a b c Supplementary Figure S1 Phenotypic characterization of mammospheres a b c     (a) Representative phase-contrast photomicrographs of adherent cells and mammospheres formed after 11 days; ×10 objective. The single cell suspension from either fresh breast cancer or various different breast cancer cell lines was cultured under adherent and mammosphere conditions (left). Sphere frequency analysis. Primary mammospheres were enzymatically dissociated with trypsin to obtain a single-cell suspension. Cells were then seeded with a single cell/well in a 96-well plate, and mammospheres formed after 11 days. The number of mammospheres was counted (Right). Results are representative of three independent experiments. We observed that the primary cells (109 mammospheres/1000 single cells) formed more mammospheres than the cancer cell lines (0.6-2.5 mammospheres/1000 single cells). (b) Incidence of tumors induced by MCF-7 adherent cells and mammosphere cells in nude mice. When just 5×102 “mammosphere cells” were inoculated into seven mice, all seven generated tumors. All the mice injected with 10- or 100-times more cells developed tumors. In contrast, none of the mice inoculated with 5×102 or 5×103 adherent cells developed tumors, and tumors developed in only two out of five animals inoculated with 5× 104 adherent cells. (c) Tumor volume was determined as described in M & M section. The volumes of the tumors from MCF-7 mammosphere cells were greater than those from adherent cells.

Autophagic flux in mammospheres and adherent cells Supplementary Figure S2 Autophagic flux in mammospheres and adherent cells a b 1.4 1.2 1 p62/Actin 0.8 0.6 0.4 0.2 (a) Immunoblot analysis of p62 levels in MCF-7 adherent cells (left) and mammospheres (right) incubated in complete medium (CM) or with EBSS for the times indicated in the presence or absence of Baf A1 (final concentration: 100 nM) for 2 h. Actin immunoblotting was used as a loading control.(b) The p62/Actin ratio was determined using Bio1D software. Results are representative of three independent experiments. Control was set as 1 in adherent untreated cells. (c) MCF-7-GFP-LC3 adherent cells (top) and mammospheres (bottom) were incubated in complete medium (CM) or with EBSS for the times indicated in the presence or absence of Baf A1 (final concentration: 100 nM) for 2 h, fixed, and then visualized by confocal microscopy. Bars: 5 µm for adherent cells; 25 µm for mammospheres. (d) The number of GFP-LC3 dots was scored on ~50 to 100 cells. The data are presented as means ± SD from 3 independent experiments and analyzed using Student’s t-test (* p<0.05, ** p<0.01 ***, p<0.001).

Autophagy flux in mammospheres and adherent cells Supplementary Figure S2 Autophagy flux in mammospheres and adherent cells c d

a b Supplementary Figure S3 Degree of autophagy inhibition in the original cells and during the mammosphere assay in Beclin 1- knock down experiments a b (a) MCF-7 adherent cells were stably transfected with shRNA-mediated scramble (shSCR) or shRNA-mediated silencing of BECN (shBECN). Immunoblot analysis of LC3 I and LC3 II levels in stable clone adherent cells (left) and mammosphere (right) incubated in complete medium in the presence or absence of Baf A1 (final concentration: 100 nM) for 2h. Actin immunoblotting was used as a loading control. (b) The LC3 II/actin ratio was determined using Bio1D software. Results are representative of three independent experiments. Control was set as 1 in adherent untreated cells (left). The autophagic flux determined as the ratio between the LC3 II levels with Baf A1 and without Baf A1 (right). (* p<0.05, ** p<0.01).

a b c d Supplementary Figure S4 Beclin 1 is critical for the maintenance of breast CSCs and progenitor cells proliferation a b c d SK-3rd/SKBR3 adherent cells were stably transfected with empty pMKO.1 vector (control) or shRNA-mediated scramble (shSCR) or shRNA-mediated silencing of Beclin 1 (shBECN1). The single cell suspension of cells from each condition was cultured under mammosphere conditions. (a) Representative phase-contrast photomicrographs of adherent cells and mammospheres formed after 11 days; bars, 100µm (b) Quantification of mammospheres per 1000 cells. (c)The representative phase-contrast photomicrographs of a single mammosphere (×10 objective). (d) Quantification of the size of a single mammosphere. All the data are presented as means ± SD from 3 independent experiments. ***, p<0.001, as compared with shSCR cells.

