Abstract

Background

   N6‐methyladenosine (m^6A) RNA modification has been demonstrated to be
   a significant regulatory process in the progression of various tumors,
   including breast cancer. Fat mass and obesity‐associated (FTO) enzyme,
   initially known as the obesity‐related protein, is the first identified
   m^6A demethylase. However, the relationship between FTO and breast
   cancer remains controversial. In this study, we aimed to elucidate the
   role and clinical significance of FTO in breast cancer and to explore
   the underlying mechanism.

Methods

   We first investigated the expression of FTO in breast cancer cell lines
   and tissues by quantitative reverse transcription‐PCR (qRT‐PCR),
   Western blotting, and immunohistochemistry. Wound healing assay and
   Transwell assay were performed to determine the migration and invasion
   abilities of SKBR3 and MDA‐MB453 cells with either knockdown or
   overexpression of FTO. RNA sequencing (RNA‐seq) was conducted to
   decipher the downstream targets of FTO. qRT‐PCR, luciferase reporter
   assay, and Western blotting were employed to confirm the existence of
   the FTO/miR‐181b‐3p/ARL5B axis. The biological function of ADP
   ribosylation factor like GTPase 5B (ARL5B) in breast cancer cells was
   evaluated by wound healing assay and Transwell invasion assay.

Results

   High FTO expression was observed in human epidermal growth factor
   receptor 2 (HER2)‐positive breast cancer, predicting advanced
   progression (tumor size [P < 0.001], nuclear grade [P = 0.001],
   peritumoral lymphovascular invasion [P < 0.001), lymph node metastasis
   [P = 0.002], and TNM stage [P = 0.001]) and poor prognosis. Moreover,
   FTO promoted cell invasion and migration in vitro. Mechanistically,
   RNA‐seq and further confirmation studies suggested that FTO
   up‐regulated ARL5B by inhibiting miR‐181b‐3p. We further verified that
   ARL5B also displayed carcinogenic activity in breast cancer cells.

Conclusion

   Our work demonstrated the carcinogenic activity of FTO in promoting the
   invasion and migration of breast cancer cells via the
   FTO/miR‐181b‐3p/ARL5B signaling pathway.

   Keywords: ARL5B, breast cancer, disease‐free survival, Fat mass and
   obesity‐associated (FTO) enzyme, human epidermal growth factor receptor
   2, m6A modification, metastasis, miR‐181b‐3p, overall survival
     __________________________________________________________________

Abbreviations

   3′ UTR
          3′ untranslated region

   ARL
          ADP ribosylation factor‐like.

   ARL5B
          ADP ribosylation factor like GTPase 5B

   ASB2
          ankyrin repeat and SOCS box containing 2

   BNIP3
          BCL2 interacting protein 3

   CCK‐8
          Cell Counting Kit‐8

   CLTA
          clathrin light chain A

   DFS
          disease‐free survival

   DGCR8
          DiGeorge syndrome critical region 8

   DMEM
          Dulbecco's modified Eagle medium

   FBS
          fetal bovine serum

   FFPE
          formalin‐fixed

   FTO
          fat mass and obesity associated

   GAPDH
          glyceraldehyde‐3‐phosphate dehydrogenase

   HE
          hematoxylin and eosin

   HER2
          human epidermal growth factor receptor 2

   IHC
          immunohistochemistry

   IRX3
          iroquois homeobox 3

   KEGG
          Kyoto Encyclopedia of Genes and Genomes

   m^6A
          N6‐methyladenosine

   MA2
          the ethyl ester form of meclofenamic acid

   MeRIP‐seq
          methylated RNA immunoprecipitation sequencing

   METTL14
          methyltransferase like 14

   METTL3
          methyltransferase like 3

   miRNA
          microRNA

   NIH
          National Institutes of Health and the Laboratory

   one‐way ANOVA
          one‐way analysis of variance

   OS
          overall survival paraffin‐embedded

   PCMTD1
          protein‐L‐isoaspartate (D‐aspartate) O‐methyltransferase domain
          containing 1

   PVDF
          polyvinylidene difluoride

   qRT‐PCR
          quantitative reverse transcription‐PCR

   RAP2C
          Ras‐related protein Rap‐2c

   RARA
          retinoic acid receptor alpha

   RFS
          recurrence‐free survival

   RIMKLB
          ribosomal modification protein rimK like family member B

   RNA‐seq
          RNA‐sequencing

   ROC
          receiver operating characteristic

   SDS‐PAGE
          sodium dodecyl sulfate‐polyacrylamide gel electrophoresis

   siRNA
          small interfering RNA

   SOWAHB
          sosondowah ankyrin repeat domain family member B

   STR
          short tandem repeat

   TCGA
          The Cancer Genome Atlas Program

1. BACKGROUND

   Breast cancer is the second leading cause of cancer‐related deaths
   among women worldwide [[42]1]. Novel interventions, including endocrine
   therapy, anti‐human epidermal growth factor receptor 2 (HER2) targeted
   therapy, and immunotherapy, have strikingly improved the survival and
   clinical outcomes for patients with breast cancer. However,
   approximately 10%‐41% of breast cancer patients experience distant
   recurrence after 5 years of endocrine therapy depending on the node
   status and grade of the primary tumor [[43]2]. Metastatic breast cancer
   accounts for the majority of cancer‐related deaths in women worldwide
   [[44]3], and the 5‐year survival rate drops to 25% once distant
   metastasis occurs [[45]4]. Numerous studies have indicated that the
   posttranscriptional regulation process is extensively involved in tumor
   biology, evolution, and pathology [[46]5, [47]6], and a better
   understanding of the specific oncogenes may reveal the molecular
   mechanisms involved in the metastasis and development of breast cancer.

