Bioinformatic analysis

The data for human glioma tissues were downloaded from TCGA (The Cancer Genome Atlas, http://tcga-data.ncbi.gov/tvga/) and GEPIA (Gene Expression Profiling Interactive Analysis), which comprised data from glioma and normal tissues. PLEKHA4 expression was analyzed in glioma and non-glioma tissues.

Gene set enrichment analysis (GSEA)

GSEA (http://www.broad.mit.edu/gsea/) was used to assess the correlation between PLEKHA4 expression and biological processes/pathways. We divided the dataset obtained from TCGA into two groups (high and low PLEKHA4 expression). We select default settings, analyzed the data to determine significance thresholds, and calculated the FDR. It is generally believed that an Enrichment fraction (NES) with |NES | > 1, p < 0.05, and a q-value (i.e. FDR) < 0.25 is considered substantially enriched.

Cell culture

The human GBM cell lines, such as T98G were acquired from the BeNa Culture Collection (BNCC, Henan, China), U87, U251, and U343 were acquired from Jiandun Biotechnology (Shanghai, China), and HBMECs were acquired from Procell (Wuhan, China). U251, U87, U343, T98G, and HBMECs were cultured in DMEM (Biosharp, Anhui, China) supplemented with 10% fetal bovine serum (FBS, Gibco, USA). Cells were placed in a 5% CO2 incubator. All cell STR profiling and mycoplasma testing were shown in the supplementary information section.

Lentiviral infection

Three PLEKHA4 shRNAs were synthesized and cloned using the Plko.1-puro cloning vector. To prepare PLEKHA4 overexpression, the PLEKHA4 sequence was sub-cloned into the Plvx-Puro lentiviral vector. T98G and U87 cells were infected with shRNA-expressing and overexpressing lentiviral supernatants, and after 72 h of infection, mRNA and protein expression was determined. The shRNA sequences were as follows:

shRNA-1: 5ʹ-GGAGAAGGAGCAACTAGAA-3ʹ, 5ʹ-TTCTAGTTGCTCCTTCTCC-3ʹ;

shRNA-2: 5ʹ-GCTACAATCCAGCTTCTAA-3ʹ, 5ʹ-TTAGAAGCTGGATTGTAGC-3ʹ;

shRNA-3: 5ʹ-GAGTCAACTTTCCACCAAA-3ʹ, 5ʹ-TTTGGTGGAAAGTTGACTC-3ʹ

qRT-PCR

TRIzol (YEASEN, Shanghai, China) was used to collect total RNA. Complementary DNA was reversed transcribed with the Hifair® II 1st strand cDNA Synthesis Kit (YEASEN, Shanghai, China), per the manufacturer’s instructions. Transcripts were amplified via qPCR using the Hieff® qPCR SYBR Green Master Mix (YEASEN, Shanghai, China). HOXD9 and PLEKHA4 primer sequences were as follows:

HOXD9: F 5ʹ -TTTGGGGTTTCGCCCTATCC-3ʹ, R 5ʹ - CTGGGGGTGAGGGGACTAAA-3ʹ;

PLEKHA4: F 5ʹ - TGTCCGACCTCCTCTGGATT-3ʹ, R 5ʹ - AGAGTGTGCCTGTGTTCTGG-3ʹ;

Actin: F 5ʹ -CCTTCCTTCCTGGGCATGG-3ʹ, R 5ʹ - GATCTTCATTGTGCTGGGTGC-3ʹ.

Western blotting

Total protein was extracted from cells, and quantified using BCA kit. Then 20 µg of protein were added to the a 10% SDS-PAGE gel for electrophoresis, after which the protein was transferred to 0.22 μm PVDF membranes. The membranes were removed and placed in 5% nonfat milk for blocking, followed by incubation with primary antibodies overnight at dilutions of 1:5000 (anti-PLEKHA4; Abcam, Ab84727, UK), 1:2000 (anti-HOXD9; CST, 55962, USA), 1:2000 (anti-STAT3; Abcam, Ab119352, UK), 1:3000 (anti-p-STAT3; Abcam, Ab76315, UK), 1:1000 (anti-SOCS1; Abcam, Ab280886, UK), and 1:5000 (anti-β-actin; Proteintech, 66009-1-Ig, Wuhan China). Subsequently, an HRP-conjugated secondary antibody at a dilution of 1:10000 (Goat anti-mouse; ZSGB-BIO, ZB-2305, Beijing, China) or (Goat anti rabbit; ZSGB-BIO, ZB-2301, Beijing, China) was incubated for 1 h at 37 °C. ECL chemiluminescence detected the protein expression, which was quantified using Image J software (National Institutes of Health, USA).

