Supplementary Figure 1. Characterization of ALDH-positive cell population in MCF-7 cells. (a) Expression level of stem cell markers in MCF-7 cells or ALDH-positive cell population by qpcr. Data represent mean ± s.d. (n=3). (b) MCF-7 cells or ALDH-positive cell population were treated with 5-fluorouracil (5-FU; 100 M) or doxorubicin (10 M). After 3 days, cell viability were analyzed by MTS assay. Data represent mean ± s.d. (n=3). (c) Tumorigenicity of MCF-7 cells or ALDH-positive cell population. Balb/c nude mice were subcutaneously injected with 1 10 5 MCF-7 cells or ALDH-positive MCF-7 cells. Tumor incidence was indicated by tumors/injections at 6 week after injection.
Supplementary Figure 2. Expression of S1PRs in MCF-7 cells. Each expression of S1PRs was analyzed by RT-PCR using MCF-7 cdna. HUVECs, which express all S1PRs, were used as a positive control.
Supplementary Figure 3. Effect of dihydro-s1p on ALDH-positive cell population in MCF-7cells. Dose-dependent effects of dihydro-s1p in the proportion of ALDH-positive cell population. Data represent mean ± s.d. (n=3).
Supplementary Figure 4. Effect of S1PR3 knockdown on mammosphere formation in MCF-7cells. (a) Expression level of S1PR3 in shrna-transduced MCF-7 cells. Data represent mean ± s.d. (n=3). (b) Mammosphere-forming efficiency of shrna-transduced MCF-7 cells after stimulation with S1P (100 nm, 72 h). The mammospheres were counted and the percentage of mammosphere-forming cells was determined as mammosphere-forming efficiency (%).Data represent mean ± s.d. (n=3).
Supplementary Figure 5. S1P-induced increase in ALDH-positive cell population in MDA-MB-231 cells. (a) Representative flow data with ALDH substrate in the presence or absence with S1P (100 nm, 3 days) in MDA-MB-231 cells. (b) After treatment with S1PR3 antagonists TY52156 (TY; 1 M), CAY10444 (CAY; 10 M) or a S1PR2 antagonist JTE013 (JTE; 10 M), MDA-MB-231 cells were stimulated with S1P and ALDH-positive cell population were then analyzed. Data represent mean ± s.d. (n=3). (c) After MDA-MB-231 cells were treated with S1P (100 nm) for 24 h, expression levels of Hes1, Gli1 and Dkk1 were analyzed by qpcr. Data represent mean ± s.d. (n=3). (d) After treatment with DAPT (5 M), MDA-MB-231 cells were stimulated with S1P and ALDH assay were then performed. Data represent mean ± s.d. (n=3).
Supplementary Figure 6. S1P-induced increase in ALDH-positive cell population in human lung cancer A549 cells, human prostate cancer LNCaP cells, human glioma U251 cells and human ovarian cancer OVCAR-5 cells. (a) Representative flow data with ALDH substrate in the presence or absence with S1P (100 nm, 3 days) in A549 cells, LNCaP cells, U251 cells or OVCAR-5 cells. (b) After pretreatment with 1 M TY52156 (TY) or 5 M DAPT, the cells were stimulated with S1P (100 nm) and ALDH assay were then performed. Data represent mean ± s.d. (n=3).
Supplementary Figure 7. Effect of S1P on Hes1 expression and ADAM17 activity in ALDH-positive cell population. (a) Stimulation with S1P (100 nm, 24 h) induced Hes1 expression in ALDH-positive MCF-7 cells. Data represent mean ± s.d. (n=3). (b) Stimulation with S1P (100 nm, 24 h) induced Hes1 expression in ALDH-positive A549, LNCaP, U251 and OVCAR-5 cells. Data represent mean ± s.d. (n=3). (c) Stimulation with S1P (100 nm, 4 h) induced ADAM17 activity in ALDH-positive MCF-7 cells. Data represent mean ± s.d. (n=3).
Supplementary Figure 8. Effects of Hedgehog and Wnt inhibitor on each target gene expression in MCF-7 cells. (a) After MCF-7 cells were treated with of 200 ng ml -1 Sonic hedgehog with or without cyclopamine (10 M), expression of Gli1 were analyzed by qpcr. (b) After MCF-7 cells were treated with of 50 ng ml -1 Wnt3a with or without PNU74654 (10 M), expression of Dkk1 were analyzed by qpcr. Data represent mean ± s.d. (n=3).
