MYCN amplification and ATRX mutations are incompatible in neuroblastoma

Aggressive cancers often have activating mutations in growth-controlling oncogenes and inactivating mutations in tumor-suppressor genes. In neuroblastoma, amplification of the MYCN oncogene and inactivation of the ATRX tumor-suppressor gene correlate with high-risk disease and poor prognosis. Here we show that ATRX mutations and MYCN amplification are mutually exclusive across all ages and stages in neuroblastoma. Using human cell lines and mouse models, we found that elevated MYCN expression and ATRX mutations are incompatible. Elevated MYCN levels promote metabolic reprogramming, mitochondrial dysfunction, reactive-oxygen species generation, and DNA-replicative stress. The combination of replicative stress caused by defects in the ATRX–histone chaperone complex, and that induced by MYCN-mediated metabolic reprogramming, leads to synthetic lethality. Therefore, ATRX and MYCN represent an unusual example, where inactivation of a tumor-suppressor gene and activation of an oncogene are incompatible. This synthetic lethality may eventually be exploited to improve outcomes for patients with high-risk neuroblastoma.


Supplementary Figure 1. Analysis of MYCN amplified ATRX mutant sample. A)
Scatterplot of the coverage relative to the median exon coverage for the 3 indicated samples. SJNBL013668 has the normal diploid copy number of the MYCN locus, SJNBL013616 has high level amplification and SJNBL013621 is the sample reported to have MYCN amplification and ATRX mutation. The genomic region is aligned with the ChromHMM viewer that is part of this study showing the location of the exons being plotted. B,C) Micrograph of Hematoxylin and Eosin staining of the SJNBL013621 sample showing regions of low and high mitotic karyorrhectic index. D) Micrograph of fluorescent in situ hybridization (FISH) for telomeres (red) with nuclear stain (blue) showing regions of ultrabright signal indicative of alternative lengthening of telomeres (ALT). E) Micrograph of FISH for MYCN (red) and a control probe (PAX3, green). Arrows indicate strong FISH signals for telomeres and MYCN. Scale bars B&C 50 µm, D&E 5 µm.

Supplementary Figure 2. Histopathologic review of genetically engineered mouse models.
Micrographs of representative sections of adrenal and paravertebral ganglion from Atrx Lox/Lox ;LSL-MYCN;DbhiCre and control littermates (LSL-MYCN;DbhiCre). Hematoxylin and Eosin stained sections as well as Ki67, synaptophysin and tyrosine hydroxylase immunohistochemical staining are shown for each genotype at 3 weeks and 1 year of age. 3 animals for each genotype at each stage were reviewed by the veterinary pathologist in a blinded manner and no differences were identified. Scale bar: 100µm.

Supplementary Figure 3. Inactivation of ATRX and DAXX. A,B) Micrograph of telomere FISH for HeLa and U2OS cells showing individual telomeres (arrows). The experiment was repeated with the same results. C)
Step plot of the number of telomeres per cell for HeLa cells (blue) and U2OS cells (green). P value was calculated using Chi-square test D) Distribution of fluorescence intensity of individual telomeres in HeLa and U2OS cells. The ultrabright telomere signal in the U2OS (arrows in panel B) and higher intensity per telomere (green in D) are hallmarks of alternative lengthening of telomeres (ALT). P value was calculated using Chisquare test. E) Box and Whiskers plot of telomere fluorescence per cell for the indicated number of telomeres for each cell line. Each Box shows the 10 th to 90 th percentiles range of data, the line represents the median and the Whiskers show the minimum and maximum data range F) Bar plot with all data points presented of the amount of telomere DNA as measured by QPCR for each of the cell lines. The analysis was done twice while the graph shows the data from one experiment in a technical duplicate G) Box and Whiskers plot of telomere fluorescence per cell for each cell line transfected with the indicated ATRX-shRNA (130 cells were analyzed per condition). Each Box shows the 10 th to 90 th percentiles range of data, the line represents the median and the Whiskers show the minimum and maximum data range H) Bar plot with all data points presented of the amount of telomere DNA as measured by QPCR for each cell line transfected with the indicated ATRX-shRNA in technical duplicates. I) Piecharts of the cell cycle distribution for each of the shRNAs for each cell line. J) Proportion of MiSeq reads from the PCR product spanning the target sequence for g-RNA 2 and g-RNA 5 in 293T, SKNBE2 and IMR32 cells after 2, 10 and 14 days in culture. A control for IMR32 (AAVS1) is shown to demonstrate that efficient CRISPR/Cas9 targeting can be achieved in those cells. K) Proportion of MiSeq reads (with at least 200,000 reads) from the PCR product spanning the target sequence for g-RNA11 targeting exon 5 of DAXX. L) Map of the location of g-RNA 11 targeting exon 5 of DAXX. M) Plot of the frequency of mutation for g-RNA 2 with deletions in green and insertions in red from MiSeq analysis of the PCR product spanning the target sequence of DAXX. Scale bar: 5 µm.

