Analysis of the effect of CCR7 on the microenvironment of mouse oral squamous cell carcinoma by single-cell RNA sequencing technology

C57BL/6-derived mouse oral cancer cells (MOC-1 and MOC-2) were purchased from Kerafast (USA) and cultured in HyClone Iscove’s modified Dulbecco’s medium (IMDM)/HyClone Ham’s Nutrient Mixture F12 at a 2:1 mixture with 5% FCS (Fisher, Scientific, Houston, TX), 1% penicillin/streptomycin, 1% amphotericin, 5 ng/ml epidermal growth factor (EGF, Millipore, Billerica, MA), 5 μg/ml insulin (Sigma, Chemical, ST. Louis, MO), and 400 ng/ml hydrocortisone as described previously [24]. PCI-37B and PCI-4B (human head and neck squamous cell carcinoma cell line) cells were donated by the University of Pittsburgh (USA) and cultured in DMEM with 10% FBS and 1% penicillin/streptomycin at 37 °C. THP-1 (human leukemia monocytic cell line) cells were purchased from Shanghai Cell Collection of Chinese Academy of Sciences and cultured in RPMI 1640 with 10% FBS, 1% penicillin/streptomycin and 0.05 mM β-mercaptoethanol at 37 °C.

Induction of the human macrophage line THP-1 was performed as previously reported [25, 26]. Briefly, THP-1 cells were cultured with PMA (100 ng/ml) for 24 h and then in RPMI 1640 complete culture medium for 48 h to induce M0 macrophages. M0 macrophages were cultured with IL-4 and IL-13 (20 ng/ml) for 48 h to induce M2 macrophages.

CCR7 knockout mice were purchased from Cyagen Biosciences (Suzhou, China). Because OSCC has the same growth trend between the flank and oral cavity of mice and the flank tumor is easier to measure, previous research generally used the mouse flank to construct an OSCC tumor model [27]. MOC-1 cells (1.0 × 10⁵) or MOC-2 cells (1.0 × 10⁶) in 100 µl of PBS were implanted subcutaneously into CCR7^−/− (KO) and wild-type (WT) mouse flanks. The length and width of the tumors were measured by using digital calipers every two days starting from the sixth day after the injection of MOC-1 or MOC-2 cells, and the tumor volume was calculated by (length × width²)/2. Then, the mice were sacrificed when the tumor volume reached 1500 mm³, and the tumor tissues were used for subsequent single-cell sequencing, immunofluorescence and flow cytometry analysis.

According to the manufacturer’s instructions, single-cell RNA sequencing libraries were constructed using a Single Cell 3’ Library and Gel Bead Kit V3 (10 × Genomics, 1,000,075, Capital Bio Technology, Beijing, China). The cells were clustered by Seurat 3.0 (R package). Dimensionality reduction was performed using PCA, and visualization was realized by t-SNE. GO, KEGG and Reactome enrichment were performed using KOBAS software with Benjamini‒Hochberg multiple testing adjustment according to the top 50 marker genes of each cluster. GSEA (Gene Set Enrichment Analysis) was performed by using GSEA software (version 2.2.2.4), which uses predefined gene sets from the Molecular Signatures Database (MSigDB v6.2) [28]. Gene set variation analysis (GSVA) was performed by using the GSVA R package based on the top 50 differential marker genes between the KO and WT groups. Gene sets came from the Molecular Signatures Database (MSigDB v6.2). Single-cell trajectories were built with the Monocle 2 R package that introduced pseudotime. SingleR (https://​bioconductor.​org/​packages/​devel/​bioc/​html/​SingleR.​html) was used to match the cell type of each single cell referring to the annotation of mouse cell types from Benayoun [29] and finally obtain the most likely cell type for each cell. Cell interaction analysis was performed using CellPhoneDB [30] between the KO and WT groups. Cell interactions were considered relevant if the p value of ligand–receptor pairs was less than 0.05.

Gene Express Profiling Interactive Analysis (GEPIA2) was used to examine the expression analysis and survival analysis of oral squamous cell carcinoma for the significantly different genes between the WT and KO groups in this work [31].

