Oligonucleotide G3T19 downregulates CD44 succinylation and inhibits astrocyte-induced glioma immune escape

Abstract

Background

Astrocytes (ACs) play a crucial role within the immune microenvironment of glioma, facilitating its progression. Protein succinylation, a post-translational modification, may serve as a significant regulatory mechanism. SsDNA (single-chain oligonucleotide) drugs with specific sequences for the treatment of tumors have entered clinical trials with remarkable effects.

Methods

G3T19 was evaluated by peroxidase-like activity assay and circular dichroism experiments. The effects of G3T19 on glioma were evaluated by flow cytometry, CCK-8, transwell migration assay, immunohistochemistry, small animal imaging, HPLC, orthotopic xenograft glioma model and human glioma organoid model. The binding of G3T19 and CD44 was examined by immunoprecipitation assay, ELONA assay, immunofluorescence, and docking conformation visualization. The effect of G3T19 on CD44 succinylation was evaluated by mass spectrometry analysis, immunoprecipitation assay, and flow cytometry.

Results

G3T19 with G-quadruplex parallel structure binds to CD44 and inhibits AC-enhanced glioma immune escape. CD44 succinylation in ACs adjacent to gliomas is enhanced at lysine 40; however, G3T19 reduces CD44 succinylation to inhibit AC-enhanced glioma immune escape. G3T19-dependent targeting of CD44 reduces NF-κB signaling, PD-L1 expression, and AC-induced glioma immune escape.

Conclusion

These results confirm that G3T19 binds CD44 and downregulates CD44 succinylation, inhibits NF-κB pathway and the expression of PD-L1, inhibits AC-induced glioma immune escape. This study identifies a new anti-tumor mechanism based on the advanced structure of a functional nucleic acid that may be useful in the development of drugs for glioma immunotherapy.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12964-026-03081-x.

Background

With the failure of many traditional treatment methods, immunotherapy has brought hope to glioma patients, among which the most widely studied immunotherapy is Programmed death receptor 1 and Programmed death ligand 1 (PD-1/PD-L1) immune checkpoint inhibitors [1–3]. Tumor cells overexpress PD-L1 to escape immune killing. PD-L1 on the surface of tumor cells binds to PD-1 on the surface of T cells, thus reducing the killing activity of T cells and allowing tumor cells to escape immune clearance. Though PD-1 and PD-L1 inhibitors have achieved good results as immunotherapy for various tumors [4–7], their efficacy in the treatment of gliomas has been less consistent.

High expression of CD44, a multifunctional cell surface adhesion receptor, significantly enhances tumor development [8–10]. In glioma, CD44 expression has been shown to correlate positively with the degree of malignancy and negatively with the survival rate [11, 12]. CD44 signaling is stimulated by its ligand, hyaluronic acid, which is a major extracellular matrix component of brain tissue and is notably upregulated in gliomas [13–15]. High expression of CD44 can promote the activation of NF-κB, which translocates into the nucleus and binds to the promoters of PD-L1 and CD44 to promote their transcription [16, 17].

In recent years, targeting the immune microenvironment of glioma has become an important goal for anti-tumor therapy. Astrocytes (ACs) are the most abundant and complex stromal cells in the glioma microenvironment, and they play an important role in the malignant progression of gliomas [18–20]. It has been confirmed that IL-6, IL-1β and other inflammatory factors are highly expressed in reactive ACs that surround glioma [21, 22]. Consequently, in the immune microenvironment of glioma, ACs may enhance immune escape and further promote malignant progression.

The crosstalk between ACs and Glioma Cells (GCs) constitutes an important process for targeting in glioma therapy. In previous research, we demonstrated that a novel nano-drug blocks the crosstalk between ACs and GCs by inhibiting the IL-6/STAT3 pathway, thereby inhibiting the growth of gliomas [23]. Additionally, we found that the single-chain oligonucleotide (ssDNA) AS1411 binds to nucleolin in reactive ACs through its G-quadruplex (G4) parallel structure, down-regulating exosome miRNA-27a secreted by reactive ACs and up-regulating the expression of the tumor suppressor INPP4B, thereby inhibiting the proliferation of GCs [24]. Single-stranded oligonucleotide (ssDNA) drugs, particularly those forming G-quadruplex (G4) structures, have attracted considerable interest for cancer therapy due to their favorable properties, including low immunogenicity, high stability, ease of chemical modification, and good tissue penetration [25, 26]. Dozens of G4-forming ssDNA drugs have entered clinical trials for various tumors [27, 28]. A unique feature of parallel G4 structures is their ability to bind hemin and form a G4/hemin DNAzyme with peroxidase-like activity [29, 30]. This catalytic activity provides a convenient colorimetric readout that can be used to rapidly assess the formation and stability of parallel G4 structures [31]. In this study, we therefore exploited this peroxidase activity as a screening tool to identify novel G4-forming ssDNAs with potential anti-glioma activity.

Post-translational protein modification, such as succinylation, acetylation, crotonylation, or lactylation, is a chemical change that may occur on proteins after translation. Post-translational modifications can change the physical and chemical properties of proteins, affect the spatial conformation and stability, participate in the normal life process of cells, and play an important role in the occurrence and development of tumors [32, 33]. Succinylation is the process by which succinyl groups are covalently bonded to amino acid residues, mainly to lysine residues. Due to the large molecular weight of the succinyl group, succinylation can cause changes in the electric charge, resulting in dramatic changes in physical and chemical properties [34, 35]. Studies have shown that down-regulation of the expression of succinyltransferase KAT2A inhibits the proliferation of GCs by reducing the succinylation level of H3K79 [36]. Therefore, succinylation is crucial for the malignant progression of glioma. Previous studies have targeted CD44 in glioma using antibodies, hyaluronidase, or small interfering RNAs, and have modulated astrocyte- glioma interactions via various approaches [11, 12, 23, 24]. However, whether CD44 undergoes succinylation in reactive astrocytes and whether this modification can be therapeutically targeted remains unexplored.

In this study, we sought to characterize additional ssDNA drugs with G4 structure that can target diverse signaling pathways in glioma. Therefore, we performed screening assays, which led to the identification of G3T19. We found that CD44 succinylation is enhanced in ACs that adjacent to glioma, leading to NF-κB signaling pathway activation and up-regulation of PD-L1 expression. We also demonstrated that G3T19 binds and inhibits CD44 succinylation via its G4 parallel structure, which inhibits NF-κB signaling pathway activation, down-regulates the transcription and expression of PD-L1 and CD44, and inhibits glioma immune escape induced by ACs. Our study provides increased understanding of the immune escape mechanisms in glioma, including the identification of a potential new drug target, as well as providing an effective new approach for the treatment of glioma.

