Introduction

Neuroblastoma is the most common extra cranial solid tumor in early childhood that arises from neuronal crest cells, which accounts for 15% of all pediatric cancer fatalities [1, 2]. Chromosome 17q21-ter gain is one of the most commonly gene copy number alteration in neuroblastoma, which shows statistically significant correlation with MYCN amplification and portends a grim prognosis [3, 4]. The insulin growth factor 2 binding protein 1, JMJD6, survivin and ALYREF genes located in this chromosomal region have been reported to facilitate the survival and proliferation of neuroblastoma cells [5,6,7,8].

SMARCE1 gene is located on chromosome 17q21.2, and encodes a core subunit of ATP-dependent chromatin remodeling complex SWI/SNF that functions in many biological processes [9, 10]. SMARCE1 contains a DNA-binding HMG domain and modulates gene expression in tumors [11, 12]. Recent researches reported that SMARCE1 functions as a transcriptional co-activator. In breast tumors, SMARCE1 formed a complex with HIF1A to cooperatively activate the PTK2 expression. Furthermore, SMARCE1 assists RUNX1 access the silencer and repress CD4 in early thymocytes [13, 14]. However, biological function of SMARCE1 in neuroblastoma and its mechanism have not been studied.

Amplification of the MYCN oncogene compose 25% of neuroblastoma patients, and correlate with adverse clinical outcomes and poor overall patient survival rates [15, 16]. MYCN is a member of the MYC oncogene family and primarily expressed in neural tissues [17, 18]. Like C-MYC, MYCN functions as a major transcriptional regulator and plays multiple regulatory roles in neuroblastoma including differentiation, proliferation, survival, self-renewal, metabolism, metastasis and angiogenesis [18]. MYCN regulates gene transcription by heterodimerizing with MAX to bind E-box DNA motifs [19, 20] and recruits co-factors to alter the structure of the chromatin [21].

In this study, we investigated the potential role of SMARCE1 in neuroblastoma. We found that SMARCE1 was highly expressed in MYCN-amplified tumors and cells. Suppression of SMARCE1 reduced the proliferation and clonogenic capacity and impaired tumorigenic potential of MYCN-amplified cells. Our data demonstrated that MYCN regulated SMARCE1 expression by binding to its promoter. Further investigation revealed that SMARCE1 and MYCN formed co-activators to regulate the transcription of three MYCN target genes (PLK1, ODC1 and E2F2). These results demonstrated that SMARCE1 is a key player in tumorigenesis of MYCN-amplified neuroblastoma.

Result

High SMARCE1 expression correlates with poor prognosis in neuroblastoma

As SMARCE1 gene is localized at chromosome 17q21-ter region (Fig. 1A), we next examined whether chromosome 17q gain affects SMARCE1 expression in database. Analysis of the TARGET [22] and Jagannathan [23] datasets showed that SMARCE1 expression in tumor with 17q gain was significantly higher than that in 17q diploid tumor (Supplementary Fig. S1A). Interestingly, analysis using datasets of Kocak [24] and Delattre [25] showed that the mRNA level of SMARCE1 was higher in MYCN-amplified tumors than that in MYCN non-amplified tumors (Supplementary Fig. S1B). Consistently, immunoblots showed that the expression of SMARCE1 was higher in neuroblastoma cell lines with MYCN-amplification (IMR32, SK-N-BE(2)C, and SK-N-DZ) than that in neuroblastoma cell lines without the amplification (SHEP1, SH-SY5Y) (Fig. 1B). The gene expression of pan-cancer cell lines (RNAseq) were further examined in Cancer Cell Line Encyclopedia/Cancer Dependency Map [26]. We found that neuroblastoma cell lines have higher level of SMARCE1 expression compared to other tumor types (Supplementary Fig. S1C).

Fig. 1: High SMARCE1 expression associates with 17q21-ter gain, MYCN-amplification, and poor patient prognosis in neuroblastoma.
figure 1

A SMARCE1 locus on chromosome 17q21-ter region. B SMARCE1 expression was analyzed by immunoblot in SK-N-BE(2)C, IMR32, SK-N-DZ, SH-SY5Y and SH-EP1 neuroblastoma cell lines. C Kaplan–Meier curves showed the probability of overall survival of patients according to SMARCE1 mRNA expression levels in the 66 MYCN-amplified neuroblastoma samples. Data are from the Kocak dataset. D Representative images of immunohistochemical staining of SMARCE1 in neuroblastoma patient tissues. E Quantitative analysis of SMARCE1 protein expression in neuroblastoma patient tissues at INSS stages 1and 2 versus stages 2B and 4. All data were shown as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001, ns not significant.