Supplementary Figure 4. Analysis of additional ATRX and DAXX deficient cell lines. A)
Immunoblot of 6 cell lines that have stable integration of the doxycycline inducible MYCN expression plasmid. Lysates were prepared 4 days after addition of doxycycline and blotted for ATRX, DAXX, MYCN and β-actin as a loading control. SKNMM cells have an in frame deletion of ATRX (IFD), WI-38 VA13 2RA , GM847 and SAOS2 cells lack ATRX while G292 cells lack DAXX. The SW13 cells were included as a wild type control for ATRX and DAXX. The experiment was repeated with similar results B) Photograph of a cresyl violet stained 6 well dish from colony assays with each of the indicated cell lines in the presence and absence of doxycycline. The control lines have the tet regulated plasmid integrated into the genome but lacking the MYCN transgene. The experiment was repeated with similar results C,D) Immunoblot of 2 glioma cell lines with stable integration of the doxycycline inducible MYCN expression plasmid. Lysates were prepared 4 days after addition of doxycycline and blotted for ATRX, MYCN and GAPDH. E) Photograph of cresyl violet stained 6 well dish from colony assay with each line. The experiment was repeated with similar results. F,G) Line graph of growth curve for the indicated cell line with the MYCN inducible vector in the presence and absence of doxycycline. Each point is the mean and standard deviation of triplicate samples. H) Immunoblot for pRPA32 (red), MYCN (green) and an internal loading control (actin in red).

Supplementary Figure 5. DNA methylation analysis. A)
Venn diagram of the hypermethylated and hypomethylated promoters (blue) in the MYCN amplified tumors relative to the non-amplified tumors. The upregulated genes are shown in purple and the downregulated genes are shown in brown. The number of overlapping genes are indicated on the diagram. B,C) Representative plots of DNA methylation from whole genome bisulfite sequencing for a representative hypermethylated promoter (DUSP23) in MYCN amplified tumors that correlated with reduced gene expression (FPKM is shown to the right) and a hypomethylated promoter (CRYGD) that correlated with elevated gene expression. Abbreviations: FPKM, fragments per kilobase of transcript per million reads.

Supplementary Figure 6. Induction of MYCN targets in SKNMM MYCN cells. A)
The top pathways in gene set enrichment analysis for SKNMM MYCN and U2OS MYCN cells in the presence of doxycycline relative to the absence of doxycycline. B) ChIP-seq for MYCN and RNA Pol II and chrom-HMM for the PNPT1 gene promoter in SKNMM MYCN cells in the presence and absence of doxycycline. The PNPT1 protein is important for importing nuclear encoded RNA into the mitochondria. C) All ChIP-seq tracks for the PNPT1 gene in the presence and absence of doxycycline. D) ChIP-seq for MYCN and RNA Pol II and chrom-HMM for the SLC7A5 gene promoter in SKNMM MYCN cells in the presence and absence of doxycycline. The SLC7A5 protein is important for glutamine transport. E) All ChIP-seq tracks for the SLC7A5 gene in the presence and absence of doxycycline.

Supplementary Figure 7. Glutamine metabolism following MYCN induction. A)
Bar plot with individual data points presented of the scoring of the number of cells per well for each cell line cultured in the presence or absence of glutamine with and without doxycycline. Each bar is the mean and standard deviation of 3-6 wells. SKNMM and U2OS cells were significantly sensitive to glutamine deprivation, Student T-test) B) Bar plot with individual data points presented of glutathione levels in the cells on day 6 in culture for each cell line in the presence and absence of doxycycline. Each bar is the mean and standard deviation of 8 wells C) Immunoblot of γ-H2AX and GAPDH in the presence and absence of doxycycline and glutamine (GLN). The fold induction is indicated over the γ-H2AX band in the presence of doxycycline relative to the absence of doxycycline in the absence of glutamine. D) Flow cytometry of cells stained with CellRox Green to measure reactive oxygen species (ROS) on day 4 and day 5 in the presence and absence of doxycycline and glutamine.

