Supplementary Table S1. Tumor samples used for analysis Sample# Age Tumor size (cm) pn- Stage Stage BNG (grade) ERα PR HER2 (FISH) Triple negative T1 46 3 N1a III 2 Pos Neg N T2 58 1 N(i-) I 3 Pos Neg N T3 51 2 N1mi II 2 Pos Pos N T4 74 2 N II 3 Pos Neg Y T5 33 3 N3a III 2 Pos Pos N T6 52 3 N3c III 2 Neg Neg Y T7 68 2 N1a II 3 Pos Pos N T8 47 2 N1mi II 2 Pos Neg Y T9 72 3 N IV 3 Neg Neg Y T1 68 2 N1mi II 2 Pos Pos N T11 67 2 N1a II 3 Pos Pos N T12 53 3 N2a III 3 Pos Neg N T13 55 2 N2a III 2 Pos Neg N T14 54 2 N(i-) II 2 Pos Pos N T15 52 3 N3a III 2 Pos Pos N T16 51 3 N II 2 Pos Neg N T17 32 4 N1 III 3 Neg Neg N yes T18 67 3 N3a III 2 Neg Neg N yes T19 55 4 N1a III 3 Neg Neg N yes T2 4 3 N3a III 2 Neg Neg N yes T21 62 2 N II 2 Neg Neg N yes T22 46 4 N1a III 3 Neg Neg N yes T23 65 2 N(i+) II 3 Neg Neg N yes T24 53 3 N2a III 3 Neg Neg N yes T25 5 2 N1mi II 2 Pos Pos N T26 63 4 N1a IV 3 Pos Pos Y T27 71 2 N(i-) II 2 Pos Pos N T28 3 2 N1a II 3 Pos Pos N T29 52 4 N3 III 3 Neg Neg N yes T3 47 2 N3b III 3 Neg Neg N yes T31 51 4 N3a III 3 Neg Neg N yes T32 4 3 N3a III 2 Neg Neg N yes T33 82 3 N II 3 Neg Neg N yes T34 54 4 N3c III 3 Neg Neg N yes T35 47 2 N3a III 3 Neg Neg N yes T36 29 2 N II 3 Neg Neg N yes T37 7 2 N II 2 Pos Neg N T38 46 4 N(i-) III 3 Neg Neg Y T39 46 4 N3 III 3 Pos Pos N T4 64 4 N3b III 3 Pos Neg N T41 63 1 N1 II 3 Pos Pos N
NMuMG Input H3K27Me3 IgG TGFβ (d) 2 4 8 2 4 8 2 4 8 E-cad Actin Supplementary Figure S1. TGF-β-induced EMT does not significantly alter the level of H3K27me3 at the E-cadherin promoter. NMuMG cells were treated with TGF-β1 (5 ng/ml) for different time intervals, and H3K27me3 at the E-cadherin promoter was analyzed by ChIP.
A Withdrawal (2d) -TGFβ +TGFβ (2 d) B Withdrawal (2d) +TGFβ (2 d) H3K9me2 TGFβ β - + WD - IgG + WD - + WD +TGFβ (12 d) NMuMG (TGFβ for 12 days) NMuMG (TGFβ for 2 days) Input +TGFβ (12 d) -TGFβ Input - H3K9me2 + WD - + WD IgG - + WD E-cad Actin C NMuMG (TGFβ for 2 days) TGFβ m + u m NMuMG (TGFβ for 12 days) WD u m u m + u m WD u m u 14 16 E-cad Days 2 4 12 Supplementary pp y Figure g S2. Changes g of cellular morphology, p gy H3K9me2 and DNA methylation y on the E-cadherin after TGF-β β with drawal from NMuMG cells treated with TGF-β for 2 and 12 days, respectively. (A) NMuMG cells were treated with TGF-β for 2 and 12 days, respectively, followed by TGF-β withdrawal for an additional 2 days. The morphological changes of these cells are shown in the phase contrast images. (B) Cells were treated as described above, H3K9me2 on the E-cadherin promoter was analyzed by ChIP assay. (C) Cells were treated as described above, DNA methylation on the E-cadherin promoter was examined by MSP analysis.
l Supplementary Data-Dong et al sig9a 5 -Aza- -dc TGFβ - + + + + E-cad NTC sisnail Vimentin N-cad Claudin3 Claudin7 Snail G9a Actin Supplementary Figure S3. G9a, Snail or non-target control (NTC) sirna was expressed in NMuMG cells followed by TGF-β treatment for 2 days or cells were treated with 5 -Aza-dC. Expression of E-cadherin, Vimentin, N-cadherin, Claudin 3 and 7, as well as Snail and G9a were examined by Western blotting. Actin was used as a loading control.
