ROR1 facilitates glioblastoma growth via stabilizing GRB2 to promote c-Fos expression in glioma stem cells

Abstract

Background

Glioma stem cells (GSCs) are the root cause of tumorigenesis, recurrence, and therapeutic resistance in glioblastoma (GBM), the most prevalent and lethal type of primary adult brain malignancy. The exploitation of novel methods targeting GSCs is crucial for the treatment of GBM. In this study, we investigate the function of the novel ROR1-GRB2-c-Fos axis in GSCs maintenance and GBM progression.

Methods

The expression characteristics of ROR1 in GBM and GSCs were assessed by bioinformatic analysis, patient specimens, and patient-derived GSCs. Lentivirus-mediated gene knockdown and overexpression were conducted to evaluate the effect of ROR1 on GSCs proliferation and self-renewal both in vitro and in vivo. The downstream signaling of ROR1 in GSCs maintenance was unbiasedly determined by RNA-seq and validated both in vitro and in vivo. Finally, rescue assays were performed to further validate the function of the ROR1-GRB2-c-Fos axis in GSCs maintenance and GBM progression.

Results

ROR1 is upregulated in GBM and preferentially expressed in GSCs. Disruption of ROR1 markedly impairs GSC proliferation and self-renewal, and inhibits GBM growth in vivo. Moreover, ROR1 stabilizes GRB2 by directly binding and reducing its lysosomal degradation, and ROR1 knockdown significantly inhibits GRB2/ERK/c-Fos signaling in GSCs. Importantly, ectopic expression of c-Fos counteracts the effects caused by ROR1 silencing both in vitro and in vivo.

Conclusions

ROR1 plays essential roles in GSCs maintenance through binding to GRB2 and activation of ERK/c-Fos signaling, which highlights the therapeutic potential of targeting the ROR1-GRB2-c-Fos axis.

Graphical Abstract

Graphical Abstract.

Key Points.

1. ROR1 is preferentially expressed in GSCs and predicts poor prognosis of glioma patients.

2. ROR1 is indispensable for GSCs proliferation and self-renewal.

3. ROR1 activates GRB2/ERK/c-Fos signaling to promote tumorigenicity of GSCs through binding to GRB2.

Importance of the Study.

Targeting glioma stem cells (GSCs) is one of the most promising strategies for treating glioblastoma (GBM). In this study, we identified that ROR1, a tyrosine kinase-like orphan receptor, was upregulated in GBM patients and predominantly expressed in GSCs. Knockdown of ROR1 disrupts GSCs proliferation and self-renewal both in vitro and in vivo. Mechanistically, ROR1 binds with GRB2 and facilitates downstream ERK/c-Fos signaling, which promotes stemness maintenance and tumorigenicity of GSCs. These results highlight the ROR1-GRB2-c-Fos axis in the maintenance of GSCs and suggest the therapeutic potential of targeting ROR1 in GBM therapy.

Glioblastoma (GBM) is the most prevalent and aggressive malignant primary brain tumor that is classified as grade 4 glioma by the World Health Organization (WHO).^1,2 Despite a variety of treatment advances including surgery, concurrent chemoradiotherapy, and tumor treatment fields, the prognosis of GBM remains extremely poor.^3,4 The median survival time of GBM patients was less than 16 months and the 5-year survival rate was less than 4%.^1,5 Furthermore, most GBM patients are resistant to chemoradiotherapy and relapse inevitably after tumor excision. GBM harbors a functional population of tumor cells with stem cell-like properties, known as glioma stem cells (GSCs).^6,7 GSCs possess self-renewal, sustained proliferation, multilineage differentiation, and robust tumorigenic capacity in vivo, which was considered as the key mediator of GBM recurrence and treatment resistance.^6,8 Identifying novel markers and mechanisms involved in GSCs maintenance and subsequently targeting GSCs could be a promising strategy to improve GBM treatment outcomes.

Altered expression levels and activities of transcription factors (TFs) are hallmarks of cancers and are essential for the maintenance of GSCs.^9,10 Numerous small molecule modulators targeting tumor-associated transcription factors have been used in clinical trials for cancer therapy.¹¹ The identification of altered transcription factors and their function in GSCs is important for elucidating the mechanism of GSCs maintenance and for the development of GSCs-targeted therapy. As a proto-oncogene activated in multiple tumors, c-Fos encodes a nuclear DNA-binding protein to form the transcription factor–activating protein 1 (AP-1) complex with c-JUN, which exerts transcription factor activity and is widely involved in tumor cell proliferation, migration, and treatment resistance.¹² The functions of c-Fos on cancer stem cells (CSCs) maintenance have also been reported in head and neck squamous cell carcinoma, lung cancer, breast cancer, and glioma.^13–15 However, the upstream and downstream mechanisms of c-Fos in GSCs maintenance are still not fully understood.

Receptor tyrosine kinases (RTKs) widely participate in various physiological and pathological processes through signal transduction, and dysregulation of RTKs implicated in various tumors.¹⁶ As a member of RTKs, the receptor tyrosine kinase-like orphan receptor 1 (ROR1) has received sustained attention since it was first reported in 1992, with a number of studies highlighting its importance in embryonic development and cancer.^17,18 During the embryonic stage, ROR1 is highly expressed and plays a key role in nervous and skeletal systems development.¹⁹ However, in adults, ROR1 expressed at very low levels in most normal tissues but abundantly expressed in various tumor tissues, which empowers the potential of ROR1 as a normal tissue-friendly therapeutic target with high specificity.²⁰ The humanized monoclonal antibody (mAb) and cell therapies including the chimeric antigen receptor T cells (CAR-T) and the chimeric antigen receptor NK cells (CAR-NK) that target ROR1 have been used in preclinical studies and clinical trials for hematological malignancies and solid tumors.^21–23 ROR1 also correlates with CSCs and promotes invasion, metastasis of CSCs in ovarian cancer, breast cancer, chronic lymphocytic leukemia, and GBM.^24–27 However, the expression profile of ROR1 in GBM and the role of ROR1 in the maintenance of GSCs remain elusive.

