MTHFD1 regulates the NADPH redox homeostasis in MYCN-amplified neuroblastoma

MYCN amplification is an independent poor prognostic factor in patients with high-risk neuroblastoma (NB). Further exploring the molecular regulatory mechanisms in MYCN-amplified NB will help to develop novel therapy targets. In this study, methylenetetrahydrofolate dehydrogenase 1 (MTHFD1) was identified as the differentially expressed gene (DEG) highly expressed in MYCN-amplified NB, and it showed a positive correlation with MYCN and was associated with a poor prognosis of NB patients. Knockdown of MTHFD1 inhibited proliferation and migration, and induced apoptosis of NB cells in vitro. Mouse model experiments validated the tumorigenic effect of MTHFD1 in NB in vivo. In terms of the mechanism, ChIP-qPCR and dual-luciferase reporter assays demonstrated that MTHFD1 was directly activated by MYCN at the transcriptional level. As an important enzyme in the folic acid metabolism pathway, MTHFD1 maintained the NADPH redox homeostasis in MYCN-amplified NB. Knockdown of MTHFD1 reduced cellular NADPH/NADP+ and GSH/GSSG ratios, increased cellular reactive oxygen species (ROS) and triggered the apoptosis of NB cells. Moreover, genetic knockdown of MTHFD1 or application of the anti-folic acid metabolism drug methotrexate (MTX) potentiated the anti-tumor effect of JQ1 both in vitro and in vivo. Taken together, MTHFD1 as an oncogene is a potential therapeutic target for MYCN-amplified NB. The combination of MTX with JQ1 is of important clinical translational significance for the treatment of patients with MYCN-amplified NB.


Introduction
Neuroblastoma (NB) generally originating from the sympathetic nervous system or adrenal glands is the most common malignancy in infants, accounting for about 6% of all cancers in childhood.MYCN ampli cation, which occurs in 20-30% of patients with NB, is an important risk strati cation factor and strongly correlated with a poor prognosis in high-risk NB [1,2].Due to the presence of an α-helix structure on the surface, it is di cult for drug inhibitors to target MYCN directly.Small molecule JQ1 is a BET inhibitor that disrupts the correlation of BET proteins with transcription factors, suppresses the expression of oncogenes, and eventually causes the cessation of tumor cell growth.JQ1 can downregulate the expression of MYCN in NB, but its effect seems unsatisfactory when used alone [3,4].Therefore, it is of great signi cance to further explore the molecular mechanism of MYCN-ampli ed NB for nding potential novel therapy targets for patients with MYCN-ampli ed NB.
Accumulating evidence has indicated the importance of metabolic abnormality in tumor progression [5][6][7][8].The process of one-carbon metabolism, which plays an irreplaceable role in tumorigenesis, mainly consists of the folate cycle and methionine cycle, through which purine, adenosine, and other metabolites are generated.As a crucial part of carbon metabolism, the folate cycle can provide much energy for tumor cell proliferation [9], which occurs in the cytoplasm and mitochondria.Methylenetetrahydrofolate dehydrogenase 1 (MTHFD1), as a key enzyme in the folate cycle, existing in the cytoplasm, contains three functional groups with different catalytic activities, which is closely related to NADPH production and adenosine purine nucleotide metabolism, including dehydrogenase, cyclohydrolase, and tetrahydrofolate synthetase [10][11][12].MTHFD1 is up-regulated in multiple tumors such as cholangiocarcinoma, colorectal cancer, melanoma, and hematologic malignancies [13][14][15], which is related to tumor apoptosis, proliferation, migration and drug resistance.However, the role of MTHFD1 in NB and its molecular mechanisms remain unclear.This study aimed to explore the clinical signi cance, biological function, and regulatory mechanisms of MTHFD1 in MYCN-ampli ed NB.Meanwhile, the effect of MTHFD1 knockdown or anti-folic acid metabolism drug methotrexate (MTX) combined with BRD4 inhibitor JQ1 was also evaluated in this study, aiming to identify the potential of MTHFD1 as a therapy target for NB and reveal the importance of targeted inhibition of folic acid metabolism pathway for NB therapy.

