Background
Oral cancer is the sixth most common cancer in the world, and oral squamous cell carcinoma (OSCC) accounts for 90% of oral malignant neoplasms [
1,
2]. Although early diagnosis and treatment are developing and improving, the five-year survival rate of patients with OSCC remains low because of distant metastasis and recurrence, and the quality of life of patients has remained a serious problem [
3,
4]. Thus, further understanding of the complex pathogenesis of OSCC is needed to develop a more effective therapy.
The tumor microenvironment (TME), a complex meshwork of extracellular matrix (ECM) macromolecules, plays an important role in tumor progression. As a main component of TME, cancer-associated fibroblasts (CAFs) are of great importance for ECM production and interstitial tissue remodeling in the disturbed homeostasis [
5], which can induce stiffness and hypoxia in the matrix [
6]. Matrix stiffness is one of the most important biomechanical characteristics of TME faced by cancer cells and stromal cells. Recently, there has been a consensus that mechanical stimulation in the TME, such as the alteration of matrix stiffness, can have an impact on the development of solid tumors by a variety of signaling pathways, and then promote invasion and metastasis [
7]. Thus, the relationship between CAFs and matrix stiffness and the mechanism of CAFs remodeling ECM are especially worthy of further study.
Collagen is the most common ECM scaffold protein in the matrix [
8], and its remodeling and cross-linking critically affect the tumor progression by enhancing ECM stiffness [
9]. Previous studies reported that collagen is associated with the development of several malignancies such as breast, colon and liver cancer [
7,
10,
11]. Increased collagen deposition and ECM density have been found in the tumor invasion frontier and have been shown to have a significant association with several important clinicopathological parameters [
12]. Lysyl oxidase (LOX), a copper-dependent amine oxidase secreted by fibrogenic cells [
13], can initiate covalent cross-linking of collagen and elastin in the ECM [
14]. The elevation of LOX expression has been found in many malignant tumors [
13,
15,
16] and correlates with enhanced cancer cell proliferation, metastasis, and angiogenesis in the TME [
17,
18]. LOX can be a molecular target for improving the efficacy of chemotherapies against different types of solid tumors [
19], as it has been reported to promote tumor progression by regulating collagen cross-linking [
17]. However, the relationship among CAFs, LOX, collagen and their mechanism in the remodeling matrix and influencing tumor progression in OSCC are still largely unknown.
A stiff microenvironment has an effect on many aspects of tumor cells, including internal molecular signaling pathways and cell behaviors [
20]. Cellular interactions with the extracellular matrix, especially the state of cell adhesion, have changed in tumor initiation and progression. Focal adhesion kinase (FAK), a prominent signaling molecule in focal adhesions, is over-phosphorylated in multiple cancer progression and is responsible for transmitting extracellular mechanical force to changes in gene expression within cancer cells [
14,
21]. The potential of mechanical stimuli such as elevated collagen and tissue stiffness is responsible for the induction of β-catenin translocation and the expression of Wnt-associated proteins [
22]. However, the molecular mechanisms driving tissue stiffness-related tumor progression have not been completely elucidated in OSCC.
In the present study, we sought to investigate the correlation between ECM remodeling mediated by CAFs-derived LOX and OSCC development. The colocalization of CAFs and LOX in tumor stroma were explored in OSCC tissues. The expression of LOX in CAFs and its effect on biological behavior of OSCC cells were investigated in vitro. Then, we studied whether LOX expressed by CAFs can significantly catalyze the cross-linking of collagen and increase matrix stiffness, thereby regulating the morphology and invasion of tumor cells in a FAK-dependent manner. Thus, our study suggests that matrix stiffness mediated by CAFs-derived LOX might present a novel therapeutic thought for OSCC patients.