In this study, we identify a novel c-Fos regulatory axis initiated by ROR1 that promotes stemness maintenance and proliferation of GSCs both in vitro and in situ tumor-bearing mice. ROR1 is preferentially expressed in GSCs and promotes GSCs maintenance by stabilizing GRB2 to facilitate the GRB2/ERK/c-Fos signaling. Furthermore, the ROR1-GRB2-c-Fos axis is activated in GBM patients and tracks with GSCs signature. Taken together, our findings uncover a novel GSCs-specific ROR1-GRB2-c-Fos axis that maintains GSCs and highlight the potential of targeting ROR1 in GBM therapy.

Materials and Methods

Cell Lines and Cell Culture

In this study, the primary GSC cell lines (T456, T387, T3359, T3691, T4121) were derived from GBM patients and provided by Dr. Jeremy Rich (UCSD) and Dr. Shideng Bao (Cleveland clinic) as kind gifts. GSCs were incubated at 37°C with 5% CO₂, and cultured in Neurobasal A media (Thermofisher) supplemented with B-27 Supplement (Life), epidermal growth factor (EGF) (20 ng/mL R&D), and basic fibroblast growth factor (bFGF) (20 ng/mL, R&D), L-glutamine and sodium pyruvate. The HEK293T cells and non-stem tumor cells (NSTCs) differentiated from GSCs were cultured in DMEM medium (Gibco) supplemented with 10% fetal bovine serum (FBS) (Gibco) and incubated at 37°C with 5% CO₂.

Animals

The 4-week-old BALB/c nude mice were bought from Beijing Vital River Laboratory Animal Technology Co., Ltd. and utilized for orthotopic xenograft studies. The detailed methods for intracranial orthotopic xenografts have been described previously.²⁸ Briefly, after T3691 and T456 GSCs were infected with lentivirus containing shNT or shRNA and the efficiency of knockdown was confirmed, the cells suspension (containing 2 × 10⁴ GSCs per mouse) were implanted into the right frontal lobes of anesthetized mice at a depth of 3 mm. After intracranial injection, all mice were maintained in the pathogen-free barrier animal facility and observed daily. When mice showed neurological signs or severe weight loss, they were anesthetized and the brain tissues were collected. The brain tissues were fixed with 4% PFA and dehydrated with 30% sucrose, then embedded with OCT compound or paraffin for further investigation. All animal experiments were approved by the University Animal Welfare Committee, Tongji Medical College, Huazhong University of Science and Technology.

Patient Samples

For immunofluorescence staining, fresh surgical specimens were obtained from patients who underwent surgery at Tongji Hospital, Tongji Medical College of Huazhong University of Science and Technology, with the approval of the hospital ethics committee and the informed consent of the patients. The pathological diagnosis of glioma patients was classified according to the 2021 edition of the WHO classification of tumors of the central nervous system.²⁹

For immunohistochemical staining, our previously established 91-tissue glioma tissue microarray³⁰ was used to determine the expression of ROR1, GRB2, and c-Fos in normal brain tissues and glioma patient tissues with different WHO grades.

Immunohistochemical Staining

Immunohistochemical (IHC) staining was performed on our previously established glioma tissue microarray using a DAB staining kit (ZSGB-Bio), the detailed protocol was previously described.³¹ Briefly, the slides were deparaffinized, hydrated, antigen retrieved, and blocked by peroxidase. After sequential incubation with primary and secondary antibodies, the target protein was labeled with DAB solution and the nuclei were counterstained with hematoxylin. The IHC Profiler plugin, an open-source plugin for the quantitative evaluation and automated scoring of immunohistochemistry images of human tissue samples, was used to calculate the positive staining ratio of each sample.³² The following primary antibodies were used for the IHC staining in this study: ROR1 (Santa Cruz, #sc-130386), CD133 (CST, #51917), CD44 (CST, #37259, RRID: AB_2750879), CD15 (CST, #4744S; RRID: AB_1264258), GRB2 (CST, #10254-2-AP, RRID: AB_2263577), c-Fos (Proteintech, #66590-1-Ig, RRID: AB_2881950).

Bioinformatics Analysis

The mRNA expression data along with the IDH1 mutation and 1p19q codeletion status and survival time of glioma patients from the TCGA and CGGA databases were obtained from GlioVis (http://gliovis.bioinfo.cnio.es).³³ The Gene Set Enrichment Analysis (GSEA) was conducted using the GSEA4.1.0 software. R software version 3.6.3 (R Core Team, R Foundation for Statistical Computing) was used in this study. The differentially expressed genes were obtained using the R package DESeq2. R package ClusterProfiler was used to utilize gene set enrichment analysis and visualization.³⁴

Statistical Analysis

GraphPad Prism software (version 8.3.0) was used for statistical analysis. All in vitro experiments were independently repeated at least 3 times. All data were shown as the mean ± SEM unless otherwise specified. Statistical differences between 2 groups were assessed using a 2-tailed unpaired Student’s t-test, while multiple comparisons were performed using 2-way analysis of variance in this study. Kaplan–Meier survival curves were plotted for the animals, and the log-rank test was employed to evaluate the differences in survival time among the groups. In this work, P < .05 was considered as statistically significant.

Results

ROR1 Is Preferentially Expressed by GSCs and Predicts Poor Prognosis of GBM Patients

To elucidate the potential function of ROR1 in GBM progression, we first detected the mRNA and protein expression levels of ROR1 in glioma patient tissues from public databases and our own glioma patient cohort (Tongji cohort).²⁸ Bioinformatic analyses showed that the mRNA expression level of ROR1 increased with glioma WHO grade in CGGA and TCGA databases (Figure 1A, C). We performed IHC staining of tissues from 81 glioma patients and 10 normal brain tissues in the Tongji cohort and calculated the ROR1 positive staining ratio of each tissue using IHCprofiler.³² The results indicated that ROR1 protein level is upregulated in glioma tissues and increased with patient WHO grades, whereas ROR1 is rarely expressed in normal brain tissues, which strongly corroborate the results from databases (Figure 1E, F). Further analyses of GBM patients with different molecular subtypes revealed that ROR1 was highly expressed in GBM patients with IDH-WT and 1p19q non-codeletion in both databases (Figure 1A, C), while these 2 molecular subtypes are associated with poor prognosis of GBM patients. Moreover, high ROR1 expression correlated with poor GBM patient survival in CGGA and TCGA databases (Figure 1B, D).

