Abstract

   FBXW7 is a tumor suppressor gene that regulates metabolism and is
   associated with the onset and progression of colorectal cancer (CRC)),
   however, the precise mechanism whereby FBXW7 participates in the
   metabolic reprogramming of CRC remains unclear. Here, the research aims
   to reveal the association between the expression of FBXW7 and clinical
   variables and to investigate the molecular mechanism by which FBXW7
   plays a critical role in the development of CRC. The clinical
   importance of FBXW7 in CRC was determined by immunohistochemistry.
   Non-targeted metabolomics was utilized to explore the role of FBXW7 in
   the metabolic regulation of CRC. Low expression of FBXW7 was associated
   with poor prognosis in individuals with CRC, both at the mRNA and
   protein levels. FBXW7 over-expression inhibited CRC cell growth, colony
   formation, migration, and invasion. Non-targeted metabolomics unveiled
   that FBXW7 over-expression directly caused the deprivation of arginine
   which led to downmodulation of mTOR signaling pathway; meanwhile,
   FBXW7-related metabolites were primarily concentrated in the mTOR
   signaling pathway. In summary, the research identified a novel
   mechanism of action of FBXW7 in CRC. The research findings provide a
   theoretical foundation for the prognostic prediction and therapeutic
   planning of CRC based on metabolic reprogramming.

Introduction

   Colorectal cancer (CRC) is the third most prevalent malignant tumor and
   the second leading cause of cancer-related mortality [[36]1,[37]2].
   China has the highest number of new cases and fatalities from CRC in
   the world and is distinguished by a younger age of commencement, a
   higher degree of malignancy, and greater resistance to medication
   [[38]3,[39]4]. At present, CRC treatment is primarily focused on
   surgery along with radiation, chemotherapy, and targeted therapy;
   however, for patients who are not diagnosed early, there is typically
   little hope of success, even when resorting to drastic surgery.
   Simultaneously, with a 5-year survival rate of only 15%, the complete
   treatment shows insufficient efficacy in patients with recurrence and
   metastasis [[40]1,[41]5]. The FDA has authorized various targeted
   medications for CRC in recent decades; however, only 20% of patients
   benefit from them [[42]6]. Most patients with advanced CRC have no
   medications available for backup treatment, and monitoring the
   metastasis and recurrence is still based on pan-tumor markers (CEA,
   CA124, and CA199) with low specificity and sensitivity [[43]7,[44]8].
   Therefore, it is necessary to investigate the pathophysiology of the
   high-incidence CRC in China and identify new diagnostic and therapeutic
   targets to improve preventive and treatment outcomes.

   Metabolic reprogramming is a hallmark of malignant tumors
   [[45]9–[46]11]. The activation of oncogenes and/or inactivation of
   tumor suppressor genes changes the metabolism of cancer cells,
   encouraging tumor development and progression [[47]12–[48]14].
   Post-translational protein modifications regulate signaling pathways
   associated with cancer metabolism [[49]15–[50]18], and ubiquitination
   is the most prevalent post-translational protein modification
   [[51]19–[52]21]. For example, AKT-PI3K-mTOR signaling pathway-related
   proteins can be ubiquitinated by multiple E3 ligases, thus regulating
   glucose and lipid metabolism in tumors [[53]22]. Recent studies have
   also implicated aberrant expression and activity of E3 ligase in the
   molecular etiology and pathophysiology of cancer incidence and
   progression [[54]23]. Therefore, understanding the involvement of E3
   ligases is key to revealing the mechanisms of metabolic reprogramming
   in cancer. FBXW7 is located on chromosome 4q31q.3 and encodes
   SKP1-CUL1-F-box-protein (SCF) E3 ubiquitin ligase, a member of the
   F-box protein family [[55]24]. It targets ubiquitination to destroy
   several oncogenes (including c-Myc, Notch1, Notch4, c-Jun, and cyclin
   E) and takes part in many crucial biological processes, including
   metabolic reprogramming, DNA damage repair, cell cycle, immune
   response, and cell death [[56]25]. However, data on FBXW7 metabolic
   regulation are far from complete.

