Promoting roles of long non-coding RNA FAM83H-AS1 in bladder cancer growth, metastasis, and angiogenesis through the c-Myc-mediated ULK3 upregulation

ABSTRACT

Long non-coding RNA (lncRNA) FAM83H-AS1 has been recently identified with oncogenic roles in many human cancers. But its role in bladder cancer (BCa) pathogenesis and the mechanisms are largely unstudied. This study aims to evaluate the roles of FAM83H-AS1 in the malignant behaviors and the angiogenesis of BCa cells and the mechanical molecules involved. High expression of FAM83H-AS1 was found in 82 BCa tissues and in BCa cell lines compared to the normal ones. FAM83H-AS1 downregulation in T24 and BK10 cells inhibited viability, colony formation, migration, invasion, and angiogenesis of BCa cells and increased cell apoptosis. FAM83H-AS1 was found to bind to the transcription factor c-Myc to activate ULK3 expression. Overexpression of ULK3 was further introduced into T24 and BK10 cells in the presence of FAM83H-AS1 silencing, which blocked the inhibitory effects of FAM83H-AS1 downregulation on BCa cell growth. The activity of the Hedgehog signaling pathway was suppressed by FAM83H-AS1 while recovered by ULK3. Suppression of the Hedgehog pathway reduced the malignant behaviors of BCa cells promoted by ULK3. The in vitro experiment results were reproduced in vivo. This study evidenced that FAM83H-AS1 upregulates ULK3 expression through the transcription factor c-Myc and promotes the progression of BCa.

1. Introduction

Bladder cancer (BCa) is the 10th most common cancer with an estimated 549,000 new cases and 200,000 deaths in 2018 worldwide, and males have a 4-time predominance in the incidence and mortality over females [1]. According to the disease states, BCa is allocated to non-muscle invasive bladder cancer (NMIBC) and a more life-threatening type, muscle-invasive bladder cancer (MIBC) with a high risk of current or distant metastasis potential [2]. Primary tumors are usually diagnosed at advanced stages because of the lack of outward signs during the early stages [3]. Endoscopic and open surgery, along with systemic immunotherapy, chemotherapy, and radiotherapy are current treatments for BCa; however, the treating outcomes are still unfavorable [4]. The survival rate of patients saw little improvement during the last decades [5]. It is thus important to develop more biomarkers for BCa prediction and treatment, which requires a better recognition of the molecular pathogenesis of BCa.

Long non-coding RNA (LncRNA) is the largest subclass of non-coding RNAs with over 200 nucleotides in length. Aberrant expression of lncRNAs is frequently observed in many cancer types and involved in all stages of cancer development from initiation to metastasis, leaving them as potential biomarkers or targets for cancer treatment [6–8] including BCa [9]. FAM83H-AS1 has recently been identified as an oncogene in several cancer types and was correlated with poor prognosis, tumor-node-metastasis (TNM), lymph node metastasis in cancer patients [10]. In BCa, FAM83H-AS1 was correlated with advanced clinical stage and muscular is invasion and dismal prognosis of patients [11]. But the exact functions and the potential mechanical molecules are largely unknown. LncRNAs have also been observed to bind to transcription factors to affect cancer progression through the mediation of following transcripts [12,13]. In this paper, we identified that FAM83H-AS1 possibly binds to the transcription factor c-Myc. c-Myc is a key nuclear phosphoprotein of the Myc-family with a proto-oncogene function and is involved in cell proliferation, transformation, and death [14]. c-Myc was suggested as a promising target of targeted and combined therapies for cancer [15] including BCa [16]. In addition, we further predicted that c-Myc could transcriptionally activate Unc-51-Like Kinase 3 (ULK3), a positive regulator of the Hedgehog pathway whose aberrant expression was correlated with multiple developmental abnormalities and several types of cancers [17]. Therefore, we speculated that FAM83H-AS1 possibly promotes BCa progression through c-Myc-mediated ULK3 upregulation and the subsequent Hedgehog signaling activation. Both animal and cell experiments were performed to evaluate the functions of FAM83H-AS1 in BCa growth, metastasis, and angiogenesis.

2. Materials and methods

2.1. Ethics statement

This study was launched with the approval of the Ethics Committee of the First Affiliated Hospital of Bengbu Medical College and performed in compliance with the Helsinki Declaration. Signed informed consent was received from each eligible participant. Animal experiments were conducted in accordance with the Ethical Guidelines for The Study of Experimental Pain in Conscious Animals initiated by National Institutes of Health (Maryland, USA). Great efforts were made to minimize the number and suffering of animals.

2.2. Antibodies and primer sequences

Primary antibodies used in this research were these against E-cadherin (#33-4000, Thermo Fisher Scientific Inc., Waltham, MA, USA), N-cadherin (#33-3900, Thermo Fisher), Snai1 (#3879, Cell Signaling Technology, Beverly, MA, USA), Slug (#ab51772, Abcam Inc., Cambridge, MA, USA), ULK3 (#ab124947, Abcam), CD31 (ab134168, Abcam), vascular endothelial growth factor A (VEGFA, #MA5-13182, Thermo Fisher), glioma-associated oncogene 1 (GLI1, #GTX106207, Genetex, San Antonio, TX, USA), patched1 (PTCH1, #GTX83771, Genetex), human hedgehog interacting protein (Hhip, #ab230271, Abcam), c-Myc (ab32381, Abcam) glyceraldehyde-3-phosphate dehydrogenase (GAPDH, #MA5-15738, Thermo Fisher), and the secondary antibody used was horseradish peroxidase (HRP)-labeled goat anti-rabbit immunoglobulin G (IgG, #ab6721, Abcam). The primer sequences used for reverse transcription quantitative polymerase chain reaction (RT-qPCR) are presented in Table 1. The sequences of short hairpin (sh) RNAs of FAM83H-AS1 and the overexpression vector of ULK3 are listed in Table 2.