   N6‐methyladenosine (m^6A) modification, with methylated adenosine at
   the N6 position, is the most abundant internal modification of
   eukaryotic mRNA and non‐coding RNA [[48]7]. Recent studies have
   illustrated the critical roles of m^6A modification in tissue
   development, stem cell self‐renewal and differentiation, biological
   rhythm, and DNA damage repair [[49]8, [50]9]. Fat mass and
   obesity‐associated (FTO) protein, which was originally defined as an
   energy metabolism‐related protein associated with increased body mass
   and obesity in children and adults [[51]10], is the first identified
   m^6A demethylase it catalyzes m^6A demethylation in an Fe(II)‐ and
   α‐ketoglutarate‐dependent manner [[52]11]. Recently, FTO was deemed an
   m^6A “eraser” and helped elucidate the involvement of
   posttranscriptional regulation by m^6A modification in complicated
   carcinogenic networks [[53]12], including acute myeloid leukemia
   [[54]13], glioblastoma [[55]14], and breast cancer [[56]15]. Although
   previous studies have shown that FTO gene polymorphisms were related to
   the incidence of breast cancer [[57]16] and that some variations
   exerted their effects on breast cancer risk by regulating the iroquois
   homeobox 3 (IRX3) [[58]17, [59]18], the impact of FTO, especially as an
   RNA demethylase, on breast cancer progression and metastasis remains
   poorly understood.

   MicroRNAs (miRNAs) are a family of short non‐coding single‐stranded
   RNAs containing 19‐22 nucleotides that promote the degradation or
   translation inhibition of mRNA by targeting specific mRNAs and forming
   RNA‐induced silencing complexes [[60]19]. Mounting studies have focused
   on m^6A involvement in miRNA alterations in cell malignant
   transformation and tumor progression [[61]20]. Methyltransferase like
   14 (METTL14) facilitates DiGeorge syndrome critical region 8 (DGCR8)
   recognition by adding m^6A modification to pri‐miR‐126, subsequently
   leading to the improvement of hepatic carcinoma [[62]21]. Moreover,
   methyltransferase like 3 (METTL3) was confirmed to accelerate
   miR‐221/222 maturation in a similar way, resulting in the proliferation
   of bladder cancer cells [[63]22]. A methylated RNA immunoprecipitation
   sequencing (MeRIP‐seq) study showed that the steady‐state levels of
   several miRNAs were affected by the knockdown of FTO and that a
   significant fraction of miRNAs contained m^6A sites [[64]23]. However,
   whether the miRNA alterations induced by FTO are related to breast
   cancer metastasis has not yet been investigated.

   Herein, we examined the expression and effect of FTO in breast cancer
   and further explored the underlying mechanism to offer novel insight
   into breast cancer pathogenesis.

2. MATERIALS AND METHODS

2.1. Patients and tumor tissues

   Formalin‐fixed, paraffin‐embedded (FFPE) specimens of tumor and paired
   non‐tumor tissues from 408 patients with HER2‐positive breast cancer
   diagnosed between January 1, 2005 and December 31, 2013, along with
   patients’ follow‐up data, were collected from Sun Yat‐sen University
   Cancer Center (Guangzhou, Guangdong, China). The inclusion criteria
   were as follows: (1) patients had pathologically diagnosed
   HER2‐positive breast cancer (2) they underwent primary tumor resection
   after diagnosis and (3) their complete follow‐up information was
   available. All follow‐up information was collected and managed
   according to the rules of Sun Yat‐sen University Cancer Center. The
   follow‐up deadline was September 20, 2019. This study was approved by
   the Ethics Committee of Sun Yat‐sen University Cancer Center. Written
   informed consent was signed by the involved patients to allow the use
   of their data in clinical researches.

2.2. Immunohistochemistry (IHC)

   All procedures were carried out as previously described [[65]24]. The
   FFPE tissues mentioned above were sectioned into 5 µm‐thick slides. For
   hematoxylin and eosin (HE) staining, slides were dewaxed and then
   stained with hematoxylin and eosin according to routine procedures. For
   IHC staining, the slides were deparaffinized, antigen retrieval was
   performed, and then the slides were blocked and processed
   consecutively. The slides were incubated with the primary antibody
   against FTO (1:500, #31687, Cell Signaling Technology, Inc., Danvers,
   MA, USA) or HER2 (1:100, #2242, Cell Signaling Technology). The
   staining was scored by two independent and experienced pathologists and
   calculated as the product of the staining intensity (0, negative; 1,
   weak; 2, moderate; and 3, strong staining) and percentage of positive
   cells (0, 0; 1, 1%‐20%; 2, 21%‐50%; 3, 51%‐70%; 4, 71%‐100%). The
   optimal cut‐off value of FTO (8.0) generated from the receiver
   operating characteristic (ROC) curve divided FTO expression into high
   or low.

2.3. Cell lines and culture

   The human breast cancer cell lines MCF7, T47D, BT474, SKBR3, MDA‐MB453,
   MDA‐MB231, and BT549 and the immortalized mammary epithelial cell line
   MCF10A were purchased from the Institute of Biochemistry and Cell
   Biology, Shanghai Institute of Biological Sciences (Shanghai, China).
   The authenticity of all cell lines was confirmed by short tandem repeat
   (STR) DNA profiling analysis. Cells were cultured in Dulbecco's
   modified eagle medium (DMEM) (Sigma‐Aldrich, Merck Millipore,
   Darmstadt, Germany) supplemented with 10% fetal bovine serum (FBS)
   (Gibco, Thermo Fisher Scientific, Inc., Waltham, MA, USA) in a
   humidified atmosphere at 37°C with 5% CO[2]. Cancer cells subcultured
   for less than 5‐6 times were used in all experiments.

2.4. RNA and plasmid transfection

   Small interfering RNAs (siRNAs) designed for FTO (siFTO) or ADP
   ribosylation factor like GTPase 5B (ARL5B) (siARL5B) and negative
   control siRNA (siCtrl) were obtained from RiboBio Co., Ltd. (Guangzhou,
   Guangdong, China). Plasmid clones expressing FTO (oeFTO), ARL5B
   (oeARL5B), or empty vector (oeVec) were purchased from Sino Biological
   Inc. (Beijing, China). All miR‐181‐3p mimics, inhibitors, and
   corresponding negative controls (miR‐NC, miR‐NC2) used for transfection
   were purchased from GenePharma Co., Ltd. (Shanghai, China).
   Transfection was conducted using Lipofectamine 3000 (Invitrogen, Thermo
   Fisher Scientific, Inc., Waltham, MA, USA) according to the
   manufacturer's instructions.

