1 Supplementary Figures
2
3 An immature subset of neuroblastoma cells synthesizes retinoic acid and
4 depends on this metabolite
5
6
7 Tim van Groningen, Camilla U. Niklasson, et al.
8
1
Supplementary Figure 1 a control 1 µM RA 5 µM RA 10 µM RA
S - 26% S - 24,9% S - 19,9% S - 19,1%
691-MES
S - 47,1% S - 28,3% S - 21,8% S - 22,5%
691-ADRN EdU
Propidium Iodide
b control 1 µM RA 5 µM RA 10 µM RA
S –21,2% S -15,9% S -11% S -11%
717-MES
S - 28,3% S - 10% S - 6,1% S - 7,4%
717-ADRN EdU
Propidium Iodide 9 Supplementary Figure 1 – related to Figure 1.
10 Proliferation of MES and ADRN cells in response to RA.
11 A, B. EdU incorporation assay and Propidium-Iodide (PI)-staining for MES and ADRN cells
12 from patient 691 (in A) and patient 717 (in B). 691-MES, 691-ADRN, 717-MES and 717-
13 ADRN cells were treated with 1, 5 and 10 μM RA for 5 days and analyzed by FACS for PI
14 (x-axis) and EdU (y-axis).
15
2
Supplementary Figure 2
a b c MES (n=9) ADRN (n=26) Retinol ALDH1A1 ALDH3B1 500 80 RDH10 kDa 400 - 55 Retinal 60 ALDH1A1 300 ALDH1A3 - 56 ALDH DEAB 40 YAP1 - 70 200 RA + RAR BMS493 GATA3 - 48 ER50089 20
mRNA expression mRNA 100 mRNA expression mRNA Total AKT - 60 RARE-luciferase 0 0
d e g 8 691-MES 691-ADRN 717-MES 717-ADRN 691-MES SH-EP2 691-ADRN 0 120 8 *** *** SH-SY5Y 100 +ROL 0 80 60 40 - ROL
Cell number (%) 20 0 0 100 150 20 f SH-EP2 +ROL -ROL -ROL -ROL DEAB (µM) 0 control +100 nM RAL +100 nM RA 20 717-MES SH-SY5Y 717-ADRN 0 120 20 *** *** NBLW-MES 100 0 691-MES 80 20 NBLW-ADRN 60 0 40 20 Cell number (%) 717-MES 0 0 100 150 DEAB (µM) 16 Supplementary Figure 2 – related to Figure 1.
17 MES neuroblastoma cells have an active RA synthesis pathway.
18 A. Schematic overview of the core RA synthesis pathway, according to15,16. Retinol is
19 converted to retinal by RDH10. Retinal is converted to RA by ALDH enzymatic activity, the
20 latter can be inhibited by DEAB. RA associates with Retinoic Acid Receptors (RARα,
21 RARβ, RARγ). The RARα inhibitor ER50891 and the pan-RAR inhibitor BMS493 can block
22 the RA/RAR transcriptional activity. A RARE-luciferase reporter gene responds to an
23 endogenous source of RA.
24 B. mRNA expression of ALDH1A1 (left) and ALDH3B1 (right) genes, measured by
25 Affymetrix profiling in MES (n = 9, in orange) and ADRN (n = 26, in blue) cell lines.
26 C. Western blot analysis of ALDH1A1, ALDH1A3, the MES-marker YAP1 and the ADRN-
27 marker GATA3 in isogenic cell line pairs 691-MES and 691-ADRN, 717-MES and 717-ADRN
28 and SH-EP2 and SH-SY5Y. Cell lines of MES or ADRN phenotype are shown in orange and
29 blue, respectively. Total AKT is shown as loading control.
30 D. ChIP-sequencing analyses of H3K27ac of the genomic regions around the ALDH1A1 or
31 ALDH3B1 genes. The y-axis represents reads per 20 million mapped sequences. The
32 genomic position of the genes on sense (green) or antisense (red) DNA strands is shown
33 on the x-axis.