Figure 1.

ROR1 is preferentially expressed by GSCs and predicts poor prognosis of GBM patients. (A and C) ROR1 mRNA expression in glioma samples with different grades, IDH1 mutation status, and 1p19q codeletion status from the CGGA database (A) and TCGA database (C). (B and D) The Kaplan–Meier survival analysis of ROR1 expression and the overall survival of GBM patients from CGGA database (B) and Gravendeel database (D). (E) IHC staining of ROR1 in glioma tissue microarray, the representative figures in tumor tissues with different grades were shown. Scale bar, 50 μm. (F) The positive staining proportions of ROR1 were quantified using the IHC profile algorithm of the ImageJ software. The relative ROR1 stain (+) of glioma patients with different WHO grades was relative to positive staining proportions of ROR1 in normal brain tissues. (G) Immunofluorescent co-staining of ROR1 (green) and the markers for GSCs (CD133, OLIG2, CD15 and CD44) (red) in GBM patient tissues. Scale Bar, 25 μm. (H) Statistical analysis of ROR1-positive staining in stemness marker+ cells in 3 different human GBM tissues, related to (G). ROR1-positive staining was observed in 76.27%, 58.00%, 64.08%, and 82.14% of CD133⁺, CD15⁺, OLIG2⁺, and CD44⁺ GSCs, respectively. (I) The protein levels of ROR1, OLIG2, SOX2, and GFAP in 5 GBM patient-derived GSCs and matched NSTCs. **P < .01; ***P < .001; ****P < .0001. Data are represented as mean ± SEM.

To further investigate the expression characteristics of ROR1 in GBM, we performed immunofluorescence (IF) staining in frozen tissues from GBM patients and found that the expression of ROR1 was positively associated with the expression of several GSCs markers including CD133, CD15, OLIG2, SOX2, and CD44 (Figure 1G, H and Supplementary Figure 1D). Pearson correlation analysis of 3 different glioma databases also showed a positive correlation between the mRNA expression of ROR1 and stemness markers (PROM1, CD44, and FUT4) (Supplementary Figure 1A–C). To verify the preferential expression of ROR1 in GSCs, we differentiate 5 GBM patient-derived GSCs into non-stem tumor cells (NSTCs) using FBS as we have described previously.²⁸ ROR1 and stem cell markers (SOX2, OLIG2) were preferentially expressed in GSCs, while the astrocyte marker GFAP (examined as the differentiated marker) was significantly upregulated in NSTCs (Figure 1I). Taken together, these results suggest that ROR1 is preferentially expressed in GSCs and predicts poor prognosis in GBM patients.

To investigate the upstream regulatory mechanism of the preferential expression and upregulation of ROR1 in GSCs, we performed an unbiased screening for transcription factors that could regulate ROR1 expression in GSCs. A computational biology framework called Lisa was used to get the top 12 co-enriched transcription factors in GSCs relative to differentiated glioma cells (DGC) and neural stem cell (NSC) (Supplementary Figure 2A).³⁵ By transcription factor binding prediction based on the JASPAR database, we found that the ROR1 promoter binding region contains MYC binding sites, which indicates that MYC may be a regulator of ROR1 expression (Supplementary Figure 2A). To validate the regulatory effect of MYC on ROR1 expression in GSCs, lentivirus-based shRNAs and inhibitors were used to silence MYC expression or inhibit MYC activity in GBM patient-derived GSCs. Both MYC knockdown and inhibition significantly reduced the expression of ROR1, suggesting that ROR1 expression in GSCs was driven by MYC (Supplementary Figure 2B–D). Moreover, since it has been reported that Notch signaling is required for the expression of ROR1 in T98G cells,²⁷ the Notch inhibitor was used to clarify whether Notch signaling is involved in ROR1 expression in patient-derived GSCs. We found that Notch inhibition had no significant effect on the protein and mRNA expression levels of ROR1 (Supplementary Figure 2E, F).

ROR1 Is Critical for GSCs Proliferation and Self-renewal

As ROR1 is preferentially expressed in GSCs, we next investigate the functional significance of ROR1 in stemness maintenance of GSCs. Two distinct shRNAs were used to silence ROR1 expression in GBM patient-derived GSCs (Figure 2A). ROR1 knockdown significantly reduced the proliferation of GSCs, as assessed by the cell-titer-based cell viability assay and the EdU proliferation assay (Figure 2B, E and Supplementary Figure 3A). The in vitro limiting dilution assay and sphereformation assay also showed that disruption of ROR1 inhibits the self-renewal capacity of GSCs (Figure 2C, D and Supplementary Figure 3B–D). Apart from shRNA-based gene knockdown, the sgRNA-based gene knockout was also used and showed similar results (Supplementary Figure 5A–C). Moreover, overexpressing ROR1 in GSCs promotes proliferation and self-renewal of GSCs, as indicated by the in vitro limiting dilution assay, cell viability assay and sphereformation assay (Figure 2F, G, I and Supplementary Figure 3G, H, J, K).

Figure 2.

