A novel stromal lncRNA signature reprograms fibroblasts to promote the growth of oral squamous cell carcinoma via LncRNA-CAF/interleukin-33

A novel stromal lncRNA signature reprograms fibroblasts to promote the growth of oral squamous... Abstract Stromal carcinoma-related fibroblasts (CAFs) are the main type of non-immune cells in the tumor microenvironment (TME). CAFs interact with cancer cells to promote tumor proliferation. Long non-coding RNAs (lncRNAs) are known to regulate cell growth, apoptosis and metastasis of cancer cells, but their role in stromal cells is unclear. Using RNA sequencing, we identified a stromal lncRNA signature during the transformation of CAFs from normal fibroblasts (NFs) in oral squamous cell carcinoma (OSCC). We uncovered an uncharacterized lncRNA, FLJ22447, which was remarkably up-regulated in CAFs, referred to LncRNA-CAF (Lnc-CAF) hereafter. Interleukin-33 (IL-33) was mainly located in the stroma and positively co-expressed with Lnc-CAF to elevate the expression of CAF markers (α-SMA, vimentin and N-cadherin) in fibroblasts. In a co-culture system, IL-33 knockdown impaired Lnc-CAF-mediated stromal fibroblast activation, leading to decreased proliferation of tumor cells. Mechanistically, Lnc-CAF up-regulated IL-33 levels and prevented p62-dependent autophagy–lysosome degradation of IL-33, which was independent of LncRNA-protein scaffold effects. Treatment with the autophagy inducer, rapamycin, impaired the proliferative effect of Lnc-CAF/IL-33 by promoting IL-33 degradation. In turn, tumor cells further increased Lnc-CAF levels in stromal fibroblasts via exosomal Lnc-CAF. In patients with OSCC, high Lnc-CAF/IL-33 expression correlated with high TNM stage (n = 140). Moreover, high Lnc-CAF expression predicted poor prognosis. In vivo, Lnc-CAF knockdown restricted tumor growth and was associated with decreased Ki-67 expression and α-SMA+ CAF in the stroma. In conclusion, we identified a stromal lncRNA signature, which reprograms NFs to CAFs via Lnc-CAF/IL-33 and promotes OSCC development. Introduction The majority of RNA species transcribed from the human genome are non-coding RNAs (ncRNAs), which lack the coding potential capability of producing functional small peptides. Two major classes of regulatory ncRNAs have been widely investigated, including microRNAs (<200 nt in length) and long ncRNAs (lncRNAs) (>200 nt in length) (1,2). They play crucial roles in diverse cellular processes from normal development to disease progression. Deregulation of lncRNAs is involved in carcinogenesis and resistance to therapy (3). In patients with gastric cancer, high levels of lncRNA-HOXA11-AS expression are associated with short survival and poor prognosis (4). We previously found that lncRNA-IL7R is a therapy-resistance factor in oral squamous cell carcinoma (OSCC), and its levels are elevated in response to chemoimmunotherapy (5). Currently, the tumor suppressor and oncogenic roles of lncRNAs are investigated mainly in cancer cells. However, cancer initiation and progression is a multistep process that involves reciprocal autocrine–paracrine communication (e.g. secreted factors and exosomes) between tumor cells and the adjacent stromal microenvironment (6,7). As the most important infiltrated non-immune cells in the stroma, cancer-associated fibroblasts (CAFs) stimulate tumor proliferation and metastasis in a paracrine manner (8). The CAF are mainly derived and activated from normal fibroblasts (NFs), and bone marrow-derived mesenchymal stem cells (9), circulating fibrocytes (10), tissue adipocytes (11) and endothelial cells (12) are also acquire fibroblast markers during carcinogenesis and other inflammatory environment. In ovarian cancer cells, LINC00092 is up-regulated in response to the CAF-derived chemokine, CXCL14 and drives glycolysis and progression of ovarian cancer (13). Similarly, CAFs produce high levels of TGFβ1, which induces the expression of lncRNA-ZEB2NAT in bladder cancer cells, supporting EMT and invasion of cancer cells (14). Evidence regarding lncRNAs in tumor stroma cells is limited. LncRNA-MALAT1 in tumor-associated macrophage increases FGF2 protein secretion, inhibiting inflammatory cytokine release and promoting the proliferation and metastasis of thyroid cancer (15). However, the function of lncRNAs in CAF remains unclear. As an important source of CAFs, NFs are re-educated by the tumor microenvironment (TME) and acquire the characteristics of pro-tumorigenic CAFs (16,17). Therefore, uncovering the mechanisms underlying CAF transformation, re-orchestrating the tumor cells and stromal microenvironment, is needed. In this study, to uncover the LncRNA signature in stroma fibroblasts, we isolated and cultured NFs/CAFs derived from normal and matched tumor tissues from patients with OSCC and examined lncRNAs profiles during the NF/CAF transformation by RNA sequencing. The co-expression networks were analyzed to identify core lncRNA/mRNA interaction, and examined its oncogene or anti-tumor role in vitro and in vivo. Materials and methods Patients and tissue samples Ten paired non-tumor and tumor tissues were used to isolate primary NFs and CAFs from patients with OSCC, respectively. For RNA sequence, five tumors derived from the tongue, gums and cheeks were included in this study during 2015–2016. The patients were 48–66 years old with T (1,2), N(0), M(0) stage, without lymph node metastasis and underwent no preoperative chemotherapy and/or radiotherapy. All the five patients were alive and had no recurrence during this experiment. For the analysis of Lnc-CAF and IL-33 expression, we constructed a cohort including 140 fresh primary OSCC tissue samples. All patients diagnosed with primary OSCC were confirmed by hematoxylin and eosin staining by experienced pathologists from the Department of Pathology at Nanjing Stomatology Hospital. OSCC tissues were evaluated according to the WHO classification and International Cancer Control (UICC) tumor–node–metastasis (TNM) staging system. Ethical approval for this study was obtained from the Research Ethics Committee of Nanjing Stomatology Hospital. Patients who were diagnosed with autoimmune or other malignant diseases as well as pregnant or lactating individuals were excluded from this study. No patient underwent preoperative chemotherapy and/or radiotherapy. Twenty-one patients were lost to follow-up and 119 patients were followed-up until 8 may 2017. Cell lines, mice and reagents The human OSCC cell line, HSC-3, was obtained from Professor Yvonne L. Kapila (Michigan University, MI). The cells were characterized by mycoplasma detection, DNA fingerprinting, isozyme detection and cell viability by the provider. No further authentication of the cell line was conducted. The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS (Life Technologies, Carlsbad, CA) at 37°C, 5% CO2 condition. Four- to five-week-old male BALB/c-nu/nu T cell-deficient mice were purchased from Cavens (Changzhou, China). siRNA-Lnc-CAF/IL-33, pcDNA3.1-Lnc-CAF/IL-33 or negative control were purchased from RiboBio (Guangzhou, China). Detailed information on the antibodies and reagents used in this study is provided in Supplementary Table 1, available at Carcinogenesis Online. Isolation of cancer-associated fibroblasts and RNA sequence Sterile fresh OSCC tissues and its corresponding normal tissues was collected from surgery and washed with phosphate-buffered solution (PBS) and antibiotics, and then removed the epithelial and adipose tissues. The specimens were cut into small pieces and digested by enzyme mix (Collagenase, Neutral protease, Hyaluronidase) for 30 min. The remaining small tissues were placed in synthetic DMEM/F12 basic medium and incubated at 37°C. The medium was replaced every 2–3 days and the epithelial cells were easily removed via trypsinase. The remaining cells were fibroblasts and were collected by further digestion with trypsinase for 2 min. We performed the RNA-seq with the help of Novel Bioinformatics Co., Ltd (Shanghai, China). Firstly, total RNA was extracted by Trizol reagent (Invitrogen) separately. The RNA quality was checked by Bioanalyzer 2200 (Aligent) and kept at −80°C. The RNA with RIN >8.0 is right for cDNA library construction. Secondly, the complementary DNA (cDNA) libraries for single-end sequencing were prepared using Ion Total RNA-Seq Kit v2.0 (Life Technologies). The cDNA libraries were then processed for the Proton Sequencing process according to the commercially available protocols. Thirdly, mapping of single-end reads. The clean reads were then aligned to human genome (version: GRCh38.p1) using the MapSplice program (v2.2.0). Finally, pathway analysis was performed and we also applied EBseq algorithm to filter the differentially expressed genes, the significant analysis and false discovery rate analysis were performed. Besides, we presented gene co-expression Networks to find the relations among different mRNA and LncRNA (Supplementary Figure 1 and Table 2, available at Carcinogenesis Online). The bioinformatics procedures and algorithms were included in Supplementary Materials and Methods, available at Carcinogenesis Online. Immunohistochemistry and immunofluorescence assays Immunohistochemistry and immunofluorescence assays were performed as previously described (18,19). Briefly, the fibroblasts, HSC3 or the specimen frozen sections were collected and covered with a depth of 2–3 mm with 4% formaldehyde for 15 min and washed with PBS. Then cells were permeabilized with 100% methanol for 10 min at −20°C and blocked with 3% bovine serum albumin and were incubated with primary antibodies (dilution: 1:100 or 200) overnight at 4°C. After rinsing three times in PBS, incubated coverslips in fluorochrome-conjugated secondary antibody (Dilution: 1:400) for 1–2 h at room temperature in dark and then stained with DAPI (Bioword, China). The FISH assay was performed by the manufacturer’s protocol (RiboBio, Guangzhou, China), U6 and 18S were used for cytoplasmic and nuclear positive control. Finally, the coverslips were mounted onto the glass slides with neutral gum and observed by FV10i confocal microscope (OLYMPUS, Japan). Immunohistochemistry Protein expression was analyzed immunohistochemically on 2-μm-thick, formalin-fixed and paraffin-embedded specimen sections. Slides were incubated in three washes of xylene for 5 min each and were followed by two washes of 100% ethanol for 10 min, 95% ethanol for 10 min and ddH2O for 5 min each. Antigen unmasking was prepared by boiling in pH 9.0, 10 mM Tris/1 mM EDTA, blocked with 3% hydrogen peroxide for 10 min at room temperature and washed. Then anti-IL-33 antibody (diluted ×100), anti-Ki-67 antibody (diluted ×400) anti-α-SMA antibody (diluted ×200) and anti-CD31 antibody (diluted ×200) were incubated the FFPE specimen sections at 4°C overnight and then the EnVision Detection System kit (DAKO, Denmark) was used for the DAB chromogen followed by nuclear staining using hematoxylin. Neutral gum was used to cover the sliders and dry at room temperature for counting. RNA isolation and qRT-PCR Total RNA was extracted using Trizol reagent (Invitrogen) according to the standard RNA isolation protocol. Quantitative real-time RT-PCR (qRT-PCR) was performed, and cDNA synthesis was performed using a PrimeScript RT Reagent Kit (Takara, China). The PCR amplification was performed with the conditions of 95°C for 10 s, 40 cycles of 94°C for 30 s, 60°C for 30 s and 72°C for 30 s on an ABI 7900 system (Applied Biosystems, USA) with SYBR green real-time PCR Master Mix (Takara, China). The relative levels of genes were calculated by the 2−ΔΔCt method. The expression levels of α-SMA, FSP-1, IL-33, Ki-67 and Lnc-CAF were normalized to GAPDH for gene expression analysis. Each experiment was performed in triplicate. Transwell assay About 2 × 105 NFs/CAFs cells were in the lower chamber and tumor cells were plated in the upper chamber of a non-coated transwell insert with 0.45 nm aperture. After the indirect co-culture for 48 h and cells were stained by crystal violet and counted under an inverted microscope. Five random views were selected to count the cells. Cell treatment and transfection The degradation of IL-33 protein in CAFs was induced by the protein synthesis inhibitor, cycloheximide (CHX) (50 μg/ml). The pathways of IL-33 degradation were analyzed by the autophagy–lysosome inhibitors, chloroquine (CQ) (25 μM) and 3-methyladenine (3-MA) (2 mM) and a proteasome inhibitor, MG132 (5 μM) before the knockdown of Lnc-CAF. The neutralizing antibody for IL-33 was used at 1 μg/ml. The recombinant human IL-33 was used in 10 and 50 ng/ml. According to the manufacturer’s instructions, the NF/CAFs were seeded into 12 or 6-well plate and then cells were transfected with siRNA-Lnc-CAF/IL-33, pcDNA3.1-Lnc-CAF/IL-33 or negative control by Lipofectamine 2000 (Invitrogen). After transfection for the indicated time, the cells were harvested for further experiments. Preparation of cell extracts and immunoblotting For immunoblotting analysis, cells were collected with PBS and then lysed in RIPA buffer with protease inhibitors and phosphatase inhibitors (Roche Applied Science, Basel, Switzerland). Detailed procedures for immunoblotting were previously described (20,21). Flow cytometry assay For the identification of pan-CK-positive HSC3 cells, FITC-anti-human pan-CK (Cat No.ab78478, Abcam, Cambridge, MA) was used to stain HSC3 cells. For the identification of CD63+ exosomes, PE-anti-human CD63 (Cat No. 353003, Biolegend, San Diego, CA) was used according to the manufacturer’s instructions and quantified by flow cytometry on a FACS Calibur instrument. Immunoprecipitation (IP) assay CAFs were transfected with siRNA-Lnc-CAF or negative control following the manufacturer’s protocol. The cells