Background
Head and neck squamous cell carcinoma (HNSCC) is an aggressive malignancy with high local recurrence and lymph node metastasis that affects 600,000 new patients worldwide per year [
1]. Despite great progress in genomic and molecular research, the treatment effects on patients with HNSCC remains unsatisfactory [
2]. Exploration of the complex molecular mechanisms underlying HNSCC occurrence and development is still urgent.
Accumulating studies show that long noncoding RNAs (lncRNAs) participate in diverse physiological and pathological processes, such as cell cycle, apoptosis and metabolism [
3‐
6], as well as reprogramming of induced pluripotent stem cells [
7], organ fibrosis [
8], and initiation of cancer [
9]. HOX transcript antisense RNA (HOTAIR), a well-established lncRNA, is overexpressed in multiple human cancers and acts as a potential prognostic biomarker [
10,
11]. HOTAIR interacts with enhancer of zeste homolog 2 to promote the growth of HNSCC cells [
12]. Overexpression of lncRNA H19 predicts poor prognosis in colorectal cancer [
13], gastric cancer [
14], and HNSCC patients [
15]. An accumulating number of lncRNAs have been confirmed to play a critical role in cancer progression. Long intergenic noncoding RNA p21 (lincRNA-p21) was first identified as a repressor of the p53-dependent apoptotic response after DNA damage in mouse embryonic fibroblasts harboring wild-type p53 [
16]. LincRNA-p21 also provides feedback to enhance wild-type p53 transcriptional activity in vascular smooth muscle cells [
17]. Moreover, 85% of HNSCC patients harbor mutations in the
TP53 gene [
18,
19]. Mutation of the
TP53 gene can not only result in loss of wild-type p53 function or exert a dominant-negative effect over the remaining wild-type allele but also lead to a gain in oncogenic properties that promote tumor growth [
20]. As a transcriptional factor, p53 not only transcribes messenger RNAs but also noncoding RNAs. Whether lincRNA-p21 participates in carcinogenesis and whether its regulation is dependent on p53 status in HNSCC are still unknown.
In this study, we demonstrated that lincRNA-p21 is transcriptionally regulated by the mutant p53/nuclear transcription factor Y subunit alpha (NF-YA) complex. Low lincRNA-p21 expression promoted aggressive progression in HNSCC in vitro and in vivo. Meanwhile, lincRNA-p21 inhibited Janus kinase 2 (JAK2)/signal transducer and activator of transcription 3 (STAT3) signaling by binding to STAT3 and suppressing its transcriptional activation, which is a novel mechanism of lincRNA-p21. Our findings provide insight into how the p53/lincRNA-p21/STAT3 axis contributes to HNSCC development and indicate that lincRNA-p21 may serve as a novel therapeutic target for HNSCC.
Methods
RNAscope, fluorescence in situ hybridization and immunohistochemistry assay
We obtained 70 HNSCC tissues and 9 normal oral mucosal tissues from patients who had undergone surgery between 2007 and 2008 and who were diagnosed by pathological examination. No local or systemic treatment was conducted in these patients before surgery. The tissues were embedded into a tissue microarray. Fresh tumor specimens were collected at surgery and stored at − 80 °C until use. This study was approved by the Ethics Committee of the Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine.
The RNAscope probe targeting lincRNA-p21 was designed and synthesized by Advanced Cell Diagnostics company, and detection of lincRNA-p21 expression was performed on an HNSCC tissue array using an RNAscope 2.5 High Definition (HD)-BROWN Assay kit according to the manufacturer’s instructions (Advanced Cell Diagnostics, Newark, CA, USA). The images were acquired with a Pannoramic MIDI Viewer (3D HISTECH Ltd., Budapest, Hungary). The signals were visually scored based on the average number of dots per cell using the following criteria: 0 (no staining or < 1 dot/10 cells), 1 (1–3 dots/cell), 2 (4–9 dots/cell. None or very few dot clusters), 3 (10–15 dots/cell and < 10% of the dots presented in clusters), 4 (> 15 dots/cell and > 10% of the dots presented in clusters).
