Abstract

   Cancer stem cells (CSCs) are significant factors in the treatment
   resistance and recurrence of glioblastoma. Chromobox protein homolog 7
   (CBX7) can inhibit the progression of various tumors, but its impact on
   the stem cell-like properties of glioblastoma cells remains unclear.
   Clinically, low levels of CBX7 are associated with poor prognosis and
   increased distant metastasis in glioblastoma patients, and this low
   expression is caused by methylation of the CBX7 promoter. Our current
   research indicates that CBX7 plays a key role in suppressing the
   stem-like phenotype of glioblastoma. In this study, through
   bioinformatics analysis, we found that CBX7 is the most significantly
   downregulated member of the CBX family in glioblastoma and is closely
   associated with the stem-like phenotype of glioblastoma cells. We show
   that CBX7 promotes the degradation of myosin heavy chain 9 (MYH9)
   protein through the ubiquitin-proteasome pathway via the polycomb
   repressive complex 1 (PRC1) and suppresses the stem-like phenotype of
   glioblastoma cells by inhibiting the nuclear factor kappa-B (NF-κB)
   signaling pathway. Furthermore, overexpression of MYH9 in glioblastoma
   cells reverses the inhibitory effects of CBX7 on migration,
   proliferation, invasion, and stemness of glioblastoma cells. In
   summary, CBX7 acts as a tumor suppressor by inhibiting the stem
   cell-like characteristics of glioblastoma. The CBX7-MYH9-NF-κB
   signaling axis may serve as a potential therapeutic target for
   glioblastoma.

   Subject terms: CNS cancer, Neural stem cells

Introduction

   Glioblastoma is a highly aggressive and lethal brain tumor with poor
   prognosis and limited treatment options. Central to the resilience and
   recurrence of glioblastoma are glioblastoma stem cells (GSCs), a
   subpopulation of cells capable of self-renewal, differentiation, and
   resistance to conventional therapies. Understanding the regulatory
   mechanisms governing GSC properties is essential for developing
   effective glioblastoma treatments.

   The Chromobox (CBX) family of proteins are essential components of the
   Polycomb repressive complex 1 (PRC1) and play crucial roles in
   regulating stem cell identity and function [[38]1, [39]2]. For
   instance, the absence of CBX1 leads to alterations in constitutive
   pericentric heterochromatin, reductions in histone 3 lysine 9
   trimethylation, and decreased levels of heterochromatic 1γ, which in
   turn disrupt the structural integrity and function of nucleoli during
   the cell cycle [[40]3]. CBX2 has been shown to enhance the
   sphere-forming ability of glioblastoma by regulating the
   Phosphatidylinositol 3-Kinase/Akt Pathway signaling pathway [[41]4].
   CBX3, on the other hand, promotes the differentiation of stem cells
   into smooth muscle cells through its interaction with diaphanous
   homolog 1/serum response factor [[42]5]. CBX4 maintains the
   slow-cycling, undifferentiated state of human epidermal stem cells by
   inhibiting stem cell activation via SUMOylation [[43]6]. Furthermore,
   CBX5 is known to regulate the stem cell-like phenotype of lung cancer
   cells and can serve as a prognostic marker for lung cancer, also
   playing a key role in the differentiation of endothelial progenitor
   cells into endothelial cells [[44]7, [45]8]. The deletion of CBX6 leads
   to a significant reduction in the expression of several critical
   pluripotency genes, such as Rex1, Klf4, NANOG, and ERR-β [[46]9],
   thereby triggering cellular differentiation. CBX7 not only induces the
   generation of pluripotent stem cells but also promotes the self-renewal
   of hematopoietic stem cells and progenitors through both canonical and
   non-canonical pathways [[47]10]. Lastly, CBX8 contributes to the
   stem-like phenotype of hepatocellular carcinoma by regulating bone
   morphogenetic protein 4 activation of the Mitogen-Activated Protein
   Kinase pathway [[48]11]. In summary, the CBX family plays pivotal roles
   in the biological regulation of tumor stem cells [[49]6, [50]12].

   In this context, we aim to investigate the relationship between CBX
   family proteins and GSCs. Among the CBX family members, CBX7 plays
   different roles in various types of cancer. In bladder cancer, CBX7
   inhibits the Aldo-keto reductase family 1 member B10 and subsequently
   stimulates Extracellular Signal-Regulated Kinase signaling, thereby
   suppressing bladder cancer progression [[51]13]. However, in gastric
   cancer, CBX7 has been found to suppress p16 and activate the AKT-NF-kB
   signaling pathway to promote the stemness phenotype of gastric cancer
   cells [[52]14, [53]15]. In breast cancer, CBX7 inhibits cell growth and
   metastasis through blocking the binding of Twist1 to EphA2 promoter
   [[54]16]. Despite literature reporting that CBX7 suppresses
   glioblastoma migration through the activation of the Hippo signaling
   pathway [[55]17], the potential mechanisms by which it regulates GSCs
   in glioblastoma remain unclear. Our research focuses on elucidating the
   regulatory mechanisms of CBX7 in GSCs and further identifying and
   validating its direct target genes in glioblastoma.

   In this study, we report the regulatory function and mechanism of CBX7
   in glioblastoma cells stemness. Our findings indicate that CBX7
   suppresses stem cell-like biological behaviors in glioblastoma by
   inhibiting the NF-κB signaling pathway. In vitro experiments
   demonstrated that CBX7 can inhibit glioblastoma cell stemness. In vivo
   experiments confirmed that CBX7 suppresses tumor formation.
   Furthermore, we identified MYH9 as a novel downstream target gene of
   CBX7. Our results show that MYH9 drives the NF-κB signaling pathway to
   promote glioblastoma progression. We also clarified that the
   hypermethylation of the CBX7 promoter is mediated by DNA
   methyltransferases (DNMT1, DNMT3A). Immunohistochemistry (IHC) analysis
   of clinical glioblastoma tissue samples revealed that CBX7 expression
   can serve as an independent prognostic factor for glioblastoma
   patients. Overall, our study highlights the critical role of CBX7 in
   regulating glioblastoma cell stemness and provides new insights into
   its potential as a therapeutic target and prognostic marker.

