WDR82-Mediated H3K4me3 Is Associated with Tumor Proliferation and Therapeutic Efficacy in Pediatric High-Grade Gliomas

Simple Summary Pediatric high-grade gliomas bearing either an H3G34V/R or H3K27M mutation are incurable brain tumors with unique epigenomes. Current epigenomic studies on pHGGs primarily focus on these mutations; however, the function of other important crosstalk histone posttranslational modifications must be determined to elucidate tumor mechanisms. Here, we show that WDR82-mediated H3K4me3 is an important determinant of pediatric glioma malignancy and therapeutic response, and thus a potential epigenetic therapeutic target for children with malignant gliomas. Abstract Pediatric high-grade gliomas (pHGGs) are common malignant brain tumors without effective treatment and poor patient survival. Abnormal posttranslational modification at the histone H3 tail plays critical roles in tumor cell malignancy. We have previously shown that the trimethylation of lysine 4 at histone H3 (H3K4me3) plays a significant role in pediatric ependymoma malignancy and is associated with tumor therapeutic sensitivity. Here, we show that H3K4me3 and its methyltransferase WDR82 are elevated in pHGGs. A reduction in H3K4me3 by downregulating WDR82 decreases H3K4me3 promoter occupancy and the expression of genes associated with stem cell features, cell proliferation, the cell cycle, and DNA damage repair. A reduction in WDR82-mediated H3K4me3 increases the response of pediatric glioma cells to chemotherapy. These findings suggest that WDR82-mediated H3K4me3 is an important determinant of pediatric glioma malignancy and therapeutic response. This highlights the need for a more thorough understanding of the potential of WDR82 as an epigenetic target to increase therapeutic efficacy and improve the prognosis for children with malignant gliomas.


Introduction
Brain tumors, the most common form of solid tumor in children under the age of 15, are responsible for approximately 20% of all childhood cancers. Five-year survival, following diagnosis and treatment of a primary malignant brain tumor, is~30%. Among pediatric brain tumors, high-grade gliomas (pHGGs), including diffuse intrinsic pontine gliomas (DIPGs), are especially devastating, with an average patient survival of less than 15 months [1]. In the last decade, pHGGs have been extensively characterized and remarkable progress has been made in understanding the mechanisms associated with the tumors at the molecular level.
H3K4me3 levels play a role in determining the pathogenesis of various human cancers [12,13], including glial-derived brain tumors [14]. Additionally, H3K4me3 has prognostic utility for multiple cancers [12,15]. High levels of H3K4me3 are associated with the 5 regions of virtually all active genes, and a strong positive correlation exists between this modification, transcription rates, and active polymerase II occupancy, which is critical for transcriptional activity in a variety of eukaryotic species [16,17]. We have recently shown that high H3K4me3 is associated with poor survival in pediatric ependymoma [14]. Its role in pediatric gliomas has not been characterized.
A human SET domain containing 1A/B protein complexes (hSETD1A/B-COMPASS) binds to DNA containing unmethylated CpG motifs and is responsible for trimethylating H3K4 [18]. Human SETD1A/B proteins exhibit a large non-overlapping subnuclear distribution, indicating their distinct localization at subsets of target genes [19]. These complexes are identical, with the exception of the catalytic component [18,19]. Each complex contains seven units including SETD1A or SETD1B, ASH2, CXXC finger protein 1 (CFP1), DPY30, RBBP5, WDR5, and WDR82 [18,19]. WDR82 is a unique subunit of hSETD1A/B [20]. WDR82 trimethylates H3K4 through recruitment of hSETD1A/B. WDR82 interacts with hSETD1A/B via an N-terminal RNA recognition motif (RRM) within the latter to mark transcription start sites (TSS) of active genes [17,20,21]. During early embryonic development, WDR82 is crucial for H3K4me3 at the promoter of Oct4, whose expression is associated with embryonic stem cells [22]. WDR82 is also a component of the PTW/PP1 phosphatase complex, which is involved in the control of the chromatin structure and cellcycle progression during the transition from mitosis to interphase [23]. Elevated WDR82 and/or H3K4me3 is associated with therapeutic sensitivity in breast, cervical, and ovarian cancers as well as glioblastomas [24][25][26]. Recent studies show that WDR82 expression is associated with therapeutic sensitivity to platinum drugs [24,27]. However, its role in pediatric gliomas has not been investigated.
In this study, H3K4me3 levels and modifiers for this epigenetic mark were mapped using in-house pediatric glioma specimens and via in silico analysis of multiple publicly available databases. WDR82 and H3K4me3 were found to be elevated in pHGGs and associated with reduced chemotherapeutic sensitivity. A reduction in H3K4me3 by the downregulation of WDR82 decreased H3K4me3 promoter occupancy and gene expression and increased the response of pediatric glioma cells to chemotherapy. Our data suggest that H3K4me3 status is an important determinant of pediatric glioma malignancy and therapeutic response, and that WDR82, which regulates H3K4me3, is a promising target to increase therapeutic efficacy and improve the prognosis for children with malignant gliomas.

