AAA237, an SKP2 inhibitor, suppresses glioblastoma by inducing BNIP3-dependent autophagy through the mTOR pathway

Cells were seeded in the 96-well plate at a density of 4 × 10³ cells/well and cultured at 37 °C with 0, 0.03, 0.1, 0.3, 1, 3, 10, 30, 100 μM AAA237 for 48 h and 72 h. After treatment, 10 μL of CCK-8 reagent was added into each well and incubated for 1 h. Then A450 was measured.

The supernatants of U251 and LN229 cell lysates were divided into two equal portions, one of which was used as the control group (DMSO), and the other was used as the drug incubation group and incubated at room temperature for 2 h. Then the two supernatants were divided into six equal portions, and were heated for 10 min at 45, 50, 55, 60, 65, and 70 °C, respectively, and cooled, and the cooled supernatants were added with loading buffer and were heated at 70 °C for 10 min, and then the SKP2 expression was detected by Western blot at the end of the process.

U251 and LN229 cells were inoculated in 96-well plates at a density of 4000 per well, and after incubation until the second day, the experimental group incubated the cells with AAA237 at concentrations of 0.3, 1, and 3 μM for 48 h and 72 h, respectively, and the control group was incubated with DMSO. The cells were then stained with EdU, Apollo 567, Hoechst 33342 and finally photographed with a fluorescence microscope (Nikon Eclipse Ti-U) following the steps in the kit.

About 3000 U251 and LN229 cells were put into six-well plate. After continuously being treated with AAA237 at 0, 0.3, 1, 3 μM for 2 weeks. After being washed, the clones were photographed. To do the 3D-matrigel experiment, the lower layer of gel was laid down first, and after waiting for about 20 min for the lower layer to solidify, the upper layer containing about 3000 U251 and LN229 cells was added. After 3–4 weeks of incubation, the formation of cell spheres can be observed under the microscope.

To determine the ability of the cells to migrate and invade after the administration of AAA237, 24-well plates with transwell chambers with an 8 μm pore size (Corning Costar, USA) were used. 400 μL of the suspension containing 2 × 10⁵ cells/mL cells was added to the upper chamber, and the plates were incubated for 3–4 h (the invasion required pre-spreading of Matrigel Matrix with pre-cooled serum-free and culture medium in a 1:7 ratio in the upper chamber). Once the cells were attached to the bottom, the culture medium was gently aspirated off of the upper chamber, and 200–300 μL of serum-free culture was added. 600 μL of 1640 culture medium containing 10–20% FBS was added to the lower chamber, and the plates were incubated for 19 (migration) or 24 h (invasion). After the incubation, the cells that successfully traversed the polycarbonate membrane were subjected to fixation using a 4% paraformaldehyde solution (G1101, Servicebio, China) for a duration of 15 min. Subsequently, these fixed cells were stained with a 1% crystal violet solution for a period of 30 min. The cells were visualized using a microscope manufactured by Nikon, a company based in Tokyo, Japan. Five fields per chamber were assessed for the numbers of migrated/invaded cells.

Supernatants from the cell lysates were obtained by RIPA after being centrifuged. The blots were subsequently subjected to blocking in a 5% fat-free milk solution for a duration of 2 h at room temperature, followed by gentle agitation overnight at 4 °C with the primary antibodies. The principal antibodies used in this study were Actin (Proteintech, Rosemont, USA), SKP2, P27, P21, BNIP3, p-mTOR, mTOR, P62, Beclin1, ATG5, and LC3II (Cell Signalling Technology, Danvers, USA) (1000x dilution) (the detailed information of antibodies was shown in Additional file 1: Table S1). The membranes were cut horizontally. After being washed, the blots underwent incubation with the appropriate HRP-linked secondary antibody (Cell Signaling, Danvers, USA). In this research, the P62 protein was stripped and re-probed after being probed of Beclin1, the P27 protein was stripped and re-probed after being probed of P21, the mTOR protein was stripped and re-probed after being probed of p-mTOR.

Next-generation sequencing was performed by Expandbiotech Corporation (Beijing, China) on cells lysed in TRIzol. Using the DESeq2 R package (1.20.0), differential expression analysis was conducted.

TRIzol Reagent was used to get total RNA, and a PrimeScript RT Reagent Kit was used to make cDNA. Real-time RT-PCR was used with the SYBR Premix Ex TaqIIKit to find the expression of differently expressed genes (the detailed information of sequence of primers was shown in Additional file 1: Table S2).

