NLRP3 inflammasome activation promotes inflammation-induced carcinogenesis in head and neck squamous cell carcinoma

SCCHN cell lines CAL27, SCC9, SCC25, and FaDu were purchased from the American Type Culture Collection and cultured according to the manufacturer’s online instructions. The immortalized oral keratinocyte line HIOEC from primary normal human oral epithelial cells infected with HPV16E6E7 (established in the Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine) was cultured in a defined keratinocyte serum-free medium (KSFM; GIBCO BRL, USA) [19]. LPS, ATP and MCC950 (PZ0280) were purchased from Sigma-Aldrich and the treatment of LPS, ATP and MCC950 were carried out as previously described [20]. The following antibodies were used in this study: Rabbit Polyclonal anti- NLRP3 (Atlas Antibodies AB, Sweden), Goat Polyclonal anti- ASC (GeneTex Inc., US); Rabbit Polyclonal anti- Caspase-1, Rabbit Monoclonal anti- IL-1β (Cell Signaling Technology, US); Rabbit Monoclonal anti- IL-18 (Abcam, UK); Rabbit Polyclonal anti- ALDH1 and Rabbit Polyclonal anti- BMI1 (GeneTex Inc., US); Mouse Monoclonal anti- CD44 (Cell Signaling Technology, US); Rabbit monoclonal anti- CD44(epitomics, UK).

The SCCHN tissue microarrays of humans used in this study were obtained from January 2008 and August 2014 in the Department of Oral and Maxillofacial Surgery, School and Hospital of Stomatology Wuhan University. The clinical stages of their SCCHN were classified according to the guidelines of the International Union Against Cancer (UICC 2002), and histological grading was determined according to the scheme of the World Health Organisation. Custom made tissue arrays of formalin-fixed tissues from SCCHN mentioned above were constructed with 1.5 mm core. These tissue microarray (T12–412-1 and T12–412-2) slides included 64 confirmed cases of SCCHN, 38 normal oral mucosa and 12 oral epithelial dysplasia. All slides were scanned at 400 magnification using an Aperio CS Scanscope (Aperio, CA) and quantified using the available Aperio algorithms. Written approval was obtained from all patients before the initiation of this study. All protocols dealing with the patients conformed to the ethical guidelines of the Helsinki Declaration and were approved by the Medical Ethics Committee of Hospital of Stomatology Wuhan University.

Time inducible tissue specific Tgfbr1/Pten 2cKO mice (K14-Cre ^ERtam; Tgfbr1 ^flox/flox; Pten ^flox/flox), Tgfbr1 cKO mice (K14-Cre ^ERtam; Tgfbr1 ^flox/flox), Pten cKO mice (K14-Cre^ERtam; Pten ^flox/flox) were maintained and genotyped according to published protocols [21]. The tamoxifen treatment procedure has been previously described [21]. Tgfbr1 and Pten knockout mice were fully penetrated and developed oral and head neck carcinoma in 3–6 weeks. Tgfbr1/Pten 2cKO mice were baseline induced with 2 mg of tamoxifen for five consecutive days to delete Tgfbr1 and Pten. 14 days after gavage, 10 mg/kg MCC950 was administered intraperitoneally every day for the first 3 days and every other day for 20 consecutive days (Fig. 3a). Tumor growth was assessed every day after tamoxifen gavage. All the mice were maintained in FVBN/CD1/129/C57 mixed background. All animal studies were approved and supervised by the Animal Care and Use Committee of Wuhan University, which were carried out according to the Institutional Guidelines for the use of laboratory animals in specific pathogen free Animal Laboratory of Wuhan University School and Hospital of Stomatology.

For NLRP3 inflammasome stimulation, cells were primed with LPS for 6 h (10 ng/mL, Sigma-Aldrich). Then medium was replaced with serum free medium (SFM) containing DMSO (1:1000), MCC950 (10, 25, 50 nM) for 1 h, and subsequently incubated with adenosine ATP for 1 h (5 nM, Sigma-Aldrich) before collecting the supernatants [20, 22]. The concentrations of IL-1β in the culture supernatants were analyzed using commercially available IL-1β Enzyme-linked immunosorbent assay (ELISA, Dakewe BioTech, China) kits according to the manufacturer’s instructions.