34 E. Bright field images of 691-MES, 691-ADRN, 717-MES and 717-ADRN cultured in
35 neural stem cell medium with retinol (+ROL, upper panels) or without retinol (-ROL, lower
36 panels) for 14 days. Scale bars represent 50 µm.
37 F. Bright field images of 691-MES (upper panels) and 717-MES (lower panels) cells
38 cultured in the presence (+) or absence (-) of retinol (ROL) for 18 days. At day 14, 100 nM
39 Retinal (RAL) or 100 nM RA was added as indicated. DMSO was added as control. Scale
40 bar represents 50 µm.
41 G. CyQuant cell viability assay of isogenic cell-line pairs 691-MES/-ADRN (left) and 717-
42 MES/-ADRN (right) with increasing concentrations of the ALDH-inhibitor DEAB. MES cell-
43 lines are shown in orange and ADRN cell-lines in blue. Two-sided Student’s t-test assuming
3
44 equal variance was used to calculate statistical significance, *** p < 0.001. Source data are
45 provided as a Source Data File.
46
4
Supplementary Figure 3
a b e 691-MES 691-ADRN 717-MES 717-ADRN 80 7 7 7 - DEAB 72.3 +/ 4.1% 7 - DEAB - DEAB - DEAB 10 10 10 - 10 + 13.0 / 1.7% + + 60 - 72.8 / 2.4% 31.0 /- 3.4% 6 6 6 6 - 10 10 10 10 40 5 5 5 5 SSC-A SSC-A
10 20 10 10 10 Aldefluor+ (%) Aldefluor+ 4 4 4 4 0 10 10 10 10 2 10 3 10 4 10 5 10 6 10 7 1010 2 10 3 10 4 10 5 10 6 10 7 10 2 10 3 10 4 10 5 10 6 10 7 10 2 10 3 10 4 10 5 10 6 10 7 1 2 3 4 7 7 7 + DEAB 0.1% 7 + DEAB + DEAB + DEAB 10 10 10 10 0.1% 0.1% 0.1% 6 6 6 6 10 10 10 10 5 5 5 5 SSC-A SSC-A 10 10 10 10 4 4 4 4 10 10 10 10 2 10 3 10 4 10 5 10 6 10 7 1010 2 10 3 10 4 10 5 10 6 10 7 10 2 10 3 10 4 10 5 10 6 10 7 10 2 10 3 10 4 10 5 10 6 10 7 Aldefluor signal Aldefluor signal Aldefluor signal Aldefluor signal
c d SH-EP2 SH-SY5Y NBLW-MES NBLW-ADRN 7 7 7 - DEAB 7 - DEAB - DEAB - DEAB 10 10 10 + 10 44.3 /- 5.0% + + + 5.6 / 4.1% 34.5 / 5.4% 3.5 /- 1.3% 6 6 6 6 - - 10 10 10 10 5 5 5 5 SSC-A SSC-A 10 10 10 10 4 4 4 4 10 10 10 10 2 10 3 10 4 10 5 10 6 10 7 1010 2 10 3 10 4 10 5 10 6 10 7 10 2 10 3 10 4 10 5 10 6 10 7 10 2 10 3 10 4 10 5 10 6 10 7 7 7 7 + DEAB 7 + DEAB + DEAB + DEAB 10 10 10 0.1% 10 0.1% 0.1% 0.1% 6 6 6 6 10 10 10 10 5 5 5 5 SSC-A SSC-A 10 10 10 10 4 4 4 4 10 10 10 10 2 10 3 10 4 10 5 10 6 10 7 1010 2 10 3 10 4 10 5 10 6 10 7 10 2 10 3 10 4 10 5 10 6 10 7 10 2 10 3 10 4 10 5 10 6 10 7 Aldefluor signal Aldefluor signal Aldefluor signal Aldefluor signal 47 Supplementary Figure 3 – related to Figure 1.