Table 1.

Primer sequences for RT-qPCR

Gene	Primer sequence (5'-3')
FAM83H-AS1	F: TCCCAATAAACAGGGCAGAC
	R: CAAGATCACCACACCCCTCT
E-cadherin	F: GCCTCCTGAAAAGAGAGTGGAAG
	R: TGGCAGTGTCTCTCCAAATCCG
N-cadherin	F: CCTCCAGAGTTTACTGCCATGAC
	R: GTAGGATCTCCGCCACTGATTC
Snai1	F: TGCCCTCAAGATGCACATCCGA
	R: GGGACAGGAGAAGGGCTTCTC
ULK3	F: CTACGCCAAGAAGGACACTCGT
	R: ATCTCCGTGAGGAGGTTCTCCA
Slug	F: ATCTGCGGCAAGGCGTTTTCCA
	R: GAGCCCTCAGATTTGACCTGTC
CD31	F: AAGTGGAGTCCAGCCGCATATC
	R: ATGGAGCAGGACAGGTTCAGTC
VEGFA	F: TTGCCTTGCTGCTCTACCTCCA
	R: GATGGCAGTAGCTGCGCTGATA
GAPDH	F: GAGAAGGCTGGGGCTCATTT
	R: AGTGATGGCATGGACTGTGG
U6	F: CTCGCTTCGGCAGCACA
	R: AACGCTTCACGAATTTGCGT

RT-qPCR, reverse transcription quantitative polymerase chain reaction; ULK3, UNC-51-like Kinase 3; VEGFA, vascular endothelial growth factor A; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; F, forward; R, reverse.

Table 2.

Sequences of shRNAs of FAM83H-AS1 and overexpression vector of ULK3

Vectors	Sequence (5'-3')
shFAM83H-AS1-#1	F: CACCGGTCTCTGATGTTGGTGTTAATTCAAGAGATTAACACCAACATCAGAGACCTTTTTTG
	R: GATCCAAAAAAGGTCTCTGATGTTGGTGTTAATCTCTTGAATTAACACCAACATCAGAGACC
shFAM83H-AS1-#2	F: GACCGATTCAGCTGGCAGCTAAGATTCAAGAGATCTTAGCTGCCAGCTGAATGCTTTTTTG
	R: GATCCAAAAAAGCATTCAGCTGGCAGCTAAGATCTCTTGAATCTTAGCTGCCAGCTGAATGC
shFAM83H-AS1-#3	F: CACCGCAGCTGTGAGTCTGAATTTCTTCAAGAGAGAAATTCAGACTCACAGATGCTTTTTTG
	R: GATCCAAAAAAGCAGCTGTGAGTCTGAATTTCTCTCTTGAAGAAATTCAGACTCACAGCTGC
shScramble	F: CACCGTTCTCCGAACGTGTCACGTCAAGAGATTACGTGACACGTTCGGAGAATTTTTTG
	R: GATCCAAAAAATTCTCCGAACGTGTCACGTAATCTCTTGACGTGACACGTTCGGAGAAC
oe-ULK3	F: TTTTGTAATACGACTCACTATAGGGCGGCCGGGAATTCGTCGACTGGATCCGGTACCGAGGAGATCTGCCGC CGCGATCGCC
	R: ACGCGTACGCGGCCGCTCGAGCAGAAACTCATCTCAGAAGAGGATCTGGCAGCAAATGATATCCTGGATTA CAAGGATGACGACGATAAGGTTTAA

shRNA, short hairpin RNA; ULK3; ULK3, UNC-51-like Kinase 3; oe, overexpression; F: forward, R: reverse.

2.3. Clinical sample collection

A total of 82 BCa patients underwent tumorectomy from January 2013 to June 2014 in The First Affiliated Hospital of Bengbu Medical College were enrolled. The tumor tissue and the adjacent normal tissue samples were collected. All included patients never received chemotherapy, chemotherapy or molecular-targeted therapy. Among the patients, 38 were older than 50 years old (48.3%) and 44 were younger (51.6%), and 47 were males (53.2%) and 35 were females (46.8%). There were no significant differences regarding the age and gender distribution of patients. All tissue samples were collected during surgery and instantly soaked in liquid nitrogen, and then preserved at −80°C. The TNM staging was performed according to the standard initiated by American Joint Committee on Cancer on 2010. The clinical baseline characteristics of patients are given in Table 3.

Table 3.

Clinical baseline characteristics of patients with BCa

Characteristics	Number
Total	82
Age (years)
≥ 50	38
< 20	44
Sex
Male	47
Female	35
Clinical Stage
I + II	43
III + IV	39
Lymph node metastasis
with	26
without	56
Tumor differentiation
poor	27
moderate	39
well	16

BCa, bladder cancer.

2.4. RT-qPCR

Total RNA from BCa tissues and cells was extracted using the TRIzol Reagent (Invitrogen, Thermo Fisher). Then, 1 μg total RNA was reversely transcribed to cDNA using a PrimeScript RT kit (Perfect Real Time, Takara, Holdings Inc., Kyoto, Japan). Next, qPCR was performed at a reaction volume of 25 μL on a CFX96 real-time qPCR Detection System (Bio-Rad, Hercules, CA, USA). GAPDH and U6 were used as the internal references. Relative gene expression was evaluated using the 2^−ΔΔCt method [18], and the average value of three independent experiments was obtained.

2.5. Immunohistochemical staining

Collected tumor samples were fixed in 4% paraformaldehyde overnight, dehydrated, embedded in paraffin, and cut into consecutive sections (6 μm thick) for staining. After that, the sections were dewaxed, rehydrated, and heated by microwave for antigen retrieval in 0.001 mol/L sodium citrate buffer. Next, the cells were treated with 3% H₂O₂ for 10 min, and then blocked with 5% bovine serum albumin (BSA) in phosphate-buffered saline (PBS) at 20°C for 30 min. The antibodies were incubated with the primary antibodies at 4°C overnight, and then with the HRP-labeled secondary antibody at 37°C for 2 h. Next, the tissues on slides were further stained with 3,3ʹ-diaminobenzidine at 37°C for 30 min, and hematoxylin was used for nuclear staining at 37°C for 30 min.