For the sphere-forming assay, single-cell suspensions were resuspended in culture media with 1% N2 supplement (Gibco), 10 ng/mL bFGF (Invitrogen), and 10 ng/mL EGF (Gibco) and plated in ultra-low attachment plates as our previously reported [23]. Irnverted microscope and 96 well plate (Corning, US) were used to count the tumor sphere number. For colony formation assay, single-cell suspensions were seeded in 6 well plate (NEST Biotechnology Co. LTD.) at a density of 500 cells/well with 10% FBS contained-medium. After 7 days incubation, the colonies were fixed with 4% paraformaldehyde and stained by crystal violet. The numbers of colonies were counted. Each assay was performed in triplicate.

Tumors from Tgfbr1/Pten 2cKO mice were dissected and fixed as previously described [21], and slides were stained with the appropriate antibody using a standard immunohistochemical staining protocol as previously described [24, 25]. The immunohistochemical staining was scanned using an Aperio ScanScope CS whole slice scanner (Vista, CA, USA) with background subtraction. The positive result was quantified using Aperio Quantification software for membrane, nuclear, or pixel quantification and histoscore were calculated using formula (3+) × 3 + (2+) × 2 + (1+) × 1 as previous described [26].

Cells were seeded on a cover glass slide chamber (Millipore, USA). After fixing with 4% paraformaldehyde at room temperature for 15 min, cells were treated with 0.5% triton X-100 and blocked with 2.5% BSA at room temperature for 1 h, and then incubated with primary antibody mentioned above overnight at 4 °C. Cells were then incubated with secondary fluorescent antibodies (DyLight 488 anti-rabbit, DyLight 594 anti-rabbit; Thermo Scientific, USA) with DAPI (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) for 1 h in the dark at room temperature. The slides were observed by a confocal laser scanning microscope (FV300, Olympus Life Science).

The Western blotting analysis was conducted as previously described [27]. Briefly, cultured cells, tumors and normal mucosa from the buccal mucosa and tongue were collected from mice, then the protein lysates were generated using M-PER or RIPA reagent (Pierce, Rockford, IL) containing a complete mini protease inhibitor cocktail and phosphate inhibitors (Roche, Branchburg, NJ). After denaturation the total protein was separated using 10% SDS–polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride membranes (Millipore). The blots were then blocked with 5% non-fat dry milk at room temperature for 1 h, and incubated overnight with the corresponding primary antibodies at dilutions recommended by the suppliers at 4 °C, finally by incubation with horseradish peroxidase-conjugated secondary antibody (Pierce, Rockford, IL). Next, the blots were detected using an enhanced chemiluminescence detection kit (West Pico, Thermo). GAPDH was detected on the same membrane and used as a loading control.

Statistical data analysis was performed with GraphPad Prism 6 (GraphPad Software, Inc., La Jolla, CA) statistical packages. We analyzed the data between 2 experimental groups using unpaired t test and between multiple groups using a one-way ANOVA test. Overall survival curves were estimated by the Kaplan–Meier method and compared by the log-rank test. All data were presented as mean ± SEM, statistical significance was defined as the p-value was <0.05. *P < 0.05; ** P < 0.01; *** P < 0.001; ns, not significant.

To determine whether NLRP3 inflamasome was associated with human HNSCC, we examined their expression by immunohistochemistry and found that NLRP3, Caspase-1, ASC, IL-18 were highly expressed in the cytoplasm of cancer cells, and the expression of them was negative or low in the normal mucosa (Fig. 1a). BMI1, ALDH1 and CD44 have been proposed as putative markers for the identification and isolation of CSCs in head and neck cancer [28, 29]. Our results showed that the protein expression of ALDH1 was positively stained in the cytoplasm, whereas BMI1 was in the cytoplasm and nucleus in SCCHN (Fig. 1a); CD44 was mainly expressed in the membrane of cancer cells but also in the basal layer of normal mucosa (Fig. 1a). These data indicated that BMI1, ALDH1 and CD44 were upregulated in the SCCHN tissues, but the correlation between NLRP3 inflammasome and CSCs remains unclear.