Cell lines and clinical samples
Human NB cell lines SK-N-BE (2) were cultured in MEM/F12 (1:1) supplemented with 10% fetal bovine serum (FBS), human NB cell lines IMR32 were cultured in MEM supplemented with 10% FBS, and human NB cell lines SK-N-AS were cultured in DMEM supplemented with 10% FBS.All cell lines had been authenticated and free from mycoplasma.Human tissue samples were obtained from Sun Yat-sen University Cancer Center under protocols approved by the Institutional Review Board.Written informed consent was obtained from each tissue donor and all procedures were conducted following the medical ethical guidelines.Demographics of NB patients from our center are summarized in Supplemental Table 1.

Immunohistochemistry (IHC) assay
The IHC assay was performed on 57 NB tissue samples obtained from our center, as previously described [16].Anti-MTHFD1 mouse mAb was used in this assay.The staining results were evaluated based on the intensity and the proportion of positive-stained tumor cells.The intensity was scored as follows: 0negative; 1 -weak; 2 -moderate; 3 -strong.The proportion of positive cells was scored as follows: 0 -less than 25%; 1-25-50%; 2-50-75%; 3-75-100%.The composite staining score (the product of the above two scores) of 0-4 was considered a low expression, and that of 5-9 was considered a high expression.
Transfection, lentiviral transduction, and RNAi SK-N-BE(2) cell lines stably expressing MTHFD1 shRNA were generated through lentiviral transduction followed by puromycin selection.To obtain the shRNA-expressing virus, shRNA vectors were cotransfected with the lentivirus packaging plasmids into HEK293T cells using Lipofectamine 3000 (Invitrogen, USA).Fresh media were added after 24 hours, and viral supernatants were collected at 48 hours.Target cells were infected with viral supernatant (diluted at 1:1 with fresh media; 8 µg/mL polybrene), added with fresh DMEM 24 hours later and selected with 1 µg/mL puromycin.SK-N-AS cells overexpressing MYCN were generated using the same packaging system and MYCN plasmid.For MTHFD1 and MYCN knockdown, siRNA targeting each gene (RiboBio, Guangzhou, China) was transfected into SK-N-BE(2) or IMR32 cells using Lipofectamine RNAiMAX (Invitrogen, USA).
Realtime quantitative polymerase chain reaction (RT-qPCR) Total RNA was isolated with TRIzol according to the manufacturer's instruction, and subjected to RT-PCR using SYBR PrimeScript RT Master Mix, with β-actin as an internal control.The primer sequences are shown in Supplemental Table 2.

Cell proliferation assay and IC50 assay
The cell proliferation rate was measured using Cell Counting Kit 8 (CCK-8, Dojindo, Shanghai, China).Speci cally, cells were seeded at a density of 2000 cells per well in a 96-well plate.24 hours later, the number of viable cells was detected with Cell Titer Glo reagent daily.In IC50 assay, cells were seeded at a density of 2000 cells per well in a 96-well plate for 24 hours, and then treated with JQ1 for 96 hours.The number of viable cells was detected with Cell Titer Glo reagent.

Colony formation assay
In colony formation assay, 5×10 4 cells were seeded in a 6-well plate and incubated with indicated compounds for 10-14 days until the obvious colony was formed.Then the plate was gently washed and stained with crystal violet for colony visualization.

Cell apoptosis analysis
In cell apoptosis assay, Annexin V-FITC/PI Apoptosis Detection Kit (4A Biotech, Beijing, China) was used according to the manufacturer's instructions.Brie y, cells were collected, washed, resuspended in binding buffer, sequentially stained with Annexin V-FITC and PI, and immediately analyzed by ow cytometry (SP6800, Sony, Japan).

ChIP assay
Cells were xed with 1% formalin and sonicated.Then sheared DNA was incubated with antibodies against MYCN (Santa Cruz Biotechnology, sc-53993) or IgG control (Santa Cruz Biotechnology, USA, sc-2027).DNA-protein-antibody complexes were incubated with Protein A Agarose/Salmon Sperm DNA Beads (Merck, Germany, CS204457).The beads were washed sequentially with gradient salt buffer and eluted in 1% SDS/NaHCO 3 .Finally, ChIP DNA was analyzed by qPCR.