Results

Frequent downregulation of CBX7 predicts poor prognosis in glioblastoma
patients

   To investigate the potential research value of different CBX family
   members in glioblastoma patients, we first analyzed the expression of
   CBX 1-8 in tumor and normal tissues using the GTEx dataset. Our
   analysis revealed that CBX7 was the most significantly downregulated
   member of the CBX family (Fig. [56]1A, [57]B). We further examined CBX7
   expression in primary tissues, downloading sequencing data of primary
   glioblastoma tissues from the CGGA and TCGA datasets via the Gliovis
   database ([58]http://gliovis.bioinfo.cnio.es/). This analysis showed
   that CBX7 mRNA levels were significantly downregulated in glioblastoma
   compared to normal tissues (Fig. [59]1C). Further analysis of glioma
   datasets from the TCGA and CGGA datasets indicated that CBX7 mRNA
   levels were downregulated in both low-grade and high-grade gliomas
   (Fig. [60]1D). CBX7 mRNA levels were observed to be lowest in grade 4
   gliomas (p < 0.01). Survival analysis predicted that patients with
   higher CBX7 expression had significantly longer overall survival (OS)
   compared to those with lower CBX7 expression (Fig. [61]1E, F; log-rank
   survival analysis). Additionally, based on pathological
   characteristics, glioblastoma is divided into four different subtypes:
   proneural, neural, classical, and mesenchymal, which indicate different
   survival rates for glioblastoma patients [[62]18]. We examined the
   association between CBX7 expression and these glioblastoma subtypes and
   found that CBX7 expression varied among different subtypes and was
   significantly different from the normal group (Fig. [63]1H). Isocitrate
   dehydrogenase (IDH) mutations are a critical event in glioblastoma
   progression and is widely used as a diagnostic and prognostic marker
   for glioblastoma patients [[64]19]. We explored the potential
   relationship between CBX7 and IDH mutation in glioblastoma in the TCGA
   and CGGA datasets, finding that CBX7 expression (p < 0.05) was
   positively correlated with IDH mutation, suggesting that CBX7 is a
   potential biomarker for predicting glioblastoma survival and IDH
   classification (Fig. [65]1G). These data collectively indicate that
   CBX7 is closely related to the pathological subtypes of glioblastoma
   and can be an important marker for glioblastoma classification (Fig.
   [66]1G, H). To further determine the clinical significance of CBX7 in
   human glioblastoma, we performed immunohistochemistry (IHC) on a
   glioblastoma tissue microarray containing 180 tumor samples.
   Representative images show that CBX7 was primarily localized in the
   cytoplasm, and quantitative analysis demonstrated that CBX7 expression
   varied among different grades of gliomas, with lower expression in
   higher-grade gliomas (Fig. [67]1I, J). Survival analysis based on
   median CBX7 levels showed that higher CBX7 levels were associated with
   better clinical outcomes, regardless of age and gender (Fig. [68]1K).
   In summary, our data suggest that CBX7 is frequently downregulated in
   glioblastoma and is significantly associated with poor prognosis in
   glioblastoma patients.

Fig. 1. Frequent downregulation of CBX7 predicts poor prognosis in
glioblastoma patients.

   [69]Fig. 1
   [70]Open in a new tab

   A Clustering heat map of mRNA expression of the CBX family in
   TCGA-GTEx-GBM between tissues of Normal and Tumor. Heat map
   visualization performed using the ComplexHeatmap package. B Analysis of
   differences in mRNA expression of CBX family members in tumor and
   normal tissues. Data derived from the TCGA-GTEx-GBM dataset, and
   visualized using the ggplot2 package. Significant differences are
   observed across all members with increases in CBX1, 2, 3, 4, 5 and 8
   and decreases in CBX6 and 7 (n = 1323, ***p < 0.001, Wilcoxon rank sum
   test). C The difference in mRNA expression of CBX7 in glioblastoma was
   analyzed using The Cancer Genome Atlas (TCGA) and the Chinese Glioma
   Genome Atlas (CGGA) databases. The data were obtained from the Gliovis
   online website (p values as indicated, Student’s t test). D Statistical
   analysis of CBX7 mRNA expression across different grades of gliomas
   derived from the datasets indicated in (C). (p values as indicated,
   two-tailed Student’s t test). E, F Overall survival analysis based on
   CBX7 mRNA expression in the indicated glioblastoma datasets
   (***p < 0.001, Kaplan–Meier survival test). G Box plot graphs show the
   differences in expression of CBX7 at the mRNA level of TCGA mutant
   glioma samples (Mut) and wild-type glioma samples (WT) (p values as
   indicated, Student’s t test). H. The box plot shows the expression of
   the CBX7 gene in different subtypes of glioblastoma in the TCGA
   database. I Immunohistochemical (IHC) analysis of CBX7 protein
   expression in a tissue array containing different grades of primary
   glioblastoma tissues. Scale bar = 20 μm. J CBX7 protein was frequently
   decreased in high grade glioblastoma compared with low grade
   glioblastoma (p values as indicated, Student’s t test). K Overall
   survival analysis based on CBX7 expression in glioblastoma tissues.
   Groups were ranked according to CBX7 IHC scores. The median expression
   of CBX7 was used as a cutoff (Log-rank χ^2 = 5.823, ***p < 0.001).

CBX7 inhibits stemness in glioblastoma cells

   To clarify whether CBX7 exerts tumor-suppressive biological functions
   in glioblastoma, we established CBX7 overexpression and knockdown
   models based on CBX7 protein levels in glioblastoma cell lines for
   subsequent studies (Fig. [71]S1A, Supporting Information). Using RNAi
   technology and corresponding controls, we constructed knockdown (KD)
   models in A172 and T98G cells with relatively high endogenous CBX7
   expression, while CBX7 overexpression (OE) models were generated in
   U251 and LN229 cell lines, which have lower CBX7 expression (Fig.
   [72]S1B, Supporting Information), and verified by PCR (Fig. [73]S1C,
   Supporting Information). Cell growth experiments demonstrated that CBX7
   overexpression significantly inhibited the growth of glioblastoma cell
   lines (Fig. [74]2A). Subsequent functional experiments showed that CBX7
   overexpression reduced clonogenicity and significantly decreased the
   migration and invasion potential of glioblastoma cells (Fig. [75]2B–D).
   Conversely, CBX7 knockdown promoted cell proliferation, colony
   formation, and significantly enhanced the migration and invasion
   potential of glioblastoma cells (Fig. [76]S1D–G, Supporting
   Information). Since the CBX family plays a crucial role in regulating
   tumor cell stemness, we hypothesized that CBX7 might exert its
   biological function by regulating the stem cell phenotype in
   glioblastoma. To investigate this, we first selected a stemness dataset
   ([77]GSE54791) related to glioblastoma patients from the GEO database
   to examine the correlation between the CBX family and glioblastoma stem
   cells. We found that compared to differentiated glioblastoma cells
   (DGC), expression of CBX2, CBX3, and CBX5 were upregulated, while CBX1,
   CBX4, CBX6, CBX7, and CBX8 were downregulated. Among them, CBX7 was the
   most significantly downregulated gene in tumor-propagating cells (TPC)
   (Fig. [78]2E). Next, we analyzed the correlation between stemness genes
   and CBX7 expression in glioblastoma using online clinical data. Based
   on gene set enrichment analysis (GSEA) of the combined CGGA-GBM and
   TCGA-GBM datasets, the ssGSEA enrichment score of HALLMARK_STEMNESS
   (stemness score) was significantly negatively correlated with CBX7 mRNA
   expression in glioblastoma (Fig. [79]2F). We first verified that the
   selected glioblastoma cell lines all had the potential for spheroidal
   stemness (Fig. [80]S1G, Supporting Information). We also performed a 3D
   tumor sphere assay and observed the results at the same time points.
   The results showed that sphere-forming ability was reduced in
   CBX7-overexpressing U251 and LN229 cells, whereas it was significantly
   enhanced in CBX7-knockdown T98G and A172 cell lines (Fig. [81]2H, I,
   Fig. [82]S1I, Supporting Information). Western blot results further
   demonstrated that the expression of stem cell markers (CD44, SOX2,
   ALDH1A3) decreased with CBX7 overexpression and increased with CBX7
   knockdown (Fig. [83]2G, Fig. [84]S2J, Supporting Information). In
   conclusion, these data suggest that CBX7 can inhibit stemness in
   glioblastoma cells.

Fig. 2. CBX7 inhibited glioblastoma cell stemness.

