Background
Head and neck squamous cell carcinoma (HNSCC), including oral squamous cell carcinoma (OSCC), is a common malignancy worldwide. The OSCC patients five-year survival rate is generally around 50% [
1]. Despite that, the fact that tissue assay and cell line analysis have yielded abundant findings that have advanced significantly the diagnosis and treatment of OSCC, the survival of OSCC patients has not been improved profoundly in the last few decades. Human OSCC cell lines have contributed significantly helping to elucidate the mechanisms associated with OSCC pathogenesis. However, the use of immunocompromised murine systems, which has allowed the xenografic growth of human OSCC cell lines, has not been sufficient when trying to address the in vivo neoplastic process and treatment responses; this is due to the lack of an immune microenvironment [
2]. The establishment of murine OSCC cell lines and their subsequent preclinical assessment using immunocompetent recipients would be an important step to facilitating the use of a translational approach.
Several mouse OSCC cell lines have been established using C3H, BALB/c and DBA/2 mouse strains, and these have yielded profound pathogenic insights into OSCC [
3‐
5]. In this context, the mouse strain C57BL/6 is the one that is most commonly used for gene engineering; this has led to rapid progress in terms of genetic engineering mouse models. Thus, the induction of OSCC using a combination of a genetically engineered background and chemical carcinogen treatment will open up a door to understanding the exogenous and endogenous factors that contribute to OSCC in C57BL/6 mice [
2]. miRNA disruption plays an important role in OSCC pathogenesis [
6‐
17]. We have established
miR-31 and
miR-211 transgenic (Tg) mouse lines that have these transgenes overexpressed in the mouse basal keratinocytes [
7,
8,
18]. These mice show higher frequency and faster OSCC tumor induction following 4-nitroquinoline 1-oxide (4NQO) treatment [
7,
8,
18,
19]. By means of these models, we have uncovered new suppressors that are targeted by these oncogenic miRNAs and unraveled the involvement of DNA defects and the enrichment of oxidative stress in OSCC progression. In addition, due to the rapid tumor induction and fluorescent tumor labeling in these mice, the models have been used to enable new developments in image diagnosis [
20].
Xenotransplantation requires a relatively shorter time period to obtain a full-blown tumor than chemical treatment [
2]. In addition, tumor xenografts have more homogeneous characteristics compared to chemically induced lesions. Xenografts of human cancer cells into immuno-compromised mice have helped with the functional elucidation of tumor growth and its interception. However, being able to carry out orthotopic xenotransplantation of mouse OSCC cells into immunocompetent syngeneic mice would help us to obtain a better and a more comprehensive understanding of tumor complexity, which in part is due to the presence of a relevant tumor microenvironment and appropriate immuno-modulation [
21]. This study establishes, for the first time, four murine OSCC cells lines; these were obtained from 4NQO treated transgenic mice. The genetic disruption and aggressiveness of these cell lines, their tumorigenicity, their ability to bring about both local regional metastasis and distal metastasis in C57BL/6 syngeneic mice are defined in the present study. These cell lines and the linked immunocompetent animal model that we have established will facilitate the investigation of therapies that can be used to treat OSCC.
Methods
Induction of OSCC from K14-EGFP-miR-211 Tg mice
miR-211 is an oncogenic miRNA associated with OSCC [
7,
8,
14,
15]. The K14-EGFP-
miR-211 Tg mouse has been established in C57BL/6 previously using the murine pri-
miR-211 sequence tagged with a green fluorescence protein (GFP) [
7]. For OSCC induction, 100 μg/ml of 4NQO was added to the drinking water of 6–8 week-old mice for 16 weeks. Mice were sacrificed at a time point when their body weight loss was > 1/3, when tumors had begun to interfere with their food uptake, when they showed weakness, or when at a specific endpoint [
7,
11,
18].
Establishing the MOC-L cell lines
The human OSCC SAS cell line was obtained from the JCRB cell bank, while the FaDu and 293 T ell lines were obtained from the ATCC cell bank. Human cell lines were authenticated by short tandem repeat analysis. Mouse OSCC lesions on the dorsal tongue surface were dissected and introduced into the primary culture. The cells were maintained in complete DMEM medium (Thermo Scientific, Waltham, MA) containing 10% FBS (Biological Industries), 2 mM L-glutamine (Biological Industries) and 1% pen-strep-ampho (Biological Industries). The cultivation conditions are described in Additional file
1: Table S1. Defined keratinocyte serum-free medium (KSFM; Thermo Scientific) was used during culture to limit the growth of cells other than keratinocytes. A region of the
PTGER2 gene that shows high homology between the human and mouse sequences was analyzed using species-specific PCR (primers listed in Additional file
1: Table S2) in order to confirm the origin of the cell lines [
22]. Stable cell subclones expressing luciferase were achieved by transfecting the pGL4.5 Luc plasmid (Promega, Madison, WI) into a given cell line and this was then followed by selection. All reagents were purchased from Sigma-Aldrich (St Louise, MO) unless specified otherwise.