2.5. RNA m^6A quantification

   Total RNA was extracted using an RNeasy FFPE kit (Qiagen, Hilden,
   Germany) for tissues or TRIzol (Thermo Fisher Scientific, Inc.) for
   cells following the manufacturer's protocol. The relative m^6A content
   was measured by using the EpiQuik m^6A RNA Methylation Quantification
   kit (colorimetric) (EpiGentek Group Inc., Farmingdale, NY, USA).
   Briefly, 200 ng of qualified RNA was added to the assay wells provided
   by the kit. Capture antibody, detection antibody, and enhancer solution
   were sequentially added according to the manufacturer's instructions.
   Absorbance at 450 nm (A [450]) was measured using a microplate reader
   (SpectraMax M5 Molecular Devices, LLC, Sunnyvale, CA, USA) to calculate
   the m^6A levels.

2.6. Treatment with FTO inhibitor

   FTO selective inhibitor, the ethyl ester form of meclofenamic acid
   (MA2) [[66]25], was a gift from Professor Cai‐Guang Yang (Chinese
   Academy of Sciences Key Laboratory of Receptor Research, Shanghai
   Institute of Materia Medical, Shanghai, China). A stock solution of MA2
   (60 mmol/L) was dissolved in DMSO and stored at ‐20°C. After incubation
   with gradient concentrations of MA2 (40, 60, 80, and 100 µmol/L) for 48
   h, SKBR3 and MDA‐MB453 cells were collected for further studies. We
   found that MA2 at 80 µmol/L had the optimal inhibitory effect in the
   wound healing experiment. Therefore, 80 µmol/L MA2 was used for
   subsequent studies.

2.7. Quantitative reverse transcription‐PCR (qRT‐PCR)

   For mRNA or miRNA level examination, total RNA from tissues or cells
   was isolated by using the RNeasy FFPE kit or TRIzol. cDNA was
   synthesized using the PrimeScript RT reagent kit (Takara, Kusatsu,
   Japan), and the stem‐loop RT primer method was applied in miRNA reverse
   transcription. qRT‐PCR was performed using the SYBR‐Green Master mix
   (Takara) in a CFX96 touch system (Bio‐Rad Laboratories, Inc., Hercules,
   CA, USA). Glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) or U6 was
   used as an endogenous control for mRNA or miRNA detection, and the fold
   change was calculated with the relative quantification method
   (2^−ΔΔCt). Related primer sequences are listed in Supplementary Table
   S1.

2.8. Western blotting

   Briefly, breast cancer cells, after transfection or without processing,
   were harvested and lysed. Protein concentration was determined with the
   BCA protein assay kit (Thermo Fisher Scientific, Inc.). Equal amounts
   of protein were separated by 10% or 15% sodium dodecyl
   sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE) and transferred
   onto polyvinylidene difluoride (PVDF) membranes (Thermo Fisher
   Scientific, Inc.). Membranes were incubated with primary antibodies at
   4°C overnight and then secondary antibodies at room temperature for 2
   h. After washing, signals were detected using a ChemiDoc™ imaging
   system (Bio‐Rad Laboratories, Inc.). The following antibodies were
   purchased: anti‐FTO (1:5000, #31687, Cell Signaling Technology),
   anti‐ARL5B (1:1000, #CSB‐PA002091GA01HU, Cusabio Technology LLC, Wuhan,
   Hubei, China), anti‐GAPDH (1:5000, #BS60630, Bioworld Technology, Co,
   Ltd., Nanjing, Jiangsu, China), and anti‐β‐actin (1:5000, #BS‐0061R,
   Santa Cruz Biotechnology Inc., Dallas, TX, USA) and secondary
   antibodies anti‐rabbit IgG (#7074) and anti‐mouse IgG (#7076) (1:5000,
   Cell Signaling Technology).

2.9. Transwell assay

   For the Transwell invasion assay, briefly, 5 × 10^4‐1 × 10^5 cells in
   200 µL serum‐free DMEM were reseeded into the top of the insert of a
   Boyden chamber (Corning Inc., Corning, NY, USA) with 300 µg/mL Matrigel
   (BD Bioscience, San Jose, CA, USA), while 600 µL medium with 10% FBS
   was loaded into the well below. After 20‐24 h incubation, invasive
   cells that passed through the filter were fixed with 0.1%
   paraformaldehyde (Solarbio Science & Technology Co., Ltd, Beijing,
   China) and stained with 0.1% crystal violet solution (Solarbio Science
   & Technology Co., Ltd). For the Transwell migration assay, all
   procedures were similar but without the incubation of Matrigel. The
   cells that passed through the filter were imaged at 100 × magnification
   in six random fields, and measured using the ImageJ software (National
   Institutes of Health and the Laboratory, Bethesda, MD, USA).

2.10. Wound healing assay

   As previously described [[67]26], cultured cells were seeded into
   24‐well plates at a density of 2 × 10^5‐4 × 10^5 cells per well. A 10
   µL sterile pipette tip was used to scratch the cell layer and form a
   wound. The closure of the gap was imaged at designated time intervals
   under 100 × field, and the wound closure speed was evaluated using the
   ImageJ software.

2.11. Cell viability assay

   A total of 3000‐6000 transfected cells were seeded into 96‐well plates.
   At indicated time points, 10 µL of Cell Counting Kit‐8 (CCK‐8) reagent
   (Dojindo, Kumamoto, Japan) was added following the manufacturer's
   instructions. A [450] was measured on a microplate reader (SpectraMax
   M5 Molecular Devices, LLC).

2.12. RNA sequencing (RNA‐seq)

   RNA‐seq was performed by RiboBio Co., Ltd. Total RNA from SKBR3 cells
   transfected with siFTO or siCtrl was extracted with TRIzol. RNA
   sequencing libraries were constructed using the TruSeq Small RNA
   Library Prep kit (Illumina, San Diego, CA, USA) according to the
   manufacturer's instructions. RNAs were single‐end sequenced on an
   Illumina HiSeq™ 2500 Sequencer. All reads were mapped to human genome
   19 using BWA tools and annotated by miRBase version 21
   ([68]https://www.mirbase.org) using miRDeep2
   ([69]https://www.mdc-berlin.de/8551903/en/). RNA‐seq data were analyzed
   using the “edgeR” package in R software (The Free Software Foundation,
   Boston, MA, USA). Differential expression profiling was filtered with a
   threshold of |log2 (fold change)| ≥ 1 and P value < 0.05.