Luciferase assay
Cells were plated in a 96-well plate before transfection.Empty pGL3 luciferase vector, pGL3 expressing MTHFD1 or pGL3 MTHFD1 mutant was transiently co-transfected into HEK293T cells, using a control Renilla luciferase plasmid (pRL-TK).The test plasmid: control plasmid ratio was 50:1.Luciferase activities were measured 48 hours later using Dual-Luciferase Reporter Gene Assay Kit (Vazyme, Nanjing, China, DL101-01).Fire y luciferase activities were normalized to the value of Renilla luciferase control and described by the average of triplicates.

Animal experiments
All animal experiments were conducted on 3 to 4-week-old female NCG mice in accordance with animal protocols approved by the Institutional Animal Care and Use Committee (IACUC) of Sun Yat-sen University.In MTHFD1 knockdown xenograft experiment, 1×10 7 SK-N-BE(2) cells with stable MTHFD1 knockdown were implanted subcutaneously into the mice in the ank.Then subcutaneous tumors were harvested for ex vivo imaging followed by H&E or IHC staining for histological analysis.For the combined administration model, the mice were implanted with 1×10 7 wide-type SK-N-BE(2) cells.After about 2 weeks, the tumor could be touched.When the tumor volume reached about 100 mm 3 , the mice were randomly divided into four groups for drug administration.For the combination of shMTFHD1 and JQ1, JQ1 was intraperitoneally injected at 50 mg/kg/day for 14 consecutive days.For the combination of MTX and JQ1, JQ1 and MTX were intraperitoneally injected at 50 mg/kg/day and 20 mg/kg/day, respectively, for 14 consecutive days.Tumor volumes were measured using an electronic caliper and calculated using the following formula: tumor volume = length (mm) × width (mm) × width (mm) × 0.5.The tumor length reaching 2 cm in some mice could be regarded as the endpoint.

Bioinformatics analysis
The TARGET database of NB was searched (https://ocg.cancer.gov/programs/target),and the RNA sequencing data and basic clinical information of patients were included.Patients were divided into two groups: MYCN ampli cation group and MYCN non-ampli cation group.The limma package was used to screen the DEG with |logFC| >0.378, and an adjusted P-value < 0.001 was de ned as the threshold of screening.Then R language was used to perform the KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis, Spearman correlation analysis and Kaplan-Meier survival prognosis analysis.

Statistical analysis
All statistical analyses were performed using GraphPad Prism 8.0 software by two-way ANOVA, Student's t-test, or χ 2 test.Kaplan-Meier survival analysis followed by log-rank test was performed.All functional assays were independently repeated at least 3 times and the results were expressed as mean ± SD. p < 0.05 was considered statistically signi cant.