Mouse tumor tissues were fixed in 4% paraformaldehyde for 24 h before dehydration and paraffin embedding. Slice the tumor tissue at a thickness of 4 µm. After repairing the antigen in citrate buffer, the slices were incubated with 3% H₂O₂, washed with PBS, and then sealed in 10% goat serum. The cells were incubated overnight at 4 °C with the following antibodies: CD206 (1:800, 24595S, CST, USA) and F4/80 (1:500, NB600-404SS, Novus, USA). On the second day, after secondary antibody incubation, DAB staining and hematoxylin staining were performed, and the sections were observed under a microscope.

Briefly, fresh mouse tumor tissues were quick-frozen embedded with OCT, and then the tumor tissue was sliced at a thickness of 6 µm. Next, the tissue sections were blocked with the prepared antibody blocking buffer for one hour. Then, diluted antibody was added [F4/80 (1:100, NB600-404SS, Novus, USA; CD206 (1:800, 24595S, CST, USA); MKP-1 (1:500, sc-373841, SANTA CRUZ, USA)] and incubated overnight at 4 °C. After the antibody was removed, the cells were washed with PBS three times, and fluorescently labeled antibody was added [goat anti-rat IgG (1:100, 112–545-003, Jackson, USA); goat anti-mouse IgG (1:100, 115–165-003, Jackson, USA)] and incubated for two hours at room temperature. The cell nuclei were stained with DAPI for 5 min after washing with PBS. The sections were observed under a confocal laser scanning microscope (OLYMPUS FV3000, Japan) (800 ×).

Tumor dissociation was performed using a Keygen tissue dissociation kit (KG829, Keygen Biotech, China) according to the manufacturer's instructions. Cells were sorted by utilizing CD45 MicroBeads (130–052-301, Miltenyi Biotec, Germany) and a MACS cell separation system (Miltenyi Biotec, Germany). Nonspecific staining through Fc receptor binding was blocked by incubation with 50 μl of rat anti-mouse CD16/CD32 (553,141, BD, USA). The following murine-specific flow cytometry antibodies were used: CD11b-APC (1:50, 130–113-231, Miltenyi Biotec, Germany), MHC class II-FITC (1:20, 130–102-168, Miltenyi Biotec, Germany), and CD206-PE (1:100, Clone: MR6F3, eBioscience, USA).

Total RNA was extracted from the tumor tissues and cell lines using TRIzol Reagent (Takara, Tokyo, Japan) according to the manufacturer’s instructions. The mRNA quality was determined by A260/A280 (between 1.8 and 2.2) and A260/230 (> 1.7) ratios. Then, the RNA samples were reverse transcribed into cDNA using the PrimeScript™ RT reagent Kit (Takara, Tokyo, Japan), and real-time PCR was performed using TB Green® Premix Ex Taq™ II (Takara, Tokyo, Japan) according to the manufacturer's instructions. The primer sequences are shown in Table S1. The data were analyzed by the 2^−ΔΔCt method.

Cell proliferation was measured using a Cell Counting Kit-8 (CCK-8) (Dojindo Laboratories, Kumamoto, Japan) according to the manufacturer's instructions. The optical density was measured with a microplate reader (Bio-Rad, Hercules, CA, USA) at a wavelength of 450 nm.

According to the experimental grouping, 2.0 × 10⁶ cells were placed on a 6-well plate, and the cell status was observed the next day. Three cell scratches perpendicular to the labeled horizontal line were made in parallel within each well using a 200 µl pipette tip. The cells were washed with PBS and then replaced with macrophage supernatant from different treatment groups. Each group of cells was placed under an inverted microscope. At least 3 fields of view were selected for each group to take photos. ImageJ software was used to calculate the change in scratch area for each group of cells at different time points (0 h, 24 h, 48 h) and calculate the wound healing rate: (initial scratch area -24 h/48 h scratch area)/initial scratch area × 100%.