2.6. Cell culture and treatment

BCa cell lines T24, SW780, HT-1197, BK10, and HTB-9 and normal urothelial cell line SVHUC-1 were obtained from Cancer Cell Bank of Chinese Academy of Medical Sciences (Shanghai, China). T24 and BK10 cells were cultured in fetal bovine serum (FBS)-supplemented Dulbecco’s modified Eagle’s medium (DMEM) containing 100 U/L penicillin-streptomycin at 37°C and 5% CO₂. Well-growing T24 and BK10 cells were transfected with shRNAs targeting FAM83H-AS1 (shFAM83H-AS1-1, shFAM83H-AS12, shFAM83H-AS13) or ULK3 overexpressing vector using the Lipofectamine^TM 2000 (Invitrogen) as per the kit’s instructions. After transfection, the FAM83H-AS1 and ULK3 expression was determined using RT-qPCR or western blot analysis.

2.7. Cell counting kit-8 (CCK-8) method

A CCK-8 kit (Dojindo, Laboratories, Kumamoto, Japan) was used. In brief, cells were seeded in 96-well plates at 5 × 10³ cells per well. At 0 h, 24 h, 48 h, 72 h, and 96 h, respectively, each well was added with 100 μL serum-free medium and 10 μL CCK-8 solution for another 2-h incubation at 37°C. After that, the optical density (OD) value at 450 nm was determined using a microplate reader (Thermo Fisher Scientific).

2.8. Colony formation assay

Cells were seeded into 6-well plates at 500 cells per well and incubated at 37°C with 5% CO₂. Two weeks later, the cells were fixed in 4% paraformaldehyde for 30 min and and stained with 0.1% crystal violet for 15 min, and the number of formed colonies (over 50 cells) was calculated.

2.9. Flow cytometry

Transfected cells were washed in pre-cooled PBS, and then centrifuged at 2000 rpm for 10 minutes. Next, the cells were resuspended in 500 μL binding buffer and treated with 5 μL Annexin V-fluorescein isothiocyanate and 5 μL propidium iodide (PI, BD Biosciences, San Diego, CA, USA) in the dark for 20 min. Thereafter, the number of apoptotic cells was determined using a flow cytometer (BD Biosciences) and the apoptosis rate was calculated. For cell cycle assays, the harvested cells were stained by PI using a CycleTEST PLUS DNA kit (BD Biosciences). The ratio of cells in the G0/G1, S or G2/M phases was calculated and compared according to FACScan analysis.

2.10. Terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick end labeling (TUNEL)

A TUNEL kit (Roche Applied Science, Mannheim GmbH, Penzberg, Germany) was used in compliance with the instructions of the manufacturer’s instructions. Then, the samples were imaged using a laser scanning confocal microscope (Zeiss LSM 510 META, Carl Zeiss, MicroImaging, Inc., Thornwood, NY, USA).

2.11. Transwell assay

Transwell assays were performed to measure invasion and migration of cells. As for cell migration, each apical chamber (Corning, NY, USA) was loaded with 1 × 10⁵ cells, and each basolateral chamber was filled with 500 μL 10% FBS-supplemented DMEM. As for the invasion assay, the apical chambers were pre-coated with Matrigel (BD Biosciences). After a 12-h (migration) or 24-h (invasion) warm incubation at 37°C, the cells in the apical chambers were wiped out, and the migrated or invaded cells were fixed in paraformaldehyde, stained with crystal violet, and photographed under a microscope.

2.12. Tube formation assay

Human umbilical vein endothelial cells (HUVECs, China Center for Type Culture Collection) were seeded on 96-well plates at 2 × 10⁴ cells per well and pre-applied with 60 μL Matrigel. After normal incubation for 24 h, the tube formation of cells was imaged under an inverted microscope. The length of tube brunches was quantified using the Image J v6.0 software.

2.13. Western blot assay

In brief, cells were lysed in 4°C lysis buffer to extract the total protein. Then, an equal volume of protein samples was run on 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto the methanol-pretreated polyvinylidene fluoride membranes. The western blots were determined following co-incubation with primary antibodies at 4°C overnight. Next, the membranes with protein were further incubated with HRP-conjugated secondary antibody at 37°C for 1 hour. The protein bolts were visualized using the enhanced chemiluminescence method.

2.14. Sub-cellular localization of FAM83H-AS1

A fluorescence in situ hybridization (FISH) assay was first performed to evaluate the sub-cellular localization of FAM83H-AS1 in BK10 and T24 cells. Well-growing T24 and B10 cells (5 × 10⁵ cells/mL) were fixed in 4% paraformaldehyde at 20°C for 30 min. The probe of FAM83H-AS1 for RNA FISH was acquired from RiboBio Co., Ltd. (Guangzhou, Guangdong China). After incubation with proteinase K, the cells were dehydrated in ethanol and then hybridized with the fluorescence-labeled FAM83H-AS1 probes. Nuclei of cells were stained by 4ʹ, 6-diamidino-2-phenylindole (Life Technologies, Carlsbad, USA), and the fluorescence was scanned under the microscope. A PARIS^TM (Invitrogen, Thermo Fisher) was used to further determine the sub-cellular localization detection as per the kit’s instructions. Distribution of RNA in nuclear and in cytoplasm was determined by RT-qPCR.

2.15. RNA immunoprecipitation (RIP) assay

A Magna RIP kit (Millipore, Corp., Billerica, MA, USA) was used as per the instructions. In brief, cells were lysed in RIP lysis buffer, and then co-incubated with the RIP buffer containing the magnet bead-conjugated human anti-c-Myc or control IgG at 37°C for 1 h. The precipitated RNA was determined by RT-qPCR to measure the expression of FAM83H-AS1.