Phenotypic assays
Cells were seeded into 96-well plates at a density of 2500 cells per well, and the cell proliferation over the following 4 days was then analyzed by MTT assay. The cell growth curves of the various cell lines were plotted in order to calculate the population doubling time of each cell line. The cell migration and invasion assays were performed using 0.8 μm 24-well Transwell chambers (Merck Millipore, Billerica, MA). For the migration assays, cells were collected and seeded into the upper chamber of a Transwell at a density of 1 × 10
5 cells per well. For the invasion assays, 50 μl of 10% Matrigel (BD Biosciences, San Jose, CA) was used to coat each Transwell chamber and then 2 × 10
5 cells were seeded onto the Matrigel-coated Transwell. After incubating at 37 °C for 48 h, the Transwell membranes were fixed and stained with 10 μg/ml Hoechst 33258 for 10 min. Images of the migrated or invading cells were captured using a fluorescence microscope. The cell numbers in each picture were counted and then normalized to allow the fold-change to be calculated. For colony formation assays, cells were seeded into 6-well plates at a density of 100 cells per well to allow for colony formation. Ten days later, the colonies were washed twice with PBS, fixed with ethanol and stained with 0.05% crystal violet. Colonies were counted by microscopy [
7,
17].
Western blot analysis
Western blot analysis followed the approach of previous protocols [
7,
17]. The primary antibodies are listed in Additional file
1: Table S3. Signals were revealed by Western Lightening Chemiluminescence Reagent Plus kit (Thermo Scientific) and detected using a FUJIFILM LAS-4000 mini luminescent image analyzer (GE Life Sciences, Piscataway, NJ).
RNA extraction and PCR analysis
Cells were collected in TriPure Isolation Reagent (Roche, Basel, Switzerland) and disrupted using 1-bromo-3-chloropropane. The lysate was collected and the upper aqueous phase, which contains the RNA, was moved. The aqueous layer was next mixed thoroughly with isopropanol and then spun at maximum speed to pellet the precipitated RNA. The resulting pellet was washed and then dissolved in nuclease-free water. The various miRNA TaqMan® assay (Thermo Scientific) probes used in this study are described in Additional file
1: Table S4. The threshold cycle (C
t) method was used. The resulting information was analyzed using the –ΔΔC
t method and values calculated relative to internal controls [
17]. A mutation of the
p53 gene was also analyzed using PCR analysis (primers listed in Additional file
1: Table S2).
Immunohistochemistry (IHC)
Formalin-fixed paraffin tissue samples were sliced into 5 μm sections, deparaffinized with xylene, rehydrated, and subjected to antigen retrieval solution. IHC was carried out using the UltraVision Quanto Detection System IHC staining kit (Thermo Scientific) by following the manufactured protocol. The anti-GFP antibody is described in Additional file
1: Table S3. The tissue sections exhibiting faint and/or scattered GFP immunoreactivity were scored +, while sections exhibiting dense and/or extensive GFP immunoreactivity were scored ++.
Xenograft transplantation
A total of 5 × 106 cells were injected into the flank of each non-obese diabetic CB17-Prkdcscid/NcrCrlBltw (NOD-SCID) mouse (National Laboratory Animal Center, Taipei, Taiwan). The cells used were mixed 1:1 with Matrigel (BD Biosciences), which gave a total volume of 200 μl. These mice were sacrificed during the 5th week. A total of 1 × 107 cells were injected into the flank of each C57BL/6 mouse (BioLASCO, Taipei, Taiwan). These cells were also mixed 1:1 with Matrigel (BD Biosciences) to give a total volume of 200 μl. However, these mice were sacrificed during the 2nd week.
For the orthotopic xenografts, 5 × 106 cells in a total volume of 100 μl were injected into the central portion of the tongue of C57BL/6 mice (BioLASCO). These mice were sacrificed between the 3rd and 6th week in order to detect the rate of growth of the primary tumor and the presence of neck metastasis. In addition, nude mice (National Laboratory Animal Center) were also used to assess the neck metastasis capability and this was done by injecting 5 × 106 of cells (in 100 μl volume) into the central portion of the tongue. These mice were sacrificed between the 7th and 9th week. The tongue tumors and the neck tissues from the mice were isolated to allow histopathological and IHC evaluation.