A Continuation o of Figure 4B B Input for Figure 4C Input for Figure 4D C IgG IP DA157 M BT T2 DA231 M M DA157 BT T2 M DA231 Dnmt 1 Dnmt 1 Dnmt 3b G9a Snail Dnmt 3b G9a Input Dnmt 3a Dnmt 3b Snail G9a Input Dnmt 3a Dnmt 3b G9a Snail Input Dnmt 1 Dnmt 3a Dnmt 3b Snail Supplementary Figure S4. Snail forms a complex with G9a and Dnmts. (A)Snail forms a complex with G9a and Dnmts. Endogenous Dnmt3b was immunoprecipitated from MDA-MB157, BT2 and MDA-MB231 cells, and bound endogenous Snail and G9a were examined by Western blotting. Input lysate used for the immunoprecipitation (Fig. 4B) was shown on the bottom panel. (B) Input lysate used for the immunoprecipitation in Figure 4C was shown. (C)Input lysate used for the immunoprecipitation in Figure 4D was shown.
A GST Pull-down IP-Snail Input B Snail 1 kda 75 GST-G9a G9a 5 GST-G9a 5 MW (kda) 1 75 37 37 25 Snail 25 Supplementary Figure S5. Snail interacts directly with G9a. (A)Full-length Snail was purified as a bacterially expressed GST fusion protein followed by excision and removal of the GST protein using TEV protease. Purified full-length Snail was incubated with various GST-G9a deletion mutants. After G9a was pulled-down, bound Snail was examined by Western blotting. (B) Purified full-length Snail was incubated with the various deletions of GST-G9a described above. After Snail was immunoprecipitated, the bound G9a deletion mutants were examined by Western blotting.
IgG Supplementary Data-Dong et al Cont rol BIX2 2194 Cont rol BIX2 2194 Dnmt 1 Dnmt 3a IP: G9a Dnmt 3b Snail G9a Dnmt 1 Dnmt 3a Input Dnmt 3b Snail G9a Supplementary Figure S6. Inhibition of G9a activity does not alter the interaction of G9a with Snail and Dnmts. Supplementary Figure S6. Inhibition of G9a activity does not alter the interaction of G9a with Snail and Dnmts. The G9a inhibitor BIX1294 (2.5 µm) was added to the freshly-prepared MDA-MB157 lysate during immunoprecipitation. After immunoprecipitating endogenous G9a, bound Snail and Dnmts were analyzed by Western blotting. Input lysate used for the immunoprecipitation was shown on the bottom panel.
Luminal CLBC (Basal B) BT47 74 MCF 7 T47D ZR75 MDA AMB361 BT54 49 HS57 78 SUM 1315 MDA AMB435 MDA AMB231 MDA AMB157 H3K9me1 H3K9me2 H3K9me3 H3K9Ac H3 Supplementary Figure S7. There is no significant difference in global H3K9 methylation and acetylation between luminal and CLBC cell lines. Cell extracts were prepared from five luminal and six CLBC cell lines, and the global level of H3K9 methylation and acetylation were analyzed by Western blotting.
Luminal CLBC (Basal B) Supplementary Data-Dong et al BT4 74 MCF F7 T47D ZR7 75 MDA AMB361 BT5 549 HS5 578 SUM M1315 MDA AMB435 MDA AMB231 % in nput MDA AMB157.6.4.2 IP-Snail Supplementary Figure S8. G9a-related repressive marks are enriched at the E-cadherin promoter in CLBC cell lines. Quantitative real-time PCR was performed to analyze ChIP samples from Fig. 6B. Results from three independent d experiments were presented (mean ± SD from three separate experiments)..6 IP-G9a % input.4.2 1.2 IP-H3K9me2 % input % inpu ut.9.6.3.6.4.2 IP-H3Kac
Luminal Luminal Triple-negative Luminal Triple-negative Samples: L1 L2 L3 L4 L5 L6 L7 L8 L9 L1 L11 L12 L13 L14L15L16 T1 T2 T3 T4 T5 T6 T7 T8 L17 L18 L19L2 T9 T1 T11T12T13T14T15T16 E-cad Claudin3 ERα Snail G9a Actin E-cad Claudin3 ERα Luminal Samples: L21 L22 L23 L24 L25 Snail G9a Supplementary Figure S9. Lysates of fresh frozen tumor samples from 25 cases of luminal breast cancer and 16 cases of triple-negative breast cancer were analyzed for the expression of E-cadherin, Claudin 3, ERα, Snail and G9a by Western blotting. Actin was used a loading control. Actin
IP-G9a IP-Snail IP-Dnmt1 E-cad input.15.12.12.12.9.9.9.6.6.6 NTC sisnail sig9a %.3.3.3 Supplementary Figure S1. Quantitative real-time PCR was performed to analyze ChIP samples from Fig. 7B. Results from three independent experiments are presented (mean ± SD from three separate experiments).