Fluorescence in situ hybridization (FISH) was strictly performed according to the manufacture’s protocol (Ribobio Company, Guangzhou, China). Briefly, target-specific probe was designed, synthesized and incubated overnight at 4 °C. 18S RNA and U6 were used as control to indicate the cytoplasm and nucleus.
Immunohistochemistry (IHC) was performed as described in our previous study [
21], with primary antibodies for p-STAT3 (9145) and Ki67 (9129) (1:200 dilution for IHC, Cell Signaling Technology, Danvers, MA, USA). The staining intensity was assessed as follows: 0 = negative; 1 = weak; 2 = moderate; 3 = strong. The staining score was calculated by multiplying the staining intensity and the percentage of positive cells.
Cell culture
The HNSCC cell lines HN4, HN6 and HN30 were kindly provided by Professor Mao Li, University of Maryland and were verified by STR genotyping, and HEK293T, Cal27, SCC25, Detroit562, MCF7 and MDA-MB-231 cell lines were purchased from Type Culture Collection of Chinese Academy of Sciences (Shanghai, China). Normal oral primary keratinocytes were cultured from gingival tissues after tooth extraction from healthy patients. An informed consent was signed by the patients. SCC25 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM)/F12, and the other cell lines were maintained in DMEM (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum, 1% glutamine, and 1% penicillin-streptomycin. Cells were cultured in a standard humidified atmosphere of 5% CO2 at 37 °C.
Cytoplasmic and nuclear RNA isolation and real-time PCR
Cytoplasmic and nuclear RNA was extracted using an RNA Purification Kit (21,000, Norgen Biotek, Thorold, ON, Canada) according to the manufacturer’s instructions. Real-time PCR (qRT-PCR) was performed using a StepOnePlus Real-time PCR system (Thermo Fisher, Waltham, MA, USA) as previously described [
22] and following the manufacturer’s instructions (Takara, Dalian, China). The primer sequences are listed in Additional file
1: Table S1.
Plasmid construction, siRNA, lentivirus and cell transfection
pcDNA3.1 vector containing the human lincRNA-p21 sequence was synthesized by OBiO Technology (Shanghai, China). Plasmids encoding wild-type p53 or mutant p53s (P151S, R175H, G245C and R282W) and lentivirus, including for lincRNA-p21 and si-lincRNA-p21, were synthesized by Gene Pharma Co., Suzhou, China. The lincRNA-p21 promoter reporter gene was cloned into the vector pGL4.0-basic (OBiO Technology, Shanghai, China). Small interfering RNA (siRNA) was synthesized by Biotend Biotechnology, Shanghai, China. The siRNA#3 sequence of lincRNA-21 was packaged into a pGLV-h1-GFP-puro vector for in vivo experiments. Cells were transfected with siRNAs or plasmids using Lipofectamine™ 3000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Treatments were administered 24 h after transfection. The sequences are provided in Additional file
1: Table S1.
Chromatin immunoprecipitation assays
The chromatin immunoprecipitation (ChIP) assay was strictly performed according to the protocol of a SimpleChIP Enzymatic Chromatin IP kit (CST, Danvers, MA, USA) according to the manufacturer’s instructions as described in our previous study [
22,
23]. The primers for the lincRNA-p21 promoter were synthetized by Sangon Biotech (Shanghai, China) and are described in Additional file
1: Table S1.
Dual-luciferase reporter assay
Luciferase assays were performed as described in our previous study [
22]. Briefly, HEK293T cells were cotransfected with each lincRNA-p21 promoter-luciferase construct and pRL-TK promoter Renilla luciferase construct in the vector group, wild-type p53 group and mutant p53 group for 48 h. Whole-cell lysates were extracted, and luciferase activity was determined using a dual luciferase reporter assay system (Beyotime, Shanghai, China) according to the manufacturer’s instructions. A pRL-TK promoter Renilla luciferase construct was used as an internal control.
