Abstract
Background
Medulloblastoma (MB) is the most common pediatric malignant brain tumor. Infiltration of tumor-associated macrophages (TAMs) and m6A modification of RNA are correlated with poor prognosis and tumor progression in the Sonic Hedgehog (SHH) subtype (SHH-MB). However, the relationship between TAMs infiltration in SHH-MB and m6A modification status during tumor progression remains unclear.
Methods
Expression of m6A modification-related proteins was assessed in 40 cases of SHH-MB. Genes affected by TAM-derived exosomes were identified with methylated RNA immunoprecipitation sequencing. Mechanisms of m6A modification of FOXD1 were evaluated and combinatorial treatment with AAV2/9-shFOXD1 and PD-1 inhibitors was investigated in the NeuroD2:SmoA1 mouse model.
Results
TAMs infiltration led to decreased METTL14 expression, which was mediated by TAM-derived exosomes containing METTL14-specific microRNAs. In turn, this led to lower levels of m6A modifications. Through a screen, FOXD1 was identified as a critical downstream target of TAM-derived exosomes, and its expression level was correlated with poor prognosis in SHH-MBs. Importantly, knockdown of FOXD1 in SHH-MB cells significantly promoted the release of chemokines CXCL10/11, resulting in CD8+ T cell recruitment. Furthermore, treatment with AAV2/9-shFOXD1 significantly enhanced the antitumor effect of the PD-1 inhibitor in transgenic SHH-MB mice.
Conclusions
Our study revealed for the first time that TAM-derived exosomes modulate m6A levels in SHH-MB, which promotes tumor progression via FOXD1. We identified FOXD1 as a novel therapeutic target whose inhibition sensitizes SHH-MB to immune checkpoint blockade.
Keywords: FOXD1, immunotherapy, medulloblastoma, m6A, TAMs
Graphical Abstract
Graphical Abstract.
Key Points.
TAM-derived exosomes reduce m6A modification levels in SHH-MB.
FOXD1 upregulated by m6A modification in SHH-MB suppresses cytotoxic T cell recruitment.
FOXD1 knockdown enhances the sensitivity of SHH-MB tumors to immunotherapy.
Importance of the Study.
Infiltration of tumor-associated macrophages (TAMs) is associated with poor prognosis, and N6-methyladenosine epigenetic modification is known to be correlated with tumor progression in medulloblastoma (MB). However, the association of TAMs infiltration with m6A modification status remains poorly understood. This is, to the best of our knowledge, the first study to demonstrate how TAM influences m6A levels in the SHH subtype of MB. We show that m6A modifications upregulate FOXD1, which modulates the immune microenvironment. Importantly, we show that in vivo combinatorial treatment with FOXD1 increased the sensitivity of SHH-MB to PD-1 inhibition, highlighting a potential novel therapeutic strategy for MB.
Medulloblastoma (MB) is the most common malignant pediatric brain tumor occurring in the cerebellum.1 Currently, MB is classified into 4 molecular subgroups: Wingless and Int-1 (WNT), Sonic Hedgehog (SHH; with or without TP53 mutation), Group3 (G3), and Group4 (G4).2 Approximately 30% of medulloblastomas are classified as SHH.3 We showed previously that the SHH subtype is characterized by a higher level of infiltration of tumor-associated macrophages (TAMs) in the tumor microenvironment (TME) than other subtypes.4,5 Elevated TAMs can lead to the formation of an inhibitory immune microenvironment (IME), resulting in reduced sensitivity to immunotherapy,6,7 but the underlying mechanism is unclear.
Exosomes are double-layered membrane extracellular vesicles of 30–150-nm diameter, and they play crucial roles in intercellular communication within the IME.8 Exosomes derived from TAMs carry microRNAs and proteins that significantly impact cellular proliferation, tumor invasion, immune suppression, and therapeutic resistance.4,9 However, the mechanism of how TAM-derived exosomes affect tumor cells in SHH-MB remains unknown.
MB tumors undergo significant epigenetic alterations.10–13 N6-methyladenosine (m6A) RNA methylation represents one of the most prominent chemical modifications in messenger and non-coding RNA in eukaryotic cells,14 and m6A can remodel the TME in various tumors, including gliomas.15,16 A recent study showed that high m6A levels in MB tumor cells activate the SHH pathway and promote malignant progression.17 The role of TAMs in regulating this posttranscriptional modification in MB remains to be clarified.
The present study reports that TAM-derived exosomal miRNAs downregulate METTL14 expression in tumor cells, leading to decreased overall m6A modification and elevated FOXD1 expression in SHH-MB. In turn, FOXD1 reduces the release of cytokines for CD8+ T cell recruitment, resulting in the formation of an inhibitory TME in SHH-MB. In vivo experiments confirmed that FOXD1 knockdown significantly enhances the tumor response to PD-1 inhibition, uncovering a novel combinatorial treatment strategy for SHH-MB.
Materials and Methods
Patients and Specimens
This study was approved by the Institutional Review Board of Peking University (No. IRB00006761-M2023465). All samples underwent diagnosis and histological subtyping by 2 neuropathologists (Q.C. and D.Z.). The NanoString nCounter Analysis System (NanoString Technologies) was used in the identification of molecular subtypes.12,13 Characteristics of patients and tumor are summarized in the Supplementary Table.1
Cell Culture
Human Daoy (HTB-186, ATCC) MB and THP-1 (TIB-202, ATCC) cell lines were purchased from the Cell Resource Center, Peking Union Medical College. The human ONS-76 (IFO50355, JCRB) MB cell line was kindly provided by Prof. Xiu-Wu Bian of Southwest Hospital of Army Medical University. Daoy, ONS-76, and THP-1 cells were cultured as previously described.5
Establishment of Stable Cell Lines
pCDH-CMV-MCS-EF1-Puro was used as the overexpression vector, and pLKO.1-puro as the knockdown vector. For virus packaging, HEK293T cells were employed, and the virus was collected 48 hours after cell infection. Oligonucleotides of short hairpin RNAs (shRNAs) are listed in the Supplementary Table.2
Immunoblotting Analysis
Protein extraction, electrophoresis, immunoblotting, and chemiluminescence detection were performed using a gel imaging system (Bio-Rad).
Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR)
Total RNA was extracted using TRIzol (Invitrogen), followed by reverse transcription, and cDNA products were amplified for further analysis. Primers for the targeted genes are listed in the Supplementary Table.3
Mouse Model Assay
The NeuroD2:SmoA1 mouse model (#008831) is a well-established model of SHH-MB,5 these mice were purchased from Jackson Labs. The mice received STM2457 or rAAV2/9-shFOXD1, in combination with anti-PD-1 antibodies or isotype controls (Mab isotype), and tumor progression was monitored via weekly MRI. All animal care followed ethical guidelines and was approved by the Animal Care Committee of Peking University Health Science Center (permit number BCJH0201).
Luciferase Reporter Assay
The target sequence was cloned downstream of firefly luciferase (F-Luc). After co-transfection with pmirGLO and shVector for 48 hours, F-Luc and renilla luciferase (R-luc) were assayed using the Dual-Glo Luciferase Reporter Gene Assay Kit (11404ES60).
Exosome Isolation, Analysis, and Tracking
Isolation, identification, and tracking of exosomes were performed as previously described.18 For in vitro assays, 1 × 108 exosomes (as determined by NanoSight analysis) were added to 0.1 × 106 cells.
Immunohistochemistry and Immunofluorescence
Paraffin-embedded sections were subject to antigen retrieval, blocking, and immunolabeling, followed by 3,3′-diaminobenzidine staining or fluorescent secondary antibody labeling. Imaging was performed using a ZEISS LSM 700 microscope.
