Porphyromonas gingivalis Promotes Immunoevasion of Oral Cancer by Protecting Cancer from Macrophage Attack

Oral cancer is the eighth most prevalent cancer worldwide, and more than 90% of oral cancers are oral squamous cell carcinoma (OSCC) (1, 2). According to the global cancer statistics, ∼354,864 new cases and 177,384 deaths from oral cavity cancer were estimated in 2018 worldwide (3), and the 5-y survival rates for individuals with oral cancer are <50% (4). In addition, OSCC has a high chance for recurrence and metastasis, and patients who survive the first occurrence of oral cancer are at a higher risk of developing a second primary oral cancer. OSCC recurrence rates can range between 32.7 and 44.9% (5, 6). Therefore, exploring OSCC pathogenesis is critical for ensuring better prevention and improving patients’ survival.

Accumulated evidence indicates that chronic periodontitis is a risk factor for oral cancer (7–9). Further, Porphyromonas gingivalis, a major opportunistic pathogen responsible for chronic periodontitis, may be an important mediator for the development of OSCC (10). Strong association between P. gingivalis and OSCC has been found in clinical investigations (11, 12). We collected more clinical OSCC samples, and the result showed that the Porphyromonas species was found in a higher proportion in tumor inner areas than on the tumor surface (Supplemental Fig. 1). Numerous studies have revealed the mechanisms by which P. gingivalis promotes oral carcinogenesis and cancer development; P. gingivalis was reported to inhibit programmed cell death or apoptosis in primary gingival epithelial cells (13–15), accelerate gingival epithelial cell proliferation (16, 17), promote metastatic dissemination and invasion of OSCC cells (18), and induce chronic inflammation, which is anticipated to be a potential pathway in bacteria contributing to carcinogenesis (19). However, these studies mainly focused on the direct effect of P. gingivalis on the activities of primary epithelial and OSCC cells in vitro. The immune system is responsible for mediating cancer development in immunocompetent hosts (20), whereas the initiation and perpetuation of cancer undergo the evasion of immune system (21). Few studies have focused on the role of P. gingivalis in oral cancer immunity.

Tumor microenvironment contains various immune cells, including macrophages, neutrophils, myeloid-derived suppressor cells, NK cells, dendritic cells, mast cells, and adaptive immune cells (T and B lymphocytes) (22). Among the diverse immune cells, macrophages play a major role in the recognition and clearance of cancer cells (23). Avoiding phagocytosis by tumor-associated macrophages (TAMs) is needed for the growth and metastasis of solid tumors (24), and programmed cell removal by macrophages can regulate cancer cell survival (25). Many studies have evaluated the role of P. gingivalis in the physiological activities of macrophages (26–28); however, whether these activities affect the ability of macrophages to attack cancer cells is seldom explored to date. Therefore, given the importance of macrophages in cancer immunity, we aimed at exploring the impact of P. gingivalis on macrophages mediating OSCC development.

P. gingivalis ATCC 33277 was anaerobically cultured in brain heart infusion broth (Difco, Sparks, MD) containing defibrinated blood of sheep (5%), hemin (0.5%), and vitamin K (0.1%) at 37°C (90% N₂, 5% H₂, 5% CO₂). Bacterial cells were precultured overnight and adjusted to a concentration of 2 × 10⁸ CFU/ml.

OSCC cell line of Cal-27 was cultured in high-glucose DMEM (HyClone, Logan, UT) containing 10% FBS (Life Technologies, Waltham, MA) and 1% penicillin-streptomycin antibiotic mixture (PS; HyClone). Bone marrow–derived macrophages from BALB/c mice aged 5–6 wk were cultured in IMDM (HyClone) containing 10% FBS and 1% PS. An M-CSF (PeproTech) was used for stimulating the differentiation of monocytes into macrophages. Cells were cultured in an incubator with 5% CO₂ and 95% air at 37°C.