Supplementary Figure 8. Mitochondrial stress and glycolysis stress for SKNMM MYCN cells. A) Bar plot with individual data points presented of 3 independent experiments showing the mean and standard deviation for each experiment and each sample at basal and treated
conditions. The oxygen consumption rate is plotted in this panel. B) The extracellular acidification rate (ECAR) is plotted for the same samples as shown in (A). For the data in A and B, the experiment was performed using conditions for mitochondrial stress testing on the Seahorse instrument. C) Representative experiment out of three independent ones showing the ECAR for SKNMM MYCN cells in the presence and absence of doxycycline grown under glycolysis stress conditions. Each point represents the average and the standard error of the mean for the technical replicates in the same experiment. D) Bar plot with individual data points presented of 3 independent experiments for the glycolysis stress test conditions for the indicated cell lines.

Supplementary Figure 9. Electron microscopy of mitochondria in SKNMM MYCN cells. A,B)
Micrograph of SKNMM MYCN cell mitochondria after 4 days in culture in the absence of doxycycline. The mitochondrial matrix and cristae are labeled as well as the inner and outer mitochondrial membranes (im and om). C,D) Micrograph of SKNMM MYCN cell mitochondria after 4 days in culture in the presence of doxycycline. The cristae are disorganized and swollen and the mitochondrial membranes are disrupted (arrows). E,F) Bar plot of scoring of mitochondrial morphology for the U2OS MYCN and SKNBE2 MYCN cells in the presence or absence of doxycycline after 4 days in culture.

Supplementary Figure 10. Oxidative stress contributes to synthetic lethality. A)
Bar plot with individual data points presented of the relative viability of SKNMM MYCN cells in the presence and absence of doxycycline on day 4 in culture in the presence of different concentrations of hydroxyurea. Each bar is the mean and standard deviation of 8 wells. P-value was calculated using two-tailed Student T-test B) Plot of DNA content per cell and relative proportion of cells in G0/G1, S and G2/M phases of the cell cycle on day 4 in culture for SKNMM MYCN and SKNFI MYCN cells in the presence and absence of doxycycline. The piechart shows the proportion of each phase for the untreated (-DOX, gray lines) and in the presence of doxycycline (red). C) Bar plot with individual data points presented of SKNMM MYCN cells on day 7 after addition of doxycycline in the presence and absence of N-Acetyl-L-cysteine. Each bar is the mean and standard deviation of 4 different biologic replicates. P value was calculated using two-tailed Student T-test D) The same samples analyzed in (C) were also analyzed by flow cytometry using CellRox Green to measure reactive oxygen species (ROS). The panel on the left shows +/-doxycycline and the panel on the right shows +/-N-Acetyl-L-cysteine in the presence of doxycycline. E-G) Representative dose response curves for SKNMM cells on day 3 of culture in the presence and absence of doxycycline for a drug that had no effect on viability (F) and two drugs that potentiated killing my MYCN. The dashed line indicates the EC50 and the area under the curve is shown for each graph. H) Spline plot (see Supplemental Information) for one of two drugs that partially rescued the cell death induced by MYCN in SKNMM cells. Figure 11. Neuroblastoma differentiation and CUX2. A) Micrograph of 11 NB cell lines cultured in the presence and absence of retinoic acid for 6 days to induce differentiation. The same cultures were used for the RNA-seq analysis to determine if CUX2 was induced with differentiation. B) Bar plot with individual data points of the mean and standard deviation of cell survival measured with CellTiterGlo for SKNMM MYCN cells in the presence and absence of doxycycline (DOX). There was a significant increase in viability when NAC was added to the culture or CUX2 was ectopically expressed but no added benefit of the combination. P values were calculated using two-tailed Student T-test, n=8 biologic replicates. Scale bars 10 µm.

SUPPLEMENTARY NOTE 1 Testing mutual exclusivity between ATRX-mutations and MYCN amplification
To test the mutual exclusivity between ATRX-mutations and MYCN amplification in neuroblastoma, we merged the COG, TARGET, and PCGP data.
Below are results of the analysis looking at the relationship between ATRX and MYCN in the subgroup of patients with Stage 4, >5 years of age: If we look just at Stage 4, 5-12 years of age: We also looked at stratified analyses using the Cochrane-Mantel-Haenszel test.