Liquid Chromatography–Mass Spectrometry (LC–MS/MS) Assay for m6A Quantification
Quantification of m6A was performed according to a previously established protocol with slight modifications.16 mRNA was isolated, hydrolyzed with nucleases, and then analyzed by LC-MS/MS.
m6A RNA Immunoprecipitation and qRT-PCR
Total RNA was extracted using Trizol, and 2 μg of RNA was subjected to m6A immunoprecipitation (IP) with an m6A-specific antibody (#202003, Synaptic Systems). Total RNA and m6A-IP RNA (2 ng each) were used as templates for qRT-PCR. The housekeeping gene HPRT1 was used as an internal control because it lacks m6A peaks.16
Methylated RNA Immunoprecipitation Sequencing
After treatment of Daoy cells with THP-1-derived macrophage exosomes, total RNA was extracted and m6A RNA was immunoprecipitated (#202003, Synaptic Systems) as described above. Sequencing, library preparation, and data analysis were performed by Wuhan Seqhealth Co., Ltd. (China) according to the manufacturer’s instructions.19
RNA Stability
To assess the mRNA stability of FOXD1 in Daoy cells under various treatment conditions, cells were treated with actinomycin D (HY-17559) at 5 μg/ml. Subsequently, cells were collected at specific time points, and RNA was isolated for qRT-PCR analysis. The half-life (t1/2) of FOXD1 mRNA was calculated by t1/2 = ln(2)/ slope, using ACTB as a normalization control.
Database Analysis
The miRNA sequencing data of M2-macrophage exosomes from THP-1 cells were obtained from GEO dataset GSE97467. Kaplan–Meier survival analysis and correlation analysis involving FOXD1 were performed on GEO dataset GSE85217, using the R2 Genomics Analysis and Visualization Platform (r2.amc.nl).
Statistical Analyses
Data points are presented as mean ± SD from at least 3 independent experiments. Unpaired Student’s t-tests were used to compare 2 groups. One-way or 2-way analysis of variance (ANOVA) followed by Bonferroni tests for multiple comparisons were used for comparisons of multiple groups and factors. All statistical tests were 2-sided tests. GraphPad Prism 9.0 was utilized for data analysis. P-value < .05 was considered statistically significant.
Results
TAMs in the IME Affected m6A Methylation
To determine whether TAMs have an impact on the RNA epigenetic regulation in MB tumor cells, we conducted a quantitative analysis of TAM infiltration across MB subtypes. TAM infiltration levels were significantly higher in SHH MB tumors compared to other subtypes (Figure 1A). Immunohistochemistry (IHC) staining of tissue sections of 40 SHH-MB tumors and 5 normal cerebellar tissue samples revealed that METTL14 expression was not significantly different between SHH-MB and normal cerebellar tissues (Figure S1A, B). However, higher TAM infiltration was associated with decreased METTL14 expression (Figure 1B, C). Conversely, METTL3 expression was significantly increased in SHH-MB compared with normal cerebellar tissues (Figure S1C, D). However, METTL3 expression was not correlated with TAM infiltration across SHH-MB tumors (Figure 1D, E). Expressions of FTO and ALKBH5 (demethylases for m6A) were significantly reduced in SHH-MB compared with normal cerebellar tissue (Figure S1E-H), but their expressions were not correlated with TAM infiltration (Figure 1F-I). These results suggest that the extent of TAM infiltration may be associated with changes in m6A methylation within SHH-MB primary tumors.
Figure 1.
M2-type macrophage infiltration and METTL14 expression levels are negatively correlated in SHH-MB. (A) Representative images of CD163 expression in WNT, SHH, G3, and G4 MB—specimens (left). Quantification of M2-like macrophage (CD163) numbers in a 20X field of view for each of the 4 MB subtypes (right); (B, D, F, H) IHC staining of METTL14 (B), METTL3 (D), FTO (F), ALKBH5 (H) and CD163 was performed in the same SHH-MB sample section. (C, E, G, I) Linear regression analysis revealed correlations between CD163 and (C) METTL14, (E) METTL3, (G) FTO, and (I) ALKBH5. *P < .05, **P < .01, ***P < .001, ****P < .0001, NS not significant, by Wilcox test.
Exosomes From TAMs Promote SHH-MB Tumor Progression In Vitro
We derived M2 macrophages from THP-1 cells (Figure S2A), collected exosomes from the M2 macrophages, and validated the exosomes by multiple methods. The structures of the exosomes were confirmed by TEM, which revealed distinctive cup-shaped, double-membraned structures (Figure S2B). Nanoparticle tracking analysis showed the diameters of the exosomes to be approximately 118.3 nm (Figure S2C). The expressions of exosomal markers (TSG101, CD63) were confirmed by western blotting, using cytoplasmic protein calnexin as a negative control (NC; Figure S2D).
We also confirmed the successful delivery of exosomes to tumor cells using an exosome-uptake assay with PKH67 labeling, which showed green fluorescence in the cytoplasm of tumor cells (Figure S2E). This fluorescence signal disappeared when Dynasore (an endocytosis inhibitor) was added. These results demonstrate that TAM exosomes were internalized by the tumor cells. Henceforth, exosomes released by M0 and M2 macrophages derived from THP-1 will be referred to as M0-EXO and M2-EXO, respectively.
Treatments of Daoy and ONS-76 cells with macrophage-derived exosomes demonstrated that M2-EXO significantly promoted proliferation, migration, invasion, and clonogenic ability of tumor cells (Figure S2F-I). These effects can be blocked by endocytosis inhibition with Dynasore, which confirms that the observed effects are due to exosome uptake by the tumor cells. These in vitro experiments therefore show that M2-EXO can significantly accelerate the tumor progression of SHH-MB.
TAM-Derived Exosomal miRNAs Target METTL14 to Decrease Global m6A Level
To further explore whether exosomes derived from M2 macrophages (TAMs) can induce the dysregulation of m6A methylation in SHH-MB, we quantified m6A modification levels in tumor cells after M2-EXO versus M0-EXO treatment. The results show that M2-EXO treatment caused m6A levels in Daoy cells to decrease significantly by 69% compared with M0-EXO treatment (Figure 2A).
Figure 2.
miR-320e/196b-5p/628-5p from exosomes of M2-type macrophages can downregulate overall m6A modification in SHH-MB cells. (A, B) Daoy and ONS-76 cells were treated with M0-EXO or M2-EXO for 2 days. (A) the m6A/A ratio of total mRNA was determined by LC–MS/MS.(B) Protein and mRNA expressions of METTL3/14, FTO, and ALKBH5 were determined by western blot (left) and qPCR (right) analysis, respectively. (C) A diagram showing high miRNA expression in M2-type macrophage-derived exosomes compared with SHH-MB, and the database-predicted miRNA-targeting relationship with METTL14. (D) Western blot and qPCR analysis showing the ability of 5 miRNAs to degrade METTL14 at the RNA and protein levels. (E) Schematic representation of wild type and mutant-type binding sites between the 3“UTR of METTL14 and miR-320e. (F) Relative luciferase activity of 3”UTR-METTL14-luc constructs in Daoy cells after transfection of miR-320e mimics/NC (negative control). (G) The level of miR-320e/196b-5p/628-5p in Daoy cells was evaluated by qPCR before and after treatment with M2-EXO. (H, I) Daoy and ONS-76 cells were incubated with M0-EXO, M2-EXO, or M2-EXO plus co-transfection with miR-320e/196b-5p/628-5p inhibitors for 2 days. (H) The m6A/A ratio of total mRNA was determined by LC-MS/MS. (I) METTL14 mRNA and protein expression levels were determined by qPCR (left) and western blot (right) analyses, respectively. Data are presented as means ± SD from 3 independent experiments. *P < .05, **P < .01, ***P < .001, ****P < .0001, NS, not significant.