Beclin 1 is critical for SK-3rd mammosphere tumorigenicity Supplementary Figure S5 Beclin 1 is critical for SK-3rd mammosphere tumorigenicity (a) SK-3rd adherent cells were stably transfected with shRNA-mediated scramble (shSCR) or shRNA-mediated silencing of Beclin 1 (shBECN1). The single cell suspension from each condition was cultured under mammosphere conditions. Representative phase-contrast photomicrographs are shown of tumors formed by injecting the indicated number of adherent and mammosphere cells into nude mice. (b) Tumor volume was determined as described in Materials and methods section. (c) Incidence of tumors formed by injecting SK-3rd mammosphere and adherent cells in nude mice following the shRNA-mediated silencing of Beclin 1.

Beclin 1 is critical for SKBR3 mammosphere tumorigenicity Supplementary Figure S5 Beclin 1 is critical for SKBR3 mammosphere tumorigenicity (d) Representative phase-contrast photomicrographs of tumors formed by injecting the SKBR3 cells into nude mice as described in (a). (e) Tumor volume was determined as described in (b). (f) Incidence of tumors formed by injecting SKBR3 mammosphere and adherent cells in nude mice following the shRNA-mediated silencing of Beclin 1.

Autophagy is critical for maintenance of breast CSCs Supplementary Figure S6 Autophagy is critical for maintenance of breast CSCs a b (a) Representative phase-contrast photomicrographs of mammospheres formed after 11 days; ×10 objective. The single cell suspension from MCF-7 mammospheres was seeded at a single cell/well in a 96-well plate in the presence of the inhibitors indicated. (b) The number (right) and size (left) of the mammospheres were quantified. Results are representative of three independent experiments.

Supplemental Figure 7 ATG7 is critical for maintenance of Breast CSCs and proliferation of progenitor cells a b MCF-7 MCF-7 mammospheres shSCR shATG7 shSCR shATG7 Atg7 GAPDH B c C shSCR shATG7 shSCR shATG7 MCF-7 adherent cells were stably transfected with shRNA-mediated scramble (shSCR) or shRNA-mediated silencing of Atg7 (shATG7). (a) Immunoblot analysis of Atg7 levels in stable clone adherent cells. GAPDH immunoblotting was used as a loading control. (b) Representative phase-contrast photomicrographs of mammospheres formed after 8 days; ×10 objective. The single cell suspension from each condition was cultured under mammosphere conditions and analyzed under microscopy. (c) The size and the number of the mammospheres were quantified.

d e Supplementary Figure 7 Degree of autophagy inhibition in the original cells and during the mammosphere assay in Atg7- knock down experiments d e (d) Immunoblot analysis of LC3 I and LC3 II levels in stable clone adherent cells (left) and mammosphere (right) incubated in complete medium in the presence or absence of Baf A1 (final concentration: 100 nM) for 2h. Actin immunoblotting was used as a loading control. (e) The LC3 II/actin ratio (left) and autophagic flux (right) was determined as described in supplemantary Figure S3b. (* p<0.05, *** p<0.001).

a b SSC-A ALDH1 (Green Fluorescence) Supplementary Figure S8 Quantification of ALDH+ cells numbers in MCF-7 SCR and MCF-7 shATG7 stable clones at an earlier time point following detachment a SSC-A ALDH1 (Green Fluorescence) b (a) ALDH activity in MCF-7 shSCR (left) and MCF-7 shATG7 (right) stable clones and 24 h after detachment was measured by flow cytometry using the ALDEFLUOR reagent in the presence or absence of the ALDH1 inhibitor diethylamino-benzyaldehyde (DEAB). The frames represent gates that depict ALDH1+ cells that were created based on cells treated with DEAB and then applied to untreated cells. Facs profiles of DEAB control (with DEAB), and ALDH staining (without DEAB). Numbers indicate the percentage of the ALDH1-positive cells; SSC: side Scatter. (b) Grap representation of percentage of the ALDH1-positive cell population identified by the ALDEFLUOR assay in MCF-7 shSCR and MCF-7 shATG stable clones in complete medium and 24 h after detachment in mammosphere medium.