MDA-MB157 MDA-MB231 Control sirna + - - - + - - - G9a sirna - + - + - + - + Snail sirna - - + + - - + + G9a Snail β-actin Supplementary Figure S11. Snail, G9a or NTC sirna was expressed in MDA-MB157 and MDA-MB231 cells. After 48 h, expression of Snail and G9a were analyzed by Western blotting. Actin was used a loading control.
A B E-cad dherin mrn NA (qrt-pcr R) 3 2.5 2 1.5 1.5 NTC sisnail sig9a sisnail+g9a rase (%) E-cad lucife 4 NTC sisnail sig9a 3 sisnail+g9a 2 1 Supplementary Figure S12 (A)G9a, Snail or NTC sirna wereexpressedinmda-mb157 expressed cells. After 48 h, E-cadherin mrna was analyzed by quantitative real-time-pcr. (B) G9a, Snail or NTC sirna were co-expressed with the E-cadherin promoter luciferase construct in MDA-MB157 cells. After 48 h, luciferase activities were normalized and determined (mean ± SD in three separate experiments).
shntc shg9a-1 shg9a-2 actin F-a Supplementary Figure S13. Formation of lamellipodia in MDA-MB231 cells and stable transfectants with knockdown of G9a was analyzed by immunofluorescent staining with actin-phalloidin. Nuclei were stained with DAPI (blue).
A B 9 shntc shg9a-1.4 Cell l number (1 4 ) 6 3 shg9a-2 OD=57nm 3.3.2.1 24 48 72 h shntc shg9a-1 shg9a-2 Supplementary Figure S14. Knockdown of G9a expression does not affect the growth and proliferation of MDA-MB231 cells. (A)Cell growth rates of MDA-MB231 cells with stably expressing control vector or G9a shrna were analyzed by the cell count assay for a period of 3 days. Statistical analyses from two independent experiments with triplicate samples are plotted. (B) Cell proliferation rates of MDA-MB231 cells with stably expressing control vector or G9a shrna were examined by MTT assays. Statistical analyses from three independent experiments with triplicate samples are shown on the bar graph.
MDA-MB157 A B 12 NTC 12 sig9a 1 1 NTC sig9a Migration (%) 8 6 4 Invasion n (%) 8 6 4 2 2 Supplementary Figure S15. Knockdown of G9a expression inhibits migration and invasion of MDA-MB157 cells. (A)The migratory ability of MDA-MB157 cells with and without of knockdown of G9a expression was analyzed by wound healing assay as described in the Materials and Methods. Statistical analysis for the cell migration is shown on the bar graph (mean + SD from three independent experiments). (B) The invasiveness of MDA-MB157 cells with and without of knockdown of G9a expression was analyzed with a modified Boyden Chamber invasion assay as described in the Materials and Methods. The percentage of invasive cells is shown in the bar graph (mean ± SD in three separate experiments).
shntc shg9a-1 shg9a-2 4 tal Flux (pho oton/sec) 35 3 25 2 15 1 Tot 5 Supplementary Figure S16. MDA-MB231 cells with stably expressing control vector or G9a shrna were injected into the tail vein of SCID mice. To ensure an even number of cell injections from three groups, luciferase bioluminescence imaging measured photon flux after initial injection were taken and quantified.