Cell Lines and Cultures
Pediatric SJ-GBM2 glioblastoma cells were obtained from the Children's Oncology Group Cell Culture and Xenograft Repository. Pediatric KNS42 cells were purchased from the JCRB (Japanese Cancer Research Resources) cell bank. The human H3K27M DIPG cell line SF8628 was generously provided by Dr. Rintaro Hashizume (Department of Pediatrics, Lurie Children's Hospital of Chicago). All cells were propagated as monolayers in complete medium consisting of Dulbecco's modified Eagle's medium (DMEM, Cat#11965-092) supplemented with 10% fetal bovine serum (FBS, Cat#10082147) from Gibco (Thermo Fisher Scientific, Waltham, MA, USA), at 37 • C with 5% CO 2 .

Plasmid Transduction
Tumor cells were plated a day before transduction, in 6-well plates at 30-40% confluency, to avoid cell clumps and uneven distribution that can reduce the efficacy of viral transduction. Individual wells of confluent cells were treated with packaged lentivirus (100-500 µL) and polybrene 4-8 µg/mL, followed by incubation for 6-14 h. Cell viability and protein expression were monitored. The transduction process was repeated until an optimal amount of GFP protein was expressed. GFP protein expression was checked after 48-72 h using a fluorescent microscope to visualize the fluorescent tag present in the construct. Cells were then selected with antibiotics. In cases where the construct needed a substrate such as doxycycline (Dox) for induction, Dox was added and protein expression was monitored after 24 h, 48 h, and 72 h.

In Silico Public Dataset Analysis
Two expression profiling datasets (GSE50161, GSE73038, GSE68015, and GSE36245) were downloaded from the GEO database at the NCBI. Original gene expression profiles of glioma were obtained from these datasets. Expression of individual genes was identified with GEO2R. Adult low-and high-grade glioma data in TCGA were processed with the Gliovis online portal (http://gliovis.bioinfo.cnio.es/, accessed on 15 March 2022). Data were further analyzed with GraphPad Prism 9.0 for gene expression, gene expression correlation, and survival analysis (GraphPad Software, Inc. La Jolla, CA, USA).

RNA-seq
Total RNA was prepared using the RNeasy Mini Kit (Qiagen, Germantown, MD, USA. Cat#74106) as per manufacturer's instructions. Stranded mRNA-seq was conducted in the Northwestern University Sequence (NUseq) Core Facility. Briefly, total RNA samples were checked for quality using RINs generated from Agilent Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA). RNA quantity was determined with the Qubit fluorometer (Qubit 2.0, Thermofisher, Waltham, MA, USA). The Illumina TruSeq Stranded mRNA Library Preparation Kit (Cat#20020595, Illumina, Inc., San Diego, CA, USA) was used to prepare sequencing libraries from 1 mg of high-quality RNA samples (RIN = 10). The Kit procedure was performed without modification. The procedure includes mRNA purification and fragmentation, cDNA synthesis, 3 end adenylation, Illumina adapter ligation, and library PCR amplification and validation. Illumina HiSeq 4000 sequencer (San Diego, CA, USA) was used to sequence the libraries with the production of singleend, 50 bp reads at the depth of 30-40 M reads per sample. The quality of reads, in FASTQ format, was evaluated using FastQC. Adapters were trimmed and reads of poor quality or aligning to rRNA sequences were filtered. The cleaned reads were aligned to the H. sapiens genome using STAR [28]. Read counts for each gene were calculated using HTseq-count, in conjunction with a gene annotation file for hg38 obtained from UCSC (http://genome.ucsc.edu, accessed on 14 March 2021). Differential expression was determined using DESeq2. The cutoff for determining genes, which showed significant differential expression, was an FDR-adjusted p-value less than 0.05.