   In this study, a tissue microarray containing 51 pairs of CRC and
   normal tissues was used to reveal the association between the
   expression of FBXW7 and clinical variables via immunohistochemistry
   (IHC). Through non-targeted metabolomic detection in
   FBXW7-overexpressing CRC cell lines, the molecular mechanism by which
   FBXW7 plays a critical role in the development of CRC was explored. The
   low FBXW7 expression was found to be related to poor prognosis in
   patients with CRC and that FBXW7 functions as a tumor suppressor gene
   in CRC cells, inhibiting the development of CRC via the mechanism:
   reduction in arginine production through metabolic reprogramming and
   direct down-regulation of mToR signaling. Most notably, metabolic
   enzymes were identified in tumor samples from 20 patients with CRC.
   Using data from The Cancer Genome Atlas (TCGA) Program database, the
   relationship between FBXW7 and CRC-related metabolic genes was
   verified. The research findings indicate a putative mechanism by which
   FBXW7 loss-of-function contributes to the progression of CRC, implying
   a potential therapeutic target for people with low expression of FBXW7.

Materials and methods

Cell culture and transfection

   The colorectal cell lines Lovo and HCT116 were stored at General
   Surgery Department of Shanxi Bethune Hospital. Cells were grown in DMEM
   high glucose media containing 10% FBS, 100U/ml penicillin, and 100g/ml
   streptomycin. Short Tandem Repeats (STR) was used to identify all cell
   lines that were free of mycoplasma infection.

   The full-length sequence of FBXW7 gene was constructed using pCDNA3
   vector. Liposome 2000 (Invitrogen, USA) was used for plasmid
   transfection, according to the product instructions. The transfection
   method was introduced in our previous study [[57]26].

Western blot

   The cells were cultured for 4 days in the presence of either 0, 50 or
   100 μM Arg. The cell precipitation was treated with a suitable amount
   of radio immunoprecipitation assay (RIPA) buffer including protease and
   phosphatase inhibitors, and the total protein was extracted by cold
   incubation. The protein concentration was determined using the BCA kit.
   The protein was transferred to a Polyvinylidene Difluoride (PVDF)
   membrane after separation on a 10% SDS-polyacrylamide gel. After
   blocking for 2 hours at room temperature with 5% skim milk, the protein
   was incubated with particular antibody overnight at 4°C. The next day,
   the second antibody was incubated at room temperature for 1 hour, and
   GelView was utilized to detect protein imprinting. Western blotting
   using FBXW7 antibody was used to evaluate the protein level in
   colorectal cell lines, and GAPDH was employed as the internal reference
   protein.

qPCR

   The quantitative real-time PCR (qRT-PCR) analysis was performed as
   previously described. The method was described in a previous study
   [[58]26]. Total RNA was extracted from cells and reverse transcribed
   into cDNA. All qPCR experiments included a no template control and were
   done in triplicate and repeated three times at least. Specific primer
   sequences were used for qPCR. The primer sequences used in this study
   are listed in Table in [59]S1 Table.

MTT assay

   The single cell solution was injected in a 96 well plate at a density
   of 4×10^3/well to detect changes in the viability of CRC cells. At 24h,
   48h, 72h, and 96h, 20μl 5mg/ml Methylthiazolyldiphenyl-tetrazolium
   bromide (MTT) (Invitrogen) was added to each well for 4 hours before
   discarding the mixture. At room temperature, 150μl Dimethyl sulfoxide
   (DMSO) was applied to each well for 15 minutes. An enzyme labeled
   equipment was used to detect absorbance at 490nm. Each experiment had
   five replicates, as well as at least three distinct experiments.

Cell migration and invasion assay

   To detect cell migration and invasion, we employed a Biocoat Matrigel
   invasion chamber. With 5×10^4 cells/well, a single cell suspension
   without FBS was injected in the top chamber of a 24-well plate, and
   full media was added to the lower chamber. The top chamber was removed
   after 24 hours and treated with 4% paraformaldehyde. The migrating
   cells to the sub-membrane surface were examined using an Olympus
   microscope (Japan) after 0.1% crystal violet staining. Four visual
   areas were chosen at random for manual counting. Before beginning the
   invasion experiment, cover the membrane surface with Matrigel (BD
   bioscience). Each experiment had three holes, and at least three
   separate experiments were carried out.

Colony formation assay

   In 6-well plates with full DMEM high glucose media, cells were planted
   at 500 cells per well and allowed to develop for 2 weeks. Cells were
   fixed with 4% paraformaldehyde and stained with 1% crystal violet on
   day 14 before being manually counted. The experiment was performed in
   triplicate and repeated three times at least.