ROR1 is critical for GSCs proliferation and self-renewal. (A) Western blot analysis of ROR1 in GSCs transduced with lentivirus expressing shNT or shROR1. ROR1 expression in T3359 and T3691 GSCs is efficiently silenced by 2 distinct shRNAs (sh#2024 and sh#2026). (B) Cell viability assay of T3359 and T3691 GSCs with ROR1 knockdown. (C) In vitro limiting dilution assay of T3359 and T3691 GSCs with ROR1 knockdown. (D) Sphereformation analysis of T3359 and T3691 GSCs with ROR1 knockdown (left), and the quantification of sphere numbers in each group (right). Scale bar, 200 μm. (E) The EdU proliferation assay (left) of T3359 and T3691 GSCs with ROR1 knockdown and the quantification of EdU⁺ cells in each group (right). Scale bar, 25 μm. (F) Cell viability assay of T3359 and T3691 GSCs with ROR1 overexpression. (G) Sphereformation analysis of T3359 and T3691 GSCs with ROR1 overexpression (left), and the quantification of relative sphere number (right). Scale bar, 200 μm. (H) The protein levels of ROR1, FLAG, and SOX2 in T3359 and T3691 GSCs with ROR1 overexpression. (I) In vitro limiting dilution assay of T3359 and T3691 GSCs with ROR1 overexpression. (J) The protein levels of ROR1, SOX2, and GFAP in T3359 and T3691 GSCs transduced with lentivirus expressing shNT or shROR1 (left). SOX2 is a GSC marker, and GFAP is an astrocyte marker. The protein levels of ROR1, Cleaved PARP, and Cleaved Caspase3 in T3359 and T3691 GSCs transduced with lentivirus expressing shNT or shROR1 (right). Cleaved PARP and Cleaved Caspase3 are markers for apoptosis. (K) H&E staining of T3691 GSCs-derived tumors in the shNT and shROR1 group 25 days post-injection (top-bottom). Areas with hyperchromatic nuclei are tumors. Scale Bar, 200 μm. Kaplan–Meier survival curve of mice injected with T3691 GSCs expressing shNT or shROR1 (n = 5 per group, half bottom). (L) Immunofluorescent staining of Ki67, Cleaved Caspase3, CD133, SOX2, GFAP, and TUNEL staining in xenografts derived from T3691 GSCs transduced with lentivirus expressing shNT or shROR1 (top-bottom). Scale Bar, 25 μm.Statistical analysis of Ki67, Cleaved Caspase3, TUNEL, CD133, SOX2, and GFAP-positive staining in xenografts (half bottom). **P < .01; ***P < .001; ****P < .0001. Data are represented as mean ± SEM.

To further confirm the essential role of ROR1 in the proliferation and self-renewal of GSCs, we explored the protein levels of apoptosis, stemness and differentiation-related markers in GSCs with ROR1 knockdown, knockout, or overexpression. The apoptosis markers (Cleaved PARP and Cleaved Caspase3) and astrocyte marker (GFAP, examined as the differentiated marker) were increased in GSCs with ROR1 knockdown or knockout (Figure 2J and Supplementary Figures 3E, F and 5D, E). While the expression of stemness marker (SOX2) was reduced in GSCs with ROR1 knockdown or knockout and enhanced in GSCs with ROR1 overexpression (Figure 2H, J and Supplementary Figures 3E, I and 5D). To prove that the shRNA was specific and there were no off-target effects, a rescue experiment using shRNA resistant form of the ROR1 gene (with target site mutated) was performed, and the cell viability assay, in vitro limiting dilution assay, sphereformation assay, and western blot showed that the self-renewal capacity of GSCs restored after ROR1 transduction (Supplementary Figure 7A–D). These data indicate that ROR1 is critical for GSCs proliferation and self-renewal.

Disrupting ROR1 Suppressed GSCs Tumor Growth and Prolonged Survival of Tumor-bearing Mice

To assess the biological significance of ROR1 expression on GSC tumorigenic potential in vivo, we generated the in situ GSCs-derived GBM mouse model using GSCs transduced with lentivirus expressing shROR1 or sgROR1 as we described previously.²⁸ When mice develop neurological signs, we anesthetized, executed, and collected the mouse brain tissues. ROR1 knockdown or knockout markedly suppressed the in vivo tumor growth of both T3691 and T456 GSCs as shown by the H&E staining of mouse brain tissues (Figure 2K and Supplementary Figures 4A and 5F). The survival analysis also indicates that disrupting ROR1 significantly prolonged the survival time of tumor-bearing mice (Figure 2K and Supplementary Figures 4B and 5G). Moreover, we utilize Doxycycline (DOX) inducible knockdown system to knockdown the ROR1 expression in tumor-bearing mice as described previously,³⁶ the results showed that DOX-inducible knockdown of ROR1 suppressed GSCs tumor growth and prolonged the survival time of tumor-bearing mice (Supplementary Figure 6A–C). Next, we verified the function of ROR1 on in vivo tumor growth of GSCs using frozen section staining of T3691 and T456 xenografts. IF staining of ki67 and Cleaved Caspase3, and TUNEL staining showed that the xenografts from GSCs with ROR1 knockdown expressed fewer Ki67 and more Cleaved Caspase3, and contained more TUNEL-positive cells (Figure 2L and Supplementary Figure 4C, E). In addition, the expression of stemness markers (CD133 and SOX2) decreased, whereas the astrocyte marker (GFAP, examined as the differentiated marker) increased in the xenografts from both T3691 and T456 GSCs expressing shROR1 (Figure 2L and Supplementary Figure 4D, F). Collectively, these results suggest that ROR1 plays a critical role in the tumorigenic capacity of GSCs in vivo.

ROR1 Promotes GRB2/ERK/c-Fos Signaling in GSCs

Because ROR1 plays a critical role in the proliferation and self-renewal of GSCs, we next want to unravel the mechanisms by which ROR1 maintains GSCs. We utilized RNA sequencing (RNA-seq)-based transcriptome sequencing of GSCs transduced with lentivirus expressing shNT or shROR1 and performed pathway enrichment analysis of the differentially expressed genes between shNT and shROR1. The analysis showed that ROR1 correlates with several pathways that have been reported to play key roles in cancer progression, including “PI3K-Akt signaling pathway,” “Pathways in cancer,” “Hippo signaling pathway,” and “MAPK signaling pathway” (Figure 3A). Among them, the MAPK signaling pathway was identified as the term with the highest rich factor and smallest P-value (Figure 3A). Furthermore, the GSEA of glioma patients in the CGGA database suggests that the mitogen-activated protein kinases (MAPK) pathway is activated in patients with higher ROR1 expression (Figure 3B). These results indicate that ROR1 may promote the maintenance of GSCs by activating the MAPK signaling pathway.