2.16. Chromatin immunoprecipitation (ChIP)-qPCR

Cells were washed twice in PBS and then warm incubated in 4% paraformaldehyde at room temperature for 10 minutes. After that, cells were lysed in cell lysis buffer at 4°C for 1 h, and then the cell lysates were ultrasonicated to chromatin fragments (500–800 bp in length). Next, the samples were pre-purified by protein A-agarose (Roche) at 4°C and then incubated with anti-cMyc (ab32072, Abcam) overnight. The immune precipitates were collected using 10% (v/v) protein A-agarose. A ULK3 primer was used for PCR to analyze the binding relationship between c-Myc and the promoter of ULK3.

2.17. Animal experiments

Specific-pathogen-free female BALB/c nude mice (3–5 weeks old) were purchased from the Vital River Experimental Animal Technology Co., Ltd. (Beijing, China). Cells with stable transfection (1.5 × 10⁶ cells/mL) were implanted into the mouse abdomen through subcutaneous injection. From the 10th d after implantation, the volume (V) of tumors was evaluated every 5 d by the following formula: V = (a × b²)/2, where “a” indicates the tumor length while “b” indicates the width. The mice were euthanized on the 35th d through intraperitoneal injection of overdose of pentobarbital sodium (150 mg/kg), and the tumors were collected and weighed for immunohistochemical staining. As for metastasis assay, cells with stable transfection (2 × 10⁷ cells/mL) were injected into mice through caudal veins. The mice were euthanized on the 45th d, and the liver and lung tissues were collected for hematoxylin and eosin (HE) staining to evaluate the number of metastatic nodules.

2.18. Statistical analysis

Data were presented as mean ± standard deviation (SD) from at least three independent experiments. The SPSS 25.0 software (IBM Corp. Armonk, NY, USA) was applied for data analysis. Differences were compared using the t test (two groups) and one-way or two-way analysis of variance (ANOVA, three or more groups). Correlation between FAM83H-AS1 and ULK3 expression was analyzed by Pearson’s correlation analysis. The survival curves were produced and compared using the Kaplan-Meier method. p < 0.05 was considered to show significant difference.

3. Results

3.1. FAM83H-AS1 is abundantly expressed in BCa patients and indicates poor prognosis

First, according to the RT-qPCR results, FAM83H-AS1 expression in 82 pairs of cancer tissues was much higher than that in the adjacent normal ones (Figure 1(a)). A similar trend was found by the RNA-ISH assay concerning the staining intensity of FAM83H-AS1 in the collected tissues. In the para-cancerous tissues, 38.92% was not stained, 56.09% was weakly stained, 9.76% was moderately stained, and 1.22% was strongly stained. But in the tumor tissues, 7.32% was not stained, 28.05% was weakly stained, 32.92% was moderately stained, and 31.71% was strongly stained (Figure 1(b)). The data on TCGA-BLCA Database (http://gepia.cancer-pku.cn/index.html) also found an increase FAM83H-AS1 expression in BCa tissues compared to the normal tissues (Figure 1(c)). The correlations between FAM83H-AS1 expression and the clinical presentations of BCa patients were analyzed. High expression of FAM83H-AS1 was found to be associated increased Ki67 staining intensity (Figure 1(d)). Further, according to the lymph node metastasis, the patients were allocated into two groups, and it was found that patients with lymph node metastasis showed increased expression of FAM83H-AS1 compared to those without (Figure 1(e)). The patients were allocated into stage I, stage II, stage III and stage IV in terms of the TNM staging. It was also found that FAM83H-AS1 expression was notably increased as the TNM stage increases (Figure 1(f)). In addition, patients with poor tumor differentiation rate also showed increased expression of FAM83H-AS1 compared to those with moderate or well differentiation rate (Figure 1(g)). In addition, a higher level of FAM83H-AS1 was found to lead to a worse 5-year survival rate in BCa patients (Figure 1(h)). In cell experiments, the RT-qPCR results also revealed a high expression profiling of FAM83H-AS1 in BCa cell lines T24, SW780, HT-1197, BK10, and HTB-9 compared to the SV-HUC-1 cells, especially in T24 and BK10 cells (Figure 1(i)). Therefore, these two cell lines were used for the subsequent experiments and transfected with three shRNAs of FAM83H-AS1 (shFAM83H-AS1-1, shFAM83H-AS1-2, shFAM83H-AS1-3), after which the FAM83H-AS1 expression in both cells was decreased according to RT-qPCR (Figure 1(j)). The shFAM83H-AS1-1 showed the best interfering efficacy, was used for the subsequent experiments.

Figure 1.

FAM83H-AS1 is abundantly expressed in BCa patients and indicates poor prognosis. (a), FAM83H-AS1 expression in 82 pairs of BCa tissues and the adjacent normal ones determined by RT-qPCR; (b), FAM83H-AS1 expression in BCa and the adjacent normal tissues determined by the RNA-ISH assay; (c), FAM83H-AS1 expression in BCa predicted on the TCGA-BLCA Database (http://gepia.cancer-pku.cn/index.html); (d–g), correlations between FAM83H-AS1 expression and the intensity of immunohistochemical staining of Ki67 in BCa tissues (d), the lymph node metastasis in BCa patients (e), the TNM staging of BCa patients (f) and the tumor differentiation (g); (h), correlation between FAM83H-AS1 expression and the overall 5-year survival rate of BCa patients evaluated by the Kaplan-Meier analysis; (i), FAM83H-AS1 expression in BCa cell lines (T24, SW780, HT-1197, BK10 and HTB-9) and in SV-HUC-1 cells determined by RT-qPCR; J, FAM83H-AS1 expression in T24 and BK10 cells after FAM83H-AS1 shRNA (shFAM83H-AS1-1, shFAM83H-AS1-2 and shFAM83H-AS1-3) transfection determined by RT-qPCR. Data were presented as mean ± SD from at least three independent experiments. In panels A and E, data were analyzed by paired t test; in panels B and J, data were analyzed by two-way ANOVA while data in panels D, F, G and I were compared by one-way ANOVA, and Tukey’s multiple comparison test was used for post hoc test. **p < 0.01, ***p < 0.001