48 ALDH activity marks MES-type neuroblastoma cells.
49 A-D. Flow cytometry analysis of ALDH activity measured by Aldefluor-assay in four isogenic
50 MES and ADRN cell line pairs (A) 691-MES/691-ADRN, (B) 717-MES/717-ADRN, (C) SH-
51 EP2/SH-SY5Y and (D) NBLW-MES/NBLW-ADRN in the absence (-, upper panel) or
52 presence (+, lower panel) of the ALDH-inhibitor DEAB. The gate includes Aldefluorpositive cells
53 sensitive to DEAB-treatment. A representative FACS analysis from three independent
54 Aldefluor experiments is shown. The percentages of Aldefluorpositive cells are averages of
55 three independent measurements +/- standard deviation. Aldelfuor signal is shown on the x-
56 axis, side-scatter (SSC-A) is shown on the y-axis.
57 E. Summary of Aldefluorpositive cells in the four isogenic MES and ADRN cell line pairs. Shown
58 are the average measurements of three independent FACS experiments. In each
59 experiment, 10,000-20,000 cells were analysed. The error bar indicates standard deviation.
60 MES cell lines are shown in orange, ADRN cell lines are shown in blue. The percentage of
61 Aldefluorpositive cells is shown on the y-axis. Source data are provided as a Source Data File.
62
5
Supplementary Figure 4 a c 691-MES 717-MES 691-MES 717-MES 300 *** 250 *** 120 120 250 200 100 100 200 80 80 150 *** * ** ** 150 *** 60 60 100 100 40 40 *** *** *** 50 50 20 20 RARE-luciferase activity RARE-luciferase activity RARE-luciferase activity 0 0 0 RARE-luciferase activity 0 - - - - + + ER50891 control + + ER50891 control-- + + RA -- + + RA 0 0.1 1 10 0 0.1 1 10 µM ER50891 µM ER50891 b d 691-MES 717-MES 691-MES 717-MES 300 250 *** 120 120 250 100 100 *** 200 200 80 80 150 150 *** *** 60 60 100 100 40 40 50 50 20 *** *** *** 20 *** *** *** RARE-luciferase activity RARE-luciferase activity RARE-luciferase activity 0 0 0 RARE-luciferase activity 0 - + - + BMS493 - + - + BMS493 control-- +RA+ + RA control-- +RA+ + RA 0 0.1 1 10 0 0.1 1 10 µM BMS493 µM BMS493 e f 691-MES 717-MES shALDH1A3 shALDH1A3 140 140 691-MES 717-MES * ** shALDH1A3 shALDH1A3 120 120 100 100 - + IPTG - + IPTG 80 80 - 56 - 56 ALDH1A3 ALDH1A3 60 60
β-actin - 45 β-actin - 45 40 40 % RARE-luciferase … % RARE-luciferase % RARE-luciferase … % RARE-luciferase 20 20 RARE-luciferase activity 0 RARE-luciferase activity 0 - + IPTG - + IPTG 63 Supplementary Figure 4 – related to Figure 1.
64 Inhibition of a RARE-reporter gene with endogenous activity in MES cells.
65 A, B. RA reporter assay using a 3xRARE-luciferase construct in 691-MES and 717-MES
66 cells. Cells were incubated for T=24 hours in the presence (+) or absence (-) of 100 nM RA
67 in combination with 10 µM of the RARα inhibitor ER50891 (A) or 10 µM of the pan-RAR
68 inhibitor BMS493 (B).
69 C, D. RA reporter assay (3xRARE-luciferase reporter gene) in 691-MES or 717-MES cells
70 that were cultured in increasing concentrations of (C) the RARα inhibitor ER50891 or (D) the
71 pan-RAR inhibitor BMS493.
72 E. Western blot analysis of ALDH1A3 in 691-MES (left) and in 717-MES (right) cells with
73 IPTG-inducible shRNA targeting ALDH1A3. β-actin is shown as loading control. Molecular
74 weight (in kDa) is indicated on the right.