ChIP-seq
Fragmental DNA for ChIP-seq was prepared using SimpleChIP ® Plus Enzymatic Chromatin IP Kit (Magnetic Beads) (Cell Signaling Technologies, Danvers, MA, USA, Cat#9005) per manufacturer's instructions. ChIP-seq was conducted in the NUseq Core Facility. Briefly, ChIP and input DNA samples were checked for quality and quantity using Illumina adapter ligation, and library PCR amplification and validation. Illumina HiSeq 4000 sequencer was used to sequence the libraries with the production of single-end, 50 bp reads at the depth of 30-40 M reads per sample. After sequencing, the quality of reads, in FASTQ format, will be evaluated using FastQC. Adapters will be trimmed; reads of poor quality will be filtered. The cleaned reads will be aligned to the H. sapiens genome (hg38) using Bowtie [29]. Peak calling and differential peak analysis will be performed using HOMER (http://homer.ucsd.edu/homer/index.html, accessed on 26 January 2021).

Cell Proliferation Assay
SJ-GBM2 and SF8628 cells were used to determine if reducing WDR82 through inducible knockdown affected cell viability; 1 × 10 4 cells/100µL were plated in 96-well plates with complete cell culture medium with or without Dox (2 µg/mL), and subjected to 3-(4, 5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS, Promega, Madison, WI, USA) assay. Pediatric high-grade glioma KNS42 cells were seeded into 6-well tissue culture plates and allowed to adhere. Attached cells were treated with or without Dox (SKU#D9891-10G, Sigma-Aldrich) at 2 µg/mL. Cells were incubated with Dox for 2 weeks, at which time, colonies were counted following staining with methylene blue (0.66% solution in 95% ethanol). Plating efficiencies were calculated as the ratio of the number of colonies formed to the number of cells seeded. The cell lines were used for different assays due to their distinct growth patterns.

In Vivo Studies
Briefly, 6-8-week-old athymic nude mice were purchased from Taconic. All mice were housed under aseptic conditions, which included filtered air and sterilized food, water, bedding, and cages. The Institutional Animal Care and Use Committee (IACUC) approved all animal protocols. SJ-GBM2 (1 × 10 5 ) or KNS42 (2 × 10 5 ) cells, transduced with sh-WDR82 vectors, or non-transduced controls, were implanted into the right striatum. Mice were given 1% sucrose or 1% sucrose + 2 mg/mL Dox in their drinking water [30], which was replaced every 2-3 days. Animals were monitored daily and body weight was measured every 2-3 days. Mice were euthanized with CO 2 asphyxiation followed by cervical dislocation when they became moribund (e.g., >20% weight loss, neurologic symptoms, or evidence of pain/distress). Brains were harvested and fixed with 10% paraformaldehyde in PBS overnight and switched to PBS prior to embedding. The tissue was sectioned onto slides and stained with H&E. Tissue preparation was conducted at the Mouse Histology and Phenotyping Laboratory (MHPL), Northwestern University Feinberg School of Medicine.

Statistical Analysis
Cell proliferation results were read on a Synergy 2 Microplate Reader (BioTek Instruments Inc., Winooski, VT, USA). Cell survival is presented as a percentage of viable cells compared to the viable cell number in the corresponding control, with the control set as 100%. P values were calculated using two-way ANOVA, with p < 0.05 considered significant. Statistical tests were 2-sided. Kaplan-Meier and t-tests were performed to compare survival between groups. Graph generation and statistical analyses were performed with GraphPad Prism 9 software (GraphPad Software, Inc. La Jolla, CA, USA).

H3K4me3 Levels Increase with Histopathological Malignancy in Pediatric Gliomas
To investigate whether a correlation exists between clinicopathologic variables and H3K4me3, forty-six pediatric low-grade gliomas (WHO grades I and II) and ten pediatric high-grade gliomas (WHO grades III and IV) FFPE specimens were IHC-stained for H3K4me3 (Supplementary Table S1). H3K4me3 was predominantly observed in cell nuclei with the frequency of immunopositive cells ranging from 0% to 100%. Patient gender and tumor location are not associated with H3K4me3 levels. Patient age and WHO grade positively correlated with H3K4me3 levels. IHC results showed H3K4me3 significantly increased in pHGGs, relative to pLGGs ( Figure 1A,B). Multivariate Cox proportional hazards analysis showed that progression-free survival of patients with high levels of H3K4me3 (IHC score ≥ 3) was significantly shorter relative to patients with low levels of H3K4me3 (IHC score ≤ 2) ( Figure 1C). Western blots of protein extracts from tissue samples confirmed higher levels of H3K4me3 in pHGGs vs. pLGGs ( Figure 1D). These results indicate that H3K4me3 IHC staining is a potential predictor of pediatric glioma malignancy, especially in pHGGs, irrespective of the presence or absence of histone H3 mutations.
To investigate whether a correlation exists between clinicopathologic variables and H3K4me3, forty-six pediatric low-grade gliomas (WHO grades I and II) and ten pediatric high-grade gliomas (WHO grades III and IV) FFPE specimens were IHC-stained for H3K4me3 (Supplementary Table S1). H3K4me3 was predominantly observed in cell nuclei with the frequency of immunopositive cells ranging from 0% to 100%. Patient gender and tumor location are not associated with H3K4me3 levels. Patient age and WHO grade positively correlated with H3K4me3 levels. IHC results showed H3K4me3 significantly increased in pHGGs, relative to pLGGs ( Figure 1A,B). Multivariate Cox proportional hazards analysis showed that progression-free survival of patients with high levels of H3K4me3 (IHC score ≥ 3) was significantly shorter relative to patients with low levels of H3K4me3 (IHC score ≤ 2) ( Figure 1C). Western blots of protein extracts from tissue samples confirmed higher levels of H3K4me3 in pHGGs vs. pLGGs ( Figure 1D). These results indicate that H3K4me3 IHC staining is a potential predictor of pediatric glioma malignancy, especially in pHGGs, irrespective of the presence or absence of histone H3 mutations.