3.2. Knockdown of FAM83H-AS1 inhibits BCa cell growth but promotes cell apoptosis and cell cycle arrest

Following the findings above, we further found that the viability of BK10 and T24 cells was notably decreased upon FAM83H-AS1 downregulation according to the CCK-8 assay (Figure 2(a)). The proliferation activity of cells was determined by the colony formation assay. After sh-FAM83H-AS1 transfection, the number of cell colonies formed by BK10 and T24 cells was significantly reduced, indicating that silencing of FAM83H-AS1 reduced proliferation of these two cell lines (Figure 2(b)). The cell apoptosis was then determined by TUNEL assay. In BK10 and T24 cells, after FAM83H-AS1, the number of cells labeled with green fluorescence was increased, indicating increased apoptotic bodies and TUNEL positive rate in BK10 and T24 cells after FAM83H-AS1 downregulation (Figure 2(c)). The flow cytometry was additionally performed to examine the apoptosis rate of cells. After FAM83H-AS1 downregulation, the number of apoptotic BK10 and T24 cells (cells in the fourth quadrant) was increased (Figure 2(d)). In addition, the flow cytometry also suggested that knockdown of FAM83H-AS1 led to an increase in cell cycle arrest at G0/G1 phases in BK10 and T24 cells, which triggered the stress reaction and apoptosis signaling pathway in cells and induced apoptosis and PI positive rate in cells. (Figure 2(e)). These results collectively validate an important role of FAM83H-AS1 in BCa cell growth and survival.

Figure 2.

Knockdown of FAM83H-AS1 inhibits BCa cell growth but promotes cell apoptosis and cell cycle arrest. (a), viability of BK10 and T24 cells determined by CCK-8 assay; (b), number of cell colonies evaluated by the colony formation assay; (c–d), number of apoptotic cells determined by TUNEL (c) and flow cytometry (d), respectively; (e), cell cycle distribution assessed by flow cytometry. Data were presented as mean ± SD from at least three independent experiments. In all panels, data were compared by two-way ANOVA followed by Tukey’s multiple comparison test. **p < 0.01, ***p < 0.001

3.3. Knockdown of FAM83H-AS1 inhibits BCa cell metastasis in vitro and angiogenesis of HUVECs

The metastatic potential of BCa cells was evaluated as well. Importantly, after FAM83H-AS1 downregulation, the RT-qPCR and western blot analysis identified a decline in mRNA and protein levels of mesenchymal cell markers N-cadherin, Snai1 and Slug while an increase in epithelial cell marker E-cadherin (Figure 3(a,b)), which indicated that sh-FAM83H-AS1 reduced the epithelial-mesenchymal transition (EMT) process of BK10 and T24 cells. This was further examined by the morphological changes in cells under the microscope, where we found the spindle-shaped cells (mesenchymal shape) were decreased while the square cells (epithelial shape) were increased (Figure 3(c)). In addition, the Transwell assays suggested the number of invaded/migrated cells from the apical chambers was decreased upon FAM83H-AS1 silencing. As for migration assay, cells were loaded in the apical chambers. After 12 h, the number of cells migrated into the basolateral chambers was notably reduced after FAM83H-AS1 knockdown (Figure 3(d)). Likewise, the number of BK10 and T24 cells invaded into the basolateral chambers was significantly reduced as well after a 24-h culture in the setting of pre-coating Matrigel in the apical chambers (Figure 3(e)). Artificial downregulation of FAM83H-AS1 was further introduced in HUVECs, after which the mRNA and protein levels of CD31 and VEGFA (indicators of angiogenesis) in cells, according to the RT-qPCR and western blot assays, were decreased (Figure 3(f,g)). In a more direct and noticeable manner, a tube formation assay was performed to observe the angiogenesis ability of the HUVECs. Knockdown of FAM83H-AS1 in HUVECs led to a notable decline in the number and the area of formed tubes, which validated that sh-FAM83H-AS1 suppressed the angiogenesis ability of cells (Figure 3(h)).

Figure 3.

Knockdown of FAM83H-AS1 inhibits BCa cell metastasis in vitro and angiogenesis of HUVECs. (a,b), mRNA (a) and protein (b) expression of E-cadherin, N-cadherin, Snai1 and Slug in BK10 and T24 cells determined by RT-qPCR and western blot analysis, respectively; (c), cell morphological changes observed under an optical microscope (200 × magnification); (d,e), migration (d) and invasion (e) abilities of T24 and BK10 cells measured by Transwell assays; F-G, mRNA (f) and protein (g) level of CD31 and VEGFA in cells determined by RT-qPCR and western blot analysis, respectively; (h), number of formed tubes by HUVECs determined by tube formation assay. Data were presented as mean ± SD from at least three independent experiments. In all panels, data were analyzed by two-way ANOVA and Tukey’s multiple comparison test. **p < 0.01, ***p < 0.001