75 F. RA reporter assay (3xRARE-luciferase reporter gene) in 691-MES and 717-MES cells with
76 inducible shRNA targeting ALDH1A3. The normalized luciferase activities in A-D and G-H
77 are ratios between firefly-luciferase values of the 3xRARE reporter and renilla-luciferase
78 values of the transfection control. Error bars denote standard deviation. Two-sided Student’s
79 t-test assuming equal variance was used to calculate statistical significance, * p < 0.05, ** p <
80 0.01, *** p < 0.001. Source data for A-F are provided as a Source Data file.
81
6
Supplementary Figure 5
a b 691-MES 691-MES 717-MES 1200 *** control - + + - + + DEAB 50 µM 1 µM RA -- + -- + RA 1 µM 1000 800 = 62) =
600 n 400 ** -30 3 200 Score Migrated cells/HPF Migrated
0 1 2 3 4 5 6 7 8 9 10 11 12 - 1 5 10 - - - 1 5 10 - - µM BMS493 - --- 5 10 - --- 5 10 µM ER50891 MES RA-induced ( RA-induced MES
717-MES 1400 ** control c 1200 1 µM RA 691-MES 717-MES 1000 800 100 *** 120 ** 600 * 80 100 400 80 200
Migrated cells/HPF Migrated 60
0 1 2 3 4 5 6 7 8 9 10 11 12 60 - 1 5 10 - - - 1 5 10 - - µM BMS493 40 - --- 5 10 - --- 5 10 µM ER50891 40 20 20 Migrated cells/HPF Migrated cells/HPF 0 0 01 50 2100 3 01 50 2100 3 DEAB (µM) DEAB (µM) 82 Supplementary Figure 5 – related to Figure 2.
83 ALDH-activity and RA induce a motility programme in MES cells.
84 A. Transwell migration assay of 691-MES and 717-MES cells in the absence (white) or
85 presence (grey) of 1 µM RA in combination with increasing concentrations of the RARα
86 inhibitor ER50891 or the pan-RAR inhibitor BMS493. Cells were allowed to migrate for 48
87 hours. Source data are provided as a Source Data File.
88 B. Heatmap visualization of the MES-RAinduced gene signature in 691-MES (left) or 717-MES
89 cells (right) treated for 72 hours with 50 µM DEAB or DMSO with or without 1 µM RA and
90 analyzed by Affymetrix mRNA profiling. The list of RA-target genes in MES cells is available
91 from Supplementary Table 2.
92 C. Cell migration assay of 691-MES and 717-MES cells treated with the ALDH-inhibitor
93 DEAB. Cells were pre-incubated with DEAB for 4 days prior to seeding in a Boyden
94 chamber to assess migration. Error bars in A and C denote standard deviation. Two-sided
95 Student’s t-test assuming equal variance was used to calculate statistical significance, * p <
96 0.05, ** p < 0.01, *** p < 0.001. Source data are provided as a Source Data File.
97
7
Supplementary Figure 6
a 691-ADRN 717-ADRN
0 1 2 3 1 2 3 0 1 2 3 1 2 3 Time (days) -- - - + + + -- - - + + + RA 1 µM =98) n ADRN-RA_up ( = 67) = n
ADRN-RA_up ADRN-RA_down ADRN-RA_down ( -3 3 Score
b RA-induced genes
MES ADRN
52 10 88 98 Supplementary Figure 6 – related to Figure 2.
99 RA induces distinct gene-sets in MES and ADRN cells.
100 A. Heatmap visualization of genes regulated by RA in ADRN cells. Expression is shown in
101 691-ADRN and 717-ADRN cells cultured for indicated time points in the presence (+) or
102 absence (-) of 1 µM RA. Cumulative signature scores for both up- and down-regulated RA
103 responsive gene-sets are indicated at the bottom for each time point. Time is indicated in
104 days. The list of RA-target genes in ADRN cells is available from Supplementary Table 2.
105 B. Venn diagram depicting the overlap of MES-specific RA-induced genes (ROL-RA
106 responsive in MES cell lines 691-MES and 717-MES) and ADRN-specific RA-induced genes
107 (RA responsive in ADRN cell lines 691-ADRN and 717-ADRN).