2.13. Bioinformatics analysis

   Recurrence‐free survival (RFS) curves were generated via the
   Kaplan‐Meier plotter online database ([70]http://kmplot.com/analysis/).
   By employing miRNA databases (TargetScan,
   [71]http://www.targetscan.org/vert_72/ miRTarBase,
   [72]http://mirtarbase.mbc.nctu.edu.tw miRDB, [73]http://www.mirdb.org/
   and miRWalk, [74]http://mirwalk.umm.uni heidelberg.de/), we obtained a
   set of target molecules of miR‐181b‐3p. The Kyoto Encyclopedia of Genes
   and Genomes (KEGG) biological pathway enrichment analysis was performed
   to investigate the involved pathways of miR‐181b‐3p target molecules by
   applying KOBAS 3.0 ([75]http://kobas.cbi.pku.edu.cn/). Afterwards, the
   Venn diagram online tool
   ([76]http://bioinformatics.psb.ugent.be/webtools/Venn/) was utilized to
   narrow down the scope into six target molecules. The Oncomine database
   ([77]http://www.oncomine.org) was used to examine the expression levels
   of miR‐181b‐3p target molecules in normal and invasive breast cancer
   tissues. The binding site of miR‐181b‐3p to the ARL5B 3′‐untranslated
   region (3′‐UTR) was predicted using the miRNA databases mentioned
   above. A scatter plot of the relative correlation between FTO and ARL5B
   expression was performed using the R software. The “RTCGA. mRNA” and
   “dplyr” packages were used to download and manage data from The Cancer
   Genome Atlas Program (TCGA, [78]https://portal.gdc.cancer.gov/). The
   “ggpubr” package was used to draw the scatter plot.

2.14. Luciferase reporter assay

   DNA fragments containing the wild‐type or mutant 3′‐UTR of ARL5B were
   separately inserted into the pmirGLO vector (Public Protein/Plasmid
   Library, Nanjing, Jiangsu, China). HEK‐293T cells were co‐transfected
   with plasmids containing ARL5B 3′‐UTR wild‐type or mutant fragments and
   miR‐181b‐3p mimics or miRNC using Lipofectamine 3000. After 48 h of
   co‐transfection, luciferase reporter assays were performed using the
   Dual Luciferase Reporter Assay System (Promega Corporation, Madison,
   WI, USA). The ratio of luminescence signals from firefly luciferase to
   those from Renilla luciferase was calculated to assess the binding of
   miR‐181b‐3p and ARL5B.

2.15. Statistical analyses

   Statistical analyses were performed using IBM SPSS Statistics version
   22.0 (SPSS Inc., Chicago, IL, USA) and GraphPad Prism 7.0 (GraphPad
   Software Inc., La Jolla, CA, USA). Data are presented as the mean ±
   standard deviation. Significant differences were calculated by
   two‐tailed Student's t‐test or one‐way analysis of variance (one‐way
   ANOVA). The Kaplan‐Meier method and log‐rank test were applied to
   determine disease‐free survival (DFS) and overall survival (OS), which
   were defined as the intervals from surgery to the occurrence of disease
   or death of any cause. The chi‐squared test was utilized to detect the
   association between FTO expression and clinicopathological factors.
   Experiments were repeated independently at least three times. P
   value < 0.05 was considered statistically significant.

3. RESULTS

3.1. The expression of FTO and its clinical significance in HER2‐positive
breast cancer

   We first examined the expression of FTO in breast cancer cells compared
   with MCF10A cells, and FTO was observed to be highly expressed
   exclusively in the HER2‐positive cell lines SKBR3 and MDA‐MB453
   (Figure [79]1A). Moreover, compared with non‐tumor tissues,
   HER2‐positive breast cancer tissues showed increased mRNA levels of FTO
   (Figure [80]1B). Due to the m^6A demethylase function of FTO, we found
   that m^6A was irregularly down‐regulated in HER2‐positive breast cancer
   tissues (Figure [81]1C). In addition, IHC staining was conducted and
   demonstrated that FTO was expressed at higher levels in HER2‐positive
   breast cancer tissues than in non‐tumor tissues (Figure [82]1D‐1E).

FIGURE 1.

   FIGURE 1
   [83]Open in a new tab

   FTO is highly expressed in HER2‐positive breast cancer tissues and
   predicts poor prognosis. A Relative expression of FTO protein in breast
   cancer cell lines compared with the immortalized mammary epithelial
   cell line MCF10A. B Relative mRNA level of FTO in HER2‐positive breast
   cancer tissues and paired non‐tumor tissues (n = 12). C Relative m^6A
   content of total RNA in paired HER2‐positive breast cancer tissues and
   adjacent non‐tumor tissues (n = 12). D Representative IHC images of
   HER2 and FTO expression in tumor and non‐tumor tissues. Scale bar:
   100 µm. E IHC score of FTO in non‐tumor and breast cancer tumor tissues
   (n = 408). F Kaplan‐Meier DFS and OS curves stratified by FTO
   expression in 408 patients with HER2‐positive breast cancer. G
   Kaplan‐Meier analysis of RFS based on FTO expression in HER2‐positive
   breast cancer from the online Kaplan‐Meier plotter database. *
   P < 0.05, ** P < 0.01.

   Abbreviations: FTO, fat mass and obesity‐associated; HER2, human
   epidermal growth factor receptor 2; m^6A, N6‐methyladenosine; HE,
   hematoxylin and eosin; IHC, immunohistochemistry; DFS, disease‐free
   survival; OS, overall survival; RFS, recurrence‐free survival

   The clinical significance of aberrantly expressed FTO in HER2‐positive
   breast cancer was analyzed, and the results showed that FTO expression
   was significantly associated with tumor size (P < 0.001), nuclear grade
   (P = 0.001), peritumoral lymphovascular invasion (P < 0.001), lymph
   node metastasis (P = 0.002), and TNM stage (P = 0.001) (Table [84]1).
   Kaplan‐Meier DFS and OS curves suggested that high expression of FTO
   predicted poor prognosis (Figure [85]1F). Similarly, RFS of
   HER2‐positive breast cancer cases from the Kaplan‐Meier plotter online
   database implied that high levels of FTO were related to increased
   tumor recurrence risk (Figure [86]1G).