Electrophoretic mobility shift assay
Electrophoretic mobility shift assay (EMSA) was performed as described previously [
24] and according to the instructions of an EMSA kit (Beyotime, Shanghai, China). HN6, Cal27 and HN30 cells were stimulated with or without Doxorubicin (DOX, 500 nM) for 24 h. Total nuclear cell lysate was incubated with 1 pmol biotin-labeled DNA probes (1000 fmol/μl working fluid concentration) with or without 150 pmol cold competitor probes (50 pmol/μl concentration) consisting of a 16-bp region of the lincRNA-p21 promoter. The biotin-labeled probe sequence was Bio-5′-TTGTCCTTGCCTTTGCTTCCCTAGGGT-3′. The unlabeled cold probe sequence was 5′-CTTGTCCTTGCCTTTGCTTCCCTAGGGTG-3′.
Western blot and immunoprecipitation
Western blotting was performed as described in our previous study [
22]. DOX, MG132 and cryptotanshinone were purchased from Selleck (Houston, TX, USA). IL6 was ordered from PeproTech (Rocky Hill, NJ, USA). The antibodies used in this study were as follows: Rb (181616), p-Rb (184796), E2F1 (179445), Caspase-3 (13847) and cleaved Caspases-3 (32042) antibodies were purchased from Abcam (Cambridge, MA, UK). ERK1/2 (4695), p-ERK1/2 (Thr202/Tyr204) (4370), AKT (4691), p-AKT (Ser473) (4060), Cdc2 (28439), Cyclin B1 (4135), Cyclin D1 (2978), PARP (9532), cleaved PARP (5625), p53 (2524), MMP2 (40994), MMP9 (13667), STAT3 (9139), p-STAT3 (Tyr705) (9145), JAK2 (3230) and p-JAK2 (Tyr1007/1008) (3771) antibodies were from Cell Signaling Technology (Danvers, MA, USA). NF-YA (NBP2–19533) was from NOVUS Biologicals (Littleton, USA). GAPDH antibody (60004–1, Proteintech, Rocky Hill, NJ, USA) were used as an internal control. Immunoreactive bands were scanned and analysed using an Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln, NE, USA).
Immunoprecipitation was operated by incubation with anti-p53 or anti-STAT3 monoclonal antibody with cell lysate at 4 °C rotation overnight. The reaction mix was further incubated with Protein A/G agarose beads for 4 h. The precipitated proteins were detected by immunoblotting with the corresponding antibodies. Anti-GAPDH antibody was used for non-specific control.
Immunofluorescence assay
Cells were plated on coverslips, and incubated at 4 °C overnight with antibodies specific for p-STAT3 (1:100), NF-YA (1:500) and p53 (1:1000). Then, cells were incubated with Cy3 conjugate goat anti-rabbit IgG (SA00009–2, 1:100) or FITC conjugate goat anti-mouse IgG (SA00003–1) (Proteintech, Hubei, China) and then stained with 4′, 6-diamidino-2-phenylindole (DAPI). Labeled cells were visualized on an Axio Vert. A1 microscope (Carl Zeiss, Germany).
Cellular proliferation and invasion assay
HN6 and Cal27 cells were transfected with siRNA or plasmids of lincRNA-p21 for 24 h, and then seeded in the plates. As described in our previous study [
25], cell proliferation experiments were performed using the Cell Counting Kit (CCK8; Dojindo, Kumamoto, Japan) assays. The colony-forming assay was performed to monitor the cloning capability of HN6 and Cal27 cells [
22]. Cell migration and invasion assays were performed using a Transwell technique with uncoated polycarbonate inserts (Millipore, Darmstadt, Germany) or BioCoat™ inserts (BD Biosciences, Franklin Lake, NJ, USA). Medium without FBS (2 × 10
4 cells/200 μl) were added into the upper portion of a migration (uncoated insert) or invasion (matrigel-coated insert) chamber, with 500 μl DMEM containing 10% FBS added into the lower chamber. After crystal violet staining, the crossed cells were dissolved in 33% of acetic acid. OD 570 nm absorbance was measured.