were then harvested and crosslinked with 1% formaldehyde at room temperature. After sonication and removal of cell debris, 2 mg anti-p62 mAb (Cat No.18420, Proteintech, Rosemont, IL) was added together with 10 ml Protein G Sepharose slurry and incubated overnight at 4°C. IL-33 immunoprecipitates (IPs) were washed three times. The proteins were then released from the beads by heating with Laemmli buffer, followed by SDS-PAGE and western blot analysis. Precipitated proteins were detected with anti-p62 (Cat No.18420, Proteintech) and anti–IL-33 mAb Nessy-1 (Cat No. ALX-804–840, Enzo), respectively. RNA binding protein immunoprecipitation The RNA immunoprecipitation assay (RIP) assay was performed by using the Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (Millipore, Billerica, MA) according to the manufacturer’s instructions. Briefly, cells were harvested and lysed in RIP lysis buffer. RNA was immunoprecipitated with an antibody against IL-33 and protein A/G magnetic beads. The magnetic bead bound complexes were immobilized with a magnet and unbound materials were washed off. Then, RNAs of Lnc-CAF were extracted and analyzed by qRT-PCR. Exosome isolation and characterization Exosomes from the culture supernatants were isolated by differential centrifugation as we previously described (20,21). Tumor mouse model To investigate the role of Lnc-CAF in the TME of OSCC, CAFs or HSC3 cells were infected with lentivirus-RNAi-Lnc-CAF (MOI = 100) or negative control. The two types of cells infected with lentivirus-RNAi-control were used as control groups in this study. The cells infected with lentivirus-RNAi-Lnc-CAF were used as test groups. The HSC3 cell lines used alone is the negative control group in this study. These conditional cells were subcutaneously co-injected in the rear flank of nude mice (eight per group) as indicated description. The animal study was approved by institutional animal research committee of Nanjing University and that animals were cared for following the guidelines for use and care of laboratory animals. Statistical analyses Statistical analyses were performed on Statistical package for social sciences version 16.0 (SPSS 16.0, SPSS Inc., Chicago, IL) and Prism statistical software package (GraphPad Software Inc., San Diego, CA). The relationships between the presence of Lnc-CAF/IL-33 and clinicopathological characteristics were determined by unpaired t-test. Unpaired t-test or Mann–Whitney U test was used to compare two groups and the differences between more than two groups were analyzed by the Kruskal–Wallis test. Differences were considered statistically significant with P < 0.05. Results Characteristics of NFs and CAFs in the OSCC microenvironment To determine LncRNA profiles during the transformation of stromal fibroblasts, we first collected 10 paired non-tumor and tumor tissues from patients with OSCC and fibroblasts were isolated and cultured. To confirm the identity of CAFs, the phenotype and function of the paired NFs/CAFs were evaluated. CAFs as stromal mesenchymal cells are often identified by the expression of αSMA and other markers, including fibroblast specific protein (FSP-1) (17). The expression of CAF markers, α-SMA and FSP-1, were higher in CAFs than those in NFs (Figure 1A). OSCC cells were derived from the epithelium and used as control cells. We further analyzed the expression of cytokeratin, a specific marker for epithelial cells. NFs and CAFs were negative for cytokeratin, but the OSCC cell line, HSC3, showed strong positive staining (Figure 1A), indicating that we successfully isolated tissue-derived stromal fibroblasts. Furthermore, α-SMA and FSP-1 mRNA levels were also upregulated in CAFs (Figure 1B). Evidence demonstrated that CAFs promote tumor progression by regulating proliferation, migration, and therapy-induced apoptosis. Our previous findings showed that CAFs derived from patients with OSCC inhibited chemoimmunotherapy-induced cell death (5). Here, we confirmed that, when compared with HSC3 cells alone, CAFs stimulated tumor proliferation in vitro, although NFs also showed a very weak ability to induce tumor cell growth (Figure 1C). After the phenotype and function validation of paired NFs/CAFs, five tumors derived from the tongue, gums, and cheeks were included in this study during 2015–2016 were selected for RNA-sequence. The patients were 48–66 years old with T (1,2), N(0), M(0) stage, without lymph node metastasis and underwent no preoperative chemotherapy and/or radiotherapy. All the five patients were alive and had no recurrence during this experiment. Figure 1. View largeDownload slide Characteristics of normal fibroblasts (NFs) and carcinoma-related fibroblasts (CAFs) in oral squamous cell carcinoma (OSCC). (A) Three paired NFs and CAFs were isolated from three independent patients with OSCC and cultured. Their phenotype was determined by staining for the CAF markers, α-SMA (Green) and FSP-1 (Red), via immunofluorescence. The OSCC cell line, HSC3, was used as control cells and fibroblasts were negative for the epithelial cells marker, pan-cytokeratin (Red). (B) α-SMA and FSP-1 mRNA expression was determined by Q-PCR. (C) The transwell assay was performed to estimate the proliferation of HSC3 cells induced by NFs and CAFs. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Figure 1. View largeDownload slide Characteristics of normal fibroblasts (NFs) and carcinoma-related fibroblasts (CAFs) in oral squamous cell carcinoma (OSCC). (A) Three paired NFs and CAFs were isolated from three independent patients with OSCC and cultured. Their phenotype was determined by staining for the CAF markers, α-SMA (Green) and FSP-1 (Red), via immunofluorescence. The OSCC cell line, HSC3, was used as control cells and fibroblasts were negative for the epithelial cells marker, pan-cytokeratin (Red). (B) α-SMA and FSP-1 mRNA expression was determined by Q-PCR. (C) The transwell assay was performed to estimate the proliferation of HSC3 cells induced by NFs and CAFs. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Lnc-CAF/IL-33 is up-regulated during NF/CAF transformation The results of RNA sequencing indicated that the mRNA/lncRNA profiles significantly changed during NF/CAF transformation. Differentially expressed mRNAs and lncRNAs with a fold change above 2 and a false discovery rate less than 0.05 were selected for co-expression network analysis. An uncharacterized intergenic lncRNA LOC400221, also known as FLJ22447 (NCBI Gene ID: 400221) was upregulated in CAFs (referred to as Lnc-CAF) (Figure 2A). Lnc-CAF is located in 14q23.1-q23.2 which is reported to harbor cancer risk locus and associated with shorter survival of patients in prostate cancer (22), colorectal cancer (23), lung cancer (24) and head and neck squamous cell carcinomas (25). More importantly, the analysis of co-expression networks indicated that Lnc-CAF was located in the core position of differentially expressed genes. And according to its level in fibroblasts and its fold-change, Lnc-CAF is highly expressed in fibroblasts (Supplementary Table 2, available at Carcinogenesis Online). Therefore, Lnc-CAF was chosen, and we speculated that it was involved in the NF/CAF transformation (Supplementary Figure 1A, available at Carcinogenesis Online). Meanwhile, the five paired NFs and CAFs isolated from five independent patients with OSCC were cultured for validation of five differentially expressed genes (ASMA, YKL-40, IL-33, LOC100506114 and Lnc-CAF) by real-time PCR (Supplementary Figure 1B, available at Carcinogenesis Online). Figure 2. View largeDownload slide Lnc-CAF and IL-33 are up-regulated in activated CAFs. (A) The gene expression between CAFs and NFs (vertical axis) and average expression of genes in CAFs versus those in NFs (horizontal axis) are presented as a Bland-Altman plot. Data are from our RNA-seq analysis. Highlighted in red are 35 lncRNAs with significant changes in expression (fold > 2, false discovery rate < 0.05). (B) The cytoplasmic and nuclear fractions of CAFs were prepared. After RNA extraction, the levels of Lnc-CAF, U6 and 18S transcripts were determined by Q-PCR. (C) The expression pattern of Lnc-CAF in NFs, CAFs, and OSCC tumor samples was determined by FISH and immunofluorescence. (D) The co-expression network of Lnc-CAF and IL-33 was analyzed. (E) The expression of IL-33 in the tumor nest and stroma of OSCC samples was determined by immunohistochemistry, T, tumor nest. S, stroma. ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Figure 2. View largeDownload slide Lnc-CAF and IL-33 are up-regulated in activated CAFs. (A) The gene expression between CAFs and NFs (vertical axis) and average expression of genes in CAFs versus those in NFs (horizontal axis) are presented as a Bland-Altman plot. Data are from our RNA-seq analysis. Highlighted in red are 35 lncRNAs with significant changes in expression (fold > 2, false discovery rate < 0.05). (B) The cytoplasmic and nuclear fractions of CAFs were prepared. After RNA extraction, the levels of Lnc-CAF, U6 and 18S transcripts were determined by Q-PCR. (C) The expression pattern of Lnc-CAF in NFs, CAFs, and OSCC tumor samples was determined by FISH and immunofluorescence. (D) The co-expression network of Lnc-CAF and IL-33 was analyzed. (E) The expression of IL-33 in the tumor nest and stroma of OSCC samples was determined by immunohistochemistry, T, tumor nest. S, stroma. ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. To determine the subcellular location of Lnc-CAF, the cytoplasmic and nuclear fractions were isolated. Lnc-CAF was detected in both the cytoplasm and nucleus in CAFs (Figure 2B). FISH of Lnc-CAF was performed in NFs, CAFs and OSCC samples. Clinical evidence implicated that both tumor cells and stromal fibroblasts expressed Lnc-CAF, but the expression of Lnc-CAF seemed to be higher in tumor nests (pan-CK-positive) than in CAFs (α-SMA-positive) (Figure 2C). Interestingly, as an important cytokine involved in inflammation and cancer, IL-33 was associated with Lnc-CAF in co-expression networks and was up-regulated in CAFs (Figure 2D). In clinical samples, IL-33 was mainly expressed in the tumor stroma, but rarely in the tumor nest, which indicated that IL-33 is indispensable for the stromal microenvironment (Figure 2E). These findings indicated that enhanced stromal Lnc-CAF/IL-33 signals were associated with a CAF-dependent TME in clinical specimens. Upregulated Lnc-CAF promotes the CAF phenotype via IL-33, supporting tumor cell proliferation We next analyzed the function of Lnc-CAF/IL-33 signals in fibroblasts. Lnc-CAF was silenced by siRNA in CAFs and overexpressed in NFs. Two siRNAs were designed and siRNA-02 was chosen for the subsequent experiments (Figure 3A). Effective knockdown of Lnc-CAF in CAFs (Figure 3B) inhibited the expression of α-SMA, vimentin and N-cadherin, indicating that Lnc-CAF promoted the mesenchymal phenotype of CAFs (Figure 3C). Besides, forced expression of Lnc-CAF in NFs induced the expression of α-SMA, vimentin, and N-cadherin and enhanced the mesenchymal phenotype of CAFs (Figure 3D and E). The results indicated that Lnc-CAF is required for the maintenance of the stromal phenotype of CAF. Figure 3. View largeDownload slide Lnc-CAF activates CAF phenotype via IL-33. (A) Two siRNAs against Lnc-CAF were transfected into CAFs to select the most efficient one. (B, C) Different concentrations of siRNA-LncRNA-CAF-002 were transfected into CAFs and the expression of LncRNA-CAF, α-SMA, vimentin, and N-cadherin was determined by Q-PCR or western blotting. (D, E) Overexpression of Lnc-CAF in NFs via transfection of pcDNA 3.1-Lnc-CAF and the CAF phenotype was analyzed by western blotting. (F) The correlation between the level of Lnc-CAF and IL-33 was analyzed in eight paired NFs and CAFs. (G) After overexpression of Lnc-CAF in NFs, the siRNA-IL-33 was transfected and the CAF phenotype was analyzed. (H, I) NFs were treated with recombinant human IL-33 at the indicated dose. The expression levels of Lnc-CAF, α-SMA, vimentin, and N-cadherin were determined. (J) NFs were transfected with pcDNA 3.1-Lnc-CAF and/or siRNA-IL-33, and the transfected NFs were directly co-cultured with HSC3 cells. The pan-CK-positive cells were referred to as HSC3 cells and the proliferation of HSC3 cells was determined by pan-CK staining by flow cytometry. HSC3 cells alone were used as positive control and the IgG-isotype antibody was used as a negative control. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Figure 3. View largeDownload slide Lnc-CAF activates CAF phenotype via IL-33. (A) Two siRNAs against Lnc-CAF were transfected into CAFs to select the most efficient one. (B, C) Different concentrations of siRNA-LncRNA-CAF-002 were transfected into CAFs and the expression of LncRNA-CAF, α-SMA, vimentin, and N-cadherin was determined by Q-PCR or western blotting. (D, E) Overexpression of Lnc-CAF in NFs via transfection of pcDNA 3.1-Lnc-CAF and the CAF phenotype was analyzed by western blotting. (F) The correlation between the level of Lnc-CAF and IL-33 was analyzed in eight paired NFs and CAFs. (G) After overexpression of Lnc-CAF in NFs, the siRNA-IL-33 was transfected and the CAF phenotype was analyzed. (H, I) NFs were treated with recombinant human IL-33 at the indicated dose. The expression levels of Lnc-CAF, α-SMA, vimentin, and N-cadherin were determined. (J) NFs were transfected with pcDNA 3.1-Lnc-CAF and/or siRNA-IL-33, and the transfected NFs were directly co-cultured with HSC3 cells. The pan-CK-positive cells were referred to as HSC3 cells and the proliferation of HSC3 cells was determined by pan-CK staining by flow cytometry. HSC3 cells alone were used as positive control and the IgG-isotype antibody was used as a negative control. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. IL-33 level was positively correlated with Lnc-CAF level in eight paired NFs and CAFs (Figure 3F). Thus, we next determined whether IL-33 was indispensable for Lnc-CAF-induced phenotype of CAFs. The results showed that, although Lnc-CAF activated the expression of α-SMA, vimentin and N-cadherin, IL-33 knockdown reversed the stromal phenotype regulated by Lnc-CAF, suggesting that Lnc-CAF maintains the stromal phenotype of CAFs via IL-33 (Figure 3G). Given that IL-33 receptor (IL-33R or ST2) was first identified on the surface of fibroblasts and mast cells, we speculated an autocrine pathway of IL-33/ST2 in fibroblasts for IL-33-dependent CAF phenotype. Therefore, NFs were treated with recombinant human IL-33 for 24 h, which led to increased Lnc-CAF expression in NFs (Figure 3H) as well as increased expression of α-SMA, vimentin and N-cadherin (Figure 3I). Initial signaling complex of MyD88-IRAK4-TRAF6 signals are activated by the formation of the IL-33 receptor complex. Here, IL-33 induced the phosphorylation of IRAK4, which was accompanied by enhanced CAF phenotype (Figure 3I). These data indicated that IL-33 uses multiple pathways to regulate the activation of the CAF phenotype. Furthermore, we examine the impacts of Lnc-CAF/IL-33 signaling in the stroma on tumor cells. We first overexpressed Lnc-CAF and/or siRNA-IL-33 in cultured NFs, which were then directly co-cultured with HSC3 cells. HSC3 cells alone were used as positive controls. NFs overexpressing Lnc-CAF enhanced the proliferation of HSC3 cells (pan-CK+ cells: 65%) when compared with normal NFs (pan-CK+ cells: 46%), but this effect was attenuated by IL-33 knockdown (pan-CK+ cells: 48%) (Figure 3J). These findings suggest that up-regulated Lnc-CAF/IL-33 signaling in stromal fibroblasts enhances the proliferation of HSC3 cells. Lnc-CAF prevents the degradation of IL-33 by the p62-dependent autophagy–lysosome pathway, supporting tumor growth To identify the mechanism by which high abundance of Lnc-CAF/IL-33 was maintained in CAFs, we knocked down Lnc-CAF in CAFs. Lnc-CAF knockdown in CAFs led to reduced expression of IL-33. In cycloheximide-chase assays, the protein synthesis inhibitor, cycloheximide (CHX), could further accelerate the degradation of IL-33 protein when Lnc-CAF was knocked down (Figure 4A). This result suggested that Lnc-CAF was involved in the maintenance of IL-33 protein stability. Because lncRNAs can act as decoys, scaffolds, guides or enhancers (26), we speculated that Lnc-CAF functioned as a lncRNA scaffold to maintain the stability of IL-33 protein and to inhibit its degradation. Figure 4. View largeDownload slide Lnc-CAF improves IL-33 stability via inhibition of p62-dependent autophagy–lysosome degradation. (A) The degradation of IL-33 protein in CAFs was induced by the protein synthesis inhibitor, cycloheximide (CHX) (50 μg/ml) for the indicated time, the level of IL-33 was normalized to that of GAPDH. (B) The autophagy–lysosome inhibitors, chloroquine (CQ) (25 μM) and 3-methyladenine (3-MA) (2 mM), and a proteasome inhibitor, MG132 (5 μM), were added to CAFs knocked down for Lnc-CAF to inhibit the degradation of IL-33. (C) The autophagy level in CAFs knocked down for Lnc-CAF was analyzed by western blot. (D) The IL-33/p62 complex was pulldown by an anti-p62 antibody and the IL-33 level was determined. (E) RNA immunoprecipitation (RIP) assay was performed to determine the direct interaction between Lnc-CAF and IL-33. An Anti-IL-33 antibody was used to pulldown the RNA/protein complex and the IgG isotype antibody was used as a negative control. (F) NFs were transfected with pcDNA 3.1-IL-33 and/or treated with rapamycin, and directly co-cultured with HSC3 cells. The proliferation of HSC3 cells was analyzed via pan-CK staining by flow cytometry. The IgG-isotype antibody was used as a negative control. *P < 0.05, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Figure 4. View largeDownload slide Lnc-CAF improves IL-33 stability via inhibition of p62-dependent autophagy–lysosome degradation. (A) The degradation of IL-33 protein in CAFs was induced by the protein synthesis inhibitor, cycloheximide (CHX) (50 μg/ml) for the indicated time, the level of IL-33 was normalized to that of GAPDH. (B) The autophagy–lysosome inhibitors, chloroquine (CQ) (25 μM) and 3-methyladenine (3-MA) (2 mM), and a proteasome inhibitor, MG132 (5 μM), were added to CAFs knocked down for Lnc-CAF to inhibit the degradation of IL-33. (C) The autophagy level in CAFs knocked down for Lnc-CAF was analyzed by western blot. (D) The IL-33/p62 complex was pulldown by an anti-p62 antibody and the IL-33 level was determined. (E) RNA immunoprecipitation (RIP) assay was performed to determine the direct interaction between Lnc-CAF and IL-33. An Anti-IL-33 antibody was used to pulldown the RNA/protein complex and the IgG isotype antibody was used as a negative control. (F) NFs were transfected with pcDNA 3.1-IL-33 and/or treated with rapamycin, and directly co-cultured with HSC3 cells. The proliferation of HSC3 cells was analyzed via pan-CK staining by flow cytometry. The IgG-isotype antibody was used as a negative control. *P < 0.05, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. CAFs were transfected with siRNA-Lnc-CAF, followed by treatment with various inhibitors of protein degradation pathways to investigate the mechanisms underlying the regulation of IL-33 stability by Lnc-CAF. The autophagy–lysosome inhibitors, chloroquine (CQ) and 3-methyladenine (3-MA), but not MG132, a proteasome inhibitor, remarkably inhibited the degradation of IL-33 in CAFs knocked down for Lnc-CAF, suggesting that Lnc-CAF deficiency resulted in the degradation of IL-33 via the autophagy–lysosome pathway (Figure 4B). Interestingly, we observed a significant increase of the LC3II/LC3I ratio in CAFs knocked down for Lnc-CAF, implying that Lnc-CAF deficiency induced the initiation of autophagy (Figure 4C). p62 is a cargo protein involved in the protein degradation via selective autophagy. After Lnc-CAF knockdown, the interaction between p62 and IL-33 was increased, leading to the degradation of IL-33 via selective autophagy (Figure 4D). To further provide direct evidence of the role of Lnc-CAF as a lncRNA scaffold, RIP assay showed that the Lnc-CAF level in the co-immunoprecipitated complex with the anti-IL-33 antibody was not significantly different from that observed with the IgG isotype antibody, suggesting that Lnc-CAF-induced stability of IL-33 was not associated with its lncRNA scaffold function (Figure 4E). In the NF/HSC3 cell co-culture system, overexpression of IL-33 in NFs promoted the proliferation of tumor cells (pan-CK+ cells: 75%), but the induction of autophagy by rapamycin impaired this proliferative effect by promoting IL-33 degradation (pan-CK+ cells: 47%) (Figure 4F). Tumor cell-derived exosomal Lnc-CAF increases Lnc-CAF levels in stromal fibroblasts in a positive feed-back loop During the carcinogenesis or therapy resistance, secreted factors and exosomes in the stroma are essential mediators for the reciprocal autocrine–paracrine communication between tumor cells and the adjacent stromal microenvironment (27,28). We here investigated the impacts of OSCC-derived exosomes on stromal fibroblasts. The HSC3 cells were treated with or without the exosome secretion blocker, GW4869, and labeled with CM-Dil (red) for NF/HSC3 co-culture. CM-Dil was used to label the membranes and lipids (Figure 5A). We observed an increase in Dil+ NFs when NFs were co-cultured with normal HSC3 cells, but the GW4869-treated HSC3 cells secreted less exosomes, which resulted in decreased Dil+ NFs (Figure 5B). These data indicated that OSCC-derived NFs could effectively ingest tumor cell-derived exosomes. Thus, NFs were cultured in the conditional tumor cell-derived culture medium (CM). The results showed that Lnc-CAF/IL-33 expression was induced by the conditional tumor cell-derived CM, but not by CM from cells treated with GW4869 (Figure 5C). IL-33 promotes the activation of CAFs. An IL-33 neutralizing antibody also inhibited CM-induced Lnc-CAF/IL-33 expression in NFs (Figure 5C), suggesting that both the exosomes and IL-33 in tumor cell-derived CM contributed to the elevated Lnc-CAF /IL-33 expression in NFs. Figure 5. View largeDownload slide Tumor cells promote the upregulation of Lnc-CAF in NFs in a positive feedback loop. (A, B) HSC3 cells were pre-treated with or without the exosome secretion blocker, GW4869 and labeled with CM-Dil (red). NFs were directly co-cultured with labeled HSC3 cells for 18 h and the CM-Dil positive cells were analyzed. The ratio of CM-Dil+/pan-CK- NFs was calculated. (C) NFs were treated with the conditioned tumor cell culture medium for 48 h and the expression of Lnc-CAF and IL-33 was assessed by Q-PCR. The neutralizing antibody for IL-33 was used at 1 μg/mL. (D, E) Representative images of HSC3-derived exosomes are shown. Tracking analysis using qNano showed that the average size was approximately 95 nm and the CD63 and Alix expression in exosomes derived from HSC3 cells was estimated by western blotting. (F) The expression of Lnc-CAF in exosomes was analyzed and normalized to that of GAPDH. (G, H) The exosomes were labeled with PKH67 (green) or anti-CD63 antibody and added to NFs (5 μg/ml), and the level of Lnc-CAF in NFs was estimated by Q-PCR. *P < 0.05, **P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Figure 5. View largeDownload slide Tumor cells promote the upregulation of Lnc-CAF in NFs in a positive feedback loop. (A, B) HSC3 cells were pre-treated with or without the exosome secretion blocker, GW4869 and labeled with CM-Dil (red). NFs were directly co-cultured with labeled HSC3 cells for 18 h and the CM-Dil positive cells were analyzed. The ratio of CM-Dil+/pan-CK- NFs was calculated. (C) NFs were treated with the conditioned tumor cell culture medium for 48 h and the expression of Lnc-CAF and IL-33 was assessed by Q-PCR. The neutralizing antibody for IL-33 was used at 1 μg/mL. (D, E) Representative images of HSC3-derived exosomes are shown. Tracking analysis using qNano showed that the average size was approximately 95 nm and the CD63 and Alix expression in exosomes derived from HSC3 cells was estimated by western blotting. (F) The expression of Lnc-CAF in exosomes was analyzed and normalized to that of GAPDH. (G, H) The exosomes were labeled with PKH67 (green) or anti-CD63 antibody and added to NFs (5 μg/ml), and the level of Lnc-CAF in NFs was estimated by Q-PCR. *P < 0.05, **P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. To confirm the roles of tumor-derived exosomal Lnc-CAF in the stroma, the HSC3 cell-derived exosomes were isolated and their size and phenotype were determined (Figure 5D and E). High levels of Lnc-CAF were detected in exosomes (Figure 5F). PKH67 (green) was used to label exosomes. PKH67+ exosomes could be taken up by NFs and up-regulated the expression of Lnc-CAF in NFs (Figure 5G and H). This effect was also observed in CAFs (unpublished data), indicating that tumor cells could secrete exosomal Lnc-CAF into the stroma and induce Lnc-CAF expression in stromal fibroblasts, leading to an Lnc-CAF-mediated positive feed-back loop between NFs/CAFs and tumor cells. High Lnc-CAF/IL-33 levels correlate with high tumor stage in patients with OSCC Considering the high level of Lnc-CAF/IL-33 in the TME, we further analyzed the correlation between aberrant Lnc-CAF/IL-33 expression and the clinicopathological characteristics of patients with OSCC (n = 140). Patients with high TNM stage, but not with LNM, often harbored high expression levels of Lnc-CAF and IL-33 (Figure 6A and B). Importantly, the expression of the proliferation index, Ki-67, was higher in patients with higher expression of Lnc-CAF/IL-33, which confirmed that the enhanced Lnc-CAF/IL-33 signature facilitated the growth of OSCC in vivo. Additionally, the prognostic values of Lnc-CAF/IL-33 were estimated; high level of Lnc-CAF, but not IL-33, predicted poor clinical survival outcome of patients with OSCC (Figure 6C and D). Figure 6. View largeDownload slide The significance of high Lnc-CAF levels in patients with OSCC and xenograft mouse model. (A, B) The relationship between Lnc-CAF/IL-33 expression and clinicopathological characteristics in patients with OSCC was analyzed, including TNM stage, Ki-67 and lymph node metastasis. (C, D) The prognostic value of Lnc-CAF/IL-33 was determined. (E) Knockdown of Lnc-CAF in CAFs and/or HSC3 cells and the xenograft mouse model of OSCC was established. Knocked down cells were subcutaneously injected in the rear flank of the nude mice (eight per group). The mean tumor size (mm3) was determined. (F) The micro-vessel density (MDV) was estimated by CD31 immunohistochemistry (200×); the expression of the proliferation index, Ki-67 (400×) and α-SMA (200×) in CAFs was analyzed. (G) The mRNA expression of Lnc-CAF and IL-33 was determined by Q-PCR. (H) Model summarizing the pro-tumorigenic effect of Lnc-CAF/IL-33 in the tumor microenvironment of OSCC. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Figure 6. View largeDownload slide The significance of high Lnc-CAF levels in patients with OSCC and xenograft mouse model. (A, B) The relationship between Lnc-CAF/IL-33 expression and clinicopathological characteristics in patients with OSCC was analyzed, including TNM stage, Ki-67 and lymph node metastasis. (C, D) The prognostic value of Lnc-CAF/IL-33 was determined. (E) Knockdown of Lnc-CAF in CAFs and/or HSC3 cells and the xenograft mouse model of OSCC was established. Knocked down cells were subcutaneously injected in the rear flank of the nude mice (eight per group). The mean tumor size (mm3) was determined. (F) The micro-vessel density (MDV) was estimated by CD31 immunohistochemistry (200×); the expression of the proliferation index, Ki-67 (400×) and α-SMA (200×) in CAFs was analyzed. (G) The mRNA expression of Lnc-CAF and IL-33 was determined by Q-PCR. (H) Model summarizing the pro-tumorigenic effect of Lnc-CAF/IL-33 in the tumor microenvironment of OSCC. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Lnc-CAF knockdown impairs OSCC tumor growth To further examine the tumorigenic role of Lnc-CAF in vivo, we produced two cell lines, CAFs and HSC3 cells knocked down for Lnc-CAF by a lentiviral vector, RNAi-Lnc-CAF. Lnc-CAF knockdown had no effect on cell viability. Lnc-CAF knockdown was performed in CAFs to impair the pro-tumor characteristics of stromal fibroblasts, while it was performed in HSC3 cells to inhibit the tumor cell-derived stromal support (e.g. cytokines and exosomes). Overall, Lnc-CAF knockdown in these two cell types was performed to inhibit the role of Lnc-CAF in the TME. The results showed that co-culture of CAFs with HSC3 cells promoted tumor growth, while Lnc-CAF knockdown in CAF or HSC3 antagonized this pro-tumor effect. Additionally, Lnc-CAF knockdown in both CAFs and HSC3 cells maximized the antitumor effect (Figure 6E). Besides, the proliferation and angiogenesis indices, Ki-67 and CD31, respectively, were determined in vivo. Mice injected with both CAFs and HSC3 knocked down for Lnc-CAF showed the lowest expression of Ki-67, CD31 (Figure 6F), Lnc-CAF and IL-33 (Figure 6G), which could be explained by the significant reduction of α-SMA+ CAFs in the TME (Figure 6F). Thus, herein, we identified a novel LncRNA signal, Lnc-CAF/IL-33, in the stroma and tumor cells, which supports the growth of tumor cells via a positive feedback loop in the TME (Figure 6H). Discussion In this study, we examined lncRNA profiles during NF/CAF transformation and identified an uncharacterized intergenic lncRNA, Lnc-CAF, which was elevated in CAFs. Lnc-CAF increases the expression of α-SMA, vimentin and N-cadherin and promotes the activation of CAFs, leading to tumor proliferation, which predicted poor clinical outcome in patients with OSCC. In the co-expression networks obtained after RNA sequencing, Lnc-CAF was directly associated with upregulated IL-33 in CAFs. As a member of the IL-1 family, IL-33 was a cytokine ‘favoring’ immune responses with T helper type 2 (Th2) bias during allergic inflammation, parasitic infections and cancer (29). However, the role of IL-33 in carcinogenesis was unclear. Maywald et al. (30) reported that IL-33 expression was induced in the tumor epithelium of adenomas and carcinomas, leading to the activation of ST2+ myofibroblasts and master cells for the expression of extracellular matrix components and growth factors associated with polyposis and intestinal tumor progression. IL-33 is abundantly released in breast cancer tissues to reduce apoptosis and sustain the survival of myeloid-derived suppressor cells (MDSCs), which is associated with the activation of NF-kB and MAPK signaling for the induction of autocrine secretion of GM-CSF (31). In this study, IL-33 expression was upregulated to promote the activation phenotype of CAFs, resulting in CAF-supported tumor growth in OSCC. These findings are consistent with previous reports in head and neck squamous cell carcinoma, indicating that CAFs abundantly express IL-33 to induce EMT-mediated invasiveness of tumor cells (32). Different studies showed that IL-33 expression is regulated via different pathways in a cell type and stimulator–specific fashion. TLR3 and TLR4 agonists could significantly induce IL-33 expression in peritoneal macrophages via the cAMP-CREB pathway (33). In this study, upregulated Lnc-CAF promoted IL-33 stability by preventing IL-33 degradation through the p62-dependent autophagy–lysosome pathway, although it was not related to the lncRNA scaffold function. Exosomes are small membrane vesicles with a size of 50–100 nm that contain ncRNAs, mRNAs, and proteins. TME-derived exosomes can reprogram stromal cells and tumor cells (34). In chronic lymphocytic leukemia (CLL), CLL-derived exosomes contain miR-21 and -146a and can be actively incorporated by endothelial and mesenchymal stem cells, inducing an inflammatory phenotype of CAFs (35). Under nutrient deprivation or nutrient stressed conditions, CAF-derived exosomes contain amino acids, lipids, and TCA-cycle intermediates, which could be utilized by cancer cells for central carbon metabolism, promoting tumor growth (36). We observed that OSCC cells secreted exosomes containing Lnc-CAF to stromal fibroblasts and promoted the up-regulation of Lnc-CAF levels for the activation of CAFs, which formed a positive regulation loop to induce tumor proliferation. Thus, the expression of Lnc-CAF is not limited to CAF, but also tumor cells. In summary, in this study, we report a specific LncRNA profile during the activation of CAFs in OSCC and identified an uncharacterized intergenic lncRNA, Lnc-CAF, which was significantly elevated in CAFs and could stabilize and up-regulate cytokine IL-33 to reprogram CAFs, supporting tumor growth. While further studies on the interaction way between Lnc-CAF and IL-33 are need to be investigated. Supplementary material Supplementary material is available at Carcinogenesis online. Funding This work was supported by the National Natural Science Foundation of China (81402238, 81772880, 81271698, 31370899, 81702680) and the fundamental research funds for the central universities (021414380325), the Nanjing Medical Science and technique development foundation (YKK16164, QRX17083) and the Jiangsu Provincial key Medicine discipline (since 2017). Abbreviations CAFs carcinoma-related fibroblasts lncRNA long non-coding RNA NF normal fibroblast ncRNA non-coding RNA OSCC oral squamous cell carcinoma PBS phosphate-buffered solution TME tumor microenvironment Acknowledgements We thank Hospital of Stomatology, Nanjing for their support and for providing clinical samples. L.D., YH.N. and YY.H. designed experiments with valuable help from X.F.H., J.R. and Q.G.H. L.D. performed and analyzed data with valuable help from Y.L. and D.Y.Z., H.W. L.D. and YY.H wrote the manuscript. XF.H and L.D. collected surgical specimens. YY.H. and YH.N. oversaw the overall project. YX.S helped to revise the manuscript. Conflict of Interest Statement: None declared. References 1. Schmitt, A.M.et al.  ( 2016) Long noncoding RNAs in cancer pathways. Cancer Cell , 29, 452– 463. Google Scholar CrossRef Search ADS PubMed  2. Sun, W. et al. (2014) Sequencing the head and neck cancer genome: implications for therapy. Ann. N. Y. Acad. Sci., 1333, 33–42. 3. Huarte, M. ( 2015) The emerging role of lncRNAs in cancer. Nat. Med ., 21, 1253– 1261. Google Scholar CrossRef Search ADS PubMed  4. Liu, Z.et al.  ( 2017) Over-expressed long noncoding RNA HOXA11-AS promotes cell cycle progression and metastasis in gastric cancer. Mol. Cancer , 16, 82. Google Scholar CrossRef Search ADS PubMed  5. Ding, L.et al.  ( 2017) The TLR3 agonist inhibit drug efflux and sequentially consolidates low-dose cisplatin-based chemoimmunotherapy while reducing side effects. Mol. Cancer Ther ., 16, 1068– 1079. Google Scholar CrossRef Search ADS PubMed  6. Casey, S.C.et al.  ( 2015) Cancer prevention and therapy through the modulation of the tumor microenvironment. Semin. Cancer Biol ., 35, S199– S223. Google Scholar CrossRef Search ADS PubMed  7. Salo, T.et al.  ( 2014) Insights into the role of components of the tumor microenvironment in oral carcinoma call for new therapeutic approaches. Exp. Cell Res ., 325, 58– 64. Google Scholar CrossRef Search ADS PubMed  8. Mitra, A.K.et al.  ( 2012) MicroRNAs reprogram normal fibroblasts into cancer-associated fibroblasts in ovarian cancer. Cancer Discov ., 2, 1100– 1108. Google Scholar CrossRef Search ADS PubMed  9. Kidd, S.et al.  ( 2009) Direct evidence of mesenchymal stem cell tropism for tumor and wounding microenvironments using in vivo bioluminescent imaging. Stem Cells , 27, 2614– 2623. Google Scholar CrossRef Search ADS PubMed  10. Abe, R.et al.  ( 2001) Peripheral blood fibrocytes: differentiation pathway and migration to wound sites. J. Immunol ., 166, 7556– 7562. Google Scholar CrossRef Search ADS PubMed  11. Kidd, S.et al.  ( 2012) Origins of the tumor microenvironment: quantitative assessment of adipose-derived and bone marrow-derived stroma. PLoS One , 7, e30563. Google Scholar CrossRef Search ADS PubMed  12. Zeisberg, E.M.et al.  ( 2008) Fibroblasts in kidney fibrosis emerge via endothelial-to-mesenchymal transition. J. Am. Soc. Nephrol ., 19, 2282– 2287. Google Scholar CrossRef Search ADS PubMed  13. Zhao, L.et al.  ( 2017) Long noncoding RNA LINC00092 acts in cancer-associated fibroblasts to drive glycolysis and progression of ovarian cancer. Cancer Res ., 77, 1369– 1382. Google Scholar CrossRef Search ADS PubMed  14. Zhuang, J.et al.  ( 2015) TGFβ1 secreted by cancer-associated fibroblasts induces epithelial-mesenchymal transition of bladder cancer cells through lncRNA-ZEB2NAT. Sci. Rep ., 5, 11924. Google Scholar CrossRef Search ADS PubMed  15. Huang, J.K.et al.  ( 2017) LncRNA-MALAT1 promotes angiogenesis of thyroid cancer by modulating tumor-associated macrophage FGF2 protein secretion. J. Cell. Biochem ., 118, 4821– 4830. Google Scholar CrossRef Search ADS PubMed  16. Wu, M.H.et al.  ( 2011) Targeting galectin-1 in carcinoma-associated fibroblasts inhibits oral squamous cell carcinoma metastasis by downregulating MCP-1/CCL2 expression. Clin. Cancer Res ., 17, 1306– 1316. Google Scholar CrossRef Search ADS PubMed  17. Marsh, T.et al.  ( 2013) Fibroblasts as architects of cancer pathogenesis. Biochim. Biophys. Acta , 1832, 1070– 1078. Google Scholar CrossRef Search ADS PubMed  18. Liang, D.et al.  ( 2015) Activated STING enhances Tregs infiltration in the HPV-related carcinogenesis of tongue squamous cells via the c-jun/CCL22 signal. Biochim. Biophys. Acta , 1852, 2494– 2503. Google Scholar CrossRef Search ADS PubMed  19. Ni, Y.H.et al.  ( 2015) Distinct expression patterns of Toll-like receptor 7 in tumour cells and fibroblast-like cells in oral squamous cell carcinoma. Histopathology , 67, 730– 739. Google Scholar CrossRef Search ADS PubMed  20. Song, Y.et al.  ( 2017) Exosomal miR-146a contributes to the enhanced therapeutic efficacy of interleukin-1β-primed mesenchymal stem cells against sepsis. Stem Cells , 35, 1208– 1221. Google Scholar CrossRef Search ADS PubMed  21. Sponheim, J.et al.  ( 2010) Inflammatory bowel disease-associated interleukin-33 is preferentially expressed in ulceration-associated myofibroblasts. Am. J. Pathol ., 177, 2804– 2815. Google Scholar CrossRef Search ADS PubMed  22. Wang, M.et al.  ( 2015) Large-scale association analysis in Asians identifies new susceptibility loci for prostate cancer. Nat. Commun ., 6, 8469. Google Scholar CrossRef Search ADS PubMed  23. Lemire, M.et al.  ( 2015) A genome-wide association study for colorectal cancer identifies a risk locus in 14q23.1. Hum. Genet ., 134, 1249– 1262. Google Scholar CrossRef Search ADS PubMed  24. Micke, P.et al.  ( 2011) Gene copy number aberrations are associated with survival in histologic subgroups of non-small cell lung cancer. J. Thorac. Oncol ., 6, 1833– 1840. Google Scholar CrossRef Search ADS PubMed  25. Bockmühl, U.et al.  ( 2000) Genetic imbalances with impact on survival in head and neck cancer patients. Am. J. Pathol ., 157, 369– 375. Google Scholar CrossRef Search ADS PubMed  26. Ulitsky, I.et al.  ( 2013) lincRNAs: genomics, evolution, and mechanisms. Cell , 154, 26– 46. Google Scholar CrossRef Search ADS PubMed  27. Luga, V.et al.  ( 2013) Tumor-stroma interaction: revealing fibroblast-secreted exosomes as potent regulators of Wnt-planar cell polarity signaling in cancer metastasis. Cancer Res ., 73, 6843– 6847. Google Scholar CrossRef Search ADS PubMed  28. Webber, J.P.et al.  ( 2015) Differentiation of tumour-promoting stromal myofibroblasts by cancer exosomes. Oncogene , 34, 290– 302. Google Scholar CrossRef Search ADS PubMed  29. Martin, N.T.et al.  ( 2016) Interleukin 33 is a guardian of barriers and a local alarmin. Nat. Immunol ., 17, 122– 131. Google Scholar CrossRef Search ADS PubMed  30. Maywald, R.L.et al.  ( 2015) IL-33 activates tumor stroma to promote intestinal polyposis. Proc. Natl. Acad. Sci. USA , 112, E2487– E2496. Google Scholar CrossRef Search ADS   31. Xiao, P.et al.  ( 2016) Interleukin 33 in tumor microenvironment is crucial for the accumulation and function of myeloid-derived suppressor cells. Oncoimmunology , 5, e1063772. Google Scholar CrossRef Search ADS PubMed  32. Chen, S.F.et al.  ( 2013) The paracrine effect of cancer-associated fibroblast-induced interleukin-33 regulates the invasiveness of head and neck squamous cell carcinoma. J. Pathol ., 231, 180– 189. Google Scholar CrossRef Search ADS PubMed  33. Polumuri, S.K.et al.  ( 2012) Transcriptional regulation of murine IL-33 by TLR and non-TLR agonists. J. Immunol ., 189, 50– 60. Google Scholar CrossRef Search ADS PubMed  34. Yu, S.et al.  ( 2015) Tumor-derived exosomes in cancer progression and treatment failure. Oncotarget , 6, 37151– 37168. Google Scholar PubMed  35. Paggetti, J.et al.  ( 2015) Exosomes released by chronic lymphocytic leukemia cells induce the transition of stromal cells into cancer-associated fibroblasts. Blood , 126, 1106– 1117. Google Scholar CrossRef Search ADS PubMed  36. Zhao, H.et al.  ( 2016) Tumor microenvironment derived exosomes pleiotropically modulate cancer cell metabolism. Elife , 5, e10250. Google Scholar PubMed  © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Carcinogenesis Oxford University Press

A novel stromal lncRNA signature reprograms fibroblasts to promote the growth of oral squamous cell carcinoma via LncRNA-CAF/interleukin-33

Loading next page...