TABLE 1.

   Associations between FTO expression and the clinicopathological
   characteristics of HER2‐positive breast cancer patients
   FTO expression [cases (%)]
   Characteristic Whole cohort [cases (%)] Low High P value
   Total 408 308 100
   Age at surgery (years) 0.396
   ≤50 219 (53.7) 169 (54.9) 50 (50.0)
   > 50 189 (46.3) 139 (45.1) 50 (50.0)
   Tumor size (cm) < 0.001
   ≤2.0 127 (31.1) 103 (33.5) 24 (24.0)
   2.1‐5.0 242 (59.3) 191 (62.0) 51 (51.0)
   > 5.0 39 (9.6) 14 (4.5) 25 (25.0)
   Nuclear grade[87]^* 0.001
   II 250 (61.3) 203 (65.9) 47 (47.0)
   III 158 (38.7) 105 (34.1) 53 (53.0)
   LVI < 0.001
   No 356 (87.3) 286 (92.9) 70 (70.0)
   Yes 52 (12.7) 22 (7.1) 30 (30.0)
   LNM 0.002
   No 231 (56.6) 188 (61.0) 43 (43.0)
   Yes 177 (43.4) 120 (39.0) 57 (57.0)
   TNM stage[88]^# 0.001
   I 79 (19.4) 64 (20.8) 15 (15.0)
   II 246 (60.3) 195 (63.3) 51 (51.0)
   III 77 (18.9) 46 (14.9) 31 (31.0)
   IV 6 (1.5) 3 (1.0) 3 (3.0)
   [89]Open in a new tab

   Abbreviations: FTO, fat mass and obesity‐associated; HER2, human
   epidermal growth factor receptor 2; LVI, peritumoral lymphovascular
   invasion; LNM, lymph node metastasis.

   The χ^2 test was used for the comparison between low and high FTO
   expression groups.
   ^*

   The nuclear grade was stratified according to the World Health
   Organization (WHO) classification of breast cancer: grade I, nuclei are
   similar in size (< 1.5 ×) to benign epithelial cells grade II, nuclei
   are larger (1.5‐2 × the size of benign epithelial cell nuclei) grade
   III, nuclei are markedly larger in size (> 2 × the size of benign
   epithelial cell nuclei).
   ^#

   The TNM stage was classified according to the 8th edition of the
   American Joint Committee on Cancer (AJCC 8th) staging system.

3.2. The pro‐tumor role of FTO in HER2‐positive breast cancer cells

   Knockdown and overexpression of FTO were carried out in SKBR3 and
   MDA‐MB453 cells, and corresponding changes in the expression of FTO
   were confirmed in both cell lines (Figure [90]2A‐2B). By wound healing
   assay in SKBR3 cells and Transwell migration assay in MDA‐MB453 cells,
   cell migration was decreased upon FTO knockdown but accelerated upon
   FTO overexpression (Figure [91]2C‐2D). A wound healing assay in
   FTO‐silenced MDA‐MB453 cells also showed inhibited cell migration
   (Supplementary Figure S1A). In addition, we observed inhibition of cell
   migration with MA2 treatment, and the inhibition effect was maximal at
   80 µmol/L (Figure [92]2E). Correspondingly, FTO inhibition by siRNA or
   MA2 was found to impair cell invasion (Figure [93]2F and [94]2G), while
   the results were reversed when FTO was overexpressed (Figure [95]2H).
   However, cell proliferation was not distinctly changed after FTO
   alteration (Supplementary Figure S1B). Furthermore, m^6A quantification
   assay showed that inhibition of FTO increased m^6A levels, while FTO
   overexpression decreased m^6A abundance (Figure [96]2I).

FIGURE 2.

   FIGURE 2
   [97]Open in a new tab

   FTO enhances the migration and invasion of HER2‐positive breast cancer
   cells. A‐B The transfection efficiency of FTO siRNA or overexpression
   plasmid in SKBR3 and MDA‐MB453 cells confirmed by qRT‐PCR (A) and
   Western blotting (B). C‐D The detection of cell migration by Wound
   healing assay in SKBR3 cells and by Transwell migration assay in
   MDA‐MB453 cells. The healing percent and migration cells were measured
   at 24 h after the transfection of FTO siRNA or overexpression plasmid.
   E Wound healing assay of SKBR3 and MDA‐MB453 cells pretreated with
   gradient concentrations of MA2 for 48 h. DMSO was used as solvent
   control. F‐G Transwell invasion assay of SKBR3 and MDA‐MB453 cells
   treated with FTO siRNA or MA2 (80 µmol/L). H Transwell invasion assay
   of SKBR3 and MDA‐MB453 cells transfected with FTO plasmid. Experiments
   were terminated after seeding for 24 h. I Relative m^6A content of
   total RNA after FTO knockdown, overexpression, or FTO inhibitor MA2
   treatment. Scale bar: 200 µm. * P < 0.05, ** P < 0.01, *** P < 0.001.

   Abbreviations: siCtrl, negative control of siRNA; siFTO, FTO siRNA;
   eVec, empty plasmid vector; oeFTO, FTO clone plasmid; qRT‐PCR,
   quantitative reverse transcription‐PCR; MA2, the ethyl ester form of
   meclofenamic acid other abbreviations as in Figure [98]1

3.3. Identification of FTO downstream targets by RNA‐seq

   To explore potential targets of FTO, we performed RNA‐seq on
   FTO‐knockdown SKBR3 cells. RNA‐seq analysis revealed that 19 miRNAs
   were significantly down‐regulated, while 10 miRNAs were significantly
   up‐regulated (Figure [99]3A) the top six up‐regulated or down‐regulated
   miRNAs are listed in Figure [100]3B. The KEGG pathway enrichment
   analysis of miRNA target genes showed that metabolic pathways were
   intensely clustered, followed by pathways in cancer and endocytosis
   (Figure [101]3C).

FIGURE 3.

   FIGURE 3
   [102]Open in a new tab

   Identification of miRNAs and pathways related to FTO alteration by
   RNA‐sequencing. A The discrepancy in miRNA expression in FTO‐silenced
   SKBR3 cells compared with negative control cells. B The top six
   up‐regulated or down‐regulated miRNAs. C The pathway enrichment
   analysis of miRNA target genes in FTO knockdown SKBR3 cells compared
   with negative control cells.