Increasing body of evidence from studies on various tumor types showed that chronic inflammation may regulate CSCs and enhance the development and function of the cancer cells [30]. However, the relationship between NLRP3 inflammasome and CSCs has not been investigated in human SCCHN. Interestingly, that data showed that the expression of BMI1 was statistically associated with NLRP3 (P < 0.001, r = 0.3169), Caspase-1 (P < 0.001, r = 0.3365) and IL-18 (P < 0.001, r = 0.3447) but not with ASC (P > 0.05, r = 0.1705); ALDH1 was statistically associated with NLRP3 (P < 0.001, r = 0.6613), Caspase-1 (P < 0.001, r = 0.6934), ASC (P < 0.001, r = 0.4821) and IL-18 (P < 0.001, r = 0.6632), and CD44 was statistically associated with NLRP3 (P < 0.001, r = 0.4206), Caspase-1 (P < 0.001, r = 0.5161), ASC (P < 0.001, r = 0.4615) and IL-18 (P < 0.001, r = 0.6014) as shown by the results of the Pearson correlation coefficient test (Fig. 1b-d). Results from hierarchy clustering and linear regression analyses showed that the expression of tumor-associated ASC and IL-18 are closer to ALDH1 expression, NLRP3 and Caspase-1 are close to ALDH1 expression, and NLRP3, Caspase-1, ASC, and IL-18 are close to CD44 and BMI1 expression (Fig. 1e).

NLRP3 inflammasome is a key player in the progression of cancers, but its role in tumorigenesis and tumor environment are complex. A spontaneous de novo SCCHN mice model for tumorigenesis studies was applied to further determine the potential role of NLRP3 inflammasome in tumor initiation effect. First we immunostained the proteins of the NLRP3 inflammasome components to investigate whether NLRP3 inflammasome was activated in Tgfbr1/Pten 2cKO mice. Our results revealed intense staining of NLRP3, ASC, Caspase-1, and IL-18 expression in Tgfbr1/Pten 2cKO SCCHN mice tumor lysates compared with control wild type mice tongues (Fig. 2a), these results indicated that the NLRP3 inflammasome was activated, and very high level of NLRP3 inflammasome was found in Tgfbr1/Pten 2cKO SCCHN mouse model. Moreover, we found the expression of BMI1, ALDH1 and CD44 was also upregulated in the Tgfbr1/Pten 2cKO SCCHN mice tumor (Fig. 2a). Additionally, the Pearson correlation coefficient test results shown that BMI1 expression was closely related to NLRP3 inflammasome in SCCHN mice model (NLRP3, P < 0.01, r = 0.5791; Caspase-1, P < 0.05, r = 0.4699; ASC, P < 0.001, r = 0.6401; IL-18 P < 0.001, r = 0.6311; Fig. 2b), ALDH1 was the same (NLRP3, P < 0.01, r = 0.6656; Caspase-1, P < 0.05, r = 0.4307; ASC, P < 0.001, r = 0.6793; IL-18 P < 0.01, r = 0.6075; Fig. 2c); and CD44 was also correlated with NLRP3 inflammasome (NLRP3, P < 0.01, r = 0.5334; Caspase-1, P < 0.05, r = 0.421; ASC, P < 0.05, r = 0.5147; IL-18 P < 0.01, r = 0.612; Fig. 2d). In addition, our western blot results showed that NLRP3, Caspase-1, IL-18 and IL-1β were highly expressed in the Tgfbr1/Pten 2cKO mice SCCHN tumor lysates compared with control wild type tongues, and the expression of BMI1, ALDH1 and CD44 were consistent with NLRP3 inflammasome (Fig. 2e and f).

We previously reported that the loss of Tgfbr1 and Pten results in cellular senescence evasion, cancer-related inflammation, and expansion of the CSCs in the basilar epithelial layer in SCCHN of the Tgfbr1/Pten 2cKO mice. The inflammation-related marker STAT3, CXCL1 [31, 32], and CSCs-related markers CD44, CD133, Sox2, and Oct4 were overexpressed within Tgfbr1/Pten 2cKO mouse SCCHN model [32‐34]. We determined whether MCC950 treatment could modulate the inflammation environment in 2cKO SCCHN model. A specific small molecule inhibitor of NLRP3 inflammasome MCC950 was applied (Fig. 3a). We found that MCC950 treatment (n = 6) significantly reduced tumor burden of the head–neck (Fig. 3b) compared with control group (n = 6). Moreover, MCC950 treatment caused no weight loss compared with the PBS treatment group (Fig. 3c). In addition, the tumor growth curve of the head and neck demonstrated that MCC950 treatment effectively reduced the tumor burden (Fig. 3d).