To clarify whether the TAM-derived exosomes directly impact m6A levels in SHH-MB tumor cells, we measured the protein expression of key players in m6A modification. M2-EXO significantly reduced METTL14 expression in both Daoy and ONS-76 cells, but it had no apparent effect on METTL3, FTO, or ALKBH5 expression (Figure 2B) or on the enzymatic activity of METTL3 (Figure S3A).
We then hypothesized that some cargos packed within M2-EXO, such as miRNAs, can regulate the expression of METTL14. To test this hypothesis, we first collected publicly available miRNA sequencing data of exosomes released by THP-1-derived M0 or M2 macrophages (GSE97467). We then identified miRNAs that are highly expressed in M2-EXO and excluded miRNAs that are expressed in Daoy cells (using in-house generated data5), forming a set of 243 exosomal miRNAs that are specific to M2-EXO (Figure 2C). We cross-referenced these miRNA candidates with multiple miRNA target databases (StarBase, miRDB, miMap, and miWalk), and we identified 3 miRNAs (miR-196b-5p, miR-320e, and miR-628-5p) that potentially target METTL14.
Next, we experimentally validated that the candidate miRNAs directly downregulate METTL14 expression. Transfections with any of the 3 candidates caused a reduction in METTL14 expression at both the RNA and protein levels (Figure 2D), whereas irrelevant miRNAs (miR-409-3p and miR-34b-5p) did not influence METTL14 expression. These results indicate that miR-196b-5p, miR-320e, and miR-628-5p are bona fide candidates for targeting METTL14. Additionally, none of these miRNAs significantly affected METTL3, FTO, or ALKBH5 expression (Figure S3B). We further validated that miR-320e targets METTL14 using a dual-luciferase reporter assay (Figure 2E). The 3′-untranslated region (3′UTR) of METTL14 was cloned downstream of the luciferase reporter, which was transfected into Daoy and ONS-76 cells alongside miR-320e. miR-320e co-transfection significantly reduced the activity of the luciferase reporter in the construct with the wild type (WT) METTL14 3′UTR, and this effect was abolished when the miRNA binding site in the METTL14 3′UTR was mutated (Figure 2F). Similar results were obtained for miR-196b-5p and miR-628-5p (Figure S3C). These results provide validation that miR-320e, miR-196-5p, and miR-628-5p directly downregulate METTL14 expression.
We performed further validation and rescue experiments to confirm that M2 macrophage-derived exosomes modulate m6A levels in tumor cells via miRNA-mediated inhibition of METTL14. First, qRT-PCR results show that M2-EXO treatment significantly upregulated the expression levels of the METTL14-targeting miRNAs in Daoy and ONS-76 cells (Figure 2G and Figure S3D). Second, we show that M2-EXO-mediated suppression of m6A levels in tumor cells can be rescued by co-transfection of miR-320e, miR-196b-5p, and miR-628-5p inhibitors into M2-polarized THP-1 cells before exosome collection and treatment of the tumor cells (Figure 2H). Furthermore, joint inhibition of these 3 miRNAs in the M2-polarized cells before exosome collection also rescued the effect of M2-EXO on the RNA and protein expression levels of METTL14 (Figure 2I). Third, exosomes derived from M2 macrophages pretreated with miRNA inhibitors also have significantly attenuated effects on tumor cell proliferation, migration, invasion, and clonogenicity (Figure S3E-H). These results confirm that exosomal miRNAs (miR-320e, miR-196b-5p, and miR-628-5p) released by M2 macrophages inhibit METTL14 and consequently reduce m6A levels in SHH-MB tumor cells, leading to increased cellular proliferation, migration, invasion, and clonogencity.
Finally, we conducted additional rescue experiments to demonstrate that the effects exerted by M2-EXO on the tumor cells are specifically mediated by METTL14. We established METTL14 knock-out Daoy and ONS-76 cells and verified the knock-out by Western blotting (Figure S3I). As expected, M2-EXO treatment did not have an effect on the proliferation and invasion of MELL14−/− tumor cells (Figure S3J, K left). Additionally, overexpression of METTL14 in tumor cells after treatment with M2-EXO significantly reversed the effects of M2-EXO on tumor cell proliferation and invasion (Figure S3J, K; right).
Therefore, our results collectively indicate that M2 macrophages-derived exosomal miRNAs, which are taken up by MB tumor cells, directly downregulate METTL14 expression and reduce m6A levels in the tumor cells, leading to increased cancer cell proliferation, migration, and invasion.
m6A Hypomethylation of FOXD1 Increases Protein Expression and Promotes SHH-MB Malignancy
To comprehensively characterize the changes in m6A RNA methylation levels induced by M2 macrophage-derived exosomes, we used methylated RNA immunoprecipitation sequencing (MeRIP-seq) to characterize the m6A methylomes after M2-EXO versus M0-EXO treatment in Daoy cells. Globally, M2-EXO treatment caused a significant decrease in m6A methylation of the 5′ untranslated and coding sequence regions (5′UTR and CDS, respectively; Figure 3A). We also characterized the RNA-seq expression profiles of Daoy cells following exosome treatment, and we integrated the RNA expression alterations with the m6A methylation changes. We then categorized genes into 4 groups based on RNA-seq expression and m6A methylation responses to M2-EXO vs. M0-EXO treatment: hyper-up (hypermethylation m6A level and upregulated RNA expression), hyper-down (hypermethylated and downregulated), hypo-up (hypomethylated and upregulated), and hypo-down (hypomethylated and downregulated; Figure 3B). Differentially expressed genes under M2-EXO versus M0-EXO treatment were identified based on fold change >2 and P < .05, and differentially m6A methylated genes were identified based on fold change >3 and P < .05. Since M2-EXO causes an overall reduction in m6A level in tumor cells, we predicted that genes with hypomethylated m6A and significant differential RNA expression would be potential targets of exosome-mediated m6A modulation. These candidate targets include hypo-up genes (THBS1, FOXD1, TACR2) and hypo-down genes (KRT75 and ADAM28).
Figure 3.
M2 macrophage-derived exosomes affect FOXD1 expression via m6A modification. (A) The density distribution of m6A peaks across mRNA transcripts after THP-1-derived M0/2-type macrophages exosome treatment of Daoy cells. Regions of the 5′UTR, CDS, and 3′UTR were split into 100 segments, and the percentages of total m6A peaks within each segment were determined. (B) Distribution of peaks (fold change > 1 or < − 1, P < .05) with significant changes at both the RNA expression and m6A level after exposure of Daoy cells to M0/2-type macrophage-derived exosomes. (C) qPCR analysis of alterations in the RNA level of 5 representative genes in Daoy and ONS-76 cells. (D) Metascape enrichment analysis of the 944 genes exhibiting 2-fold change in m6A expression (downregulation) after exposure of Daoy cells to exosomes derived from M2-type macrophages compared with those from M0-type macrophages. (E) Kaplan–Meier survival curves of overall survival probability based on FOXD1 mRNA expression in patients with SHH-MB. Log-rank analysis was used to compare the differences between the 2 groups. Data were extracted from GEO (GSE85217). (F) Representative images of IHC staining of FOXD1 in SHH-MB and paired normal cerebellum samples (left), statistical analysis (right). (G) m6A peaks were enriched in the CDS of FOXD1 genes from m6A RNA immunoprecipitation (RIP) sequencing data. For each track, immunoprecipitation (IP) is indicated with blue and red, while the corresponding input is indicated with gray. Gene structure is shown under the track. Two biological replicates are shown for each group. Squares indicate decreases of m6A peaks in Daoy cells after M2-type macrophage-derived exosome treatment. (H) m6A RIP-qPCR and (I) protein expression analysis of FOXD1 in Daoy cells after M0/2-type macrophage-derived exosome treatment. Data are presented as means ± SD from 3 independent experiments. *P < .05, **P < .01, ***P < .001, ****P < .0001, NS, not significant.