   Abbreviations: miRNAs, microRNAs; FTO, fat mass and obesity‐associated

   We selected miR‐181b‐3p, the miRNA with the greatest change, to test
   the relationship between FTO and miRNAs in HER2‐positive breast cancer.
   qRT‐PCR analysis confirmed that FTO inhibition using siRNA or MA2
   resulted in significantly increased levels of miR‐181b‐3p, while FTO
   overexpression reduced miR‐181b‐3p expression (Figure [103]4).
   miR‐181b‐3p depletion mimicked the FTO‐mediated promotion of breast
   cancer cell migration and invasion (Figure [104]5A‐5B), whereas
   miR‐181b‐3p overexpression yielded opposite results
   (Figure [105]5C‐5D). Dysregulation of miR‐181b‐3p also significantly
   reversed the inhibitory effect of MA2 (Figure [106]5E and [107]5F).

FIGURE 4.

   FIGURE 4
   [108]Open in a new tab

   The regulation of FTO on miR‐181b‐3p expression. The relative level of
   miR‐181b‐3p in SKBR3 and MDA‐MB453 cells treated with FTO siRNA, FTO
   overexpression plasmid, or MA2 (80 µmol/L). *P < 0.05, **P < 0.01,
   ***P < 0.001. Abbreviations are the same as in Figure [109]2

FIGURE 5.

   FIGURE 5
   [110]Open in a new tab

   miR‐181b‐3p inhibits the pro‐tumor effects of FTO in HER2‐positive
   breast cancer cells. A‐B Wound healing assay and Transwell invasion
   assay in control or FTO‐depleted SKBR3 cells with or without the
   presence of miR‐181b‐3p inhibitors. C‐D Wound healing assay and
   Transwell invasion assay in control or FTO‐overexpressing MDA‐MB453
   cells with or without the presence of miR181b‐3p mimics. E‐F Wound
   healing assay in DMSO or MA2 (80 µmol/L) treated SKBR3 and MDA‐MB453
   cells with or without the presence of miR‐181b‐3p inhibitors. *
   P < 0.05, ** P < 0.01, *** P < 0.001. Scale bar: 200 µm. Abbreviations
   are the same as in Figure [111]2

3.4. Potential of ARL5B as a novel target of miR‐181b‐3p and its regulation
by FTO

   Based on the prediction of miR‐181b‐3p targets by bioinformatics
   analysis, six candidate molecules were identified (Figure [112]6A):
   ARL5B, Ras‐related protein Rap‐2c (RAP2C), clathrin light chain A
   (CLTA), protein‐L‐isoaspartate (D‐aspartate) O‐methyltransferase domain
   containing 1 (PCMTD1), ribosomal modification protein rimK like family
   member B (RIMKLB), and sosondowah ankyrin repeat domain family member B
   (SOWAHB), of which ARL5B and RIMKLB mRNA were found to be significantly
   down‐regulated in FTO‐silenced cells according to RNA‐seq analysis
   (data not shown). Further confirmation by qRT‐PCR demonstrated that the
   expression of ARL5B and PCMTD1 was significantly inhibited with
   transfection of miR‐181b‐3p mimics (Figure [113]6B and [114]6C), while
   ARL5B, RAP2C, PCMTD1, and RIMKLB were elevated with transfection of
   miR‐181b‐3p inhibitors (Figure [115]6D and [116]6E). Similarly, the
   expression of all candidate molecules except SOWAHB was decreased upon
   FTO knockdown (Figure [117]6F and [118]6G), whereas only ARL5B, CLTA,
   and RIMKLB showed up‐regulation upon FTO overexpression (Figure [119]6H
   and [120]6I), while ARL5B, RAP2C, CLTA, and RIMKLB were substantially
   reduced upon MA2 treatment (Figure [121]6J and [122]6K). The expression
   of five candidate molecules in breast cancer tissues was also
   investigated (no SOWAHB data online), of which only ARL5B and RAP2C
   were aberrantly up‐regulated in invasive breast tumor (Supplementary
   Figure S2). Kaplan‐Meier analysis in breast cancer patients suggested
   that high levels of ARL5B, RAP2C, and CLTA and low levels of PCMTD1 and
   RIMKLB were significantly associated with short RFS (Supplementary
   Figure S3). Overall, ARL5B was identified as the downstream target of
   interest.

FIGURE 6.

   FIGURE 6
   [123]Open in a new tab

   FTO positively mediates the expression of miR‐181b‐3p downstream
   targets. A The Venn diagram refined six potential targets of
   miR‐181b‐3p: ARL5B, RAP2C, CLTA, PCMTD1, RIMKLB, and SOWAHB. B‐C
   qRT‐PCR analysis of the expression of the six candidate molecules in
   SKBR3 and MDA‐MB453 cells with the transfection of miR‐181b‐3p mimics.
   D‐E qRT‐PCR analysis of the expression of the six candidate molecules
   in SKBR3 and MDA‐MB453 cells with the transfection of miR‐181b‐3p
   inhibitors. F‐K The mRNA levels of candidate molecules in SKBR3 and
   MDA‐MB453 cells treated with FTO siRNA (F‐G), FTO plasmid (H‐I), or MA2
   (80 µmol/L, J‐K). * P < 0.05, ** P < 0.01, *** P < 0.001.