Antibiotic-mediated inactivation and heat inactivation were used in the current study. P. gingivalis was incubated in IMDM containing 10% PS for 2 h at 37°C or incubated in PBS for 2 h at 80°C to induce cell death. The dead cells were centrifugated and resuspended in IMDM with 2% heat-inactivated FBS when needed.

In a flow cytometry tube, 2 × 10⁸ CFU of living or inactivated P. gingivalis was incubated with 2 × 10⁶ CFSE (Thermo Fisher Scientific), labeled Cal-27 cells, and 10 μg/ml anti-CD47 (B6H12; Abcam) at 37°C for 2 h. Following this, 2 × 10⁶ macrophages were added to the medium to phagocytose tumor cells. All cells were suspended in IMDM with 2% heat-inactivated FBS. After incubation for another 2 h, macrophages were stained with allophycocyanin anti-mouse F4/80 (BioLegend), according to the manufacturer’s protocol. Phagocytosis was evaluated by flow cytometry (Cytomics FC 500; Beckman Coulter) via the detection of CFSE⁺ macrophages.

In a 24-well plate, 2.5 × 10⁵ Cal-27 cells and 2.5 × 10⁵ macrophages were infected with inactivated P. gingivalis at a multiplicity of infection of 100:1 and suspended in DMEM containing 10% heat-inactivated FBS at 37°C. Anti-CD47 (10 μg/ml) was also added. After incubation overnight (∼16 h), cells at the bottom of the plate were photographed using an inverted microscope (ECLIPSE Ti; Nikon, Tokyo, Japan) and were then collected into a flow cytometry tube. The cells were successively stained with the CFSE dye and allophycocyanin anti-mouse F4/80. After dyeing, 1 × 10⁵ unlabeled Cal-27 cells were added to each tube. Phagocytosis was then evaluated by flow cytometry by determining the ratio of CFSE⁺ Cal-27 cells to unlabeled cells.

Twenty-four specific pathogen-free female NOD-SCID mice, aged 5 wk, were randomly divided into four groups (six mice per group). The mice were s.c. injected with 2 × 10⁶ Cal-27 cells suspended in 100 μl DMEM. After 1 wk, the mice were infected with 100 μl of antibiotics-inactivated P. gingivalis (OD₆₀₀ = 1) suspended in physiological saline near two tumor sites or were i.p. injected with 100 μg of anti-CD47 three times a week. The tumor volume was measured at the treatment time points of day 14 and 18. P. gingivalis was incubated in IMDM containing 10% PS overnight for the in vivo experiments in the current study.

Role of P. gingivalis on TAM polarization was also evaluated. For this, 5-wk-old female NOD-SCID mice were randomly divided into two groups (five mice per group). The mice were s.c. injected with 2 × 10⁶ GFP⁺ Cal-27 cells. Two weeks later, mice were infected with 100 μl of antibiotics-inactivated P. gingivalis (OD₆₀₀ = 1) near two different tumor sites three times a week. The animals were killed at the treatment time point of days 7 and 14 postinfection, and tumors were aseptically dissected and processed for TAMs polarization analysis. Briefly, the tumors were shred and incubated in medium 199 (HyClone) containing with 50 μg/ml liberase (Liberase TM; Roche Diagnostics) for 4 h at 37°C. Single cells were obtained using a 70-μm filter and transported into a flow cytometry tube. The cells were then stained with PE anti-mouse MHC class II (MHC II; Thermo Fisher Scientific), allophycocyanin anti-mouse F4/80 (BioLegend), and PE/Cy7 anti-mouse CD206 (BioLegend), according to the manufacturer’s protocol. GFP⁻ F4/80⁺ cells were gated as TAMs by flow cytometry, and the TAM polarization was evaluated by detecting MHC II and CD206.