Based on the BNIP3 gene sequence and the principles of small molecule interfering RNA (siRNA) design, we designed and synthesized an siRNA targeting the human BNIP3 gene, 5ʹ-AAGGAACACGAGCGUCAUGAAdTdT-3ʹ. Logarithmic growth phase cells were selected and transfected with liposomal lipofectamin 3000-mediated interfering chain BNIP3 siRNA for U251 and LN229, respectively. After 48 h of transfection, mTOR protein level was measured by Western Blot.

As per the guidelines provided by the manufacturer, the cells were subjected to infection using a tandem mRFP-GFP-LC3 adenovirus. Briefly, 5×10⁴ U251 and LN229 cells were planted on a Glass Bottom Cell Culture Dish (801002, NEST, China) and infected for 24 h with tandem mRFP-GFP-LC3 adenovirus (1×10⁸ TU/mL) at MOI = 1. Laser confocal microscopy was used to examine cells after they had been infected. Autophagosomes are shown by yellow puncta, whereas autolysosomes are represented by red puncta.

Cells were fixed for 30 min at room temperature using an electron microscope fixative (G1102, Servicebio, China) and then postfixed for 2 h with 1% OsO4. After being dehydrated in gradient ethanol, the samples were infiltrated and embedded in epoxy resin (ZXBR, Spon 812). A transmission electron microscope (HITACHI HT7800; Tokyo, Japan) was used to photograph the ultrastructure of U251 and LN229 cells.

The changes in cell morphology were imaged to determine the effect of AAA237 on the viability and proliferation of GBM cells. At the 48-h time point, the administration of AAA237 resulted in a dose-dependent alteration in the morphology of U251 and LN229 cells (Fig. 1A). The IC₅₀ values for AAA237 on U251 cells at 48 h and 72 h were 0.485 μM, and 0.418 μM; and for LN229 cells, 0.407 μM, and 0.378 μM at 48, and 72 h (Fig. 1B, C) respectively. In order to get a deeper understanding of the effects of AAA237 on the survival and growth of glioblastoma multiforme cells, a CCK-8 experiment was carried out. The findings of the study revealed that the compound AAA237 had a suppressive impact on the growth of U251 and LN229 cells, with the degree of inhibition being dependent on both the dosage and duration of treatment (Fig. 1D, E). Altogether, the results demonstrated that AAA237 could inhibit the growth of GBM cells in vitro.

Skp2, a well-studied Skp2-SCF E3 ligase complex member, can attach K48-linked and K63-linked ubiquitin chains to a wide variety of substrates [25]. In several types of human cancer, high levels of Skp2 expression are linked to poor prognosis and unfavorable therapy outcomes [26]. AAA237 has apparent interaction with Skp2 protein. The Cellular Thermal Shift Assay (CETSA) assay, which utilizes the generation of a melting curve for a specific protein in cell lysates, was employed to ascertain the status of Skp2 as an intrinsic target of AAA237 inside the cellular milieu. The melting temperature (Tm) of a protein is significantly altered when a compound like AAA237 binds to it.

The experimental results demonstrated that AAA237 exhibited a notable enhancement in the thermal stability of Skp2 compared to the control group in U251 and LN229 cell lines, indicating that it binds directly to intracellular Skp2 (Fig. 2A–D). Subsequently, an examination was conducted to determine the possible impact of AAA237 on the protein expression of Skp2. The results shown that AAA237 exhibited a time-dependent inhibition of Skp2 protein expression in U251 and LN229, and the experimental intervention of AAA237 resulted in an upregulation of the expression levels of p21 and p27 in U251 and LN229 cell lines (Fig. 2E, F). These two proteins, p21 and p27, are well recognized as prominent substrates of Skp2.

Collectively, based on these findings, the binding of AAA237 to Skp2 would inhibit Skp2 expression and the degradation of p21 and p27.

The EdU-DNA synthesis test was performed to check the suppressive effects of AAA237 on cell growth. The results of this investigation indicate that the application of AAA237 at doses of 0.3, 1, and 3 μM for periods of 48 and 72 h resulted in a decrease in the proliferation rate of U251 and LN229 cells. Moreover, the inhibitory effect was seen to be dependent on both the dosage and duration of treatment (Fig. 3A–D). In line with these findings, the number of colonies that formed in cells treated with 0.3, 1, or 3 μM AAA237 was less than in the control cells in both Matrigel (Fig. 3E) and soft agar (Fig. 3F). Moreover, the compound AAA237 demonstrated in U251 and LN229 cells, there is a dose-dependent reduction of migration and invasion (Fig. 3G–J). Overall, in a dose-dependent manner, AAA237 demonstrated the ability to inhibit colony formation, migration, and invasion of GBM cells.