   Abbreviations: miR‐NC, negative control of miR‐181b‐3p mimics; miR‐NC2,
   negative control of miR‐181b‐3p inhibitors; ARL5B, ADP ribosylation
   factor like GTPase 5B; RAP2C, Ras‐related protein Rap‐2c; CLTA,
   clathrin light chain A; PCMTD1, protein‐L‐isoaspartate (D‐aspartate)
   O‐methyltransferase domain containing 1; RIMKLB, ribosomal modification
   protein rimK like family member B; SOWAHB, sosondowah ankyrin repeat
   domain family member B; other abbreviations as in Figure [124]2

   We next focused on the relationship between ARL5B and FTO or
   miR‐181b‐3p. The results showed that FTO reinforced ARL5B expression,
   while miR‐181b‐3p inhibited ARL5B expression (Figure [125]7A‐7C).
   Correlation analysis revealed that the expression levels of FTO and
   ARL5B were positively correlated (Figure [126]7D), and the dual
   luciferase reporter assay confirmed that miR‐181b‐3p bound directly to
   the ARL5B 3′‐UTR (Figure [127]7E‐7F). Furthermore, by transfecting
   siARL5B into FTO‐overexpressing cells or ARL5B clone plasmid into
   FTO‐silenced cells, ARL5B was verified to be able to drive the
   migration and invasion effects induced by FTO (Figure [128]8A‐8D).
   Overexpressing ARL5B after MA2 treatment also rescued the inhibitory
   effect of MA2 (Figure [129]8E and 8F). Additionally, we knocked down or
   enhanced the expression of ARL5B in SKBR3 and MDA‐MB453 cells. Based on
   satisfactory transfection efficiency (Figure [130]9A‐9D), ARL5B
   silencing markedly reduced tumor cell migration and invasion, while
   forced expression of ARL5B yielded the opposite effects
   (Figure [131]9E‐9H). In summary, these data indicate that FTO could
   accelerate the migration and invasion of HER2‐positive breast cancer
   cells via the FTO/miR‐181b‐3p/ARL5B signaling pathway (Figure [132]10).

FIGURE 7.

   FIGURE 7
   [133]Open in a new tab

   ARL5B is the target of miR‐181b‐3p and positively regulated by FTO. A‐B
   The protein level of ARL5B in SKBR3 and MDA‐MB453 cells with
   transfection of miR‐181b‐3p mimics or inhibitors. C The protein level
   of ARL5B in FTO‐silenced or ‐overexpressing SKBR3 and MDA‐MB453 cells.
   D The scatter plot between the expression of FTO and ARL5B in
   HER2‐positive breast cancer (data from TCGA, R^2 = 0.036, P = 0.010). E
   Schematic illustration of miR‐181b‐3p and ARL5B 3′ UTR‐binding sites
   accompanied with wild‐type and mutant‐type sequences of ARL5B 3′‐UTR
   reporter plasmid. The shaded area denoted the complementary bases and
   corresponding mutated nucleobases. F Luciferase reporter gene assay in
   HEK‐293T cells following co‐transfection of ARL5B wild type or mutant
   type and miR‐181b‐3p mimics or miR‐NC, respectively (*** P < 0.001).

   Abbreviations: TCGA, The Cancer Genome Atlas Program; FPKM: Fragments
   Per Kilobase Million; 3′‐UTR, 3′‐untranslated region; other
   abbreviations as in Figure [134]6

FIGURE 8.

   FIGURE 8
   [135]Open in a new tab

   ARL5B can mediate the migration and invasion of breast cancer cells
   induced by FTO. A‐B Wound healing assay and Transwell invasion assay in
   control or FTO‐overexpressing SKBR3 cells with or without the presence
   of siARL5B. C‐D Wound healing assay and Transwell invasion assay in
   control or FTO depleted SKBR3 and MDA‐MB453 cells with or without ARL5B
   forced expression. E‐F Wound healing assay in DMSO or MA2 (80 µmol/L)
   treated SKBR3 and MDA‐MB453 cells with or without the presence of ARL5B
   overexpression. * P < 0.05, ** P < 0.01, *** P < 0.001. Scale bar:
   200 µm.

   Abbreviations: ARL5B, ADP ribosylation factor like GTPase 5B; siARL5B,
   ARL5B siRNA; oeARL5B, ARL5B clone plasmid; other abbreviations as in
   Figure [136]2

FIGURE 9.

   FIGURE 9
   [137]Open in a new tab

   ARL5B plays an oncogenic role in HER2‐positive breast cancer cells. A‐D
   The transfection efficiency of ARL5B siRNA and overexpression plasmid
   in SKBR3 and MDA‐MB453 cells by qRT‐PCR and Western blotting. E‐F Wound
   healing assay of SKBR3 and MDA‐MB453 cells with ARL5B depletion or
   overexpression. Experiments were terminated after scratch for 24 h. G‐H
   Transwell invasion assay in ARL5B‐silenced or ‐overexpressing SKBR3 and
   MDA‐MB453 cells. * P < 0.05, ** P < 0.01, *** P < 0.001. Scale bar:
   200 µm. Abbreviations are the same as in Figure [138]8

FIGURE 10.

   FIGURE 10
   [139]Open in a new tab

   The schematic model of the role and underlying mechanism of FTO in
   breast cancer.

   Abbreviations: FTO, fat mass and obesity‐associated; ARL5B, ADP
   ribosylation factor like GTPase 5B

4. DISCUSSION

   The contribution of FTO to the risk of various cancers shown in
   epidemiological studies [[140]12], along with its aberrant expression
   in breast cancer, indicates the possible involvement of FTO in
   mediating the pathological development of breast cancer. In the present
   study, we discovered that FTO was highly expressed in HER2‐positive
   breast cancer cells and tissues. Statistical analysis of clinical
   characteristics and survival of HER2‐positive breast cancer patients
   suggested that a high level of FTO was significantly associated with
   poor clinical outcomes. Mechanistically, in SKBR3 and MDA‐MB453 cells,
   FTO accelerated cell migration and invasion by targeting ARL5B via
   miR‐181b‐3p down‐regulation. Furthermore, ARL5B was shown to promote
   breast tumor cell migration and invasion. Together, these findings
   reveal the prognostic value of FTO and provide novel insight into the
   role of the FTO/miR‐181b‐3p/ARL5B axis in the pathogenesis of breast
   cancer.