All animal study procedures were approved by the ethics committee of West China School of Medicine, Sichuan University (2018105A), and all experiments were performed according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

In a flow cytometry tube, 2 × 10⁸ CFU of antibiotics-inactivated P. gingivalis was incubated with 2 × 10⁶ Cal-27 cells or macrophages for 2 and 16 h at 37°C. The cells were then collected by centrifugation at 1000 rpm for 5 min. Total RNA isolation was then performed following the instructions provided with TRIzol reagent (Invitrogen, Carlsbad, CA). RNA quality and concentration were measured using a NanoDrop 2000 (Thermo Fisher Scientific), and RNA integrity was assessed using a Bioanalyzer 2100 System (Agilent Technologies). rRNA was deleted using a Ribo-Zero rRNA Removal Kit (Illumina) and RNA fragmentation was performed using divalent cations in an Illumina proprietary fragmentation buffer.

Random oligonucleotides and SuperScript III were used to synthesize the first-strand cDNA, followed by DNA polymerase I and RNase H for the synthesis of the second-strand cDNA. The remaining overhangs were converted to blunt ends through exonuclease/polymerase treatment, and the enzymes were then removed. After adenylation of the 3′ ends of the DNA fragments, Illumina PE adapter oligonucleotides were ligated to prepare for hybridization. cDNA fragments of the preferred 300 bp length were purified using the AMPure XP System (Beckman Coulter). DNA fragments with ligated adaptor molecules on both ends were selectively enriched using Illumina PCR Primer Cocktail in a 15-cycle PCR. Products were purified (AMPure XP System) and quantified following the Agilent high-sensitivity DNA assay on a Bioanalyzer 2100 System (Agilent). Finally, the sequencing library was sequenced on a Hiseq platform (Illumina; Shanghai Personal Biotechnology).

Raw data (raw reads) were estimated in the FASTQ format. After filtration of low quality reads and reads containing adapters or poly-N, high-quality reads were aligned against the Cal-27 cell reference genome (http://asia.ensembl.org/Homo_sapiens/Info/Index) and macrophage reference genome (http://asia.ensembl.org/Mus_musculus/Info/Index) using Bowtie2 (http://bowtie-bio.sourceforge.net/index.shtml), and gene expressions were quantified using HTSeq 0.6.1p2 (http://htseq.readthedocs.io/). The values for reads per kilo base per million reads were calculated to normalize the expression values.

After RNA sequencing (RNA-seq), the differentially expressed genes were analyzed using DESeq (version 1.18.0) based on p values <0.05. Raw transcriptome reads reported in this paper have been deposited in the National Center for Biotechnology Information Sequence Read Archive under accession numbers PRJNA552947 (https://www.ncbi.nlm.nih.gov/sra/PRJNA552947) and PRJNA604716 (https://www.ncbi.nlm.nih.gov/sra/PRJNA604716).

To assess the reliability of RNA-seq, we performed quantitative RT-PCR (qRT-PCR) for several selected genes (Cal-27: cd47, il-1r2, and suprabasin; macrophages: il-1α, ccl-3, and ccl5). Cal-27 cells and macrophages were stimulated by antibiotics-inactivated P. gingivalis for 16 h at 37°C. Cells were then collected, and total RNA was isolated following the instructions provided with the TRIzol reagent (Invitrogen). RNA quality and purity were determined using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA). cDNA was then synthesized using a PrimeScript RT Reagent Kit with gDNA Eraser (Takara Clontech, Shiga, Japan), according to the manufacturer’s instructions. Quantitative real-time PCR was performed using LightCycler-480 (Roche Diagnostics, Bromma, Sweden) with TB Green Reagent (Takara, Dalian, China), following the manufacturer’s instructions. A 20-μl mixture of 10 μl of TB Green qPCR Mix (Takara), 1.6 μl of PCR primer mix, 2 μl of diluted template cDNA, and 6.4 μl of deionized distilled water was processed for RT-PCR. The relative fold changes (2^−ΔΔCt) were normalized to 18S expression. The PCR primers used in the study are listed in Table I.