Flow cytometry apoptosis detection

   Flow cytometry apoptosis detection. CRC cells were detached with
   EDTA-free trypsin, centrifuged to collect cell pellet, washed with PBS,
   resuspended with 500 μl binding buffer, mixed with 5 μl of Annexin
   V-FITC (Solarbio life Science, China), and 10 μl of Propidium Iodide
   (Solarbio Life Science, China). After incubation in the dark at room
   temperature for 20 min, the cells were detected by flow cytometry
   (Beckman Coulter Life Science, USA). We did this experiment at least
   three times.

Immunohistochemistry

   Shanghai Xinchao Biological Co., Ltd. provided Tissue Microarrays (TMA)
   (HcolA180Su14). To detect FBXW7, immunohistochemistry (IHC) was used.
   Briefly, tissues were fixed in 4% paraformaldehyde and embedded in
   paraffin. Sections were treated with a particular antibody (Abcam) at
   an optimal dilution (1:200) overnight at 4°C. The slides were then
   stained with hematoxylin and stained with PV8000 (Zhongshan) and the
   Diaminobenzidine (DAB) detection kit (Maixin). All pictures were taken
   at Axio scan z1. The quantity of the protein of interest was determined
   using the Aperio Nuclear v.9 program. SPSS 19.0 was used for
   statistical analysis.

Sample preparation

   The cell precipitation was extracted in 50 μL of 15% phosphoric acid
   with 10 μL of 75 μg/mL 4-methylvaleric acid solution as IS and 140 μL
   ether. Subsequently, the samples were centrifuged at 4°C for 10 min at
   12000 rpm after vortexing for 1 min and the supernatant was transferred
   into the vial prior to LC-MS analysis.

LC-MS analysis

   The LC analysis was performed on a ACQUITY UPLC System (Waters,
   Milford, MA, USA). Chromatography was carried out with an ACQUITY UPLC
   ® HSS T3 (150×2.1 mm, 1.8 μm) (Waters, Milford, MA, USA). The column
   maintained at 40°C. The flow rate and injection volume were set at 0.25
   mL/min and 2 μL, respectively. For LC-ESI (+)-MS analysis, the mobile
   phases consisted of (C) 0.1% formic acid in acetonitrile (v/v) and (D)
   0.1% formic acid in water (v/v). Separation was conducted under the
   following gradient: 0~1 min, 2% C; 1~9 min, 2%~50% C; 9~12 min, 50%~98%
   C; 12~13.5 min, 98% C; 13.5~14 min, 98%~2% C; 14~20 min, 2% C. For
   LC-ESI (-)-MS analysis, the analytes were carried out with (A)
   acetonitrile and (B) ammonium formate (5mM). Separation was conducted
   under the following gradient: 0~1 min, 2% A; 1~9 min, 2%~50% A; 9~12
   min, 50%~98% A; 12~13.5 min, 98% A; 13.5~14 min, 98%~2% A; 14~17 min,
   2% A.

   Mass spectrometric detection of metabolites was performed on Q Exactive
   (Thermo Fisher Scientific, USA) with ESI ion source. Simultaneous MS1
   and MS/MS (Full MS-ddMS2 mode, data-dependent MS/MS) acquisition was
   used. The parameters were as follows: sheath gas pressure, 30 arb; aux
   gas flow, 10 arb; spray voltage, 3.50 kV and -2.50 kV for ESI(+) and
   ESI(-), respectively; capillary temperature, 325°C; MS1 range, m/z
   81–1000; MS1 resolving power, 70000 FWHM; number of data dependant
   scans per cycle, 10; MS/MS resolving power, 17500 FWHM; normalized
   collision energy, 30%; dynamic exclusion time, automatic.

Statistical analysis

   According to receiver operating characteristic (ROCs) curves, all
   samples and subgroup samples were classified as FBXW7[high] or
   FBXW7[low]. Rank sum and chi-squared (χ^2) tests were used to examine
   differences in the expression levels of FBXW7 in colorectal tissues
   with other clinic-pathological characteristics. To evaluate the
   cumulative survival time across various FBXW7 groups with
   clinic-pathological variables, Kaplan-Meier estimator and the log-rank
   test were performed. A Cox proportional hazards regression model was
   utilized to perform univariate and multivariate survival analyses. All
   tests were conducted at least three times, and the results were
   expressed as mean ± SD. For statistical examination of the data
   comparing the two groups, Student’s t-test was utilized. For the data
   above the two groups, one-way analysis of variance (ANOVA) and
   Dunnett’s test were performed. Spearman correlation analysis was
   performed in GraphPad Prism 7 software to elucidate the relationship
   between FBXW7 and the expression of other genes. The SPSS 19.0
   statistical software program (IBS SPSS, Armonk, NY, USA) was employed
   to process all data. A p value of less than 0.05 was considered
   statistically significant.

