CCAT1/FABP5 promotes tumour progression through mediating fatty acid metabolism and stabilizing PI3K/AKT/mTOR signalling in lung adenocarcinoma

Abstract Long non‐coding RNA (lncRNA) colon cancer associated transcript 1 (CCAT1) has been identified as an oncogene in many cancers, but its role in lung adenocarcinoma (LUAD) remains to be further investigated. We identified the upregulation of CCAT1 in LUAD tissues and LUAD cells. Through RNA pull‐down and mass spectrometry analysis, we obtained the interacting proteins with CCAT1 and discovered their functional relation with ‘signal transduction’, ‘energy pathways’ and ‘metabolism’ and revealed the potential of CCAT1 on fatty acid (FA) metabolism. For mechanism exploration, we uncovered the mediation of CCAT1 on the translocation of fatty acid binding protein 5 (FABP5) into nucleus by confirming their interaction and localization. Also, CCAT1 was discovered to promote the formation of the transcription complex by RXR and PPARγ so as to activate the transcription of CD36, PDK1 and VEGFA. Moreover, we found that CCAT1 regulated the activity of AKT by promoting the ubiquitination of FKBP51 through binding with USP49. Subsequently, cell function assays revealed the enhancement of CCAT1 on LUAD cell proliferation and angiogenesis in vitro and in vivo. Collectively, CCAT1 regulated cell proliferation and angiogenesis through regulating FA metabolism in LUAD, providing a novel target for LUAD treatment.

with protein-coding transcripts, such as RNA polymerase II mediated transcription, canonical splice site motifs based splicing and frequent 3'-end polyadenylation. 8 Functional characterization has shown that lncRNAs exhibit various biological functions such as transcriptional regulation, titration of miRNAs or proteins, and bridging of proteins or chromatin regions. 9 To date, although only a fraction of functional lncRNAs have been well characterized, their participation at every level of the multi-level regulated gene expression pathway has been revealed. 10 For instance, lncRNAs have been implicated to modulate gene expression through controlling protein synthesis, RNA maturation and transport, and transcriptionally silencing gene via regulating the chromatin structure. 11 Through different regulatory mechanisms, lncRNAs are involved in the regulation on numerous biological processes, including cell cycle, proliferation, differentiation and apoptosis. 12 Accordingly, profiling of lncRNAs has led us to the discovery of their abnormal expression pattern in various types of cancer, 13 indicating the potential value of lncRNAs as cancer biomarkers or therapeutic targets. Despite these advances, most ln-cRNAs remain partially uncharacterized, especially in lung adenocarcinoma. Colon cancer-associated transcript-1 (CCAT1), firstly identified by Nissan et al., was highly expressed in colorectal cancer. 15 Recently, the oncogenic property and regulatory mechanism of CCAT1 has been reported in various types of cancer, including squamous cancer 16 and oesophageal squamous cell carcinoma. 17 However, the biological functions and regulatory mechanism of CCAT1 in the control of LUAD tumorigenesis and the association with FA metabolism have not been well elucidated.
In our present study, we found CCAT1 was significantly upregulated in LUAD cell lines and identified its correlation with FA metabolism. Also, CCAT1 was revealed to drive the transcription complex formed by RXR and PPARγ therefore activated the transcription of CD36, PDK1 and VEGFA. Moreover, we uncovered the regulation of CCAT1 on the activity of AKT by promoting the ubiquitination of FKBP51 through the interaction with USP49. Gain-and loss-offunction assays revealed the oncogenic effect of CCAT1 on LUAD cell proliferation and angiogenesis. Finally, in vivo assays confirmed the results of in vitro assays. In conclusion, current study revealed the role of CCAT1 on the FA metabolism in lung adenocarcinoma, which provided a novel therapeutic target for LUAD patients.

| Tissue collection
Ten lung cancer tissues and paired non-cancerous tissues were dissected from patients in Nanjing First Hospital. The tissues were stored under −80℃ after dissection for subsequent use. All patients who underwent the surgical removal of tumours had not received any other radio-or chemo-therapy and had signed the written informed consent. Assays on tissues were approved by the ethic committee of Nanjing First Hospital.