Figure 3.

ROR1 promotes GRB2/ERK/c-Fos signaling in GSCs. (A) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of differentially expressed genes between GSCs transduced with lentivirus expressing shNT or shROR1. (B) GSEA for MAPK pathway in ROR1-high compared to ROR1-low patients in CGGA database. NES, normalized enrichment score; FDR, false discovery rate. (C) Western blot analysis of ROR1, GRB2, KRAS, p-C-RAF, p-ERK, ERK, and c-Fos expression in T3359 and T456 GSCs transduced with lentivirus expressing shNT or shROR1. (D) Western blot analysis of ROR1, GRB2, KRAS, p-C-RAF, p-ERK, ERK, and c-Fos expression in T3359 and T456 GSCs with ROR1 overexpression. (E) GSEA for KEGG cell cycle pathway in ROR1-high comparing to ROR1-low (upper) and FOS-high comparing to FOS-low (bottom) patients in CGGA database. NES, normalized enrichment score; FDR, false discovery rate. (F) The protein levels of ROR1, c-Fos, p21, CCND1, CDK2, and CDK6 in T3359 and T456 GSCs transduced with lentivirus expressing shNT or shROR1. p21, CCND1, CDK2, and CDK6 are cell cycle related proteins. (G) Western blot analysis of ROR1, c-Fos, p21, CCND1, CDK2, and CDK6 expression in T3359 and T456 GSCs with ROR1 overexpression. p21, CCND1, CDK2, and CDK6 are key regulators of cell cycle. (H) Immunofluorescent co-staining of ROR1 (red) and c-Fos, CCND1, p21, CDK2, and CDK6 (green) in xenografts derived from T3691 GSCs transduced with lentivirus expressing shNT or shROR1. Scale Bar, 25 μm. The expression of c-Fos, CCND1, CDK2, and CDK6 decreased, whereas the p21 increased with ROR1 knockdown. (I) Statistical analysis of c-Fos, CCND1, p21, CDK2, and CDK6-positive staining in xenografts of (H). **P < .01; ***P < .001. Data are represented as mean ± SEM.

The MAPK pathway is hyperactivated and plays a central role in human cancers³⁷. The activation of MAPK pathway participates in GBM proliferation, invasion, metastasis, tumor stemness, and treatment resistance.³⁸ To clarify the detailed regulation mechanism of ROR1 on the MAPK pathway, the immunoblotting analysis was performed to detect the expression of key components of this pathway. Disruption of ROR1 significantly reduced the expression/phosphorylation level of GRB2, KRAS, c-RAF, and ERK, which ultimately inhibits the expression of c-Fos (Figure 3C and Supplementary Figures 8B and 9A). In addition, overexpression of ROR1 activates the GRB2-ERK pathway and promotes c-Fos expression (Figure 3D and Supplementary Figure 8A). Because c-Fos is widely involved in tumor cell proliferation, migration, and treatment resistance which is largely based on its regulation of cell cycle,³⁹ we next wonder if ROR1 facilitates GSCs maintenance through c-Fos-mediated cell cycle regulation. GSEA analysis showed that the cell cycle pathway is activated both in patients with higher ROR1/FOS expression (Figure 3E). Consistent with this in silico data, the protein levels of key positive cell cycle regulators (CCND1, CDK2, and CDK6) were decreased after ROR1 knockdown or knockout in GSCs, whereas p21, which function as a negative cell cycle regulator, was increased (Figure 3F and Supplementary Figures 8A and 9B–C). Moreover, the rescue experiment using shRNA resistant form of the ROR1 gene (with target site mutated) showed that the protein expression of these markers in GSCs restored after ROR1 transduction (Supplementary Figure 7D). As expected, ROR1 overexpression upregulates the protein expression of CCND1, CDK2, and CDK6, and downregulates the p21 (Figure 3G). Importantly, the regulation of ROR1 on the c-Fos-cell cycle axis was further confirmed by the IF staining of GSCs-derived xenografts and GBM patient tissues (Figure 3H, I and Supplementary Figure 8C–E). Taken together, these findings suggest that ROR1 promotes GRB2/ERK/c-Fos signaling and cell cycle in GSCs.

ROR1 Binds to GRB2 and Reduces GRB2 Lysosomal Degradation

GRB2 is an adaptor protein that acts as an intermediate between cell-surface receptors and downstream signaling cascade.⁴⁰ The dysregulation of GRB2 has been discovered in various tumor progression.⁴⁰ GRB2 activates the downstream RAS-RAF-ERK cascade to promote malignant behaviors of tumor cells through binding with several phosphor-tyrosine residues including multiple RTKs on the cell membrane.^41,42 Because the intracellular domain of ROR1 consists of a tyrosine kinase-like domain, we hypothesized that GRB2 may act as an adaptor protein to mediate the activation of signaling pathways downstream of ROR1 in GSCs. However, it is unclear whether ROR1 promotes downstream signaling pathways to maintain GSCs through direct binding to GRB2. To shed light on this, we using lentivirus-based shRNAs to silencing ROR1 and GRB2 in GSCs. Disruption of ROR1 reduces the protein level of GRB2, whereas the mRNA level of GRB2 did not change significantly after ROR1 silencing (Figure 4A, D). Moreover, GRB2 knockdown also results in decreased ROR1 protein expression (Figure 4B), and ROR1 overexpression increases the GRB2 protein level in GSCs (Figure 4C). These data suggest that ROR1 may regulates GRB2 through protein interactions.

Figure 4.