WDR82 Plays a Distinct Role in the Tumor Biology of Pediatric Glioma
H3K4 is reversibly modified by methyltransferases and demethylases. To identify the key modifier(s) of H3K4 in pHGGs, gene expression profiling with microarray analysis

WDR82 Plays a Distinct Role in the Tumor Biology of Pediatric Glioma
H3K4 is reversibly modified by methyltransferases and demethylases. To identify the key modifier(s) of H3K4 in pHGGs, gene expression profiling with microarray analysis revealed that WDR82, among all H3K4 modifiers, is differentially expressed at significant levels and correlates with WHO-grade malignancy, regardless of histone mutation subtypes (Figure 2A,B). These results are supported by in silico analysis of a glioma gene expression profiling dataset (GSE50161, Supplementary Figure S1A). Geneontology function analysis indicates that WDR82 expression is related to multiple biological processes, molecular functions, and cellular signaling pathways ( Figure 2C). The level of WDR82 shows an inverse correlation with patient survival ( Figure 2D). Interestingly, in silico analysis results from The Cancer Genome Atlas (TCGA) show that WDR82 does not correlate with WHOgrade malignancy and survival in adult gliomas ( Figure 2E,F). In combination, these results indicate that WDR82 expression is specifically relevant to the tumor biology of pHGGs.
Altered gene expression following WDR82 inducible knockdown. We have shown that WDR82 is highly expressed in pHGGs vs. pLGGs. To investigate the role of WDR82 in pHGGs, pHGG cell lines SJ-GBM2, SF8628, and KNS42 were transduced with SMARTvector inducible lentiviral shRNA targeting WDR82 (shRNA#1 and shRNA#2) or non-targeting control vector (Supplementary Figure S2A). The results show that GFP is expressed (green  Figure S2B). We further checked WDR82 expression in these cells by real-time PCR (Supplementary Figure S2C) and Western blot (Supplementary Figure S2D). WDR82 level significantly decreased.
revealed that WDR82, among all H3K4 modifiers, is differentially expressed at significant levels and correlates with WHO-grade malignancy, regardless of histone mutation subtypes (Figure 2A,B). These results are supported by in silico analysis of a glioma gene expression profiling dataset (GSE50161, Supplementary Figure S1A). Geneontology function analysis indicates that WDR82 expression is related to multiple biological processes, molecular functions, and cellular signaling pathways ( Figure 2C). The level of WDR82 shows an inverse correlation with patient survival ( Figure 2D). Interestingly, in silico analysis results from The Cancer Genome Atlas (TCGA) show that WDR82 does not correlate with WHO-grade malignancy and survival in adult gliomas ( Figure 2E,F). In combination, these results indicate that WDR82 expression is specifically relevant to the tumor biology of pHGGs. To identify altered gene expression, with WDR82 knockdown, total RNA was extracted from transduced cells in the absence or presence of Dox (2 µg/mL) at day 5 and sequenced (RNA-seq and ChIP-seq with H3K4me3 antibody immunoprecipitation). The RNA-seq results from SJ-GBM2 cells indicate that 129 genes were differentially (cutoff: 1.5 times) expressed in cells transduced with shRNA#1 and shRNA#2, in comparison to non-significant gene expression changes with non-target shRNA (shNT) controls in the presence or absence of Dox ( Figure 3A-C). We further analyzed the function of these genes using the Gene Ontology online portal (http://geneontology.org/, accessed on 15 March 2021). The results showed that reducing WDR82 significantly decreases the expression of genes associated with mitosis, proliferation, and DNA repair ( Figure 3D). ChIP-seq results indicate that LIN9 and NUP43 promoter H3K4me3 levels decreased in SJ-GBM2 cells transduced with inducible shRNA#2 against WDR82 following Dox induction ( Figure 3E).