Results

Low expression of FBXW7 protein is associated with poor prognosis in patients
with CRC

   IHC was utilized to determine the expression of FBXW7 protein in 51 CRC
   and matched normal tissues using a tissue microarray. The findings
   revealed that FBXW7 protein was localized in the nucleus and cytoplasm
   of CRC and normal tissues ([60]Fig 1A). FBXW7 has three subtypes
   (FBXW7α, FBXW7β, and FBXW7γ) that are found in the nucleus, cytoplasm,
   and nucleolus, respectively. The most expressed subtype, FBXW7α,
   inhibits tumors via ubiquitination of oncogenic proteins. Therefore,
   the ratio of FBXW7 nuclear to cytoplasmic expression in tumor tissues
   was analyzed based on the ROC curve to elucidate the association
   between FBXW7 protein levels and ANOVA in CRC, and the optimal critical
   value of 1.875 with the maximal Youden index was determined. All CRC
   cases were divided into two groups based on FBXW7 levels: FBXW7[low] (≤
   1.875) and FBXW7[high] (> 1.875). The OS for FBXW7[low] patients ranged
   from 7 to 78 months (median, 69 months). The FBXW7[high] OS ranged from
   13 to 77 months, with a median of 73 months. The rank-sum test
   indicated that patients in the FBXW7[low] group had a shorter total
   cumulative survival time than those in the FBXW7[high] group (p =
   0.029). Kaplan-Meier survival analysis also suggested that the OS of
   the FBXW7[high] group was significantly longer than that of the
   FBXW7[low] group (p = 0.041), and the cumulative survival time of
   FBXW7[high] patients was significantly longer than that of FBXW7[low]
   patients in male (p = 0.027), aged 60 years old (p = 0.010), at
   pathological grade 3 (p = 0.039), and with TNM stage I + II (p = 0.045)
   ([61]Fig 1B). Furthermore, the research discovered that while there was
   no significant difference in the cumulative survival time between
   FBXW7[high] and FBXW7[low] patients without lymph node metastases,
   FBXW7[high] patients had a longer cumulative survival time than
   FBXW7[low] patients.

Fig 1. Prognostic values of CRC patients with different expression levels of
FBXW7 protein.

   [62]Fig 1
   [63]Open in a new tab

   (A) Representative immunohistochemistry images of FBXW7 expression in
   tumor tissues and normal tissues from paraffin-embedded formalin-fixed
   CRC tissue microarrays including 51 tumors and corresponding non-tumor
   tissues. Scale bars represent 200 μm. (B) Kaplan-Meier survival curves
   of patients with different expression levels of FBXW7 in overall
   population and in patients with different genders, ages, tumor grades,
   lymph node metastases, and TNM stages.

Low expression of FBXW7 mRNA is associated with poor prognosis in patients
with CRC

   In addition, TCGA data for 279 patients with CRC were examined and the
   mRNA expression level of FBXW7 was found to be related to age, sex, TNM
   stage, T pathology, and lymph node metastases. According to ROC curve
   analysis, samples were split into two groups: FBXW7[low] (≤ 0.42985)
   and FBXW7[high] (≥ 0.42985). Kaplan-Meier survival analysis revealed
   that FBXW7[low] patients had a shorter survival time than FBXW7[high]
   patients. The OS showed statistically significant relationships with
   the age, sex, lymph node metastasis, and TNM stage. In particular, the
   OS of FBXW7[low] patients was shorter than that of FBXW7[high] patients
   in CRC samples from patients aged > 60 years, of male sex, without
   lymph node metastases, and with TNM stage I + II ([64]Fig 2A).

Fig 2. Prognostic values of CRC patients with different FBXW7 mRNA expression
levels.

   [65]Fig 2
   [66]Open in a new tab

   (A) Kaplan-Meier survival curves of patients with different FBXW7
   expression levels in overall population and in patients with different
   genders, ages, pathology T, tumor grades, lymph node metastases, and
   TNM stages. (B) Multivariate and univariate analysis by the Cox
   proportional hazards regression model in overall population.

   The Cox proportional hazards regression model for univariate and
   multivariate analyses was adopted to assess the predictive value of
   FBXW7 mRNA levels for cumulative survival of patients with CRC.
   Univariate analysis indicated that T pathology (HR, 1.92; 95% CI,
   1.136–3.238; p = 0.015) and low FBXW7 expression (HR, 1.83; 95% CI,
   1.129–2.969; p = 0.014) were significant risk factors for OS. Further
   multivariate analysis confirmed that FBXW7 was an independent predictor
   for predicting the overall cumulative survival status (HR, 2.202; 95%
   CI, 1.305–3.716; p = 0.003) ([67]Fig 2B).