Full size image

To validate the clinically relevance of SMARCE1 in patients with neuroblastoma. Kaplan–Meier analyses of overall survival probability of patients in Kocak, SEQC [27] and Versteeg [28] datasets were performed. The results showed that high SMARCE1 expression was associated with poor prognosis (Supplementary Fig. S1D). This is consistent with the literature reported previously [29, 30]. Furthermore, high SMARCE1 expression in the MYCN-amplified neuroblastoma patients was significantly correlated with worse outcome in Kocak and SEQC datasets (Fig. 1C and Supplementary Fig. S1E). We next examined SMARCE1 expression in 3 normal brain and 27 neuroblastoma samples by immunohistochemical staining. The quantified results showed SMARCE1 overexpression in neuroblastoma (Fig. 1D and Supplementary Fig. S1F). Moreover, high SMARCE1 protein expression was associated with the International Neuroblastoma Staging System (INSS) stage, an established risk factor (Fig. 1E). Similarly, SMARCE1 expression was closely associated with neuroblastoma INSS stage in Kocak datasets (Supplementary Fig. S1G).

SMARCE1 is required for proliferation and tumorigenicity of MYCN-amplified neuroblastoma cell

To determine whether SMARCE1 is required for neuroblastoma cell proliferation and survival, we designed two independent short hairpin RNAs (shRNAs) (shSMARCE1-1# and shSMARCE1-2#) targeting different regions of SMARCE1 mRNA. SK-N-BE(2)C and IMR32 cells were stably transfected with the shRNA vectors for effective knockdown of SMARCE1 expression (Fig. 2A). Consequently, cell proliferation was significantly suppressed after down-regulating SMARCE1 expression (Fig. 2B). Moreover, colony-formation assay showed that colony formation was diminished in the downregulated SMARCE1 group (Supplementary Fig. S2A). To determine whether downregulation of SMARCE1 expression has any effects on the cell cycle, we examined the cell cycle by flow cytometry. Silence of SMARCE1 resulted in G2/M-phase cell-cycle arrest of SK-N-BE(2)C and IMR32 cells (Supplementary Fig. S2B). In addition, the expression of CDK1 and Cyclin B1, G2/M related proteins, was markedly reduced in shSMARCE1 cells (Supplementary Fig. S2C).

Fig. 2: SMARCE1 affects the proliferation and tumorigenicity of neuroblastoma cells.
figure 2

A Real-time qPCR and immunoblot analyses of SMARCE1 in SK-N-BE(2)C and IMR32 cells with SMARCE1 shRNA knockdown. B Cell viability of SMARCE1 shRNA and the GFP control in SK-N-BE(2)C and IMR32 cells were assessed. C, D Xenograft assay of BE(2)-C and IMR32 cells with inducible expression of shGFP or shSMARCE1. Tumor weight was analyzed by scatterplot with horizontal lines indicating the mean. Tumor volume was measured and indicated by curves. E–G IHC staining of SMARCE1 and Ki67 for the tumors isolated from the indicated treatment. E Representative image of the staining; F, G percentage of SMARCE1 and Ki67 positive cells were quantified from three independent samples. All data were shown as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001, ns not significant.

Full size image

Then, we implanted neuroblastoma cells subcutaneously in nonobese diabetic (NOD)/severe combined immunodeficient (SCID) mice. Both the volume and weight of the xenografts in the shSMARCE1 SK-N-BE(2)C and IMR32 cells were markedly smaller than those in the shGFP control cells (Fig. 2C, D). Immunohistochemical staining of the subcutaneous tumors indicated that Ki-67 expression, a proliferation marker, was significantly reduced in tumors expressing shSMARCE1 than that expressing shGFP (Fig. 2E–G). Taken together, these results underlined the strong oncogenic potential of SMARCE1 expression in neuroblastoma, and supported that SMARCE1 is critical to maintain clonogenic capacity, proliferation and survival of neuroblastoma cell.