WDR82-Mediated H3K4me3 Is Associated with Mitotic and Cell-Cycle-Related Gene Regulation
To investigate if changes in gene expression link to WDR82-mediated H3K4me3 alteration at their promoters, representative genes were selected. The correlation between gene expression and WDR82 was plotted using RNA-seq results from pediatric glioma

WDR82-Mediated H3K4me3 Is Associated with Mitotic and Cell-Cycle-Related Gene Regulation
To investigate if changes in gene expression link to WDR82-mediated H3K4me3 alteration at their promoters, representative genes were selected. The correlation between gene expression and WDR82 was plotted using RNA-seq results from pediatric glioma tissue specimens. Promoter H3K4me3 levels were mapped using ChIP-seq results SJ-GBM2 cells treated with the Dox inducible shRNA vector against WDR82 (shRNA#2), in the absence and presence of Dox. The representative results showed WDR82 did not correlate with the expression of NUP62 and FBOX10, genes for which levels increased in pediatric glioma tissue specimens ( Figure 4A). Additionally, H3K4me3 at their promoters was unaltered ( Figure 4B). In the group of genes with decreased expression, there was a correlation with WDR82 in pediatric glioma tissue specimens, and representative results are shown for NUP43 and PKYMT1 ( Figure 4C). Furthermore, H3K4me3 promoter levels were undetectable in SJ-GBM2 cells treated with inducible shRNA#2 against WDR82, following Dox or non-Dox treatment ( Figure 4D). was unaltered ( Figure 4B). In the group of genes with decreased expression, there was a correlation with WDR82 in pediatric glioma tissue specimens, and representative results are shown for NUP43 and PKYMT1 ( Figure 4C). Furthermore, H3K4me3 promoter levels were undetectable in SJ-GBM2 cells treated with inducible shRNA#2 against WDR82, following Dox or non-Dox treatment ( Figure 4D).

WDR82 Knockdown Decreases In Vitro Cell Viability and DNA Damage Repair in pHGG Cells
To investigate whether WDR82 levels affect pHGG cell growth, equal numbers of cells were plated in 96-well plates in the absence and presence of Dox (2 µg/mL), with relative cell numbers compared at 0~7 days by MTS assay. OD values for Dox-treated groups were normalized based on the values of corresponding untreated cells. Inducible knockdown of WDR82, following Dox treatment, significantly suppresses cell growth in SJ-GBM2 and SF8628 cells ( Figure 5A). Inducible knockdown of WDR82 also affects pHGG KNS42 cell proliferation, as indicated by colony formation assay ( Figure 5B,C). KNS42 cells were plated (125, 250, and 500 cells) in 6-well plates and cultured in the absence or presence of Dox (2 µg/mL) for two weeks prior to counting colonies stained with methylene blue (0.66% solution in 95% ethanol).
CDK2 is associated with CCND1, which regulates cell proliferation, the cell cycle, and DNA damage repair in childhood cancer through the regulation of Rb [31]. CDK2 decreased in SJ-GBM2 cells treated with inducible shRNA#2 against WDR82, following

WDR82 Knockdown Decreases In Vitro Cell Viability and DNA Damage Repair in pHGG Cells
To investigate whether WDR82 levels affect pHGG cell growth, equal numbers of cells were plated in 96-well plates in the absence and presence of Dox (2 µg/mL), with relative cell numbers compared at 0~7 days by MTS assay. OD values for Dox-treated groups were normalized based on the values of corresponding untreated cells. Inducible knockdown of WDR82, following Dox treatment, significantly suppresses cell growth in SJ-GBM2 and SF8628 cells ( Figure 5A). Inducible knockdown of WDR82 also affects pHGG KNS42 cell proliferation, as indicated by colony formation assay ( Figure 5B,C). KNS42 cells were plated (125, 250, and 500 cells) in 6-well plates and cultured in the absence or presence of Dox (2 µg/mL) for two weeks prior to counting colonies stained with methylene blue (0.66% solution in 95% ethanol). inducible shRNA#2 against WDR82, following Dox vs. non-Dox treatment ( Figure 5E). These results indicate that WDR82-mediated H3K4me3 alteration is associated with cell proliferation, the cell cycle, and DNA damage repair ( Figure 5F).