| RNA isolation and quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR)
A total amount of RNA was isolated from cells applying TRIzol reagent following the standard protocols. The reverse transcription of total RNA was performed by Revert Aid First Strand cDNA Synthesis Kit followed by quantification using spectrophotometry.
The Maxima SYBR Green/ROX qPCR Master Mix was employed for real-time polymerase chain reaction assay with GAPDH or U1 used as the internal control.

| Cell transfection
Specific si-RNAs targeting CCAT1 was applied to silence CCAT1 whereas pcDNA3.1 with CCAT1 sequence was used to overexpress CCAT1. siCtrl and pcDNA3.1 empty vectors served as negative controls. All of the plasmid vectors above were acquired from Shanghai Integrated Biotech Solutions Co., Ltd. (Shanghai, China). Transfection of these vectors into cells as demand was carried out with the use of Lipofectamine 2000 or siRNA mate (Invitrogen).

| Pull-down, mass spectrometry and RNA immunoprecipitation assays
For the RNA pull-down assay, Magnetic RNA-Protein Pull-Down kit (Pierce, MA, USA) was applied referring to the manufacturer's directions. Protein bands were then silver-stained. Bands of which repeatedly enriched on the biotin-labelled probe with CCAT1 rather than antisense CCAT1 were analysed by mass spectrometry (MS) and validated using western blotting. For the RNA immunoprecipitation (RIP), Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, MA, USA) was employed following the manufacturer's guide.

| Chromatin immunoprecipitation (ChIP)
For the ChIP assay, EZ-Magna ChIP TMA Chromatin Immunoprecipitation Kit (Millipore, Billerica, MA, USA) was used according to the manufacturer's protocols. The formaldehyde crosslinked chromatin from cells was sonicated for the generation of DNA fragments between 200 to 300 bp. Antibodies against PPAR and RXR (Abcam) were used to precipitate the chromatin DNA. The quantification of the enrichment of PDK1, VEGFA and CD36 promoter were performed by qRT-PCR. Anti-IgG served as control.

| CoIP
For the CoIP assay, total cells were lysed 30 min with cell lysis buffer under 4℃. Then, the lysate was incubated with the antibodies against FKBP51 and UBI overnight. After adding protein G sepharose agitate, the sample 4℃ for 2 h. Following washing, the immunoprecipitants were separated using SDS-PAGE and analysed via immunoblotting together with indicated antibodies.

| Luciferase reporter assay
The luciferase reporter vector with wild type or mutant PDK1, VEGFA and CD36 promoter were transfected with pcDNA3.1/ Relative luciferase activities were normalized to Renilla luciferase activity.

| Immunofluorescence (IF)
4% of formaldehyde was used to fix the cells for 15 min. Then, pepsin treated cells were dehydrated the using ethanol and permeabi-

| In situ hybridization (ISH)
The ISH probes were generated by Exiqon (Shanghai, China). ISH Kit (Boster Bio-Engineering Company, Wuhan, China) was used to carry out ISH. The stained tissues were observed and then scored respectively by two pathologists who were blinded to the clinical parameters.

| Protein Extraction, Western blotting and antibodies
For western blot assay, cell lysates were added with RIPA buffer with protease and phosphatase inhibitor (Thermo Fisher Scientific, Houston, TX). Following the centrifuging at 12,000 g under 4℃ for 12 min, the supernatants were examined by

| Subcellular fractionation
PARIS Kit (Invitrogen) was used to separate the nuclear and cytosolic fractions under the manufacturer's guides. RNAs in cytoplasm and nuclei were isolated and then extracted from cells, and the levels of CCAT1 and FABP5 in the cytoplasm and nuclei were detected using qRT-PCR or WB. TERC, Histone3 and GAPDH were examined as fractionation indicators.