For the in vivo real-time tumor imaging, 5 × 106 cells were injected into the tongue central portion of C57BL/6 mice (BioLASCO). An IVIS Spectrum system (Caliper Life Science, Hopkinton, MA) was used for detecting tumor growth and this was done once a week. The mice were sacrificed during the 8th week. Tongue tumors and neck tissue were isolated for histopathological evaluation.
For the distant metastasis assay, 1 × 106 cells were injected intravenously into C57BL/6 mice (BioLASCO). These mice were sacrificed when body weight loss > 1/3 or between the 7th and the 9th week. An autopsy was performed on each mouse in order to evaluate the metastasis and the presence of tumor foci in each animal’s various organs.
To test the therapeutic efficacy of cisplatin (CDDP), cells that had been mixed 2:1 with Matrigel (BD Biosciences) to give a total volume of 150 μl were injected into C57BL/6 mice (BioLASCO) subcutaneously. When tumor volume had reached about 0.2 mm
3 during the 3rd week after tumor injection, 2 mg/kg CDDP dissolved in 0.9% saline was intraperitoneally injected every other day for four weeks. In parallel, the control group received an equal volume injection of 0.9% saline [
23,
24]. The tumor sizes were measured every week.
The primary tumors, the neck regions of the mice and the lungs of the mice were photographed using Illumatool Bright Light System (LT-9500; TLS, Sarasota, FL) in order to visualize the primary tumors and the various metastasis lesions that were exhibiting green fluorescence. All mice were sacrificed by CO2 inhalation at the end point of the experiment. All resected tissues were subjected to histopathological evaluation. The tumor volumes were calculated using the formula = 0.5xaxb2, the parameters having been measured at a micro-scale under light microscopy or using a gauge. For this calculation, a was the longest diameter and b was the shortest diameter. All animal studies were carried out followed the guideline of National Yang-Ming University Institutional Animal Care and Use Committee (IACUC).
Statistical analyses
The results are presented as means ± SE. Mann-Whitney tests, t-tests, two-way ANOVA tests, correlation analysis, and Kaplan-Meier survival analysis were used to compare the differences between the various groups of results. A p-value of less than 0.05 was considered statistically significant.
Discussion
To establish an immunocompetent mouse model is critical to defining pathogenesis, developing preclinical tests and investigating host defense during carcinogenesis. We have successfully established Tg mice lines that exogenously express
miR-31 and
miR-211 in basal keratinocytes. These mouse lines show increased susceptibility to 4NQO induction of tumors due to their defective DNA repair and their oxidative homeostasis [
7,
8,
18].
miR-211 has been reported to be an oncogenic miRNA in OSCC that targets TGF-ẞRII, TCF12 and BIN1, which affects the Myc pathway, antioxidative activation and EGFR/MAPK axis, respectively [
7,
8,
15]. Some of these regulatory cascades have been shown to be recapitulated during the oral carcinogenesis process using Tg models [
7,
8]. In this study, we established multiple cell lines from the OSCC induced by 4NQO in K14-EGFP-
miR-211 mice. Although these cell lines should have the same transgene background and had been exposed to equivalent doses of 4NQO during tumor induction, their tremendously genetic heterogeneity and their diverse oncogenic properties would seem to reflect the acquisition of different genetic insults during the establishment and progression of the OSCC; this simulates the diversity and complexity of human OSCC cell lines. In general, more than 80% of human
p53 mutations in malignancies are clustered between codons 125 and 300, which corresponds mainly with the DNA binding domain [
25]. In addition, more than 40% of HNSCC tumors carry a
p53 mutation. This study identifies three
p53 hotspot mutations involving guanine nucleotides within the four newly established cell lines. Some or all of the
p53 mutations are likely to have been acquired during the 4NQO treatment because 4NQO preferentially targets guanine and therefore the mutation would probably occur before the establishment of these three cell lines [
26].
The MOC-L4 cell line shows the weakest oncogenic properties, especially in terms of proliferation and colony formation, which is in agreement with the low expression levels of proliferation, EMT and stemness markers, together with the absence of a
p53 mutation and low level of
miR-196a expression. By way of contrast, MOC-L1 and MOC-L2 have the potent oncogenic ability, particularly in terms of colony formation and growth in the syngeneic and immune compromised mice lines. These two cell lines show a higher expression of stemness markers compared to MOC-L3 and MOC-L4. MOC-L1 is the most aggressive mouse cell line that we have established. It is characterized by the potent in vitro oncogenicity and in vivo induction of primary xenografts. Relative to MOC-L2, the orthotopic MOC-L1 xenografic tumors give rise to neck nodal metastasis; and on the tail vein injection of MOC-L1, there was extensive lung metastasis and rapid mortality of the syngeneic recipient mice. As a copious induction of EMT markers and stemness markers is present in this cell line, these factors are likely to underlie its high level of aggressiveness. Cancer cell dormancy is thought to be present in tumor progression, which suggests being as a growth arrest and an evolutionary mechanism in metastasis [
27]. Whether MOC-L2 is in a dormant state that requires additional factors to activate its metastasis machinery needs specification [
28].
