CHAF1A promotes NB aggressiveness. a) CHAF1A is turned on upon DOX induction (1 µg mL⁻¹ for 72 h) in SHEP cells. Validation of CHAF1A overexpression by qPCR and western blotting is shown. Date are mean ± SD (n = 3); ****p < 0.0001; two‐sided unpaired t‐test. b) Proliferation assay in SHEP‐CHAF1A cells. Cells were cultured in normoxic and hypoxic (1% O₂) conditions for 0–96 h. Cell number was assessed by Cell Counting Kit‐8 and are indicated by absorbance (450 nm). Mean ± SD (n = 4); ****p < 0.0001; two‐way ANOVA with Sidak's multiple comparisons test. c) Migration and invasion analyses of SHEP‐CHAF1A cells upon induction of CHAF1A (48 and 72 h). Mean ± SD (n = 5–10); ****p < 0.0001; two‐sided unpaired t‐test. d) Cell cycle analysis of SHEP‐CHAF1A cells upon induction of CHAF1A (3 and 6 days). Mean ± SD (n = 4); **p < 0.01, ***p < 0.001, ****p < 0.0001; two‐way ANOVA with Dunnett's multiple comparisons test. e,f) GSEA Hallmark analysis in patients with high and low CHAF1A expression in two independent patient cohorts. False discovery rate (FDR) is computed using a Benjamini–Hochberg corrected two‐sided homoscedastic t‐test. Pathways are ranked by −Log₁₀ FDR (FDR < 0.25). g) Tumor formation upon activation of CHAF1A in an orthotopic mouse model. Low‐tumorigenic NB SHEP cells were injected into the renal capsule of NCr nude mice. Four‐week‐old mice were treated with control (n = 11) or DOX‐containing diet (0.625 g kg⁻¹, n = 12) for five weeks. Tumor incidence and tumor weights are shown. Data are the mean ± SEM; comparison of tumor incidence between CHAF1A OFF and CHAF1A ON mice was computed by two‐sided Fisher's exact test, p = 0.012.

CHAF1A blocks RA‐induced cell differentiation. a) Bright field images of neurite outgrowth and quantification of neurite length. RA‐sensitive NGP cells were treated with RA (5 μм) in the presence or absence of CHAF1A induction for 72 h. Neurite length was quantified by Image J2 and presented as mean ± SEM (n > 300, two biological replicates); ****p < 0.0001; two‐way ANOVA with Tukey's multiple comparisons test. Scale bar = 50 µm. b) TUJ1 immunofluorescence staining. Scale bar = 50 µm. c) qPCR analysis of neuron‐specific marker genes (MAPT, GAP43, and NGFR). Mean ± SD (n = 3); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; two‐sided unpaired t‐test. d) RA treatment (10 μм) and CHAF1A conditional KD (0–10 days) in RA‐resistant SK‐N‐AS cells. Neurite length is quantified by Image J2 and presented as mean ± SEM (n > 150, two biological replicates); ****p < 0.0001; one‐way ANOVA with Dunnett's multiple comparisons test. Scale bar = 100 µm. e) TUJ1 immunofluorescence staining. Scale bar = 50 µm. f) qPCR analysis of neuron‐specific genes (MAPT, GAP43, and NGFR). Mean ± SD (n = 2); *p < 0.05, **p < 0.01, ***p < 0.001; one‐way ANOVA with Dunnett's multiple comparisons test. KD = knockdown. FC = fold change.