Methods

Cell culture

Glioma cell lines (U87 MG, U251 and GL261-luc), Human normal astrocytes (HA) and Jurkat T lymphoma cells were obtained from the China Center for Type Culture Collection (CCTCC, Wuhan, China). Normal primary ACs and CD44 gene knockout ACs were derived from the cerebral cortex of 3-day-old C57BL/6 or C57BL/6 gene knockout mice, respectively. T lymphocytes were activated with immune factors, including anti-CD3 and anti-CD28 antibody, prior to processing. The cells were cultured in DMEM/F12 medium with 10% fetal bovine serum (Cat# mu001SR, QmSuero/Tsingmu Biotechnology, Wuhan, China) and 1% antimycotic and maintained in an incubator at 37 °C under a 5% CO₂ atmosphere. Authentication of all cell lines was performed using DNA fingerprinting, and they were confirmed to be mycoplasma-free. The human organoids used in this study were commercially obtained from D1 Medical Technology (Shanghai, China). The biological source and collection process of the original human tissue samples was reviewed and approved by the Institutional Review Boards of Fudan University Shanghai Cancer Center (20241038-2). The studies were conducted in accordance with recognized ethical guidelines (Declaration of Helsinki). Informed consent was obtained from all of the participants. Organoids were maintained in commercial glioma organoid amplification medium (D1 Medical Technology, K211M16).

Assay for G-quadruplex/hemin DNAzyme peroxidase-like activity

The principle and catalytic system of G-quadruplex/hemin DNAzyme as previously described [37]. In brief, G-quadruplex sequences (100 nM) were incubated with 50 mM Tris (pH 7.5), 150 mM NH₄Cl, 1 µM hemin and 0.0003% Triton X-100 at 42 °C for 10 min. Subsequently, H₂O₂ and 2,2′-azino-bis 3-ethylbenzthiazoline-6-sulfonic acid (ABTS) were added. The absorbance of the catalytic product was detected at 414 nm using an Epoch™ spectrophotometer. The most active sequence (G3T19) was identified, and its complementary sequence (Controlled Random Oligonucleotides, CRO) was used as a control. The nucleic acid sequences used in the experiments are shown in Table 1.

Table 1.

Oligonucleotide sequence

Oligonucleotide	Sequence (5′-3′)
G4	GGGGGGGGGGGGGGGG
G3A	GGGAGGGAGGGAGGGA
G3G	GGGGGGGGGGGGGGGG
G3C	GGGCGGGCGGGCGGGC
G3T	GGGTGGGTGGGTGGGT
G3T19	CTGGGTGGGTGGGTGGGTC
CRO	TACTCATCCACGCACTCAG
AS1411	GGTGGTGGTGGTTGTGGTGGTGGTGG
[B7]-3-0	ATTGGGAGGGATTGGGTGGG
CatG4	TGGGTAGGGCGGGTTGGGAAA
EAD2	CTGGGAGGGAGGGAGGGA
WY-A6	TGGGTTGGGTTGGGTGGGT
PS2.M	GTGGGTAGGGCGGGTTGG

To evaluate the tissue distribution of G3T19, mice were injected with biotin-labeled G3T19 and euthanized after 24 h. Major organs were homogenized and ultrasonically lysed to release the oligonucleotide. The lysate supernatant was incubated with streptavidin-coated beads, followed by washing and resuspension. The beads were then subjected to the peroxidase-colorimetric reaction as described in above and absorbance was measured at 414 nm. A standard curve was generated with known concentrations of biotin-labeled G3T19 (0-200 nM) processed identically, enabling quantification of G3T19 in tissue samples.

Circular dichroism experiments

G3T19 or CRO (1 µM) was incubated with 50 mM Tris, pH 7.5, and 10 mM KCl at 45 °C for 10 min. The sample was transferred to a quartz dish with a path length of 1 cm and observed using a jasco-810 spectropolarimeter. The corresponding parameters are as previously described [24].

High-performance liquid chromatography

Chromatography was performed under the following conditions: C18 column (4.6 mm×250 mm, 5 μm), column temperature 30 °C, flow rate 0.8 mL/min, injection volume 10 µL, and detection wavelengths 260 nm and 414 nm. The Mobile Phase A was acetonitrile and the mobile phase B was 20 mM KH₂PO₄. The solvent reaction was activated by adding 100 µM G3T19 into the sample buffer containing 50 mM Tris (pH 7.5), 150 mM NH₄Cl, 1 µM hemin, and 0.0003% Triton X-100 in a water bath at 45 °C for 10 min, followed by injection of the sample into a glass bottle.

EdU cell proliferation assay

Cell proliferation assays were performed according to the protocol in the EdU Cell Proliferation Assay Kit (Beyotime, China). Glima cells (GCs) or ACs in logarithmic growth phase were evenly plated in 6-well plates and treated with G3T19 after adherence. Preheated 2×EdU working solution (20 µM; 37 °C) and culture medium were added to the 6-well plates in equal proportion, and the cells were labeled for 2 h. Nuclear staining was performed with Hoechst 33,342, and fluorescence photography was performed using an inverted microscope (Olympus, Japan).

Transwell migration assay

Cell migration was assessed using Transwell chambers (8 μm pore size, 24 well specification; Corning). In the upper chamber, 400 µL serum-free cell suspension containing 1 × 10⁵ GCs was added. For the single culture group, 600 µL serum-containing medium was added, and for the co-culture group, 600 µL ACs suspension with G3T19 or CRO was added in the lower chambers. After 24 h, the wells were cleaned with pre-cooling buffer and the residual cells were removed with a cotton swab. The cells were then fixed with paraformaldehyde, soaked with 0.1% crystal violet for 20 min, and washed. Five random high power field images were collected under a microscope.

Flow cytometry

For transwell co-culture, the detailed methods were found in [24]. For cytokine analysis, 400 µL of cell suspension containing 1 × 10⁶ cells were added with 5 µL flow cytometry antibody and incubated at room temperature in the dark for 40 min. The antibodies used for flow cytometry, purchased from Elabscience, included: APC Anti-Human/Mouse CD44 Antibody (Clone IM7; Cat# E-AB-F1100E), APC Anti-Human CD274/PD-L1 Antibody (Clone 29E.2A3; Cat# E-AB-F1133E), and APC Anti-Mouse CD274/PD-L1 Antibody (Clone 10 F.9G2; Cat# E-AB-F1132E). Flow cytometry analysis was performed using a full-spectrum flow cytometer (SA3800, Sony Biotechnology, San Jose, CA, USA).