A shntc shg9a-1 shg9a-2 D shntc shg9a-1 shg9a-2 shntc shg9a-1 E shg9a-2 G9a E-cad Vimentin B Mig ration (%) C Invas ion (%) β-actin 1 8 6 4 2 1 8 6 4 2 shntc shg9a-1 shg9a-2 shntc shg9a-1 shg9a-2 shntc shg9a-1 shg9a-2 Tu umor weight (g).8.7.6.5.4.3.2.1 shntc shg9a-1 shg9a-2 ases nodule Ma acro-metast Num mber of lung 25 2 15 1 5 shntc shg9a-1 shg9a-2
Supplementary Figure S17. Knockdown of G9a expression inhibits breast cancer MDA-MB435 cell migration and invasion in vitro and suppresses tumor growth and lung colonization in vivo. (A)MDA-MB435 cells stably expressing control vector or G9a shrna were examined for the expression of G9a, E-cadherin and vimentin by Western blotting. (B) The migratory ability of MDA-MB435 cells and the corresponding stable transfectants with knockdown of G9a expression were analyzed by wound healing assay as described in Materials and Methods. Statistical analysis for the cell migration is shown in the bar graph (mean + SD from three independent experiments). (C) The invasiveness of MDA-MB435 cells with stably expressing control vector or G9a shrna was analyzed with a modified Boyden Chamber invasion i assay as described d in Materials and Methods. The percentage of invasive i cells is shown in the bar graph (mean ± SD in three separate experiments). (D) MDA-MB435 cells with stably expressing control vector or G9a shrna were injected into the mammary fat pad of SCID mice. The growth of breast tumors was measured weekly by caliber. Mice were sacrificed when tumor of control group reached 1.5 cm. Tumors were removed and weighed (the mean of 6 animals + SEM). (E)MDA-MB435 MB435 cells with stably expressing control vector or G9a shrna were injected into the tail vein of SCID mice. Mice were sacrificed after 12 weeks; lung metastatic nodules were examined macroscopically or detected in paraffin-embedded sections stained with hematoxylin and eosin. The number of macro-metastatic nodules was presented (the mean of 6 animals + SEM).
NTC shg9a Clau2 Clau4 Clau6 Clau7 Clau2 Clau4 Clau6 Clau7 Clau9 Relative ex xpression (% %) 2 16 12 8 4 18 15 12 9 6 3 25 2 15 1 5 3 25 2 15 1 5 NTC shg9a Clau1 Clau11 Clau19 Clau2 Actin ive expressi ion (%) Relat Clau9 Clau1 Clau11 35 2 4 4 3 25 15 3 3 2 15 1 2 2 1 5 1 1 5 Clau19 35 3 25 2 15 1 5 Clau2 Supplementary Figure S18. Knockdown of G9a restored claudin expression in MDA-MB231 cells. The differentially expressed claudins from MDA-MB231 vector-control cells and the corresponding stable clone with knockdown of G9a expression were analyzed by either quantitative RT-PCR (left panel) or real-time PCR. Results from three independent experiments of real-time PCR were presented on the right panel (mean ± SD from three separate experiments).
1 1 3 shntc shg9a uciferase ac ctivity (%) L 8 6 4 2 uciferase ac ctivity (%) L 8 6 4 2 uciferase ac ctivity (%) L 25 2 15 1 5 Wnt/β-cat TGFβ E-cad Supplementary Figure S19. Activation of the β-catenin, TGF-β and cellular adhesion pathways in MDA-MB231 cells and corresponding stable transfectants with knockdown of G9a expression was analyzed by luciferase reporter assays (mean ± SD in three separate experiments).
A GSE11121 Supplementary Data-Dong et al ability Prob.6.8 1..2.4. p=1.999e-8 1 2 3 4 5 6 7 B Survival (days) Supplementary Figure S2 Identification of a 9-gene prognostic signature using breast cancer dataset GSE11121 Supplementary Figure S2. Identification of a 9-gene prognostic signature using breast cancer dataset GSE11121. (A)Kaplan Meier survival curve separates the tumors (from GSE11121 data set, training set) into three groups with expression of a 9-gene prognostic signature. (B) Heatmap of the 9-gene signature expression in 2 lymph node-negative breast cancer patients. Top bar, tumor grade (1=blue, 2=pink, 3=red).
A GSE234 lity Probabi 1. B.6.8..2.4 p=.469 1 2 3 4 5 Survival (days) Supplementary Figure S21. Validation of the 9-gene prognostic signature using breast cancer dataset GSE234. (A) Kaplan-Meier plot of patient survival as stratified using the 9-gene signature. (B) Heatmap of the 9-gene signature expression in 286 lymph node-negative breast cancer patients. Top bar: ER status (white, ER positive; black, ER negative).
A GSE1456-Disease-free Survival GSE1456-Overall Survival Supplementary Data-Dong et al 1. bility.6.8 Proba.2.4 1. p=.919 Probab bility.2.4.6.8 p=.197.. 5 1 15 2 25 3 5 1 15 2 25 3 B Survival (days) Survival (days) Supplementary Figure S22. Validation of the 9-gene prognostic signature using breast cancer dataset GSE1456. (A) Kaplan-Meier plot of disease-free and overall patient survival as stratified using the 9-gene signature. (B) Heatmap of the 9-gene signature expression in 159 breast cancer patients. Top bars: tumor grade (1=blue, 2=pink, 3=red) and tumor subtypes (normal-like, blue, luminal-a yellow, luminal-b orange, basal, pink, HER2, red).