CHAF1A blocks NC differentiation. a) Schematic presentation of early NCC events during zebrafish development. hpf = hours post fertilization. b) Spatial‐temporal expression of sox9b, crestin, and chaf1a in 16 hpf and 24 hpf embryos by hybridization chain reaction (HCR). A (anterior), P (posterior), D (dorsal), and V (ventral) axes shown in upper left corner. c) Spatial‐temporal expression of chaf1a and mycn in 16 hpf and 24 hpf embryos by HCR. d) Top: tSNE plots with relative expression levels of chaf1a and mycn in NCC and NCC derivatives at 48–50 hpf (red = chaf1a, blue = mycn, magenta = both). Bottom: HCR against chaf1a and mycn in 48 hpf embryos. Arrowheads: populations co‐expressing chaf1a and mycn. e) Top: tSNE plots with relative expression of chaf1a and elavl3 in NCC and NCC‐derivatives at 48–50 hpf and at 68–70 hpf (red = chaf1a, blue = elavl3, magenta = both). Bottom: HCR against chaf1a and elavl3 in 48 hpf and 70 hpf embryos. White arrowheads: developing cranial ganglia (elavl3+); red arrowheads: surrounding non‐neuronal tissue (chaf1a+). f) Top: tSNE plots with relative expression of mycn and elavl3 in NCC and NCC‐derivatives at 68–70 hpf (red = mycn, blue = elavl3, magenta = both). Bottom: HCR against mycn and elavl3 in 70 hpf embryos. White arrowheads: developing cranial ganglia (elavl3+). g) Diagram for ectopic expression of human CHAF1A in zebrafish NCCs. h) Percentage of GFP+/mCherry+ or GFP+/CHAF1A+ clones that also express Elavl3. i) Representative image from a sox10: mCherry‐IRES‐EGFP and a sox10:CHAF1A‐IRES‐EGFP injected embryo. Markers: EGFP (green), gene of interest (GOI) either mCherry or CHAF1A (red), and Elavl3 (cyan). White arrowheads: GFP+/GOI+, GFP+/mCherry+, or GFP+/CHAF1A+ clones; tan arrowheads: GFP+/GOI+/Elavl3+ clones. j) CHAF1A expression in neurons versus NCCs in a hESCs‐derived NCC induction and differentiation model. Left: schematic representation of the hNC model. Middle: TFAP2A and TUJ1 immunofluorescence staining in NCCs and neurons, respectively. Right: qPCR analysis of NCC markers (SOX9 and TFAP2A) and neuron markers (TUBB3, which encodes TUJ1, and MAPT) in NCCs and neurons. Data are presented as the mean ± SD (n = 2); two‐sided unpaired t‐test, *p < 0.05, **p < 0.01. k) Left: schematic representation of NCC induction and RA‐induced differentiation into mature neurons with or without CHAF1A overexpression. Middle: immunofluorescence staining of TUJ1 with or without CHAF1A overexpression. Right: percentages of TUJ1 positive cells are quantified with Image J2. Mean ± SD (n = 6); two‐sided unpaired t‐test, ****p < 0.0001. Scale bars = 100 µm in (a–f), (j), (k), and uncropped images in (i); Scale bars = 25 µm for cropped images in (i). y = yolk sac, e = developing eye, b = developing brain, sc = developing spinal cord.

CHAF1A blocks NC differentiation. a) Schematic presentation of early NCC events during zebrafish development. hpf = hours post fertilization. b) Spatial‐temporal expression of sox9b, crestin, and chaf1a in 16 hpf and 24 hpf embryos by hybridization chain reaction (HCR). A (anterior), P (posterior), D (dorsal), and V (ventral) axes shown in upper left corner. c) Spatial‐temporal expression of chaf1a and mycn in 16 hpf and 24 hpf embryos by HCR. d) Top: tSNE plots with relative expression levels of chaf1a and mycn in NCC and NCC derivatives at 48–50 hpf (red = chaf1a, blue = mycn, magenta = both). Bottom: HCR against chaf1a and mycn in 48 hpf embryos. Arrowheads: populations co‐expressing chaf1a and mycn. e) Top: tSNE plots with relative expression of chaf1a and elavl3 in NCC and NCC‐derivatives at 48–50 hpf and at 68–70 hpf (red = chaf1a, blue = elavl3, magenta = both). Bottom: HCR against chaf1a and elavl3 in 48 hpf and 70 hpf embryos. White arrowheads: developing cranial ganglia (elavl3+); red arrowheads: surrounding non‐neuronal tissue (chaf1a+). f) Top: tSNE plots with relative expression of mycn and elavl3 in NCC and NCC‐derivatives at 68–70 hpf (red = mycn, blue = elavl3, magenta = both). Bottom: HCR against mycn and elavl3 in 70 hpf embryos. White arrowheads: developing cranial ganglia (elavl3+). g) Diagram for ectopic expression of human CHAF1A in zebrafish NCCs. h) Percentage of GFP+/mCherry+ or GFP+/CHAF1A+ clones that also express Elavl3. i) Representative image from a sox10: mCherry‐IRES‐EGFP and a sox10:CHAF1A‐IRES‐EGFP injected embryo. Markers: EGFP (green), gene of interest (GOI) either mCherry or CHAF1A (red), and Elavl3 (cyan). White arrowheads: GFP+/GOI+, GFP+/mCherry+, or GFP+/CHAF1A+ clones; tan arrowheads: GFP+/GOI+/Elavl3+ clones. j) CHAF1A expression in neurons versus NCCs in a hESCs‐derived NCC induction and differentiation model. Left: schematic representation of the hNC model. Middle: TFAP2A and TUJ1 immunofluorescence staining in NCCs and neurons, respectively. Right: qPCR analysis of NCC markers (SOX9 and TFAP2A) and neuron markers (TUBB3, which encodes TUJ1, and MAPT) in NCCs and neurons. Data are presented as the mean ± SD (n = 2); two‐sided unpaired t‐test, *p < 0.05, **p < 0.01. k) Left: schematic representation of NCC induction and RA‐induced differentiation into mature neurons with or without CHAF1A overexpression. Middle: immunofluorescence staining of TUJ1 with or without CHAF1A overexpression. Right: percentages of TUJ1 positive cells are quantified with Image J2. Mean ± SD (n = 6); two‐sided unpaired t‐test, ****p < 0.0001. Scale bars = 100 µm in (a–f), (j), (k), and uncropped images in (i); Scale bars = 25 µm for cropped images in (i). y = yolk sac, e = developing eye, b = developing brain, sc = developing spinal cord.