Overexpression of FBXW7 inhibits the proliferation, invasion, and migration
of CRC cells

   To clarify biological functions of FBXW7 in CRC, endogenous expression
   levels of FBXW7 in several CRC cell lines were measured by western blot
   assay. The different CRC cell lines showed different expression levels
   ([68]S1 Fig). Among them, HCT-8 and Lovo CRC cells with low endogenous
   expression were used for transfecting FBXW7 over-expression plasmid.
   The efficiency of FBXW7 over-expression, as detected by western blot,
   was 4.98 and 5.3, respectively ([69]Fig 3A). Functional analysis
   indicated that FBXW7 may decrease the proliferation, invasion, and
   migration ability of CRC cells ([70]Fig 3B–3D). The induction of
   apoptosis in cells over-expressing FBXW7 was detected by Annexin
   V-FITC/PI double staining using flow cytometry; the data suggested that
   the number of apoptotic cells was significantly increased ([71]Fig 3E),
   implying that FBXW7 also functions as a tumor suppressor gene in CRC.

Fig 3. FBXW7 acts as a depression-gene affecting CRC cell growth, colony
formation, cell migration, and apoptosis.

   [72]Fig 3
   [73]Open in a new tab

   (A) Over-expression efficiency of FBXW7 in HCT8 and Lovo cells was
   assayed by western blot. (B) FBXW7 over-expression inhibited the
   proliferation of HCT8 and Lovo cells. (C) The number of colony
   formation in CRC cells was reduced in the FBXW7 over-expression group
   compared to the control. (D) FBXW7 over-expression markedly inhibited
   HCT8 and Lovo cell migration. (E) The proportion of apoptosis in HCT8
   cells was significantly increased in the FBXW7 over-expression group.
   All data were presented as the mean ± standard deviation and three
   independent experiments. * p < 0.05, ** p < 0.01, *** p < 0.001.

Pathway and metabolite alterations in CRC cells are affected by FBXW7
over-expression

   Non-targeted metabolomic analysis was performed on FBXW7-overexpressing
   Lovo and control cells to detect distinct metabolites and expound the
   molecular mechanism of action of FBXW7 in CRC. In positive-ion mode,
   1075 differential metabolites were identified, whereas 2934
   differential metabolites were found in negative-ion mode. The principal
   component analysis (PCA) was performed on the dataset. The PCA score
   plot revealed a pattern of group separation. In the score plot, QC
   samples were grouped in a compact space, indicative of satisfactory
   stability and repeatability of the system ([74]S2 Fig). An OPLS-DA
   score plot was created using this dataset. The FBXW7 over-expression
   group differed considerably from the control group and demonstrated
   adequate modeling and prediction abilities (R^2[X] = 0.541, R^2[Y] =
   0.999, Q^2[Cum] = 0.787) ([75]Fig 4A). The statistical test was carried
   out to determine the p-value, the OPLS-DA dimension reduction method
   was used to compute the variable significance in projection (VIP), and
   the fold change was utilized to calculate the multiple differences
   across groups. The strength and interpretability of the content of each
   metabolite component in the sample classification and discrimination
   were assessed, and metabolite screening was performed. The criteria for
   screening differential metabolites were VIP > 1, p < 0.05, and fc >
   1.5. The first-level material list of samples revealed 1075
   differential metabolites, of which 381 were up-regulated and 694 were
   down-regulated ([76]Fig 4B). The metabolites were retrieved using the
   Human Metabolome Database (HMDB) ([77]http://www.hmdb.ca), Massbank
   ([78]http://www.massbank.jp/), LipidMaps
   ([79]http://www.lipidmaps.org), mzCloud ([80]https://www.mzcloud.org)
   confirmation annotation, and MS/MS fragment patterns. There were 133
   distinct metabolites, of which 24 were up-regulated and 109 were
   down-regulated. The signaling pathways enriched by differential
   metabolites, such as the mToR signaling pathway, central carbon
   metabolism in cancer, pyrimidine metabolism, and nicotine addiction,
   indicate that FBXW7 may govern the incidence of CRC via these pathways
   ([81]Fig 4C). Among these, the mTOR signaling pathway showed the
   highest effect value, indicating that the metabolites detected in this
   pathway contributed the most.