Reagents

TMZ (MCE, HY-17364, china) was purchased from MedChemExpress (MCE, https://www.medchemexpress.cn/). We prepared a 10 mM TMZ solution with dimethyl sulfoxide (DMSO) and separated it to avoid repeated freeze-thaw cycles. It was stored at −20 °C. All antibody information was shown in Table S1.

Cell proliferation

Cell Counting Kit-8 (CCK8, Beyotime, C0039, China) was used to detect cell growth per the manufacturer’s instructions. An aliquot of the 5 × 103 cell suspension was added in triplicate to a 96-well plate and grown at 37 °C. Add 10 µl CCK8 reagents into each well at 0, 12, 24, 48, and 72 h of incubation, followed by incubation at 37 °C for 1.5 h. The absorbance was measured at 450 nm using an enzyme label analyzer (PERLONG, DNM-9602, Beijing, China).

Cell apoptosis

The Annexin V-FITC staining kit (Beyotime, Shanghai, China) was used to detect GBM cell apoptosis. According to the manufacturer’s instructions, we collected cells and re-suspended them with a staining solution, followed by incubation them with annexin V-FITC and Propidium Iodide (PI) in the dark for 15 min. A Calibur Flow Cytometer (Beckman, CytoFLEX, USA) was used to detect apoptosis, where Annexin V+/PI- cells were deemed to be early apoptotic cells.

Glucose uptake assays

We collected cells to detect glucose uptake. The cells were incubated in a culture medium containing 2-NBDG (2-(N-(7-nitrobenz-2-oxa-1,3-dizol-4-yl)amino)-2-deoxyglucose, Bio Vision, USA). After incubation, the cells were collected and centrifuged, re-suspended with PBS, and finally analyzed with FlowLogic FCS analysis software (Backman Coulter, USA).

ATP and lactate assays

ATP and lactate were measured using the ATP and lactate kits (Nanjing Jian-cheng Bioengineering, Nanjing, China). After transfecting T98G or U87 cells with PLEKH4A-shRNA or the PLEKHA4 overexpression vector, the cells were collected. The assays were conducted according to the instructions of the kit. Finally, the fluorescence microplate reader (BioTek, SYNERGY H1) was used for detection. ATP and lactate concentrations were computed using the following formula:

$$\mathrm\,(}\mathrm/\mathrm)=(}_}-}_})/(}_}-}_})\times \mathrm\times (\mathrm/\mathrm)$$

$$}(})=(}}_}}-}}_}})/(}}_}}-}}_}})\times }\times }$$

Luciferase reporter assay

The dual-luciferase PLEKHA4 and HOXD9 overexpression vectors were co-transfected into U87 cells. After 48 h, the relative luciferase activity and Renilla luciferase activity were detected, and the obtained data were analyzed and processed to obtain the differences between experimental and control group. All samples were detected in triplicates.

Immunofluorescence

Cells were fixed with 4% formaldehyde for 30 min, permeabilized with 0.5% Triton X-100 for 10 min, blocked with 1% BSA for 1 h, and incubated with primary antibody (anti-γ-H2AX, Abcam, Ab81299, 1:100, UK) at 4 °C overnight. Then, we added the diluted fluorescent secondary antibody (Alexa Fluor 488-labeled goat-Rabbit IgG (H + L), Beyotime, A0423, China) and incubated the cells for 1 h. Finally, the cells were treated with a mixture of anti-quenching sealing agent and DAPI, followed by imaging under a fluorescence microscope. The cell nucleus was dyed blue, and the positive results were dyed green.

Xenograft models

Eighteen male nude mice (18–22 g, 6 weeks old) were procured from Hangzhou Ziyuan Experimental Animal Technology Company (Hangzhou, China). The animals were retained at 22 ± 1 °C, under a 12 h light-dark cycle, and provided water ad libitum. Mice were randomly divided into three groups: shNC, shPLEKHA4-1 and shPLKHA4-3 groups (n = 6 per group), with 2 × 106 T98G cells injected into each mouse in their left axilla. Once the tumor grew, the tumor’s length and width were measured to calculate its volume every three days. After 33 days, the mice were euthanized, and the tumors were removed and weighed. The volume was calculated based on the measured length and width, along with the growth curve. Tumor volume (mm3) = 0.5 × length × width2.