Methods
Patients and specimens
Ninety six cases of primary OSCC tissue samples were recruited from 2010 to 2015 from patients treated with surgery at the Department of Oral and Maxillofacial Surgery, the Affiliated Stomatological Hospital of Nanjing Medical University: 59 were male (61.46%) and 37 were female (38.54%), with an average age of 60.4 years (range, 35–91 years). The primary tumors were located in the tongue (n = 32), gingiva (n = 21), buccal mucosa (n = 26), and other sites (n = 17), such as the palate, lower lip, jaw, soft palate, oropharynx, or floor of the mouth. All patients underwent extensive resection of the primary tumor and underwent classic radical neck dissection or selective regional lymph node dissection. None of the patients had received chemotherapy or radiotherapy before the surgery. All patients had complete clinical data and clear pathological diagnosis, including sex, age, tumor location, tumor size, lymph node metastasis, distant metastasis, pathological grade, clinical stage, postoperative recurrence, survival and follow-up information. Clinical stage and TNM classification were defined by the International Union Against Cancer (UICC) and the American Joint Commission on Cancer (AJCC). The pathological classification was based on the World Health Organization (WHO) criteria. This study was approved by the ethics committee of Nanjing Medical University and was conducted in accordance with the World Medical Association Declaration of Helsinki. Written informed consent was obtained from all the patients.
Multiple immunohistochemical staining
Multiplex IHC staining was performed using Polychromatic fluorescent IHC Kit (CAT#:10080100050, PANOVUE, China). Formalin-fixed, paraffin-embedded primary OSCC tissue specimens were cut into 4 μm-thick sections, cleared with xylene, and rehydrated through a graded alcohol series. After antigen retrieval and blocking, the sections were incubated overnight with primary antibodies against as follows: LOX (1:100; Abcam, Cambridge, MA, USA), α-SMA (1:200; Abcam), EPCAM (1:200; Abcam), FAP (1:250; Abcam). Followed by secondary antibodies, fluorophore (FITC, Cy3, Cy5)-conjugated TSA signal amplification buffer were added. After three sequential reactions, sections were counterstained with DAPI and mounted with enhanced anti-fluorescence quencher. Observation was under laser confocal microscope (Zeiss, Oberkochen, Germany).
For immunohistochemical staining, after incubated overnight with primary antibodies against LOX (1:100; Abcam) and α-SMA (1:200; Abcam), all sections were incubated with HRP-polymer anti-rabbit Kit and a DAB Detection Kit (Maixin Biotech, Fuzhou, China) followed by counterstaining with hematoxylin. The sections were dehydrated in gradient alcohol, clarified in xylene, mounted with neutral gum. Observation was under microscope (Leica Microsystems, Mannheim, Germany).
Evaluation of immunoreactivity
The staining intensity of tissue sections was individually evaluated by two pathologists who were blinded to the patients' clinical data. Immunoreactivity of LOX and α-SMA in tumor stroma was estimated according to the classification system described by Fuji et al. [
23]: 0, negative (no staining); 1, sparse (a small population of discrete stained fibroblasts); 2, focal (irregular and intermittent areas of stained fibroblasts); and 3, abundant (widespread and consecutive areas of stained fibroblasts). The specimens were categorized into two subgroups according to immunoreactivity score: low expression (0–1) and high expression (2–3).
Fibroblast isolation and preparation of conditioned media
Fresh oral cancer tissues and matched normal tissues were collected under sterile conditions. All samples were obtained with informed consent from the patients, and the fibroblast isolation protocol was approved by the Nanjing Medical University Ethics Committee. All tissue samples were snap-frozen in liquid nitrogen and stored at − 80 °C until fibroblast isolation. Fresh tissues were washed with PBS (penicillin 100 U/mL, streptomycin 100 μg/mL) three times and cut into approximately 1 mm3 size. The tissues were spread evenly on the culture dish and placed upside down in a primary incubator for 4 h to adhere. The tissues were maintained in DMEM (Gibco, Grand Island, NY, USA) with 15% FBS (Gibco) for 1–2 weeks. Purified fibroblasts were obtained by differential enzyme digestion and maintained in DMEM with 10% FBS. Cells were seeded into T-25 culture flasks and grown in 5 ml serum-free media for 48 h until the cells were ~ 80% confluent. Conditioned media were collected and centrifuged at 2000 rmp for 30 min to remove cellular debris.