WDR82 Knockdown Increases Therapeutic Sensitivity against DNA-Damaging Agents and Radiation Sensitivity
Increased WDR82 expression is associated with tumor malignancy grade (Figure 2A). Recent studies have shown that the overexpression of WDR82 promotes high H3K4me3 in tumors following chemotherapy [24,27]. Our RNA-seq results from SJ-GBM2 cells indicate that reduction of WDR82 impairs DNA repair and increases apoptosis. We hypothesized that reduction of WDR82 could increase sensitivity to chemotherapeutic agents that break double-stranded DNA, and to radiation. To verify this hypothesis, the effect of inducible WDR82 knockdown was investigated in SJ-GBM2 and SF8628 cells treated with cisplatin (CDDP), a first-line clinical drug for treating pHGG patients. Inducible WDR82 knockdown by Dox increased sensitivity to CDDP (Supplementary Figure S3A). We also tested whether inducible WDR82 knockdown by Dox increases cell response to radiation following a clonogenic assay protocol as described [32]. WDR82 reduction increased radiation sensitivity in KNS42 cells, as indicated by a comparison of colony numbers (Supplementary Figure S3B) and calculating the dose of enhancement factors (DEFs) at 10% survival (Supplementary Figure S3C). CDK2 is associated with CCND1, which regulates cell proliferation, the cell cycle, and DNA damage repair in childhood cancer through the regulation of Rb [31]. CDK2 decreased in SJ-GBM2 cells treated with inducible shRNA#2 against WDR82, following Dox vs. non-Dox treatment ( Figure 3C). We hypothesized that CDK2 and CCND1 are regulated by the WDR82 mediation of promoter H3K4me3. To verify this hypothesis, CDK2, CCND1, and WDR82 expressions were examined in pediatric gliomas. A significant correlation was identified ( Figure 5D). ChIP-seq results showed that at the promoters of these two genes, H3K4me3 decreased in SJ-GBM2 cells, treated with inducible shRNA#2 against WDR82, following Dox vs. non-Dox treatment ( Figure 5E). These results indicate that WDR82-mediated H3K4me3 alteration is associated with cell proliferation, the cell cycle, and DNA damage repair ( Figure 5F).

WDR82 Knockdown Increases Therapeutic Sensitivity against DNA-Damaging Agents and Radiation Sensitivity
Increased WDR82 expression is associated with tumor malignancy grade (Figure 2A). Recent studies have shown that the overexpression of WDR82 promotes high H3K4me3 in tumors following chemotherapy [24,27]. Our RNA-seq results from SJ-GBM2 cells indicate that reduction of WDR82 impairs DNA repair and increases apoptosis. We hypothesized that reduction of WDR82 could increase sensitivity to chemotherapeutic agents that break double-stranded DNA, and to radiation. To verify this hypothesis, the effect of inducible WDR82 knockdown was investigated in SJ-GBM2 and SF8628 cells treated with cisplatin (CDDP), a first-line clinical drug for treating pHGG patients. Inducible WDR82 knockdown by Dox increased sensitivity to CDDP (Supplementary Figure S3A). We also tested whether inducible WDR82 knockdown by Dox increases cell response to radiation following a clonogenic assay protocol as described [32]. WDR82 reduction increased radiation sensitivity in KNS42 cells, as indicated by a comparison of colony numbers (Supplementary Figure S3B) and calculating the dose of enhancement factors (DEFs) at 10% survival (Supplementary Figure S3C).

WDR82 Knockdown Decreases In Vivo pHGG Tumor Growth and Extends Survival of pHGG-Tumor-Bearing Mice
WDR82 contributes to tumorigenesis, malignant phenotype, and tumor proliferation of multiple human cancers [33,34]. To investigate if WDR82 and H3K4me3 correlate to proliferation in pediatric glioma specimens, PCNA, MIK67, WDR82, and H3K4me3 expressions were examined. WDR82 correlated with PCNA and MIK67 ( Figure 6A) and Ki67 was associated with H3K4me3 ( Figure 6B). To investigate if WDR82 knockdown decreases pHGG tumor growth in vivo, 1 × 10 5 cells (SJ-GBM2 transduced with shNT and WDR82 shRNA#2) were inoculated intracranially into the right striatum of 6-8-week-old athymic nude mice. Mouse housing and tissue processing as described under in vivo studies. Tumor size and mitotic cells are not significantly different in non-targeted shRNA (shNT) groups with or without Dox. In contrast, tumor size is smaller and there are fewer mitotic cells in the group treated with Dox, as indicated by comparing results from animals given Dox vs. 1% sucrose, (Figure 6C-E). The results also showed that survival was significantly extended in mice in the WDR82 shRNA#2 + Dox group ( Figure 6F). These results indicate that suppression of WDR82 decreases cell growth and extends the survival of tumor-bearing animals.
(shNT) groups with or without Dox. In contrast, tumor size is smaller and there are fewer mitotic cells in the group treated with Dox, as indicated by comparing results from animals given Dox vs. 1% sucrose, (Figure 6C-E). The results also showed that survival was significantly extended in mice in the WDR82 shRNA#2 + Dox group ( Figure 6F). These results indicate that suppression of WDR82 decreases cell growth and extends the survival of tumor-bearing animals.