To examine the mechanism by which AAA237 affected GBM, an RNA-seq experiment was conducted to identify genes that exhibit differential expression in U251 and LN229 cells following treatment with 0.3 μM AAA237 for a duration of 48 h. In comparison to gene expression in control U251 cells, the cells treated with AAA237 had 360 up-regulated genes and 353 down-regulated genes (Fig. 4A). The findings derived from the KEGG pathway enrichment analysis indicated that the genes exhibited enrichment in pathways that are closely linked to the process of mitophagy, FoxO signaling pathway and MAPK signaling pathway, etc. (Fig. 4B). The findings obtained from the Gene Ontology (GO) enrichment analysis of the DEGs in U251 cells indicated that the genes enriched in the biological processes (BP) category were associated with the control of neuron death and response to hypoxia (Fig. 4C). In the cell components (CC) category, the DEGs were involved in the collagen trimer and vesicle lumen (Fig. 4D). The DEGs identified in the molecular function (MF) category showed involvement in extracellular matrix structural component and growth factor receptor binding (Fig. 4E). Similarly, in comparison to gene expression in control LN229 cells, the cells treated with AAA237 had 1019 up-regulated genes and 1425 down-regulated genes (Fig. 4F). The results derived from the KEGG pathway enrichment analysis indicated that the genes exhibited enrichment in autophagy, MAPK signaling pathway, etc. (Fig. 4G). GO enrichment analysis of the DEGs in U251 cells indicated that the genes enriched in the BP category were associated with cellular response to hypoxia (Fig. 4H). In the CC category, the DEGs were involved in the cell-cell junction (Fig. 4I). The DEGs identified in the MF category showed involvement in channel activity (Fig. 4J).

In comparison to gene expression in control LN229 cells, the cells treated with AAA237 had 1019 up-regulated genes and 1425 down-regulated genes (Fig. 4F). The outcomes of the KEGG pathway enrichment analysis indicated that the genes exhibited enrichment in pathways linked to mitophagy, PI3K-AKT signaling pathway and HIF-1 signaling pathway (Fig. 4G). The findings obtained from the Gene Ontology (GO) enrichment analysis of the differentially expressed genes (DEGs) in LN229 cells demonstrated that the genes enriched in the biological process (BP) category are mostly associated with the cellular response to hypoxia (Fig. 4H). In the CC category, the DEGs were involved in the basement membrane and perikaryon (Fig. 4I). The DEGs under the MF category were shown to be associated with channel activity and L-ascorbic acid binding (Fig. 4J).

BNIP3 is a protein related to the BH3-only family, which induces both cell death and autophagy [27]. The evidence suggests that BNIP3 induces cell death through multiple mechanisms [27, 28]. BNIP3 can initiate or enhance autophagy and its variation, mitophagy [29]. We overlapped the DEGs of RNA-seq results in U251 and LN229 after administering AAA237. Surprisingly, BNIP3 was upregulated as a hub gene in both U251 and LN229 cells after the administration of AAA237(the detailed information of overlapped DEGs of RNA-seq in U251 and LN229 cells was shown in Additional file 1: Table S3).

To confirm the mRNA and protein level of BNIP3 after the administration of AAA237, qRT-PCR and Western blot were performed. Consistent with the RNA-seq results, the mRNA level of BNIP3 was up-regulated as compared to the control group in a dose-dependent manner in both U251 (Fig. 5A–C) and LN229 (Fig. 5D–F). Meanwhile, the protein level of BNIP3 was upregulated in comparison with the control group in both U251 (Fig. 5G) and LN229 (Fig. 5H). Considering that previous studies have shown that BNIP3 has an inhibitory effect on mTOR, Western blot was conducted to verify the effect of BNIP3 on mTOR. The results showed that BNIP3 negatively regulated mTOR (Figure S1). Taken together, the mechanism of AAA237 on GBM was related to the change of BNIP3.

Transmission electron microscopy was used to analyze the morphology of U251 and LN229 cells after they were treated with AAA237 for 48 h in order to ascertain whether the compound might cause autophagy in GBM. As a positive control, rapamycin was also used to incubate the cells. In contrast to the untreated control group, the administration of either rapamycin or AAA237 resulted in the observed development of autolysosomes in U251 (Fig. 6A) and LN229 (Fig. 6B) cells. In addition, we utilized the mRFP-GFP-LC3 indicator system to assess alterations in autophagy flux, to investigate the impact of AAA237 on the dynamic fusion process between autophagosomes and lysosomes. The result showed that AAA237 could induce a significant increase of GFP-LC3 puncta in U251 (Fig. 6C) and LN229 (Fig. 6D) cells, indicating that AAA237 promoted the fusion of lysosomes and autophagosomes, resulting in the formation of autolysosomes. According to these findings, AAA237 increased the union of phagosomes and lysosomes, which resulted in cell autophagy.