MYCN regulates SMARCE1 transcription by binding to its promoter

Myc oncoproteins bind the canonical and non-canonical E-box sequences and stimulate transcription of target gene [31]. Bioinformatic analysis identified a non-canonical E-box upstream of the transcription start site of SMARCE1 gene (Supplementary Fig. S3A) and showed positive correlation between mRNA levels of MYCN and SMARCE1 (Supplementary Fig. S3B), which implied MYCN is involved in the regulation of SMARCE1 expression. To probe this possibility, we evaluated the effect of MYCN on SMARCE1 expression. Downregulation of MYCN led to reduced SMARCE1 mRNA and protein level in SK-N-BE(2)C and IMR32 cells (Fig. 3A, B). In addition, a dual-luciferase report assay was designed to examine the impact of knocking down MYCN on the SMARCE1 target promoter activity, and the results showed that the SMARCE1 promoter activity was markedly decreased upon silence of MYCN in SK-N-BE(2)C and IMR32 cells (Fig. 3C, D). Then, we performed chromatin immunoprecipitation (ChIP) assays using MYCN antibody in SK-N-BE(2)C and IMR32 cells. The experiments showed MYCN protein was enriched in the −782 to −408 bp region (Amplicon C) of the SMARCE1 promoter (Fig. 3E, F), consisting with previously published ChIP results [32, 33] that MYCN binds to the SMARCE1 promoter in other neuroblastoma cells (Supplementary Fig. S3C).

Fig. 3: MYCN directly upregulates SMARCE1 expression by binding to the SMARCE1 promoter.
figure 3

A Real-time qPCR analysis of MYCN and SMARCE1 expression upon MYCN shRNA knockdown in SK-N-BE(2)C and IMR32 cells. B Immunoblot detection of SMARCE1 protein levels after MYCN shRNA knockdown. C, D Relative luciferase activity of reporter constructs containing the SMARCE1 promoter in SK-N-BE(2) and IMR32 cells transfected with MYCN shRNA. E Upper panel, a total of five sets of primers were designed within the human SMARCE1 promoter; lower panel, ChIP-qPCR analysis using anti-IgG and anti-MYCN antibodies in SK-N-BE(2) cells. F ChIP-qPCR analysis using anti-IgG and anti-MYCN antibodies in IMR32 cells. G Predicted sequences of E-box in SMARCE1 promoter by JASPAR. H Luciferase assays performed using SMARCE1 wild-type and SMARCE1 mutant promoter. All data were shown as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001, ns not significant.

Full size image

As the predicted E-box is located in the −782 to −408 bp region of the SMARCE1 promoter, we hypothesized this site is crucial for the MYCN binding and function. To test this hypothesis, we mutated the E-box sequence and tested its response to the knockdown of MYCN (Fig. 3G, H). The dual-luciferase report assay showed that the SMARCE1 promoter activity was significantly reduced upon MYCN knockdown, while such response was disappeared when the mutant SMARCE1 promoter sequence was used. These data indicated that MYCN regulated SMARCE1 transcription by binding to the non-canonical E-box of SMARCE1 promoter.

SMARCE1 interacts with MYCN

It has been reported that SWI/SNF complex interacts with c-MYC and is necessary for c-MYC-mediated transactivation [34], which inspired us to investigate the potential interaction between SMARCE1 and MYCN. We first performed Mass spectrometric assay for SMARCE1-interacting proteins. This assay identified MYCN as a bona fide interacting partner with SMARCE1 (Supplementary Fig. S4A). In addition, a number of previously reported SMARCE1-interacting proteins, including the SMARCC2, SMARCA4 and ILF3, were enriched in the interactome [35,36,37]. We then performed co-immunoprecipitation with MYCN or SMARCE1 antibody in SK-N-BE(2)C cells. The immunoblot experiments showed SMARCE1 was co-immunoprecipitated with MYCN, and vice versa (Fig. 4A). Similarly, co-immunoprecipitation between SMARCE1 and MYCN was observed in HEK-293FT cells transfected with Flag-SMARCE1 and HA-MYCN (Fig. 4B). To test whether the interaction between SMARCE1 and MYCN was direct, we performed GST pull-down assays. The result showed that SMARCE1 directly interacted with MYCN in vitro (Fig. 4C). Then we used a proximity ligation assay (PLA) to visualize the protein-protein interaction in vivo. The obvious PLA dot signal in the two neuroblastoma cell lines underpinned the interaction between SMARCE1 and MYCN (Fig. 4D and Supplementary Fig. S4B).