   [85]Fig. 2
   [86]Open in a new tab

   A Cell growth assay of control (Vector) and CBX7 transfected LN229 and
   U251 cells using the Cell Counting Kit-8 demonstrated that the
   expression of CBX7 significantly decreases cell viability at all time
   points measured (mean ± SD, n = 4, **p < 0.01, ***p < 0.001, Student’s
   t test). B Colony formation (left) of control cells or those
   ectopically expressing CBX7 and the accompanying statistical analysis
   (right) (n = 4, **p < 0.01, ***p < 0.001, Student’s t test). C, D.
   Transwell invasion and migration assay of the indicated cells (left)
   and statistical analysis (right; n = 4, **p < 0.01, ***p < 0.001,
   Student’s t test,) demonstrate that transfection with CBX7 results in
   decreased levels of invasion and migration. E Clustering heat map of
   the expression of CBX family members at the mRNA level in non-stem
   glioma cells (DGC) and tumor propagating cells (TPC). Differentially
   expressed genes were calculated using the limma R package (p-values as
   indicated, Student’s t-test). Results demonstrate that CBX7 was the
   most significantly downregulated gene amongst the CBX family. F
   Correlation analysis between CBX7 mRNA levels and cell stemness scores
   in the specified dataset reveal a significant negative correlation
   between CBX7 mRNA expression and glioblastoma (p-values as indicated,
   Student’s t test). G Western blot analysis of CD44, ALDH1A3 and SOX2 in
   LN229 and U251 demonstrate a downregulation of these proteins following
   ectopic expression of CBX7. Tubulin served as a loading control. H, I.
   3D-tumor spheroid growth was recorded (left) and quantitatively
   analyzed (right; n = 8, **p < 0.01, Student’s t test) and demonstrated
   that CBX7 expression resulted in significantly reduced sphere-forming
   capability in both U251 and LN229 cells. Scale bars = 200 μm.

CBX7 promotes ubiquitin-dependent degradation of MYH9

   The downstream target genes of CBX7 in glioblastoma remain undefined.
   To address this, we transiently transfected U251 cells with a
   Flag-tagged CBX7 overexpressing adenoviral vector. Using
   co-immunoprecipitation (co-IP) with an anti-Flag antibody or IgG as a
   control, followed by liquid chromatography-tandem mass spectrometry
   (LC-MS/MS), we identified MYH9 as a potential target, as it had the
   highest peptide match score with CBX7 (Fig. [87]3A, Supplementary Table
   [88]S1). To validate the mass spectrometry results, we performed co-IP
   in U251 and LN229 cells, confirming the endogenous interaction between
   CBX7 and MYH9 (Fig. [89]3B). Additionally, co-transfection of CBX7 and
   MYH9 in 293 T cells demonstrated the exogenous interaction between CBX7
   and MYH9 (Fig. [90]3C). Western blot analysis revealed that CBX7
   overexpression decreased MYH9 protein levels, whereas CBX7 knockdown
   increased MYH9 protein levels (Fig. [91]3D). However, analysis of the
   Gliovis database showed no significant correlation between CBX7 and
   MYH9 at the mRNA level (Fig. [92]S2A, [93]S2B, Supporting Information).
   RT-qPCR of transfected U251 and LN229 cells also indicated that CBX7
   overexpression or knockdown did not affect MYH9 transcription levels
   (Fig. [94]S2C, [95]S2D, Supporting Information). We then treated U251
   and LN229 cells with cycloheximide (CHX) to block protein synthesis and
   assessed MYH9 protein stability. The results showed that CBX7
   overexpression significantly promoted the degradation of endogenous
   MYH9 protein (Fig. [96]3F, G). To determine whether CBX7-mediated MYH9
   degradation occurs through the ubiquitin-proteasome or lysosome
   pathway, we transfected 293T cells with either a CBX7 overexpression
   plasmid or an empty vector, followed by treatment with the proteasome
   inhibitor MG132 or the lysosome inhibitor chloroquine (CQ). The results
   showed that MG132 reversed CBX7-mediated MYH9 degradation, while CQ did
   not, suggesting that CBX7 promotes MYH9 degradation via the
   ubiquitin-proteasome pathway (Fig. [97]3E). Additionally,
   co-transfection of CBX7, MYH9, and Ubiquitin (Ub) plasmids in U251 and
   LN229 cells revealed that CBX7 overexpression significantly increased
   the polyubiquitination levels of exogenous MYH9 (Fig. [98]3H). In
   conclusion, our results demonstrate that CBX7 promotes the
   polyubiquitination and degradation of MYH9 via the ubiquitin-proteasome
   pathway.

Fig. 3. CBX7 promotes MYH9 degradation in the form of ubiquitination
modification.

   [99]Fig. 3
   [100]Open in a new tab

   A LC-MS/MS (Left) with an anti-FLAG-tag antibody and IgG in designated
   cells identified MYH9 as a potential binding target, as it had the
   highest peptide match score (sorted by Score Sequest HT(Right)) with
   CBX7. B Co-immunoprecipitation (Co-IP) followed by western blotting
   using anti-CBX7 or anti-MYH9 antibodies confirmed the endogenous
   association of CBX7 and MYH9 in U251 and LN229 cells. C Co-IP followed
   by western blotting using anti-Flag or anti-HA antibodies revealed the
   exogenous association of CBX7 and MYH9 in U251 and LN229 cells. D
   Expression of Flag-tagged CBX7 in either U251 or LN229 cells resulted
   in the downregulation of endogenous MYH9 levels. Knockdown of
   endogenous CBX7 using the shCBX7 silencing vector (#1) resulted in an
   upregulation of endogenous CBX7. NC refers to a scrambled silencing
   sequence that was used as a transfection control. Tubulin was used as a
   loading control. E Transfection of 293T cells with CBX7-overexpressing
   vector resulted in decreased expression of endogenous MYH9 and this was
   reversed following treatment with the proteasome inhibitor MG132 (left
   panel) whilst treatment with the lysosome inhibitor chloroquine (CQ)
   had no effect on MYH9 levels in the cells (right panel). F, G Treatment
   of CBX7-overexpressing LN229 (F) and U251 (G) cells with cycloheximide
   (CHX), used to block new protein synthesis, revealed that CBX7
   overexpression significantly promoted the degradation of endogenous
   MYH9 protein. The relative levels of MYH9 protein at the indicated
   timepoints are shown in the panel on the right. H Co-transfection of
   U251 and LN229 cells with indicated plasmids revealed that CBX7
   overexpression substantially increased the polyubiquitination levels of
   endogenous MYH9. Cell lysates were subjected to denaturing-IP, and the
   ubiquitination of MYH9 and the indicated proteins were detected by
   western blotting.

CBX7 promotes ubiquitination and degradation of MYH9 in a PRC1-dependent
manner

   Degradation of proteins via the ubiquitin-proteasome system (UPS)
   generally involves two stages. The first stage is the interaction
   between the protein substrate, ubiquitin, and the ubiquitylating
   enzymes E1, E2, and E3. The second stage involves the degradation of
   the ubiquitin-modified protein substrate by the proteasome [[101]20].
   By co-transfecting 293T cells with CBX7, MYH9, K48-Ub, and K63-Ub
   plasmids, we found that CBX7 could promote K48-type ubiquitination of
   MYH9 and inhibit K63-type ubiquitination (Fig. [102]S2E, Supporting
   Information). Given that CBX7 cannot function as an E3 ubiquitin ligase
   by itself, we hypothesized that CBX7 might promote MYH9 degradation by
   recruiting other E3 ubiquitin ligases. It is well-known that CBX7, as a
   key component of the classical PRC1 complex, can bind the E3 ubiquitin
   ligase RING1A/B to monoubiquitinate histone H2A. By screening and
   analyzing the MS data, we found that RING1A (RNF1) and RING1B (RNF2)
   were among the high-scoring differentially expressed genes (Fig.
   [103]3A). Therefore, we first considered that CBX7 could function in
   the form of the PRC1 complex, with RING1A/B acting as the E3 ligase. To
   determine whether RING1A/B performed the E3 function, we used the PRC1
   inhibitor RB3 to verify whether it hindered the
   ubiquitination-promoting effect of CBX7 on MYH9. The results showed
   that the addition of the RB3 inhibitor reversed the
   ubiquitination-promoting effect of CBX7 on MYH9 (Fig. [104]S2F,
   Supporting Information). These findings indicate that CBX7 promotes
   K48-type ubiquitination and inhibits K63-type ubiquitination of MYH9,
   and that ubiquitination and degradation of MYH9 are facilitated in a
   PRC1-dependent manner.