Supplementary Figure 9. Effects of overexpression of NICD on ALDH-positive cell population in MCF-7 cells. (a) After overexpression of N2ICD, N3ICD or N4ICD, ALDH-positive cell population was analyzed in MCF-7 cells. Data represent mean ± s.d. (n=3). (b) Immunoblotting of N3ICD in MCF-7 cells treated S1P (100 nm) or LPA (100 nm). N3ICD-overexpressed cell lysates were used as a control.
Supplementary Figure 10. Knockdown by sirna against Notch1 in MCF-7 cells. Expression level of Notch1 mrna (left) and protein (right) in sirna-transfected MCF-7 cells was analyzed by qpcr and immunoblotting. Data represent mean ± s.d. (n=3).
Supplementary Figure 11. S1P-induced ALDH-positive population is independent on Notch ligands in MCF-7 cells. (a) After treatment with S1P (100 nm), expression level of Notch ligands were analyzed by qpcr. Data represent mean ± s.d. (n=3). (b) Effects of sirna against JAG1, JAG2, Dll1 and Dll4 on DFO-induced target gene expression. Data represent mean ± s.d. (n=3). (c) Effect of neutralizing antibody to JAG1 on JAG1-Fc-induced Hes1 expression. Data represent mean ± s.d. (n=3). (d) Effect of neutralizing antibody to JAG1 on S1P-induced increase in ALDH-positive cell population. Data represent mean ± s.d. (n=3).
Supplementary Figure 12. S1P-induced activation of -secretase via S1PR3 and G i in MCF-7 cells. (a) After stimulation with 100 nm S1P or 100 nm LPA for 4 h, -secretase activity was analyzed in MCF-7 cells. (b) After stimulation with S1P in the presence or absence of CAY10444 (10 M) or PTX (0.1 g ml -1 ), -secretase activity was analyzed in MCF-7 cells. (c) Effects of constitutively active (CA) mutants of G i or G 12 on γ-secretase activity in MCF-7 cells. Data represent mean ± s.d. (n=3).
Supplementary Figure 13. The p38mapk inhibitor SB203580 inhibited the S1P-induced increase in ALDH-positive cell population via ADAM17 in MCF-7 cells. (a) MCF-7 cells were stimulated with S1P (100 nm) for the indicated times and were then immunoblotted with antibodies to phospho-p38mapk or p38mapk. Data represent mean ± s.d. (n=3). (b) Effects of SB203580 (1 M), U0126 (1 M) or LY294002 (20 M) on S1P-induced ADAM17 activation. Data represent mean ± s.d. (n=3). (c) Effects of SB203580 (1 M) or LY294002 (20 M) on S1P-induced Hes1 expression. Data represent mean ± s.d. (n=3). (d) Effects of SB203580 (1 M), U0126 (1 M) or LY294002 (20 M) on S1P-induced increase in ALDH-positive cell population. Data represent mean ± s.d. (n=3). (e) Effect of LY294002 (LY; 20 M) on E 2 (1 M)-induced Akt phosphorylation. (f) Effect of SB203580 on S1P-induced association between p38mapk and ADAM17.
Supplementary Figure 14. Effects of SphK activity on overexpression of SphK1 or SphK2 in MCF-7 cells. (a) After transfection with Flag-SphK1, HA-SphK2 or empty vector, expression levels of SphK were determined in MCF-7 cells by immunoblotting. (b) SphK1 activity was measured in SphK1- or SphK2-overexpressed MCF-7 cells. Data represent mean ± s.d. (n=3). (c) SphK2 activity was measured in SphK1- or SphK2-overexpressed MCF-7 cells. Data represent mean ± s.d. (n=3).
Supplementary Figure 15. Effect of overexpression of SphK on ALDH-positive cell population in MDA-MB-231 cells. After MDA-MB-231 cells were transfected with SphK1 or SphK2, ALDH assay were performed. Data represent mean ± s.d. (n=3).
Supplementary Figure 16. Role of SphK1 in ALDH-positive cell population. (a) Effects of ABCC1 sirna and Spns2 sirna on the SphK1-induced ADAM17 activity, Hes1 expression and ALDH-positive cell population. Data represent mean ± s.d. (n=3). (b) Effect of ABCC1 inhibitor MK571 (5 M) on the SphK1-induced ADAM17 activation, Hes1 expression and ALDH-positive cell population. Data represent mean ± s.d. (n=3).