MTHFD1 was up-regulated in MYCN-ampli ed NB and correlated with the poor prognosis of NB patients
To identify the potential pathogenic genes associated with MYCN-ampli ed NB, publicly available genomic data from the TCGA database were rst analyzed, which included 66 cases of MYCN-ampli ed NB and 177 cases of MYCN-non-ampli ed NB.The DEGs between the two groups were screened out, including 697 up-regulated genes and 1298 down-regulated genes.The heat map and volcano plot of these DEGs are shown in Fig, 1A and 1B, respectively.The KEGG functional enrichment analysis on the 697 signi cantly up-regulated DEGs suggested that they were mainly enriched in folate-carbon metabolism, cell cycle, DNA replication, cysteine and methionine metabolism, glycine-serine and threonine metabolism, p53 pathway, etc. (Fig. 1C).
Folic acid metabolism is an important part of one-carbon metabolism, and is associated with high invasiveness signature of high-risk NB [17].However, the regulatory mechanism of one-carbon metabolism in NB and its relationship with MYCN have not been clari ed yet, thus the one-carbon metabolism pathway was selected for further study.Eight genes have been found to be enriched in this pathway: MTHFD1, MTHFD2, ATIC, MTHFD2L, SHMT2, DHFR, MTHFD1L and TYMS.Among them, MTHFD1 is reported to be a potential oncogene for multiple cancers [18], but its role and molecular regulatory mechanism in NB remain unclear.On the above basis, MTHFD1 was selected as the object of interest for further study.
Based on the TCGA database, the expression of MTHFD1 was compared between the MYCN-ampli ed and the MYCN-non-ampli ed groups.The results showed a signi cantly higher expression of MTHFD1 in the MYCN-ampli ed group (P < 0.0001, Fig. 1D).Besides, a high expression of MTHFD1 implied a worse prognosis (P = 0.0022, Fig. 1E).As shown by Spearman correlation analysis based on the TCGA database, MTHFD1 was positively correlated with MYCN at the transcriptional level in NB (P < 0.01, Fig. 1F).The results of RT-qPCR on 21 fresh tumor tissue samples of NB patients from our center further indicated a positive correlation between MTHFD1 and MYCN at the transcriptional level (R = 0.8179, P < 0.0001, Fig. 1G).In addition, 6 NB patients with a high expression of MYCN protein also had a relatively higher level of MTHFD1 (Fig. 1H).Based on the above ndings, it is speculated that MTHFD1 expression is positively correlated with MYCN in NB.
Then, the expression levels of MTHFD1 were validated in 2 MYCN-ampli ed NB cell lines (SK-N-BE(2) and IMR32) and 3 MYCN-non-ampli ed NB cell lines (SK-N-AS, SK-N-SH and SHSY-5Y).It was found by RT-qPCR and WB that the expression level of MTHFD1 in MYCN-ampli ed NB cells was signi cantly higher than that in MYCN-non-ampli ed NB cells, indicating a positive correlation between MTHFD1 and MYCN in NB cell lines (Fig. 1I and 1J).
To further verify the clinical signi cance and prognostic value of MTHFD1 expression level in NB, IHC staining was performed on 57 tumor samples from NB patients at our center (Fig. 1K).Then Kaplan-Meier survival analyses (Fig. 1L) were conducted.The results revealed that NB patients exhibiting high expression levels of MTHFD1 had worse OS and PFS than those with low expression levels of MTHFD1.Besides, the correlation between MTHFD1 expression and the clinicopathological characteristics was also analyzed among 57 NB patients (Supplemental Table 3).The results indicated that there was no signi cant relationship between MTHFD1 expression and patient's age, gender, INSS stage or COG risk classi cation.However, the ampli cation status of MYCN was possibly related to the expression level of MTFHD1 since the P-value was 0.058, which ought to be veri ed with a larger sample size of tumor tissues.According to the results above, it is concluded that MTHFD1 is up-regulated in MYCN-ampli ed NB, and a high MTHFD1 expression is associated with a poor prognosis in NB patients.

MTHFD1 regulated the proliferation, apoptosis and migration of NB cells in vitro
To investigate the potential function of MTHFD1 in NB, 2 MYCN-ampli ed NB cells (SK-N-BE(2) and IMR32) were selected to validate the effects of MTHFD1 knockdown by shRNA on their biological phenotypes.The knockdown e ciency of MTHFD1 was rstly veri ed with WB, and shMTHFD1#1 and shMTHFD1#2 were selected for further veri cation (Fig. 2A).The results of CCK-8 assay indicated that NB cells proliferated at a signi cantly slower rate after genetic knockdown of MTHFD1 (Fig. 2B).Colony formation assay also revealed the anti-proliferation effects on SK-N-BE(2) and IMR32 cells after MTHFD1 suppression (Fig. 2C and 2D).Meanwhile, knockdown of MTHFD1 signi cantly induced the apoptosis of SK-N-BE(2) and IMR32 cells, as proved by ow cytometry (Fig. 2E and 2F).The migration of NB cells was obviously weakened after knockdown of MTHFD1 (Fig. 2G and 2H).To sum up, MTHFD1 promotes NB tumorigenesis in vitro by improving the proliferation and migration and inhibiting apoptosis of MYCNampli ed NB cells.

