SK-N-BE (2) was obtained from Procell Life Science &Technology (Procell, Hubei, China); HL7702 was obtained from COBIOER (Cobioer Biosciences, Jiangsu, China); N2a and SH-SY5Y were kind gifts from Dr. Yuxian Shen (Anhui Medical University, Anhui, China); 10A, HUVEC and HepG2 were kind gifts from Dr. Kai Xue, Dr. Lei Shi, and Prof. Cong Li (Dalian Medical University, Liaoning, China), respectively; SK-N-BE (2) cell was maintained in Dulbecco’s modified Eagle’s medium (DMEM)/Ham’s F12 medium; HL7702 was maintained in RPMI 1640 medium; N2a,SH-SY5Y, and 10A cells were maintained in DMEM high glucose medium and HepG2 cells were maintained in Eagle’s Minimum Essential Medium (EMEM) in a humidified incubator with 5% CO₂ at 37°C. The culture medium was supplemented with 10% fetal bovine serum FBS (AusGeneX), 100 U/ml penicillin, and 100 mg/ml streptomycin. By the end of the study, the passage number of any cell line mentioned above was below 20.

The formalin-fixed-paraffin-embedded tissue samples, including one neuroblastoma, one ganglioneuroblastoma, and four olfactory neuroblastoma tissues were obtained from the First Affiliated Hospital of University of Science and Technology of China (USTC); eight neuroblastoma, one ganglioneuroblastoma, three olfactory neuroblastoma, and six retinoblastoma samples were obtained from the Second Affiliated Hospital of Dalian Medical University (DMU). We collected all patient-derived specimens under protocols approved by the Institutional Review Boards of the Second Affiliated Hospital of DMU and the First Affiliated Hospital of USTC.

Simvastatin, fluvastatin, and rosuvastatin was purchased from MCE (Shanghai, China). Abiraterone acetate was obtained from Aladdin (Shanghai, China). Polyclonal anti-β-actin, polyclonal anti-Ki67, polyclonal anti-HMGCR, polyclonal anti-SREBP1, polyclonal anti-Ki67, and anti-AR (N-20) antibodies were obtained from Santa Cruz (TX, USA). Monoclonal anti- SCAP, Monoclonal anti-CYP17A1 were obtained from Proteintech (Hubei, China), Monoclonal anti-SREBP2 were obtained from Abcam (Shanghai, China).

N2a, SH-SY5Y, SK-N-BE (2), 10A, HUVEC, HL7702 cells were plated in 96-well plates at 8,000 cells/well, respectively, in complete media in triplicate wells for each dose and cultured for 18 h. The neuroblastoma cell lines, N2a, SH-SY5Y, and SK-N-BE (2), and non-cancer cell lines, 10A, HUVEC, and HL7702, were treated with simvastatin and fluvastatin in a range of doses from 0 to 20 μM and with AA from 0 to 30 μM; N2a, SH-SY5Y SK-N-BE (2), were treated with rosuvastatin in a range of doses from 0 to 100 μM for 72 h. Statins dissolved in DMSO and AA dissolved in ethanol. After 72 h treatment, the medium was added 0.5 mg/ml MTT reagent and incubated at 37°C for 4 h. Subsequently, the supernatant was aspirated, and cells were lysed in 150 µl DMSO, then shaken 10 min at 37°C. The optical density (OD) was measured at 570 and 630 nm using a plate reader.

Cell lysis, protein extraction, and immunoblotting were performed as described previously (22).