Discussion
In a previous study, we established that H3K4me3 is associated with WHO-grade malignancy in pediatric ependymomas [14]. In this work, we found that H3K4me3 is also associated with WHO grade in pediatric gliomas, with higher levels of H3K4me3 in HGGs vs. LGGs ( Figure 1A). Patients with higher tumor levels of H3K4me3 had a shorter progression-free survival ( Figure 1C). Furthermore, WDR82, an H3K4me3 methyltransferase, is elevated in pHGGs vs.
LGGs, which is inversely correlated with clinical prognosis (Figure 2). Furthermore, we also found that WDR82-meditated H3K4me3 alters gene expression related to a variety of biological functions including stem cell features, cell proliferation, the cell cycle, and DNA damage repair. The results highlight the necessity of a thorough understanding of the biological roles of WDR82 and H3K4me3 in pediatric glioma. This knowledge will establish if WDR82 and H3K4me3 are potential epigenetic targets to suppress tumor progression, increase therapeutic efficacy, and improve outcomes for children with malignant gliomas.
The discovery of recurrent histone mutations in children with pHGGs in the last decade revealed that PTMs on histone tails are one of the milestone changes whose role must be elucidated in cancer research. In pediatric gliomas, H3G34V/R in hemisphere tumors and H3K27M in midline gliomas including DIPG are common histone mutations [2,3]. These mutations crosstalk with other histone modifications and bring about abnormal gene expression involved in pediatric malignant glioma tumorigenesis. Recently, H3K36me2 and H4K16Ac were discovered to be novel epigenetic signatures of DIPG [35]. H3K4me3, an important histone mark for developmental neurogenesis, was relatively unaffected regardless of these mutations as shown by Western blot analysis [8][9][10]. However, ChIP-seq analysis to investigate promoter H3K27me3 and H3K4me3 in wild-type (WT) and H3K27M NSCs and DIPGs showed elevated promoter H3K4me3 in H3K27M-induced genes such as Lin28b, Igf2bp2, Plag1, Pbx3, Eya1, etc. These genes regulate neuroprogenitor cell proliferation and differentiation and are associated with the DIPG oncogenic signature [36]. Another study used ChIP-seq to map promoter H3K27me3 and H3K4me3 in DIPGs with and without H3K27M knockdown, promoter H3K4me3 changed in differentially expressed genes, with increased levels in K27M-induced upregulated genes, for example, RORB and VIM [37]. H3K4me3 is activated in pHGGs with an H3.3G34R/V mutation in comparison with WT tumors [38]. Pediatric diffuse midline gliomas with an H3K27M mutation have lower levels of methionine adenosyltransferase 2A (MAT2A) protein; depletion of residual MAT2A reduces global H3K4me3 [39]. In this study, we determined that H3K4me3 is elevated in pHGGs vs. pLGGs (Figure 1 and Supplementary Table S1), and its reduction via decreasing WDR82 alters the expression of genes involved in cell proliferation, the cell cycle, and DNA damage repair. Altogether, these findings suggest H3K4me3 impacts pHGGs, which merits further investigation.
Among the six methyltransferases for H3K4me3, SETD1A is associated with WHO malignancy in pediatric gliomas ( Figure 1D and Figure S1A). Investigation of its subunits WDR82, ASH2L DPY30 CXXC1, RBBP5, and WDR5 showed that WDR82 and WDR5 were significantly associated with WHO-grade malignancy in pediatric gliomas ( Figure 1C, Figure 2 and Figure S1C). WDR5 is also associated with malignancy in adult gliomas [40]. WDR82 is relatively high in glioma vs. other cancers (Supplementary Figure S4), which is the sole subunit in SETD1A-COMPASS, specifically associated with pediatric glioma WHO-grade malignancy ( Figure 2A) and patient survival ( Figure 2D), regardless of histone mutation status. It is not correlated with adult glioma WHO-grade malignancy or patient survival ( Figure 2E,F). Due to this specificity for pediatric gliomas, WDR82-mediated H3K4me3 is the focus of this study.
WDR82 activity is a key factor in maintaining embryonic [34] and cancer stem cells [41][42][43]. In this work, sphere formation was impaired in shWDR82-transduced cells following Dox treatment, in comparison to non-Dox treatment and shNT groups, respectively (Supplementary Figure S5A,B). Results from in silico analysis of dataset GSE50161 showed a positive correlation between WDR82 and the expression of SOX2, (Supplementary Figure S5C), a glioma stem cell determinant [41], and one of the genes decreased in Figure 2B. ChIP-seq results from SJ-GBM2 cells show that WDR82 knockdown causes promoter H3K4me3 reduction at the SOX2 gene (Supplementary Figure S5D). These results are consistent with previous findings [22,42] and provide further confirmation that WDR82-mediated H3K4me3 is associated with stem cell characteristics.
In the present study, WDR82-mediated H3K4me3 alters the gene expression of NUP43 [43], LIN9 [44], PKMYT1 [45], CDK2 [46], LOX [47], and NUP62 [48], genes associated with mitosis, the cell cycle, and DNA damage repair in pHGGs (Figures 3-5). WDR82 association with these genes is a novel finding; however, the results are consistent with discoveries in embryonic stem cells and various human cancer cells. For instance, WDR82-mediated H3K4me3 is responsible for facilitating M-phase progression in mixed-lineage leukemia. WDR82 knockdown increases mitotic cells [49]. WDR82 modulates cell cycle progression through the regulation of the B-cell translocation gene 2 (BTG2). Depletion of WDR82 induces the expression of BTG2, an anti-proliferative protein [50]. WDR82-mediated H3K4me3 reduction induces apoptosis in embryo stem cells [22]. WDR82 and/or H3K4me3 are associated with chemotherapeutic sensitivity in breast, cervical, and ovarian cancers, and adult glioblastoma cells [24][25][26]. WDR82 is an important binding partner with TOX4 in HeLa cells following cisplatin treatment [24,27]. Our results, showing that inducible reduction of WDR82 sensitizes the pediatric glioma cell response to cisplatin and radiation therapy (Supplementary Figure S3), are also in line with alterations in gene expression observed in this work. Taken together these findings demonstrate that WDR82 regulates the gene expression associated with mitosis, proliferation, and the cell cycle, as well as apoptosis and chemotherapeutic response in pHGGs, and thus may be a potential therapeutic target.

