Microenvironment contamination is mediated by direct cell-cell interactions and diffusible particles but not free dye molecules. HL60 cells were stained using 4μM DiI for 10 minutes in PBS without FCS at 37°C, washed, and plated in standard liquid culture conditions. As a control, eGFP-bright cells were treated in identical conditions without staining. Before and after culture, cell viability was higher than 90% (not shown). After 24-hour culture, cells were centrifuged (300g for 7 minutes at 4°C). The upper third part of the supernatant was harvested for ultracentrifugation (compare below); the remaining one was discarded. Cells were washed and 100 000 cells from each condition were plated on a confluent layer of HUVECs in a well of a 24-well plate. The supernatant from the first centrifugation step was further ultracentrifuged (61 600g for 2 hours 30 minutes at 4°C). The resulting supernatant (ultra-SN) and pellet (ultrapellet) were harvested separately. We determined the amount of ultra-SN and ultrapellet that would be equivalent to the number of cells plated on HUVECs and added this amount into separate confluent HUVEC wells. After 24 hours of culture, hematopoietic and stromal cells were recovered and analyzed by FACS after staining with anti–human CD45-APC and anti–human CD33-APC monoclonal antibody. DiI contamination of stromal cells (CD45⁻) after direct coculture or after exposure to similar amounts of ultrapellet or ultra-SN is displayed. No eGFP signal was found associated with stromal cells, whatever the condition (upper line). These data are representative of 2 experiments, which also included DiR.