Phagocytosis experiments were independently repeated thrice. Comparisons between groups during tumor volume analysis were made by ANOVA. Comparisons between groups were analyzed by unpaired t test for TAMs polarization analysis and qRT-PCR assay. The data are presented as mean ± SD, and the results were considered statistically significant at a p value < 0.05.

Macrophages play an important role in the clearance of foreign, aged, and transformed cells given its function of recognition and phagocytosis. We studied the effect of P. gingivalis on Cal-27 cell–phagocytizing ability of macrophages. For this purpose, we added anti-CD47 Ab to enhance phagocytosis for better observing the function of P. gingivalis in mediating phagocytosis of macrophages. The result showed that the proportion of CFSE⁺ macrophages was decreased with living P. gingivalis treatment for 2 h (7.6 versus 11.9%), suggesting a decrease in the proportion of macrophages that had phagocytized Cal-27 cells; this indicated an inhibitory effect of P. gingivalis on phagocytosis of Cal-27 cells by macrophages (Fig. 1A). To determine the factor that inhibited phagocytosis (the physiological activity or the membrane-component molecule of P. gingivalis), we then used inactivated P. gingivalis. We observed that antibiotics-inactivated P. gingivalis could also inhibit phagocytosis, and phagocytosis inhibition disappeared when P. gingivalis was heated (Fig. 1B), indicating that effector components in P. gingivalis would lose function once heated. Therefore, we speculated that membrane-component molecules, such as proteins, functioned to inhibit phagocytosis. To test whether phagocytosis inhibition was specific to P. gingivalis, we repeated the phagocytosis experiment with another periodontal pathogen, Fusobacterium nucleatum (ATCC 25586). However, F. nucleatum failed to inhibit phagocytosis of macrophages (Fig. 1C), suggesting the specific role of P. gingivalis in phagocytosis inhibition.

To test whether the phagocytosis inhibition by P. gingivalis was a short-term effect, we next cocultured P. gingivalis with Cal-27 cells and macrophages overnight (∼16 h). In this experiment (Fig. 2A), CFSE⁺ F4/80⁻ cells represented unphagocytized Cal-27 cells, and the ratio of unphagocytized Cal-27 cells (H4) to unlabeled cells (H3) reflected the relative amount of unphagocytized Cal-27 cells after an overnight incubation. This ratio was larger in the antibiotics-inactivated P. gingivalis group than in the control group (1.57 versus 0.48), which indicated that a higher proportion of unphagocytized Cal-27 cells remained in antibiotics-inactivated P. gingivalis group, suggesting the weakened ability of macrophages to phagocytize Cal-27 cells. Fig. 2B showed more cells remained in the bottom of the plate after an overnight incubation with antibiotics-inactivated P. gingivalis treatment. No obvious difference was found between the control and heat-inactivated groups. The result showed a trend similar to that of the phagocytosis assay for 2 h.

All phagocytosis experiments were independently repeated thrice (see supplementary material for the results of the three repeated experiments; Supplemental Figs. 2–5).

After demonstrating that P. gingivalis protected Cal-27 cells from macrophage attack, next we established a s.c. transplantation model of OSCC in mice and explored the role of P. gingivalis in tumor development in vivo. Based on the results from phagocytosis assay, antibiotics-inactivated P. gingivalis was used and sustainably injected around tumor sites. As expected, P. gingivalis–infected OSCC displayed accelerated growth compared with noninfected OSCC (Fig. 3A). Tumor volume was significantly larger after P. gingivalis induction at the treatment points of day 14 and 18. In the current study, the mice were also injected with anti-CD47 Ab, and tumor growth markedly retarded after anti-CD47 injection, irrespective of whether the mice were infected with P. gingivalis. These results suggested that P. gingivalis promoted in vivo OSCC growth, whereas anti-CD47 inhibited tumor growth.