 
/lp/ou_press/a-novel-stromal-lncrna-signature-reprograms-fibroblasts-to-promote-the-h4ZGJqj0Pj
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.
ISSN
0143-3334
eISSN
1460-2180
D.O.I.
10.1093/carcin/bgy006
Publisher site
See Article on Publisher Site

Abstract

Abstract Stromal carcinoma-related fibroblasts (CAFs) are the main type of non-immune cells in the tumor microenvironment (TME). CAFs interact with cancer cells to promote tumor proliferation. Long non-coding RNAs (lncRNAs) are known to regulate cell growth, apoptosis and metastasis of cancer cells, but their role in stromal cells is unclear. Using RNA sequencing, we identified a stromal lncRNA signature during the transformation of CAFs from normal fibroblasts (NFs) in oral squamous cell carcinoma (OSCC). We uncovered an uncharacterized lncRNA, FLJ22447, which was remarkably up-regulated in CAFs, referred to LncRNA-CAF (Lnc-CAF) hereafter. Interleukin-33 (IL-33) was mainly located in the stroma and positively co-expressed with Lnc-CAF to elevate the expression of CAF markers (α-SMA, vimentin and N-cadherin) in fibroblasts. In a co-culture system, IL-33 knockdown impaired Lnc-CAF-mediated stromal fibroblast activation, leading to decreased proliferation of tumor cells. Mechanistically, Lnc-CAF up-regulated IL-33 levels and prevented p62-dependent autophagy–lysosome degradation of IL-33, which was independent of LncRNA-protein scaffold effects. Treatment with the autophagy inducer, rapamycin, impaired the proliferative effect of Lnc-CAF/IL-33 by promoting IL-33 degradation. In turn, tumor cells further increased Lnc-CAF levels in stromal fibroblasts via exosomal Lnc-CAF. In patients with OSCC, high Lnc-CAF/IL-33 expression correlated with high TNM stage (n = 140). Moreover, high Lnc-CAF expression predicted poor prognosis. In vivo, Lnc-CAF knockdown restricted tumor growth and was associated with decreased Ki-67 expression and α-SMA+ CAF in the stroma. In conclusion, we identified a stromal lncRNA signature, which reprograms NFs to CAFs via Lnc-CAF/IL-33 and promotes OSCC development. Introduction The majority of RNA species transcribed from the human genome are non-coding RNAs (ncRNAs), which lack the coding potential capability of producing functional small peptides. Two major classes of regulatory ncRNAs have been widely investigated, including microRNAs (<200 nt in length) and long ncRNAs (lncRNAs) (>200 nt in length) (1,2). They play crucial roles in diverse cellular processes from normal development to disease progression. Deregulation of lncRNAs is involved in carcinogenesis and resistance to therapy (3). In patients with gastric cancer, high levels of lncRNA-HOXA11-AS expression are associated with short survival and poor prognosis (4). We previously found that lncRNA-IL7R is a therapy-resistance factor in oral squamous cell carcinoma (OSCC), and its levels are elevated in response to chemoimmunotherapy (5). Currently, the tumor suppressor and oncogenic roles of lncRNAs are investigated mainly in cancer cells. However, cancer initiation and progression is a multistep process that involves reciprocal autocrine–paracrine communication (e.g. secreted factors and exosomes) between tumor cells and the adjacent stromal microenvironment (6,7). As the most important infiltrated non-immune cells in the stroma, cancer-associated fibroblasts (CAFs) stimulate tumor proliferation and metastasis in a paracrine manner (8). The CAF are mainly derived and activated from normal fibroblasts (NFs), and bone marrow-derived mesenchymal stem cells (9), circulating fibrocytes (10), tissue adipocytes (11) and endothelial cells (12) are also acquire fibroblast markers during carcinogenesis and other inflammatory environment. In ovarian cancer cells, LINC00092 is up-regulated in response to the CAF-derived chemokine, CXCL14 and drives glycolysis and progression of ovarian cancer (13). Similarly, CAFs produce high levels of TGFβ1, which induces the expression of lncRNA-ZEB2NAT in bladder cancer cells, supporting EMT and invasion of cancer cells (14). Evidence regarding lncRNAs in tumor stroma cells is limited. LncRNA-MALAT1 in tumor-associated macrophage increases FGF2 protein secretion, inhibiting inflammatory cytokine release and promoting the proliferation and metastasis of thyroid cancer (15). However, the function of lncRNAs in CAF remains unclear. As an important source of CAFs, NFs are re-educated by the tumor microenvironment (TME) and acquire the characteristics of pro-tumorigenic CAFs (16,17). Therefore, uncovering the mechanisms underlying CAF transformation, re-orchestrating the tumor cells and stromal microenvironment, is needed. In this study, to uncover the LncRNA signature in stroma fibroblasts, we isolated and cultured NFs/CAFs derived from normal and matched tumor tissues from patients with OSCC and examined lncRNAs profiles during the NF/CAF transformation by RNA sequencing. The co-expression networks were analyzed to identify core lncRNA/mRNA interaction, and examined its oncogene or anti-tumor role in vitro and in vivo. Materials and methods Patients and tissue samples Ten paired non-tumor and tumor tissues were used to isolate primary NFs and CAFs from patients with OSCC, respectively. For RNA sequence, five tumors derived from the tongue, gums and cheeks were included in this study during 2015–2016. The patients were 48–66 years old with T (1,2), N(0), M(0) stage, without lymph node metastasis and underwent no preoperative chemotherapy and/or radiotherapy. All the five patients were alive and had no recurrence during this experiment. For the analysis of Lnc-CAF and IL-33 expression, we constructed a cohort including 140 fresh primary OSCC tissue samples. All patients diagnosed with primary OSCC were confirmed by hematoxylin and eosin staining by experienced pathologists from the Department of Pathology at Nanjing Stomatology Hospital. OSCC tissues were evaluated according to the WHO classification and International Cancer Control (UICC) tumor–node–metastasis (TNM) staging system. Ethical approval for this study was obtained from the Research Ethics Committee of Nanjing Stomatology Hospital. Patients who were diagnosed with autoimmune or other malignant diseases as well as pregnant or lactating individuals were excluded from this study. No patient underwent preoperative chemotherapy and/or radiotherapy. Twenty-one patients were lost to follow-up and 119 patients were followed-up until 8 may 2017. Cell lines, mice and reagents The human OSCC cell line, HSC-3, was obtained from Professor Yvonne L. Kapila (Michigan University, MI). The cells were characterized by mycoplasma detection, DNA fingerprinting, isozyme detection and cell viability by the provider. No further authentication of the cell line was conducted. The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS (Life Technologies, Carlsbad, CA) at 37°C, 5% CO2 condition. Four- to five-week-old male BALB/c-nu/nu T cell-deficient mice were purchased from Cavens (Changzhou, China). siRNA-Lnc-CAF/IL-33, pcDNA3.1-Lnc-CAF/IL-33 or negative control were purchased from RiboBio (Guangzhou, China). Detailed information on the antibodies and reagents used in this study is provided in Supplementary Table 1, available at Carcinogenesis Online. Isolation of cancer-associated fibroblasts and RNA sequence Sterile fresh OSCC tissues and its corresponding normal tissues was collected from surgery and washed with phosphate-buffered solution (PBS) and antibiotics, and then removed the epithelial and adipose tissues. The specimens were cut into small pieces and digested by enzyme mix (Collagenase, Neutral protease, Hyaluronidase) for 30 min. The remaining small tissues were placed in synthetic DMEM/F12 basic medium and incubated at 37°C. The medium was replaced every 2–3 days and the epithelial cells were easily removed via trypsinase. The remaining cells were fibroblasts and were collected by further digestion with trypsinase for 2 min. We performed the RNA-seq with the help of Novel Bioinformatics Co., Ltd (Shanghai, China). Firstly, total RNA was extracted by Trizol reagent (Invitrogen) separately. The RNA quality was checked by Bioanalyzer 2200 (Aligent) and kept at −80°C. The RNA with RIN >8.0 is right for cDNA library construction. Secondly, the complementary DNA (cDNA) libraries for single-end sequencing were prepared using Ion Total RNA-Seq Kit v2.0 (Life Technologies). The cDNA libraries were then processed for the Proton Sequencing process according to the commercially available protocols. Thirdly, mapping of single-end reads. The clean reads were then aligned to human genome (version: GRCh38.p1) using the MapSplice program (v2.2.0). Finally, pathway analysis was performed and we also applied EBseq algorithm to filter the differentially expressed genes, the significant analysis and false discovery rate analysis were performed. Besides, we presented gene co-expression Networks to find the relations among different mRNA and LncRNA (Supplementary Figure 1 and Table 2, available at Carcinogenesis Online). The bioinformatics procedures and algorithms were included in Supplementary Materials and Methods, available at Carcinogenesis Online. Immunohistochemistry and immunofluorescence assays Immunohistochemistry and immunofluorescence assays were performed as previously described (18,19). Briefly, the fibroblasts, HSC3 or the specimen frozen sections were collected and covered with a depth of 2–3 mm with 4% formaldehyde for 15 min and washed with PBS. Then cells were permeabilized with 100% methanol for 10 min at −20°C and blocked with 3% bovine serum albumin and were incubated with primary antibodies (dilution: 1:100 or 200) overnight at 4°C. After rinsing three times in PBS, incubated coverslips in fluorochrome-conjugated secondary antibody (Dilution: 1:400) for 1–2 h at room temperature in dark and then stained with DAPI (Bioword, China). The FISH assay was performed by the manufacturer’s protocol (RiboBio, Guangzhou, China), U6 and 18S were used for cytoplasmic and nuclear positive control. Finally, the coverslips were mounted onto the glass slides with neutral gum and observed by FV10i confocal microscope (OLYMPUS, Japan). Immunohistochemistry Protein expression was analyzed immunohistochemically on 2-μm-thick, formalin-fixed and paraffin-embedded specimen sections. Slides were incubated in three washes of xylene for 5 min each and were followed by two washes of 100% ethanol for 10 min, 95% ethanol for 10 min and ddH2O for 5 min each. Antigen unmasking was prepared by boiling in pH 9.0, 10 mM Tris/1 mM EDTA, blocked with 3% hydrogen peroxide for 10 min at room temperature and washed. Then anti-IL-33 antibody (diluted ×100), anti-Ki-67 antibody (diluted ×400) anti-α-SMA antibody (diluted ×200) and anti-CD31 antibody (diluted ×200) were incubated the FFPE specimen sections at 4°C overnight and then the EnVision Detection System kit (DAKO, Denmark) was used for the DAB chromogen followed by nuclear staining using hematoxylin. Neutral gum was used to cover the sliders and dry at room temperature for counting. RNA isolation and qRT-PCR Total RNA was extracted using Trizol reagent (Invitrogen) according to the standard RNA isolation protocol. Quantitative real-time RT-PCR (qRT-PCR) was performed, and cDNA synthesis was performed using a PrimeScript RT Reagent Kit (Takara, China). The PCR amplification was performed with the conditions of 95°C for 10 s, 40 cycles of 94°C for 30 s, 60°C for 30 s and 72°C for 30 s on an ABI 7900 system (Applied Biosystems, USA) with SYBR green real-time PCR Master Mix (Takara, China). The relative levels of genes were calculated by the 2−ΔΔCt method. The expression levels of α-SMA, FSP-1, IL-33, Ki-67 and Lnc-CAF were normalized to GAPDH for gene expression analysis. Each experiment was performed in triplicate. Transwell assay About 2 × 105 NFs/CAFs cells were in the lower chamber and tumor cells were plated in the upper chamber of a non-coated transwell insert with 0.45 nm aperture. After the indirect co-culture for 48 h and cells were stained by crystal violet and counted under an inverted microscope. Five random views were selected to count the cells. Cell treatment and transfection The degradation of IL-33 protein in CAFs was induced by the protein synthesis inhibitor, cycloheximide (CHX) (50 μg/ml). The pathways of IL-33 degradation were analyzed by the autophagy–lysosome inhibitors, chloroquine (CQ) (25 μM) and 3-methyladenine (3-MA) (2 mM) and a proteasome inhibitor, MG132 (5 μM) before the knockdown of Lnc-CAF. The neutralizing antibody for IL-33 was used at 1 μg/ml. The recombinant human IL-33 was used in 10 and 50 ng/ml. According to the manufacturer’s instructions, the NF/CAFs were seeded into 12 or 6-well plate and then cells were transfected with siRNA-Lnc-CAF/IL-33, pcDNA3.1-Lnc-CAF/IL-33 or negative control by Lipofectamine 2000 (Invitrogen). After transfection for the indicated time, the cells were harvested for further experiments. Preparation of cell extracts and immunoblotting For immunoblotting analysis, cells were collected with PBS and then lysed in RIPA buffer with protease inhibitors and phosphatase inhibitors (Roche Applied Science, Basel, Switzerland). Detailed procedures for immunoblotting were previously described (20,21). Flow cytometry assay For the identification of pan-CK-positive HSC3 cells, FITC-anti-human pan-CK (Cat No.ab78478, Abcam, Cambridge, MA) was used to stain HSC3 cells. For the identification of CD63+ exosomes, PE-anti-human CD63 (Cat No. 353003, Biolegend, San Diego, CA) was used according to the manufacturer’s instructions and quantified by flow cytometry on a FACS Calibur instrument. Immunoprecipitation (IP) assay CAFs were transfected with siRNA-Lnc-CAF or negative control following