mTOR, a serine/threonine kinase, serves as a pivotal regulator of cellular metabolism, exerting control over cell growth and proliferation in response to various stimuli. Notably, the signaling pathway associated with mTOR is disrupted in several clinical disorders [30‐32]. mTOR plays a crucial role in regulating autophagy [33, 34]. Considering that the previous transcriptome data suggested that the downstream mechanism may be associated with the PI3K pathway, the mTOR level has been checked in U251 and LN229 cells after the administration of AAA237. Meanwhile, Western blot was carried out to check critical proteins associated with autophagy induced by AAA237. These proteins included p-mTOR, mTOR, Beclin1, P62, LC3B and ATG5. According to the results (Fig. 6E, F), p-mTOR and mTOR were reduced. In contrast, the protein level of LC3-II, a recognized indicator of autophagy, exhibited a dose-dependent elevation subsequent to treatment with escalating doses of AAA237, suggesting that the medication elicited an autophagic reaction. The upregulation of Beclin1 expression resulted in an enhancement in phagophore and autophagosome production. The expression of ATG5 was found to be elevated, leading to an enhanced conversion of LC3-I to LC3-II. Additionally, a drop in p62 levels indicated unimpeded autophagy flow downstream, since p62 serves as a substrate for autophagolysosome destruction. These results above suggested that the induction of autophagy by AAA237 is achieved through the regulation of the mTOR-mediated pathway.

To elucidate how AAA237 activated autophagy, whether 3-Methyladenine (3-MA), an inhibitor of PI3K [35], could reverse AAA237-induced autophagy was examined. 3-MA is widely used as an autophagy inhibitor by inhibiting classIII PI3K [36, 37]. Firstly, the IC₅₀ of 3-MA in U251 and LN229 was calculated to choose a concentration with no obvious toxicity and no discernible impact on the cells. The IC₅₀ of 3-MA at 48 h in U251 and LN229 was 8.02 mM, 7.136 mM respectively (Fig. 6G, H). Consequently, we employed the 5 mM concentration of 3-MA in subsequent experiments. The CCK-8 assay was used to examine whether 3-MA could reverse the inhibition of cell proliferation caused by AAA237. Interestingly, AAA237-induced cell growth suppression was reversed by 3-MA (Fig. 6I, J). Simultaneously, morphology of GBM cells revealed that 3-MA reversed AAA237’s lethal impact on U251 and LN229 cells (Fig. 6K). Moreover, an EdU-DNA synthesis test was performed to verify the changeover of 3-MA in autophagy. Results showed that 3-MA neutralized the suppression of proliferation in U251 and LN229 cells (Fig. 6L, M). To confirm how 3-MA affected the autophagy flux resulting from AAA237 in GBM, the expression of LC3 and p62 was checked by Western blot. Results showed that the reduction of p-mTOR, mTOR, and p62 was reversed, while 3-MA reversed the increase of Beclin1, ATG5 and LC3II in U251 and LN229 (Fig. 6N, O). To sum up, 3-MA reversed the autophagy caused by AAA237. To summarise the two parts above, AAA237 can cause the onset of autophagy in GBM cells, whereas 3-MA can somewhat reverse the process.

To investigate the effect of compound AAA237 on tumor growth in vivo, LN229 cells were injected into the brains of mice in an orthotopic model and tumor growth was monitored by magnetic resonance imaging (MRI) (Fig. 7A). The AAA237 (15 and 45 mg/kg) treatment groups effectively retards the progression of tumors in situ (Fig. 7B) which confirmed the in vitro results. The mice in the AAA237 treatment groups had substantially decreased ultimate tumor volume and relative brain weight compared to the animals in the Vehicle group (Fig. 7C, D). Throughout the 21-day period of AAA237 treatment, no discernible instances of weight loss or aberrant behavior were observed (Fig. 7E). Meanwhile, AAA237 had no side effects on the major organs of the test mice (Fig. 7F). The results of H&E staining revealed that the tumor tissue in the Vehicle group exhibited a higher density compared to that in the AAA237 treatment group. Additionally, the protein expression of Ki67 was seen to decrease following the administration of AAA237 (Fig. 7G). Western Blot results of tumor tissues in mouse brain showed that when AAA237 was administered, the level of SKP2 was down-regulated, whereas the levels of SKP2 substrates P27 and P21 were up-regulated (Fig. 7H). Meanwhile, BNIP3 was upregulated in mouse intracerebral tumor tissues upon administration of AAA237 (Fig. 7I). Moreover, autophagy-related markers in mouse intracerebral tumor tissue showed that autophagy occurred when AAA237 was administered (Fig. 7J). Collectively, all the data indicated that the administration of AAA237 effectively suppressed the development of glioma tumors in vivo.