Fig 4. Key metabolic pathway components altered in FBXW7 over-expression
cells.

   [82]Fig 4
   [83]Open in a new tab

   (A) An OPLS-DA score plot was created using the dataset of LC-MS. (B)
   Volcano plot of differential metabolites. (C) Pathway-enrichment
   analysis of differential metabolites.

   Specifically, the FBXW7-OE group contained less L-arginine than the
   control group. Arginine has been shown to directly activate the mToR
   signaling pathway ([84]Fig 5A). Therefore, the metabolites involved in
   arginine production were investigated. The quantities of isocitrate,
   citrate, citrulline, succinate, and glutamate associated with arginine
   synthesis decreased ([85]Fig 5B). Furthermore, qPCR was utilized to
   assess the expression of relevant metabolic enzyme-coding genes in the
   cell types before and after FBXW7 over-expression. The results revealed
   that the expression levels of the arginine synthesis metabolic enzymes,
   aminoacylase-1 (ACY1), argininosuccinate lyase (ASL),
   glutamic-oxaloacetic transaminase 1 (GOT1), GOT2, isocitrate
   dehydrogenase 2 (IDH2), malate dehydrogenase 2 (MDH2), and succinate
   dehydrogenase complex flavoprotein subunit A (SDHA), were significantly
   down-regulated in FBXW7 over-expression cells ([86]Fig 5C), indicating
   that the inhibitory effect of FBXW7 on arginine production may arise
   from changes in the expression of the associated metabolic enzymes. To
   test formally whether FBXW7 regulates mTORC1 by reducing the quantities
   of arginine, CRC cells engineered to over-express FBXW7 were leveraged.
   In contrast to normal cells, the abundance of phosphorylated (active
   form) forms of S6K1 in FBXW7-OE cells diminished, however, increasing
   extracellular concentrations of Arg from 0 to 100 μM in a
   dose-dependent manner enhanced the mTORC1 activity, as detected by the
   phosphorylation of its substrate S6K1 ([87]Fig 5D).

Fig 5. The change in the expression of mTOR pathway-related molecules.

   [88]Fig 5
   [89]Open in a new tab

   (A) mTOR signaling pathway-related metabolic pathways. (B) The
   metabolite changes of arginine biosynthesis and TCA cycle. (C) q-PCR
   was utilized to detect the mRNA level of metabolic enzymes in arginine
   biosynthesis and TCA cycle. (D) FBXW7 regulates mTORC1 by reducing the
   quantities of arginine. Statistical analysis was performed with a
   two-sided t test. * p < 0.05, ** p < 0.01, *** p <0.001.

The expression of genes involved in arginine biosynthesis is negatively
correlated with that of FBXW7

   The relationship between FBXW7 and arginine-related metabolic
   enzyme-coding genes in CRC was explored using TCGA data. In the 329 CRC
   tissue samples, the expression of FBXW7 was negatively correlated with
   those of ACY1, ASL, ARG, IDH2, MDH2, SDHA, GOT1, GOT2, and GPT2
   ([90]Fig 6A). Furthermore, IHC tests were carried out on 20 CRC
   tissues; the findings revealed that the expression of FBXW7 was
   negatively correlated with that of ASL, MDH2, and IDH2 at the protein
   level ([91]Fig 6B). These findings suggest that FBXW7 may down-regulate
   arginine production in CRC cells via metabolic reprogramming and limit
   CRC growth by decreasing the activation of mToR pathway.

Fig 6. FBXW7 is negatively correlated with ASL and MDH2 in CRC.

   [92]Fig 6
   [93]Open in a new tab

   (A) FBXW7 mRNA was negatively correlated with ACY1, ASL, GOT2, MDH2,
   NAGS, SDHA, and IDH2 in 329 CRC samples based on TCGA data. (B)
   Immunohistochemistry was used to detect the expression of FBXW7, MDH2,
   and ASL in CRC samples. Statistical analysis was performed with
   Spearman correlation and a two-sided t test.

Discussion

   The present study found that FBXW7 may have predictive significance in
   patients with CRC. Non-targeted metabolomics and in vitro tests
   revealed the anti-tumor activity of FBXW7 and that FBXW7
   loss-of-function may lead to a malignant phenotype of CRC. To the best
   of our knowledge, the research findings show, for the first time, that
   FBXW7 may not only directly ubiquitinate the mTOR protein, leading to
   mTOR protein degradation, but also decrease arginine production via
   metabolic reprogramming, suppressing mTOR pathway activation and
   consequently the occurrence and progression of CRC.