Cell culture
HNSCC cell lines Cal27 and HN6 were used. Cal27 and HN6 cell lines were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). All cancerous cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM)/F12 (Invitrogen) supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin/streptomycin at 37 °C in a 5% CO2-humidified incubator. All cells were routinely tested for mycoplasma at regular intervals throughout the whole course of the study. To inhibit of FAK phosphorylation, cells were treated with FAK inhibitor 14 (FAKi, Santa Cruz Biotechnology, Santa Cruz, CA, USA) at different concentrations.
Lentiviral construct and transfection
Human LOX-encoding lentiviral vectors were constructed by GeneChem Co., Ltd (Shanghai, China). CAFs cells were transfected with a specific shRNA lentivirus targeting LOX knockdown, and the scrambled sequence was co-transfected as a negative control. All transfection procedures were performed according to the manufacturer’s instructions. Cells in the control group (CAFs-NC) and the experimental group (CAFs-shLOX) were cultured at 37 °C in a 5% CO2 incubator and changed to complete medium 12–16 h after transfection. Stable cells were selected using 2 mg/ml puromycin. Fluorescence microscopy was used to observe transfection efficiency, and real-time polymerase chain reaction (PCR) and western blot were used to estimate the efficiency of LOX knockdown.
RNA isolation and quantitative reverse transcription-quantitative PCR (qRT-PCR)
Total RNA was extracted from cells with TRIzol reagent (Invitrogen) and equal quantities of cDNA were reverse transcribed using Superscript (Vazyme, Nanjing, China) according to the manufacturer’s instructions. qPCR was performed using the PCR System 7900 (Applied Biosystems, Foster City, CA, USA) with SYBR Green Master Mix (TaKaRa, Shiga, Japan). The primers for GAPDH were: forward 5′-GAAGGTGAAGGTCGGAGTC-3′ and reverse 5′-GAGATGGTGATGGGATTTC-3′; LOX: forward 5′-TTCTTACCCAGCCGACCAAGATA-3′ and reverse 5′-GTGTTGGCATCAAGCAGGTCA-3′; α-SMA: forward 5′-CTGGCCGAGATCTCACTGACTA-3′ and reverse 5′-GCCCATCAGGCAACTCGTAA-3′; FAP: forward 5′-GACCCACGCTCTGAAGACAGAATTA-3′ and reverse 5′-AGCAGAGGTGGCAACTCCAAATAC-3′; FSP-1: forward 5′-CAGATAAGCAGCCCAGGAAGA-3′ and reverse 5′-AAGGAGCCAGGGTGGAAAA-3′; E-cadherin: forward 5′-TACACTGCCCAGGAGCCAGA-3′ and reverse 5′-TGGCACCAGTGTCCGGATTA-3′; N-cadherin: forward 5′-TGGACCATCACTCGGCTTA-3′ and reverse 5′-ACACTGGCAAACCTTCAC-3′; vimentin: forward 5′-TGAGTACCGGAGACAGGTGCAG-3′ and reverse 5′-TAGCAGCTTCAACGGCAAAGTTC-3′. The following program was performed using a two-step cycling protocol: an initial denaturation at 95 °C for 30 s, followed by 40 cycles of 95 °C for 5 s, 60 °C for 31 s. Each sample in each experiment was performed in triplicate and the results were expressed as the mean of three independent experiments. Relative expression levels of related genes were quantified and compared to the internal control GAPDH and analyzed using the 2−ΔΔCT method.
Protein extraction and nuclear, cytoplasmic protein fractionation
Protein extraction was performed using common methods. Briefly, cells were lysed with an appropriate radioimmunoprecipitation assay (RIPA) buffer (Beyotime, Shanghai, China) containing protease inhibitor cocktail and phosphatase inhibitor cocktail (Beyotime, Shanghai, China), incubated for 15 min at 4 °C, and scraped using a cell scraper. For cells cultured on collagen, cells were agitated on a rocking platform at 4 °C for 30 min and then removed the lysis buffer. Nuclear and cytoplasmic proteins were isolated using the Invent Nuclear and Cytoplasmic Extraction Reagents Kit (Invent Biotechnologies, Plymouth, MN, USA) according to the manufacturer’s instructions.