| EdU Incorporation Assay
Total cells were plated onto 96-well cell culture plates with a con-

| Quantification of neutral lipids
The lipophilic fluorescence dye BODIPY 493/503 (Invitrogen) was employed for monitoring the neutral lipid accumulation in A549 cells.

| Tube formation
For the tube formation assay, each well of the 96-well plate was added with 50 μl of Matrigel (BD Biosciences, San Jose, CA, USA).
Following the polymerizing for 30 min at 37℃, the harvested HUVECs from co-culture system were seeded into the Matrigel with a concentration of 3 × 10 4 cells/well. After 6 h incubation in a humid incubator with 5% CO 2 at 37℃, the tubules were observed using an IX71 inverted microscope.

| Animal models
Animal assays had obtained the approval of the Ethics Committee of Nanjing First Hospital. BALB/C nude mice aged 6-8 weeks were injected with luciferase A549 cells with lv-CCAT1 or lv-vector transfection for the generation of the xenografts. After 4 weeks, the tumours were dissected from mice and subjected to the observation of metastasis and HE staining. The tumour volume was examined referring to the equation as follows: volume = length × width 2 × 0.5.

| Immunohistochemistry (IHC)
For Immunohistochemistry assay, after the deparaffinizing the slides, the antigen retrieval was conducted in a steam cooker with 1 mM EDTA for 1.5 min. Antibody against FABP5 (Cell Signaling Technology, Danvers, MA, USA) was added and subjected to overnight incubation at 4℃. And the secondary antibody was added at room temperature for 15 min. 3-amino-9-ethylcarbazole or diaminobenzidine served as chromogens. Counterstain the slides using haematoxylin before mounting.

| Statistical analysis
The experiments were repeated three times, and the resulting data are presented after statistical processing. Data from multiple independent experiments are expressed as mean ± standard deviation.
One-way analysis of variance followed by Dunnett's or Tukey's multiple comparison test was performed to determine the significance of differences among groups. Statistical analysis was carried out on SPSS 13.0 software (SPSS; North Chicago, IL, USA). Differences were considered statistically significant at p < 0.05 in all experiments. *p < 0.05, **p < 0.01, ***p < 0.001.

| Screened CCAT1-associated RNA-binding proteins (RBPs) involved in lung adenocarcinoma cells
To probe the biological role of CCAT1 in LUAD, we first analysed the association between CCAT1 expression and the survival rate of LUAD patients from TCGA-PANCANCER data. As shown in Figure 1A, high level of CCAT1 was negatively correlated with the overall survival (OS) (Logrank p = 0.0027, n (high) = 232, n (low) = 272) and disease free survival (DFS) (Logrank p = 0.0061, n (high) = 232, n (low) = 272). And we found that high level of CCAT1 was associated with advanced stage (F value = 4.64, Pr(>F) = 0.00329, Figure 1B). ISH results presented a higher level of CCAT1 in cancer tissues than in para-carcinoma tissues ( Figure 1C). Then, we measured the expression level of CCAT1 in several LUAD cell lines. As shown in Figure 1D, the level of CCAT1 was significantly increased in LUAD cell lines. To determine the possible function and mechanism of CCAT1 in LUAD cells, we performed RNA pull-down assay in A549 cell line and then analysed the retrieved proteins by Mass Spectrometry. As elucidated in Figure 1E and Table S1, a total of 66 proteins were identified to be binding partners with CCAT1. Then, we analysed the coding genes of these proteins by using the GO database (http://www.geneo ntolo gy.org). Enrichment of these genes represents a measure of functions with significance, including 'signal transduction', 'Energy pathways' and 'Metabolism' ( Figure 1F).
Based on the analysis of GO, we selected fatty acid binding protein 5 (FABP5), which is critical factor for fatty acid (FA) metabolism and the mass spectrum peak of FABP5 was presented in Figure 1G.
Moreover, protein-protein interaction analysis revealed the interaction of FABP5 with FA metabolism-related proteins such as peroxisome proliferator activated receptor (PPAR) and retinoid X receptor (RXR) ( Figure 1H and Figure S1A). All these results indicated a possible regulation of CCAT1 on FA metabolism.