Jurkat T lymphoma cells proliferation assay

Jurkat T Cells were labeled following the protocol provided in the carboxyfluorescein succinimide ester (CFSE) kit. Subsequently, co-culture conditions were established in transwell chambers depending on the experiment. About 1 × 10⁶ ACs were added into the upper chamber, and GCs were pre-plated in the lower chamber, T cells are suspended in the culture medium. The cells were co-cultured with 24 h, Then, Collected T cells, and the cells proliferation was measured by flow cytometry.

Western blotting

Western blotting was performed according to conventional protocol using the following antibodies: anti-GAPDH (2118, Cell Signaling Technology), anti-PD-L1 (66248-1-lg, Proteintech), anti-CD44 (37259, Cell Signaling Technology), anti-Phospho-NF-κB-P65 (3033, Cell Signaling Technology), and anti-phospho-NF-κB1 p105/p50 (Cat# 13586, Cell Signaling Technology).

Enzyme-linked oligonucleotide assay (ELONA)

CD44 protein was coated onto ELISA plates by incubation at 4 °C overnight. After washing with buffer, the plates were blocked with 100 µL salmon sperm DNA blocking solution (100 µg/mL) at 37 °C for 2 h. Biotin modified G3T19 was added to each well, and K⁺ and Na⁺ were added to Tris buffer to form the parallel structure of G3T19 via incubation at 37 °C for 1 h. The plates were rinsed with buffer, and then 1:5000 dilution of horseradish peroxidase labeled streptavidin (HRP-SA) was added, and the plates were incubated at 37 °C for 30 min. Finally, 100 µL chromogenic solution was added, the plates were incubated at 37 °C for 5–20 min, and 100 µL stop solution was added to each well. The OD value at 450 nm was detected using a microplate reader.

Immunoprecipitation

Approximately 1 × 10⁷ ACs from each group were collected and lysed in 500 µL of IP lysis buffer on ice for 30 min. The protein supernatants were obtained by high-speed centrifugation. Activated streptavidin-coated beads (HRP-beads) were incubated with biotin-labeled G3T19 and CRO. After incubation, the beads were collected and rinsed to remove unbound oligonucleotides, followed by low-speed centrifugation. The protein supernatant was then added to the beads, and protein binding was carried out with shaking at 4 °C overnight. The bead-bound complexes were collected again by low-speed centrifugation and dissociated by high-temperature denaturation in SDS sample buffer. Finally, the supernatant was collected via high-speed centrifugation for subsequent immunoblotting analysis.

To analyze the total succinylation levels of CD44 in reactive ACs with or without G3T19, ACs were treated as indicated for 24 h. Whole cells were harvested and suspended in ice-cold lysis buffer. Lysates were sonicated on ice and incubated for 30 min before being centrifuged at 12,000 rpm for 30 min at 4 °C. Then, the protein A/G immunoprecipitation kit (Cat# IK-1004, Biolinkedin, China) was used according to the manufacturer’s instructions. The resulting immunoprecipitates were analyzed by immunoblotting using succinylated pan antibody (Cat# PTM-401) from PTM Biolabs (Hangzhou, China).

Cell transfection

A total of 2 × 10⁵/well cells were cultured in 6-well plates for 6–12 h. When the cell density reached 70–80%, transient transfection was performed according to the protocol provided by the UltraFection 2.0 (4 A Biotech, China) transfection kit. According to the species source and target gene CD44 or HYAL2, human and mouse siRNA and negative control siRNA were designed. The siRNA sequences, designed and synthesized by Tsingke Technology Co., Ltd. (Wuhan, China), are shown in Table 2.

Table 2.

siRNA sequence

Gene CD44	Sense Sequence (5′-3′)
siRNA1 (Mouse)	UAAACAUAUAAAUUUAGUCAA
siRNA2 (Mouse)	UAUAAAGAGAGUAUGUAACC
siRNA3 (Mouse)	AAAAACAAAGAUAAAUAGCUU
Negative control (Mouse)	UUCUCCGAACGUGUCACGU
siRNA1 (Human)	UAUUCCACGUGGAGAAAAA
siRNA2 (Human)	GCAGAUCGAUUUGAAUAUATT
siRNA3 (Human)	GUAUGACACAUAUUGCUUCTTGTG
Negative control (Human)	UUCUCCGAACGUGUCACGUTT
Gene HYAL2 siRNA (Human)	CACCUA AUGAGGGUUUUGUTT
Negative control (Human)	UUCUC GAACGUGUCACGUTT

Chromatin immunoprecipitation (ChIP) PCR assay

ACs were categorized into three groups (AC, Cul.AC, and G3T19/Cul.AC) and processed using the ChIP Assay Kit (Millipore, MA, USA) as per the manufacturer’s instructions. Chromatin fragments were immunoprecipitated with antibodies specific to NF-κB p50 or control IgG (Millipore). Subsequently, the samples were reverse cross-linked, purified, and analyzed via qRT-PCR. The primers used for amplifying the mouse PD-L1 promoter regions were as follows: PD-L1 promoter, 5’-CCCTAGCAGGAAGCACAT-3’(forward) and 5’ -CCGAGGCTGACAGGTATT-3’ (reverse).

RNA extraction, cDNA synthesis, and quantitative real-time PCR

Total RNA was isolated from cells using Trizol reagent (Invitrogen, USA). RNA was reverse transcribed into cDNA according to the instructions of PrimeScript RT reagent (Takara, Shiga, Japan), and cDNA was amplified using specific primers. The primer sequences, designed and synthesized by Tsingke Technology Co., Ltd., are shown in Table 3. Real-Time fluorescent quantitative PCR (qPCR) was performed according to the TB Green^® Fast qPCR Mix (Takara) instructions using the CFX96 Real-Time PCR Detection System (Bio-Rad, USA).

Table 3.

Primer sequence

Gene	Sequence (5′-3′)
GAPDH (Mouse)	5’-CTGAGCAGACCGGTGTCACATC-3’
GAPDH (Human)	5’-GGAGTCCACTGGCGTCTTCA-3’
CD44 (Mouse)	5’-TCACAAGGCAAGCACTTCAC-3’
CD44 (Human)	5’-CACGTGGAATACACCTGCAA-3’
PD-L1 (Mouse)	5’-GCTTTAATCTTCGAAACTCTTCCC-3’
PD-L1 (Human)	5’-ACAGCTGAATTGGTCATCCC-3’

Protein-DNA (CD44-G4) docking simulation

To study the binding region and interaction mode between CD44 protein and G4 molecules, a professional protein-protein and protein-DNA /RNA docking HDCOK program was adopted. The docking score was calculated based on the ITScorePP or ITScorePR iterative scoring function. Negative docking scores were indicative of stronger interactions. Given our observation that the docking score of the protein-protein /RNA/DNA complex in the PDB averaged at about − 200, we defined a docking score dependent confidence score to represent the putative binding of two molecules as follows: Confidence score = 1.0/ [1.0 + e0.02*(Docking Score + 150)]. Confidence scores > 0.7 indicated that the two molecules are very likely to combine; confidence scores between 0.5 and 0.7 indicated that the two molecules are likely to combine; and confidence scores below 0.5 indicated that the two molecules are less likely to bind.