   The FBXW7 protein is a member of the F-box protein family and the
   target protein recognition component of the SCF E3 ubiquitin ligase
   complex [[94]25]. FBXW7 is related to the survival rate of patients
   with cancer. Through a meta-analysis, Shang discovered that FBXW7
   mutations or low expression might predict the OS rate of patients with
   CRC [[95]27]. The expression of FBXW7 was shown to be lower in
   metastatic melanoma than in original melanoma and related to a poor
   5-year survival rate [[96]28]. Meta-analysis and bioinformatics
   unveiled that FBXW7 expression was associated with gastric cancer
   prognosis [[97]29]. The expression of FBXW7 in CRC and matched normal
   tissues was initially examined using a tissue microarray and discovered
   that the FBXW7 protein was expressed in both the nucleus and cytoplasm.
   A low ratio of nuclear/cytoplasmic expression of FBXW7 was found to be
   related to poor outcome of individuals with CRC when combined with
   clinicopathological variables. Further univariate and multivariate COX
   regression model analyses suggested that low FBXW7 expression may be
   employed as an independent predictor of prognosis in CRC patients. The
   expression of FBXW7 mRNA and the clinicopathological factors were also
   examined using TCGA data, and the findings coincided with the results
   of the tissue microarray, confirming the potential prognostic value of
   low FBXW7 expression in patients with CRC.

   Given that it is one of the major recognition factors in the
   ubiquitin-proteasome degradation pathway, many FBXW7 substrates are
   oncogenes. For example, cyclin E, c-Myc, c-Jun, and Notch are destroyed
   by the FBXW7-mediated ubiquitin-proteasome system
   [[98]25,[99]30,[100]31]. FBXW7 mutations and deletions promote the
   accumulation of these oncogenes, which have been detected in ovarian,
   breast, and CRC tissues [[101]32,[102]33]. FBXW7 mutations or deletions
   are detected in approximately 9% of CRC cases [[103]34]. Therefore, the
   role of FBXW7 in CRC cells was investigated. Over-expression studies
   were performed on two CRC cell lines with low endogenous FBXW7
   expression. The results showed that FBXW7 over-expression suppressed
   the proliferation, invasion, and migration of CRC cells, demonstrating
   that FBXW7 functions as a tumor suppressor gene in CRC.

   The mechanism by which FBXW7 functions as a tumor suppressor differs
   among cancer types. For example, in gastric cancer, FBXW7 causes
   apoptosis, arrests growth, and inhibits epithelial-mesenchymal
   transition by enhancing RhoA ubiquitination and proteasome degradation
   [[104]35]. In colon cancer, FBXW7 inhibits enolase 1 activity via the
   ubiquitin/proteasome pathway in a glycogen synthase kinase 3-dependent
   manner [[105]36]. FBXW7 suppresses tumor development and progression in
   ovarian cancer by degrading YTHDF2 and enhancing the stability of the
   pro-apoptotic gene Bcl-2-modifying factor [[106]37]. By targeting MTDH
   degradation, FBXW7 can also suppress breast cancer growth and induce
   apoptosis [[107]38]. By inhibiting FBXW7, miR-27a-3p enhances
   esophageal squamous cell carcinoma cell growth [[108]39]. However, as a
   key E3 ubiquitinated protein ligase, the significance of FBXW7 in tumor
   metabolic reprogramming has been overlooked. Ubiquitination, the most
   prevalent post-translational protein modification, regulates cancer
   metabolic signaling pathways [[109]19,[110]21]. For instance, the
   deletion of von Hippel-Lindau Cullin RING E3 ligase can stabilize the
   expression of various glycolytic enzymes promoted by hypoxia-inducible
   factor 1-alpha in cancer cells to realize aerobic glycolysis. The E3
   ligases HUWE1 and tumor necrosis factor receptor-associated factor 6
   can ubiquitinate metabolic enzymes including hexokinase 2 and affect
   tumor growth by regulating glycolysis [[111]40]. Therefore,
   understanding the role of FBXW7 in metabolic control is critical for
   analyzing CRC pathophysiology. Non-targeted metabolomics in CRC cells
   before and after FBXW7 over-expression could be detected and a series
   of differential metabolites were screened. After pathway enrichment
   analysis, the differential metabolites were found to show the most
   significant aggregation in the mTOR pathway. mTOR is a serine/threonine
   kinase that plays a role in tumor growth, angiogenesis, insulin
   resistance, fat creation, and other cellular processes, as well as in
   metabolism [[112]41]. By boosting protein, lipid, and nucleotide
   synthesis and blocking catabolism, mTORC1 promotes glucose metabolism
   and accelerates cell growth and division [[113]42]. The stability of
   the mToR also affects cytokine production in T cells and helps to
   regulate immunosuppression [[114]43]. Therefore, mTOR has been
   identified as a novel target for tumor therapy. Rapamycin and its
   equivalents have been used to treat solid tumors, organ
   transplantation, rheumatoid arthritis, and other disorders [[115]44].
   The activation of the mTOR signaling pathway is controlled by four
   signals: growth factors, energy status, oxygen content, and amino acids
   [[116]45]. Growth factors stimulate mTORC1 by activating the
   traditional insulin and Ras signaling pathways [[117]41]. The energy
   status of the cell is then transmitted to mTORC1 via AMP-activated
   protein kinase [[118]46]. Hypoxia exerts numerous effects on mTORC1
   activity, including lowered ATP levels. Arginine is an important amino
   acid for mTORC1 activation and can well regulate mTORC1 signaling
   [[119]47]. The metabolic studies unveiled that the amount of arginine
   in FBXW7-overexpressing CRC cells was dramatically reduced, which may
   explain why the mToR signaling pathway was inhibited.