ROR1 binds to GRB2 and reduces GRB2 lysosomal degradation. (A) The protein levels of ROR1 and GRB2 in T3359 and T456 GSCs transduced with lentivirus expressing shNT or shROR1. (B) The protein levels of GRB2 and ROR1 in T3359 and T456 GSCs transduced with lentivirus expressing shNT or shGRB2. (C) Western blot analysis of ROR1 and GRB2 expression in T3359 and T456 GSCs with ROR1 overexpression. (D) qRT-PCR analysis of ROR1 and GRB2 expression in T3359 and T456 GSCs transduced with lentivirus expressing shNT or shROR1. (E) Co-IP analysis in T3359 and T456 GSCs. Cell lysates were immunoprecipitated with anti-ROR1/GRB2 antibodies followed by immunoblotting with anti-ROR1 and anti-GRB2 antibodies. (F) Immunofluorescent co-staining of ROR1 (green) and GRB2 (red) in GBM patient tissues. Scale Bar, 25 μm. (G) The protein levels of ROR1 and GRB2 in T3359 and T456 GSCs transduced with lentivirus expressing shNT or shROR1 and then treated with chloroquine (Cq). (H) The protein levels of ROR1 and GRB2 in T3359 and T456 GSCs transduced with lentivirus expressing shNT or shROR1 and then treated with MG132. (I) Immunofluorescent co-staining of GRB2 (green) and the lysosomal marker LAMP1 (red) in GSCs transduced with lentivirus expressing shNT or shROR1 and then treated with DMSO (ctrl) or chloroquine. (J) Immunofluorescent co-staining of ROR1 (green) and GRB2 (red) in xenografts derived from T3691 GSCs transduced with lentivirus expressing shNT or shROR1. Scale Bar, 25 μm. (K) The quantification of the percentage of GRB2 co-localized with LAMP1. (L) Statistical analysis of ROR1 and GRB2 positive staining in xenografts of (J). **P < .01; ****P < .0001. Data are represented as mean ± SEM.

To further verify the interactions between ROR1 and GRB2, we performed reciprocal co-immunoprecipitation (co-IP) assay in GSCs and found that ROR1 interacts with GRB2 each other endogenously (Figure 4E). The IF co-staining of ROR1 and GRB2 in GBM patient tissues and GSCs-derived xenografts indicates that ROR1 and GRB2 co-localized at the cell membrane (Figure 4F, J and Supplementary Figure 10A–C). Moreover, the staining of GRB2 was decreased in xenografts derived from T3691 and T456 GSCs with ROR1 knockdown (Figure 4J, L and Supplementary Figure 10A, C). Collectively, these results demonstrate the interaction and co-expression of ROR1 and GRB2 in GSCs and GBM patient tissues. Given that the knockdown of ROR1 did not alter GRB2 mRNA levels, we hypothesized that ROR1 binds to GRB2 to inhibit its degradation. To elucidate the detailed mechanism by which ROR1 binds and promotes GRB2, we used chloroquine (Cq, inhibitor of lysosomal degradation pathway) and MG132 (inhibitor of proteasome degradation pathway) to treat GSCs with ROR1 knockdown. Interestingly, treatment of GSCs with Cq partially rescued the GRB2 protein levels upon ROR1 knockdown (Figure 4G), whereas MG132 treatment had no significant effect on GRB2 protein expression in GSCs with ROR1 knockdown (Figure 4H). In agreement with these immunoblotting results, the lysosomal distribution of GRB2 increased when ROR1 was knockdown, while Cq treatment blocks this process, as shown by IF co-staining of GRB2 and LAMP1 (lysosome marker) in GSCs (Figure 4I, K). Taken together, these findings confirm that ROR1 binds to GRB2 and reduces GRB2 lysosomal degradation, which ultimately facilitating the downstream signaling pathway of GRB2.

c-Fos Overexpression Rescues the Effects Caused by ROR1 Knockdown

Our previous findings suggest that ROR1 promotes GSCs maintenance by stabilizing GRB2 and facilitating the GRB2-ERK-c-Fos signaling axis. Due to the important role of c-Fos in GBM progression which functions as a transcription factor, we then investigated whether overexpression of c-Fos after ROR1 knockdown could rescue the effects caused by ROR1 silencing in GSCs. The FOS overexpression plasmid was constructed and delivered into GSCs through lentivirus, and the efficiency of overexpression in T3691 and T456 GSCs was validated by qRT-PCR and immunoblots (Supplementary Figure 11A, B). As expected, in GSCs with ROR1 knockdown, reduced tumorsphere number and cell viability were partially rescued when FOS was ectopically overexpressed (Figure 5A–C), at the same time, the protein levels of CCND1, CDK2, and CDK6 were also partially rescued while the apoptosis markers (Cleaved PARP and Cleaved Caspase3) were decreased (Figure 5E). These results indicate that c-Fos overexpression rescues the effects caused by ROR1 knockdown in vitro.

Figure 5.

c-Fos overexpression rescues the effects caused by ROR1 knockdown. (A) Sphereformation analysis of T3359 and T456 GSCs with ROR1 knockdown, with/without FOS overexpression. Scale bar, 200 μm. (B) The quantification of relative sphere number in (A). (C) Cell viability assay of T3359 and T456 GSCs with ROR1 knockdown, with/without FOS overexpression. (D) H&E staining of T3691 GSCs-derived tumors 25 days after transplantation. GSCs with ROR1 knockdown and with/without lentivirus-mediated FOS overexpression were used for intracranial injection. Areas with hyperchromatic nuclei are tumors. Scale Bar, 200 μm. (E) Western blot analysis of ROR1, c-Fos, cell cycle regulators (CCND1, CDK2, and CDK6), and apoptosis markers (Cleaved PARP and Cleaved Caspase3) in T3691 and T456 GSCs with ROR1 knockdown and with/without lentivirus-mediated FOS overexpression. (F) Kaplan–Meier survival curve of mice injected with T3691 GSCs with ROR1 knockdown and with/without lentivirus-mediated FOS overexpression (n = 5 per group). (G) Immunofluorescent co-staining of ROR1 and Ki67, Cleaved Caspase3 in T3691 GSCs-derived xenografts with ROR1 knockdown and with/without lentivirus-mediated FOS overexpression. Scale Bar, 25 μm. (H) Statistical analysis of Ki67 and Cleaved Caspase3 positive staining in xenografts of (G). *P < .05; **P < .01; ***P < .001; ****P < .0001. Data are represented as mean ± SEM.