   Tan et al. [[141]27] discovered that FTO expression was significantly
   associated with clinicopathological characteristics of HER2‐positive
   breast cancer patients, which aligns with our current results. In
   addition to our findings in SKBR3 and MDA‐MB453 cells, Niu et al.
   [[142]15] proposed that FTO induced breast cancer cell proliferation
   and metastasis by epigenetically demethylating the 3′‐UTR of the
   pro‐apoptosis factor BCL2 interacting protein 3 (BNIP3) in MCF7 and
   MDA‐MB231 cells. This observation implied the intricate involvement of
   FTO in different breast cancer molecular subtypes, which is worthy of
   further exploration. Su et al. [[143]28] discovered that
   R‐2‐hydroxyglutarate exerted antitumor activity by targeting the
   FTO/m^6A/MYC/CEBPA signaling pathway, leading to suppressed
   proliferation or survival of leukemia and glioma cells expressing high
   levels of FTO. Li et al. [[144]13] showed that FTO reduced the
   transcription of ankyrin repeat and SOCS box containing 2 (ASB2) and
   retinoic acid receptor alpha (RARA) by serving as an m^6A demethylase,
   promoting proliferation and inhibiting apoptosis in acute myeloid
   leukemia. Similarly, treatment with the FTO inhibitor MA2 and
   subsequent m^6A detection in the present study suggested that FTO
   exerted a pro‐tumor function by removing the m^6A modification.
   Moreover, based on the study by Berulava et al. [[145]23] searching for
   the possible consensus sequence motifs for m^6A modification in m^6A
   immunoprecipitated miRNAs, we succeeded in mapping the miR‐181b‐3p
   sequence to all the most discriminating motifs, and the mapped sequence
   was enriched in the center section. The miR‐181b‐3p sequence also
   contains the “RRACH” motif which has been identified as a highly
   conserved common sequence for m^6A modification of mRNA and long
   non‐coding RNA [[146]29, [147]30]. These results may imply the direct
   interaction between FTO and miR‐181b‐3p. However, MeRIP‐seq should be
   applied to fully clarify the specific m^6A modification site of
   miR‐181b‐3p that is modified by FTO. Based on the crucial role of FTO
   in metabolism, although no difference in cell viability was detected in
   our research, we do not exclude the possibility that FTO is involved in
   breast cancer cell proliferation and apoptosis. Further studies are
   urged to fully unveil the comprehensive role of FTO in breast cancer.

   In the present study, we confirmed that FTO could reduce the expression
   of miR‐181b‐3p, which inhibited the pro‐tumor effects of FTO. The
   miR‐181 family has been clarified to be critical for cell cycle
   progression and differentiation [[148]31]. miR‐181b‐3p was found to
   promote tumor epithelial‐mesenchymal transition by snail stabilization
   in MCF7, BT549, and MDA‐MB231 cells [[149]32], whereas miR‐181b
   inhibited the migration of MCF7 and MDA‐MB231 cells [[150]33]. In the
   present study, miR‐181b‐3p inhibited the migration and invasion of
   HER2‐positive breast cancer cells by targeting ARL5B. The fact that
   miR‐181b‐3p could exhibit diverse biological roles in different
   cellular contexts indicates that miRNAs are under delicate and
   intricate control. The illumination of its regulatory mechanism will be
   helpful to clarify the clinical significance of miR‐181b‐3p in
   different circumstances [[151]34].

   To date, ARL5B has been recognized as a member of the ADP ribosylation
   factor‐like (ARL) family belonging to the RAS superfamily [[152]35]. In
   addition to being conserved, ARL5B is the only small G protein present
   on mature lysosomes [[153]36, [154]37]. Previous studies have
   demonstrated that ARL5B overexpression facilitated lysosome motility,
   resulting in lysosome dispersion and accumulation at the cell periphery
   [[155]38, [156]39]. Lysosomes move toward the cell periphery along
   microtubules in a kinesin‐1‐SKIP‐ARL5B‐dependent manner, which is
   essential for focal adhesion turnover and cathepsin exocytosis in the
   process of cell migration [[157]40, [158]41]. The guided outward
   movement of lysosomes is considered to be a critical determinant of
   lysosomal function [[159]42, [160]43]. The present research discovered
   that ARL5B enhanced breast tumor cell migration and invasion, implying
   an oncogenic role in breast cancer. Moreover, in SKBR3 and MDA‐MB453
   cells, Rafn et al. [[161]44] detected that decreased lysosomal outward
   trafficking resulted in inhibited cancer cell invasion. Additionally,
   ARL5B silencing reduced lysosome dispersion and subsequently decreased
   cell invasion in prostate cancer [[162]45]. Taken together, these
   results suggest that ARL5B may enhance the migration and invasion of
   breast cancer cells by accelerating the redistribution of lysosomes to
   the cell periphery.

   In our present study, we failed to pinpoint the exact m^6A modification
   site in miR‐181b‐3p by MeRIP‐seq. In addition, the explicit mechanism
   of how ARL5B drives cell invasion and migration has not been clearly
   explored. Moreover, an in vivo study is essential to confirm the
   cancer‐promoting role of FTO and ARL5B at the integral level. Further
   studies are needed to fully unfold the clinical value of FTO,
   miR‐181b‐3p, and ARL5B in breast cancer.

5. CONCLUSIONS

   In the current study, we provide clinical and in vitro evidence that
   FTO, which is highly expressed in HER2‐positive breast cancer, could
   promote cell migration and invasion through the FTO/miR‐181b‐3p/ARL5B
   signaling pathway. Our work highlights the functional involvement of
   FTO in tumor pathogenesis and uncovers the cancer‐promoting role of
   ARL5B, indicating the potential prognostic value for breast cancer.
   Future work is urged to shed light on the clinical significance of
   miR‐181b‐3p and ARL5B in breast cancer and to search for possible
   predictive factors.

AUTHORS' CONTRIBUTIONS

   THW and PS designed and funded the study. YYX, SY, and NZ conducted the
   experiments and analyzed the data. HTL and YC performed bioinformatic
   data analyses. LW provided experimental material support. SHZ and KWZ
   provided supervision and suggestions for the implementation of the
   study. YYX and SY drafted the manuscript. THW, PS, and KWZ guided and
   reviewed the manuscript. All authors read and approved the final
   manuscript.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

   This study was approved by the Institutional Ethical Boards of Sun
   Yat‐sen University Cancer Center. Written informed consent was obtained
   from all patients.

CONSENT FOR PUBLICATION

   Not applicable.

AVAILABILITY OF DATA AND MATERIALS

   Not applicable.Sad

COMPETING INTERESTS

   The authors declare that they have no competing interests.

Supporting information

   Supporting Information
   [163]Click here for additional data file.^ (861KB, tif)

   Supporting Information
   [164]Click here for additional data file.^ (390.7KB, tif)

   Supporting Information
   [165]Click here for additional data file.^ (321.4KB, tif)

   Supporting Information
   [166]Click here for additional data file.^ (18.5KB, docx)

ACKNOWLEDGEMENTS