CBX7 negatively regulates the NF-κB signaling pathway through MYH9

   To further explore the downstream signaling pathways of the CBX7-MYH9
   regulatory axis, we used blebbistatin, a specific inhibitor of MYH9
   that suppresses the activation of the NF-κB signaling pathway
   [[105]21]. To verify this result, we first co-transfected 293T cells
   with a luciferase reporter gene plasmid for NF-κB and an MYH9
   overexpression plasmid. We found that MYH9 could increase the
   luciferase activity of NF-κB, but that the addition of blebbistatin
   inhibited the activating effect of MYH9 on NF-κB (Fig. [106]S3A,
   Supporting Information). Subsequently, we confirmed through western
   blot that MYH9 overexpression activated the NF-κB signaling pathway
   markers (p65, IκBα), while MYH9 knockdown inhibited their activation
   (Fig. [107]S3B, [108]S3C, Supporting Information). These experiments
   demonstrated that MYH9 promotes the activation of the NF-κB signaling
   pathway. Next, we overexpressed CBX7 in the tumorigenic glioblastoma
   cell line U251 and performed RNA sequencing (RNA-seq) (Fig. [109]4A).
   KEGG pathway enrichment analysis revealed that the NF-κB signaling
   pathway was significantly enriched following CBX7 overexpression (Fig.
   [110]4B), and further analysis indicated that the NF-κB signaling
   pathway was notably suppressed (Fig. [111]4C). GSEA also showed that
   overexpression of CBX7 resulted in the suppression of the NF-κB
   signaling pathway (Fig. [112]4D). Subsequently, RT-qPCR confirmed that
   CBX7 overexpression inhibited the expression of downstream genes of the
   NF-κB pathway (CCL19, CCL21, CXCL1, CXCL13, IL1B, CD14) (Fig. [113]4E).
   Further validation showed that CBX7 overexpression could inhibit NF-κB
   activity as detected using a luciferase assay, and this inhibitory
   effect could be reversed by an NF-κB signaling pathway activator (Fig.
   [114]4F). Additionally, western blotting confirmed that CBX7
   overexpression significantly inhibited the activation of NF-κB pathway
   markers (p65, IκBα), while CBX7 knockdown promoted their expression
   (Fig. [115]4G). Finally, to determine whether CBX7 inhibits the NF-κB
   signaling pathway by promoting the degradation of MYH9, co-transfection
   of MYH9 and CBX7 in U251 and LN229 cells revealed that MYH9 could
   reverse the inhibitory effect of CBX7 on the NF-κB signaling pathway.
   The dual luciferase reporter gene experiment also provided similar
   results (Fig. [116]S3D, [117]S3E, Supporting Information). Therefore,
   we conclude that CBX7 negatively regulates the NF-κB signaling pathway
   through MYH9.

Fig. 4. CBX7 regulates the NF-κB signaling pathway through MYH9.

   [118]Fig. 4
   [119]Open in a new tab

   A Overexpression of CBX7 in the glioblastoma cell line U251 was
   confirmed via RNA sequencing (RNA-seq) 48 h following transfection. B,
   C KEGG pathway enrichment analysis indicated that the NF-κB signaling
   pathway was significantly enriched following CBX7 overexpression (B)
   and that overexpression of CBX7 resulted in notable suppression of the
   NF-κB signaling pathway (C). D Gene set enrichment analysis (GSEA) also
   revealed that overexpression of CBX7 in U251 cells leads to an
   inhibition of the NF-κB signaling pathway. E U251 cells were infected
   with control or AD-CBX7 for 36 h. Cells were collected for RT-qPCR
   analysis. (Data presented as mean ± SD with 3 replicates, **p < 0.01;
   ***p < 0.001, Student’s t test). F A luciferase assay to detect levels
   of NF-κB transcriptional activity in transformed U251 cells revealed
   that CBX7 overexpression results in an inhibition of NF-κB activity
   whilst the addition of HY-P1860 (20 ng/ml; n = 3), an NF-κB signaling
   pathway activator, added 24 h before the assay, led to a reversal of
   the inhibitory effects of CBX7 (**p < 0.01, ***p < 0.001, ns—not
   significant, Student’s t test). G Representative western blot analysis
   of NF-κB signaling pathway markers P65, IκBα and p-P65 in U251 and
   LN229 cells transfected with the indicated plasmids. Results
   demonstrate that CBX7 overexpression (Flag-CBX7) substantially
   inhibited the activation of NF-κB pathway markers (p65, IκBα), while
   CBX7 knockdown (shCBX7) increased their expression.

CBX7 inhibits the stemness of glioblastoma cells through MYH9

   In 2021, Yao and colleagues reported that overexpression of
   Thrombospondin 1 (THBS1) promotes the proliferation, migration, and
   invasion of glioblastoma cells, while knockdown of MYH9 reverses these
   effects [[120]22]. Additionally, a 2023 study reported that gambogic
   amide forms a complex with WD repeat domain 1, Cofilin, and MYH9,
   inhibiting the invasion of patient-derived glioma cells by disrupting
   the cytoskeleton. Overexpression of WD repeat domain 1 promotes the
   proliferation of patient-derived glioma cells, indicating a synergistic
   effect with MYH9 [[121]23]. To investigate the biological functions of
   MYH9 in glioblastoma, we first discovered through a glioma chip
   (HBraG160Su01) that the expression of MYH9 increases with the
   pathological grade (Fig. [122]S4A, [123]S4B, Supporting Information).
   We then overexpressed MYH9 in U251 and LN229 cells and found that MYH9
   could promote the proliferative and sphere-forming ability of
   glioblastoma cells (Fig. [124]S4C, [125]S4D, Supporting Information).
   Moreover, overexpression of MYH9 also promoted the increase of stemness
   markers (Fig. [126]S4E, Supporting Information). These results indicate
   that MYH9 also promoted the malignant phenotype of glioblastoma cells.
   To explore whether CBX7 inhibits the stemness of glioblastoma cells by
   promoting the degradation of MYH9, we co-transfected glioblastoma cell
   lines with CBX7 and MYH9 overexpression plasmids and found that whilst
   overexpression of CBX7 significantly inhibited the growth, invasion,
   and stemness characteristics of glioblastoma cells, these effects were
   reversed by the overexpression of MYH9 (Fig. [127]5A–E). Western blot
   results showed that MYH9 could restore the inhibition of stem cell
   markers (CD44, ALDH1A3, SOX2) caused by CBX7 expression (Fig. [128]5F).
   In addition, we also demonstrated that HY-P1860, an activator of NF-κB
   signaling pathway, can promote tumor cell proliferation and CBX7 can
   inhibit its pro-cancer effect through in vitro and in vivo experiments,
   indicating that CBX7 plays a biological role in tumor suppression by
   inhibiting activation of the NF-κB signaling pathway (Fig. [129]5G, H).
   In summary, these results suggest that CBX7 inhibits the stemness
   phenotype of glioblastoma cells by downregulating MYH9.

Fig. 5. CBX7 inhibits glioblastoma cell stemness through MYH9.

