CXCL14 overexpression in stromal cells suppresses CML CD34⁺CD38⁻ cell growth in vitro and engraftment in the NSG-SGM3 mice. (A) Scheme representing the experimental design of coculture system. BM CD34⁺CD38^– cells from 5 to 10 patients with newly diagnosed CML and 7 age-matched healthy donors (NBM) were sorted and cocultured with CXCL14-overexpressed NIH3 stromal cells (NIH3-CXCL14) or control NIH3 cells (NIH3-CTRL) for 3 to 7 days and analyzed by FACS and LTC-IC assay or transplanted into sublethally irradiated NSG-SGM3 mice. (B) Total cell numbers and LTC-ICs from NBM or CML CD34⁺CD38⁻ cells cocultured with control or NIH3-CXCL14 stroma. The statistical difference was determined by paired t test. (C) The frequency of CD34⁺CD38^– and CD15⁺CD66B⁺ myeloid cells generated from NBM or CML CD34⁺CD38^– cells cocultured with control or NIH3-CXCL14 stroma. (D) Representative FACS profiles showing gating strategy of CD15⁺CD66B⁺CD33⁻ myeloid cells and CD34⁺CD38^– cells after the cocultures. The numbers in the panels are the percentage of positive cells within total live (PI^–) cells. (E) Representative FACS profiles showing gating strategy of human CD45⁺ CML cells engrafted in the NSG-SGM3 recipient mouse BM. The numbers in the panels are the percentage of positive cells within total live cells. (F) Reduced CML engraftment from the CML CD34⁺CD38^– cells cultured with NIH3-CXCL14 stroma in the BM of NSG-SGM3 mice at 10 to 12 weeks after transplantation. The tested CML cells were transplanted at 3 to 4 days after the coculture. Data were from 2 independent experiments on BM from 3 patients with CML. Each dot represents engraftment in a single recipient mouse. Horizontal bars indicate median values. The statistical difference was determined by unpaired t test.