sh_2 vs Control P < .0001. Middle: representative colonies from each group (magnification, 25×). Bottom: western blot for BMP2 in erythroleukemic blasts with and without Bmp2 knock down. (E) Methylcellulose replating assay for MPN J^VFP^−/− and J^VFP^172/– c-Kit⁺ BM cells containing sh. Control (Control), sh. Bmp2_1 (sh_1), and sh. Bmp2_2 (sh_2) (n = 4 for each group). For the first platting, 10 000 GFP⁺c-Kit⁺ cells were plated per well. For the second platting and beyond, cells from the prior platting were used for replating (10 000 cells per well). (F) Top: Methylcellulose replating assay for Bmp2-overexpressed (OE) in WT BM c-Kit⁺ cells (n = 4 for each group). For the first platting, 10 000 GFP⁺c-Kit⁺ cells were plated per well. For the second platting and beyond, cells from the prior platting were used for replating (10 000 cells per well). Second platting, P = .029; third platting, P = .0004. Bottom left: Representative CFUs in third round of replating (magnification, 25×). Bottom right: western blot for BMP2 in WT c-Kit⁺ cells with and without Bmp2 overexpression. (G) PB counts of recipients transplanted with J^VFP^−/− erythroleukemic blasts transduced with sh. Control (Control), sh. Bmp2_1 (sh_1), and sh. Bmp2_2 (sh_2) three weeks after transplantation. N = 4 to 5 mice per arm. White blood cell (WBC): sh_1 vs Control, P = .0214; sh_2 vs Control, P = .0059. HGB: sh_1 vs Control, P = .0101; sh_2 vs Control, P = .0105. (H) Percentage of GFP⁺ cells of PB in recipients transplanted with J^VFP^−/− erythroleukemic blasts transduced with sh. Control (Control), sh. Bmp2_1 (sh_1), and sh. Bmp2_2 (sh_2) three weeks after transplantation. N = 4 to 5 mice per arm. sh_1 vs Control, P = .001; sh_2 vs Control, P = .0005. (I) Kaplan-Meier survival analysis of recipients transplanted with J^VFP^−/− erythroleukemic blasts transduced with sh. Control (Control), sh. Bmp2_1 (sh_1), and sh. Bmp2_2 (sh_2). P value was determined by the log-rank test. N = 4 to 5 mice per arm. sh_1 vs Control, P = .0145; sh_2 vs Control, P = .0035. (J) Histogram representation of BMP2, SMAD1, SMAD5, ID1, and BMPR2 gene expression in 14 post-MPN AML cases (including 3 post-MPN M6 and M7) and 9 cases of polycythemia vera (PV). The false discovery rate (FDR)-adjusted P values were calculated by DESeq2. Data are represented as mean ± standard error of the mean (SEM). The unpaired t test was used to compare the mean of 2 groups in panels A, C, D, E, F, G, and H. *P < .05, **P ≤ .01, ***P ≤ .001.