5-Ethyny-2′-deoxyuridine (EdU) assay
The treated HN6 and Cal27 cells were incubated with 50 μM EdU (RiboBio, Guangzhou, China) 100 μl per well for 2 h. After fixation with paraformaldehyde and washes, the cells were treated with 100 μl of 1 × Apollo reaction for 30 min. Then, the DNA of cells were stained with DAPI for 5 min and visualized under Axio Vert. A1 fluorescence microscope.
Flow cytometry
Flow cytometry was performed using previously described methods [
22,
25]. HN6 and Cal27 cells transfected with si-lincRNA-p21, lincRNA-p21, or scramble were harvested 48 h after transfection. After incubating with reagents from the Annexin V-FITC/propidium iodide (PI) apoptosis kit (BD Biosciences, Franklin Lakes, NJ, USA), cells were analysed using a BD FORTASA flow cytometer (BD Biosciences) and the FlowJo software. For cell cycle analysis, cells were incubated using the PI/RNase staining kit (BD Biosciences). The subsequent steps were performed as described above.
Terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) assay
As described in our previous study [
21], TUNEL (Beyotime, Shanghai, China) was performed following the manufacturer’s instructions. Briefly, tissue slice were incubated with fluorescein-labelled dUTP. The nucleus was stained with DAPI. The apoptotic cells (green) were observed and analysed using Axio Vert. A1 fluorescence microscope.
RNA immunoprecipitation analysis
RNA Immunoprecipitation (RIP) assay was performed according to the manufacturer’s instructions of Millipore Magna RIP Kit (Merck KGaA, Darmstadt, Germany). HN6 cells were transfected with vector or lincRNA-p21 plasmid for 24 h and harvested. The cell lysates were incubated with RIP buffer containing magnetic beads conjugated with control IgG, STAT3, phosphorylated STAT3 antibodies overnight at 4 °C. The samples were then incubated with proteinase K to isolate immunoprecipitated RNA. The isolated RNA was reverse transcribed to cDNA and then analyzed by qRT-PCR. qRT-PCR was performed with SYBR Green PCR (Takara, Dalian, China) and primers targeting lincRNA-p21 truncations (Additional file
1: Table S1).
RNA pull-down analysis
RNA pull-down assay was performed according to the instructions provided in a Pierce™ Magnetic RNA-Protein Pull-Down Kit (20,164, Thermo Scientific, Waltham, MA, USA). LincRNA-p21 RNA was transcribed in vitro using a DNA template that included the T7 promoter according to the instructions for a Ribo™ RNAmax-T7 Transcription Kit (RiboBio, Guangzhou, China). Then, lincRNA-p21 was end-labeled with desthiobiotin using a Pierce RNA 3′ End Desthiobiotinylation Kit (20,163, Thermo Scientific). HN6 cells were treated with lincRNA-p21 plasmid for 48 h and then harvested. Biotinylated lincRNA-p21 was captured with streptavidin magnetic beads and then mixed with HN6 cell extract. The protein was eluted from the RNA-protein complex and detected by western blot using STAT3 antibody.
Animal studies
BALB/c nude mice (nu/nu, aged 4 weeks, and weighing approximately 20 g) were purchased from the Shanghai Laboratory Animal Center (Shanghai, China) and were bred in SPF facilities at Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine. The Laboratory Animal Care and Use Committees of the hospital approved all experimental procedures. The tumor xenograft model was established with Cal27 cells as described in our previous study [
26]. The Cal27 cells were (2 × 10
6) stably transduced with lentivirus. The animals were divided into 4 groups: (a) si-Scramble; (b) si-lincRNA-p21; (c) vector; (d) lincRNA-p21. The tumor sizes and animal weights were monitored once a week. The tumor volumes were calculated using the following formula: (length×width
2)/2. Mice were sacrificed, and the tumor tissues were excised after 6 weeks.