Western blot assay
Western blot assays were performed according to the common methods. Briefly, equal amounts of protein samples were separated using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA). The membranes were blocked with 5% skimmed milk at room temperature for 2 h and incubated at 4 °C overnight with primary antibodies against β-actin as a control (1:500; Proteintech), α-SMA (1:200; Abcam), FSP-1 (1:1000; Abcam), FAP (1:800; Abcam), LOX (1:1000; Abcam), E-cadherin (1:1000; CST), N-cadherin (1:1000; CST), vimentin (1:500; Santa Cruz Biotechnology), β-catenin (1:1000; CST), FAK (1:1000, CST), p-FAK(Tyr397) (1:1000, CST), c-myc (1:500, CST), cyclin D1 (1:500, CST), Lamin B1 (1:1000, CST) followed by incubation with anti-goat immunoglobulin G (IgG) horseradish peroxidase (HRP)-conjugated secondary antibodies (Zhongshan Golden Bridge Bio, Beijing, China) for 45 min at room temperature. The protein bands were detected using ECL chemiluminescence reagents (Millipore) and visualized using the ImageQuant LAS4000 Mini Imaging System (General Electrics, Louisville, Kentucky). Analyses of the bands were performed using ImageJ software.
Cell proliferation assay
Cell proliferation was assessed by CCK-8 cell assay (Cell Counting Kit-8, Dojindo, Japan). Cells (3 × 103 cells/well) were seeded into 96-well plates, and 10 μl CCK-8 reagent was added to each well 2–4 h before testing at the same time. Cell viability was determined at 0, 1, 2, 3 and 4 days after cell attachment. Three parallel repeats were performed for each sample in each experiment and the results were expressed as the mean of three independent experiments.
Cell migration and invasion assays
Cell migration and invasion assays in vitro were performed using wound healing and transwell assays. For the wound healing assay, cells were plated in six-well plates and grown to 90% confluence. Artificial wounds were created with a 200 μl sterile pipette tip across the cell surface and then the cells were incubated with CM to migrate into the open area. Images of the same area of the wound were taken at 0, 12 and 24 h to determine the wound closure. Cell invasion assays were performed using 8 μm pore size transwell chambers (Miliipore) whose upper chamber was precoated with Matrigel (Corning, Bedford, MA, USA). For the migration assays, there was no matrigel in the upper chamber. Approximately 1 × 105 cells per well were seeded in the upper chambers and incubated with CM in the lower chamber. After 24 h, the transwell chambers were fixed with 4% PFA and stained with crystal violet (Sigma-Aldrich, St. Louis, MO, USA). Cells attached to the lower layer were imaged (Olympus, Tokyo, Japan) and counted in randomly selected fields.
Preparation of collagen matrix
For the fibroblast-modified collagen, fibroblasts were resuspended in FBS, and type I collagen, 5 × DMEM and reconstitution buffer (50 mmol/L NaOH, 260 mmol/L NaHCO3, and 200 mmol/L HEPES) were sequentially added and evenly mixed. The mixture was spread in a 12-well plate, and the plate was placed in a 5% CO2 incubator at 37 °C for 30–60 min to solidify. Complete medium was added and the medium was changed every day. For the collagen treated with CAFs-CM, CM from CAFs-NC, CAFs-shLOX were collected to stimulate the formation of collagen with different stiffness. DMEM was used as blank control, and human recombinant LOX (huLOX; OriGene, Rockville, MD, USA) was used for the rescue of LOX knockdown. 12-well plates were coated with 1.5 mg/ml collagen and placed at room temperature for 30 min followed by incubation at 37 °C for 2 h, and then stimulated by DMEM, CM from CAFs-NC, CAFs-shLOX and CAFs-shLOX + huLOX for 5 days. For the collagen treated with drugs, huLOX was used to induce crosslinking of collagen. Ribose (Sigma-Aldrich, Dorset, UK), which induced crosslinking non-enzymatically, was used as a positive control and BAPN (Sigma-Aldrich, Dorset, UK) was used as an inhibitor of LOX. Then, PBS only (PBS), or PBS containing 150 ng/ml huLOX (huLOX), or 60 mM ribose (ribose), or huLOX combined with 10 mM BAPN (huLOX + BAPN) were added to each group and incubated at 37 °C for 5 days.