MTHFD1 exerted a tumorigenic effect in vivo
To further determine the tumorigenic effect of MTHFD1 in vivo, the MYCN-ampli ed NB SK-N-BE(2) cells were subcutaneously injected into the NCG mice to establish the xenograft model.The mice were divided into 3 groups: shNC, shMTHFD1#1 and shMTHFD1#2 (Fig. 2I and 2J).As shown in Fig. 2K, the tumors in shNC group grew faster than those in the other two groups.Moreover, tumor weight in shNC group was larger than that in the other two groups (Fig. 2L), suggesting an anti-proliferation effect of MTHFD1 knockdown.IHC staining validated that the tumors had lower expressions of Ki67 after MTHFD1 knockdown (Fig. 2M), which indicated a lower proliferation rate.

MTHFD1 was transcriptionally activated by MYCN in NB
The above ndings indicated that knockdown of MTHFD1 led to obvious proliferation suppression of MYCN-ampli ed NB cells both in vitro and in vivo, suggesting the tumor-promoting effect of MTHFD1 in MYCN-ampli ed NB.Meanwhile, MTHFD1 was positively correlated with MYCN at the mRNA and protein levels.As previously reported, the expression of MTHFD1 is signi cantly high in MYCN-ampli ed NB, but no further veri cation has been performed [19].As a member of the MYC transcription factor family, MYCN is dysregulated in various tumors and regulates the transcription of multiple oncogenes [20,21].Based on these ndings, it is speculated that MTHFD1 may be the target gene of MYCN.To verify our speculation, MYCN was rstly knocked down in MYCN-ampli ed NB cells.The down-regulation of MTHFD1 was found, consistent with the results of WB (Fig. 3A).Then MYCN was overexpressed in MYCN-non-ampli ed NB SK-N-AS cell lines, and the up-regulation of MTHFD1 was observed as expected (Fig. 3B).These results revealed that MYCN regulates the MTHFD1 expression in NB.
To further investigate the possible molecular mechanisms by which MYCN regulates MTHFD1 expression, the possible binding sites to MYCN at the MTHFD1 promoter regions (about 1500 bp upstream from the transcription start site) were predicted based on the PROMO online database (http://alggen.lsi.upc.es/cgibin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3/),and a potential binding site to MYCN was found (5'-CACGTG-3') (Fig. 3C).Then ChIP-qPCR was performed and it was found that MYCN was enriched at the MTHFD1 promoter region.Compared with the MYCN-non-ampli ed SK-N-AS cells, the MYCN-ampli ed SK-N-BE(2) cells showed an enhanced enrichment at the promoter region of MTHFD1 (Fig. 3D).The results of dual-luciferase reporter assay demonstrated that the activation of MTHFD1 promoter was signi cantly decreased after knockdown of MYCN (Fig. 3E), while a signi cantly enhanced promoter activity of MTHFD1 was observed after overexpression of MYCN in SK-N-AS cells (Fig. 3F).When the MTHFD1 promoter binding site (5'-CACGTG-3') was mutated to 5'-ACATGT-3', the promoter activity was signi cantly reduced in mutant group compared with that in wild type group (Fig. 3G).Rescue assay showed that knockdown of MTHFD1 with siRNA partially reversed the proliferation-promoting effect resulting from the overexpression of MYCN (Fig. 3H and 3I).Knockdown of MTHFD1 also partially reversed the anti-apoptosis effect of MYCN (Fig. 3J and 3K).The results above demonstrated that MTHFD1 is a direct target gene of MYCN and can be transcriptionally up-regulated by MYCN.