Given that M2-derived exosomes exerted similar phenotypic effects on both Daoy and ONS-76 cells, we performed qRT-PCR validation of the M2-EXO-mediated and m6A-hypomethylated candidate targets in both cell lines. FOXD1 was the only candidate target that exhibited a significant and consistent change in RNA expression after M2-EXO treatment across both cell lines (Figure 3C). Kyoto Encyclopedia of Genes and Genome enrichment analysis revealed that genes with m6A hypomethylation are enriched in the “chemotaxis” and “negative regulation of immune system process” pathways (Figure 3D), indicating that TAMs may reshape the IME of SHH-MB tumors. Taking these results into consideration, we next focused on FOXD1 as a potential downstream target of m6A methylation changes induced by exosomes derived from M2 macrophages.
Bioinformatic analysis on public SHH-MB datasets (n = 172) showed that high expression of FOXD1 is associated with poorer overall survival (Figure 3E). In our cohort of 40 tumors, IHC revealed a marked increase in FOXD1 protein expression in SHH-MB tumors versus normal cerebellum tissue (Figure 3F), and FOXD1 expression was also positively correlated with TAM infiltration (Figure S4A).
A deeper analysis of our MeRIP-seq data showed that FOXD1 mRNA is differentially m6A-methylated in the CDS region after M2-EXO treatment (Figure 3G). We further verified this specific hypomethylation in FOXD1 by m6A immunoprecipitation and qRT-PCR (Figure 3H). Furthermore, M2-EXO treatment significantly upregulated FOXD1 protein expression in Daoy and ONS-76 cells (Figure 3I). This provides further support for FOXD1 as a candidate downstream target of M2-derived exosomes.
To establish the role of FOXD1 in primary SHH-MB tumors, we performed in situ imaging analyses using multiplex IHC (mIHC) on our SHH-MB cohort. These analyses revealed that primary tumor sections with higher TAM infiltration levels exhibit downregulation of METTL14 expression, lower m6A level, and upregulation of FOXD1 (Figure S4B-C).
To ascertain that m6A methylation regulates FOXD1 expression, we performed additional knockdown experiments of m6A methylases (METTL14 and METTL3) and demethylases (FTO and ALKBH5) in Daoy and ONS-76 cells, respectively. We found that RNA and protein expression of FOXD1 negatively correlated with m6A level in vitro (Figure S4D-G).
Next, we investigated whether FOXD1 is also regulated by m6A modification in vivo using the NeuroD2: SmoA1 transgenic SHH-MB mouse model. We reduced the overall m6A levels in the tumors by treatment with STM2457, an inhibitor of m6A modification (Figure S4H). m6A reduction led to significant upregulation of FOXD1 protein expression (Figure S4I-J). We then relate these results to human tumors by analyzing the RNA expression data of SHH-MB tumors (GSE85217; n = 197). In human SHH-MB, FOXD1 expression is negatively correlated with the expression of m6A methyltransferases METTL14 and METTL3, and it is positively correlated with the expression of m6A demethylase ALKBH5 (Figure S4K). Taken together, m6A modification negatively regulates FOXD1 expression in mouse and human SHH-MB tumors.
m6A Modification Affects the Stability of the FOXD1 Transcripts
We conducted in vitro experiments to further investigate the mechanism of how METTL14 and m6A modification upregulate FOXD1 expression. We created a construct in which the luciferase reporter is under the transcriptional control of the FOXD1 promoter, and we found that METTL14 knockdown did not affect FOXD1 promoter activity. This excludes the possibility that METTL14 regulates FOXD1 at the transcriptional level (Figure S5A). We then tested whether m6A regulates the nuclear export of FOXD1 transcripts, however, knockdown of METTL14 did not affect nuclear retention of FOXD1 mRNA (Figure 4A, Figure S5B). Next, we tested whether m6A modification can regulate FOXD1 mRNA stability. To do so, we used actinomycin D (Act-D) to inhibit transcription and measured the half-life of FOXD1 mRNA transcripts. Knockdown of METTL14 significantly increased the half-life of FOXD1 transcripts (Figure 4B), whereas knockdown of ALKBH5 markedly reduced the stability of FOXD1 mRNA (Figure S5C). These data indicate that m6A modification accelerates the degradation of FOXD1 mRNA, thereby reducing FOXD1 levels in SHH-MB cells.
Figure 4.
YTHDF2 influences FOXD1 mRNA stability by recognizing m6A modification
(A) Relative levels of nuclear versus cytoplasmic FOXD1 mRNA in Daoy and ONS-76 cells with the shVector and shMETTL14, respectively. (B) After treatment with Act-D (5μg/ml for 0/1/2/4/8 h), the mRNA level of FOXD1 was measured after shVector versus shMETTL14 transfection in Daoy cells. (C) Schematic representation of the mutated (GGAC to GGCC) CDS of the pmirGLO vector, used to investigate the impact of m6A modification in this region of FOXD1 expression (left). Relative luciferase activity of F-Luc/R-Luc of pmirGLO-CDS-WT or pmirGLO-CDS-Mut-1/-2/-3 in shVector and shMETTL14 Daoy cells (right). (D) The mRNA and (E) protein levels of FOXD1 in siYTHDF2 and their corresponding control cells were measured by qRT-PCR and Western blot. (F) After treatment with Act-D (5μg/ml for 0/1/2/4/8h), the mRNA level of FOXD1 was detected in shVector and shYTHDF2 cells. (G) Daoy cells were transfected with siNC, siYTHDF2-1, or siYTHDF2-1 + (METTL14 overexpression) followed by Act-D treatment for 0/1/2/4/8h. The mRNA level of FOXD1 was checked by qRT-PCR. (H) Western blot of YTHDF2 and FOXD1 in Daoy cells transfected with vector, wild type (WT) or recognition defective (MUT) YTHDF2. (I) Lifetime of FOXD1 mRNA in Daoy cells over-expressing WT or m6A-recognition-defective YTHDF2 (W432A/W486A/W491A, MUT). Transcription was inhibited by Act-D; (J) qPCR analysis of the mRNA stability of FOXD1 in shVector and shYTHDF2 Daoy cells with or without expression of wild type (WT) or m6A-recognition-defective YTHDF2 (MUT) after transcription inhibition with Act-D; (K) Daoy cells were transfected with Flag-tagged YTHDF2 (WT or MUT). Flag-RIP-qPCR showed the interaction of FOXD1 transcripts with YTHDF2. Data are presented as means ± SD from 3 independent experiments. *P < .05, **P < .01, ***P < .001, ****P < .0001, NS, not significant.