The effects of c-Fos overexpression after ROR1 silencing on in vivo tumor growth were also investigated. GSCs transduced with lentivirus expressing shNT/shROR1 and vector/FLAG-FOS were injected into the brains of nude mice. Consistent with the effects on GSCs in vitro, c-Fos overexpression in mice with ROR1 knockdown (shROR1 + FLAG-FOS) accelerated tumor growth and shortened the survival time of mice as indicated by the H&E staining and survival analysis of mice in each group (Figure 5D, F). Moreover, the IF co-staining showed accumulated Ki67 staining and less Cleaved Caspase3 staining in xenografts from mice with ROR1 silencing and c-Fos overexpression (Figure 5G, H). Together, these results demonstrated that ectopic expression of FOS can reverse the impairment of GSCs maintenance and tumor growth caused by the disruption of ROR1, which further supports the critical role of the GSCs-specific ROR1-GRB2-ERK-c-Fos axis in proliferation and self-renewal of GSCs.

ROR1-GRB2-c-Fos Axis Is Activated in GBM Patients and Tracks With GSCs Signature

To elucidate the clinical relevance of the ROR1-GRB2-c-Fos axis in GSCs of GBM patients, we utilized the multivariable Cox regression analysis to establish a prognostic model for predicting stemness index based on the expression of 4 stemness-related genes (PROM1, MYC, CD44, and FUT4) as we have described previously.³⁰ The Concordance Index of this model is 0.66 (Supplementary Figure 12A). The stemness value was calculated for each patient in the CGGA database, and patients were divided into stemness-low and stemness-high groups using the median value of the stemness value as the cutoff. With the increase of stemness value, the expression of 4 stemness-related genes increased, and the OS of patients gradually decreased (Supplementary Figure 12B). Moreover, this stemness-based model exhibited excellent performance on survival prediction, the area under the curve (AUC) of 1- and 3-year OS are 0.67 and 0.72 (Supplementary Figure 12C).

Further analysis showed that the stemness value increased with tumor grade, and a higher stemness value indicates a worse prognosis for glioma patients (Figure 6A, B). The mRNA levels of ROR1, GRB2, and FOS were higher in stemness-high groups (Figure 6C). Moreover, the GSEA of glioma patients in the CGGA database suggests that the stemness signature is activated in patients with higher ROR1/GRB2/FOS expression (Figure 6D–F). Corresponding with these in silico results, the protein levels of GRB2 and c-Fos are upregulated in glioma tissues and increased with patient WHO grades (Supplementary Figure 13A–D), and the expression of ROR1/GRB2/c-Fos was positively correlated with stemness-related markers (CD133, CD44, and CD15) as shown by IHC staining of tissue microarray (TMA) (Figure 6G, H). Collectively, these correlative human GBM findings suggest that the ROR1-GRB2-c-Fos axis is activated in GBM patients and tracks with GSCs signature, which highlights the critical role and therapeutic significance of this signaling axis in GSCs maintenance and GBM progression.

Figure 6.

ROR1-GRB2-c-Fos axis is activated in GBM patients and tracks with GSCs signature. (A) The Kaplan–Meier survival analysis of stemness value and the overall survival of glioma patients from the CGGA database. (B) The stemness value of glioma patients with different WHO grades, patients are from the CGGA database. (C) The mRNA expression of ROR1, GRB2, and FOS in patients with different stemness values. (D) GSEA for stemness signature in ROR1-high compared to ROR1-low patients in CGGA database. NES, normalized enrichment score; FDR, false discovery rate. (E) GSEA for stemness signature in GRB2-high compared to GRB2-low patients in CGGA database. NES, normalized enrichment score; FDR, false discovery rate. (F) GSEA for stemness signature in FOS-high compared to FOS-low patients in CGGA database. NES, normalized enrichment score; FDR, false discovery rate. (G) IHC staining for ROR1, GRB2, c-Fos, CD133, CD44, and CD15 in human glioma tissue microarray (Tongji cohort). (H) The relative positive staining ratios of proteins in (G) were quantified using the IHC profile algorithm of the ImageJ software and the Pearson’s correlation test was used to determine the correlations between these 6 proteins (r and P values are shown). ****P < .0001. Data are represented as mean ± SEM.

Discussion

ROR1 is a transmembrane protein belonging to the RTK family.¹⁸ It is evolutionarily conserved and was termed “orphan receptors” since its ligand was unknown at the time of its discovery.¹⁸ The molecular structure of ROR1 comprises an extracellular domain and an intracellular domain.⁴³ The extracellular domain composed of an immunoglobulin-like domain, a cysteine-rich domain, and a kringle domain.⁴³ The intracellular domain consists of a tyrosine kinase-like domain and serine/threonine segments that flank a proline-rich domain.⁴³ In addition to Wnt5a and IGFBP5, which can act as ligands for ROR1, it has been reported that ROR1 can regulates downstream signaling transduction independently of ligands.^25,43,44 ROR1 is highly expressed in various tumors including GBM, and is expressed at extremely low levels in normal adult brain tissue, which suggests that ROR1 could be a potential target for GBM treatment.²⁰ ROR1-targeted therapies, including CAR-T cell and monoclonal antibodies, have been developed and utilized in the treatment of hematological malignancies and solid tumors.²¹ However, the expression characteristics and mechanism of ROR1 in GBM are still unclear, which largely limits the application and efficacy of ROR1-targeted therapies in the treatment of GBM. In this study, we demonstrate that ROR1 was upregulated in GBM patients and preferentially expressed in GSCs. ROR1 facilitates the proliferation and self-renewal of GSCs through GRB2/ERK/c-Fos signaling axis. Moreover, the ROR1-GRB2-c-Fos axis is activated in GBM patients and tracks with GSCs signature, suggesting that targeting ROR1 and this signaling axis is a promising therapeutic strategy for GSCs-targeted GBM treatment. Notably, we found that ROR1 knockdown significantly reduced the cell viability of GSCs in vitro, particularly in T3359 GSCs, where cell viability was consistently reduced in the ROR1 knockdown group. However, ROR1-knockdown GSCs were still able to form intracranial tumors in nude mice, suggesting that the growth conditions of GSCs differ between in vivo and in vitro. We think that the reason for this difference may be the presence of multiple microenvironmental components including TAMs and TANs within the GBM tumor microenvironment in vivo, which play important roles in the stemness maintenance and tumor growth of GSCs through mechanisms such as paracrine secretion.