Animal model

The study protocol was approved by the Experimental Animal Research Ethics Committee of Hubei University of Medicine. All animals were treated in accordance.with the guidelines of the Committee on Animals of the Hubei University of Medicine (Reference Number: 2023-068, dated May 16, 2023). C57BL/6 mice (6–8 weeks, female, specific pathogen-free) were purchased from the Laboratory Animal Resources Center at Hubei University of Medicine (Shiyan, China). To establish a mouse orthotopic glioma xenograft model, mice were fixed on a brain stereotaxic instrument. A 10 µL cell suspension containing 1 × 10⁵ GL261-luc was inoculated into the right caudate nucleus of the mouse brain, and one week after modeling, the small animal live imager was used to detect the modeling by IVI chemiluminescence (DELL, T3500/Matrx6y/caliper). After anesthetizing mice with isoflurane, 150 mg·kg^− 1 fluorescein (about 2.5 mg per 16 g) was injected intraperitoneally. The appearance of a fluorescent signal in the mouse brain tissue after 5 min indicated successful modeling.

For assessment of drug intervention in tumor bearing mice, the experimental mice were divided into five groups, each group contained six mice. After tumor bearing, groups were given PBS, G3T19, or CRO every 3 days for 3 consecutive weeks. Live imaging was performed weekly to monitor the in situ growth of the gliomas in the brain. The survival status and body weights of mice were recorded for 30 days. Eventually, the animals were euthanized, and the brain tumor tissues were prepared for HE staining. Additionally, paraffin sections were prepared for IHC staining with anti-CD44, GFAP, anti-PD-L1 and anti-Phospho-NF-κB P50 antibodies.

Statistical analysis

The data were processed with GraphPad 8.0 software. All quantitative results are presented as the mean ± standard deviation (SD). For statistical analysis, comparisons between two groups were conducted using the independent Student’s t-test, while comparisons involving three or more groups were analyzed using one-way ANOVA. *p < 0.05, **p < 0.01, ***p < 0.001.

Results

G-quadruplex G3T19 can cross the blood-brain barrier

To identify a new ssDNA drug for treating glioma, we screened candidate oligonucleotides with G-quadruplex structure (Table 1). The oligonucleotide sequences of G3T19, AS1411, PS2.M, [B7]-3-0, WY-A6, EAD2, CatG4 were screened by ABTS peroxidation activity, with the random nucleic acid sequence CRO used as control. As shown in Fig. 1A, G3T19 had the strongest color reaction, suggesting strong peroxidase activity. For verification, we investigated the conformations of G-quadruplex structures using Circular Dichroism (CD) spectroscopy. (Fig. 1B). The CD spectrum of G3T19 had a positive peak at 280 nm and a negative peak at 245 nm, indicative of a typical parallel structure, thus confirming that its G-quadruplex structure was stable. These values were greater for G3T19 than for AS1411, which was characterized to inhibit the proliferation of glioma in our previous studies [24]. Therefore, these findings suggest that G3T19 may represent a more stable G4 structure with stronger peroxidase activity than AS1411.

Fig. 1.

G-quadruplex G3T19 can cross the blood-brain barrier. A The peroxidase activity of G3T19 was compared with that of other G4 oligonucleotides. B G4 parallel structure confirmation by Circular Dichroism. C Spatial parallel structure simulation diagram of G3T19. D, E The content of G3T19 in each tissue was determined by HPLC at 24 h after mouse tail vein administration. F, G The distribution of G3T19 in major tissues of mice was detected by small animal imaging. H The distribution and content of G3T19 in mouse tissues were detected by peroxidase color reaction experiments. All experiments were repeated three times and averaged (mean ± SD) (**** p < 0.0001)

Figure 1C shows a simulation diagram of the spatial parallel structure of G3T19. To evaluate the bioavailability of G3T19, we injected it into mice and then collected tissues 24 h later for High Performance Liquid Chromatography (HPLC). According to the peak area and absorbance of the chromatograph, the G3T19 content in each tissue (brain, heart, liver, spleen, lung, kidney) was calculated. The results showed that the content of G3T19 was higher in the kidney, liver, lung, and brain; and lower in the heart and spleen (Fig. 1D and E). Delivery of G3T19 into the brain and its organ distribution (heart, liver, spleen, lung, kidney) were confirmed by fluorescent imaging (Fig. 1F and G). The biochemical indicators (Fig. S1A-D) and HE staining (Fig. S1E) consistently show that G3T19 treatment does not cause toxicity in the mouse heart, liver, or kidney, with all measured parameters remaining similar to those in the control and CRO groups. Additionally, we determined that the plasma half-life of G3T19 is 1.03 h (Fig. S1F and G). Furthermore, the amounts of residual G3T19 in the corresponding organs per unit weight demonstrate that the G3T19 distribution trend was consistent with the HPLC and fluorescent imaging system results (Fig. 1H). These results suggest that G-quadruplex G3T19 can cross the blood-brain barrier.

G3T19 significantly inhibits the growth of glioma

To determine whether G3T19 may have glioma growth inhibitory activity, we performed CCK-8. U87, U251 and ACs were subjected to treatment with different concentrations of G3T19, AS1411 or CRO (1, 2, 5 µmol/L), and also the positive control drug such as doxorubicin(dox). The results demonstrate that increasing concentrations of G3T19 decreased the proliferation ability of GCs (U87 and U251 cells). At 5 µmol/L, G3T19 significantly inhibited the proliferation of GCs, while AS1411 had no obvious inhibitory effect at this concentration. Furthermore, G3T19 had no observable effect on ACs, indicating that its inhibitory effect is specific to GCs. (Fig. 2A). And the increasing concentrations of dox decreased the proliferation ability of GCs (U87 and U251 cells) and ACs (Fig. S2A). For additional confirmation, we performed EdU-488 staining assays. The fluorescence intensity of the GCs was significantly weaker after G3T19 treatment as compared to the fluorescence intensity in the untreated or CRO groups, with no observable effect on ACs, thus verifying that G3T19 specifically inhibits the proliferation of GCs (Fig. 2B). Moreover, G3T19 increased the apoptosis levels and reduced the migration of GCs (Fig. 2C–E). At the same time, Dox specifically inhibits the proliferation of GCs and ACs (Fig. S2B). Moreover, Dox increased the apoptosis levels and reduced the migration of GCs (Fig. S2C–E).

Fig. 2.