THP-1 cells (1.0 × 10⁶) were inoculated into the lower Transwell chamber, induced according to the experimental groups, washed and cultured in 700 µl RPMI 1640 medium without FBS. A total of 5.0 × 10⁴ OSCC cells (PCI-37B, PCI-4B) were inoculated into the upper chamber of the Transwell chamber with 300 µl of RPMI 1640 medium without FBS. After coculture for 24 h, the membrane was fixed and stained. For each group, 3 fields of view were randomly selected for photography under a microscope, and the number of migrating cells in each group was counted using ImageJ software. For the invasion experiment, 100 μl of diluted Matrigel solution was evenly spread on the basement membrane of the Transwell upper chamber, and the other steps were the same as above.

The statistical analysis was performed using R software (version 4.0.5) and GraphPad Prism (version 8.0.1). All experiments were repeated at least three times independently, and data are presented as the mean ± standard deviation (SD). The significance of differences between two groups was determined by Student’s t test. A P value < 0.05 was considered to indicate statistical significance.

The expression level of CCR7 in head and neck squamous cell carcinoma (HNSCC) was higher than that in normal tissue (Fig. 1a) but differed between individual cancer stages according to the TCGA database (Fig. 1b). And it is consistent with our previous research results [32]. To clarify the role of CCR7 in the growth of OSCC, MOC-1 and MOC-2 cells were injected subcutaneously into the flanks of KO and WT mice. The results indicated that MOC-1 and MOC-2 OSCC cells were successfully engrafted into both KO and WT mice (Fig. 1c). Based on the tumor volume data, we found that the growth rate of OSCC in WT group mice was higher, and the tumor volume reached approximately 1500 mm³ (1593.86 ± 273.37) on day 20 implanted with MOC-2 cells and 600 mm³ (640.52 ± 55.32) on day 24 implanted with MOC-1 cells; however, the growth rate of OSCC in KO group mice was lower, and the tumor volume reached approximately 600 mm³ (640.46 ± 189.88) on day 20 implanted with MOC-2 cells and 200 mm³ (231.30 ± 24.67) on day 24 implanted with MOC-1 cells (Fig. 1d, e). These results indicated that CCR7 gene knockout can inhibit tumor growth and decrease the tumor burden of mice with OSCC. The immunohistochemical results for CCR7 indicate that CCR7 is expressed in tumor cells of the KO and WT groups, but in stroma cells, the expression level of CCR7 in the KO group was significantly lower than that in the WT group. (Fig. 1j).

Because the tumor cells injected into the mice were the same, we analyzed the difference in tumor growth between the KO and WT groups caused by changes in the OSCC microenvironment. Then, we performed scRNA-seq on the tumors obtained from the WT groups (B2, B9, B47) and KO groups (B30, B76, B77) (Fig. 1f, Fig. S1a, b). A total of 70,316 cells were obtained after preliminary quality control. Then, the Doublet Finder R package was used to exclude the interference of multiple cells [33], and 58,836 cells without multiple cell interference were used for subsequent analysis (Fig. S2). Data reduction and cell clustering visualization were analyzed by the Seurat R package, and 25 cell clusters were visualized by t-SNE (Fig. S1c, d). The preliminary cell clustering by Single R [29, 34] can match some immune cells, but fibroblasts and tumor cells cannot be distinguished. Then, the inferCNV R package was used to distinguish benign and malignant cells according to reports [35‐38] (Fig. 1g). Finally, eight clusters (1, 2, 5, 6, 8, 12, 14, and 17) were identified as malignant cells. The other 17 cell clusters were identified as benign cells, and then the final clustering annotation results were obtained (Fig. 1h). Eight main types of cells and their marker genes, including fibroblasts (Col1a2) [39, 40], macrophages (Arg1, Mrc1) [41], granulocytes (S100a8, S100a9) [42], T cells (Cd3g) [43], monocytes (Cd74, H2-eb1) [41, 44, 45], B cells (Cd79a, Cd79b) [46], endothelial cells (Cdh5, Pecam1) [47] and tumor cells, were finally identified (Fig. 1i). The proportion of each cell type among all cells in the KO and WT groups is shown in the supplementary materials (Table S2). Among these cells, monocytes were present in a higher proportion in the KO group than in the WT group (P = 0.046), but monocytes accounted for a small proportion of the total cells.

