N6‐methyladenosine demethylase FTO suppresses clear cell renal cell carcinoma through a novel FTO‐PGC‐1α signalling axis

Abstract The abundant and reversible N6‐methyladenosine (m6A) RNA modification and its modulators have important roles in regulating various gene expression and biological processes. Here, we demonstrate that fat mass and obesity associated (FTO), as an m6A demethylase, plays a critical anti‐tumorigenic role in clear cell renal cell carcinoma (ccRCC). FTO is suppressed in ccRCC tissue. The low expression of FTO in human ccRCC correlates with increased tumour severity and poor patient survival. The Von Hippel‐Lindau‐deficient cells expressing FTO restores mitochondrial activity, induces oxidative stress and ROS production and shows impaired tumour growth, through increasing expression of PGC‐1α by reducing m6A levels in its mRNA transcripts. Our work demonstrates the functional importance of the m6A methylation and its modulator, and uncovers a critical FTO‐PGC‐1α axis for developing effective therapeutic strategies in the treatment of ccRCC.

Von Hippel-Lindau (VHL) mutations or deletions are the most frequent genetic alterations in ccRCC. [6][7][8] The Vhl gene, encoding an E3 ubiquitin ligase, is essential for oxygen-dependent degradation of HIFα family. 9 However, in hypoxia, HIFα degradation is inhibited, leading to HIFα stabilization, increased nuclear localization of HIFα and transcription of various target genes, including the vascular endothelial growth factor. Consequently, VHL loss of function in ccRCC leads to constitutive activation of HIF-α, which promotes tumorigenesis through transcriptional activation of genes mediating angiogenesis, 10 anti-apoptosis 11 and metabolism. 12 Constitutive activation of HIF transcription factors have been placed a high value on metabolic reprogramming in ccRCC. HIFα drive a gene expression program that increases glycolytic activity while inhibiting mitochondrial function. 13,14 In ccRCC, low mitochondrial content is associated with tumour aggressiveness, suggested that inhibition of mitochondrial function may play a key role in ccRCC progression. 4 However, the mechanism underlying reduced mitochondrial content in ccRCC, and its subsequent events, remains less understood.
The PPARg coactivators (PGC) are a family of transcriptional coactivators (consisting of PGC-1α, PGC-1β and PRC) that mediate mitochondrial biogenesis and oxidative phosphorylation. But PGC-1α, as a central regulator of mitochondrial function, cannot be compensated for by the other family members. 15,16 Accumulating evidence indicates PGC-1α play a dualistic role in cancer, with reports of tumour suppression and pro-tumorigenic effects of PGC-1α expression in variant cancer types. [17][18][19] A convincingly better understanding of the role of PGC-1α in variant tumour types will be highly significant in exploring whether this target will be amenable as anticancer agents in ccRCCs.
N6-Methyladenosine (m6A) is the most abundant internal post-transcriptional modification in mRNA occurring particularly at the beginning of the 3′-UTR near the stop codon, usually embedded within the consensus motif RR(m6)ACH (R=G or A,H=U, A or C). [20][21][22][23] This reversible modification, installed by a methyltransferase complex consisting of the proteins methyltransferase-like 3 (METTL3), METTL14 and Wilms tumour 1 associated protein (WTAP) identified as m6A "writers", is "erased" by fat mass and obesity-associated protein (FTO) and AlkB homolog 5 (ALKBH5). 24,25 Recent studies have shown that m6A modification in mRNAs plays critical roles in mRNA splicing and translation efficiency. [26][27][28][29] However, m6A modification has been most strongly related to increased mRNA instability. 30 Other groups reported that suppressor of cytokine signalling 2(SOCS2) mRNA degradation mediated by METTL3 promotes liver cancer progression, 31 that ALKBH5 maintains cancer stem-like cells phenotype by sustaining NANOG and FOXM1 mRNA stability 32,33 and that FTO-regulated ADAM19 promotes the tumorigenesis of brain tumour, 34 suggesting the biological significance of the mRNA m6A methylation and its modulators in cancer development. However, the impact of FTO, especially as an RNA demethylase, in mitochondrial biogenesis, oxidative stress and ccRCC progression remain elusive. Importantly, recent reports suggest that FTO regulate mitochondria content through mediating mitochondrial fusion, fission and biogenesis-associated genes expression as a N6methyladenosine RNA demethylase. 35,36 To the point, we aim to explore the biological function of FTO in the post-transcriptional modification of mitochondrial biogenesis and its subsequent influence in ccRCC and also explore the underlying molecular mechanism through identifying its key mRNA targets.