Finally, we sought to identify m6A modification sites in the FOXD1 transcript. The SRAMP software found 3 m6A modification sites with high confidence scores (Figure S5D), and we proceeded to experimentally validate these predictions. Downstream of the firefly luciferase (F-Luc) reporter, we inserted the WT FOXD1 coding sequence, or 3 different mutant variants (MUT1/2/3) with each candidate m6A site mutated (GGAC to GGCC; Figure 4C). Knockdown of METTL14 significantly increased reporter activity in the FOXD1 WT construct by ~2.8-fold but it did not affect reporter activity in the MUT1 construct with a mutation in the first m6A candidate site (Figure 4C), which suggests that the first m6A site in the FOXD1 coding sequence may be critical for regulation of mRNA stability. We further measured the mRNA stability of the reporter constructs with wild-type FOXD1 sequence versus mutant sequences. Indeed, mutation at the first candidate m6A site abrogated the difference in mRNA stability following METTL14 knockdown (Figure S5E). Together, our data indicate that FOXD1 mRNA stability is regulated by methylation at an experimentally validated m6A site, which may potentially influence the secondary mRNA structure (Figure S5F).
YTHDF2 Promotes the Degradation of FOXD1 Transcripts in an m6A-Dependent Manner
To determine how m6A regulates FOXD1 mRNA stability, we manipulated genes of the YTH and IGF2BP families, which are well-established m6A “readers” that recognize methylated transcripts and regulate their stability.20 YTHDF1 and YTHDF2 knockdown both significantly increased FOXD1 mRNA level, while YTHDF3 knockdown significantly decreased FOXD1 mRNA (Figure 4D, Figure S5G). In contrast, knockdown of members of the IGF2BP family had no effect on FOXD1 mRNA level (Figure S5H). Furthermore, knockdown of YTHDF2 significantly increased mRNA stability and protein expression of FOXD1 (Figure 4E-F). However, knockdown of YTHDF1 had no effect on FOXD1 mRNA stability or protein level (Figure S5I-J). Furthermore, overexpression of METTL14 and knockdown of ALKBH5 can both rescue the effect of YTHDF2 knockdown on FOXD1 mRNA stability, which validates that the regulation of FOXD1 by YTHDF2 requires m6A modification (Figure 4G, Figure S5K). Collectively, these findings suggest that YTHDF2 recognizes m6A-methylated transcripts of FOXD1 and target them for degradation.
To further validate that YTHDF2 regulates m6A-methylated FOXD1, we constructed m6A recognition-deficient YTHDF2 mutants.21 We mutated tryptophan residues at positions 432, 486, and 491 within the YTH domain of YTHDF2 because these residues play crucial roles in m6A recognition across multiple vertebrate species.22,23 As expected, overexpression of the wild-type YTHDF2 reduced FOXD1 protein level, whereas overexpression of the triple-mutant YTHDF2 did not (Figure S5L, Figure 4H). We then performed additional validation experiments using the m6A-recognition-deficient triple-mutant YTHDF2 (YTHDF2-MUT), which contained W432A, W486A, and W491A mutations.
To further clarify whether YTHDF2 regulates FOXD1 level by mediating mRNA degradation, we performed RNA stability assays following YTHDF2-WT versus YTHDF2-MUT overexpression. Compared against mock control, YTHDF2-WT accelerated the mRNA degradation of FOXD1 with the control, while YTHDF2-MUT had no impact on FOXD1 mRNA degradation (Figure 4I). Following YTHDF2 knockdown, YTHDF2-WT overexpression rescued FOXD1 mRNA degradation, while YTHDF2-MUT overexpression failed to do so (Figure 4J). Additionally, we performed RNA-immunoprecipitation using anti-Flag antibodies after YTHDF2-WT-Flag versus YTHDF2-MUT-Flag overexpression. The results revealed that wild-type YTHDF2 could enrich FOXD1 mRNA, while the m6A-recognition-deficient YTHDF2-MUT did not (Figure 4K, Figure S5M). Additionally, knockdown of METTL14 significantly suppressed YTHDF2-mediated enrichment of FOXD1 mRNA (Figure S5N-O). Altogether, our results show that YTHDF2 directly recognizes m6A modifications on FOXD1 transcripts, thereby promoting their degradation and reducing the protein level of FOXD1.
FOXD1 Promotes the Formation of an Immunosuppressive Microenvironment in SHH-MB
To investigate the role of FOXD1 in SHH-MB, we knocked down FOXD1 in Daoy cells (Figure 5A) and performed RNA-seq. FOXD1 Knockdown SHH-MB led to the significant upregulation of 2423 genes and downregulation of 2363 genes (Figure 5B-C and Figure S6A). A total of 2181 genes are consistently differently expressed across independent shRNAs against FOXD1 (Figure 5D). Metascape pathway analysis of these genes revealed enrichment in pathways related to cytokine regulation of the immune system, interferon signaling, and cellular response to cytokine stimulus (Figure 5E). Seven of these genes—CXCL10/11, IFNL1, CCL5, ISG15, CCL20, and TNFAIP3—are involved in immune cell recruitment and activation (Figure 5C). We analyzed public SHH-MB RNA expression datasets and found that FOXD1 expression is negatively correlated with the expression of each of these 6 candidate targets (Figure S6B), consistent with our data.
Figure 5.
FOXD1 promotes the formation of an inhibitory IME in SHH-MB. (A) Protein expression of FOXD1 in shVector and shFOXD1-1/2/3 Daoy cells. (B) Heatmap of normalized gene expression of the shVector and shFOXD1-2/3 in Daoy cells. Blue and red indicate downregulated and upregulated gene expression, respectively. (C) Volcano diagram displaying upregulated and downregulated genes in shFOXD1-3 compared to shVector Daoy cells. Genes associated with inflammation are marked with text labels (D) Illustration of the overlapping transcripts between shFOXD1-2 and shFOXD1-3 Daoy cells. (E) Metascape enrichment analysis of 2181 genes from Figure 5D in the transcriptome of FOXD1-knockdown Daoy cells. (F) qRT-PCR and (G) ELISA analysis of CXCL10/11, IFNL1, TNFAIP3, CCL5, CCL20 and ISG15 secretion by Daoy cells after treatment with control vector, shFOXD1 or shFOXD1 + FOXD1 overexpression. (H, I) Relative expression levels of the 7 gene targets in Daoy cells after exposure to THP-1-derived M0/2-type macrophage exosomes with/without FOXD1 knockdown as detected by (H) qRT-PCR and (I) ELISA analyses. Data are presented as means ± SD from 3 independent experiments. *P < .05, **P < .01, ***P < .001, ****P < .0001, NS, not significant.
We then verified the regulation of the 6 candidate FOXD1 targets by knockdown, qRT-PCR, and rescue experiments. Knockdown of FOXD1 caused significant increases in the expression of the candidates at the RNA level, and this effect can be reversed by FOXD1 overexpression (Figure 5F). Similarly, ELISA analysis revealed that FOXD1 knockdown increased the secreted levels of the cytokine candidate targets by Daoy cells, and this effect can also be reversed by FOXD1 overexpression in cell (Figure 5G). These cytokines include CXCL10/11, IFNL1, and CCL5, and they are known to activate T cells.
To demonstrate that M2 macrophage-derived exosomes ultimately regulate the candidate FOXD1 targets, we repeated the qRT-PCR and ELISA experiments following exosome treatment of Daoy cells. As expected, M2-EXO treatment lowered the RNA level and secreted protein level of the candidate FOXD1 targets, and this exosome-mediated effect can be rescued by FOXD1 knockdown (Figure 5H-I).
These results collectively suggest that following TAM-derived exosome-induced upregulation of FOXD1 in SHH-MB tumor cells, FOXD1 acts downstream by inhibiting the release of chemotactic and inflammatory factors.