   [130]Fig. 5
   [131]Open in a new tab

   A Proliferation of glioblastoma cells transfected with the indicated
   plasmids was measured using CCK-8 (n = 4, **p < 0.01, ***p < 0.001,
   Student’s t test). Results reveal that transfection with a
   CBX7-overexpressing plasmid alone results in a significant reduction in
   stemness characteristics, whereas co-transfection with CBX7 and MYH9
   partially reversed the inhibitory effects of CBX7. B, D Transwell
   migration assays (B) and the statistical analysis (D; n = 6,
   **p < 0.01, Student’s t test) reveal significant differences when CBX7
   is co-transfected with MYH9 in both LN229 and U251 cells compared to
   when CBX7 is transfected by itself, which leads to significant
   decreases in stemness characteristics. C, E 3D-tumor spheroid growth
   was recorded (C) and quantitatively analyzed (E; n = 12, **p < 0.01,
   ***p < 0.001, Student’s t test) and illustrates the inhibitory effects
   of CBX7 compared to CBX7 co-transfected with MYH9 on spheroid growth.
   Scale bars, 200 μm. F Representative western blots of stem cell markers
   CD44, ALDHIA3, SOX2 in U251 and LN229 cells transfected with indicated
   plasmids. Transfection with CBX7 decreases expression of the stem cell
   markers whilst co-transfection of CBX7 and MYH9 stabilizes expression
   of the markers. G Measurement of proliferative capacity of CBX7
   transfected cells revealed that whilst CBX7 significantly inhibited
   cell proliferation, treatment with an activator of the NF-κB signaling
   pathway, HY-P1860, significantly reversed the inhibitory effects.
   Proliferation was measured using the CCK-8 assay (n = 6, **p < 0.01,
   ***p < 0.001, Student’s t test). H To determine the in vivo effects of
   CBX7, we obtained bioluminescent images of intracranial xenografts
   derived from the implantation of the indicated cells (left) and
   statistical analysis of tumor volume (right; n = 9, Student’s t test,
   **p < 0.01, ***p < 0.001) revealed significant differences in the
   inhibitory effects of CBX7, indicating that CBX7 plays a biological
   role in tumor suppression by inhibiting activation of the NF-κB
   signaling pathway.

CBX7 inhibits the growth of tumor cells in vivo

   To determine the significance of CBX7 in human glioblastoma, we used a
   3D culture method to derive primary cells from patient-derived
   glioblastoma tissue. Second-generation 3D cultures were used for
   xenograft models and subsequent analysis (Fig. [132]6A). To assess the
   correlation between CBX7 and glioblastoma tumorigenesis, we implanted
   second-generation 3D cultures subcutaneously into nude mice and
   measured CBX7 mRNA expression in the subcutaneous tumors by PCR. We
   found that CBX7 mRNA expression levels in these cells somewhat were
   inversely correlated with tumor volume; xenografts with higher CBX7
   expression grew more slowly than those with lower CBX7 expression
   (except for G6), indicating that higher CBX7 levels are associated with
   reduced tumorigenicity (Fig. [133]6B, C). We further analyzed frozen
   sections of subcutaneous tumors derived from G1 and G2 cells, stained
   for proliferation markers, and found that tumors with high CBX7
   expression exhibited lower proliferative capacity (Fig. [134]6D).
   Subsequently, we injected CBX7-overexpressing cells into the axillae of
   nude mice and observed a significant reduction in tumor volume and size
   in the CBX7-overexpressing group (Fig. [135]6E). To further validate
   the tumor-suppressive role of CBX7 in vivo, we performed intracranial
   xenograft tumor growth assays using luciferase-labeled U251 cells and
   control vectors in nude mice. The results showed that CBX7
   overexpression significantly inhibited intracranial tumor development
   (Fig. [136]6F). Additionally, hematoxylin and eosin (H&E) staining of
   brain sections followed by volumetric measurements confirmed that the
   intracranial tumors were significantly smaller in the CBX7
   overexpression group (Fig. [137]6G). In summary, these data support the
   tumor-suppressive function of CBX7 in vivo.

Fig. 6. CBX7 inhibits the growth of tumor cells in vivo.

   [138]Fig. 6
   [139]Open in a new tab

   A Illustration of the 3D-PDX model of patient derived glioblastoma
   tissues. The patient-derived glioblastoma cells were cultured in 3D and
   the second passages were used for xenograft models and subsequent
   analysis. B, C. Xenograft tumors derived from 3D-cultured
   patient-derived glioblastoma cells (B) and CBX7 mRNA expression levels
   in the xenografts (C). Analysis revealed that CBX7 mRNA expression
   levels in these cells were somewhat inversely correlated with tumor
   volume (except for G6), indicating that higher CBX7 levels are
   associated with reduced tumorigenicity. D Cell immunofluorescence
   analysis of Ki-67 expression (a marker of proliferation) in tumors
   derived from primary glioblastoma cell lines (left; at 20x
   magnification). Statistical analysis revealed that tumors formed by the
   glioblastoma primary cell line G1 exhibit a higher proliferative
   capacity (right; ***p < 0.001, Student’s t test). E Representative
   images of subcutaneous xenografts at day 21 following U251 injection.
   The right panel illustrates the tumor weight analysis, revealing that
   U251 cells overexpressing CBX7 resulted in a significant reduction in
   tumor volume and size (n = 5, **p < 0.01, Student’s t test). F
   Bioluminescent images of intracranial xenografts derived from the
   implantation of either vector or CBX7 expressing cells (left) and
   statistical analysis of tumor volume revealed that CBX7 overexpression
   significantly inhibited intracranial tumor development (right; n = 6,
   **p < 0.01, Student’s t test). G Representative hematoxylin and eosin
   (H&E) stained images demonstrating that overexpression of CBX7 reduces
   the volumes of xenograft tumors by more than 2-fold.

High promoter methylation leads to frequent downregulation of CBX7 in
glioblastoma

   Studies have reported that the downregulation of CBX7 in glioblastoma
   is caused by hypermethylation of its promoter [[140]17, [141]24], but
   the specific DNA methyltransferases (DNMTs) mediating CBX7 promoter
   hypermethylation have not been elucidated. To determine the cause of
   hypermethylation of the CBX7 promoter, we analyzed methylation data
   from the CGGA database and found that the methylation level of CBX7 was
   higher in tumor compared to normal tissues (Fig. [142]7A). Moreover,
   survival analysis showed that patients with lower methylation levels
   had better survival outcomes (Fig. [143]7B). Subsequently, a negative
   correlation between CBX7 mRNA and its methylation levels were observed
   in the TCGA database (Fig. [144]7C), indicating that promoter
   methylation plays a significant regulatory role in the expression of
   CBX7 in glioblastoma. Following this, we treated the U251 and LN229
   cell lines with the DNA demethylating agent 5-Aza-CdR and observed an
   increase in CBX7 protein expression (Fig. [145]7D, H). Since members of
   the DNA methyltransferase (DNMT) family, including DNMT1, DNMT3A, and
   DNMT3B, play a crucial role in maintaining DNA methylation, we
   established knockdown models for these DNMTs in U251 and LN229 cells.
   Western blot analysis showed that the knockdown of DNMT1 and DNMT3A
   increased CBX7 protein levels, while DNMT3B did not have a similar
   effect (Fig. [146]7E–G), and PCR results also reached the same
   conclusion (Fig. [147]7I). These results indicate that DNMT1 and DNMT3A
   mediate the hypermethylation of the CBX7 promoter, leading to its
   frequent downregulation in glioblastoma. To further verify whether CBX7
   is hypermethylated in glioma cell lines and tissues, we predicted the
   CpG islands in the CBX7 promoter region and designed corresponding
   primers (Fig. [148]7J). Pyrosequencing revealed that the methylation
   rate of CBX7 was higher in glioma cells (Fig. [149]7K), and clinical
   sample tests also showed that the methylation rate of CBX7 was higher
   in cancer tissues compared to adjacent non-cancerous tissues (Fig.
   [150]7L). In summary, we show that downregulation of CBX7 in
   glioblastoma is caused by promoter hypermethylation that is mediated by
   DNMT1 and DNMT3A.