Fig. 4: SMARCE1 forms a protein complex with MYCN.
figure 4

A Endogenous interaction between MYCN with SMARCE1 was confirmed by immunoprecipitation using anti-MYCN antibody and anti-SMARCE1 antibody in SK-N-BE(2)C cells. B HEK-293FT cells were ectopically overexpressed HA-MYCN and Flag-SMARCE1. The interaction was detected by co-immunoprecipitation using anti-HA antibody and anti-Flag antibody. C GST pull-down assay of MYCN using the indicated proteins expressed in bacteria. D Protein-protein interactions between SMARCE1 and MYCN were evaluated by using Duolink in situ PLA technology. Scale bars, 5 μm. E Schematic representation of the SMARCE1 deletion mutants. F Representative immunoblot analysis for ectopically expressed MYCN from HEK-293FT cells after IP of the indicated Flag-SMARCE1 mutants. G Schematic representation of the MYCN deletion mutants. H Representative immunoblot analysis for ectopically expressed SMARCE1 from HEK-293FT cells after IP of the indicated HA-MYCN mutants.

Full size image

To better understand the interaction, we generated a series of truncated mutants of SMARCE1, including N-terminal truncated mutant, C-terminal truncated mutant, HMG domain-deletion mutant and coiled-coil domain-deletion mutant (Fig. 4D). Co-immunoprecipitation revealed C-terminal truncated and HMG domain-deletion mutants were unable to detect the co-precipitation of MYCN, demonstrating that HMG domain was crucial for the binding to MYCN (Fig. 4E). On the other hand, we constructed a series of MYCN truncating mutants and found only full length MYCN (1-464) and MYCN-M (141-345) were able to co-immunoprecipitated with SMARCE1 (Fig. 4F, G), suggesting the small region (amino acids 141-350) of MYCN is responsible for binding to SMARCE1.

SMARCE1 and MYCN cooperate to regulate the expression PLK1, ODC1 and E2F2

We next examined whether SMARCE1 is necessary for MYCN-mediated transcriptional regulation. Gene set enrichment analysis of the GSE16476 [38] cohort showed enriched MYCN target genes, PLK1 pathway and E2F pathway in samples with relatively high SMARCE1 expression (Fig. 5A). We focused on the effect of SMARCE1 on the three known target genes of Myc, PLK1, ODC1 and E2F2 [33, 22,23,24], whose high expression was associated with poor prognosis of neuroblastoma patients in Kocak and Versteeg datasets (Fig. 5B and Supplementary Fig. S5A). Quantitative PCR and immunoblotting assay showed that the mRNA and protein level of PLK1, ODC1 and E2F2 was significantly reduced after knockdown of SMARCE1 in SK-N-BE(2)C and IMR32 cells (Fig. 5C, D). Of course, the mRNA and protein level of MYCN did not change after knockdown of SMARCE1 in SK-N-BE(2)C and IMR32 cells. In addition, the mRNA and protein level of three target genes were restored upon re-expression of wild-type SMARCE1 in SK-N-BE(2)C cells (Supplementary Fig. S5B, C). Consistently, immunohistochemical staining of subcutaneous tumor tissues showed that the intensity of PLK1, ODC1 and E2F2 signal in the shSMARCE1-treated group was significantly reduced compared to that of shGFP-treated group (Supplementary Fig. S5D). We also analyzed the correlation between the expression of SMARCE1 and the three-target genes in Versteeg dataset and TARGET database. The results showed that SMARCE1 expression positively correlated with PLK1, ODC1 and E2F2 expression (Fig. 5E and Supplementary Fig. S5E).

Fig. 5: SMARCE1 induces the expression of the MYCN target genes PLK1, ODC1 and E2F2.
figure 5

A GSEA plots depicting enrichment of the MYCN target gene signature and PLK1 pathway in the transcriptional profiling of samples with high SMARCE1 expression in the GSE16476 cohort. FDR false discovery rate, NES normalized enrichment score. B Kaplan–Meier curves showed the probability of overall survival of patients according to PLK1, ODC1 and E2F2 mRNA expression levels in the 88 neuroblastoma samples in the Versteeg dataset. C Real-time qPCR analysis of PLK1, ODC1, E2F2 and MYCN expression upon SMARCE1 shRNA knockdown in SK-N-BE(2)C and IMR32 cells. D Immunoblot detection of PLK1, ODC1, E2F2 and MYCN protein levels after SMARCE1 shRNA knockdown. E The expression of SMARCE1 positively correlated with the expression of PLK1, ODC1 and E2F2 in the Target public database. All data were shown as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001, ns not significant.