As mentioned above, eight malignant cell clusters (1, 2, 5, 6, 8, 12, 14, and 17) were identified as tumor cells. The Find Clusters R package was used to analyze the different genes of the cell subset. Fig. S3a shows the top 20 differentially expressed genes in the KO group. Subsequent functional enrichment analysis was based on the top 50 upregulated and top 50 downregulated differentially expressed genes in the KO group. KEGG enrichment analysis showed that “MAPK signaling pathways”, “cell adhesion molecules (CAMs)” and “ECM-receptor interaction” were enriched in the genes that were upregulated in the KO group compared with the WT group, and these signaling pathways were closely related to the proliferation and metastasis of OSCC [48‐50] (Fig. 2a); “ribosome”, “nod-like receptor signaling pathway” and “osteoclast differentiation” were enriched in genes downregulated in KO vs. WT tumor cells (Fig. 2b). GSEA results showed that “cell adhesion molecule (CAM)” and “cytokine‒cytokine receptor interaction signaling pathway” were enriched in genes upregulated between the KO group and the WT group (Fig. 2c, d, and the main enriched genes are shown in supplementary materials (Fig. S4a, b)). GSVA results showed that “alanine aspartate and glutamate metabolism”, “MAPK signaling pathway” and “ECM receptor interaction pathway” were enriched in the KO group, while “fructose and mannose metabolism”, “glycerophospholipid metabolism” and “fatty acid metabolism pathway” were enriched in the WT group (Fig. 2e). The functional enrichment analysis of tumor cells indicated that CCR7 gene knockout not only affected the enrichment of key signaling pathways in tumor cells but also affected the metabolic pathways of tumor cells.

To study the effect of CCR7 gene knockout on the interaction between various cells in OSCC, we performed cell communication analysis on various cells in the WT group and KO group via CellPhoneDB. We found that tumor cells had strong interactions with macrophages, endothelial cells, monocytes and T cells in both the KO and WT groups (Fig. 2f, i); however, the receptor‒ligand pairs between the KO and WT groups were different (Fig. 2g, j). Through CellPhoneDB analysis, we obtained the top 20 expressed receptor‒ligand pairs between different types of cells in the KO group (Fig. 2h) and WT group (Fig. 2k). SCENIC software was used to analyze the differences in transcription factor enrichment between the KO and WT groups as described previously [51], and the results showed that the top 5 transcription factors in the KO group were Grhl2, Grhl1, Maz, Foxo1 and Esrra; the top 5 transcription factors in the WT group were Nebfk2, Cr3bl2, E2f3, Atf5 and Creb3l1 (Fig. 2l, m).

To further dissect the mechanisms of tumor progression, trajectory analysis for tumor clusters was carried out using the Monocle 2 R package. Trajectory analysis diagrams were obtained in different clusters, different states and different samples according to the pseudotime (Fig. 3a-c). We selected the top 100 genes with the most significant differences and similar trends during the pseudotime process for clustering visualization. Figure 3d shows the GO functional enrichment in six clusters. Next, we conducted BEAM analysis on node one, which had a significant impact on cell trajectory. The results indicated that genes highly expressed in the pre-branch node were mainly enriched in “positive regulation of exit from mitosis” and “Protein‒RNA complex assembly”; genes enriched in “response to stimulus”, “negative regulation of endopeptidase activity”, “Wound healing” and “regulation of secretion” were highly expressed in the cell fate1 node; genes enriched in “cell junction organization”, “Extracellular matrix organization”, “regulation of epithelial cell migration” and “regulation of MAPK cascade” were highly expressed in the cell fate2 node (Fig. 3e). Dynamic changes in the top 5 representative genes (Bcam, Cdh1, Ifrd1, Mxd1, and Timp1) that determine the fate of tumor cells during the pseudotime process are shown in Fig. 3f. The GEPIA2 analysis results indicate that the top 5 key genes (Bcam, Cdh1, Ifrd1, Mxd1, and Timp1) based on the trajectory analysis of tumor cells can affect the survival of tumor patients (Fig. 3g).