MTHFD1 maintained redox homeostasis in NB
MTHFD1 has been found to induce NADPH production and reduce the ROS content in cholangiocarcinoma cells [22].MTHFD1 can also promote tumorigenesis and lead to drug resistance by regulating NADPH homeostasis in the one-carbon metabolism process in acute myeloid leukemia [23].The in uence of MTHFD1 on the folic acid metabolism in NB has not been elucidated yet.Here the role of MTHFD1 in regulating NADPH redox homeostasis in NB was investigated.The results revealed that suppression of MTHFD1 reduced both NADPH/NADP + and GSH/GSSG ratios (Fig. 4A and 4B), and obviously increased cellular ROS content in SK-N-BE(2) and IMR32 cells (Fig. 4C and 4D).
Excessive intracellular ROS can induce cytotoxicity and lead to apoptosis of tumor cells [24,25].The results revealed that MTHFD1 knockdown induced apoptosis of NB cells (Fig. 2E and 2F).To investigate whether MTHFD1 affects apoptosis through redox homeostasis in NB, the antioxidant N-Acetyl-L-cysteine (NAC) was used after MTHFD1 knockdown in NB cells.It was found that the enhanced apoptosis caused by MTHFD1 knockdown could be partially reversed by NAC in NB cells (Fig. 4E and 4F), indicating that MTHFD1 affects apoptosis through regulating redox homeostasis.

Knockdown of MTHFD1 enhanced the anti-tumor effect of JQ1 in NB
MTX is an inhibitor of folic acid metabolism, which can lead to a decrease in the expression level of MTHFD1 [26].MYCN-ampli ed NB cells are more sensitive to MTX than MYCN-non-ampli ed NB cells [27,28].Besides, the BRD4 inhibitor JQ1 can reduce the expression of MYCN and inhibit tumor growth in MYCN-ampli ed NB [29].It was reported that the combined use of MTX and JQ1 exerts a synergistic antitumor effect in hematologic tumors [30].Based on the above background, the combination effect of inhibiting the folic acid metabolism by silencing MTHFD1 and decreasing MYCN expression by JQ1 in the treatment of MYCN-ampli ed NB was investigated.
The IC50 value calculated in the CCK-8 assay showed that the MYCN-ampli ed NB cells (SK-N-BE(2) and IMR32) had lower IC50 values, suggesting that they are more sensitive to JQ1 (Fig. 5A).JQ1 at a concentration of 0.1 µM was selected for further experiments.The growth curves revealed that JQ1 signi cantly enhanced the growth inhibition induced by MTHFD1 knockdown in a time-dependent manner (Fig. 5B).JQ1 could also distinctly enhance the apoptosis induced by MTHFD1 suppression in SK-N-BE(2) and IMR32 cells (Fig. 5C and 5D).Furthermore, mouse model experiments validated that the inhibitory effect of MTHFD1 knockdown on proliferation was signi cantly strengthened in NB cells treated with JQ1 (Fig. 5E-G).

MTX augmented the anti-tumor effect of JQ1 in NB
Next, the combination effect of JQ1 and the folic acid inhibitor MTX was explored.MTX at a working concentration of 0.15 µM was used.The results of experiments in vitro demonstrated that in JQ1 + MTX group, the inhibitory effect on cell proliferation was more obvious than that in single drug group (Fig. 6A), and the proportion of apoptotic SK-N-BE(2) and IMR32 cells was dramatically increased (Fig. 6B and 6C).Consistently, experiments in vivo also validated that the combination of JQ1 and MTX caused signi cant proliferation suppression of MYCN-ampli ed NB (Fig. 6D-F).
To sum up, MTHFD1 is an oncogene in MYCN-ampli ed NB, and it regulates redox homeostasis and promotes the malignant progression of NB.In terms of mechanism, MTHFD1 is transcriptionally upregulated by MYCN.Inhibition of MTHFD1 augments the anti-tumor effect of JQ1 in MYCN-ampli ed NB (Fig. 6G).