Full size image

Similarly, dual-luciferase assays using the PLK1, ODC1 and E2F2 promoters displayed that SMARCE1 deficiency significantly downregulated the activity of the three promoters (Fig. 6A–C). We further carried out ChIP experiments to determine the exact mechanism of transcriptional regulation. Using SMARCE1 antibody, we found SMARCE1 protein was mainly enriched within −132 to +156 bp region (Amplicon E) of PLK1, −335 to +35 bp region (Amplicon D) of ODC1, and −257 to +55 bp region (Amplicon D) of E2F2 (Fig. 6D–F). Similarly, MYCN was mainly recruited to Amplicon E of PLK1, Amplicon D of ODC1, and Amplicon D of E2F2 when used MYCN antibody in the ChIP experiments (Supplementary Fig. S6A). Supported of this, canonical and non-canonical E-boxes are found in these regions (Supplementary Fig. S6B). Analysis of publicly available ChIP sequencing data [39, 32, 33] revealed that MYCN and SMARCE1 could bind to the PLK1, ODC1 and E2F2 promoter and the binding peaks overlapped in other cancer cells (Supplementary Fig. S6C). Importantly, SMARCE1 knockdown attenuated the promoter occupancy of MYCN, suggesting that the recruitment of MYCN protein to promoters requires SMARCE1 protein (Fig. 6G and Supplementary Fig. S6D). Re-ChIP (sequential ChIP) experiments further demonstrated that SMARCE1 cooperated with MYCN to occupy the PLK1, ODC1 and E2F2 promoters (Fig. 6H).

Fig. 6: SMARCE1 and MYCN form co-activators that regulates PLK1, ODC1 and E2F2 transcription.
figure 6

A–C Relative luciferase activity of reporter constructs containing the PLK1, ODC1 or E2F2 promoter in SK-N-BE(2) cells transfected with SMARCE1 shRNA. D Upper panel, a total of six sets of primers were designed within the human PLK1 promoter; lower panel, ChIP-qPCR analysis using anti-IgG and anti-SMARCE1 antibodies in SK-N-BE(2) cells. E, F Upper panel, a total of five sets of primers were designed within the human ODC1 or E2F2 promoter; lower panel, ChIP-qPCR analysis using anti-IgG and anti-SMARCE1 antibodies in SK-N-BE(2) cells. G ChIP analysis with MYCN and SMARCE1 antibodies upon SMARCE1 shRNA knockdown in SK-N-BE(2)C cells. H SK-N-BE(2) cells analyzed by re-ChIP to detect mutual binding of SMARCE1 and MYCN to promoters. The first round (initial ChIP) antibody (IgG or MYCN) is indicated below the plots and the second round (re-ChIP) antibody (IgG or SMARCE1) is shown in the color key. All data were shown as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001, ns not significant.

Full size image

Finally, we overexpressed PLK1, ODC1 or E2F2 in SMARCE1-depleted SK-N-BE(2)C cells (Fig. 7A). As a result, the cell proliferation and colony formation were significantly restored upon the re-introduction of PLK1, ODC1 or E2F2 in the SMARCE1-depleted SK-N-BE(2)C cells (Fig. 7B–D). Next, we performed subcutaneous tumorigenesis assays to test tumorigenic potential in vivo, and the result indicated that overexpression of PLK1, ODC1 or E2F2 restored tumor growth after SMARCE1 knockdown (Fig. 7E–G).

Fig. 7: SMARCE1 promotes neuroblastoma progression via regulating PLK1, ODC1 and E2F2 expression.
figure 7

A Immunoblot detection of SMARCE1, PLK1, ODC1 or E2F2 protein levels after overexpressing with PLK1, ODC1 or E2F2 in SMARCE1 shRNA knockdown SK-N-BE(2)C cells. B SK-N-BE(2)C cells stably expressing the negative control or SMARCE1 shRNA were further infected with a lentivirus overexpressing PLK1, ODC1 or E2F2. Cell viability was assessed. C, D The colony-formation ability was tested after overexpressing PLK1, ODC1 or E2F2 in SMARCE1 shRNA knockdown SK-N-BE(2)C cells. E–G Xenograft assay was performed after overexpressing PLK1, ODC1 or E2F2 in SMARCE1 shRNA knockdown SK-N-BE(2)C cells. Tumor weight was analyzed by scatterplot with horizontal lines indicating the mean. Tumor volume was measured and indicated by curves. H Model for MYCN-SMARCE1 cooperation in neuroblastoma. All data were shown as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001, ns not significant.