Supplementary Figure 17. Gating strategy for flow cytometry analysis of ALDH-stained samples in MCF-7 cells. Cells derived from disintegrated xenograft tumor samples were stained ALDH assay and APC-conjugated human TRA-1-85-specific antibodies. All samples were stained with 7-AAD just before flow cytometry. All samples were analyzed by sequential gating including main population (G1), single cells (G2, G3), viable (7-AAD-negative) cells (G4), human (TRA-1-85-positive) cells (G5). Negative controls with DEAB were used to establish a gate G6 (ALDH-positive) and G7 (ALDH-negative).
Supplementary Figure 18. Co-expression of S1PR3 and ALDH1 in tissue section from xenograft tumor. (a) Representative immunohistochemical images of S1PR3 (green) and ALDH1 (red) in sections of tumor xenografts derived from vector-, SphK1-, and SphK2-overexpressing ALDH-positive cells. Nuclei were counterstained by DAPI (blue). Bar = 30 m. (b) The number of S1PR3 and ALDH1-positive cells in xenografts derived from vector-, SphK1-, and SphK2-overexpressing ALDH-positive cells. The number of double-positive cells were counted in 5 fields per sample. Data represent mean ± s.d. (n=5).
Supplementary Figure 19. Original immunoblot data. Boxes indicate cropped images used in the figures.
Supplementary Figure 19. Continued
Supplementary Figure 19. Continued
Patients Age ER Her2 EGFR 1 40 + - - 2 45 + - - 3 45 - - - Supplementary Table 1. Characteristics of the patients
application forward reverse GAPDH qpcr GTCTCCTCTGACTTCAACAGCG ACCACCCTGTTGCTGTAGCCAA S1PR2 qpcr GCAAGTTCCACTCGGCAATGT AGCCAGAGAGCAAGGTATTGGC S1PR3 qpcr ATCCTGCCCCTCTACTCCAAGA GTGCGTAGAGGATCACGATGGT Oct3/4 qpcr ACATCAAAGCTCTGCAGAAAGAA CTGAATACCTTCCCAAATAGAACCC Nanog qpcr CAGAAGGCCTCAGCACCTAC ATTGTTCCAGGTCTGGTTGC c-myc qpcr CACGAAACTTTGCCCATAGC GCAAGGAGAGCCTTTCAGAG Gli1 qpcr GTGCAAGTCAAGCCAGAACA ATAGGGGCCTGACTGGAGAT Hes1 qpcr AGCGGGCGCAGATGAC CGTTCATGCACTCGCTGAA Dkk1 qpcr GGGCGGGAATAAGTACCAG CATAGCGTGACGCATGCAG Jagged1 qpcr CGGGATTTGGTTAATGGTTATC ATAGTCACTGGCACGGTTGTAGCAC Jagged2 qpcr GGTCGTACTTGCACTCACAATACC GTAGCAAGGCAGAGGGTTGC Dll1 qpcr CCAAGCCCTGCAAGAATGGA GGTGGGCAGGTACAGGAGTA Dll3 qpcr GAGACACCCAGGTCCTTTGA CAGTGGCAGATGTAGGCAGA Dll4 qpcr GTGGGTCAGAACTGGTTATGGA TGCAGATGACCCGGTAAGAGT Notch1 qpcr CACTGTGGGCGGGTCC GTTGTATTGGTTCGGCACCAT ABCC1 qpcr ATGTCACGTGGAATACCAGC GAAGACTGAACTCCCTTCCT Spns2 qpcr TTACTGGCTCCAGCGTGA TGATCATGCCCAGGACAG S1PR1 RT-PCR GACTCTGCTGGCAAATTCAAGCGAC ACCCTTCCCAGTGCATTGTTCACAG S1PR2 RT-PCR GTCATCCTCTGTTGCGCCATTGTG AGGCTGAAGACAGAGGCCGAGAGC S1PR3 RT-PCR GCCCCATCCTCTTCAAGGCTCAGT GTGGGGCAGGTCTTCCTTGACCTT S1PR4 RT-PCR TCCAGCCTTCTGCCCCTCTACTCC CAGGAAGGCCAGCAGGATCATCAG S1PR5 RT-PCR ACAACTACACCGGCAAGCTC GCCCCGACAGTAGGATGTT GAPDH RT-PCR CGGAGTCAACGGATTGGTCGTAT AGCCTTCTCCATGGTGGTGAAGAC Supplementary Table 2. List of primer sequences for PCR