GRB2 is a 25 kDa adaptor protein that comprises one Src homology 2 (SH2) domain and 2 Src homology 3 (SH3) domains.⁴⁵ GRB2 recruits several phosphor-tyrosine residues including RTKs (eg, platelet-derived growth factor receptor and epidermal growth factor receptor) and non-RTKs (eg, insulin receptor substrate 1 and focal adhesion kinase), which activates the extracellular signaling pathways to regulate multiple cellular processes.⁴⁵ The upregulation of GRB2 has been found in a variety of tumors including breast cancer, chronic myelogenous leukemia, hepatocellular carcinoma, bladder cancer, and lung cancer.⁴⁰ GRB2 antagonist peptides targeting the SH2 domain and SH3 domain have been investigated and serve as novel strategies for cancer therapies.⁴⁰ However, the expression pattern and function of GRB2 in GBM are less reported. In this study, we performed IHC staining of glioma patients TMA of Tongji cohort and found that the expression of GRB2 was upregulated in GBM and increased with glioma grade. GRB2 was stabilized by ROR1 through inhibition of lysosomal degradation in GSCs. Elevated GRB2 level facilitates its downstream ERK/c-Fos pathway, which was indispensable for GSCs self-renewal and GBM growth. Our findings provide new insights into the function of GRB2 in GSCs maintenance and GBM progression.

The function and regulator mechanisms of c-Fos in GSCs maintenance and GBM growth were also investigated in this study. We demonstrated that c-Fos acts as a transcription factor at the end of the ROR1/GRB2/ERK/c-Fos signaling axis, and promotes the proliferation of GSCs through cell cycle regulation. The ectopic expression of FOS enhances the proliferation and self-renewal capacity of GSCs, and reverses the effect of ROR1 knockdown in GSCs both in vitro and in vivo. Moreover, the IHC staining of glioma patients TMA of the Tongji cohort showed that the protein level of c-Fos was upregulated in GBM and increased with glioma grade. Collectively, these data indicate the significant role of c-Fos in GSCs maintenance and the pro-tumoral signaling pathway initiated by ROR1. TFs were historically classified as “undruggable” since most TFs do not have defined ligand-binding pockets typically.⁴⁶ However, with the deeper understanding of the biomedical properties and structure of TFs, a considerable number of small molecules targeting TFs have been identified or synthesized, and have been approved for clinical trials.⁴⁶ In this case, the critical function of c-Fos in GSCs maintenance and GBM growth revealed by this study highlights the potential of c-Fos as an attractive target for GBM treatment.

As the most common brain malignant tumor, GBM was hallmarked by multi-dimensional heterogeneity.⁴⁷ The inter-tumor heterogeneity between different GBM patients, the intra-tumor heterogeneity, and the differences between cells, animal models, and patient tumors contribute to the formation of refractory GBM.⁴⁷ To assess the therapeutic significance of the ROR1-GRB2-c-Fos axis in GBM patients, we performed multiple bioinformatic analyses of 970 glioma patients in the CGGA database and IHC staining of 91 patient tissues in the Tongji cohort. The results suggest that the ROR1-GRB2-c-Fos axis is activated in GBM patients and tracks with GSCs signature, which further provides a theoretical basis for the development of therapeutic strategies targeting this axis.

In conclusion, this study identifies that ROR1 is preferentially expressed in GSCs and is indispensable for GSCs proliferation, self-renewal, and in vitro tumor growth. ROR1 activates GRB2/c-Fos signaling to promote tumorigenicity of GSCs through binding to GRB2 and stabilizes GRB2 by reducing its lysosome degradation. In addition, c-Fos overexpression rescues the effects caused by ROR1 knockdown in GSCs, and the ROR1-GRB2-c-Fos axis is activated in GBM patients and tracks with GSCs signature. These results indicate that the ROR1-GRB2-c-Fos axis may be a novel critical paradigm in GSCs maintenance and an attractive therapeutic target in GBM treatment.

Supplementary material

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

Conflict of interest statement

The authors declare no competing interests.

Funding

This work was supported by grants from the National Key R&D Program of China, MOST (2023YFC2510000), the National Natural Science Foundation of China (82072805, 81974452, 82403476), the Hubei Association for Science and Technology Young Talents Support Projects (2024), the Hubei Natural Science Foundation (2020CFB678, 2023AFB135), the China Postdoctoral Science Foundation (2022M711253), Wuhan Science and Technology Major Project (2021022002023426), Huazhong University of Science and Technology Independent Innovation Research Fund Project (2019kfyXJJS187), the postdoctoral innovation research position funding of Hubei province and the Program for the HUST Academic Frontier Youth Team.

Author contributions

Conceptualization, H.Z., L.C., X.Y., and K.S.; methodology, validation, data acquisition, and visualization, H.Z., L.C., D.L., X.M., Z.C., H.F., R.L., Y.Z., H.M., and J.L.; manuscript writing and editing, H.Z., L.C., X.Y., and K.S.; funding acquisition, H.Z., S.Z., X.Y., and K.S.; supervision, S.Z., X.Y., and K.S.; all authors reviewed and approved the final manuscript.

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary information files.