As mentioned above, the marker gene expression of endothelial cells, monocytes, B cells, granulocytes, T cells, fibroblasts and macrophages is shown in the supplementary materials (Fig. S5a-g, S6a-g). The top 20 differentially expressed genes in these seven stromal cell types in the KO group are shown in Fig. S3b-h. Subsequent functional enrichment analysis was based on the top 50 upregulated and top 50 downregulated differentially expressed genes in the KO group. The KEGG results showed the positively enriched signaling pathways in the KO group and the negatively enriched pathways in the WT group in endothelial cells (Fig. S7a), monocytes (Fig. S7b), B cells (Fig. S7c), granulocytes (Fig. S7d), T cells (Fig. S7e), fibroblasts (Fig. S7f) and macrophages (Fig. S7g). The GSVA results showed different enriched signaling pathways between the KO and WT groups in endothelial cells (Fig. S8), monocytes (Fig. S9), B cells (Fig. S10), granulocytes (Fig. S11), T cells (Fig. S12), fibroblasts (Fig. S13) and macrophages (Fig. S14). The GSEA results were as follows: in endothelial cells, genes in the “cell-cycle signaling pathway” were positively enriched in the KO and WT groups, and genes in the “VEGF signaling pathway” were negatively enriched in the KO and WT groups (Fig. S15a-d); in monocytes, the “Toll-like signaling pathway” and “nod-like signaling pathway” were positively enriched in the WT and KO groups (Fig. S16a-d); in B cells, “cytokine‒cytokine receptor interaction” and “ECM receptor interaction signaling pathway” were positively enriched in the KO but negatively enriched in the WT group (Fig. S17a-d); in granulocytes, the “nod-like receptor signaling pathway” and “Toll-like receptor signaling pathway” was positively enriched in the KO and WT groups (Fig. S18a-d); in T cells, the “chemokine signaling pathway” and “JAK-STAT signaling pathway” were positively enriched in the KO group and negatively enriched in the WT group (Fig. S19a-d); in fibroblasts, the “TGF-β signaling pathway” and “MAPK signaling pathway” were positively enriched in the KO group vs. the WT group (Fig. S20a-d); in macrophages, the “TGF-β signaling pathway” were negatively enriched in KO and WT groups and “chemokine receptor signaling pathway” were positively enriched in the KO group, but negatively enriched in the WT group (Fig. S21a-d).

We subsequently performed trajectory analysis in T cells, granulocytes, and macrophages, which accounted for a significant proportion of inflammatory cells based on scRNA-seq results. We obtained trajectory diagrams for different clusters, different samples and different states during pseudotime in T cells (Fig. S22a-c), granulocytes (Fig. S23a-c) and macrophages (Fig. S24a-c). As mentioned above, the top 100 genes with the most significant differences and similar trends during pseudotime were selected and divided into six clusters. For T cells, GO enrichment analysis results for six clusters are shown in Fig. S22d. The BEAM analysis results on node one indicated that “cell killing”, “negative regulation of viral transcription”, “humoral immune response”, and “negative regulation of angiogenesis” were enriched in the pre-branch node; “positive regulation of leukocyte activation”, “cellular process”, “leukocyte aggregation”, and “regulation of prostaglandin biosynthetic process” were enriched in the cell fate 2 node; and “regulation of inflammatory response”, “regulation of macrophage activation”, and “regulation of leukocyte activation” were enriched in the cell fate 1 node (Fig. S22e). Dynamic changes in the top 5 representative genes (Ccl5, Gzma, Gzmc, Gzd, Gzmf) that determine the fate of T cells during the pseudotime process are shown in Fig. S22f. In granulocytes, GO enrichment results in six clusters are shown in Fig. S23d. BEAM analysis results on node one indicated that “Cytokine production involved in immune response”, “Eosinophil chemotaxis”, and “regulation of TRAIL-activated apoptotic signaling pathway” were enriched in prebranch; “NF-κB signaling pathway”, “positive regulation of leukocyte migration”, “negative regulation of cell cycle G1/S phase transition”, and “negative regulation of cysteine-type endopeptidase activity” were enriched in the cell fate 2 node; and “MAPK signaling pathway”, “positive regulation of angiogenesis” and “cellular response to hypoxia” were enriched in the cell fate 1 node (Fig. S23e). Dynamic changes in the top 5 representative genes (Ccl3, Ccl4, Ccl6, Hk2, Slpi) that determine the fate of granulocytes during pseudotime are shown in Fig. S23f. For macrophages, Fig. S24d shows the GO enrichment result in six clusters. The BEAM analysis results on node one indicated that “neutrophil chemotaxis”, “positive regulation of MAPK cascade”, “negative regulation of catalytic activity”, and “negative regulation of metallopeptidase activity” were enriched in the pre-branch node; “tissue remodeling”, “extracellular matrix organization” and “skeletal system development” were enriched in cell fate 1; and “JAK-STAT signaling pathway”, “positive regulation of leukocyte activation”, and “positive regulation of tumor necrosis factor production” were enriched in the cell fate 2 node (Fig. S24e). The dynamic changes in the top 5 representative genes (Acp5, Cd52, Ckb, S100a4, S100a6) that determine the fate of macrophages during the pseudotime process are shown in Fig. S24f, and GEPIA2 analysis showed that changes in the top 5 key genes can affect the survival time of tumor patients (Fig. S24g). Next, we performed reclustering analysis on T cells and granulocytes. We obtained six cell subpopulations in T cells (Fig. S22g) and seven cell subpopulations in granulocytes (Fig. S23g). The cell number, gene number and UMI number of six T-cell subpopulations and seven granulocyte subpopulations are shown in Fig. S22h and Fig. S23h.