| CCAT1 mediated the translocation of FABP5 into nucleus
To investigate the regulatory relationship between CCAT1 and FABP5, we analysed the interaction between CCAT1 and FABP5. The combination between CCAT1 and FABP5 was confirmed using RIP assays both in A549 and SPC-A1 cell lines (Figure 2A). Since lncRNAs exhibited their functions dependent on their location, we then evaluated the subcellular localization of CCAT1 and FABP5. As shown in  Figure 2C). Since the function of FABP5 is to translocate FA into nucleus, we transfected SPC-A1 cells with CCAT1 expression vector and A549 cells with CCAT1 siRNA ( Figure S2A) and measured the level of FABP5 both in cytoplasm and nucleus using immunofluorescence and western blot assays. As illustrated in Figure 2D

| CCAT1 functioned synergistically with FABP5 to translocate FA into nucleus to induce PPAR-RXR transcriptional complex and activate CD36, VEGF and PDK1
It is well known that FA is essential for the activation of PPAR&RXR transcriptional complex and that PPAR&RXR mediated lipid metabolism. 18 Therefore, we analysed the potential binding sites on pro-  Figure 3A). Furtherly, luciferase reporter assay proved the activation of PPAR and RXR on the transcription of PDK1, VEGFA and CD36 ( Figure 3B). Subsequently, we measured the expression of CD36, VEGFA and PDK1, in response to CCAT1 dysregulation. As revealed in Figure 3C, the protein levels of CD36, VEGFA and PDK1 were obviously increased in

CCAT1-overexpressed cells and decreased in CCAT1-silenced cells.
Moreover, we assessed the levels of CD36 and PDK1 in the LUAD tissues with high or low CCAT1 expression level. Consistently, LUAD tissues with high CCAT1 level presented a higher positive staining rates of CD36 and PDK1 than LUAD tissues with low CCAT1 level ( Figure 3D). All these findings revealed that CCAT1 functioned

| CCAT1 participated in reprogramming fatty acid metabolism
Fatty acid (FA), as a crucial energy provider for cells as well as the vital source for the generation of triglyceride, is closely related to the en- overexpression resulted in reverse impacts on the levels of these key FA metabolic enzymes in A549 and SPC-A1 cells ( Figure 4B). To provide further evidence, we measured the expression of FASN, ACC1 and ACOX1 in LUAD tissues with different CCAT1 levels using IHC staining, and obtained consistent results ( Figure 4C). Thus, we concluded that CCAT1 promotes the reprogramming of FA metabolism in LUAD.