Fig. 7. DNMT1 and DNMT3A mediate hypermethylation of the CBX7 promoter.

   [151]Fig. 7
   [152]Open in a new tab

   A Beta values of Cbx7 in control brain and Glioblastoma samples, from
   CGGA data set, plotted as a cellular map (p values are as indicated,
   Student’s t test). B. Log-rank survival curve based on the methylation
   levels of CBX7 in the CGGA data set. The median expression of CBX7 was
   used as a cutoff. Results show that patients with lower methylation
   levels had significantly better survival outcomes (Log-rank
   χ^2 = 3.823, p = 0.002). C Correlation analysis, indicating a
   significant negative correlation between Cbx7 methylation and its
   expression in the TCGA dataset (Spearman’s correlation analysis, p
   value and correlation coefficient R are as indicated). D, H. Protein
   (D) and mRNA levels (H) of CBX7 were restored in U251 and LN229 cells
   following treatment with 5-Aza-CdR at 10 μM for the indicated times
   (n = 6, **p < 0.01, ***p < 0.001, Student’s t test). E–G Western
   blotting analysis indicated that the knockdown of DNMT1 (E) and DNMT3A
   (F) in U251 and LN229 cells led to a significant increase in the level
   of CBX7. However, the knockdown of DNMT3B (G) did not have the same
   effect. I CBX7 mRNA levels were examined in DNMT1, DNMT3A, or
   DNMT3B-silenced U251 and LN229 cells by qRT-PCR (n = 8, **p < 0.01,
   ***p < 0.001, ns, not significant). Results revealed that whilst
   knockdown of DNMT1 and DNMT3A resulted in significant increases in the
   relative levels of CBX7 mRNA, knockdown of DNMT3B had no effect on CBX7
   mRNA levels. J Design of primers targeting the promoter region of CBX7
   for pyrosequencing. The solid box contains the primer sequences, and
   nucleotides in red indicate the 12 CG sites detected by the primers. K,
   L Methylation status of the CBX7 promoter in glioblastoma cell lines
   (K) and paired (L) glioblastoma tissues (T) and their adjacent
   noncancerous tissues (N). Percentage methylation of promoter was
   calculated as total percent of methylated cytosines from the 12 sites
   (n = 5, ***p < 0.001, Student’s t test). Pyrosequencing revealed that
   the methylation rate of CBX7 was higher in glioma cells (K) and in
   cancer tissues compared to adjacent non-cancerous tissues (L).

Discussion

   CBX7 is an important component of PRC1 and is involved in the
   development of various cancers, including glioblastoma [[153]17,
   [154]25]. In this study, query of the TCGA database revealed that CBX7
   is the most significantly downregulated member of the CBX family in
   glioblastoma compared to normal tissues. Functional experiments showed
   that overexpression of CBX7 significantly inhibited glioblastoma cell
   growth, colony formation, migration, invasion, and stemness
   proliferation in vitro, and notably reduced the growth rate of
   intracranial and subcutaneous tumors in vivo. Mechanistically, CBX7
   promotes the degradation of MYH9 protein in the form of the PRC1
   complex, thereby inhibiting the NF-κB signaling pathway. The inhibitory
   effects of CBX7 on glioblastoma cell proliferation, migration,
   invasion, and stemness are reversed following MYH9 overexpression.
   Additionally, our study revealed that DNMT1 and DNMT3A mediate the
   hypermethylation of the CBX7 promoter, leading to its frequent
   silencing in glioblastoma cell lines and primary tumor tissues.
   Clinically, low levels of CBX7 expression are associated with poor
   prognosis in glioblastoma patients, increased self-renewal capacity,
   and enhanced proliferation of cancer cells. Our findings suggest that
   DNMT-mediated methylation is a promising target for glioblastoma
   treatment.

   The stem cell phenotype refers to a series of characteristics exhibited
   by stem cells or cells with stem cell-like properties, which enable
   them to self-renew and differentiate into various cell types [[155]26].
   In glioblastoma, the stem cell phenotype is often considered to play a
   critical role in tumor initiation, drug resistance, and recurrence,
   making it an important focus of glioblastoma research [[156]27]. In
   this study, based on the TCGA public database, we explored the
   differential expression of CBX family members in stemness gene sets and
   found that CBX7 was significantly downregulated in tumor-propagating
   cells (TPCs) compared to differentiated glioblastoma cells (DGCs), with
   the most pronounced downregulation among CBX family members. Subsequent
   in vitro functional experiments demonstrated that overexpression of
   CBX7 significantly inhibited tumor sphere growth, and western blot
   results also showed decreased expression of stemness markers. Our
   findings complement the understanding of the role of CBX7 in the
   stemness phenotype in glioblastoma.

   We conducted mass spectrophotometry analysis of glioblastoma cells and
   identified MYH9 as a novel regulatory target of CBX7. While CBX
   proteins typically function as transcriptional repressors of downstream
   targets, analysis of sequencing data showed that MYH9 mRNA levels did
   not change with CBX7 overexpression, whereas CBX7 promoted MYH9 protein
   degradation. MYH9, also known as non-muscle myosin heavy chain IIA,
   belongs to the myosin II protein subfamily [[157]28]. MYH9 participates
   in cell polarity, cell-cell adhesion, and cytoskeleton maintenance
   through its interaction with actin filaments and is involved in cancer
   invasion and metastasis [[158]29]. MYH9 promotes tumor stemness by
   regulating various signaling pathways. For instance, MYH9 can
   ubiquitinate and degrade GSK3β, disrupting the β-catenin destruction
   complex and inducing stem-like phenotypes, epithelial-mesenchymal
   transition (EMT), and c-Jun signaling [[159]30]. In non-small cell lung
   cancer, MYH9 activates the mTOR signaling pathway, enhancing stemness
   features [[160]31]. MYH9 also serves as a marker and prognostic
   indicator for esophageal cancer stem cells, promoting tumorigenesis via
   the PI3K/AKT/mTOR axis [[161]32]. Moreover, MYH9 plays a crucial role
   in maintaining and supporting hematopoietic stem cells and their
   progenitors [[162]33]. Here, we reveal that CBX7 interacts with MYH9 to
   promote its degradation, thereby inhibiting glioblastoma cell stemness.

   Recent research on MYH9 has focused on post-translational modifications
   [[163]30, [164]34]. For example, in nasopharyngeal carcinoma, DNAJA4
   directly interacts with MYH9, recruiting PSMD2 to promote MYH9
   ubiquitination and degradation, thereby inhibiting nasopharyngeal
   carcinoma migration and EMT [[165]35]. In triple-negative breast
   cancer, circRNA EIF6-224aa directly interacts with MYH9, inhibiting
   MYH9 degradation through the ubiquitin-proteasome pathway and
   activating the Wnt/β-catenin signaling pathway [[166]36]. Conversely,
   lncRNA TPRG1-AS1 directly interacts with MYH9, promoting its
   degradation via the ubiquitin-proteasome pathway, thus hindering
   F-actin stress fiber formation and ultimately inhibiting human aortic
   smooth muscle cell migration [[167]37]. In this study, we found that
   CBX7 interacts with MYH9, with Ring1A/B acting as E3 ligases, promoting
   MYH9 degradation through the ubiquitin-proteasome pathway in
   glioblastoma. Reportedly, Ring1A/B are two key subunits of the PRC1
   complex, playing crucial roles in biological processes. They possess E3
   ligase activity, catalyzing the monoubiquitination of histone H2A
   (H2AK119ub1), thus participating in chromatin structure regulation and
   repressing gene expression [[168]38]. Ring1A/B also interact with
   Polycomb group (PcG) RING finger protein (PCGF) family members to form
   different PRC1 subcomplexes [[169]38]. Here we reveal that CBX7, as
   part of the PRC1 complex, promotes MYH9 ubiquitination and degradation,
   thereby inhibiting the glioblastoma stem cell phenotype, with Ring1A/B
   serving as E3 ligases. However, further exploration is needed to
   determine the specific domains involved in the CBX7-MYH9 interaction
   and the specific sites of MYH9 ubiquitination by Ring1A/B.