We performed reclustering analysis on macrophages in the KO group and WT group and obtained 9 clusters (Fig. 4a). The number of genes, cells and UMI in each cluster are shown in Fig. 4b-e. According to the marker gene expression of CD206 (also named Mrc1), clusters 2, 3 and 5 were defined as M2 macrophages (Fig. 4f, g). ScRNA-seq results showed that the proportion of M2 macrophages in inflammatory cells in the KO group was lower than that in the WT group (P < 0.05, Fig. 4h). Immunohistochemical analysis and immunofluorescence staining both demonstrated that there were fewer M2 macrophages (F4/80, CD206) in KO tumor tissue than in WT tumor tissue implanted with MOC-1 or MOC-2 cells (Fig. 4i-l). Flow cytometry analysis showed that the proportion of M2 cells/inflammatory cells in the KO group was significantly lower than that in the WT group (Fig. 4m, n). These results indicated that CCR7 gene knockout can reduce the infiltration of M2 macrophages in OSCC.

In the in vitro experiment, THP-1 cells were induced into M0 macrophages by PMA and then induced into M2 macrophages by IL-4 + IL-13 as described previously. The real-time PCR results showed that the expression levels of the M2 markers CD206, ARG-1, IL-10 and TGF-β in the M2 group were higher than those in the M0 group, indicating induction success (Fig. 5a-d). This induction was significantly inhibited after the addition of CCR7 mAb, indicating that CCR7 can block M2 polarization (Fig. 5a-d). The abilities of cell migration and invasion were increased significantly when PCI-4B and PCI-37B cells were co-cultured with M2 macrophages; however, when we pretreated M2 macrophages with CCR7 mAb, the migration and invasion abilities of tumor cells (PCI-4B, PCI-37B) were significantly decreased (Fig. 6a, b). Similar results were observed in the wound healing assay and CCK-8 assay, indicating that CCR7-induced M2 polarization can promote tumor cell survival, migration and invasion (Fig. 6c-f).

To study the mechanism of CCR7, we compared the gene changes in different cells between the KO and WT groups according to the scRNA-seq results. We found that in all stromal cells, including B cells, T cells, endothelial cells, granulocytes, macrophages, monocytes and fibroblasts, CCR7 knockout generally induced high Dusp1 mRNA expression (Fig. 7a-g). In an in vitro experiment, the real-time PCR results showed that CCR7 mAb can promote Dusp1 expression in the M2 macrophages we induced (Fig. 7h). Immunofluorescence staining also showed that MKP-1 (encoded by Dusp1 mRNA) was largely located in M2 macrophages and was highly expressed in the CCR7 knockout group (Fig. 7i, j). Since our previous studies demonstrated that Dusp1 gene deficiency can promote the growth of OSCC and enhance M2 macrophage polarization [52], we suggest that CCR7 promotion of M2 polarization is dependent on Dusp1 loss.