Fig 2: The effects of CD164 on stem cell marker and sphere formation. a) A quantitative mRNA analysis of ALDH-1, CD44, Nanog, Nestin, Oct4 and Sox2 in hOSE and hOSE-CD164 cells was performed by qRT-PCR, and mRNA levels are expressed relative to individual GAPDH mRNA. The presented data are the means of three experiments (means ± S.D.; n = 3). b) Spheroid formation of hOSE and hOSE-CD164 and shCXCR4 cells after culture in ultralow plates for 7 days. **p < 0.01. c) Western blot analysis of hOSE and hOSE.CD164 cell lysates were subject to against antibodies for CSC markers, including Nanog, Oct4 and Sox-2. a-tubulin was used as a loading control.

Fig 3: Analysis of cell growth and sphere formation in M1-2 and NHRI-HN1 cells. (A) Representative growth curves for M1-2 and NHRI-HN1 cells analyzed by MTS assay. The results were expressed as the fold change to the results at day 1. (B) Cell cycle analysis of M1-2 and NHRI-HN1 cells by Fluorescence Activated Cell Sorting (FACS). (C) Sphere-forming activity of M1-2 and NHRI-HN1 cells was assessed by sphere culture. Left: The presentative pictures for sphere formation at 50× magnification. Right: The average sphere diameter and number of spheres are shown. (D) Analysis of CSCs in M1-2 and NHRI-HN1 cells by FACS using anti-Prominin-1 and anti-CD44 antibodies. Representative data are shown in the left panel. The relative level of Prominin-1+/CD44+ cells is shown in the right panel. The Prominin-1+/CD44+ ratio was calculated by normalizing the mean percentage of Prominin-1+/CD44+ NHRI-HN1 cells to that of M1-2 cells. (E) Immunoblot analysis of stemness-related transcription factors Klf4, Nanog, and Oct4. The protein levels were normalized to internal control (a-tubulin). The ratios were calculated by dividing the normalized protein levels of NHRI-HN1 cells by those of M1-2 cells. Error bars represent SE; * p < 0.05; ** p < 0.01; *** p < 0.001.

Fig 4: Exogenous EFNA4-Fc enhanced cell migration and sphere formation of OEC-M1 cells. (A) Representative growth curves of OEC-M1 cells treated with 0.1 or 0.5 µg/ml EFNA4-Fc (n = 4) or 0.5 µg/ml IgG control (n = 4) based on MTS assay data. (B) Left: Representative CFSE cell division assay data from OEC-M1 cells treated with 0.1 or 0.5 µg/ml EFNA4-Fc (n = 2) or 0.5 µg/ml IgG control (n = 2). The interval bar indicates cell division signal on day 3. Right: The percentage of the cell population within the interval bar limits. (C) Left: Representative cell death analysis via PI/Annexin V double staining of OEC-M1 cells treated with 0.1 or 0.5 µg/ml EFNA4-Fc (n = 2) or 0.5 µg/ml IgG control (n = 2). Right: The percentage of cell death, including quadrants Q1, Q2, and Q3. (D) Left: Representative images of migrated cells. Scale bars, 100 µm. Right: Relative migration activity was determined by normalizing the mean number of migrated cells per field of EFNA4-Fc treated cells (n = 10) to that of the IgG-treated cells (n = 10). (E) Left: Representative images of tumorspheres. Scale bars, 100 µm. Right: Sphere formation in OEC-M1 cells treated with 0.1 or 0.5 µg/ml EFNA4-Fc or 0.5 µg/ml IgG control in sphere culture. Relative sphere formation activity in EFNA4-treated cells (n = 2) was determined by normalizing the mean number of spheres per well to that of the IgG-treated cells (n = 2). (F) Relative levels of TWIST, SNAIL, SLUG, OCT4, NANOG, and SOX2 mRNA in OEC-M1 cells treated with 0.1 or 0.5 µg/ml EFNA4-Fc or 0.5 µg/ml IgG as determined by qRT-PCR. The amplifications were first normalized to ß-actin (internal control). The relative mRNA expression in EFNA4-treated OEC-M1 cells (n = 3) was then normalized to that in IgG-treated cells (n = 3). Bars represent SE. *p < 0.05; **p < 0.01; ***p < 0.001.

Fig 5: Immunofluorescence staining showing NANOG nuclear reactivity; OCT4 reactivity is predominantly nuclear and scant in the cytoplasm; SOX2 immunoreactivity is localized to both nucleus and cytoplasm (Cy3: cyanine dye; DAPI: 4',6-diamidino-2-fenilindol).