| Human samples, tissue microarray and cell lines
All ccRCC samples from Figure 1A-C and Figure 6A were collected from patients who had ccRCC resection performed at Peking University Shenzhen Hospital. The samples were used for subsequent RNA extraction or immunohistochemistry (IHC). All human materials were obtained with informed consent. The ccRCC tissue microarray (#HKid-CRC060CS-01) was purchased from Shanghai Outdo Biotech CO. Ltd  Table S1.
Lentivirus for FTO, FTO-mut as well as control were packaged with psPAX2, pMD2G (Addgene, Watertown, MA, USA) and pCDH-puro into HEK-293T cells. To establish stable cell lines, the concentrated lentivirus were directly added into cancer cells and incubated at 37°C for 48 hours before they were washed out with phosphate-buffered saline (PBS). Finally, cells were selected with 2.5 mg/mL puromycin for 4 days.

| RT-qPCR and gene-specific m6A qPCR
Total RNA was isolated from cultured cells using TRIzol (Invitrogen, Waltham, MA, USA) according to the manufacturer's instructions. The cDNA synthesis was performed using the Transcript First-Strand cDNA Synthesis SuperMix Kit (TransGen Biotech, Beijing, China). Quantitative real-time PCR (qRT-PCR) was performed with SYBR Premix Ex Taq II (Tli RNaseH Plus) (Takara, Dalian, China, #RR820). Relative genes expression was tested by 2 ΔΔCt normalized to GAPDH; gene -specific m6A qPCR were conducted as described previously. 32,37 Relative m6Agenes expression was tested by 2ΔΔCt normalized to hypoxanthine guanine phosphoribosyl transferase (HPRT) according to the reason that HPRT mRNA did not have m6A peaks from the m6A seq data. 30 Realtime PCR was performed with a Roche 480 thermal cycler. The primer sequences used are provided in Table S1.

| Cell proliferation, apoptosis and colony formation assays
Cells were collected and seeded with required concentration, and then apoptosis and proliferation were assessed using TransDetect Annexin V-EGFP/PI Cell Apoptosis Detection Kit (Transgene) and Cell Counting Kit-8 (CCK-8; Transgene) following the manufacturer's instructions, respectively. For colony formation assays, 1000 cells/well were plated onto six-well plates, which were incubated at 37°C and 5% CO 2 until colonies were formed. After 10-15 days, colonies were fixed using 0.05% crystal violet in 4% paraformaldehyde and counted.

| Measurement of PGC-1α mRNA stability
FTO-overexpressing cell lines, mutFTO-overexpressing cell lines and control cell lines were cultured in six-well plates. Then actinomycin D (Calbiochem, Burlington, MA, USA) was added to 8 μg/mL at 0, 3 and 6 hours before cell collection, followed by RNA extraction and real-time PCR as described earlier.

| Luciferase reporter and mutagenesis assays
The DNA fragments of PGC-1α-3′UTR containing the wild-type m6A motifs as well as mutant motifs (m6A was replaced by C) were directly synthesized from GeneCreate (Wuhan, China

| ATP measurements assay
Cells ATP measurements were obtained using Enhanced ATP Assay Kit (Beyotime, Shanghai, China) following the manufacturer's instructions. Membranes were washed 5 minutes with TBST for three times, incubated for 1 hour with required secondary antibodies conjugated to horseradish peroxidase and developed by chemiluminescent substrates.

| Statistical analysis
Data in graphs are presented as mean ± SD or mean ± SEM. Differences between two groups or multiple groups were analysed by Student's t test and ANOVA, respectively. All statistical analyses were performed and P values were obtained using the GraphPad Prism software 6.0 or SPSS 20 (SPSS Inc., Chicago, IL, USA). P values <0.05 were considered significant.