   To elucidate the downstream signaling pathways regulated by CBX7-MYH9,
   we prioritized the NF-κB signaling pathway in our transcriptome
   sequencing analysis. The NF-κB signaling pathway is a key regulatory
   network involved in immune responses, inflammation, cell proliferation,
   and differentiation in glioblastoma [[170]39]. For instance, TRIM22 can
   activate the NF-κB signaling pathway by degrading IκBα, promoting
   glioblastoma cell proliferation [[171]40]. Additionally, the NF-κB
   pathway can exacerbate the Warburg effect, leading to increased
   expression of LINC01127, activation of the MAPK4-JNK pathway, and
   promotion of glioblastoma stem cells [[172]41]. Inhibition of the NF-κB
   pathway by curcumin significantly reduces glioblastoma cell
   proliferation and invasion and induces early G2/M cell cycle arrest
   [[173]42]. In this study, we found that overexpression of CBX7
   significantly inhibits, while MYH9 overexpression activates the NF-κB
   signaling pathway. The inhibitory effect of CBX7 on the NF-κB signaling
   pathway can be reversed by MYH9.

   Overall, our findings reveal that DNA methyltransferases (DNMT1 and
   DNMT3A) mediate the hypermethylation of the CBX7 promoter, leading to
   its frequent downregulation in glioblastoma tissues and cell lines. The
   downregulation of CBX7 is associated with poor clinical outcomes.
   Additionally, we have demonstrated for the first time that CBX7, as
   part of the PRC1 complex, promotes MYH9 ubiquitination and degradation,
   inhibiting stemness, migration, invasion, and colony formation in
   glioblastoma cells. This newly identified CBX7-MYH9-NF-κB axis provides
   a novel therapeutic strategy for glioblastoma.

Materials and methods

Online cancer database analysis

   We downloaded publicly available clinical and gene expression data from
   The Cancer Genome Atlas (TCGA) dataset
   ([174]http://tcga-data.nci.ngov/tcga/) and the Chinese Glioma Genome
   Atlas (CGGA) database ([175]http://www.cgga.org.cn). Pan-cancer data on
   CBX7 gene expression were obtained from TCGA. In addition, stem and
   non-stem glioma data were obtained from [176]GSE54791. Differentially
   expressed genes were calculated using the limma package in R. Survival
   analysis was performed using the survival package in R, and the cohort
   was divided into high expression and low expression groups based on the
   best Cut-off value.

Clinical samples and immunohistochemical analysis

   Immunohistochemical (IHC) staining was used to analyze the expression
   of CBX7 protein in glioblastoma tissues on a tissue microarray. Tissue
   microarrays were constructed by Shanghai Zhuoli Biotechnology Co., Ltd
   (ZL-BraG180sur01). The Ethics Committee of Shanghai Zhuoli
   Biotechnology Co., Ltd. Approved the study, and all subjects provided
   consent. Briefly, tissue slides were incubated with a rabbit anti-CBX7
   antibody (1:5000; Abcam, Cambridge, UK), an anti-rabbit secondary
   antibody from Zymed Systems (1:10000; Invitrogen, Carlsbad, CA) and
   3,3’-diaminobenzidine (DAB) to visualize the IHC labeling. The slides
   were lightly counterstained with cresyl violet. Normal rabbit IgG was
   used to confirm the specificity of the IHC. The expression of CBX7 was
   semi-quantitatively scored based on an established immunoreactivity
   scoring (IRS) system, with some modifications [[177]43]. In summary,
   the IRS was calculated as the product of the proportion score
   (percentage of positive cells) and the staining intensity score (0-3).
   The proportion score represents the percentage of positively stained
   cells, whilst the intensity score represents the average intensity of
   the staining (0: no staining; 1: yellow; 2: claybank; and 3: tawny).
   Two pathologists blinded to the patient categories reviewed and scored
   the slides.

Cell lines, primary cell preparation, and culture conditions

   Human glioblastoma cell lines LN229, A172, and T98G were obtained from
   the American Type Culture Collection (ATCC). The U251 cell line and the
   human embryonic kidney cell line 293T were purchased from Pricell
   (Wuhan, China). All cells were cultured in Dulbecco’s Modified Eagle
   Medium (DMEM) supplemented with 10% Fetal Bovine Serum (FBS; CellBio,
   Australia). All cell lines were authenticated by short tandem repeat
   (STR) analysis and tested for mycoplasma contamination. For the
   patient-derived glioblastoma cell cultures, fresh glioblastoma tissue
   was collected immediately following tumor resection, washed, minced,
   and enzymatically dissociated. The tumor cells were resuspended and
   cultured in GBO medium to allow the formation of tumor spheres
   [[178]44]. The primary spheres were cultured for 10 days to obtain a
   sufficient number of cells for passaging. Cells from the third or
   fourth generation were used for subsequent experiments, including tumor
   xenograft and western blot analysis.

Vector construction and transduction

   The full-length cDNA encoding human CBX7 and MYH9 were amplified by PCR
   and verified by DNA sequencing. The confirmed CBX7 cDNA was inserted
   into the lentiviral vector CMV-MCS-3FLAG-SV40-EGFP (Genechem, Shanghai,
   China). The MYH9 cDNA was inserted into the lentiviral vector with an
   HA tag, pCDH-CMV-MCS-EF1-PURO (Transheep, Shanghai, China). To achieve
   stable cell expression of CBX7, the indicated cells were transduced
   with the established lentivirus or the corresponding construct (Mock),
   and then selected with 2 μg/mL puromycin (Sangon Biotech, Shanghai,
   China). The short hairpin RNA (shRNA) targeting CBX7 and MYH9 were
   inserted into the vectors pGPU6/GFP/Neo-Homo-365 and
   pGPU6/GFP/Neo-Homo-259 respectively (GenePharm, Suzhou, Jiangsu,
   China).

Immunoprecipitation assays

   Immunoprecipitation (IP) experiments were conducted to analyze protein
   interactions. Briefly, cells were transfected with specified plasmids,
   and whole-cell lysates were precipitated using protein A/G microbeads
   (Santa Cruz, CA) with the specified antibodies. The precipitates were
   analyzed by western blotting using the indicated antibodies.

Cell growth and colony formation assay

   For the cell growth assay, a total of 1000 cells were seeded into a
   96-well plate and monitored at specified time points using the Cell
   Counting Kit-8 (CCK-8) according to the manufacturer’s instructions
   (GlpBio, CA, USA). For the colony formation assay, 800 cells per well
   were seeded into six-well plates and maintained in complete medium for
   14 days. Colonies were fixed with 4% paraformaldehyde (PFA) (Sangon,
   Shanghai, China), stained with 5% cresyl violet (Sangon, Shanghai,
   China), and the number of colonies was counted using an inverted
   microscope.

EdU labeling assay

   As previously described [[179]45], DNA replication was analyzed using
   the 5-Ethynyl-2′-deoxyuridine (EdU) incorporation method. Briefly,
   cells were cultured in a 48-well plate for 24 h, incubated with EdU
   (Invitrogen) for 4 h, and then fixed. The staining procedure was
   performed according to the manufacturer’s instructions (Invitrogen,
   Carlsbad, CA). After staining, coverslips were mounted with Gelmount
   containing Hoechst 33342. The quantitative results are expressed as the
   percentage of EdU positive cells relative to the total number of
   cells/nuclei.