Total RNA was extracted from cell pellets or fresh tissues using the TRIzol reagent (Invitrogen, USA). RT-PCR was undertaken with TransScript First-Strand cDNA Synthesis SuperMix (Transgen Biotech, Beijing, China) and 2 × EasyTaq PCR SuperMix (cwBiotech, Beijing, China) according to the manufacturer’s instructions. The following primer pairs were used: SCAP (Forward; 5′ CCCCAGGCTATGACTTCAGC-3′) and (Reverse: 5′-CCAAGCT CCAGATGGAACCC-3′), SREBF1 (Forward: 5′-CTGTTCCTGTGTGACCTGCT-3′) and (Reverse: 5′-CATGTAGGAACACCCTCCGC-3′), SREBF2 (Forward: 5′-CTGGGAGACATCG ACGAGAT-3′) and (Reverse: 5′-GACCTGGGTGAATGACCGTT-3′), HMGCR (Forward: 5′- GTCATTCCAGCCAAGGTTGT-3′) and (Reverse: 5′-GGGACCACTTGCTTCCATTA-3′), CYP17A1 (Forward:5′-TTCAGCCGCACACCAACTAT-3′) and (Reverse: 5′-GGATTCAAG AAACGCTCAGGC-3′), ACTB (β-actin, Forward: GCTCGTCGTCGACAACGGCT) and (Reverse: CAAACATGATCTGGGTCATCTTCTCT).

N2a, SH-SY5Y, and SK-N-BE (2) cells were cultured in six-well plates (5 × 10⁵cells/well) and incubated till they reached 90–100% confluence. The cells were then maintained in phenol red-free DMEM/DMEM-F12 with 2.5%FBS, in order to minimize the cell proliferation. A sterile 20 µl tip was used to create scratch wounds of the same width. The plates were then washed twice with phosphate-buffered saline (PBS) to remove the detached cells. Photos were taken at 0, 24, and 48 h, and the area covered by the cells enumerated the closure of the wounds. Each experiment was performed in triplicate.

Immunohistological analysis was performed as previously described (23). The dilution and incubation with antibodies were as follows: anti-AR 1:250, anti-SCAP 1:100, anti-SREBP1 1:20, anti-SREBP2 1:100, anti-HMGCR 1:10, anti-CYP17A1 1:200, and anti-Ki67 1:250. The sections were then counterstained with hematoxylin.

Four weeks old BALB/c nude mice of both sexes were purchased from Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China) and maintained under specific pathogen-free conditions. Male mice were castrated 5 days before cell inoculation. A total of 5 × 10⁶ SH-SY5Y cells were suspended in PBS to a final volume of 100 μl and injected subcutaneously into the flanks of a 6-week-old nude mouse. Then, all the mice were divided randomly into four groups: control group (Placebo, n = 7), simvastatin group (Sim, n = 6), abiraterone acetate group (AA, n = 6), and combination group (Com, n = 6) including three male mice in each group. Mice were then treated orally when the volume of the tumor reached 100 mm³ estimated with calipers as described previously (23) with drugs at the following doses and routes: simvastatin 20 mg/kg/day; AA 150 mg/kg/day; combined simvastatin 10 mg/kg/day with AA 75 mg/kg/day. Tumor was harvested after 10 days of dosing, then fixed half of a tumor in 4% buffered formaldehyde for IHC and the other half was stored −80°C for RT-PCR and Western Blots analysis.

All mice experiments were carried out with ethical committee approval and met the standards required by the Dalian Medical University Animal Care and Use Committee guidelines.

The study was based on a cohort of neuroblastoma patients described in the Oncomine database, analyzed the impact of two clusters of six genes (AR, SCAP, SREBF1, SREBF2, HMGCR, CYP17A1, or LC3, ULK2, ATG8L, ATG12, ATG14, ATG21) on the survival of NB patients. Data from patients aged 60 months or less (n = 133) were focused on for analysis. The index weight coefficient for grouping and 95% confidence interval were shown in Supplementary Table S1 . Cox regression was used to calculate the regression coefficient of each cluster of six genes and calculate the risk score for each patient ( Supplementary Tables S1A–D ). By median ranking, the patients were divided into high- and low-risk groups. The risk score for each patient was derived by multiplying the expression level of a gene by its corresponding coefficient (risk score = sum of Cox coefficient of gene multiplied by expression value of gene) (24).