Knockdown of FOXD1 Enhances anti-PD-1 Immunotherapy in NeuroD2:SmoA1 SHH-MB Mice
We have previously shown that low CD8+ T cell infiltration into SHH-MB tumors is associated with poorer response to immunotherapy.24 To test whether the inhibition of FOXD1 can enhance the sensitivity of SHH-MB to immune checkpoint blockade, we injected rAAV2/9 adeno-associated virus encoding shFOXD1 in situ into the cerebellum of NeuroD2: SmoA1 mice, either alone or combined with anti-PD-1 inhibitor (Figure 6A). We then measured the transduction efficiency in the molecular, Purkinje cell, and granule cell layers, as well as the tumor mass within the cerebellum (Figure S7A).
Figure 6.
Knockdown of FOXD1 enhances the therapeutic efficacy of anti-PD-1 antibodies via CD8+T cell recruitment. (A) Schematic diagram of in vivo administration of rAAV2/9-shFOXD1 and PD-1-blocking antibodies, both alone and in combination (BIW, biweekly; n = 6 mice). (B) Representative MRI images obtained at day 60. (n = 6 for each group). (C) Tumor volume of mice during administration of antibodies. Tumor volume on Day 0 was set as 1, and tumor volume at each time point is shown after normalization to this value. (D) Kaplan–Meier survival fraction plot for the 4 groups (n = 6 for each group). (E) Representative IHC staining images of CD8a, CD163, CXCL10, and CXCL11 of each treatment group. (F, G) Quantification of CD8+T cells (CD8a) and M2 macrophages (CD163) in each 20X field. (H, I) Statistical analysis of IHC staining scores of (H) CXCL10 and (I) CXCL11. Data are presented as means ± SEM. Between-group differences were evaluated using one-way ANOVA analysis of variance in (C, F-I) or the log-rank (Mantel-Cox) test (D).
Magnetic resonance imaging (Figure 6B and Figure S7B) revealed that knockdown of FOXD1 significantly inhibited tumor growth. Combination therapy with FOXD1 and PD-1 blockade further reduced tumor volume, whereas anti-PD-1 treatment alone failed to control tumor growth (Figure 6C). Importantly, the combination of FOXD1 and PD-1 blockade significantly prolonged the survival of SHH-MB mice compared to PD-1 blockade alone (Figure 6D).
We then examined histological sections of the SHH-MB mouse tumors following rAAV2/9-shFOXD1 in situ injection. IHC analysis of the tumor-adjacent region showed that FOXD1 expression was significantly reduced following rAAV2/9-shFOXD1 treatment compared with the control (Figure S7C-D). Moreover, FOXD1 is highly expressed in the tumor region compared to the normal cerebellar granule layer (Figure S7C).
To corroborate our earlier in vitro results for FOXD1 knockdown, we performed IHC labeling of CD8+ T cells and M2 macrophages in the SHH-MB tumors from each treatment group. Comparing shFOXD1 treatments vs other groups, tumor infiltration of cytotoxic T cells increased while infiltration of M2 macrophages decreased (Figure 6E-G). To determine whether the recruitment of CD8+ T cells into SHH-MB tumors is related to chemokine release, we performed IHC staining for CXCL10 and CXCL11, which are classical T-cell chemokines. Indeed, compared to other groups, shFOXD1 treatments increased the expression of CXCL10 and CXCL11 in the tumors (Figure 6E, H-I).
In conclusion, our study provides both in vitro and in vivo results to demonstrate that the knockdown of FOXD1 can reshape the IME in SHH-MB, leading to the recruitment of peripheral CD8+ T cells into the tumor, thereby enhancing the effectiveness of the PD-1 inhibition.
Discussion
Previous studies have identified high TAM infiltration to be a characteristic feature of SHH-MB, suggesting that its tumor microenvironment is immunosuppressive.5,25,26 However, the mechanism by which TAMs promote tumor progression remains unclear. The present study demonstrates that TAMs secrete exosomes containing METTL14-targeting miRNAs, which are taken up by SHH-MB tumor cells, leading to downregulation of global m6A levels in the tumor cells, which, in turn, upregulates FOXD1 expression in an m6A-dependent manner. FOXD1 upregulation, in turn, contributes to the formation of an immunosuppressive microenvironment.
The dysregulation of m6A modification facilitates the tumor progression of SHH-MB.17 METTL3 is a key methyltransferase in the m6A “writer” complex, and its activity is implicated in many cancers, including MBs.16,17,27 We show here that METTL3 expression is elevated in SHH-MBs compared to the normal cerebellum, consistent with previous results.17 Aside from METTL3, the m6A “writer” complex also includes METTL14 and WTAP, and the integrity of the complex is determined in part by METTL14 protein stability.28–30 Therefore, in our study, the reduction in m6A modification levels depends on the decreased expression of METTL14. Our results in a large cohort of SHH-MB show that METTL14 expression is negatively correlated with TAM infiltration, which provided the first clues regarding how the immune microenvironment may regulate m6A methylation status of primary SHH-MB tumors.
Our results show that following exposure to M2 macrophage-derived exosomal miRNA, the overall m6A level in the tumor cells decreased. Our results here contrast with previous findings of hypermethylated m6A status in SHH-MBs.17 This discrepancy may be attributed to different levels of TAM infiltration, since we demonstrate that TAM-derived exosomes modulate METTL14 expression in tumor cells, thereby impacting their overall m6A level.
Additionally, we postulated that, in the context of SHH-MB, some molecules within TAMs may selectively regulate METTL14 expression without affecting METTL3. Supporting this idea, prior studies have suggested that different cells within the brain microenvironment may regulate the overall m6A modification level of target cells via exosomal exchange.31–33 Indeed, our experimental results in this study collectively confirm this hypothesis.
Methyltransferases deposit m6A in a specific subset of cellular transcripts, primarily near the stop codon and, less frequently, within the long exons.34 We demonstrate that m6A modification induced in SHH-MB cells by M2-EXO exposure occurred mainly in the 5′UTR and CDS of transcripts (Figure 3A). By joint RNA-seq and MeRIP-seq analysis, we identified FOXD1 as a key oncogene that is significantly m6A hypomethylated and upregulated after exposure to TAM-derived exosomes. Consistent with our results, a previous study also found a negative correlation between m6A modification and FOXD1 expression in the context of renal ischemia-reperfusion injury.35 Our study provides deeper insight into the molecular basis of m6A-dependent regulation of FOXD1 expression. Specifically, we found that m6A modification in the CDS reduced the mRNA stability of FOXD1 in SHH-MB, and this regulation is mediated by the m6A reader protein YTHDF2.
As a member of the Forkhead Box transcription factor (FOX) family, FOXD1 regulates a wide range of downstream targets, resulting in the coordinated regulation of various cancer hallmarks, including phenotypic plasticity, evasion of immune surveillance, and provision of proliferative signals.36 It has been shown that FOXD1 reshapes immune homeostasis by modulating the activity of inflammatory factors.37 Our functional enrichment analysis and in vitro assays suggest that FOXD1 in SHH-MB cells inhibits the recruitment of CD8+ T cells by suppressing the release of chemokines. (Figure 5). Our data highlights FOXD1 as a key player that reshapes the IME in SHH-MB.
Immune checkpoint inhibitors are widely used for cancer immunotherapy. However, the lower expression levels of PD-L1 in MB compared with other cancers suggest that PD-L1 checkpoint inhibitors may not be an optimal choice for MB treatment.6,7 Recent studies show that PD-1+ T cells infiltrate into the tumors of some subtypes of MB, which would suggest that PD-1 blockade may be a promising strategy.38 In the present study, we demonstrate that combinatorial treatment with a PD-1 inhibitor and AAV2/9-shFOXD1 effectively reduces SHH-MB tumor volume and significantly prolongs the survival of SHH-MB mice. Our results thus identify FOXD1 as a promising target for enhancing SHH-MB response to immunotherapy.