TAMs are one of the most important components of tumor-infiltrating cells, and they play a major role in tumor immunosurveillance and immunoevasion (23). Macrophages can be classified into M1 and M2 types. M1 macrophages express high levels of MHC molecules and can mediate antitumor immune responses. In contrast, M2 macrophages express low levels of MHC II molecules and are considered to be tumor promotive (22). To explore the role of P. gingivalis in the functional polarization of TAMs, we analyzed the composition of TAMs in OSCC by flow cytometry. MHC II⁺ CD206⁻ cells were sorted as M1 macrophages, whereas MHC II⁻ CD206⁺ cells were sorted as M2 macrophages. The ratio of M2 to M1 macrophages was calculated (Fig. 3B, 3C), and it showed that the M2/M1 ratio was higher in the P. gingivalis group than in the control group at both time points, day 7 and 14, indicating that P. gingivalis could promote M2 TAMs polarization.

RNA-seq was performed to determine the possible molecular mechanism by which P. gingivalis promotes OSCC development. Cal-27 cells or macrophages were stimulated by antibiotics-inactivated P. gingivalis and the expression levels of several cytokines were evaluated between infected and noninfected cells. The transcriptome of Cal-27 cells was slightly affected by P. gingivalis treatment for 2 h (Fig. 4). Few protumor molecules were upregulated (IL-1R2, suprabasin, and TGF-α) and an antitumor cytokine was downregulated (TNF-α). When this stimulation time was prolonged to 16 h, a higher number of protumor molecules (suprabasin, IL-1R2, IL-18, and TGF-α) were upregulated, and a higher number of antitumor cytokines (TNF-β, IFN-β, TRAIL, and TNF-α) were downregulated. The transcriptome of macrophages was markedly affected by P. gingivalis at both time points of 2 and 16 h, with the upregulation of several protumor cytokines (IL-1α, IL-6, IL-10, IL-18, CCL-2, CCL-3, CCL4, CCL-5, and CCL7). To assess the reliability of RNA-seq, we performed qRT-PCR on several selected genes (Cal-27: cd47, il-1r2, and suprabasin; macrophages: il-1α, ccl-3, and ccl5). qRT-PCR showed that P. gingivalis–infected Cal-27 cells and macrophages exhibited higher mRNA levels of protumor molecules (cd47, il-1r2, suprabasin, il-1α, ccl-3, and ccl5) than uninfected cells (Fig. 5).

A close relationship between microbes and cancer has been reported in many previous studies, indicating that intratumoral microbes can affect cancer growth and metastasis through multiple ways (29). Abundance analysis from clinical samples has revealed that Porphyromonas spp. are found at higher levels in intratumoral areas. This result indicates the ability of Porphyromonas spp. to invade into tumor interiors, which may help P. gingivalis to mediate OSCC growth. Evading immune destruction is a hallmark feature of cancer (21); few studies have focused on the role of P. gingivalis in immunoevasion of oral cancer. In the current study, we identified a mechanism by which P. gingivalis facilitates immunoevasion of OSCC. We report that P. gingivalis inhibits macrophages from attacking Cal-27 cells via membrane-component molecules.

Accumulated studies have demonstrated a strong association between P. gingivalis and OSCC; P. gingivalis was reported to increase the tongue lesion size and multiplicity in a 4NQO-induced oral carcinoma model (30). However to date, the promotive effect of P. gingivalis on oral cancer growth has not, to our knowledge, been verified in vivo. Thus, we have found, for the first time to our knowledge, that antibiotics-inactivated P. gingivalis infection promoted oral tumor growth in mice. Our result was in line with the observations in the phagocytosis experiment, in which we found that P. gingivalis could inhibit macrophages from phagocytizing Cal-27 cells. Macrophages serve as a key mediator in cancer development because inhibition of TAMs attacking is required for the growth and metastasis of solid tumors (24); therefore, we detected the effect of P. gingivalis on TAM polarization. Of the two classifications of macrophages, M1 and M2, M1 macrophages express high levels of MHC II, secrete TNF-α, and participate in synthesis of reactive oxygen species and NO. Accordingly, M1 macrophages can efficiently recognize and destroy cancer cells. M2 macrophages, on the contrary, display protumor properties through angiogenesis, tissue remodeling, and suppression of adaptive immunity (31). Thus, the ratio of M2/M1 was evaluated in P. gingivalis–infected tumors, to indicate the promotive effect of P. gingivalis on M2 TAM polarization; this may be another reason for tumor growth acceleration by P. gingivalis infection.