G3T19 significantly inhibits the growth of glioma. A The proliferation ability of GCs (U87 and U251) and ACs treated with different concentrations (1 µmol/L, 2 µmol/L, 5 µmol/L) of G3T19, AS1411 and CRO was detected by CCK-8 assay. B The proliferation of U87, U251 and ACs was detected by EdU-488 staining after 5 µmol/L treatment with G3T19, AS1411 and CRO. C, D Apoptosis of GCs was detected by flow cytometry after Annexin V-FITC treatment. E G3T19 reduced GC migration as analyzed by transwell migration assays. F, G Cy5-labeled G3T19 take up by ACs, GCs and reactive ACs was detected by flow cytometry. All experiments were repeated three times and averaged (mean ± SD) (*p < 0.05, **p < 0.01, ***p < 0.001)

Next, we evaluated the effect of AC/GC co-culture on the G3T19 uptake. The same number of single-culture ACs, ACs co-cultured with GCs (Reactive ACs) and GCs co-cultured with ACs were incubated with Cy-5 fluorescently labeled G3T19, and the uptake of G3T19 by flow cytometry was detected. Among the three types of cells, the ACs co-cultured with gliomas had the largest uptake of G3T19 (Fig. 2F and G). These results reveal that G3T19 can inhibit the growth of glioma at lower concentrations than AS1411 can, but had no observable effect on ACs. While dox is toxic to both glioma and ACs, and that co-culture with GCs stimulates AC uptake of G3T19.

G3T19 reduces the pro-GC activity of ACs

Studies have demonstrated that there is a mutually beneficial interaction between GCs and ACs in the glioma microenvironment, whereby gliomas can stimulate the activation of normal ACs, and ACs can promote the growth of gliomas [23, 38]. Firstly, we co-cultured astrocytes and GCs (U87, U251) in transwell chambers, with astrocytes seeded in the upper chamber and GCs seeded in the lower chamber. After 24 h of co-culture, the GCs were collected for detection. To evaluate the effect of G3T19 in AC/GC co-cultures, we detected the proliferation, migration, and apoptosis of GCs after co-culture with ACs in transwell chambers. The results demonstrate that the proliferation of GCs was significantly increased after co-culture with ACs, but that G3T19 treatment inhibited the proliferation of GCs either in monoculture or co-culture (Fig. 3A and B). Similar results were observed in GC migration assays (Fig. 3C). Conversely, co-culture with ACs inhibited GC apoptosis, while the apoptosis rate of GCs increased significantly after G3T19 treatment (Fig. 3D and E). These results suggest that G3T19 weakens the ability of ACs to promote the proliferation and migration and alleviate the apoptosis of GCs.

Fig. 3.

G3T19 reduces the pro-GC activity of ACs. A, B The effect of G3T19 on GC proliferation was evaluated in the presence or absence of ACs by CFSE staining assays. C The effect of G3T19 on GC migration was measured in the presence or absence of ACs by transwell migration assays. D, E The effect of G3T19 on GC apoptosis was measured in the presence or absence of ACs by flow cytometry assays. All experiments were repeated three times and averaged (mean ± SD) (*p < 0.05, **p < 0.01, ***p < 0.001)

ACs enhance glioma immune escape, while G3T19 mitigates this effect

To evaluate the effect of G3T19 on immune escape, we detected the expression of PD-L1 after monoculture or co-culture of the same number of ACs and U87 cells in transwells. The PD-L1 expression level was higher in ACs that were co-cultured with U87 than in AC or U87 monocultures and in U87 that were co-cultured with ACs, as assessed by flow cytometry (Fig. 4A and B). Furthermore, G3T19 inhibited the expression of PD-L1 in co-cultures of ACs as evaluated by Western blotting and flow cytometry (Fig. 4C–E). AC/U87 co-cultures were also more effective than AC or U87 monocultures in inhibiting the proliferation of Jurkat T cells (Fig. 4F and G), which was reversed by G3T19 (Fig. 4H and I). These results suggest that ACs enhance glioma immune escape, while G3T19 mitigates this effect.

Fig. 4.

ACs enhance glioma immune escape, while G3T19 mitigated this effect. A, B The expression level of PD-L1 in ACs or U87 cells was detected by flow cytometry. C, D Flow cytometry assessment of the effect of G3T19 in reducing the expression of PD-L1 in reactive ACs. E Western blot assessment of the effect of G3T19 in reducing the expression of PD-L1 in reactive ACs. F, G The proliferation ability of Jurkat cells in co-cultures with ACs and U87 cells was detected by CFSE staining and flow cytometry. αPD-L1 was applied as a positive control. H, I Flow cytometry assessment of the effect of G3T19 in promoting the proliferation of Jurkat cells in co-cultures with ACs and U87 cells. All experiments were repeated three times and averaged (mean ± SD) (*p < 0.05, **p < 0.01, ***p < 0.001, **** p < 0.0001)

G3T19 binds to CD44 and down-regulates its expression

Studies have shown that CD44 is a positive regulator of PD-L1 expression [16]. Therefore, we considered the possibility that G3T19 may regulate PD-L1 expression via binding to CD44. Consistent with this possibility, IP assays and ELONA assays each demonstrated that G3T19 binds to CD44 (Fig. 5A and B). The G3T19-CD44 interaction was predicted by visualized docking, and molecular docking calculations were performed at three times. The docking Score of the three conformations was − 238.32 kcal/mol, − 232.56 kcal/mol, and − 235.80 kcal/mol respectively, and the mean values and standard deviations is -235.56 ± 2.36 kcal/mol. The docking Score of the optimum dock1 was − 238.32 kcal/mol, with a Confidence Score of 0.854, indicating high reliability. The optimal model results predicted that the G4 molecule mainly binds in the C-terminal cavity of the CD44 protein. Additionally, analysis of the interaction mode of the binding region revealed that a total of 12 groups of hydrogen bond interactions were formed between the CD44 protein and the G4 molecule; the main amino acids involved in the formation of hydrogen bond in the CD44 protein were His92, Ala99, Asn100, Asn101, Thr102, Tyr105, Arg186, Ser187, Thr189 and Gly192, while the bases involved in hydrogen bond formation in the G4 molecule were C1, G3, T6, G8, G11, G12 and T18 (Fig. 5C and D, Fig.S3). Then the in vitro validation of the molecular docking was performed. Point mutations at C1, G3, G8, G11, and T18 in G3T19 significantly reduced the binding ability of G3T19-CD44 (Fig. 5E), These results suggest that these sites are critical for receptor recognition and function and the parallel G-quadruplex structure of G3T19 binds to CD44 primarily via hydrogen bonding.

Fig. 5.