Statistical analysis
Statistical analyses were performed using SPSS 13.0 software (SPSS Inc., Chicago, IL, USA). GraphPad Prism version 6 (GraphPad Software, San Diego, CA, USA) was used to plot the data. Student’s t-test and one-way analysis of variance (ANOVA) were performed to assess the significance of differences. P < 0.05 was considered statistically significant (* P < 0.05, and ** P < 0.01). All values are expressed as the means ± standard error.
Discussion
In this study, we first investigated the difference between wild-type and mutant p53 in regulating lincRNA-p21 expression in HNSCC. Moreover, high lincRNA-p21 expression exerted tumor suppressor activity by inhibiting JAK2/STAT3 signaling activation, which suggest that lincRNA-p21 was deeply involved in aggressive progression in HNSCC.
LincRNA-p21 is a relatively novel lncRNA that plays a significant role in the initiation and development of multiple cancers. Previous studies have reported that lincRNA-p21 was significantly downregulated in hepatocellular carcinoma [
31], colorectal cancer [
32], gastric cancer [
33], prostate cancer [
34], non-small cell lung cancer [
35], diffuse large B cell lymphoma [
36], and chronic lymphocytic leukemia [
37]. In addition, the exosomal lincRNA-p21 levels in urine were significantly higher in prostate cancer than in benign prostatic hyperplasia [
38]. In our study, we first demonstrated that lincRNA-p21 was reduced in HNSCC tissues and cell lines compared with normal mucosa. Although we did not obtain the survival data of the patients in our study, we found that HNSCC patients with low lincRNA-p21 expression had worse prognosis using TCGA database. The relationship between the expression level of lincRNA-p21 and prognosis is heterogeneous in different types of tumors. High expression of lincRNA-p21 was linked to favorable survival in diffuse large B lymphoma patients and hepatocellular carcinoma patients, whereas lincRNA-p21 acted as a tumor suppressor, inhibiting cell growth, arresting cycle progression and facilitating apoptosis [
36,
39]. However, in lung cancer, patients with high lincRNA-p21 levels had a worse prognosis than those with low levels [
35]. We showed that lincRNA-p21 inhibited HNSCC cell growth in vitro and in vivo. Furthermore, lincRNA-p21 induced G1 phase arrest and cell apoptosis. Thus, lincRNA-p21 is considered as a tumor suppressor in most of tumors reported thus far, but how lincRNA-p21 promotes tumor progression in HNSCC needs further study.
p53 is a sequence-specific DNA-binding protein that interacts with promoter regions of target genes and regulates transcription. Accumulating evidences indicate that p53 not only transcribes messenger RNAs but also noncoding RNAs, including microRNAs and long noncoding RNAs. Liu et al. [
40] reported that lncRNA loc285194 was induced by p53 and inhibited tumor cell growth. Under glucose starvation conditions, p53 directly upregulated lncRNA TRINGS and protected cancer cells from necrosis [
41]. Consistent with the function of p53-mediated transcription,
TP53 missense mutation mainly occurs in its DNA-binding domain [
42]. Mutations that perturb p53 function or disrupt the upstream or downstream regulatory network of p53 have been found in more than half of all cancer cases [
43]. Even in those cancers that retain wild-type p53, the function of p53 is often inactivated due to alterations in its regulators and/or mediators [
44]. However, the molecular mechanisms whereby mutant p53 regulates gene expression are still unclear. Previous studies have reported that four “hot spot” p53 mutation sites (i.e., P151S, R175H, G245C and R282W) promoted invasive growth of HNSCC cells [
19]. We chose these four mutant p53 types for further study. In the current study, we showed that both wild-type p53 and mutant p53 promoted the expression of lincRNA-p21 in HNSCC cells, especially after stimulation with DOX. The interaction between p53 and lincRNA-p21 was affirmed by p53 binding to the predicted site of the promoter region of lincRNA-p21 and by p53 causing significant induction of lincRNA-p21 promoter activity, as determined by a luciferase reporter assay. An intriguing question is why both wild-type p53 and mutant p53 promote lincRNA-p21 expression. This may be attributed to the lower expression of lincRNA-p21 in HNSCC patients harboring mutant p53 than in normal tissues with wild-type p53. Mutant p53 loses the ability to bind DNA sequences specific to wild-type p53, but it is widely accepted that mutant p53 is still able to influence gene expression through other transcriptional co-factors, such as p63, p73, nuclear transcription factor Y (NF-Y) and vitamin D receptor [
45]. We propose that certain molecules must be involved in regulating lincRNA-p21 expression mediated by mutant p53. Our results demonstrated that mutant p53 transcriptionally regulates lincRNA-p21 depending on NF-YA. Knockdown of NF-YA inhibited the binding affinity between p53-R175H and the lincRNA-p21 promoter but did not affect wild-type p53 function. Knockdown NF-YA expression reversed tumor suppressor activation of lincRNA-p21 in mutant p53 cells, not wild-type p53 cells. These results illustrated that NF-YA plays a critical role in regulation of lincRNA-p21 by mutant p53 but not wild-type p53 in HNSCC.