Twenty-four nude mice were randomly divided into four groups (shNC + Vehicle, shNC + TMZ, shPLEKHA4 + Vehicle, shPLEKHA4 + TMZ) and inoculated with T98G-shNC or T98G-shPLEKHA4 stable transgenic cell lines (n = 6 per group). Then, 2 × 106/200 µl cell suspension was subcutaneously injected into the right forearm axilla of nude mice. After inoculation, different groups of nude mice were raised in cages, regularly fed with water and feed, and the bedding was changed regularly. One week later, the shNC + TMZ and shPLEKHA4 + TMZ groups were injected with TMZ (50 mg/kg) via intraperitoneal injection for five consecutive days, while the remaining groups were injected with an equal amount of solvent (intraperitoneal injection). Then, we measured the size of the tumor with a caliper every three days. After the administration, each group of experimental mice was euthanized, and the tumors were immediately removed and imaged. We weighed the tumor and measured the volume, then we fixed the tumor in 4% paraformaldehyde to prepare it for subsequent sectioning and pathology.

Intracranial models

6-week-old female Balb/c nude mice were used to construct an in intracranial model of glioma. The experimental process is as follows: 1 × 106 U87 cells stably expressing firefly luciferase (Fluc) were injected into the mouse brain 2 mm laterally, 2 mm posteriorly, and 2 mm deep using a stereotactic device. 10 days after implantation, small animal imaging was performed to confirm tumor occurrence, the mice were randomly divided into eight groups (n = 3 per group), then STAT3 inhibitor AG490 was administered via intraperitoneal injection every 2 days for 2 groups, and then intraperitoneal injection of TMZ (25 mg/kg) every 2 days for 2 groups, intraperitoneal injection of the same volume of solvent for other groups. The experimental groups are as follows: Vector + Vehicle, Vector + AG490, oePLEKHA4 + Vehicle, oePLEKHA4 + AG490, shNC + Vehicle, shNC + TMZ, shPLEKHA4 + Vehicle, shPLEKHA4 + TMZ.

IVIS imaging

Fluorescent drugs were injected into every mouse via intraperitoneal injection, after injection, the mice were anesthetized with isoflurance (RWD, China) at a set time point, then performed live fluorescence imaging using a small animal live imaging device (IVIS Lumina LT Series III, USA). The data were analyzed with the IVIS software (Living Imaging Software for IVIS).

Hematoxylin and Eosin (HE) staining

The tumor tissues were embedded in paraffin blocks and cut into 5 µm thick sections. The paraffin sections were baked for 45 min and placed in xylene I, xylene II and xylene III solution for dewaxing. The dewaxed sections were then hydrated in gradient alcohol (alcohol concentration: 100%, 95%, 85%, and 75%) for 5 min each. Then, we performed HE staining, controlling the staining time. Finally, we observed and imaged the sections under a microscope, and analyzed the collected images.

Immunofluorescence

The tumor tissues were embedded in paraffin blocks and cut into 5 µm thick sections. The sections were baked for 45 min and deparaffinized them in xylene I, xylene II and xylene III solution for 10 min each. Then placed in gradient alcohol (100%, 95%, 85%, and 75% alcohol concentration) for 5 min each. Then, antigen repair was performed using a 0.02 M sodium citrate buffer solution at high temperature for 15 min. Incubated with primary antibody (Ki67, proteintech, 27309-1-AP, 1:100, China) at 4 °C overnight. Incubated the fluorescent secondary antibody (Alexa Fluor 555-labeled donkey-rabbit IgG (H + L), Beyotime, A0453, China) at room temperature for 1 h. The sections were Sealed with DAPI containing quenching sealing agent. Immunofluoresence microscopy observation and photography (400×).

Tunel staining

We embeded the tumor tissues in paraffin blocks, and cut them into 5 µm thick sections. We baked and dewaxed the sections separately, and then TUNEL staining (Roche, 11684817910, Switzerland) was performed. The sections were incubated at 37 °C for 1 h, followed by staining and sealing with a mixture of anti-quenching sealing agent and DAPI. Finally, we observed and imaged the sections under a fluorescence microscope, and we analyzed the collected images to calculate the apoptosis rate.

Statistical analysis

All the data were presented as mean ± SD. Groups differences were determined using repeated ANOVA tests followed by Bonferroni correction and Student’s t-tests. Statistical significance was set at P < 0.05. For cell experiments such as qPCR, Western blotting, cell proliferation and cells apoptosis, we conducted three independent cultures. In each independent culture, we conducted 3 repeated experiments to evaluate the reproducibility of the experiment.