Since all of our cell lines are tagged with GFP, this facilitates cell tracing in the lymph nodes and distal organs by fluorescence detection. Furthermore, scattered tumor cells that could be missed during a standard histopathological examination are able to be identified with more certainty by means of fluorescence imaging and/or immunostaining. The fact that this syngeneic mouse model will allow for the profiling of the immune microenvironment, which will help clarifying the disparities between immune-competent mice and immune-compromised mice in neck nodal metastasis. This, in turn, should help addressing the roles of the host defense in counteracting local metastasis, is intriguing. We also further engineered the MOC-L1 cell line to express Luc, which broadens its use in terms of real-time monitoring. Despite the fact that the cell line’s tumorigenic potential is reduced in this cell subclone, its orthotopic growth potential and its ability to produce neck/distal metastases are maintained in this cell subclone. Lingual xenografts that grow to a huge size are prone to induce neck metastasis, but such a primary tumor burden may cause a high mortality risk. As MOC-L1 cell xenografts resulted in high mortality in NOD-SCID mice, we confirmed the local metastasis competence of MOC-L1 cells in nude mice. However, due to the lack of native immune modulation, surveying of immune-escape and evaluation of treatment efficacy do not seem to work well in this system. Therefore, xenotransplantation into other oral sites that will interfere with swallowing to a lesser extent needs to be further developed.
The growth of the subcutaneous xenografts of the MOC-L1 cell line was repressed by CDDP therapy. Analysis indicated that there was a reduced
miR-196b expression in these tumors relative to untreated tumors.
miR-196b has been reported to be aberrantly expressed in several cancers [
6,
10,
13,
16,
29]. Overexpression of
miR-196b is commonly found in many digestive tract cancers that have a poor prognosis [
6,
29]. It is intriguing to note that there was a decrease in
miR-196b expression in the CDDP treated mice. These findings seem to suggest that
miR-196b could be a putative target of CDDP in the MOC-L1 cell line. Furthermore, since expression of
miR-196b is correlated with the weight of the xenografts, and since
miR-196b is known to be a circulatory biomarker that can be used for the early detection of OSCC [
13], the monitoring of
miR-196b using body fluids from this immune-competent mouse model could help contributing to marker validation [
11].
Although K14-EGFP-
miR-31 Tg mice also exhibited a high susceptibility for 4NQO-induced tongue carcinogenesis [
18], the attempts to establish a cell line from the tongue carcinomas of these mice have not succeeded hitherto. However, by crossing K14-EGFP-
miR-31 mice with K14-EGFP-
miR-211 mice, we have generated mouse lines with concordant
miR-31 and
miR-211 overexpression in squamous epithelium. The oral neoplasms spontaneously emerge from these mice would enable the establishment of new cell lines to specify the pathogenic interaction between
miR-31 and
miR-211 [
7,
18]. Methylation status,
ETS2 gene and homeobox genes are involved in the regulation of
miR-196b expression [
16,
30,
31]. Since
miR-211 is an oxidative stress inducer [
7], and that the oxidation states affect DNA methylation [
16], and the expression of
ETS2 and homeobox gene members [
32,
33], the potential connection between
miR-211 and
miR-196b in oral carcinogenesis through such factors remained to be addressed. Furthermore, simultaneous inhibition of both
miR-211 and
miR-196b to treat OSCC may be a strategy that will help to abrogate OSCC. The roles of
miR-211 in OSCC pathogenesis remain elusive. This study reports four new murine OSCC cell lines exhibiting
miR-211 overexpression. The mechanistic complex of
miR-211-target gene regulatory axis associating with other genetic alterations, drug responses and oncogenic behaviors can be further specified in these cell models.
The use of cancer organoid lines derived from primary human OSCC can help to predict the therapeutic responses [
34]. The usefulness of the syngeneic tumorigenesis models established in preclinical assessment has been defined in a preliminary way. This model should also help to facilitate the therapeutic trials of new drugs. Immunotherapies against OSCC can reduce the toxicities from the conventional radiotherapies and chemotherapies [
35]. Till far, most immunotherapeutic approaches on OSCC have been limited in patient studies, because appropriate mouse models that recapitulate the pathological and immunological diversity in human counterparts are lacking. The syngeneic tumorigenesis models established during this study should help to insight tumor immunity and to develop therapeutic strategies.