This study provides compelling in vitro and in vivo evidence to show that TAMs infiltration modulates overall m6A level in SHH-MB tumors, leading to increased FOXD1 levels, which then promote immunosuppression and tumor progression. Although the underlying molecular mechanisms would require even deeper investigation than what we present here, our results have identified a promising approach to reshape the immunosuppressive TME of SHH-MB via FOXD1 inhibition, thus paving a new avenue for future therapeutic development.
Supplementary Material
Acknowledgments
We appreciate XiuWu Bian’s lab for providing the ONS-76 cell line. We are also grateful to Juntao Bie and Yutong Li from Jianyuan Luo’s lab, as well as Chao Shi from Caihong Yun’s lab, for their assistance with the experiments. We thank Luzheng Xu from the Peking University Medical and Health Analysis Center for assistance with MRI analysis for SHH-MB mice. We extend our thanks to Yuan Wang from the State Key Laboratory of Natural and Biomimetic Drugs metabolomics platform for LC/MS-MS analysis. Additionally, we thank Panxi Sun from Shihezi University for help with the mice model experiments.
Contributor Information
Yantao Liu, Beijing Key Laboratory of Research and Transformation of Biomarkers for Neurodegenerative Diseases, Peking University Third Hospital, Peking University Health Science Center, Beijing, China; Department of Neuropathology, Beijing Neurosurgical Institute, Tiantan Hospital, Capital Medical University, Beijing, China; Department of Pathology, School of Basic Medical Sciences, Peking University Third Hospital, Peking University Health Science Center, Beijing, China.
Yu Peng, Beijing Key Laboratory of Research and Transformation of Biomarkers for Neurodegenerative Diseases, Peking University Third Hospital, Peking University Health Science Center, Beijing, China; Department of Neuropathology, Beijing Neurosurgical Institute, Tiantan Hospital, Capital Medical University, Beijing, China; Department of Pathology, School of Basic Medical Sciences, Peking University Third Hospital, Peking University Health Science Center, Beijing, China.
Chen Song, Department of Cancer Biology, Dana-Farber Cancer Institute, Boston, Massachusetts, USA; Department of Medical Genetics, Center for Medical Genetics, Peking University Health Science Center, Beijing, China.
Zongran Liu, CAS Key Laboratory for Biomedical Effects of Nanomaterials and Nanosafety, National Center for Nanoscience and Technology, Beijing, China.
Xiaolong Yang, CAS Key Laboratory for Biomedical Effects of Nanomaterials and Nanosafety, National Center for Nanoscience and Technology, Beijing, China.
Shuqing Bian, School of Information, Renmin University of China, Beijing, P.R. China.
Xiaolin Xiao, Department of Pathology, School of Basic Medical Sciences, Peking University Third Hospital, Peking University Health Science Center, Beijing, China.
Jing Wang, Department of Neuropathology, Beijing Neurosurgical Institute, Tiantan Hospital, Capital Medical University, Beijing, China.
Ziwen Sun, Beijing Key Laboratory of Research and Transformation of Biomarkers for Neurodegenerative Diseases, Peking University Third Hospital, Peking University Health Science Center, Beijing, China; Department of Neuropathology, Beijing Neurosurgical Institute, Tiantan Hospital, Capital Medical University, Beijing, China; Department of Pathology, School of Basic Medical Sciences, Peking University Third Hospital, Peking University Health Science Center, Beijing, China.
Xiaodan Liu, Beijing Key Laboratory of Research and Transformation of Biomarkers for Neurodegenerative Diseases, Peking University Third Hospital, Peking University Health Science Center, Beijing, China; Department of Pathology, School of Basic Medical Sciences, Peking University Third Hospital, Peking University Health Science Center, Beijing, China.
Bao Yang, Department of Neuro-surgery, Tiantan hospital, Capital University of Medical Science, Beijing, China.
David J H Shih, School of Biomedical Sciences, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong SAR, China.
Jianyuan Luo, Beijing Key Laboratory of Protein Posttranslational Modifications and Cell Function, Department of Biochemistry and Molecular Biology, Peking University Health Science Center, Beijing, China; Department of Medical Genetics, Center for Medical Genetics, Peking University Health Science Center, Beijing, China.
Hui Liang, Department of Pathology, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, China.
Qing Chang, Beijing Key Laboratory of Research and Transformation of Biomarkers for Neurodegenerative Diseases, Peking University Third Hospital, Peking University Health Science Center, Beijing, China; Department of Neuropathology, Beijing Neurosurgical Institute, Tiantan Hospital, Capital Medical University, Beijing, China.
Funding
This study was supported by the Beijing Natural Science Foundation (No.7232098, No.7192095), Beijing Health Foundation Commission (11000023T000002044300-4) and the National Natural Science Foundation of China (No.81972353, No.81101900, No.30540008) to Q.C.
Conflict of interest statement
The authors declare no conflict of interest.
Authorship statement
Y.L. designed and performed the research, analyzed and interpreted data, and wrote the manuscript. Z.L. performed TEM characterization of exosomes. S.B. conducted data processing, analysis, and visualization of m6A-seq, RNA-seq, and GEO dataset. X.Y. guided the design of in vivo animal experiments and performed statistical analysis. Y.P. made the graphical abstract of this article using Adobe Illustrator. X.L. classified molecular subtypes of medulloblastoma clinical samples. C.S. and X.X. contributed ideas and provided valuable discussions for this study. H.L., J.W., Y.P., and Z.S.. provided assistance in some experiments. B.Y. provided clinical samples and patients’ information. Q.C., H.L., and J.L. conceived and directed the project. Q.C. and D.S. revised the manuscript. All the authors discussed the results and commented on the manuscript.
Data availability
All data generated or analyzed during this study are included in this published article and its supplementary information files. Further information and requests for reagents should be directed to and will be fulfilled by the Lead Contact Qing Chang (changqing055@bjni.org.cn).