3.4. FAM83H-AS1 binds to transcription factor c-Myc in nucleus to activate ULK3 expression

To further identify the possible molecular mechanism of action, we first predicted the sub-cellular localization of FAM83H-AS1 on a Bioinformatic Software LncAtlas (http://lncatlas.crg.eu/). However, though it was suggested to be localized in both cytoplasm and in nuclear in many cell types, there was limited information concerning its localization in BCa cells (Figure 4(a)). Thereafter, the cytoplastic and nuclear RNA separation and FISH assays were performed and identified a nuclear sub-localization of FAM83H-AS1 in BK10 and T24 cells (Figure 4(b,c)). Therefore, we speculated that FAM83H-AS1 possibly exerts its function through mediating transcription factors. Importantly, according to the data on LncMap (http://bio-bigdata.hrbmu.edu.cn/LncMAP/), FAM83H-AS1 was suggested to bind to c-Myc to transcriptionally activate ULK3 (Figure 4(d)). Subsequently, the binding relationship between FAM83H-AS1 and c-Myc was validated through the RIP assay, which found an abundancy in FAM83H-AS1 expression in the compounds enriched by anti-c-Myc compared to anti-IgG (Figure 4(e)). In addition, the binding relationship between c-Myc and the promoter region of ULK3 was first predicted on JASPER (http://jasper.genereg.net/) with several binding sites identified (Figure 4(f,g)). The subsequent ChIP-qPCR assay validated a binding relationship between c-Myc and the Site 4 at the promoter region of ULK3 (Figure 4(h)). Further, RT-qPCR identified increased ULK3 expression in 82 pairs of BCa tissues compared to the adjacent normal ones (Figure 4(i)). In addition, a Pearson’s correlation analysis and the data on TCGA-BLCA suggested a positive correlation between ULK3 expression and FAM83H-AS1 expression in BCa patients (Figure 4(j,k)). Likewise, a high expression profiling of ULK3, both in terms of mRNA and protein expression, was also found in BCa cell lines as compared to that in SV-HUC-1 cells (Figure 4(l,m)). Importantly, the mRNA and protein expression of ULK3 in BK10 and T24 cells was decreased following FAM83H-AS1 downregulation (Figure 4(n,o)). To further validate the relevance of FAM83H-AS1 to ULK1, we further examined expression of ULK3 in BK10 and T24 cells in the setting of FAM83H-AS1 overexpression. After upregulation of FAM83H-AS1 in cells using the overexpression vector of FAM83H-AS1, the mRNA and protein expression of ULK3 in BK10 and T24 cells was notably increased (Supplementary Figure S1A-B). In addition, to confirm the regulatory network of c-Myc on ULK3, altered expression of c-Myc was introduced in BK10 and T24 cells. Then, the RT-qPCR and western blot analysis results showed that the mRNA and protein level of ULK3 in cells was elevated as the increase of c-Myc expression, while down-regulation of c-Myc led to a notable decline in the expression of mRNA and protein in BK10 and T24 in cells (Figure 4(p,q)).

Figure 4.

FAM83H-AS1 binds to transcription factor c-Myc to activate ULK3 expression. (a), subcellular localization of FAM83H-AS1 predicted on the LncAtals Database (http://lncatlas.crg.eu/); (b,c), subcellular localization of FAM83H-AS1 in BK10 and T24 cells determined by cytoplastic and nuclear RNA separation (b) and FISH (c) assays; (d), binding relationships between c-Myc and FAM83H-AS1 or ULK3 predicted on LncMap (http://bio-bigdata.hrbmu.edu.cn/LncMAP/); (e), binding between FAM83H-AS1 and c-Myc in nuclear of BK10 and T24 cells validated through the RIP assay; (f), binding sites between c-Myc and the promoter region of ULK3 predicted on JASPAR (http://jasper.genereg.net/); (g), conservative binding sequence of c-Myc; (h), binding relationship between c-Myc and the promoter region of ULK3 validated by the ChIP-qPCR assay; (i), ULK3 expression in 82 cancer and normal tissues determined by RT-qPCR; (j), correlation between FAM83H-AS1 and ULK3 expression in 82 BCa patients determined by Pearson’s correlation analysis; (k), correlation between FAM83H-AS1 and ULK3 expression predicted on TCGA-BLCA (http://gepia.cancer-pku.cn/index.html); (l,m), mRNA (l) and protein (m) expression of ULK3 in BCa cell lines and in SV-HUC-1 cells determined by RT-qPCR and western blot analysis, respectively; (n,o), mRNA (n) and protein (o) expression of ULK3 in BK10 and T24 cells after FAM83H-AS1 downregulation measured by RT-qPCR and western blot analysis, respectively; (p,q), mRNA (p) and protein (q) expression of ULK3 in BK10 and T24 cells after FAM83H-AS1 knockdown determined by RT-qPCR and western blot analysis, respectively. Data were presented as mean ± SD from at least three independent experiments. In panel I, data were analyzed by paired t test, and each spot indicates a respondent, ***p < 0.001, while data in other panels were compared by two-way ANOVA followed by Tukey’s multiple comparison test, **p < 0.01, ***p < 0.001

3.5. Overexpression of ULK3 reverses the inhibitory effects of shFAM83M-AS1 on BCa cell growth

Next, overexpression of ULK3 was introduced into BK10 and T24 cells in the setting of FAM83M-AS1 knockdown, and the successful overexpression was validated by the western blot analysis (Figure 5(a)). Thereafter, it was found that further upregulation of ULK3 enhanced the viability and proliferation rate of cells that was initially inhibited by shFAM83M-AS1 (Figure 5(b)). Likewise, the number of cell colonies formed by BK10 and T24 cells was increased as well upon FAM83H-AS1 downregulation (Figure 5(c)). Next, the TUNEL assay suggested that after ULK3 upregulation, the number of BK10 and T24 cells labeled with green fluorescence was decreased, which showed ULK3 reduced the formation of apoptotic bodies and TUNEL positive rate in BK10 and T24 cells (Figure 5(d)). In addition, the flow cytometry also showed that the ratio of early apoptotic cells (cells in the fourth quadrant) was reduced after further ULK3 overexpression (Figure 5(e)). The aggressiveness of BK10 and T24 cells following ULK3 upregulation was determined as well. In the migration assay, the number of cells migrated from the apical chambers into the basolateral chambers within 12 h, according to the crystal violet staining and observation under the microscope, was notably increased after ULK3 overexpression (Figure 5(f)). Likewise, the number of BK10 and T24 cells invaded into the basolateral chambers was significantly increased as well after a 24-h culture in the setting of pre-coating Matrigel in the apical chambers (Figure 5(g)). Importantly, the tube formation assay also suggested that overexpression of ULK1 in HUVECs increased the number and the area of formed tubes, indicating that ULK3 suppressed the angiogenesis ability of cells (Figure 5(h)).