CHAF1A gene expression and pathway analyses of NB cells and patients. a) Left: overlap of differentially expressed genes (DEGs, |(fc)| > = 1.25, FDR < 0.1) between control (CHAF1A OFF) and CHAF1A‐overexpressing SHEP cells (CHAF1A ON, 96 h) and CHAF1A‐correlated genes (FDR < 0.1) in patient cohort 1 (n = 249) and 2 (n = 648). Right: GO pathway enrichment analysis of the overlapped genes (ranked by −Log₁₀FDR, FDR<0.05). b) Work flow of the metabolomics analysis: global metabolomics analysis was performed by GC‐MS and LC‐MS (DiscoveryHD4 platform, Metabolon Inc.) in CHAF1A‐overexpressing SHEP cells (DOX 1 µg mL⁻¹ for 0, 24, and 72 h, n = 5). c) Metabolite enrichment analysis depicts the pathways significantly up‐ and down‐regulated by CHAF1A (DOX 24 h, FDR < 0.25); Benjamini–Hochberg corrected two‐sided homoscedastic t‐test. d) Left: schematic presentation (redrawn from Gamble et al.^[52]) of the polyamine pathway with metabolite changes in SHEP cells with or without CHAF1A overexpression for 24 h (red = upregulated metabolites, p ≤ 0.05; blue = downregulated metabolites, p ≤ 0.05). Right: polyamine levels in SHEP cells with or without CHAF1A overexpression for 24 h. Data are mean ± SD (n = 5). e) Targeted polyamine analysis in IMR32 cells with conditional KD of CHAF1A (DOX 1 µg mL⁻¹ for 5 days). Differential metabolites (FDR < 0.25) are presented in the heatmap (yellow = upregulated; blue = downregulated) (n = 4). f) Polyamine synthetic and catabolic gene expression in SHEP cells with or without CHAF1A overexpression (24 h). Data are mean ± SD (n = 2); *p < 0.05, **p < 0.01, ***p < 0.001; two‐sided unpaired t‐test. g) Polyamine gene expression in patients with high and low CHAF1A expression (average CHAF1A mRNA expression ± 1SD, Figure 1) in patient cohorts 1 and 2. Data are mean ± SEM (n = 44 in cohort 1 and n = 107 in cohort 2); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; two‐sided unpaired t‐test. h) ODC1 activity in SHEP, GIMEN, and NGP cells with or without CHAF1A overexpression (8 h). One unit is defined as the fluorescence change per minute. Data are normalized by the protein amount and presented as the fold change compared to control (mean ± SD, n = 2); *p < 0.05, **p < 0.01; two‐sided unpaired t‐test. MTA = 5'‐methylthioadenosine; AdoMet = S‐(5'‐Adenosyl)‐L‐methionine; AdoHyc = S‐(5′‐Adenosyl)‐L‐homocysteine; FC = fold change.