108
8
Supplementary Figure 7
a b c d RA induced score ALDH1A1 ALDH1A3 ALDH3B1 MES score MES MES score MES score MES score MES
Mouse SCP score Mouse SCP score Mouse SCP score Mouse SCP score
RA induced score ALDH1A1 ALDH1A3 ALDH3B1 MES score MES MES score MES MES score MES score MES
Human SCP score Human SCP score Human SCP score Human SCP score 109 Supplementary Figure 7 – related to Figure 4.
110 Analysis of RA gene signature scores in human neuroblastoma.
111 A. Analysis of mRNA signature scores for mouse SCP cells (x-axis) and human
112 neuroblastoma MES cells (y-axis) in a panel of INSS stage 4 human neuroblastoma tumors
113 (n=183). Each tumor is represented by a dot. A signature score of the RAinduced target genes
114 in MES cells is visualized for each tumor. The signature scores represent the summed z-
115 scores of genes in the signature.
116 B-D. mRNA expression (2log-transformed) of (B) ALDH1A1, (C) ALDH1A3, (D) ALDH3B1 on
117 the INSS stage 4 neuroblastoma shown in panel A.
118 E. Analysis of gene expression signature scores for human SCP cells (x-axis) and for human
119 neuroblastoma MES cells (y-axis) in a series of 183 INSS stage 4 neuroblastoma, similar to
120 the analysis in A.
121 F-H. mRNA expression (2log-transformed) of (F) ALDH1A1, (G) ALDH1A3, (H) ALDH3B1 on
122 the INSS stage 4 neuroblastoma shown in panel E.
123
9
Supplementary Figure 8
a b c SH-SY5Y xenografts (T=15 days SH-SY5Y xenografts (T=18 days treatment) treatment)
1000 1000 n.s.
3 n.s. 3 800 800 n.s. 600 600 n.s. n.s. n.s. 400 400 Relative mRNA Relative expression 200 200 Tumor volume mm Tumor volume mm 0 0 0 2,5 5 10 0 2,5 5 10 Retinoic Acid (mg/kg/day) Retinoic Acid (mg/kg/day)
d e KCNR xenografts (T=11 days treatment) KCNR xenografts (T=15 days treatment) 1400 1400 n.s. n.s. 3 3 1200 1200 n.s. 1000 1000 n.s. 800 800 n.s. n.s. 600 600 400 400 Tumor volume mm Tumor volume mm 200 200 0 0 0 2,5 5 10 0 2,5 5 10 Retinoic Acid (mg/kg/day) Retinoic Acid (mg/kg/day) 124 Supplementary Figure 8 – related to Figure 4.
125 Analysis of ALDH1A3 expression and RA-treatment in vivo.
126 A. mRNA expression of ALDH1A3 in cells from three neuroblastoma PDX models (LU-NB-1,
127 LU-NB-2, and LU-NB-3) and in the neuroblastoma cell line SK-N-BE(2)c.
128 B-D. Retinoic acid treatment of neuroblastoma xenografts. SH-SY5Y (B, C) and KCNR (D, E)
129 xenografts were treated with 0, 2.5, 5 or 10 mg/kg body weight for 5 consecutive days,
130 followed by 2 days off treatment for each week during the experiment. Average tumor volume
131 (n=3 mice per treatment group) is plotted for SH-SY5Y at T=15 and T=18 (B, C) as well as
132 for KCNR at T=11 and T=15 (D, E) days after start of treatment. Source data of B-D are
133 provided as a Source Data File.
134
10
135 Supplementary References
136
137 15. Niederreither, K. & Dolle, P. Retinoic acid in development: towards an integrated
138 view. Nat Rev Genet 9, 541-53 (2008).
139 16. Rhinn, M. & Dolle, P. Retinoic acid signalling during development. Development 139,
140 843-58 (2012).
141
142
143
11