All tests were carried out using SPSS (version 23.0; Chicago, USA). Kaplan-Meier curves were generated using GraphPad Prism 8.3. Comparisons between two cohorts were performed by the Log-Rank test.

SH-SY5Y cells and HepG2 cells were seeded in 6 cm dishes in 50% confluency. The culture medium was discarded and washed three times with PBS 12 h after seeding. High glucose DMEM or EMEM (Hyclone) supplied with 5% charcoal-stripped fetal bovine serum (cFBS, Hyclone) for a hormone-free condition. Cell culture supernatant was centrifuged at 10,000 rpm and collected 72 h after cFBS treatment and stored at −80°C until shipping to BGI, Shenzhen (https://en.genomics.cn/en-medical-diagnosiss-Testingservices-Nutritionals.html), in dry ice. The sample pretreatment, LC-MS/MS analysis, and data processing were performed according to BGI’s standard procedures (25). The relative steroid hormone level of SH-SY5Y was calibrated with HepG2 cells.

Statistical analysis was performed in GraphPad Prism Version 8 (GraphPad Software, San Diego, USA) statistical software. Data are presented as means ± standard division of at least three independent experiments. Data were analyzed by two-tailed Student’s t-test for comparisons between two groups, or one-way analysis of variance (ANOVA) with post hoc Bonferroni multiple comparison test for comparisons involving greater than two groups and p-values of p < 0.05 (∗), p < 0.01 (∗∗), p < 0.001 (∗∗∗), and p < 0.0001 (∗∗∗∗) were considered significant and highly significant, respectively.

Although the function of AR in neuroblastoma cells were investigated in the previous work, the expression of AR and other proteins in AR-SCAP-SREBPs-CYP17A1/HMGCR Axis had never been examined in clinical samples of neuroblastoma or other types of neuroblastic or neuroendocrine tumors. From 2010 to 2019, we collected clinical samples from NB (n = 9, retroperitoneal origin, mainly younger than 5 years old) (26), ganglioneuroblastoma (GNB, n = 2), olfactory neuroblastoma (ONB, n = 7), and retinoblastoma (RB, n = 6) patients in 1^st Affiliated Hospital of University of Science and Technology of Chima (USTC) and 2^nd Affiliated Hospital of Dalian Medical University (DMU) for IHC analysis ( Figures 1A–D ) to know the expression pattern of related proteins. Peripheral neuroblastic tumors are further classified based on their morphological features into NB, GNB, and ganglioneuroma (GN), known as the International Neuroblastoma Pathology Classification (INPC; Shimada system), which are relevant biologically and prognostically (27). Compared with ONB ( Figure 1C ) and RB ( Figure 1D ), the IHC staining of retroperitoneum-initiated NB showed significant higher (P < 0.05) expression of AR, SCAP, and especially the drug targets, HMGCR and CYP17A1 ( Figure 1E ). The result supported our hypothesis that AR signaling promoted the progression of neuroblastoma specifically. There was no significant difference in the intensity of SREBPs staining in the three types of neuroblastic tumors ( Figure 1E ). But in NB group SREBPs mostly localized to the nuclear area, whereas SREBPs staining appeared in the cytoplasm in ONB or RB group ( Supplementary Figure S1 ), which indicated the activation of SREBPs in NB (28). The result of IHC showed that although the three neuroblastic tumors are related, NB is special for the activation of AR-SCAP-SREBPs-CYP17A1/HMGCR axis.

It is still unknown that whether the expression of the six genes, AR, SCAP, SREBF1, SREBF2, HMGCR, and CYP17A1 are closely associated with neuroblastoma patient survival. We examined clinical outcomes in the Oncomine Neuroblastoma Dataset of 133 patients who were divided into high- and low-expression groups (the grouping method was described in the Material and Method section) according to the simultaneous expression level of the six genes, the index weight coefficient, and 95% confidence interval were shown in Supplementary Table S1 . Interestingly, upregulation of the six genes was related to unfavorable survival rate with statistical significance ( Figure 2A , p = 0.0072) in spite of MYCN-amplification status ( Figures 2B, C ). Meanwhile, the risk scores of 133 patients including MYCN-amplified (n = 39) and MYCN non-amplified (n = 93) were calculated ( Supplementary Tables S1B, C ). With this method, we evaluated whether the expression of six autophagy-related genes ( Supplementary Table S1D ) was associated with the same 133 patients’ survival to exclude the possibility that a random cluster of six genes expression might significantly impact on the survival rate ( Figure 2D ).