G3T19 binds to CD44 and down-regulates its expression. A, B G3T19 binding to CD44 was assessed by immunoprecipitation (panel A) and ELONA assay (panel B), the binding of CRO to CD44 was used as a control. C, D G3T19-CD44 interactions were predicted according to the visualized docking conformation. Purple represents the CD44 protein structure, and green represents G4. E Mutations in G3T19 that disrupt the interactions with CD44 were demonstrated by ELONA. F, G Flow cytometry of CD44 expression in AC monocultures or co-cultures after treatment with G3T19 or CRO. H Western blot analysis of CD44 and GFAP levels in AC monocultures or co-cultures after treatment with G3T19 or CRO. I, J Flow cytometry of CD44 expression in CD44-deficient ACs with or without G3T19. All experiments were repeated three times and averaged (mean ± SD) (*p < 0.05, **p < 0.01, ***p < 0.001)

To determine whether G3T19 regulates CD44 expression, we performed flow cytometry and western blot analyses. The results showed that the expression of CD44 and GFAP in ACs was significantly increased after co-culture with GCs, while G3T19 inhibited the expression of CD44 in co-cultured ACs (Fig. 5F–H). G3T19 also still partially inhibit PD-L1 in CD44-deficient cells, it suggests potential off-target effects (Fig. 5I and J). Together, these results demonstrate that the parallel G-quadruplex structure of G3T19 binds to CD44 and down-regulates its expression, but G3T19’s action is not entirely dependent on CD44.

G3T19 inhibits AC-enhanced glioma immune escape by down-regulating CD44 and the NF-κB pathway

To confirm the correlation between CD44 and PD-L1 expression, we transfected CD44-siRNA into ACs or U87 cells and measured the mRNA and protein expression of CD44 and PD-L1 by qRT-PCR and flow cytometry. The results show that the expression levels of both CD44 and PD-L1 were decreased in ACs and U87 cells after transfection of CD44-siRNA (Fig. 6A–D, Fig. S4A–D). For in vivo confirmation, we extracted ACs from CD44 gene knockout (KO-CD44) or wild type (WT) mouse brain tissues. Western blotting and qRT-PCR results verify that the protein expression and mRNA levels of PD-L1 in ACs were reduced in CD44 gene knockout mice (Fig. 6E and F). To evaluate the functional consequence of CD44 knockout, we co-cultured ACs from CD44 knockout mice with U87 cells, and then measured the PD-L1 levels and Jurkat T cells proliferation by flow cytometry. The results confirm that CD44 depletion down-regulates PD-L1 expression and thus facilitates T cell escape (Fig. 6G–J).

Fig. 6.

G3T19 inhibits AC-enhanced glioma immune escape by down-regulating CD44 and inhibiting the NF-κB pathway. A, B The mRNA levels of PD-L1 in U87 cells and ACs transfected with CD44-siRNA were detected by qRT-PCR. C, D The protein levels of PD-L1 in U87 cells and ACs transfected with CD44-siRNA were detected by flow cytometry. E The protein expression of PD-L1 in ACs from CD44 gene knockout mice (KO-CD44) or wild-type mice (WT) was detected by Western blotting. F The mRNA levels of PD-L1 in ACs of CD44 gene knockout mice were detected by qRT-PCR. G, H Flow cytometry to compare the expression levels of PD-L1 in ACs from CD44 knockout mice and wild-type mice co-cultured with U87 cells. I, J CFSE staining and flow cytometry to compare the proliferation of Jurkat T cells after co-culture with reactive ACs from CD44 knockout or wild-type mice. K The protein expression of p-P50 and p-P65 in ACs from CD44 gene knockout mice was detected by Western blotting. L ChIP assay coupled with qRT-PCR analysis to reveal the effect of co-culture with U87 cells (Cul.AC) and G3T19 on the enrichment of NF-κB p50 on PD-L1 promoters in ACs. M Western blot results showing that G3T19 reduces the expression of p-P65 and p-P50 in reactive ACs. All experiments were repeated three times and averaged (mean ± SD) (**p < 0.01, ***p < 0.001, **** p < 0.0001)

To explore the downstream effects of reduced CD44 expression in ACs, we evaluated the expression of p-P50 and p-P65 in ACs from CD44 gene knockout mice. The results suggest that down-regulation of CD44 inhibits the NF-κB pathway (Fig. 6K). Using JASPAR software, we analyzed possible binding sites between PD-L1 and NF-κB (Fig. S4E). Subsequently, we used CHIP-qPCR to confirm the interaction between NF-κB and PD-L1. The results indicate that the occupancy of the PD-L1 promoter is increased by NF-κB in ACs co-cultured with U87 cells, but this effect is decreased by G3T19 (Fig. 6L). Western blotting assays verified that G3T19 downregulates the NF-κB pathway in ACs that are co-cultured with U87 cells (Fig. 6M). Collectively, these results suggest that G3T19 inhibits AC-enhanced glioma immune escape by down-regulating CD44 and inhibiting the NF-κB pathway.

G3T19 down-regulates CD44 succinylation to inhibit AC-induced glioma immune escape

To determine whether the regulation of CD44 might be mediated in part by succinylation, we performed IP assay of CD44 with succinylated pan-antibodies. The level of CD44 succinylation were greater in ACs that were co-cultured with GCs than those that were monocultured; furthermore, G3T19 down-regulated the succinylation of CD44 in co-cultured ACs (Fig. 7A, Fig. S5A). Mass spectrometry results revealed that there was a lysine succinylation site (YAGVFHVEK_succ) at position 40 of CD44 in Reactive ACs (Fig. 7B). Therefore, four expression plasmids were constructed: lysine (K) mutated to glutamate (E) to simulate high succinylation of CD44K40E, lysine (K) mutated to arginine (R) to simulate desuccinylation of CD44K40R, overexpressing CD44 (CD44WT) and empty Vector. These plasmids were transfected into ACs, and the results confirmed that up-regulating CD44 succinylation with the CD44K40E mutant promoted the expression of PD-L1 and CD44 and the phosphorylation of p50, while down-regulating CD44 succinylation with CD44K40R had the opposite effect (Fig. 7C). For additional confirmation, the transfected ACs after co-culture with GCs with or without T cells. Flow cytometry analysis demonstrated that up-regulating CD44 succinylation enhanced the AC-induced immune escape of GCs, while down-regulating CD44 succinylation inhibited the AC-induced immune escape of GCs (Fig. 7D and E). Moreover, we shown G3T19 alone significantly reduced PD‑L1 expression in reactive astrocytes and restored Jurkat T cell proliferation (i.e., relieved immune suppression). Overexpression of CD44K40E partially but significantly rescued the effects of G3T19 – PD‑L1 levels remained higher, and T cell proliferation remained lower, compared to the empty vector + G3T19 group. These results directly demonstrate that G3T19 acts through downregulating CD44 succinylation, because a succinylation‑mimetic mutant (CD44K40E) that cannot be de‑succinylated by G3T19 partially overrides the drug’s inhibitory effects on PD‑L1 and immune escape (Fig. 7F-H, Fig.S5B). Altogether, these results indicate that G3T19 down-regulates CD44 succinylation to inhibit AC-induced glioma immune escape.