| FTO is down-regulated in ccRCC and its expression is progressively lost during cancer progression
To explore the role of FTO in ccRCC progression, we first investigated the expression levels of FTO in a RCC sample cohort consisting of 35 pairs of primary ccRCC and adjacent normal tissues by qRT-PCR, as shown in the Figure 1A, compared with the matched adjacent normal tissues, FTO was strongly down-regulated in ccRCC tissues. Furthermore, an evident decreasing trend was observed across the early stage (I-II), which also extended to late stage ccRCC (III-IV) ( Figure 1B). Consistently, FTO protein was reduced in a group of four pairs of ccRCC tissues compared with adjacent normal tissues as examined by Western blot ( Figure 1C). To confirm the reduced FTO protein expression in a larger sample set, and correlate this to clinical phenotype, we performed immunohistochemical staining (IHC) on the FTO tissue array comprised of 25 patients. IHC showed that FTO was steadily expressed in normal kidney tissues but was declined in cancer counterpart and lost in the later stage ( Figure 1D 1-2 ). To further assess the impact of FTO expression in clinical cases of ccRCC, we next analysed RNA sequencing data from over 500 patients with ccRCC from the Cancer Genome Atlas (TCGA). These data showed that low expression of FTO was markedly correlated with worse overall survival and disease-free survival than patients whose tumours expressed relatively high levels of FTO ( Figure 1E). Collectively, these results indicate that the FTO expression is frequently down-regulated in ccRCC and associated with poor prognosis, suggesting that FTO may function as a tumour suppressor in ccRCC development.
Moreover, analyses of previously published gene expression datasets and TCGA database showed that FTO was significantly down-regulated in various types of human cancer, such as breast, endometrial, uterine cervix cancer and bladder cancer (P value = 0.05) (Fig. S1A), and FTO low expression correlated with poor prognosis in human cancers, including endometrial cancer, lung cancer, rectum adenocarcinoma and pancreatic cancer (Fig. S1B), which further suggested that FTO may play an antioncogenic role in progression and development of various cancer types.

| Ectopic expression of FTO inhibits cell growth, and induces apoptosis in ccRCC
To   Figure 4A,B).
To test whether increased expression of genes correlated with mitochondria biogenesis and oxidative phosphorylation in ccRCC associated with increased mitochondria number, mitochondrial energy production and oxidative stress, we first measured mitochondria DNA content, ATP production and ROS ( Figure 4C-E). FTO-upregulating cells showed an increased mitochondrial DNA content ( Figure 4C), an elevated cell ATP levels ( Figure 4D) and an elevate ROS byproducts ( Figure 4E). Above data suggested that FTOinduced mitochondrial biogenesis and oxidative phosphorylation were closely relevant to oxidative stress in ccRCC cells.  Figure 6B). More specially, Figure 6C showed that low PGC-1α expression is associated with more severe tumours. Next, we validated that PGC-1α restored expression induced oxidative stress ( Figure 6D) and inhibited the ccRCC cell lines growth in vitro ( Figure 6E). These results suggested that low PGC-1α expression may be associated with worse patient survival.

| FTO demethylates PGC-1α mRNA and its stability is required for the regulatory role of FTO in ccRCC
We referred to GEPIA to test whether there is an association between PGC-1α expression and overall survival or disease-free survival. PGC-1α down-regulation, as expected, was related to worse disease progression and reduced overall survival in 257 and 258 patients with ccRCC, respectively ( Figure 6F).

| DISCUSSION
The data presented here show that FTO, a member of the AlkB subfamily of FeII/α-ketoglutarate-dependent dioxygenases and the first identified m6A demethylase, act as a tumour suppressor in ccRCC.
In  F I G U R E 7 A model illustrating FTOmediated PGC-1α expression and mitochondrial activity in ccRCC cells. FTO mediates demethylation of adenosine residues in the 3′-UTR of PGC-1α mRNA, leading to increased PGC-1α mRNA stability and protein expression, and increased mitochondrial biogenesis, oxidative stress and tumour suppression