Based on the hypothesis mentioned above, statin or abiraterone acetate (AA) is supposed to show a growth inhibition effect on NB cells independently. As expected, statins and AA inhibited the proliferation of three NB cell lines, N2a, SH-SY5Y (both are MYCN non-amplified) and SK-N-BE (2) (MYCN amplified), in a dose-dependent manner ( Figures 3A–C ) analyzed with MTT method. Then we calculated the half-maximal inhibitory concentrations (IC50) of statins and AA in NB and normal cell lines at 72 h ( Table 1 ). Surprisingly, compared with simvastatin and fluvastatin, rosuvastatin displayed a remarkable higher IC50 ( Supplementary Figure S2 ). Moreover, to investigate if these drugs were toxic to non-cancer cell lines, the IC50 of human hepatocyte HL-7702, as well as human breast cell line 10A and endothelial cell line HUVEC was determined. All the three non-cancer cell lines showed less sensitive to simvastatin ( Table 1 and Supplementary Figure S2 ). Further, the efficacy of half-dose combination of a statin with AA on NB cells growth inhibition was compared with that of single treatment ( Figures 3D, E ). Evident synergistic effect was observed. For instance, the inhibition rate was about 10 and 20% with 1.5 μM simvastatin and 2.5 μM AA, respectively, while the inhibition rate was 70% with combined 0.75 μM simvastatin and 1.25 μM AA in N2a cells. But, astonishingly, the combination of drugs had limited effect on the growth of normal cells ( Supplementary Figure S3 ). The results suggested that combined HMGCER and CYP17A1 inhibition synergistically restrained neuroblastoma proliferation and this strategy showed good safety.

Concerning that reduced cholesterol content in plasma membrane may affect cell migration, the wound healing assay was applied to measure the effect of drugs on the migratory inhibition of NB cells. The migratory ability of neuroblastoma cells was largely reduced by combined treatment of statin and AA (p < 0.0001, Figures 4A–D ). Whereas the migration of cells with either one statin or AA treatment at 24 or 48 h did not show significant difference from that of cells in the control group, with an exception that the migration of N2a cells treated with 1.5 μM simvastatin at 48 h was significantly higher than that of the control group (P < 0.05) ( Figure 4B ). It suggested the combination of a statin and abiraterone acetate, even at half doses, can significantly decrease NB cell migration.

Our previous study suggested that NB cells may produce androgens for the growth of themselves (6). Here the production of hormones of cultured SH-SY5Y cells in comparison with HepG2 cells were determined with LC-MS/MS analysis. Interestingly, the testosterone level of culture supernatant of SH-SY5Y cells was nearly four times that of HepG2 cells ( Supplementary Figure S4A ) after culturing for 72 h in hormone free culture medium. The difference of gene expression of a series of enzymes playing key roles in cholesterol, testosterone, and dihydrotestosterone (DHT) synthesis was determined using SH-SY5Y and HepG2 cells in normal culture condition or steroid-deprivation culture condition. Compensatory elevated expression of these enzymes on transcriptional level was observed in SH-SY5Y cells treated with culture medium supplied with 5% charcoal-stripped FBS (cFBS) in contrast to normal culture condition, for example, mRNA of HMGCR, CYP11A1, an enzyme producing pregnenolone from cholesterol, CYP17A1 and SRD5A1, an enzyme producing dihydrotestosterone (DHT) from testosterone, dramatically increased due to cFBS culturing ( Supplementary Figure S4B ). The findings above proved that NB cells produced androgens for the need of growth.