Migration assay

   A cell invasion assay was performed using Boyden Transwell chambers
   (8 mm pore size, BD Biosciences, Mountain View, CA) coated with
   Matrigel (50 μg, BD Biosciences) in a 48-well plate. Briefly, 1 × 10^5
   cells were seeded into the upper chamber, and medium containing 20% FBS
   was added to the lower chamber. After 36 h of incubation, the cells
   were fixed with 4% formaldehyde and stained with 5% cresyl violet.
   Cells on the upper side of the filter were carefully removed, and the
   cells on the underside of the filter were defined as invasive. Invasive
   cells were examined and counted using the EVOS cell imaging system
   (Life Technologies, Grand Island, NY, USA). Cell invasion was expressed
   as the percentage of migrating cells relative to the total area.

3D-tumor spheroid assay

   According to the 3D spheroid formation and growth protocol, 1000 cells
   were seeded in a spheroid microplate (Corning Inc., NY, USA). Growth of
   the 3D spheroid-cultured cells was monitored using a live-cell imaging
   system (EVOS^® FL Auto Imaging System, Life Technologies, Carlsbad, CA,
   USA). For spheroid growth assessment, images of the tumor spheroids at
   designated time points were captured and spheroid diameter was measured
   to reflect spheroid growth.

Reverse transcription PCR (RT-PCR) and Quantitative PCR (RT-qPCR)

   Following manufacturer’s instructions, cellular RNA was extracted using
   Trizol^® reagent (at no. 15596026, Invitrogen, Carlsbad, CA, USA). cDNA
   was synthesized using the TaqMan Reverse Transcription Kit (Cat no.
   4366596, Applied Biosystems) following the manufacturers protocol. PCR
   analysis for quantifying mRNA expression was performed on the
   LightCycler 480 II instrument (Roche Applied Science) using the Real
   SYBR Mixture (Cat no. 4309155, Invitrogen, Carlsbad). The endogenous
   control, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), expression
   was monitored. Each experiment was conducted in triplicate. The
   relative mRNA expression was calculated using the 2^−ΔΔCt method.
   Specific oligonucleotide primer pairs are listed in Table [180]S2,
   [181]Supporting Information.

mRNA sequencing and analysis process

   The total transcriptome RNA-seq analysis was performed on the Illumina
   Novaseq 6000 sequencing platform at OE Biotech Co., Ltd. (Shanghai,
   China) to detect the mRNA expression profile of U251 cells
   overexpressing CBX7. Clean reads were aligned to the reference genome
   using HISAT2 software, generating raw counts corresponding to each
   known gene (in total 23899), and calculating gene expression levels
   (FPKM). Differential expression gene analysis was conducted using the
   DESeq2 software, where genes meeting the criteria of a q-value < 0.05
   and fold change > 2 were defined as differentially expressed genes
   (DEGs). Kyoto Encyclopedia of Genes and Genomes Pathway enrichment
   analysis and GSEA enrichment analysis of the differentially expressed
   genes were performed based on the hypergeometric distribution algorithm
   and GSEA software.

Pyrosequencing for CBX7 methylation levels

   DNA methylation assays were performed by pyrosequencing of the CBX7
   gene. Briefly, genomic DNA was extracted from frozen patient glioma
   tissue and glioblastoma cell samples and bisulfite transformed. The
   bisulfite-converted DNA was then amplified and bound to a
   single-stranded template. Pyrosequencing runs were performed using the
   PyroMark Q96/48 ID System (QIAGEN, Hilden, Germany). Further details
   are provided elsewhere [[182]46]. As mentioned above, methylation
   levels are expressed as the ratio of methylated cytosine to the sum of
   methylated and unmethylated cytosine at the 2-cytosine position. Primer
   sequences are provided in Table [183]S3, Supporting Information.

Luciferase reporter assay

   We employed the NF-κB promoter luciferase reporter plasmid (MiaoLing;
   Wuhan, China) to evaluate the effect of CBX7 on NF-κB transcriptional
   activity. Briefly, the indicated cells were transiently transfected
   with the vector. The pRL-TKRenilla luciferase plasmid (Promega) was
   also transfected to serve as a control for the total amount of
   transfected DNA in each well in all experiments. Forty-eight hours
   post-transfection, the cells were harvested and luciferase activity was
   measured using the Dual-Luciferase^® Reporter Assay System (Promega).

Western blot assay and chemical agents

   Protein expression was determined using standard western blotting.
   Antibodies and chemical reagents used to detect the indicated proteins
   are listed in Table [184]S4, Supporting Information.

Tumor sphere formation and growth assay

   Briefly, cells were seeded in 6-well plates (1000 cells/well) and
   maintained in NSC medium (DMEM/F-12 with 20 ng/ml bFGF (PeproTech,
   Rocky Hill, NJ), 20 ng/ml EGF (PeproTech), N2 (Thermo) and B27 (Thermo)
   to facilitate sphere formation. Images of the tumor spheres were
   obtained at designated time points, and sphere diameters were measured.
   At least 10 random fields were analyzed per experiment.

Animal experiments

   Four-week-old female nude mice were purchased from the Shanghai
   Institute of Biological Sciences, Chinese Academy of Sciences, and were
   housed under specific pathogen-free conditions. All mice were randomly
   assigned. For the U251 intracranial model, cells expressing the firefly
   luciferase gene were infected with either CBX7-overexpressing
   lentivirus or control lentivirus. A total of 5 × 10^5 cells were
   stereotaxically implanted into the brains of individual mice. Tumor
   growth was monitored using bioluminescence imaging using the IVIS
   Spectrum system and quantified with Living Image Software.
   Additionally, to assess the tumor-suppressive effect of the CBX7
   construct, a total of 5 × 10^6 U251 cells expressing ectopic CBX7 or an
   empty vector were injected into the flanks of each nude mouse (n = 10),
   with tumor growth assessed by measuring tumor diameter every three
   days. At the end of the experiment, mice were euthanized by an overdose
   of anesthetic. The whole brains or tumors were extracted,
   paraffin-embedded, sectioned, and stained with hematoxylin and eosin
   (H&E). Images were captured using a laser scanning confocal microscope.
   All animal care and handling procedures were conducted in accordance
   with the National Institutes of Health’s Guide for the Care and Use of
   Laboratory Animals. All animal experiments were approved by the Ethics
   Committee of the Affiliated Wuxi People’s Hospital of Nanjing Medical
   University (No. (2024)235). Animal experiments were performed by two
   technicians who were blinded to the treatment conditions of the mice.
   Collection of human glioblastoma samples were approved by the Ethics
   Committee of the Affiliated Wuxi People’s Hospital of Nanjing Medical
   University (No. (2024)133).

Statistical analysis

   The experiments were independently repeated at least three times, and
   data are presented as mean ± standard deviation (SD). All data were
   analyzed using the SPSS 20.0 software package (IBM Corp., Boston, MA,
   USA) and were consistent with normal distribution and homogeneity of
   variance tests. The t test was used to compare differences between two
   groups, while one-way ANOVA was used for mean comparisons among
   multiple sample groups, followed by Tukey’s post hoc test for multiple
   comparisons within the groups. A P-value of less than 0.05 was
   considered statistically significant. Graphs were analyzed using
   GraphPad Prism 9.5 software (GraphPad Software, Boston, MA, USA).

Supplementary information

   [185]Supplementary Figures and legends^ (785KB, pdf)
   [186]Table S1^ (44.2KB, xlsx)
   [187]Table S2^ (9.5KB, xlsx)
   [188]Table S3^ (10.1KB, xlsx)
   [189]Table S4^ (9.2KB, xlsx)
   [190]Original western blots^ (3.6MB, png)

Acknowledgements