Inhibition of polyamine synthesis restores neuronal differentiation. a,b) Neurite length and TUJ1 immunofluorescence staining in NGP‐CHAF1A cells treated with RA (5 μм), DOX (1 µg mL⁻¹), and DFMO (0.5 mм) for 72 h. Data are mean ± SEM (n > 300); ****p < 0.0001; one‐way ANOVA with Tukey's multiple comparisons test. Scale bar = 50 µm. c) Cell cycle analysis of NGP‐CHAF1A cells treated with DOX (1 µg mL⁻¹) or DFMO (0.5 mм) for 72 h. Data are mean ± SD (n = 2); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; two‐way ANOVA with Tukey's multiple comparisons test. d) Neuronal outgrowth of IMR32 CHAF1A KD cells in the presence or absence of ODC1 overexpression. Neurite length is quantified using Image J2 and presented as mean ± SEM (n > 300); ****p < 0.0001; one‐way ANOVA with Tukey's multiple comparisons test. Scale bar = 50 µm. e) Cell cycle analysis of IMR32 CHAF1A KD cells in the presence or absence of ODC1 overexpression. Data are mean ± SD (n = 2); ****p < 0.0001; two‐way ANOVA with Tukey's multiple comparisons test. f) Cell viability of LAN5, IMR32, CHLA255, and SK‐N‐AS cells treated with increasing concentrations of DFMO single agent, RA single agent, and their combination (combo). Cell viability of LAN5 shCTRL and shCHAF1A cells treated with increasing doses of RA. Data are mean ± SD (n = 3). †synergy with CI < 1; **p < 0.01; ***p < 0.001; two‐sided unpaired t‐test. g) Apoptosis of LAN5, IMR32, CHLA255, and SK‐N‐AS cells treated with DFMO, RA, and combo (IC₅₀–₇₅). Apoptosis of LAN5 shCTRL and shCHAF1A cells treated with RA (IC₅₀). Data are mean ± SD (n = 3); * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; two‐way ANOVA with Tukey's multiple comparisons test. h) Polyamine metabolites in LAN5 cells treated with DFMO (1 mм), RA (10 μм), and combo for 5 days (n = 4). Metabolites with FDR < 0.05 in at least one comparison are shown in the heatmap (red = upregulated; blue = downregulated); two‐way ANOVA with original FDR method of Benjamini and Hochberg. The relative abundance of putrescine and spermidine are presented in box and whiskers plots. # indicates FDR < 0.05. i) Top, scheme of RA+DFMO study in LAN5 luc orthotopic xenograft model. Mice were treated with vehicle (1% methylcellulose, p.o., b.i.d., 5 days per week), RA (p.o., 40 mg kg⁻¹ b.i.d., 5 days per week), DFMO (2% in sterile water, replaced weekly), and their combination for three weeks. Bottom left, tumor weights post treatment. Mean ± SEM (n = 8–11); Mann–Whitney test. Bottom right, cleaved caspase‐3 staining and quantification in tumors. Scale bar = 20 µm. Mean ± SEM (n = 6); Mann–Whitney test. j) Top, scheme of RA study in LAN5 luc shCHAF1A versus shCTRL orthotopic xenograft model. Mice were treated with vehicle (1% methylcellulose, p.o., b.i.d., 5 days per week) or RA (p.o., 40 mg kg⁻¹ b.i.d., 5 days per week) for 3 weeks. Bottom left, tumor weights post treatment. Mean ± SEM (n = 9–10); Mann–Whitney test. Bottom right, cleaved caspase‐3 staining and quantification in tumors. Scale bar = 20 µm. Mean ± SEM (n = 5–6); Mann–Whitney test. FC = fold change; ns = not significant.

Inhibition of polyamine synthesis restores neuronal differentiation. a,b) Neurite length and TUJ1 immunofluorescence staining in NGP‐CHAF1A cells treated with RA (5 μм), DOX (1 µg mL⁻¹), and DFMO (0.5 mм) for 72 h. Data are mean ± SEM (n > 300); ****p < 0.0001; one‐way ANOVA with Tukey's multiple comparisons test. Scale bar = 50 µm. c) Cell cycle analysis of NGP‐CHAF1A cells treated with DOX (1 µg mL⁻¹) or DFMO (0.5 mм) for 72 h. Data are mean ± SD (n = 2); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; two‐way ANOVA with Tukey's multiple comparisons test. d) Neuronal outgrowth of IMR32 CHAF1A KD cells in the presence or absence of ODC1 overexpression. Neurite length is quantified using Image J2 and presented as mean ± SEM (n > 300); ****p < 0.0001; one‐way ANOVA with Tukey's multiple comparisons test. Scale bar = 50 µm. e) Cell cycle analysis of IMR32 CHAF1A KD cells in the presence or absence