Figure 5.

Overexpression of ULK3 reverses the inhibitory effects of shFAM83M-AS1 on BCa cell growth. BK10 and T24 cells with stable shFAM83M-AS1 transfection were further transfected with ULK overexpressing vector. (a), protein level of ULK3 in cells was determined by western blot analysis; (b), viability of BK10 and T24 cells measured by CCK-8 assay; (c), number of cell colonies evaluated by the colony formation assay; (d,e), apoptosis of BK10 and T24 cells determined by TUNEL (d) and flow cytometry (e); (f,g), migration (f) and invasion (g) abilities of BK10 and T24 cells measured by Transwell assays; (h), number of formed tubes by HUVECs determined by tube formation assay. Data were presented as mean ± SD from at least three independent experiments. In all panels, data were analyzed by two-way ANOVA and Tukey’s multiple comparison test. **p < 0.01, ***p < 0.001

3.6. ULK3 activates the Hedgehog signaling pathway to promote the malignant behaviors of BCa cells

ULK3 has been noted as a positive regulator of the Hedgehog signaling pathway [16]. Here, we determined the Hedgehog pathway biomarkers GLI1, PTCH1, and HhIP in BK10 and T24 cells. It was found that the levels of these proteins were notably decreased following shFAM83M-AS1 administration but then recovered after the further upregulation of ULK3 (Figure 6(a)). In addition, a Hedgehog-specific antagonist, Jervine, was administrated into BK10 cells and T24 cells in the presence of ULK3 overexpression. Then, it was found that the viability and proliferation of cells enhanced by ULK3 was then blocked by Jervine (Figure 6(b)). Still, the colony formation assay suggested that the number of colonies formed by BK10 and T24 cells was reduced after administration of Jervine (Figure 6(c)). Next, the TUNEL assay suggested that the number of apoptotic bodies in BK10 and T24 cells, namely the TUNEL positive rate in cells was enhanced by when the Hedgehog signaling pathway was inactivated by Jervine (Figure 6(d)). Again, the flow cytometry also suggested that the ratio of early apoptotic BK10 and T24 cells (cells in the fourth quadrant) was notably increased after Jervine treatment (Figure 6(e)). In addition, the Transwell assays suggested that the number of BK10 and T24 cells migrated (Figure 6(f)) and invaded (Figure 6(g)) into the basolateral chambers was suppressed after Jervine administration.

Figure 6.

ULK3 activates the Hedgehog signaling pathway to promote the malignant behaviors of BCa cells. (a), protein levels of Hedgehog pathway biomarkers GLI1, PTCH1 and HhIP in BK10 and T24 cells measured by western blot analysis; (b), viability of BK10 and T24 cells determined by CCK-8 assay; (c), number of cell colonies evaluated by the colony formation assay; (d,e), apoptosis of BK10 and T24 cells determined by TUNEL (d) and flow cytometry (e); (f,g), migration (f) and invasion (g) abilities of BK10 and T24 cells measured by Transwell assays. Data were presented as mean ± SD from at least three independent experiments. In all panels, data were analyzed by two-way ANOVA and Tukey’s multiple comparison test. **p < 0.01, ***p < 0.001

3.7. Knockdown of FAM83H-AS1 inhibits growth, metastasis and angiogenesis of BCa in vivo

BALB/c nude mice were used to validate the function of FAM83H-AS1 in BCa cell growth in vivo. Each mouse was subcutaneously injected with 2 × 10⁶ T24 or BK10 cells with stable transfection of sh (control) or shFAM83H-AS1. Then, we found that shFAM83H-AS1 suppressed the growth rate of T24 or BK10 in vivo. The increase in volume of xenograft tumors in nude mice was suppressed (Figure 7(a)), and the weight of tumor on the 35th d was decreased as well after FAM83H-AS1 suppression (Figure 7(b)). Th Scramble e expression of Ki-67 in the xenograft tumors was determined by immunohistochemical staining, which suggested that the Ki-67-positive cell rate in the tumor tissues (cells with brownish nuclei) was notably reduced (Figure 7(c)). Also, the immunohistochemical staining also suggested that the CD31-positive rate in tissues (cytoplasm stained in brown while nuclei stained in blue) was notably decreased (Figure 7(d)). In addition, the immunohistochemical staining showed VEGFA-positive rate in tissues (cytoplasm stained in brown while nuclei stained in blue) was reduced after FAM83H-AS1 knockdown (Figure 7(e)). These results indicated that downregulation of FAM83H-AS1 suppressed growth of xenograft tumors and angiogenesis of BCa in vivo. Furthermore, BK10 and T24 cells with stable transfection of shFAM83H-AS1 were injected into mice through caudal veins for metastasis assay. On the 35th d the mice were euthanized, and the lung tissue sections were collected for HE staining. It was found that injection of cells transfected with shFAM83H-AS1 led to a decline in the number of metastatic nodules in mouse lung tissues compared to injection of cells transfected with shScramble (Figure 7(f)).

Figure 7.