Discussion
MYCN ampli cation is an important risk strati cation factor for NB.Inhibition of MYCN seems to be a promising therapeutic approach to improve the outcome of patients with high-risk NB.However, due to the special α-helical structure on the surface of MYCN protein, developing small molecule inhibitors directly targeting MYCN remains challenging.At present, researchers are enthusiastically searching for key transcriptional targets at the downstream of MYCN, aiming to explore treatment options for MYCNampli ed NB.In this study, the DEGs between MYCN-ampli ed and MYCN-non-ampli ed NB patients were identi ed by bioinformatics analysis based on the TCGA database.The KEGG functional enrichment analysis showed that carbon metabolism was one of the main pathways with signi cant changes, among which MTHFD1 was the major DEG.Subsequently, bioinformatics analysis and veri cation in NB tissues and cell lines con rmed the positive correlation between MTHFD1 and MYCN.MTHFD1 was highly expressed in MYCN-ampli ed NB, and a high expression of MTHFD1 was associated with a poor prognosis of NB patients, suggesting that MTHFD1 may be associated with the malignant progression of MYCN-ampli ed NB.
Previous studies indicated that MTHFD1 is highly expressed in a variety of tumors and plays crucial roles in the malignant progression of tumors [23,31].In cholangiocarcinoma, MTHFD1 reduces the ROS content in tumor cells, and induces resistance to gemcitabine, thus enhancing the progressive phenotype of tumors [22].MTHFD1 maintains redox homeostasis in the folic acid metabolism process in metastatic colorectal cancer cells and serves as a potential therapeutic target for colorectal cancer [32].In lung cancer, knockdown of MTHFD1 signi cantly increases the percentage of apoptotic tumor cells [33].In this study, it was found that MTHFD1 played a tumor-promoting role in MYCN-ampli ed NB, and it promoted proliferation and migration but inhibited apoptosis of NB cells.Animal experiments in mice also revealed the tumorigenic role of MTHFD1 in MYCN-ampli ed NB cells in vivo.These results were consistent with previous ndings that MTHFD1 serves as an oncogene in tumors.
Recent studies have reported that metabolic reprogramming in MYCN-ampli ed NB is dependent on the activation of one-carbon metabolism pathway, and MTHFD1 is identi ed as one of the signi cant DEGs in this pathway based on the sequencing results [19,34,35].In addition, several key enzymes in the carbon metabolism process are highly expressed in NB and exert a tumor-promoting effect [36][37][38][39].
These ndings suggested that the carbon metabolism pathway is abnormally activated in MYCNampli ed NB.In this study, it was also veri ed that the carbon metabolism pathway was activated in MYCN-ampli ed NB.Subsequent experiments con rmed that MYCN could bind to the MTHFD1 promoter region and transcriptionally activate the expression of MTHFD1 in MYCN-ampli ed NB, validating the molecular mechanism of MYCN regulating MTHFD1.Targeted inhibition of MTHFD1 may provide a promising therapy for MYCN-ampli ed NB.
Folic acid metabolism belongs to carbon metabolism, during which several enzymes are considered potential tumor-speci c therapeutic targets.The folic acid metabolizing enzyme MTHFD1L can generate NADPH to defend against oxidative stress and promote tumor growth in liver cancer [40].ATF4 and c-MYC promote tumor cell growth by synergistically regulating MTHFD2, and MTHFD2 is identi ed as a biomarker or therapeutic target for prostate cancer [41].Inhibiting MTHFD1 in chronic myelogenous leukemia causes a signi cant decrease in the proliferative ability of tumor cells both in vitro and in vivo [26].Similarly, knocking down MTHFD1 can reduce the anti-oxidant stress ability of tumor cells, thus inhibiting the distant metastasis of melanoma [42].
As a key enzyme in folic acid metabolism, MTHFD1 plays an important role in maintaining redox homeostasis, NADPH production and nucleotide metabolism.NADPH is important for ROS generation and catalyzes the GSSG reaction, thereby producing GSH and then inducing ROS generation.A high level of ROS in cells causes DNA damage and mitochondrial activation, inducing cell senescence and apoptosis [43,44].However, the regulatory effect of MTHFD1 on redox homeostasis in NB has not been elucidated yet.In this study, it was veri ed that knockdown of MTHFD1 decreased both NADPH/NADP + and GSH/GSSG ratios and increased cellular ROS content in SK-N-BE(2) and IMR32 cells, which was dependent on MYCN.The results suggested that MTHFD1 maintains redox homeostasis in MYCNampli ed NB, and inhibiting MTHFD1 in tumor cells may weaken the resistance to oxidative stress, thus preventing the malignant progression of tumors.MTX is the inhibitor of folic acid metabolism, which can cause the loss of chromatin-associated MTHFD1 [22,23,26], but the anti-tumor effect of MTX alone in NB seems unsatisfactory due to high toxicity and side effects [45,46].Despite this, its anti-tumor effect in MYCN-ampli ed NB is better than that in MYCNnon-ampli ed NB.The BET inhibitor JQ1 can down-regulate the MYCN expression in MYCN-ampli ed NB [3], but its effect is also poor.The combination of MTX and JQ1 has been veri ed to synergistically inhibit the proliferation of CML cells both in vitro and in vivo [28].In this study, it was found that MTX and JQ1 synergistically exerted an anti-tumor effect in MYCN-ampli ed NB.Then the folic acid metabolism was inhibited by MTHFD1 knockdown, and it was found that MTHFD1 knockdown enhanced the anti-tumor effect of JQ1.Subsequently, the anti-tumor effect of MTX combined with JQ1 was veri ed in vivo and in vitro, which provided a preliminary basis for the treatment of MYCN-ampli ed NB with the combination of these two drugs.
In conclusion, this study revealed the tumor-promoting role of MTHFD1 in NB for the rst time, and also elucidated the molecular mechanism by which MTHFD1 was directly transcriptionally regulated by MYCN.MTHFD1 maintains redox homeostasis and protects tumor cells from oxidative stress.Genetic knockdown of MTHFD1 or application of folic acid inhibitor MTX combined with JQ1 can synergistically inhibit tumor progression in MYCN-ampli ed NB.Thus, MTHFD1 may be a potential therapy target for MYCN-ampli ed NB.