We postulated that the combination of HMGCR and CYP17A1 inhibitors might have a synergistic effect on lowering the cholesterol synthesis, the subsequent testosterone production, and then reducing the activity of AR as a transcription factor. So, the expression of AR and related proteins was determined by RT-RCR and WB assays in three NB cell lines treated with the inhibitor(s) alone or combined for 24 h. Initially, as a target gene of AR, the mRNA level of SCAP dramatically decreased in SK-N-BE (2) cells treated with combined inhibitors compared with either fluvastatin or AA alone ( Figure 5A ). More than that, target gene (HMGCR and CYP17A1) expressions of SREBP1&2 were also detected. All genes’ expressions were significantly reduced under the presence of combination of drugs ( Figure 5A ). In consistent with the results on transcription level, the expression of related proteins in combination treatment group sharply decreased when compared with that in control group. In SH-SY5Y cells, although protein level of SCAP was not obviously altered when treated with a statin, AA or their combination, protein level of CYP17A1, HMGCR, and AR markedly decreased ( Figures 5B, C ). The expressions of SREBP precursors and mature forms (SREBP-p and SREBP-n, respectively, in Figures 5B, C ) were not significantly affected by a single drug treatment, but markedly decreased due to combined treatment. These results further prove that the combination of drugs can better inhibit the growth of NB than single treatment.

Although the IC50 of fluvastatin was lower than that of simvastatin when treating SH-SY5Y cells ( Figure 3B and Table 1 ), combined with 0.5 μM AA (1/5 of IC50), 2 μM Simvastatin (1/5 of IC50) or 2 μM Fluvastatin (1/3 of IC50) reached the inhibition rate of 50% (P < 0.001). So, simvastatin showed better synergistic effect than fluvastatin with AA ( Figure 3 ). Then, documented xenographic experiments with nude mice were investigated focusing on the dosages and the inhibitory activities of simvastatin and AA ( Table 2 and Supplementary Table S2 ). As is shown in Supplementary Table S2 , the dosages of simvastatin were 0.5–200 mg/kg/day for oral gavage, 1.25–40 mg/kg/day for intradermal injection, 5–10 mg/kg/day for intraviral injection, and 2.5–11 mg/kg/day for subcutaneous injection, respectively ( Table 3 ). The concentration of AA for oral treatment was 7.5–285.7 mg/kg/day, whereas 3.5–200 mg/kg/day for intraperitoneal injection. Based on an overall consideration, we planned to orally treat the mice with simvastatin (20 mg/kg/day) or AA (150 mg/kg/day) as single treatment and the mixture of simvastatin (10 mg/kg/day) with AA (75 mg/kg/day) as the combined treatment in the following experiment.

Given the significant synergy demonstrated with combined simvastatin–AA treatment in vitro, we sought to test the efficacy of the combination strategy in neuroblastoma xenograft models. SH-SY5Y xenografts were implanted subcutaneously and allowed to grow to approximately 100 mm (3) at which point they were randomized to one of four arms for oral administration daily: vehicle, simvastatin (Sim 20 mg/kg), AA (150 mg/kg), or half-dosage combination (Sim 10 mg/kg and AA 75mg/kg). We observed tumor growth delay in all the three drug treatment groups ( Figure 6A ). It was astonishing that after 10 days of oral gavage, the mean tumor volume in the combined treatment group was about 30% of that in the control group. Further, we observed significant reduction of expressions of related genes in the combination treatment group compared with either simvastatin or AA group ( Figure 4B ). The expressions of related proteins examined by WB and IHC analysis also revealed obvious inhibition with combination treatment ( Figures 6B, C ). IHC staining intensity of Ki-67, a marker reflected the proliferation rate of tumor cells (29), and AR were both significantly weaker in combination treatment group than in the single treatment group or the control group ( Figure 6D ). The results suggested that combination of both drugs had a clinical potential to treat NB patients even with both reduced dosages.