Supplementary Figure 2: Characterization of HES3 MIXL1^GFP/+ cells exposed to uniform BMP4 concentration in the microfluidic device.

(a) Combined graph of the quantification of MIXL1, T, SOX17 and CDX2 after 48 hours of constant and uniform exposure of HES3 MIXL1^GFP/+ cells to 50ng/ml BMP4 (from Fig. 1k). (b, c, d, e) Fluorescence pictures of indicated markers for HES3 MIXL1^GFP/+ cells after 48hours of exposure to constant and uniform 50ng/ml BMP4 in the microfluidic device. These experiments were repeated independently two times (with eight samples per repetition) and showed similar results. (b) Red arrows indicate CDX2-positive MIXL1-negative cells located at the edges of the colony. White arrows indicate MIXL1/CDX2 double positive cells. (c) The vast majority of the SOX17-positive cells co-express CDX2. White arrows indicate CDX2-positive/SOX17-negative cells. Red arrows indicate SOX17-positive/CDX2-negative cells. (d) Red arrows indicate MIXL1/SOX17 double negative cells present in patches at the edges of the colony. White arrows indicate MIXL1/SOX17 double-positive cells. (e) White arrows indicate cells negative for MIXL1, T and SOX2 with wider internuclear distances located at the edges of the colony in patches of various sizes. (f) Scheme that approximates the distribution of markers in a colony of HES3 MIXL1^GFP/+ cells after 48hours of constant and uniform exposure to 50ng/ml BMP4. (g) Computational simulation showing the local percentage of phospho-SMAD1 positive cells in a colony under a uniform and constant concentration of BMP4 at different time points. Top, distribution of exogenously provided BMP4. Bottom, local percentage of phospho-SMAD1 positive cells. Scale bar 200μm.