Knockdown of FAM83H-AS1 inhibits BCa cell growth and metastasis and angiogenesis in vivo. Each BALB/c nude mouse was subcutaneously injected with 2 × 10⁶ T24 or BK10 cells with stable shScramble (control) or shFAM83H-AS1 transfection. (a), assessment of volume of the xenograft tumors in mice form the 10th d after cell implantation; (b), weight of the xenograft tumors on the 35th d; (c–e), protein expression of Ki67 (c), CD31 (d) and VEGFA (e) in xenograft tumors detected by immunohistochemical staining; (f), BALB/c nude mice were injected with BK10 or T24 cells with stable shScramble (control) or shFAM83H-AS1 transfection through caudal veins, and then the number of metastatic nodules in mouse lung tissues was determined on the 35th d. Data were presented as mean ± SD from at least three independent experiments. In all panels, data were analyzed by two-way ANOVA and Tukey’s multiple comparison test. **p < 0.01, ***p < 0.001

4. Discussion

BCa is the most prevailing malignancy in the urinary system [19] and remains a huge healthy concern. The aggressive MIBC type accounts for approximately 25% of all firstly diagnosed BCa cases and owes a less than 15% 5-year survival rate if without treatment, and 50% of patients die from metastatic disease even with the best therapeutic option [20]. In this paper, we identified a novel molecular mechanism in which FAM83H-AS1 promotes growth, metastasis, and angiogenesis of BCa through the c-Myc-mediated ULK3 upregulation and the following activity of the Hedgehog pathway.

Aberrant expression of lncRNAs has been largely concerned in the development and progression of BCa [21]. For example, UCA1, a landmark oncogenic lncRNA first identified in BCa, triggers cell proliferation, invasion, and resistance to chemo drugs [22–24]. Likewise, other lncRNAs such as small nucleolar RNA host gene 16 (SNHG16) [25,26] and FGFR3-AS1 [27] have been found to increase proliferation, EMT, migration, and invasion of BCa cells. In this paper, we validated a high-expression profile of FAM83H-AS1 in BCa tissues and cell lines compared to the normal ones, which was in line with a previous report [11]. The high expression of FAM83H-AS1 was correlated with positive Ki67 rate, a well-known proliferation biomarker [28], increased TNM staging and lymph node metastasis, as well as poor tumor differentiation and dismal survival in patients. Based on this, artificial downregulation of FAM83H-AS1 was introduced, which inhibited proliferation, invasion, and migration while promoted cell cycle arrest and apoptosis of BCa cells. In addition, the EMT of BCa cells and angiogenesis of HUVECs were decreased as well. During EMT, the typical malignant phase of tumor progression, adhesion between cells was lost, enabling invasion and metastasis of cancer cells [29]. While angiogenesis, defined as the formation of new blood vessels, is recognized as an essential event in cell proliferation and tumorigenesis and diffuse of malignant lesions [30]. The tumor-promoting functions of FAM83H-AS1 have been well noted, including enhancing chemoresistance and the common proliferation and metastasis of cancer cells such as in gastric [31], hepatocellular carcinoma [32], and ovarian cancer [33]. Importantly, in our present research, the in vitro experimental results were reproduced in vivo, where FAM83H-AS1 downregulation inhibits BCa cell growth and metastasis and angiogenesis in nude mice. Collectively, these results evidenced the oncogenic roles of FAM83H-AS1 in BCa progression.

In addition to the well-known RNA transcript networks by which lncRNAs exert versatile functions [34], lncRNAs are also increasingly noted to bind to transcription factors to mediate the following transcripts [35]. Regulation of transcription factors has also been increasingly acknowledged to be important for the etiology of human cancer [36]. Importantly, according to the prediction data obtained from the bioinformatics system and the following RIP and ChIP-qPCR assays, we confirmed that FAM83H-AS1 binds to c-Myc to transcriptionally upregulate ULK3 expression. Consequently, an increased ULK3 expression was found in BCa tissues and cells, presenting as a positive correlation with FAM83H-AS1 expression. C-Myc is one of the top five most seen transcription factors presenting increased activity in human cancers, and dysregulation of c-Myc gene was surprisingly suggested in over 70% of tumors, which affects genomic stability and overall organization of the nucleus, leaving it a great risk factor for cancer development [36–39]. Intriguingly, artificial upregulation of ULK3 in this research reversed the inhibitory effects of FAM83H-AS1 silencing on BCa cell growth, metastasis, and angiogenesis. Though the correlation between ULK3 and tumorigenesis was rarely investigated, it was reported that the ULK3 kinase is critically responsible for the activation of cancer-associated fibroblasts (CAFs), and upregulation of ULK3 was frequently observed in CAFs of many cancers while its suppression was linked to reduced tumor properties [40]. Here, we validated the oncogenic roles of this kinase in BCa. Since ULK3 is recognized as a positive regulator of the Hedgehog signaling pathway [17] we next explored the activation of this signaling in this research. It was found that the levels of Hedgehog marker proteins GLI1, PTCH1, and HhIP in BCa cells were decreased following FAM83H-AS1 knockdown but increased after ULK3 upregulation. A Hedgehog-specific antagonist, Jervine, suppressed the malignant behaviors of T24 and BL10 cells. The Hedgehog signaling is critical for embryonic development and organism homeostasis maintenance, cell proliferation, tissue patterning, and stem cell maintenance, whose abnormal persistence is directly implicated in a wide range of human cancers [41,42]. This pathway has also been concluded to play a crucial role in maintenance or formation of the tumor vasculature, namely to promote the angiogenesis in pancreatic cancer [43]. More relatively, inhibition of hedgehog was found to suppress the self-renewal ability of BCa stem cells to reduce tumorigenesis [44]. Thus, activation of this signaling pathway is possibly involved in FAM83H-AS1-mediated tumorigenic events.

To sum up, this study evidenced that FAM83H-AS1 promotes growth, metastasis, and angiogenesis of BCa cells through c-Myc-mediated ULK3 upregulation and the hedgehog activation. These findings may offer novel insights into gene-based therapies for BCa. We also hope more studies will be conducted in this field to validate our findings and to develop more ideas for BCa treatment.

Funding Statement

This work was supported by Natural Science Research Projects of Anhui Province [KJ2019A0308], National Natural Science Foundation of China [81702495] and Anhui Natural Science Foundation [1808085QH279].

Disclosure statement

All authors declare no conflict of interests in this study.