To evaluated the potential role of inflammation on CSCs markers during SCCHN progression and the use of the novel anti-inflammatory NLRP3 inflammasome inhibitor MCC950 arising de novo in 2cKO mouse SCCHN model. Our results from the Western blot analysis showed that the expression of NLRP3, Caspase-1, IL-1β, IL-18, BMI1, ALDH1 and CD44 was also significantly downregulated by MCC950 treatment (Fig. 4a). Moreover, the immunohistochemical staining analysis showed that MCC950 treatment remarkably reduced the expression of NLRP3, Caspase-1, ASC, IL-18, BMI1, ALDH1 and CD44 (*P < 0.05; **P < 0.01, Fig. 4b and c) compared with the control group. Therefore, NLRP3 inflammasome blockade could delay tumor initiation and progression by modulating the inflammation environment and regulating the population of CSCs in Tgfbr1/Pten 2cKO mouse SCCHN model.

Cumulative studies demonstrated that inflammasomes, including NLRP3, could be activated by microbial pathogens and cancer [20]. Two signals are required for the NLRP3 inflammasome formation and activation. One of these signals is the danger-associated molecular pattern (DAMP) including lipopolysaccharide (LPS), LPS primes the expression of NLRP3, pro-IL-1β, and pro-IL-18. The other stimulus, such as adenosine 5′-triphosphate (ATP), activates the NLRP3 inflammasome [7]. Hence, LPS and ATP was used in combination as NLRP3 inflammasome activator in our study. We detected stronger levels of NLRP3, Caspase-1, pro-IL-1β, and IL-18 in CAL27, FaDu, SCC9, and SCC25 than in the control cell (Fig. 5a). Hence, we selected CAL27 and SCC25 cell lines with high expression of NLRP3, Caspase-1, and pro-IL-1β for the following in vitro functional assay. As we can see, the level of both pro-IL-1β and IL-1β were obviously elevated after the stimulation of LPS and ATP in the CAL27 and SCC25 cell lines as shown by the results from ELISA and Western blot analysis (Fig. 5b and c). Thus, NLRP3 inflammasome was obviously activated by LPS and ATP stimulation.

Sphere-forming assay and colony formation assay were performed to investigate the potential role of NLRP3 inflammasome in the self-renewal capacity of cancer cells. MCC950 was used to further investigate whether CSCs could be reduced by blockade of NLRP3 inflammasome in SCCHN cell lines. Contrary to other NLRP3 inflammasome inhibitors, MCC950 blocks ASC oligomerization to inhibit canonical and non-canonical NLRP3 inflammasome activation rather than by reducing NLRP3 protein expression [35]. First, the level of IL-1β was obviously downregulated after stimulation with LPS and ATP and treatment with MCC950 at 10, 25, and 50 nM by ELISA in the CAL27 and SCC25 cell lines (Fig. 5c). LPS- and ATP-treated CAL27 cells were found to form more spheres and colonies than the control, which indicates LPS and ATP possible promotion of self-renewal of SCCHN cells in vitro (Fig. 5d and e). MCC950-treated group after LPS and ATP stimulation significantly showed lower number of spheres and colonies than in the LPS- and ATP-treated groups (Fig. 5d and e). BMI1, ALDH1, CD44 and the expression of NLRP3 inflammasome were also analyzed by Western blot to further confirm the effect of MCC950 on stemness. The results show that the expression of NLRP3, pro-Caspase-1(P50), and pro-IL-1β were not obviously downregulated by MCC950 after LPS and ATP treatment, because MCC950 treatment could alter the expression of IL-1β and activated form of Caspase-1(P20) but not that of NLRP3, pro-IL-1β [20]. Additionally, BMI1, ALDH1 and CD44 expression was significantly downregulated after MCC950 treatment (Fig. 5f). we performed immunofluorescence to further verify whether the inhibition of self-renewal capacity by MCC950 was through those CSCs related makers. The positive expression of BMI1, ALDH1 and CD44 was significantly upregulated by LPS and ATP treatment but downregulated by MCC950 treatment in dose-dependent compared with the control group (Fig. 5g). Thus, NLRP3 inflammasome activation could promote self-renewal capacity along with upregulated expression of stemness markers. Moreover, NLRP3 inflammasome blockade could reduce self-renewal capacity concomitant with downregulated expression of stemness markers in SCCHN cell lines.