References
- 1. Riemondy KA, Venkataraman S, Willard N, et al. Neoplastic and immune single-cell transcriptomics define subgroup-specific intra-tumoral heterogeneity of childhood medulloblastoma. Neuro Oncol. 2022;24(2):273–286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Horbinski C, Berger T, Packer RJ, Wen PY. Clinical implications of the 2021 edition of the WHO classification of central nervous system tumours. Nat Rev Neurol. 2022;18(9):515–529. [DOI] [PubMed] [Google Scholar]
- 3. Luo Z, Xia M, Shi W, et al. Human fetal cerebellar cell atlas informs medulloblastoma origin and oncogenesis. Nature. 2022;612(7941):787–794. [DOI] [PubMed] [Google Scholar]
- 4. Chen S, Chen Z, Li Z, et al. Tumor-associated macrophages promote cholangiocarcinoma progression via exosomal Circ_0020256. Cell Death Dis. 2022;13(1):94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Zhu L, Yang Y, Li H, et al. Exosomal microRNAs induce tumor-associated macrophages via PPARgamma during tumor progression in SHH medulloblastoma. Cancer Lett. 2022;535(535):215630. [DOI] [PubMed] [Google Scholar]
- 6. Wang Q, Xin X, Dai Q, et al. Medulloblastoma targeted therapy: From signaling pathways heterogeneity and current treatment dilemma to the recent advances in development of therapeutic strategies. Pharmacol Ther. 2023;250(250):108527. [DOI] [PubMed] [Google Scholar]
- 7. Hashimoto Y, Penas-Prado M, Zhou S, et al. Rethinking medulloblastoma from a targeted therapeutics perspective. J Neurooncol. 2018;139(3):713–720. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Xu Z, Chen Y, Ma L, et al. Role of exosomal non-coding RNAs from tumor cells and tumor-associated macrophages in the tumor microenvironment. Mol Ther. 2022;30(10):3133–3154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Zhong W, Lu Y, Han X, et al. Upregulation of exosome secretion from tumor-associated macrophages plays a key role in the suppression of anti-tumor immunity. Cell Rep. 2023;42(10):113224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Northcott PA, Robinson GW, Kratz CP, et al. Medulloblastoma. Nat Rev Dis Primers. 2019;5(1):11. [DOI] [PubMed] [Google Scholar]
- 11. Hau P, Frappaz D, Hovey E, et al. Development of randomized trials in adults with medulloblastoma-the example of EORTC 1634-BTG/NOA-23. Cancers (Basel). 2021;13(14):3451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Li Y, Jiang T, Shao L, et al. Mir-449a, a potential diagnostic biomarker for WNT group of medulloblastoma. J Neurooncol. 2016;129(3):423–431. [DOI] [PubMed] [Google Scholar]
- 13. Li YX, Yu ZW, Jiang T, et al. SNCA, a novel biomarker for Group 4 medulloblastomas, can inhibit tumor invasion and induce apoptosis. Cancer Sci. 2018;109(4):1263–1275. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Zhao BS, Roundtree IA, He C. Post-transcriptional gene regulation by mRNA modifications. Nat Rev Mol Cell Biol. 2017;18(1):31–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Xu Z, Jiang J, Wang S. The critical role of RNA m(6)A methylation in gliomas: Targeting the hallmarks of cancer. Cell Mol Neurobiol. 2023;43(5):1697–1718. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Qiao K, Liu Y, Xu Z, et al. RNA m6A methylation promotes the formation of vasculogenic mimicry in hepatocellular carcinoma via Hippo pathway. Angiogenesis. 2021;24(1):83–96. [DOI] [PubMed] [Google Scholar]
- 17. Zhang ZW, Teng X, Zhao F, et al. METTL3 regulates m(6)A methylation of PTCH1 and GLI2 in Sonic hedgehog signaling to promote tumor progression in SHH-medulloblastoma. Cell Rep. 2022;41(4):111530. [DOI] [PubMed] [Google Scholar]
- 18. Zhu LY, Wu XY, Liu XD, et al. Aggressive medulloblastoma-derived exosomal miRNAs promote in vitro invasion and migration of tumor cells via Ras/MAPK Pathway. J Neuropathol Exp Neurol. 2020;79(7):734–745. [DOI] [PubMed] [Google Scholar]
- 19. Wang X, Zhao BS, Roundtree IA, et al. N(6)-methyladenosine modulates messenger RNA translation efficiency. Cell. 2015;161(6):1388–1399. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Boulias K, Greer EL. Biological roles of adenine methylation in RNA. Nat Rev Genet. 2023;24(3):143–160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Fang R, Chen X, Zhang S, et al. EGFR/SRC/ERK-stabilized YTHDF2 promotes cholesterol dysregulation and invasive growth of glioblastoma. Nat Commun. 2021;12(1):177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Zhu T, Roundtree IA, Wang P, et al. Crystal structure of the YTH domain of YTHDF2 reveals mechanism for recognition of N6-methyladenosine. Cell Res. 2014;24(12):1493–1496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Zhang C, Chen Y, Sun B, et al. m(6)A modulates haematopoietic stem and progenitor cell specification. Nature. 2017;549(7671):273–276. [DOI] [PubMed] [Google Scholar]
- 24. Chai X, Sun Z, Li H, et al. Characteristics of the CD8+ T cells infiltration and its mechanism in WNT subtype of meduloblastoma. J Peking Univ (Health Sciences). 2024;56(3):512–518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Tan IL, Arifa RDN, Rallapalli H, et al. CSF1R inhibition depletes tumor-associated macrophages and attenuates tumor progression in a mouse sonic Hedgehog-Medulloblastoma model. Oncogene. 2021;40(2):396–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Pokrajac NT, Tokarew NJA, Gurdita A, Ortin-Martinez A, Wallace VA. Meningeal macrophages inhibit chemokine signaling in pre-tumor cells to suppress mouse medulloblastoma initiation. Dev Cell. 2023;58(20):2015–2031.e8. [DOI] [PubMed] [Google Scholar]
- 27. Sun Y, Shen W, Hu S, et al. METTL3 promotes chemoresistance in small cell lung cancer by inducing mitophagy. J Exp Clin Cancer Res. 2023;42(1):65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Wang Y, Li Y, Toth JI, et al. N6-methyladenosine modification destabilizes developmental regulators in embryonic stem cells. Nat Cell Biol. 2014;16(2):191–198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Geula S, Moshitch-Moshkovitz S, Dominissini D, et al. Stem cells. m6A mRNA methylation facilitates resolution of naive pluripotency toward differentiation. Science. 2015;347(6225):1002–1006. [DOI] [PubMed] [Google Scholar]
- 30. Shan Y, Zhang Y, Wei Y, et al. METTL3/METTL14 maintain human nucleoli integrity by mediating SUV39H1/H2 degradation. Nat Commun. 2024;15(1):7186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Guo X, Qiu W, Wang C, et al. Neuronal activity promotes glioma progression by inducing proneural-to-mesenchymal transition in glioma stem cells. Cancer Res. 2024;84(3):372–387. [DOI] [PubMed] [Google Scholar]
- 32. Guo X, Qiu W, Li B, et al. Hypoxia-induced neuronal activity in glioma patients polarizes microglia by potentiating RNA m6A demethylation. Clin Cancer Res. 2023;30(6):1160–1174. [DOI] [PubMed] [Google Scholar]
- 33. Ding C, Yi X, Chen X, et al. Warburg effect-promoted exosomal circ_0072083 releasing up-regulates NANGO expression through multiple pathways and enhances temozolomide resistance in glioma. J Exp Clin Cancer Res. 2021;40(1):164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. He PC, He C. m(6) A RNA methylation: From mechanisms to therapeutic potential. EMBO J. 2021;40(3):e105977. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Meng F, Liu Y, Chen Q, et al. METTL3 contributes to renal ischemia-reperfusion injury by regulating Foxd1 methylation. Am J Physiol Renal Physiol. 2020;319(5):F839–F847. [DOI] [PubMed] [Google Scholar]
- 36. Cheng L, Yan H, Liu Y, Guan G, Cheng P. Dissecting multifunctional roles of forkhead box transcription factor D1 in cancers. Biochim Biophys Acta Rev Cancer. 2023;1878(6):188986. [DOI] [PubMed] [Google Scholar]
- 37. Peng SL. Interactions of Fox proteins with inflammatory transcription-factor pathways. Expert Rev Clin Immunol. 2006;2(6):869–876. [DOI] [PubMed] [Google Scholar]
- 38. Majzner RG, Simon JS, Grosso JF, et al. Assessment of programmed death-ligand 1 expression and tumor-associated immune cells in pediatric cancer tissues. Cancer. 2017;123(19):3807–3815. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data generated or analyzed during this study are included in this published article and its supplementary information files. Further information and requests for reagents should be directed to and will be fulfilled by the Lead Contact Qing Chang (changqing055@bjni.org.cn).