| CCAT1 activated and stabilized mTOR pathway through competitively binding to USP49
Interestingly, the phosphoinositide 3-kinase (PI3K)/AKT/mTOR signal pathway, one of the most commonly deregulated pathways in F I G U R E 7 Science schematic of current work: CCAT1 participates in FA metabolism through regulating the translocation of FABP5 into nucleus to stabilize PI3K/AKT/mTOR signalling, therefore promoting the progression of lung adenocarcinoma human cancers, 21 has been verified to be modulated by PDK1 and VEGFA. Due to the findings that CCAT1 could up-regulate PDK1 and VEGFA in an PPAR&RXR-dependent manner, it is reasonable to believe that CCAT1 could modulate PI3K/AKT/mTOR pathway. To make confirmation, we found that the levels of p-AKT (T308), p-AKT (Ser473), and p-mTOR were upregulated by CCAT1 overexpression and downregulated by CCAT1 silencing, with total-AKT and total-mTOR unchanged ( Figure 5A). Specifically, PDK1 directly phosphorylated AKT on T308, but AKT phosphorylation on S473 required mTOC complex 2 (mTORC2), indicating that besides PDK1, there might be another pathway participated in the modulation of CCAT1 on PI3K/AKT/mTOR pathway. Hence, we hypothesized that the interaction between CCAT1 and USP49 might modulate the stabilization of FKBP51. To prove our hypothesis, we first performed RIP assay and confirmed the combination between CCAT1 and USP49 in A549 cells ( Figure 5B). As shown in Figure 5C, the qPCR products were in concordance with the data. Moreover, we found through Western blot that USP49 was mainly abundant in cytoplasm ( Figure 5D and Figure S2B), which was consistent with the location of Ser473 phosphorylation of AKT. To explore the influence of CCAT1 on the ubiquitination of FKBP51, we transfected pcDNA3.1/CCAT1 or pcDNA3.1 vector into A549 and siCtrl or si-CCAT1 into SPAC1 cells. We carried out CoIP assay with antibodies against FKBP51 or UBI and tested the UBI or FKBP51 enrichment in the precipitates. It was found that CCAT1 upregulation contributed to the ubiquitination of FKBP51 in A549 cell lines, while CCAT1 down-regulation resulted in the deubiquitination of FKBP51 in SPC-A1 cell lines ( Figure 5E). Then, we analysed the combination between USP49 and FKBP51 in response to CCAT1 dysregulation by using co-IP. As exhibited in Figure 5F, upregulation of CCAT1 blocked the combination between FKBP51 and USP49, whereas silenced CCAT1 promoted the combination between FKBP51 and USP49. Furthermore, we determined the effect of CCAT1/USP49 axis on the stabilization of FKBP51. As illustrated in Figure 5G, CCAT1 overexpression or USP49 knock-down in A549 cells increased the protein level of FKBP51 but exhibited no effect on FKBP51 mRNA level ( Figure S2C), and such effect was vanished by the addition of proteasome inhibitor MG132, suggesting that CCAT1 regulated FKBP51 levels in a proteasome-dependent manner through competing binding with USP49.
Additionally, we identified the mTOR-binding protein RAPTOR through mass spectrometry ( Figure 5H) and protein-protein interaction analysis (Figure S2D), and the mass spectrum peak of RAPTOR was presented in Figure 5I. And we confirmed through RIP-qPCR the enrichment of CCAT1 in the Immunoprecipitate of RAPTOR antibody ( Figure 5J), validating the binding between CCAT1 and RAPTOR. Collectively, these results combined with the uncovered regulation of CCAT1 on the activation of PDK1 translation and phosphorylation of AKT suggested the close association between CCAT1 and PI3K signalling. Therefore, we demonstrated that CCAT1 could stabilize PI3K pathway not only through stimulating PDK1 translation and AKT phosphorylation, but also through binding with RAPTOR. However, the underlying mechanism remains elusive.

| CCAT1 enhanced malignant phenotypes of LUADs in vitro and vivo
Since PI3K/AKT/mTOR signalling pathway plays a central role in regulating cell proliferation and angiogenesis, we evaluated the effect of CCAT1 on LUAD cell proliferation and angiogenesis. As shown in