Cellular immunofluorescence assay
For cellular immunofluorescence staining, cells were attached to cell slides in 24-well plates. For F-actin staining, the cells were grown on collagen gels for 24 h. Then the cell slides and collagen were fixed in 4% PFA for 20 min at room temperature and permeabilized with 0.1% Triton X-100 for 15 min. Then incubated with primary antibody against E-cadherin (1:1000; CST), vimentin (1:500; Santa Cruz Biotechnology), F-actin (1:100, Beyotime) overnight at 4 °C. Nucleus were stained with 4,6-diamidino-2-phenylindole (DAPI) (Beyotime, Shanghai, China). Fluorescence images were visualized using a fluorescence microscope (Leica Microsystems, Mannheim, Germany).
Determination of matrix stiffness
The relative stiffness of collagen gels was measured using a MARS60 microinfrared rheometer (Thermo Fisher, MA, USA) under the following conditions. Briefly, the collagen gels were measured over a range from 0.5–5% strain at a fixed angular frequency of 0.5 rad/s and temperature of 21 °C on MARS60 microinfrared rheometer. The output of G’ was recorded for 20 min with a data point obtained every 60 s. The samples were found to be only minimally frequency dependent within the range of testing and showed a linear viscoelastic response within the strain range evaluated.
Three-dimensional co-culture system
Fibroblast-modified collagen was prepared as previously described. Cal27 cells were resuspended in the mixed medium and then seeded onto the surface of the solidified gel to create a 3D co-culture system. The culture medium was changed daily. After 3 days, the collagen gels were transferred to the scaffold in six-well plates for culture at the air–liquid interface. At the specific time point, the gels were fixed with 4% PFA, embedded in paraffin, and cut into 4 μm sections for H&E staining.
Statistical analysis
The correlation of α-SMA and LOX expression with clinicopathological parameters and prognosis was analyzed using ANOVA, the chi-square test, Fisher’s exact test, Mann–Whitney test, COX regression analysis and Kaplan‐Meier survival analysis. Spearman's correlation analysis was used to estimate the correlation between α-SMA and LOX in the tumor stroma. The results are represented as the means ± SEM or as specified in the figure legends. The data were analyzed using SPSS 18.0 and Prism 7 software (GraphPad Software, La Jolla, CA, USA). Comparisons between the two groups were analyzed using the 2-tailed, unpaired Student’s t test. Comparisons among the three groups were performed using one-way ANOVA followed by Tukey’s multiple comparisons test. Adjusted values of P < 0.05 were considered significant.
Discussion
Recently, it has become clear that TME promotes tumor progression [
25]. The components and mechanical properties of ECM have aroused concern as important determinants of cancer cell behavior and disease progression. As the major TME cellular component, CAFs play a key role in matrix remodeling [
6]. However, whether LOX participates in extracellular matrix remodeling induced by CAFs and affects the progression of OSCC remains unclear. Our study revealed that LOX is highly expressed in CAFs, and LOX can mediate CAFs-induced matrix remodeling by catalyzing collagen cross-linking. The increase in matrix stiffness can, in turn, promote tumor invasion and progression in an FAK-dependent manner.
CAFs are responsible for the secretion of multiple growth factors, kinases, cytokines and chemokines into the TME to facilitate tumor progression. Research identified high expression of LOX in several malignant tumors [
26]. The secretion of LOX was reported to induce the formation of a pre-metastatic niche in preclinical breast cancer models [
27]. LOX has a significant effect on ECM remodeling in the lungs and promotes lung metastasis [
28] and correlates with increased staging in renal cell carcinoma [
13]. In previous studies, the expression of LOX was usually stronger at the tumor invasive front than in the tumor center in OSCC, and higher expression of LOX was associated with advanced pathological grade [
12]. Here, we found that LOX was highly expressed in the same region of CAFs distribution in tumor stroma (Fig.