Fig. 7.

G3T19 down-regulates CD44 succinylation to inhibit AC-induced glioma immune escape. A Immunoprecipitation analysis of the total succinylation level of CD44 in ACs. The Western blot shows total succ-CD44 and CD44 proteins in ACs. Anti-IgG was used as a negative control, and GAPDH was used as a loading control. B Mass spectrometric verification of CD44 succinylation in Reactive ACs. CD44 succinylation at K40 (YAGVFHVEK_succ) is shown. C The expression of CD44, PD-L1 and p-P50 levels in ACs from CD44 wild-type and mutant groups was detected by western blotting. D, E Jurkat T cell proliferation in CD44 wild-type and mutant groups was detected by flow cytometry. F, G Flow cytometry of PD-L1 expression in ACs that overexpress CD44K40E with or without G3T19. H The proliferation of Jurkat T cells that co-cultured in ACs which overexpress CD44K40E with or without G3T19. All experiments were repeated three times and averaged (mean ± SD) (*p < 0.05, ***p < 0.001)

G3T19 inhibits glioma in vivo

To validated our results in vivo, we established mice bearing orthotopic GC xenografts. PBS, G3T19 (2 µM, 5 µM) or CRO were administered in the tail veins, and the glioma growth was monitored weekly. The results demonstrated that the fluorescence intensity of the PBS, CRO and G3T19 with 2 µM groups was progressively enhanced and the tumors were obviously enlarged at the second week, while the fluorescence intensity of the G3T19 with 5 µM group was weaker, indicating that the growth of the gliomas was slower. The fluorescence intensity of the PBS group was the strongest at the third week, while the fluorescence intensity of the G3T19 group did not obviously progress and was the weakest. (Fig. 8A and B). Furthermore, the survival of mice bearing orthotopic GC xenografts treated with G3T19 (5 µM) was significantly extended. The 30‑day survival rates were 0% in both the PBS and CRO groups, 33% in the 2 µM group, and 67% in the 5 µM G3T19 group (Fig. 8C). In addition, the GFAP staining level was markedly intensified in mice that received PBS, CRO and G3T19 with 2 µM, indicating that ACs were activated in brain tissue that interfaced with the tumors. On the other hand, treated with 5 µM G3T19, AC activation was significantly reduced. Similarly, the expression of CD44, PD-L1, and p-P50 was elevated on both sides of the brain tumor border in mice treated with PBS, CRO, and 2 µM G3T19, but was notably suppressed in mice treated with 5 µM G3T19. We also examined the infiltration of CD8⁺ T cells in bearing orthotopic GC xenografts. The results showed that, compared with normal mouse brain tissue, there was significant CD8⁺ T cell infiltration in the brain tissue of mice bearing orthotopic GC xenografts, and treatment with 5 µM G3T19 further increased the degree of CD8⁺ T cell infiltration, ELISA detected the CD8⁺ T cell activation factors TNF-α and IFN-γ in brain tissue, and the trend was consistent with the CD8⁺T immunohistochemistry results (Fig. 8D, Fig.S6). Collectively, these findings provide in vivo evidence that a high concentration of G3T19 (5 µM) effectively inhibits the expression of CD44 and glioma immune escape, as well as glioma growth.

Fig. 8.

G3T19 inhibits glioma in vivo. A, B The anti-tumor ability of G3T19 was assessed using an orthotopic xenograft glioma model. The in vivo fluorescence intensity of the glioma was analyzed using a small animal imaging system. C Survival rate analysis of mice. D The expression of PD-L1, CD44, GFAP, P50 and CD8 in brain tissues of mice was detected by immunohistochemistry. All experiments were repeated three or five times and averaged (mean ± SD) (*p < 0.05, **p < 0.01, ***p < 0.001)

G3T19 suppresses glioma in human glioma organoid

To further determine the effect of G3T19 on glioma, we treated glioma organoids (GBO) with 5 µM G3T19 or CRO. Immunofluorescence staining with acridine orange (AO, green, live cells) and propidium iodide (PI, red, dead cells) showed that compared with the control and CRO‑treated groups, the G3T19‑treated group exhibited markedly reduced green fluorescence and increased red fluorescence, indicating that G3T19 significantly inhibited GBO growth (Fig. 9A). Consistently, ATP viability assays confirmed the same trend (Fig. 9B). We then co‑cultured GBO with astrocytes. G3T19 reversed the GBO‑induced upregulation of PD‑L1, p‑P50, and CD44 succinylation in ACs, as shown by western blotting (Fig. 9C) and IP assay of CD44 with succinylated pan-antibodies (Fig. 9D). Finally, activated T cells were co‑cultured with astrocytes and GBO. G3T19 markedly enhanced T‑cell infiltration into the GBO (T‑cells merge panel) (Fig. 9E). Collectively, these results demonstrate that G3T19 suppresses human glioma organoid growth and astrocyte‑induced immune escape by downregulating CD44 succinylation.

Fig. 9.

G3T19 inhibits glioma in GBO. A The growth of GBO showed by immunofluorescence staining with acridine orange (AO, green, live cells) and propidium iodide (PI, red, dead cells). B The growth of GBO showed by ATP analysis (C) The expression of PD-L1 and p-P50 of ACs that co-cultured with GBO was detected by western blotting. D The CD44 succinylation level of ACs that co-cultured with GBO was detected by IP. E T‑cells infiltration into the GBO which detected by immunofluorescence staining (Red: PI, dead cells of GBO, Blue: ‌Hoechst, jurket T cells) All experiments were repeated three times and averaged (mean ± SD) (***p < 0.001)

Discussion

The glioma microenvironment has a complex cellular composition. Although its relationship with malignant progression has been studied, the intricate interactions remain poorly understood. Astrocytes (ACs) are major CNS cells that become rapidly activated under pathological conditions and exert multiple effects on brain tissue [39]. Although ACs express MHC and can serve as antigen‑presenting cells, they contribute to CNS immunosuppression. They secrete cytokines such as TGF‑α and CXCL12 that support brain tumor cell proliferation [40–42]. In glioma, AC–microglia interactions promote anti‑inflammatory cytokines (TGF‑β, IL‑10, G‑CSF), whereas activated monocytes and microglia reduce AC production of the pro‑inflammatory cytokine TNF‑α [43]. We show for the first time that glioma immune escape depends directly on surrounding reactive ACs (Fig. 4). Previous studies have focused on glioma cells themselves or on macrophage reprogramming in immune escape [43]. However, ACs, though not classical immune cells, are the most abundant cells surrounding gliomas. Thus, our finding highlights the AC‑associated microenvironment as a promising target for glioma immunotherapy.