the manufacturer’s protocol. The cells were then harvested and crosslinked with 1% formaldehyde at room temperature. After sonication and removal of cell debris, 2 mg anti-p62 mAb (Cat No.18420, Proteintech, Rosemont, IL) was added together with 10 ml Protein G Sepharose slurry and incubated overnight at 4°C. IL-33 immunoprecipitates (IPs) were washed three times. The proteins were then released from the beads by heating with Laemmli buffer, followed by SDS-PAGE and western blot analysis. Precipitated proteins were detected with anti-p62 (Cat No.18420, Proteintech) and anti–IL-33 mAb Nessy-1 (Cat No. ALX-804–840, Enzo), respectively. RNA binding protein immunoprecipitation The RNA immunoprecipitation assay (RIP) assay was performed by using the Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (Millipore, Billerica, MA) according to the manufacturer’s instructions. Briefly, cells were harvested and lysed in RIP lysis buffer. RNA was immunoprecipitated with an antibody against IL-33 and protein A/G magnetic beads. The magnetic bead bound complexes were immobilized with a magnet and unbound materials were washed off. Then, RNAs of Lnc-CAF were extracted and analyzed by qRT-PCR. Exosome isolation and characterization Exosomes from the culture supernatants were isolated by differential centrifugation as we previously described (20,21). Tumor mouse model To investigate the role of Lnc-CAF in the TME of OSCC, CAFs or HSC3 cells were infected with lentivirus-RNAi-Lnc-CAF (MOI = 100) or negative control. The two types of cells infected with lentivirus-RNAi-control were used as control groups in this study. The cells infected with lentivirus-RNAi-Lnc-CAF were used as test groups. The HSC3 cell lines used alone is the negative control group in this study. These conditional cells were subcutaneously co-injected in the rear flank of nude mice (eight per group) as indicated description. The animal study was approved by institutional animal research committee of Nanjing University and that animals were cared for following the guidelines for use and care of laboratory animals. Statistical analyses Statistical analyses were performed on Statistical package for social sciences version 16.0 (SPSS 16.0, SPSS Inc., Chicago, IL) and Prism statistical software package (GraphPad Software Inc., San Diego, CA). The relationships between the presence of Lnc-CAF/IL-33 and clinicopathological characteristics were determined by unpaired t-test. Unpaired t-test or Mann–Whitney U test was used to compare two groups and the differences between more than two groups were analyzed by the Kruskal–Wallis test. Differences were considered statistically significant with P < 0.05. Results Characteristics of NFs and CAFs in the OSCC microenvironment To determine LncRNA profiles during the transformation of stromal fibroblasts, we first collected 10 paired non-tumor and tumor tissues from patients with OSCC and fibroblasts were isolated and cultured. To confirm the identity of CAFs, the phenotype and function of the paired NFs/CAFs were evaluated. CAFs as stromal mesenchymal cells are often identified by the expression of αSMA and other markers, including fibroblast specific protein (FSP-1) (17). The expression of CAF markers, α-SMA and FSP-1, were higher in CAFs than those in NFs (Figure 1A). OSCC cells were derived from the epithelium and used as control cells. We further analyzed the expression of cytokeratin, a specific marker for epithelial cells. NFs and CAFs were negative for cytokeratin, but the OSCC cell line, HSC3, showed strong positive staining (Figure 1A), indicating that we successfully isolated tissue-derived stromal fibroblasts. Furthermore, α-SMA and FSP-1 mRNA levels were also upregulated in CAFs (Figure 1B). Evidence demonstrated that CAFs promote tumor progression by regulating proliferation, migration, and therapy-induced apoptosis. Our previous findings showed that CAFs derived from patients with OSCC inhibited chemoimmunotherapy-induced cell death (5). Here, we confirmed that, when compared with HSC3 cells alone, CAFs stimulated tumor proliferation in vitro, although NFs also showed a very weak ability to induce tumor cell growth (Figure 1C). After the phenotype and function validation of paired NFs/CAFs, five tumors derived from the tongue, gums, and cheeks were included in this study during 2015–2016 were selected for RNA-sequence. The patients were 48–66 years old with T (1,2), N(0), M(0) stage, without lymph node metastasis and underwent no preoperative chemotherapy and/or radiotherapy. All the five patients were alive and had no recurrence during this experiment. Figure 1. View largeDownload slide Characteristics of normal fibroblasts (NFs) and carcinoma-related fibroblasts (CAFs) in oral squamous cell carcinoma (OSCC). (A) Three paired NFs and CAFs were isolated from three independent patients with OSCC and cultured. Their phenotype was determined by staining for the CAF markers, α-SMA (Green) and FSP-1 (Red), via immunofluorescence. The OSCC cell line, HSC3, was used as control cells and fibroblasts were negative for the epithelial cells marker, pan-cytokeratin (Red). (B) α-SMA and FSP-1 mRNA expression was determined by Q-PCR. (C) The transwell assay was performed to estimate the proliferation of HSC3 cells induced by NFs and CAFs. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Figure 1. View largeDownload slide Characteristics of normal fibroblasts (NFs) and carcinoma-related fibroblasts (CAFs) in oral squamous cell carcinoma (OSCC). (A) Three paired NFs and CAFs were isolated from three independent patients with OSCC and cultured. Their phenotype was determined by staining for the CAF markers, α-SMA (Green) and FSP-1 (Red), via immunofluorescence. The OSCC cell line, HSC3, was used as control cells and fibroblasts were negative for the epithelial cells marker, pan-cytokeratin (Red). (B) α-SMA and FSP-1 mRNA expression was determined by Q-PCR. (C) The transwell assay was performed to estimate the proliferation of HSC3 cells induced by NFs and CAFs. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Lnc-CAF/IL-33 is up-regulated during NF/CAF transformation The results of RNA sequencing indicated that the mRNA/lncRNA profiles significantly changed during NF/CAF transformation. Differentially expressed mRNAs and lncRNAs with a fold change above 2 and a false discovery rate less than 0.05 were selected for co-expression network analysis. An uncharacterized intergenic lncRNA LOC400221, also known as FLJ22447 (NCBI Gene ID: 400221) was upregulated in CAFs (referred to as Lnc-CAF) (Figure 2A). Lnc-CAF is located in 14q23.1-q23.2 which is reported to harbor cancer risk locus and associated with shorter survival of patients in prostate cancer (22), colorectal cancer (23), lung cancer (24) and head and neck squamous cell carcinomas (25). More importantly, the analysis of co-expression networks indicated that Lnc-CAF was located in the core position of differentially expressed genes. And according to its level in fibroblasts and its fold-change, Lnc-CAF is highly expressed in fibroblasts (Supplementary Table 2, available at Carcinogenesis Online). Therefore, Lnc-CAF was chosen, and we speculated that it was involved in the NF/CAF transformation (Supplementary Figure 1A, available at Carcinogenesis Online). Meanwhile, the five paired NFs and CAFs isolated from five independent patients with OSCC were cultured for validation of five differentially expressed genes (ASMA, YKL-40, IL-33, LOC100506114 and Lnc-CAF) by real-time PCR (Supplementary Figure 1B, available at Carcinogenesis Online). Figure 2. View largeDownload slide Lnc-CAF and IL-33 are up-regulated in activated CAFs. (A) The gene expression between CAFs and NFs (vertical axis) and average expression of genes in CAFs versus those in NFs (horizontal axis) are presented as a Bland-Altman plot. Data are from our RNA-seq analysis. Highlighted in red are 35 lncRNAs with significant changes in expression (fold > 2, false discovery rate < 0.05). (B) The cytoplasmic and nuclear fractions of CAFs were prepared. After RNA extraction, the levels of Lnc-CAF, U6 and 18S transcripts were determined by Q-PCR. (C) The expression pattern of Lnc-CAF in NFs, CAFs, and OSCC tumor samples was determined by FISH and immunofluorescence. (D) The co-expression network of Lnc-CAF and IL-33 was analyzed. (E) The expression of IL-33 in the tumor nest and stroma of OSCC samples was determined by immunohistochemistry, T, tumor nest. S, stroma. ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Figure 2. View largeDownload slide Lnc-CAF and IL-33 are up-regulated in activated CAFs. (A) The gene expression between CAFs and NFs (vertical axis) and average expression of genes in CAFs versus those in NFs (horizontal axis) are presented as a Bland-Altman plot. Data are from our RNA-seq analysis. Highlighted in red are 35 lncRNAs with significant changes in expression (fold > 2, false discovery rate < 0.05). (B) The cytoplasmic and nuclear fractions of CAFs were prepared. After RNA extraction, the levels of Lnc-CAF, U6 and 18S transcripts were determined by Q-PCR. (C) The expression pattern of Lnc-CAF in NFs, CAFs, and OSCC tumor samples was determined by FISH and immunofluorescence. (D) The co-expression network of Lnc-CAF and IL-33 was analyzed. (E) The expression of IL-33 in the tumor nest and stroma of OSCC samples was determined by immunohistochemistry, T, tumor nest. S, stroma. ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. To determine the subcellular location of Lnc-CAF, the cytoplasmic and nuclear fractions were isolated. Lnc-CAF was detected in both the cytoplasm and nucleus in CAFs (Figure 2B). FISH of Lnc-CAF was performed in NFs, CAFs and OSCC samples. Clinical evidence implicated that both tumor cells and stromal fibroblasts expressed Lnc-CAF, but the expression of Lnc-CAF seemed to be higher in tumor nests (pan-CK-positive) than in CAFs (α-SMA-positive) (Figure 2C). Interestingly, as an important cytokine involved in inflammation and cancer, IL-33 was associated with Lnc-CAF in co-expression networks and was up-regulated in CAFs (Figure 2D). In clinical samples, IL-33 was mainly expressed in the tumor stroma, but rarely in the tumor nest, which indicated that IL-33 is indispensable for the stromal microenvironment (Figure 2E). These findings indicated that enhanced stromal Lnc-CAF/IL-33 signals were associated with a CAF-dependent TME in clinical specimens. Upregulated Lnc-CAF promotes the CAF phenotype via IL-33, supporting tumor cell proliferation We next analyzed the function of Lnc-CAF/IL-33 signals in fibroblasts. Lnc-CAF was silenced by siRNA in CAFs and overexpressed in NFs. Two siRNAs were designed and siRNA-02 was chosen for the subsequent experiments (Figure 3A). Effective knockdown of Lnc-CAF in CAFs (Figure 3B) inhibited the expression of α-SMA, vimentin and N-cadherin, indicating that Lnc-CAF promoted the mesenchymal phenotype of CAFs (Figure 3C). Besides, forced expression of Lnc-CAF in NFs induced the expression of α-SMA, vimentin, and N-cadherin and enhanced the mesenchymal phenotype of CAFs (Figure 3D and E). The results indicated that Lnc-CAF is required for the maintenance of the stromal phenotype of CAF. Figure 3. View largeDownload slide Lnc-CAF activates CAF phenotype via IL-33. (A) Two siRNAs against Lnc-CAF were transfected into CAFs to select the most efficient one. (B, C) Different concentrations of siRNA-LncRNA-CAF-002 were transfected into CAFs and the expression of LncRNA-CAF, α-SMA, vimentin, and N-cadherin was determined by Q-PCR or western blotting. (D, E) Overexpression of Lnc-CAF in NFs via transfection of pcDNA 3.1-Lnc-CAF and the CAF phenotype was analyzed by western blotting. (F) The correlation between the level of Lnc-CAF and IL-33 was analyzed in eight paired NFs and CAFs. (G) After overexpression of Lnc-CAF in NFs, the siRNA-IL-33 was transfected and the CAF phenotype was analyzed. (H, I) NFs were treated with recombinant human IL-33 at the indicated dose. The expression levels of Lnc-CAF, α-SMA, vimentin, and N-cadherin were determined. (J) NFs were transfected with pcDNA 3.1-Lnc-CAF and/or siRNA-IL-33, and the transfected NFs were directly co-cultured with HSC3 cells. The pan-CK-positive cells were referred to as HSC3 cells and the proliferation of HSC3 cells was determined by pan-CK staining by flow cytometry. HSC3 cells alone were used as positive control and the IgG-isotype antibody was used as a negative control. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Figure 3. View largeDownload slide Lnc-CAF activates CAF phenotype via IL-33. (A) Two siRNAs against Lnc-CAF were transfected into CAFs to select the most efficient one. (B, C) Different concentrations of siRNA-LncRNA-CAF-002 were transfected into CAFs and the expression of LncRNA-CAF, α-SMA, vimentin, and N-cadherin was determined by Q-PCR or western blotting. (D, E) Overexpression of Lnc-CAF in NFs via transfection of pcDNA 3.1-Lnc-CAF and the CAF phenotype was analyzed by western blotting. (F) The correlation between the level of Lnc-CAF and IL-33 was analyzed in eight paired NFs and CAFs. (G) After overexpression of Lnc-CAF in NFs, the siRNA-IL-33 was transfected and the CAF phenotype was analyzed. (H, I) NFs were treated with recombinant human IL-33 at the indicated dose. The expression levels of Lnc-CAF, α-SMA, vimentin, and N-cadherin were determined. (J) NFs were transfected with pcDNA 3.1-Lnc-CAF and/or siRNA-IL-33, and the transfected NFs were directly co-cultured with HSC3 cells. The pan-CK-positive cells were referred to as HSC3 cells and the proliferation of HSC3 cells was determined by pan-CK staining by flow cytometry. HSC3 cells alone were used as positive control and the IgG-isotype antibody was used as a negative control. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. IL-33 level was positively correlated with Lnc-CAF level in eight paired NFs and CAFs (Figure 3F). Thus, we next determined whether IL-33 was indispensable for Lnc-CAF-induced phenotype of CAFs. The results showed that, although Lnc-CAF activated the expression of α-SMA, vimentin and N-cadherin, IL-33 knockdown reversed the stromal phenotype regulated by Lnc-CAF, suggesting that Lnc-CAF maintains the stromal phenotype of CAFs via IL-33 (Figure 3G). Given that IL-33 receptor (IL-33R or ST2) was first identified on the surface of fibroblasts and mast cells, we speculated an autocrine pathway of IL-33/ST2 in fibroblasts for IL-33-dependent CAF phenotype. Therefore, NFs were treated with recombinant human IL-33 for 24 h, which led to increased Lnc-CAF expression in NFs (Figure 3H) as well as increased expression of α-SMA, vimentin and N-cadherin (Figure 3I). Initial signaling complex of MyD88-IRAK4-TRAF6 signals are activated by the formation of the IL-33 receptor complex. Here, IL-33 induced the phosphorylation of IRAK4, which was accompanied by enhanced CAF phenotype (Figure 3I). These data indicated that IL-33 uses multiple pathways to regulate the activation of the CAF phenotype. Furthermore, we examine the impacts of Lnc-CAF/IL-33 signaling in the stroma on tumor cells. We first overexpressed Lnc-CAF and/or siRNA-IL-33 in cultured NFs, which were then directly co-cultured with HSC3 cells. HSC3 cells alone were used as positive controls. NFs overexpressing Lnc-CAF enhanced the proliferation of HSC3 cells (pan-CK+ cells: 65%) when compared with normal NFs (pan-CK+ cells: 46%), but this effect was attenuated by IL-33 knockdown (pan-CK+ cells: 48%) (Figure 3J). These findings suggest that up-regulated Lnc-CAF/IL-33 signaling in stromal fibroblasts enhances the proliferation of HSC3 cells. Lnc-CAF prevents the degradation of IL-33 by the p62-dependent autophagy–lysosome pathway, supporting tumor growth To identify the mechanism by which high abundance of Lnc-CAF/IL-33 was maintained in CAFs, we knocked down Lnc-CAF in CAFs. Lnc-CAF knockdown in CAFs led to reduced expression of IL-33. In cycloheximide-chase assays, the protein synthesis inhibitor, cycloheximide (CHX), could further accelerate the degradation of IL-33 protein when Lnc-CAF was knocked down (Figure 4A). This result suggested that Lnc-CAF was involved in the maintenance of IL-33 protein stability. Because lncRNAs can act as decoys, scaffolds, guides or enhancers (26), we speculated that Lnc-CAF functioned as a lncRNA scaffold to maintain the stability of IL-33 protein and to inhibit its degradation. Figure 4. View largeDownload slide Lnc-CAF improves IL-33 stability via inhibition of p62-dependent autophagy–lysosome degradation. (A) The degradation of IL-33 protein in CAFs was induced by the protein synthesis inhibitor, cycloheximide (CHX) (50 μg/ml) for the indicated time, the level of IL-33 was normalized to that of GAPDH. (B) The autophagy–lysosome inhibitors, chloroquine (CQ) (25 μM) and 3-methyladenine (3-MA) (2 mM), and a proteasome inhibitor, MG132 (5 μM), were added to CAFs knocked down for Lnc-CAF to inhibit the degradation of IL-33. (C) The autophagy level in CAFs knocked down for Lnc-CAF was analyzed by western blot. (D) The IL-33/p62 complex was pulldown by an anti-p62 antibody and the IL-33 level was determined. (E) RNA immunoprecipitation (RIP) assay was performed to determine the direct interaction between Lnc-CAF and IL-33. An Anti-IL-33 antibody was used to pulldown the RNA/protein complex and the IgG isotype antibody was used as a negative control. (F) NFs were transfected with pcDNA 3.1-IL-33 and/or treated with rapamycin, and directly co-cultured with HSC3 cells. The proliferation of HSC3 cells was analyzed via pan-CK staining by flow cytometry. The IgG-isotype antibody was used as a negative control. *P < 0.05, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Figure 4. View largeDownload slide Lnc-CAF improves IL-33 stability via inhibition of p62-dependent autophagy–lysosome degradation. (A) The degradation of IL-33 protein in CAFs was induced by the protein synthesis inhibitor, cycloheximide (CHX) (50 μg/ml) for the indicated time, the level of IL-33 was normalized to that of GAPDH. (B) The autophagy–lysosome inhibitors, chloroquine (CQ) (25 μM) and 3-methyladenine (3-MA) (2 mM), and a proteasome inhibitor, MG132 (5 μM), were added to CAFs knocked down for Lnc-CAF to inhibit the degradation of IL-33. (C) The autophagy level in CAFs knocked down for Lnc-CAF was analyzed by western blot. (D) The IL-33/p62 complex was pulldown by an anti-p62 antibody and the IL-33 level was determined. (E) RNA immunoprecipitation (RIP) assay was performed to determine the direct interaction between Lnc-CAF and IL-33. An Anti-IL-33 antibody was used to pulldown the RNA/protein complex and the IgG isotype antibody was used as a negative control. (F) NFs were transfected with pcDNA 3.1-IL-33 and/or treated with rapamycin, and directly co-cultured with HSC3 cells. The proliferation of HSC3 cells was analyzed via pan-CK staining by flow cytometry. The IgG-isotype antibody was used as a negative control. *P < 0.05, ***P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. CAFs were transfected with siRNA-Lnc-CAF, followed by treatment with various inhibitors of protein degradation pathways to investigate the mechanisms underlying the regulation of IL-33 stability by Lnc-CAF. The autophagy–lysosome inhibitors, chloroquine (CQ) and 3-methyladenine (3-MA), but not MG132, a proteasome inhibitor, remarkably inhibited the degradation of IL-33 in CAFs knocked down for Lnc-CAF, suggesting that Lnc-CAF deficiency resulted in the degradation of IL-33 via the autophagy–lysosome pathway (Figure 4B). Interestingly, we observed a significant increase of the LC3II/LC3I ratio in CAFs knocked down for Lnc-CAF, implying that Lnc-CAF deficiency induced the initiation of autophagy (Figure 4C). p62 is a cargo protein involved in the protein degradation via selective autophagy. After Lnc-CAF knockdown, the interaction between p62 and IL-33 was increased, leading to the degradation of IL-33 via selective autophagy (Figure 4D). To further provide direct evidence of the role of Lnc-CAF as a lncRNA scaffold, RIP assay showed that the Lnc-CAF level in the co-immunoprecipitated complex with the anti-IL-33 antibody was not significantly different from that observed with the IgG isotype antibody, suggesting that Lnc-CAF-induced stability of IL-33 was not associated with its lncRNA scaffold function (Figure 4E). In the NF/HSC3 cell co-culture system, overexpression of IL-33 in NFs promoted the proliferation of tumor cells (pan-CK+ cells: 75%), but the induction of autophagy by rapamycin impaired this proliferative effect by promoting IL-33 degradation (pan-CK+ cells: 47%) (Figure 4F). Tumor cell-derived exosomal Lnc-CAF increases Lnc-CAF levels in stromal fibroblasts in a positive feed-back loop During the carcinogenesis or therapy resistance, secreted factors and exosomes in the stroma are essential mediators for the reciprocal autocrine–paracrine communication between tumor cells and the adjacent stromal microenvironment (27,28). We here investigated the impacts of OSCC-derived exosomes on stromal fibroblasts. The HSC3 cells were treated with or without the exosome secretion blocker, GW4869, and labeled with CM-Dil (red) for NF/HSC3 co-culture. CM-Dil was used to label the membranes and lipids (Figure 5A). We observed an increase in Dil+ NFs when NFs were co-cultured with normal HSC3 cells, but the GW4869-treated HSC3 cells secreted less exosomes, which resulted in decreased Dil+ NFs (Figure 5B). These data indicated that OSCC-derived NFs could effectively ingest tumor cell-derived exosomes. Thus, NFs were cultured in the conditional tumor cell-derived culture medium (CM). The results showed that Lnc-CAF/IL-33 expression was induced by the conditional tumor cell-derived CM, but not by CM from cells treated with GW4869 (Figure 5C). IL-33 promotes the activation of CAFs. An IL-33 neutralizing antibody also inhibited CM-induced Lnc-CAF/IL-33 expression in NFs (Figure 5C), suggesting that both the exosomes and IL-33 in tumor cell-derived CM contributed to the elevated Lnc-CAF /IL-33 expression in NFs. Figure 5. View largeDownload slide Tumor cells promote the upregulation of Lnc-CAF in NFs in a positive feedback loop. (A, B) HSC3 cells were pre-treated with or without the exosome secretion blocker, GW4869 and labeled with CM-Dil (red). NFs were directly co-cultured with labeled HSC3 cells for 18 h and the CM-Dil positive cells were analyzed. The ratio of CM-Dil+/pan-CK- NFs was calculated. (C) NFs were treated with the conditioned tumor cell culture medium for 48 h and the expression of Lnc-CAF and IL-33 was assessed by Q-PCR. The neutralizing antibody for IL-33 was used at 1 μg/mL. (D, E) Representative images of HSC3-derived exosomes are shown. Tracking analysis using qNano showed that the average size was approximately 95 nm and the CD63 and Alix expression in exosomes derived from HSC3 cells was estimated by western blotting. (F) The expression of Lnc-CAF in exosomes was analyzed and normalized to that of GAPDH. (G, H) The exosomes were labeled with PKH67 (green) or anti-CD63 antibody and added to NFs (5 μg/ml), and the level of Lnc-CAF in NFs was estimated by Q-PCR. *P < 0.05, **P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. Figure 5. View largeDownload slide Tumor cells promote the upregulation of Lnc-CAF in NFs in a positive feedback loop. (A, B) HSC3 cells were pre-treated with or without the exosome secretion blocker, GW4869 and labeled with CM-Dil (red). NFs were directly co-cultured with labeled HSC3 cells for 18 h and the CM-Dil positive cells were analyzed. The ratio of CM-Dil+/pan-CK- NFs was calculated. (C) NFs were treated with the conditioned tumor cell culture medium for 48 h and the expression of Lnc-CAF and IL-33 was assessed by Q-PCR. The neutralizing antibody for IL-33 was used at 1 μg/mL. (D, E) Representative images of HSC3-derived exosomes are shown. Tracking analysis using qNano showed that the average size was approximately 95 nm and the CD63 and Alix expression in exosomes derived from HSC3 cells was estimated by western blotting. (F) The expression of Lnc-CAF in exosomes was analyzed and normalized to that of GAPDH. (G, H) The exosomes were labeled with PKH67 (green) or anti-CD63 antibody and added to NFs (5 μg/ml), and the level of Lnc-CAF in NFs was estimated by Q-PCR. *P < 0.05, **P < 0.001 by unpaired t-test. Representative data from three independent experiments are shown. To confirm the roles of tumor-derived exosomal Lnc-CAF in the stroma, the HSC3 cell-derived exosomes were isolated and their size and phenotype were determined (Figure 5D and E). High levels of Lnc-CAF were detected in exosomes (Figure 5F). PKH67 (green) was used to label exosomes. PKH67+ exosomes could be taken up by NFs and up-regulated the expression of Lnc-CAF in NFs (Figure 5G and H). This effect was also observed in CAFs (unpublished data), indicating that tumor cells could secrete exosomal Lnc-CAF into the stroma and induce Lnc-CAF expression in stromal fibroblasts, leading to an Lnc-CAF-mediated positive feed-back loop between NFs/CAFs and tumor cells. High Lnc-CAF/IL-33 levels correlate with high tumor stage in patients with OSCC Considering the high level of Lnc-CAF/IL-33 in the TME, we further analyzed the correlation between aberrant Lnc-CAF/IL-33 expression and the clinicopathological characteristics of patients with OSCC (n = 140). Patients with high TNM stage, but not with LNM, often harbored high expression levels of Lnc-CAF and IL-33 (Figure 6A and B). Importantly, the expression of the proliferation index, Ki-67, was higher in patients with higher expression of Lnc-CAF/IL-33, which confirmed that the enhanced Lnc-CAF/IL-33 signature facilitated the growth of OSCC in vivo. Additionally, the prognostic values of Lnc-CAF/IL-33 were estimated; high level of Lnc-CAF, but not IL-33, predicted poor clinical survival outcome of patients with OSCC (Figure 6C and D). Figure 6. View largeDownload slide The significance of high Lnc-CAF levels in patients with OSCC and xenograft mouse model. (A, B) The relationship between Lnc-CAF/IL-33 expression and clinicopathological characteristics in patients with OSCC was analyzed, including TNM stage, Ki-67 and lymph node metastasis. (C, D) The prognostic value of Lnc-CAF/IL-33 was determined. (E) Knockdown of Lnc-CAF in CAFs and/or HSC3 cells and the xenograft mouse model of OSCC was established. Knocked down cells were subcutaneously injected in the rear flank of the nude mice (eight per group). The mean tumor size (mm3) was determined. (F) The micro-vessel density (MDV) was estimated by CD31 immunohistochemistry (200×); the expression of the proliferation index, Ki-67 (400×) and α-SMA (200×) in CAFs was analyzed. (G) The mRNA expression of Lnc-CAF and IL-33 was determined by Q-PCR. (H) Model summarizing the pro-tumorigenic effect of Lnc-CAF/IL-33 in the tumor microenvironment of OSCC. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Figure 6. View largeDownload slide The significance of high Lnc-CAF levels in patients with OSCC and xenograft mouse model. (A, B) The relationship between Lnc-CAF/IL-33 expression and clinicopathological characteristics in patients with OSCC was analyzed, including TNM stage, Ki-67 and lymph node metastasis. (C, D) The prognostic value of Lnc-CAF/IL-33 was determined. (E) Knockdown of Lnc-CAF in CAFs and/or HSC3 cells and the xenograft mouse model of OSCC was established. Knocked down cells were subcutaneously injected in the rear flank of the nude mice (eight per group). The mean tumor size (mm3) was determined. (F) The micro-vessel density (MDV) was estimated by CD31 immunohistochemistry (200×); the expression of the proliferation index, Ki-67 (400×) and α-SMA (200×) in CAFs was analyzed. (G) The mRNA expression of Lnc-CAF and IL-33 was determined by Q-PCR. (H) Model summarizing the pro-tumorigenic effect of Lnc-CAF/IL-33 in the tumor microenvironment of OSCC. *P < 0.05, **P < 0.001, ***P < 0.001 by unpaired t-test. Lnc-CAF knockdown impairs OSCC tumor growth To further examine the tumorigenic role of Lnc-CAF in vivo, we produced two cell lines, CAFs and HSC3 cells knocked down for Lnc-CAF by a lentiviral vector, RNAi-Lnc-CAF. Lnc-CAF knockdown had no effect on cell viability. Lnc-CAF knockdown was performed in CAFs to impair the pro-tumor characteristics of stromal fibroblasts, while it was performed in HSC3 cells to inhibit the tumor cell-derived stromal support (e.g. cytokines and exosomes). Overall, Lnc-CAF knockdown in these two cell types was performed to inhibit the role of Lnc-CAF in the TME. The results showed that co-culture of CAFs with HSC3 cells promoted tumor growth, while Lnc-CAF knockdown in CAF or HSC3 antagonized this pro-tumor effect. Additionally, Lnc-CAF knockdown in both CAFs and HSC3 cells maximized the antitumor effect (Figure 6E). Besides, the proliferation and angiogenesis indices, Ki-67 and CD31, respectively, were determined in vivo. Mice injected with both CAFs and HSC3 knocked down for Lnc-CAF showed the lowest expression of Ki-67, CD31 (Figure 6F), Lnc-CAF and IL-33 (Figure 6G), which could be explained by the significant reduction of α-SMA+ CAFs in the TME (Figure 6F). Thus, herein, we identified a novel LncRNA signal, Lnc-CAF/IL-33, in the stroma and tumor cells, which supports the growth of tumor cells via a positive feedback loop in the TME (Figure 6H). Discussion In this study, we examined lncRNA profiles during NF/CAF transformation and identified an uncharacterized intergenic lncRNA, Lnc-CAF, which was elevated in CAFs. Lnc-CAF increases the expression of α-SMA, vimentin and N-cadherin and promotes the activation of CAFs, leading to tumor proliferation, which predicted poor clinical outcome in patients with OSCC. In the co-expression networks obtained after RNA sequencing, Lnc-CAF was directly associated with upregulated IL-33 in CAFs. As a member of the IL-1 family, IL-33 was a cytokine ‘favoring’ immune responses with T helper type 2 (Th2) bias during allergic inflammation, parasitic infections and cancer (29). However, the role of IL-33 in carcinogenesis was unclear. Maywald et al. (30) reported that IL-33 expression was induced in the tumor epithelium of adenomas and carcinomas, leading to the activation of ST2+ myofibroblasts and master cells for the expression of extracellular matrix components and growth factors associated with polyposis and intestinal tumor progression. IL-33 is abundantly released in breast cancer tissues to reduce apoptosis and sustain the survival of myeloid-derived suppressor cells (MDSCs), which is associated with the activation of NF-kB and MAPK signaling for the induction of autocrine secretion of GM-CSF (31). In this study, IL-33 expression was upregulated to promote the activation phenotype of CAFs, resulting in CAF-supported tumor growth in OSCC. These findings are consistent with previous reports in head and neck squamous cell carcinoma, indicating that CAFs abundantly express IL-33 to induce EMT-mediated invasiveness of tumor cells (32). Different studies showed that IL-33 expression is regulated via different pathways in a cell type and stimulator–specific fashion. TLR3 and TLR4 agonists could significantly induce IL-33 expression in peritoneal macrophages via the cAMP-CREB pathway (33). In this study, upregulated Lnc-CAF promoted IL-33 stability by preventing IL-33 degradation through the p62-dependent autophagy–lysosome pathway, although it was not related to the lncRNA scaffold function. Exosomes are small membrane vesicles with a size of 50–100 nm that contain ncRNAs, mRNAs, and proteins. TME-derived exosomes can reprogram stromal cells and tumor cells (34). In chronic lymphocytic leukemia (CLL), CLL-derived exosomes contain miR-21 and -146a and can be actively incorporated by endothelial and mesenchymal stem cells, inducing an inflammatory phenotype of CAFs (35). Under nutrient deprivation or nutrient stressed conditions, CAF-derived exosomes contain amino acids, lipids, and TCA-cycle intermediates, which could be utilized by cancer cells for central carbon metabolism, promoting tumor growth (36). We observed that OSCC cells secreted exosomes containing Lnc-CAF to stromal fibroblasts and promoted the up-regulation of Lnc-CAF levels for the activation of CAFs, which formed a positive regulation loop to induce tumor proliferation. Thus, the expression of Lnc-CAF is not limited to CAF, but also tumor cells. In summary, in this study, we report a specific LncRNA profile during the activation of CAFs in OSCC and identified an uncharacterized intergenic lncRNA, Lnc-CAF, which was significantly elevated in CAFs and could stabilize and up-regulate cytokine IL-33 to reprogram CAFs, supporting tumor growth. While further studies on the interaction way between Lnc-CAF and IL-33 are need to be investigated. Supplementary material Supplementary material is available at Carcinogenesis online. Funding This work was supported by the National Natural Science Foundation of China (81402238, 81772880, 81271698, 31370899, 81702680) and the fundamental research funds for the central universities (021414380325), the Nanjing Medical Science and technique development foundation (YKK16164, QRX17083) and the Jiangsu Provincial key Medicine discipline (since 2017). Abbreviations CAFs carcinoma-related fibroblasts lncRNA long non-coding RNA NF normal fibroblast ncRNA non-coding RNA OSCC oral squamous cell carcinoma PBS phosphate-buffered solution TME tumor microenvironment Acknowledgements We thank Hospital of Stomatology, Nanjing for their support and for providing clinical samples. L.D., YH.N. and YY.H. designed experiments with valuable help from X.F.H., J.R. and Q.G.H. L.D. performed and analyzed data with valuable help from Y.L. and D.Y.Z., H.W. L.D. and YY.H wrote the manuscript. XF.H and L.D. collected surgical specimens. YY.H. and YH.N. oversaw the overall project. YX.S helped to revise the manuscript. Conflict of Interest Statement: None declared. References 1. Schmitt, A.M.et al.  ( 2016) Long noncoding RNAs in cancer pathways. Cancer Cell , 29, 452– 463. Google Scholar CrossRef Search ADS PubMed  2. Sun, W. et al. (2014) Sequencing the head and neck cancer genome: implications for therapy. Ann. N. Y. Acad. Sci., 1333, 33–42. 3. Huarte, M. ( 2015) The emerging role of lncRNAs in cancer. Nat. Med ., 21, 1253– 1261. Google Scholar CrossRef Search ADS PubMed  4. Liu, Z.et al.  ( 2017) Over-expressed long noncoding RNA HOXA11-AS promotes cell cycle progression and metastasis in gastric cancer. Mol. Cancer , 16, 82. Google Scholar CrossRef Search ADS PubMed  5. Ding, L.et al.  ( 2017) The TLR3 agonist inhibit drug efflux and sequentially consolidates low-dose cisplatin-based chemoimmunotherapy while reducing side effects. Mol. Cancer Ther ., 16, 1068– 1079. Google Scholar CrossRef Search ADS PubMed  6. Casey, S.C.et al.  ( 2015) Cancer prevention and therapy through the modulation of the tumor microenvironment. Semin. Cancer Biol ., 35, S199– S223. Google Scholar CrossRef Search ADS PubMed  7. Salo, T.et al.  ( 2014) Insights into the role of components of the tumor microenvironment in oral carcinoma call for new therapeutic approaches. Exp. Cell Res ., 325, 58– 64. Google Scholar CrossRef Search ADS PubMed  8. Mitra, A.K.et al.  ( 2012) MicroRNAs reprogram normal fibroblasts into cancer-associated fibroblasts in ovarian cancer. Cancer Discov ., 2, 1100– 1108. Google Scholar CrossRef Search ADS PubMed  9. Kidd, S.et al.  ( 2009) Direct evidence of mesenchymal stem cell tropism for tumor and wounding microenvironments using in vivo bioluminescent imaging. Stem Cells , 27, 2614– 2623. Google Scholar CrossRef Search ADS PubMed  10. Abe, R.et al.  ( 2001) Peripheral blood fibrocytes: differentiation pathway and migration to wound sites. J. Immunol ., 166, 7556– 7562. Google Scholar CrossRef Search ADS PubMed  11. Kidd, S.et al.  ( 2012) Origins of the tumor microenvironment: quantitative assessment of adipose-derived and bone marrow-derived stroma. PLoS One , 7, e30563. Google Scholar CrossRef Search ADS PubMed  12. Zeisberg, E.M.et al.  ( 2008) Fibroblasts in kidney fibrosis emerge via endothelial-to-mesenchymal transition. J. Am. Soc. Nephrol ., 19, 2282– 2287. Google Scholar CrossRef Search ADS PubMed  13. Zhao, L.et al.  ( 2017) Long noncoding RNA LINC00092 acts in cancer-associated fibroblasts to drive glycolysis and progression of ovarian cancer. Cancer Res ., 77, 1369– 1382. Google Scholar CrossRef Search ADS PubMed  14. Zhuang, J.et al.  ( 2015) TGFβ1 secreted by cancer-associated fibroblasts induces epithelial-mesenchymal transition of bladder cancer cells through lncRNA-ZEB2NAT. Sci. Rep ., 5, 11924. Google Scholar CrossRef Search ADS PubMed  15. Huang, J.K.et al.  ( 2017) LncRNA-MALAT1 promotes angiogenesis of thyroid cancer by modulating tumor-associated macrophage FGF2 protein secretion. J. Cell. Biochem ., 118, 4821– 4830. Google Scholar CrossRef Search ADS PubMed  16. Wu, M.H.et al.  ( 2011) Targeting galectin-1 in carcinoma-associated fibroblasts inhibits oral squamous cell carcinoma metastasis by downregulating MCP-1/CCL2 expression. Clin. Cancer Res ., 17, 1306– 1316. Google Scholar CrossRef Search ADS PubMed  17. Marsh, T.et al.  ( 2013) Fibroblasts as architects of cancer pathogenesis. Biochim. Biophys. Acta , 1832, 1070– 1078. Google Scholar CrossRef Search ADS PubMed  18. Liang, D.et al.  ( 2015) Activated STING enhances Tregs infiltration in the HPV-related carcinogenesis of tongue squamous cells via the c-jun/CCL22 signal. Biochim. Biophys. Acta , 1852, 2494– 2503. Google Scholar CrossRef Search ADS PubMed  19. Ni, Y.H.et al.  ( 2015) Distinct expression patterns of Toll-like receptor 7 in tumour cells and fibroblast-like cells in oral squamous cell carcinoma. Histopathology , 67, 730– 739. Google Scholar CrossRef Search ADS PubMed  20. Song, Y.et al.  ( 2017) Exosomal miR-146a contributes to the enhanced therapeutic efficacy of interleukin-1β-primed mesenchymal stem cells against sepsis. Stem Cells , 35, 1208– 1221. Google Scholar CrossRef Search ADS PubMed  21. Sponheim, J.et al.  ( 2010) Inflammatory bowel disease-associated interleukin-33 is preferentially expressed in ulceration-associated myofibroblasts. Am. J. Pathol ., 177, 2804– 2815. Google Scholar CrossRef Search ADS PubMed  22. Wang, M.et al.  ( 2015) Large-scale association analysis in Asians identifies new susceptibility loci for prostate cancer. Nat. Commun ., 6, 8469. Google Scholar CrossRef Search ADS PubMed  23. Lemire, M.et al.  ( 2015) A genome-wide association study for colorectal cancer identifies a risk locus in 14q23.1. Hum. Genet ., 134, 1249– 1262. Google Scholar CrossRef Search ADS PubMed  24. Micke, P.et al.  ( 2011) Gene copy number aberrations are associated with survival in histologic subgroups of non-small cell lung cancer. J. Thorac. Oncol ., 6, 1833– 1840. Google Scholar CrossRef Search ADS PubMed  25. Bockmühl, U.et al.  ( 2000) Genetic imbalances with impact on survival in head and neck cancer patients. Am. J. Pathol ., 157, 369– 375. Google Scholar CrossRef Search ADS PubMed  26. Ulitsky, I.et al.  ( 2013) lincRNAs: genomics, evolution, and mechanisms. Cell , 154, 26– 46. Google Scholar CrossRef Search ADS PubMed  27. Luga, V.et al.  ( 2013) Tumor-stroma interaction: revealing fibroblast-secreted exosomes as potent regulators of Wnt-planar cell polarity signaling in cancer metastasis. Cancer Res ., 73, 6843– 6847. Google Scholar CrossRef Search ADS PubMed  28. Webber, J.P.et al.  ( 2015) Differentiation of tumour-promoting stromal myofibroblasts by cancer exosomes. Oncogene , 34, 290– 302. Google Scholar CrossRef Search ADS PubMed  29. Martin, N.T.et al.  ( 2016) Interleukin 33 is a guardian of barriers and a local alarmin. Nat. Immunol ., 17, 122– 131. Google Scholar CrossRef Search ADS PubMed  30. Maywald, R.L.et al.  ( 2015) IL-33 activates tumor stroma to promote intestinal polyposis. Proc. Natl. Acad. Sci. USA , 112, E2487– E2496. Google Scholar CrossRef Search ADS   31. Xiao, P.et al.  ( 2016) Interleukin 33 in tumor microenvironment is crucial for the accumulation and function of myeloid-derived suppressor cells. Oncoimmunology , 5, e1063772. Google Scholar CrossRef Search ADS PubMed  32. Chen, S.F.et al.  ( 2013) The paracrine effect of cancer-associated fibroblast-induced interleukin-33 regulates the invasiveness of head and neck squamous cell carcinoma. J. Pathol ., 231, 180– 189. Google Scholar CrossRef Search ADS PubMed  33. Polumuri, S.K.et al.  ( 2012) Transcriptional regulation of murine IL-33 by TLR and non-TLR agonists. J. Immunol ., 189, 50– 60. Google Scholar CrossRef Search ADS PubMed  34. Yu, S.et al.  ( 2015) Tumor-derived exosomes in cancer progression and treatment failure. Oncotarget , 6, 37151– 37168. Google Scholar PubMed  35. Paggetti, J.et al.  ( 2015) Exosomes released by chronic lymphocytic leukemia cells induce the transition of stromal cells into cancer-associated fibroblasts. Blood , 126, 1106– 1117. Google Scholar CrossRef Search ADS PubMed  36. Zhao, H.et al.  ( 2016) Tumor microenvironment derived exosomes pleiotropically modulate cancer cell metabolism. Elife , 5, e10250. Google Scholar PubMed  © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.

Journal

CarcinogenesisOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off