Full size image

Discussion

Neuroblastoma is the most common malignancy of early childhood and arises in the developing sympathetic nervous system (from any neural crest element) [1]. In this study, we identified SMARCE1 as a potential target for prognosis and therapy of neuroblastoma, especially those with MYCN amplification. Mechanistically, we revealed MYCN-SMARCE1-PLK1/ODC1/E2F2 regulatory axis in the MYCN-amplified neuroblastoma cells, which is involved in the proliferation, self-renewal, and tumorigenicity of the tumor cells (Fig. 7H).

As the most common changes, MYCN amplification is closely correlated with poor outcomes and driving carcinogenic events in neuroblastoma [4]. MYC protein recruited transcriptional co-factors to regulate genes expression, such as TIP60 histone acetyltransferases or the p300/CBP acetyltransferase [40, 41]. Our data identified SMARCE1 as a novel binding protein of MYCN and demonstrated that the molecular interaction between SMARCE1 and MYCN might have important implications for efficient gene transcription. Consistent with this observation, SMARCE1 plays an important role in DNA replication and DNA repair, especially the transcriptional co-regulation [10, 37]. We confirmed that SMARCE1 protein interacted with MYCN protein at the truncated region of amino acids 141–350. It has been shown that MYC protein contains two highly conserved MYC boxes (MYC box III and IV) [42,43,44]. The boxes play a role in regulating DNA-binding activity and localize mostly in the binding regions of MYCN [45, 46]. Moreover, we identified that MYCN protein binds to the HMG DNA-binding domain of SMARCE1. Studies have implicated that HMG domain is required for SMARCE1 gene regulation [11, 12, 47]. SMARCE1 cooperated with MYCN to regulate the Myc target genes (PLK1, ODC1 and E2F2) transcription by binding to their gene promoters. Further research suggest that the MYCN-SMARCE1 interaction is an important factor that regulates PLK1 expression and cell proliferation (Supplementary Fig. S7A–C). Myc recruited co-activators to locally remodel the chromatin structure to achieve the function of transcriptional regulation [21, 48]. For example, MYCN interacted with p300/CBP to lead to an open chromatin structure that provides docking sites for additional proteins that promote transcription [40]. As a core subunit of ATP-dependent chromatin remodeling complex SWI/SNF, SMARCE1 might altered the nucleosome structure and chromatin accessibility for assisting the transcriptional regulation of MYCN. We suggested that the identity of the co-activator of SMARCE1 is biologically significant in MYCN-amplified neuroblastoma cells.

Previous studies have demonstrated that these chosen targets were under the regulation of MYCN and played an important role in neuroblastoma development [7, 49,50,51]. As essential regulators of cell cycle, PLK1 and E2F2 were enriched in correlated certain KEGG pathways [52, 53]. DOT1L was identified as a novel cofactor in MYCN–mediated transcriptional activation of E2F2 and ODC1 and neuroblastoma oncogenesis [51]. Moreover, C-MYC promoted cell proliferation and transformation by activating growth-promoting genes, including the ornithine decarboxylase (ODC1) and E2F2 [54].

By performing a dual-luciferase assay and ChIP analyses, we also found that MYCN regulated SMARCE1 expression by binding to its promoter. It is well known that MYCN activated the transcription of target genes through binding to Myc responsive element E-Boxes in promoters [31]. For example, recent evidence showed that MYCN regulated DOT1L expression through binding to Myc responsive element E-Boxes [51]. Interestingly, we identified an MYCN binding motif upstream of the transcription start site of the SMARCE1 promoter and confirmed MYCN protein enriching to the SMARCE1 promoter. Furthermore, the SMARCE1 promoter activity was not reduced after MYCN knockdown when the mutant SMARCE1 promoter sequence was used. Our data confirmed MYCN protein enriching to the SMARCE1 promoter and showed that SMARCE1 plays a key role in neuroblastoma due to the identity of a critical downstream factor of MYCN. Given the regulatory relationship between SMARCE1 and MYCN in this work, our finding also may extend implications to other tumor types with MYCN amplification or overexpression (e.g., retinoblastoma, glioblastoma, rhabdomyosarcoma and Wilms tumor) [55,56,57,58] (Supplementary Fig. S8A–D).