   Furthermore, metabolites associated with the arginine production
   pathway were discovered to be down-regulated. There are two major
   arginine biosynthesis pathways [[120]48]: glutamic acid synthesis and
   arginine synthesis. Glutamine synthetase interacts with glutamic acid
   and ATP to generate glutamine and ADP, whereas arginine synthetase
   combines with glutamine and arginine to produce arginine and urea. In
   the alternative pathway, arginine transferase catalyzes the reaction
   between arginine and ATP to form arginine phosphate and AMP, which are
   then catalyzed by arginine synthase to produce arginine and ADP.
   L-citrulline is an important precursor for arginine production. Under
   the action of ACY, ornithine is transformed to citrulline. This is then
   transformed into argininosuccinic acid, which is subsequently converted
   to arginine and fumaric acid under the action of ASL. Fumaric acid is
   involved in the citrate cycle. Under the action of IDH2, α-ketoglutaric
   acid, a major intermediate metabolite of the tricarboxylic acid cycle,
   can be transformed into isocitrate/citrate and succinic acid.
   Citrulline, succinate (succinic acid), glutamate (glutamic acid),
   citrate (citric acid), and isocitrate levels were considerably lower in
   FBXW7-overexpressing CRC cells. Simultaneously, the expression of
   metabolic enzyme-encoding genes in metabolic pathways related to
   arginine synthesis was detected and the expression levels of nine
   metabolic enzyme-encoding genes (ACY1, ASL, IDH2, MDH2, SDHA, GOT1,
   GOT2, and GPT2) were down-regulated to varying degrees.

   In addition, the aforementioned indicators in 20 pairs of human CRC
   tissues were detected to confirm the correlation between FBXW7 and
   arginine metabolism-related enzymes and found a negative correlation,
   as also confirmed by TCGA CRC data.

   In summary, this research explored the relationship between the FBXW7
   expression and clinicopathological variables in patients with CRC and
   explained its potential application value for prognosis. In the
   meantime, in vitro investigations revealed that FBXW7 may serve as a
   tumor suppressor gene in CRC, inhibiting proliferation and metastasis
   of CRC cells via two different mechanisms: reduction in arginine
   production through metabolic reprogramming and direct down-regulation
   of mToR signaling. The research findings provide a credible theoretical
   foundation for the prognostic prediction and therapeutic planning of
   CRC.

Supporting information

   S1 Fig. The different expression level in different CRC cell lines
   detected by western blot.

   (TIF)
   [121]pone.0317294.s001.tif^ (153.4KB, tif)
   S2 Fig. LC-MS was used to screen differential metabolites.

   (A)Differential metabolites in positive and negative ion modes. (B)
   Principal Component Analysis of Metabolites.

   (TIF)
   [122]pone.0317294.s002.tif^ (528.8KB, tif)
   S1 Raw images. Raw data and uncropped blots are depicted in the
   supplementary file.

   (PDF)
   [123]pone.0317294.s003.pdf^ (371.7KB, pdf)
   S1 Table. Primers used in this study.

   (XLSX)
   [124]pone.0317294.s004.xlsx^ (11.2KB, xlsx)
   S1 Dataset

   (XLSX)
   [125]pone.0317294.s005.xlsx^ (2.3MB, xlsx)

Acknowledgments