1), and there was a statistical correlation between the two (Table
1). Notably, higher expression of LOX in the tumor stroma was associated with an advanced clinical stage and poor prognosis (Tables
2,
3,
4). Moreover, we found that the expression level of LOX was much higher in CAFs than in NFs, and further confirmed that LOX expressed in CAFs functioned as an oncogene in OSCC (Fig.
2). These results suggest that the expression of LOX in tumor stroma may be a key molecule affecting the progression of OSCC.
There has been extensive evidence to prove that biophysical factors, including matrix stiffness and mechanical forces, play critical roles in the initiation and progression of malignant tumors [
29‐
31]. CAFs can produce large quantities of ECM molecules to remodel stroma, and LOX expression and collagen content have been observed to be clinically elevated in various malignant tumors [
13,
32‐
35]. In Kurie’s review, the deposition of a cross-linked collagen matrix by CAFs leads to intratumoral fibrosis [
36]. Xu et al. summarized the most abundant matrix protein polymers are collagens, whose crosslinking by LOX can create a stiffer tumor microenvironment, regulate tumor immunity, and promote metastasis [
37]. In present study, in order to investigate the effect of LOX in CAFs on matrix remodeling, we created a co-culture system and found that the contractile capacity of collagen was enhanced by CAFs, leading to an increase in collagen density and a decrease in pore size. Meanwhile, the elastic modulus of collagen in the CAFs-co-cultured group was elevated (Fig.
3). Thus, we innovatively explored the relationship between CAFs and LOX, finding that LOX may be a key factor in CAFs to remodel ECM through collagen crosslinking, and then promote cancer progression.
In the development of malignant tumors, the complexity of the epithelial–stromal interaction is due to the fact that the ECM is a dynamic mechanical microenvironment. ECM remodeling leads to functional changes in the biomechanical properties of the ECM, resulting in the activation of pathogenic signaling pathways involved in various processes such as cell morphology, proliferation, differentiation, migration and etc. [
30,
38,
39]. In breast carcinoma, mammary epithelial cell cultured within high-density collagen form larger, less organized structures and proliferate faster than that those cultured in low-density collagen [
40]. LOX has been identified as playing a critical role in establishing and mediating a prometastatic microenvironment within fibrotic tissues [
35]. In the present study, type I collagen was stimulated with CM from CAFs or drugs to form different stiffness, and the morphology of oral cancer cells cultured on high-stiffness collagen became slender. Moreover, a 3D collagen co-culture model was successfully established, in which oral cancer cells were plated on collagen with different stiffness mediated by CAFs. Our results showed that an increase in collagen stiffness could promote the depth of tumor cell invasion into the collagen matrix (Fig.
4). Such changes in the mechanical properties of the ECM arising from LOX-mediated collagen cross-linking play a critical role in the transformed phenotype of oral cancer cells.
Tumors respond to changes in ECM stiffness by integrin activation and subsequent phosphorylation of FAK, leading to the activation of protein–protein interactions and adhesion signaling to promote tumorigenesis [
41,
42]. FAK is required for the activation of the Wnt/β-catenin pathway through phosphorylation of GSK3β
Y216, which is crucial for intestinal tumorigenesis [
43]. In Xue’s study, LOXL2 secreted by CAFs activated the FAK signaling pathway and then induced EMT in human colon cancer LOVO cells, thereby promoting the invasion and metastasis of colon cancer (
44). By seeding OSCC cells on the collagen with different stiffness, we found that collagen with a stiffer surface mediated by CAFs-derived LOX can trigger the EMT process, which was consistent with cell morphological changes. A stiffer surface effectively activated the phosphorylation of FAK at tyrosine 397 and β-catenin downstream factors in tumor cells. After inhibition of FAK phosphorylation, the proliferation and invasion ability of tumor cells were impaired, and influenced the nuclear accumulation of β-catenin and the EMT process (Fig.
5). Our data indicate that FAK signaling activation could be an important factor in the accumulation of nuclear β-catenin induced by matrix stiffness. However, further investigation is needed to determine the molecular mechanism by which the activation of the FAK pathway promotes OSCC progression. This is crucial for studying the occurrence and development of tumors.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.