1 MTHFD1
AbbreviationsNB neuroblastoma MTHFD1 methylenetetrahydrofolate dehydrogenase 1 DEG differentially expressed gene (I) RT-qPCR analysis of MTHFD1 and MYCN mRNA expression in MYCN ampli ed (SK-N-BE(2) and IMR32) and MYCN non-ampli ed NB cell lines (SK-N-AS, SH-SY5Y and SK-N-SH).β-actin was used as loading control.The mRNA levels in SK-N-AS being designated as 1.0.(J) WB analysis of MTHFD1 and MYCN protein expression in MYCN ampli ed and MYCN non-ampli ed NB cell lines.The original full length of western blots were provided as Supplementary Fig 2.

Figure 2
Figure 2 (A) WB analysis evaluating the knockdown e ciency of MTHFD1 with shRNAs in SK-N-BE(2) and siRNAs in IMR32 cells.The original full length of western blots were provided as SupplementaryFig 3.

(
B) CCK-8 assay of viability in SK-N-BE(2) and IMR32 cells transfected with MTHFD1 shRNA, MTHFD1 siRNA or the scramble control.(C) Colony formation images of MTHFD1 knockdown in SK-N-BE(2) and IMR32.(D) Histogram representative of of colony formation assays in SK-N-BE(2) and IMR32 cells after MTHFD1 knockdown.(E, F) Changes of apoptosis rates in SK-N-BE(2) and IMR32 cells after MTHFD1 genetic knockdown.Apoptotic cells detected by Annexin V in three independent experiments.(G, H) Transwell migration assays demonstrating that knockdown of MTHFD1 decreases the migratory abilities of NB cells.(I) Xenograft tumors was established in NCG mice subcutaneously implanted with control and MTHFD1knockdown NB cells.(J) Photograph and comparison of excised tumor size.(K) Tumor volumes were recorded on the indicated days.(L) The tumor weights were measured.(M) Representative IHC images were stained with hematoxylin and eosin (H&E) or MTHFD1 and Ki67 antibodies (Scale bars: 100μm).Data are presented as the mean ± SD of three independent experiments.(C) and (K): Two-way-ANOVA.(E), (G), (I) and (L): Student's t test.*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.