Many groups have explored the effect of targeting MYCN expression in neuroblastoma cells and indicated that direct targeting of MYCN could be promising for neuroblastoma therapy but remaining difficult [59,60,61]. To repress neuroblastoma growth effectively, the MYC pathway could potentially be targeted at different levels [62,63,64]. One strategy is to target critical downstream transcriptional targets of MYCN [65]. In this study, as a direct downstream factor of MYCN, SMARCE1 could be used as a potential molecular target for inhibiting MYCN signaling pathway in neuroblastoma. In addition, another strategy for neuroblastoma is to target the interaction of MYCN with other proteins, either with the obligate partner Max, or with co-factors recruited to the MYCN [66]. Thus, the interaction between MYCN and SMARCE1 could serve as potential targets for therapeutic interventions.

In summary, our study extends the knowledge about the key role of SMARCE1 in the tumor progress in MYCN-amplified and 17q21-ter gained neuroblastoma. The MYCN-SMARCE1-PLK1/ODC1/E2F2 regulatory axis identified in this study is valuable and could be a potential target for the treatment of neuroblastoma by exploiting the enforced SMARCE1 addiction present in MYCN-amplified tumors.

Materials and methods

Cell culture

Neuroblastoma cell lines SK-N-BE(2)C, IMR32, SHEP1, SH-SY5Y, SK-N-DZ and HEK-293FT were obtained from ATCC (Manassas, VA, USA). Cell culture methods were consistent with those described previously [67]. IMR32 were grown in alpha-minimum essential media (Gibco, New York, NY, USA). SK-N-BE(2)C were cultured in DME/F-12 1:1 media (Gibco, New York, NY, USA). SHEP1, SH-SY5Y, SK-N-DZ and HEK-293FT were cultured in Dulbecco’s modified Eagle’s medium (VivaCell, Shanghai, China). All media were supplemented with 10% fetal bovine serum.

Antibodies

The MYCN (51705S), Cyclin B1 (12231S) and PLK1 (4513S) antibodies were purchased from Cell Signaling Technology (Shanghai, China). The SMARCE1 antibody (#A300-810A) was obtained from Bethyl (Texas, USA). The ODC1 (28728-1-AP), HA (51064-2-AP) and CDK1 (19532-1-AP) antibodies were purchased from Proteintech Group (Wuhan, China). The Ki67 (ab92742), α-Tubulin (ab7291), E2F2 (ab138515) and Lamin B1 (ab16048) antibodies were purchased from Abcam (Shanghai, China). Anti-Flag (F3165) was purchased from Sigma (St Louis, MO, USA).

Plasmids, transfection, and infection

Human MYCN and SMARCE1 shRNA were purchased from Beijing Genomics Institute (BGI, Beijing, China) and then cloned into the PLKO.1 vector. The shRNA sequences were listed in Supplementary Table 1. Plasmids encoding human SMARCE1, MYCN, PLK1, ODC1, E2F2, truncated mutants of SMARCE1 and MYCN was obtained from Youbao Corporation (Changsha, China). Transfection and infection were performed essentially as previously described [68].

Immunoblot analysis

Cells were harvested and lysed with cell lysis buffer (Beyotime). Immunoblotting was performed as previously described [69].

Quantitative reverse-transcription PCR (qRT-PCR)

Total RNA was extracted using the standard TRIZOL method. Subsequently, cDNA was generated with the Reverse Transcription Kit (Promega). Quantitative real-time-PCR was performed as previously described [70]. The sequence of the primers used in this study were listed in Supplementary Table 2.

Flow cytometry analysis

Flow cytometry analysis was performed as described previously [70, 71].

Chromatin immunoprecipitation (ChIP) assays

ChIP assays were performed as previously described [72]. ChIP assay was performed with anti-MYCN antibody, anti-SMARCE1 antibody or control mouse IgG antibody (Beyotime). Quantitative real-time PCR was performed to analyze the precipitates. Primers used for qPCR were listed in Supplementary Table 3. For re-ChIP assays, we performed sequential pull-downs, first with anti-MYCN antibody followed by anti-SMARCE1 antibody using appropriate IgG controls.