To explore the possible molecular mechanism, we analyzed the transcriptome of Cal-27 cells and macrophages affected by P. gingivalis. According to the results from RNA-seq and qRT-PCR, protumor molecules (including suprabasin and IL-1R2 in Cal-27 cells and IL-1α, CCL-3, and CCL-5 in macrophages) were upregulated. Suprabasin has been identified as a tumor endothelial cell marker and may play a role in tumor invasion/metastasis and angiogenesis (32–34). Furthermore, overexpression of suprabasin has been associated with the proliferation and tumorigenicity of esophageal squamous cell carcinoma (35). IL-1R2 has been associated with a poor survival rate in colon cancer patients and is correlated with the migration ability of colon cancer cells. IL-1R2 acts with c-Fos to enhance the transcription of IL-6 and VEGF-A, which promotes tumor malignancy and angiogenesis (36, 37). Besides, regorafenib resistance in colon cancer cells is associated with enhanced expression levels of IL-1R2 (38). IL-1α plays roles in the progression of human cervical and breast cancers (39, 40), and is associated with distant metastasis in head and neck squamous cell carcinoma (41). Furthermore, CCL3 might stimulate cancer progression by promoting leukocyte accumulation, tumor growth, and angiogenesis (42). Blockade of CCL3 impairs cell invasion of OSCC and decreases the metastasis of lung cancer and multiple myeloma in mice (42–44). CCL5 can cause changes in the composition of leukocyte cell types at the tumor site by accumulating harmful TAMs (45). CCL5 also can promote tumor invasion, migration, and tumor resistance to host immunosurveillance (46, 47). CCL5 secreted by TAMs may be a novel target in the treatment of gastric cancer (48). Accordingly, the upregulation of these protumor cytokines may be the potential molecular mechanism by which P. gingivalis promotes the development of OSCC. However, the specific role of these cytokines in the mediation of OSCC by P. gingivalis needs to be confirmed in further studies.

Interestingly, we also found that tumor growth was markedly retarded by anti-CD47 Ab. CD47 is a nonhousekeeping cell surface protein expressed on almost all cancer cells. CD47 functions as a phagocytosis inhibitor by binding to the signal-regulatory protein-α (SIRPα) expressed on phagocytes such as macrophages (49). Therefore, in the in vitro phagocytosis experiment, we added anti-CD47 Ab to enhance the phagocytosis to better observe P. gingivalis–mediated phagocytosis. We further found the inhibitory effect of anti-CD47 Ab on OSCC growth in our in vivo study, even under the circumstances of tumor growth promotion by P. gingivalis. Weissman et al. (24) pointed out that CD47 might be a therapeutic target for human solid tumors, including ovarian, breast, colon, bladder, and hepatocellular carcinoma and glioblastoma. To our knowledge, our result provided new evidence for anti-CD47 inhibiting human solid tumors. Additionally, we observed that P. gingivalis treatment upregulated the expression level of cd47 in Cal-27 cells. This also can be a candidate mechanism by which P. gingivalis promotes OSCC development.

In summary, to our knowledge in this study, we revealed a nove mechanism by which P. gingivalis promotes OSCC progression. P. gingivalis inhibited the phagocytosis of Cal-27 cells by macrophages, induced functional polarization of macrophages into M2 TAMs, upregulated the expression of genes encoding for protumor molecules in macrophages and Cal-27 cells, and promoted the formation of an immunosuppressive tumor microenvironment, thus triggering immunoevasion of OSCC and promoting OSCC growth. Further studies are needed to identify the effector components in P. gingivalis.

We thank Min Hu for technical support of flow cytometry.

The authors have no financial conflicts of interest.