of ODC1 overexpression. Data are mean ± SD (n = 2); ****p < 0.0001; two‐way ANOVA with Tukey's multiple comparisons test. f) Cell viability of LAN5, IMR32, CHLA255, and SK‐N‐AS cells treated with increasing concentrations of DFMO single agent, RA single agent, and their combination (combo). Cell viability of LAN5 shCTRL and shCHAF1A cells treated with increasing doses of RA. Data are mean ± SD (n = 3). †synergy with CI < 1; **p < 0.01; ***p < 0.001; two‐sided unpaired t‐test. g) Apoptosis of LAN5, IMR32, CHLA255, and SK‐N‐AS cells treated with DFMO, RA, and combo (IC₅₀–₇₅). Apoptosis of LAN5 shCTRL and shCHAF1A cells treated with RA (IC₅₀). Data are mean ± SD (n = 3); * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; two‐way ANOVA with Tukey's multiple comparisons test. h) Polyamine metabolites in LAN5 cells treated with DFMO (1 mм), RA (10 μм), and combo for 5 days (n = 4). Metabolites with FDR < 0.05 in at least one comparison are shown in the heatmap (red = upregulated; blue = downregulated); two‐way ANOVA with original FDR method of Benjamini and Hochberg. The relative abundance of putrescine and spermidine are presented in box and whiskers plots. # indicates FDR < 0.05. i) Top, scheme of RA+DFMO study in LAN5 luc orthotopic xenograft model. Mice were treated with vehicle (1% methylcellulose, p.o., b.i.d., 5 days per week), RA (p.o., 40 mg kg⁻¹ b.i.d., 5 days per week), DFMO (2% in sterile water, replaced weekly), and their combination for three weeks. Bottom left, tumor weights post treatment. Mean ± SEM (n = 8–11); Mann–Whitney test. Bottom right, cleaved caspase‐3 staining and quantification in tumors. Scale bar = 20 µm. Mean ± SEM (n = 6); Mann–Whitney test. j) Top, scheme of RA study in LAN5 luc shCHAF1A versus shCTRL orthotopic xenograft model. Mice were treated with vehicle (1% methylcellulose, p.o., b.i.d., 5 days per week) or RA (p.o., 40 mg kg⁻¹ b.i.d., 5 days per week) for 3 weeks. Bottom left, tumor weights post treatment. Mean ± SEM (n = 9–10); Mann–Whitney test. Bottom right, cleaved caspase‐3 staining and quantification in tumors. Scale bar = 20 µm. Mean ± SEM (n = 5–6); Mann–Whitney test. FC = fold change; ns = not significant.

Inhibition of polyamine synthesis restores neuronal differentiation. a,b) Neurite length and TUJ1 immunofluorescence staining in NGP‐CHAF1A cells treated with RA (5 μм), DOX (1 µg mL⁻¹), and DFMO (0.5 mм) for 72 h. Data are mean ± SEM (n > 300); ****p < 0.0001; one‐way ANOVA with Tukey's multiple comparisons test. Scale bar = 50 µm. c) Cell cycle analysis of NGP‐CHAF1A cells treated with DOX (1 µg mL⁻¹) or DFMO (0.5 mм) for 72 h. Data are mean ± SD (n = 2); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; two‐way ANOVA with Tukey's multiple comparisons test. d) Neuronal outgrowth of IMR32 CHAF1A KD cells in the presence or absence of ODC1 overexpression. Neurite length is quantified using Image J2 and presented as mean ± SEM (n > 300); ****p < 0.0001; one‐way ANOVA with Tukey's multiple comparisons test. Scale bar = 50 µm. e) Cell cycle analysis of IMR32 CHAF1A KD cells in the presence or absence of ODC1 overexpression. Data are mean ± SD (n = 2); ****p < 0.0001; two‐way ANOVA with Tukey's multiple comparisons test. f) Cell viability of LAN5, IMR32, CHLA255, and SK‐N‐AS cells treated with increasing concentrations of DFMO single agent, RA single agent, and their combination (combo). Cell viability of LAN5 shCTRL and shCHAF1A cells treated with increasing doses of RA. Data are mean ± SD (n = 3). †synergy with CI < 1; **p < 0.01; ***p < 0.001; two‐sided unpaired t‐test. g) Apoptosis of LAN5, IMR32, CHLA255, and SK‐N‐AS cells treated with DFMO, RA, and combo (IC₅₀–₇₅). Apoptosis of LAN5 shCTRL and shCHAF1A cells treated with RA (IC₅₀). Data are mean ± SD (n = 3); * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; two‐way ANOVA with Tukey's multiple comparisons test. h) Polyamine metabolites in LAN5 cells treated with DFMO (1 mм), RA (10 μм), and combo for 5 days (n = 4). Metabolites with FDR < 0.05 in at least one comparison are shown in the heatmap (red = upregulated; blue = downregulated); two‐way ANOVA with original FDR method of Benjamini and Hochberg. The relative abundance of putrescine and spermidine are presented in box and whiskers plots. # indicates FDR < 0.05. i) Top, scheme of RA+DFMO study in LAN5 luc orthotopic xenograft model. Mice were treated with vehicle (1% methylcellulose, p.o., b.i.d., 5 days per week), RA (p.o., 40 mg kg⁻¹ b.i.d., 5 days per week), DFMO (2% in sterile water, replaced weekly), and their combination for three weeks. Bottom left, tumor weights post treatment. Mean ± SEM (n = 8–11); Mann–Whitney test. Bottom right, cleaved caspase‐3 staining and quantification in tumors. Scale bar = 20 µm. Mean ± SEM (n = 6); Mann–Whitney test. j) Top, scheme of RA study in LAN5 luc shCHAF1A versus shCTRL orthotopic xenograft model. Mice were treated with vehicle (1% methylcellulose, p.o., b.i.d., 5 days per week) or RA (p.o., 40 mg kg⁻¹ b.i.d., 5 days per week) for 3 weeks. Bottom left, tumor weights post treatment. Mean ± SEM (n = 9–10); Mann–Whitney test. Bottom right, cleaved caspase‐3 staining and quantification in tumors. Scale bar = 20 µm. Mean ± SEM (n = 5–6); Mann–Whitney test. FC = fold change; ns = not significant.

CHAF1A is a direct target of MYCN. a,b) mRNA and protein expression of CHAF1A in TET‐21/N cells when MYCN is turned off upon DOX treatment (2 µg mL⁻¹, 24–96 h). GAPDH is used as housekeeping gene, CypB as protein loading control. Data are mean ± SD (n = 2–3); * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; one‐way ANOVA with Dunnett's multiple comparisons test. c) MYCN ChIP‐qPCR assays in TET‐21/N cells. Input (white bars) and MYCN‐ChIP (black bars) samples were analyzed by qPCR using specific primers for CHAF1A (Table S5, Supporting Information). Data from two independent experiments are shown (mean ± SEM, n = 2). d) mRNA expression of CHAF1A, MYCN, MYCN targets and polyamine genes in LAN5 cells upon CHAF1A KD (DOX 1 µg mL⁻¹ for 2–5 days). GAPDH served as control. Mean ± SD (n = 3); * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; two‐way ANOVA with Dunnett's multiple comparisons test. e) Protein expression of CHAF1A, MYCN, and ODC1 in LAN5 cells upon CHAF1A KD (DOX 1 µg mL⁻¹ for 0–10 days). CypB served as protein loading control. Mean ± SD (n = 2); * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; one‐way ANOVA with Dunnett's multiple comparisons test. f) Correlation of CHAF1A, MYCN, and MYCN signature scores in patient cohort 1 (n = 249) and cohort 2 (n = 648). Signatures are defined in the methods section. FC = fold change.

CHAF1A is a direct target of MYCN. a,b) mRNA and protein expression of CHAF1A in TET‐21/N cells when MYCN is turned off upon DOX treatment (2 µg mL⁻¹, 24–96 h). GAPDH is used as housekeeping gene, CypB as protein loading control. Data are mean ± SD (n = 2–3); * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; one‐way ANOVA with Dunnett's multiple comparisons test. c) MYCN ChIP‐qPCR assays in TET‐21/N cells. Input (white bars) and MYCN‐ChIP (black bars) samples were analyzed by qPCR using specific primers for CHAF1A (Table S5, Supporting Information). Data from two independent experiments are shown (mean ± SEM, n = 2). d) mRNA expression of CHAF1A, MYCN, MYCN targets and polyamine genes in LAN5 cells upon CHAF1A KD (DOX 1 µg mL⁻¹ for 2–5 days). GAPDH served as control. Mean ± SD (n = 3); * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; two‐way ANOVA with Dunnett's multiple comparisons test. e) Protein expression of CHAF1A, MYCN, and ODC1 in LAN5 cells upon CHAF1A KD (DOX 1 µg mL⁻¹ for 0–10 days). CypB served as protein loading control. Mean ± SD (n = 2); * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001; one‐way ANOVA with Dunnett's multiple comparisons test. f) Correlation of CHAF1A, MYCN, and MYCN signature scores in patient cohort 1 (n = 249) and cohort 2 (n = 648). Signatures are defined in the methods section. FC = fold change.