Conclusions
In summary, H3K4me3 and WDR82 are elevated in pHGGs vs. pLGGs. WDR82mediated H3K4me3 is associated with the expression of genes involved in regulating stem cell features, cell mitosis and proliferation, the cell cycle, and DNA damage repair. Reduction of WDR82 increased the response of pediatric glioma cells to chemotherapy. Inducible reduction of WDR82 decreases pHGG tumor cell growth in vivo and extends animal survival. These findings suggest that WDR82-mediated H3K4me3 is a significant factor in pediatric glioma, and further investigation of WDR82 as a promising epigenetic therapeutic target for pHGG is warranted.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cancers15133429/s1, Figure S1: In silico analysis of expression of subunits of human SET-COMPASS complexes in pediatric gliomas. Figure S2: Establishment and validation of a WDR82-expression-inducible knockdown system. Figure S3: WDR82 knockdown increases chemotherapeutic and radiation sensitivity. Figure S4: WDR82 expression in human cancers presented with TCGA dataset. Figure S5: WDR82 is associated with stem cell characteristics. Table  S1: Clinicopathological information for pediatric primary gliomas IHC stained for H3K4me3.
Author Contributions: Conception and design: G.X. and T.T. Development of methodology: B.M.-F., G.X. and C.D.J. Acquisition of data (clinic database and pathology reviewing, immunofluorescence, real-time PCR, Western blot and flow cytometry, radiation therapy, etc.): N.W., S.N., Y.W., G.X. and T.T. Analysis and interpretation of data (e.g., statistical analysis, biostatistics, and computational analysis): C.J., G.X. and T.T. Writing-review, and/or revision of the manuscript: B.M.-F., G.X., C.J. and T.T. Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R.H., G.X. and T.T. Study supervision: C.D.J. and T.T. All authors have read and agreed to the published version of the manuscript. Informed Consent Statement: Patient consent was waived due to no inclusion of individual patient's identifying information.

Data Availability Statement:
The RNA-and ChIP-sequence data are available upon request.