LncRNAs impact cellular functions through various mechanisms, such as interaction with chromatin remodeling proteins, genomic DNA, mRNA or proteins in the nucleus or cytoplasm [
46]. The basic structural and interactive capabilities of lncRNAs with other cellular biomolecules can help distinguish and specifically reveal their central roles in tumorigenesis. Previous investigations have described diverse cellular functions of lincRNA-p21. Hypoxia-induced lincRNA-p21 attenuated VHL-mediated HIF-1α ubiquitination and promoted glycolysis [
4]. LincRNA-p21 acted in concert with hnRNP-K as a coactivator to promote p53-mediated expression of p21 [
3]. In the absence of HuR, lincRNA-p21 associated with CTNNB1 and JUNB mRNA, and repressed their translation through a mechanism that involves reduced polysome size [
47]. How lincRNA-p21 promotes malignant progression in HNSCC is still obscure. The interactions of lncRNAs with RNA, DNA and protein are involved in various levels of regulation, including transcriptional repression by binding to polycomb repressive complex 2 and serving as a sponge to titrate miRNAs, thus participating in posttranscriptional processing. All the classic molecular mechanisms of lncRNAs, such as guiding, scaffolding and decoying, are ultimately executed through interactions with proteins. The JAK-STAT pathway was originally discovered in the context of interleukin-6 (IL-6)-mediated downstream signaling [
48]. In fact, the hyperactivation of STAT3 signaling occurs in the majority of human cancers and is correlated with a poor prognosis [
29]. Our data showed that p-STAT3 expression was higher in HNSCC tissues and was negatively correlated with lincRNA-p21. Ectopic expression of lincRNA-p21 repressed the IL6-induced STAT3 phosphorylation. LincRNA-p21 was mainly distributed in the cytoplasm of HNSCC cells. Moreover, an RIP assay showed that STAT3 but not p-STAT3 interacted with endogenous lincRNA-p21 under normal conditions; however, p-STAT3 mainly combined with lincRNA-p21 in lincRNA-p21-overexpressed cells. Under normal physiological conditions, STAT3 expression was high in HN6 cells, whereas p-STAT3 was relatively low (Fig.
6a). So lincRAN-21 mainly binds to STAT3 in HN6 cells, which may attribute to competitive advantage of expression. While ectopic expression of lincRNA-p21 was more sensitive to p-STAT3 to bind and inhibit its expression. This is why overexpression of lincRNA-p21 can significantly inhibit p-STAT3 expression, while slightly in STAT3 in HN6 cells. This binding shift can explain the inhibition of phosphorylation of STAT3 after lincRNA-p21 overexpression in HNSCC. The interaction of biotinylated lincRNA-p21 with STAT3 was further confirmed by an RNA pull-down assay. Ectopic expression of lincRNA-p21 also transcriptionally impeded STAT3-regulated gene expression. These data demonstrated that lincRNA-p21 interacts with STAT3 and blocks STAT3 phosphorylation, thereafter repressing the STAT3-regulated downstream genes in HNSCC. Activation of JAK2/STAT3 signaling may serve as the main mechanism that drives tumor progression of low expression of lincRNA-p21 in HNSCC.