| DISCUSS ION
CCAT1 was initially reported in colorectal cancer as a diagnostic biomarker. 24,25 Later, with accumulating deep researches, the role of CCAT1 in cancers has been increasingly identified, such as in oesophageal cancer and gallbladder carcinoma. 17,26 The upregulation of CCAT1 in lung adenocarcinoma has been identified, but its regulatory mechanism remains poorly understood. Fatty acid is known to be an energy provider for bodies, playing a central role in tumorigenesis and tumour progression. 27 Its biological activities require the aid of fatty epidermal acid binding proteins, which exert important regulatory roles in various metabolism-related diseases. 28 However, its association with cancers, especially with lung adenocarcinoma, remains in the dark. Present study discovered that CCAT1 was involved in the FA metabolism through binding with FABP5 so as to regulate downstream gene expression, which therefore promoted LUAD via PI3K/AKT/mTOR signalling.
The reprogramming of cellular energy metabolism, as emerging hallmarks of cancer, supports the infinite multiplication and metastatic progression of cancer cells. 29 Multiple molecular regulatory mechanisms, intrinsic and extrinsic, converge to alter core cellular metabolism. Metabolic changes are a common feature of cancerous tissues. The most commonly observed metabolic phenotype in tumour cells is the Warburg effect, which is characterized with an increased uptake of glucose and the switch from ATP generation through oxidative phosphorylation to ATP generation through glycolysis. 30 Emerging documents indicate that the Warburg effect pull-down with CCAT1 biotin-labelled probes followed by mass spectrometry analysis, we identified the interaction between CCAT1 and FABP5, which was later confirmed by RIP-qPCR assay. Additionally, though documentary research, we found that FA is not only an important source for body energy, but also a stimulator of downstream gene translation depending on PPARγ. Through IF and subcellular fractionation, we validated that CCAT1 could mediate the translocation of FABP5 into nucleus.
It has been known that CD36 is an important receptor of FA on cytomembrane which mediates FA into cells. PDK1 is the pivotal protein in PI3K signalling activating the phosphorylation of AKT (T308), which is acknowledged to be closely associated with tumour progression. And VEGFA is an angiogenesis-related protein which relates to the activation of PI3K signalling. Present study verified through bioinformatics prediction and ChIP the binding of PPARγ and RXR to the promoters of CD36, PDK1 and VEGFA. And luciferase reporter assay presented the promotion of PPARγ/RXR on the luciferase activity of their promoters, indicating that PPARγ/RXR promoted the expression of these genes through activating their transcription. Also, in vitro assay proved that CCAT1 overexpression contributed to the upregulation of the three genes.
The phosphoinositide 3-kinase (PI3K)/AKT/mTOR signal pathway is identified as one of the most commonly deregulated pathways in human cancers. 21 Genetic aberrations that modulate this pathway, like inactivation of PTEN or activated mutations of PIK3CA, have been documented in virtually all epithelial tumours. 35 The 3-phosphoinositide-dependent protein kinase-1 (PDK1), whose activation is known to be a consequence of the accumulation of phosphatidylinositol-3,4,5-trisphosphate (PIP3), is considered as a critical component of the PI3K pathway. PDK1 is a master regulator of AGC kinase members playing multiple roles in various physiologic processes, including metabolism, growth, proliferation and survival. 36 It has been known that PDK1 is a key protein in PI3K signalling promoting the phosphorylation of T308 on AKT, which is closely associated with the development of tumours and plays an important role in multiple signalling for the tumorigenesis. Ser473. 23 In current study, we identified USP49 in the retrieved proteins by pull-down, and confirmed that it is a regulator of the ubiquitination of FKBP51. FKBP51 could inhibit PI3K pathway through phosphorylating AKT S473 via suppressing mTOR2.
In our study, CCAT1 was found to block the deubiquitination of FKBP51 by recruiting USP49, which maintained the activity of AKT. And another protein identified by mass spectrometry RAPTOR help us validated that CCAT1 positively regulated PI3K signalling. RIP assay proved the interaction between CCAT1 and RAPTOR. However, through what mechanism CCAT1 regulated RAPTOR and mTOR1 remains to be uncovered. Furthermore, for the influence of CCAT1 on fat acid metabolism, we verified that CCAT1 silencing suppressed the expression of fat acid metabolism-related proteins such as FASN, ACC1 and ACOX1. 37 Previously, we also found that CCAT1 played an essential role in regulating the drug resistance of LUAD, so we speculated that it might function through PI3K signalling, and we will deepen this study in the future.
In conclusion, the promotion of CCAT1 on fat acid metabolism not only provides necessary energies for the biological processes of cancer cells, but also contributes significantly to angiogenesis and activation of tumour-related signalling pathways (Figure 7). These findings provide valuable diagnostic and prognostic targets for LUAD patients, and unfold a new outlook for the research on the mechanism underlying LUAD.

ACK N OWLED G EM ENTS
The authors acknowledge all members involved in this study. The work was supported by National Natural Science Foundation of China

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data obtained from this work have been presented within the manuscript and its related files.