Small‑molecule ssDNAs are non‑immunogenic and have non‑toxic side effects in clinical trials, making them particularly suitable for immunocompromised glioma patients [26]. We previously showed that AS1411 is a promising aptamer‑like ligand with high targeting efficiency and cellular uptake. AS1411 reduces the pro‑glioma activity of ACs and disrupts exosome‑miRNA‑27a‑mediated crosstalk between glioma cells (GCs) and ACs [24]. Here, we identified another G‑quadruplex (G4) ‑ forming ssDNA, G3T19. Compared with AS1411, G3T19 exhibits stronger peroxidase activity, greater anti‑proliferative effects on glioma cells, and higher uptake by reactive ACs surrounding gliomas (Figs. 1, 2 and 3). Furthermore, G3T19 downregulates PD‑L1 expression in reactive ACs, thereby inhibiting AC‑induced glioma immune escape (Fig. 4).

We propose that the high biological activity of G3T19 arises from its G4 structure [28–30]. We used mass spectrometry to screen for G3T19‑binding proteins, an approach distinct from the traditional SELEX method [44]. This approach allowed more targeted screening based on the drug’s biological function. We verified G3T19 binding to CD44 by co‑immunoprecipitation and colocalization. Molecular docking predicted the interaction conformation and showed that G3T19 binds CD44 through its parallel G4 structure (Fig. 5A-E). These findings provide new insights into how nucleic acid advanced structures can modulate biological functions. On the other hand, we acknowledge the off- target activity of G3T19, which is not uncommon for aptamer‑based drugs, particularly at higher concentrations or prolonged exposure. (Fig. 5I and J).

CD44 has been extensively studied as a therapeutic target in glioma, and astrocyte–glioma crosstalk is a well‑recognized driver of tumor progression. Previous work focused on CD44 expression, glycosylation, or hyaluronan interaction, using antibodies, hyaluronidase, or RNAi to block its function [13–15]. Our experiments also showed that HYAL2‑generated soluble HA fragments critically activate astrocytic CD44 in transwell co‑cultures (Fig. S7). However, other glioma‑derived soluble factors, such as SPP1 or IL‑6, may also contribute [23]. Future studies with neutralizing antibodies and recombinant proteins are needed to define each ligand’s contribution. However, two critical aspects remain unexplored. First, it is unknown whether CD44 undergoes succinylation in reactive astrocytes. Succinylation is a newly recognized modification that can alter protein charge and conformation. It is also unclear whether this modification actively contributes to immune escape. Second, although G4‑forming oligonucleotides are known for antiproliferative effects, their ability to directly bind CD44 and modulate its post‑translational modification has not been reported. Our study fills these gaps. We identify lysine 40 succinylation of CD44 as a functionally relevant modification that drives PD‑L1 expression via NF‑κB in astrocytes. We also demonstrate that G3T19 acts not merely by reducing CD44 levels, but by specifically downregulating this succinylation. Rescue experiments using the succinylation‑mimetic mutant CD44K40E provide direct causal evidence (Figs. 6 and 7). Thus, our work builds on established CD44 targeting and microenvironment modulation, but introduces a previously unrecognized regulatory layer—CD44 succinylation—as a therapeutically actionable node in glioma immune evasion.

G3T19 has demonstrated therapeutic potential in vitro and in vivo, supporting its further investigation for glioma treatment. Animal studies are essential for demonstrating anti-tumor efficacy. However, our previous work has shown that drugs can modulate interactions between glioblastoma cells and multiple stromal cell types, including astrocytes, monocytes, macrophages, and dendritic cells [23]. Consequently, even if anti-tumor efficacy is observed in animal models, attributing the effects to specific cell types remains challenging (Fig. 8). To further validate G3T19’s action, we employed glioma organoid models and found that G3T19 significantly inhibited organoid growth and concurrently downregulated CD44 succinylation and PD-L1 expression in astrocytes co-cultured with the organoids, providing preliminary evidence for the relevance of this modification in a three-dimensional microenvironment (Fig. 9). Nevertheless, organoid models lack intact vascular structures and immune cell components and cannot fully recapitulate the complex intercellular communication in vivo; thus, these findings should be interpreted with caution. Similarly, the Transwell co-culture system used in our study permits only soluble factor exchange, fails to model direct cell–cell contacts, and includes only astrocytes among stromal cells, thereby not representing the synergistic contributions of multiple cell types within the tumor microenvironment. This further complicates the precise attribution of G3T19’s cell-specific effects. Additionally, the binding sites predicted by molecular docking may be incomplete, and cryo-EM structural determination of the G3T19–CD44 interface would be valuable. Collectively, these represent the limitations of our study. To overcome these issues, future studies should incorporate co‑culture systems that combine patient‑derived organoids with autologous immune cells, or alternatively, utilize humanized immune system mouse models to evaluate the immunomodulatory effects of G3T19 under conditions that simultaneously recapitulate both stromal and immune compartments. Despite these limitations, the consistency of our findings across multiple models—including 2D co‑cultures, organoids, and xenografts—strengthens the core conclusion. Nevertheless, we acknowledge that full validation will require the integration of more physiologically relevant systems, and our ongoing work is directed toward these goals.

Conclusions

Oligonucleotide G3T19 binds to CD44 via its G-quadruplex parallel structure, downregulates CD44 succinylation, inhibits the NF-κB pathway and the expression of PD-L1, and inhibits AC-induced glioma immune escape. These findings elucidate a key mechanism of interaction between GCs and ACs in glioma (Fig. 10). Based on our results, the advanced nucleic acid structure of G3T19 might offer a new approach for modulating the immune microenvironment in glioma therapy.

Fig. 10.

Schematic representation of the mechanism of action of G3T19. Oligonucleotide G3T19 binds CD44 with G-quadruplex parallel structure and downregulates CD44 succinylation, inhibits the entry of the transcription factor P50 into the nucleus in the NF-κB pathway, down -regulating the expression of PD-L1 in astrocytes, finally inhibiting astrocytes enhanced glioma immune escape

Author contributions

CZ, SX and LC conceived and designed the study. CZ wrote main manuscript text. SX, WX performed most of the experiments. ZS, LY, WD, ZW, GC, and DC carried out initial data analyses and performed partial of the experiments. All authors contributed to drafting the manuscript. All authors have reviewed and approved the final submitted manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (No. 82574441, 82472302), Hubei Provincial Natural Science Foundation (No. 2024AFB802).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The study is approved by the Ethics Committee of Hubei University of Medicine.

Competing interests

The authors declare no competing interests.