GST pull-down and co-immunoprecipitation assays

GST pull-down assay was performed as described previously [73]. Briefly, SHEP1 or HEK-293FT cells were stably transfected with relevant proteins and lysed. The supernatant lysates were incubated overnight with GST or GST-bound protein immobilized on glutathione agarose beads. The beads were washed 4–5 times with buffer. The products were analyzed by immunoblotting analysis.

For endogenous MYCN or SMARCE1 immunoprecipitation assays, protein was extracted from SK-N-BE(2)C cells and co-immunoprecipitation was carried out using anti-MYCN antibody, anti-SMARCE1 antibody or control IgG as a negative control, followed by immunoblotting analysis. The appropriate bands were excised from protein gels were identified using MS/MS analysis.

For immunoprecipitation assays with ectopically expressed Flag and HA-tagged proteins, protein was extracted from HEK-293FT cells and co-immunoprecipitation was carried out using anti-Flag antibody, anti-HA antibody or control IgG as a negative control, followed by immunoblotting analysis.

Luciferase reporter assay

The SMARCE1, PLK1, ODC1, E2F2 and SMARCE1 mutant promoter sequences were synthesized by the BGI and subsequently ligated into the pGL3-basic vector (Promega). Luciferase reporter assay was performed as previously described [68].

Immunohistochemistry

Paraffin sections were dewaxed, hydrated and repaired with antigen. For antigen retrieval, sections were performed by microwave heating for 20 min in 10 mM citrate buffer (pH 6.0), followed incubating with SMARCE1, PLK1, ODC1, E2F2 and Ki-67 antibodies at 4 °C overnight. The target antigen was visualized using a horseradish peroxidase detection system (GBI Labs).

Cell growth and plate clone formation assays

The 2-(4,5-dimethyltriazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT; Sigma) colorimetric, plate clone formation assays were conducted as previously described [69, 74].

Animal studies

Five-week-old female nude (NOD/SCID) mice were purchased and housed in an SPF room. Total 1.5 × 106 SK-N-BE(2)C or IMR32 cells were collected and inoculated subcutaneously into the flanks of mice. Randomization and single-blinding were used for measurement. Tumor volumes were measured every 3 days with a caliper. Six weeks after injection of neuroblastoma cells, mice were euthanized and the tumors were resected. Animal studies were performed in accordance with the Guidelines of the Institute for Laboratory Animal Research, Southwest University (Chongqing, China).

Duolink PLA

This technique was performed using Duolink In Situ Red Starter Kit (#DUO92101, Sigma-Aldrich) according to the manufacturer’s instructions. The detailed technique can be found online (www.olink.com/products-services/duolink/how-useduolink). Briefly, neuroblastoma cells were fixed in 4% paraformaldehyde for 20 min at room temperature and blocked for 1 h with Duolink blocking solution at 37 °C. Then, cells were incubated with primary antibodies overnight at 4 °C and conjugated with oligonucleotides (PLA probe MINUS and PLA probe PLUS) for 120 min at 37 °C. Finally, cells were incubated sequentially with ligation solution and amplification solution followed by analysis with fluorescence microscopy.

Patient data analysis and patient tumor tissues

Patient data and gene expression datasets were obtained from R2: microarray analysis and visualization platform (https://hgserver1.amc.nl/cgi-bin/r2/main.cgi). Kaplan–Meier analysis and survival curves were carried out by using GraphPad Prism (version 6.0, GraphPad Software, San Diego, CA, USA). All cutoff values for separating high and low expression groups were determined by the online R2 database algorithm.

Tumor samples were obtained from Chaoying Biotechnology Co., Ltd. (Henan, China). Tissue analysis was approved by the Ethics Committee of Southwest University of China. All the patients provided written informed consent to participate.

ChIP-seq analysis

We obtained several other publicly available ChIP-seq datasets (GSM2113542, GSM2214113, GSM2113529, GSM2113526 and ENCSR157TCS) [32, 33, 39] to complement our experimental results. These files were aligned to the human genome (GRCh38) by using Integrative Genomics Viewer (v2.11.1).

Statistical analysis

All observations were confirmed by at least three independent experiments, and both sample size and significance analysis are indicated in the legends. All the results acquired in this study were presented as the means ± standard deviation. Two-tailed Student’s t test was performed for paired samples; p < 0.05 was considered statistically significant.