Mouse models

The animal studies were implemented with the approval of the Institutional Animal Care and Use Committee of the University of Pennsylvania. All animals used in this study were a syngeneic, immunocompetent C57BL/6J mouse strain (The Jackson Laboratory, US). To generate the mice with macrophage-lineage conditional knockout of RGS12 (cKO), Rgs12flox/flox(fl/fl) mice were mated with LysM-Cre mice.40

Oral cancer models

Eight-week-old LysM-Cre control and RGS12 cKO littermates (n = 10) were generated to induce oral cancer models. The carcinogen 4-Nitroquinoline 1-oxide (4NQO, MilliporeSigma, US) stock solution was made at a concentration of 5 mg/mL. The mice were provided drinking water with a final concentration of 100 µg/mL 4NQO for 8 or 16 weeks and further changed to clean water for additional 4 weeks to observe the tongue tumorigenesis.43

Histology

Mouse tongue tissues from each group were processed with an OCT embedding medium (Fisher Scientific, Hampton, NH, US) and kept at −80 °C. Frozen tongue tissues from LysM-Cre control and RGS12 cKO mice were sliced into 8-µm sections and performed a hematoxylin and eosin (H&E) stain to evaluate cancer pathology.

Immunohistochemistry (IHC)

Human oral cancer tissue slides (OR601b) were obtained from the Biomax company (Derwood, US). Immunohistochemistry (IHC) of human oral cancer tissues was deparaffinized and hydrated and subjected to antigen retrieval. Then sections were incubated with RGS12 antibody (1:500, MilliporeSigma HPA054646, US), followed by a standardized avidin/biotin polymer technology using IHC Kit (Santa Cruz Biotechnologies sc-398545, US) according to the instruction.

Immunofluorescence (IF)

Frozen tongue tissues or TAMs were fixed with 4% paraformaldehyde solution (MilliporeSigma, US) for 10 min at room temperature. Then, the sections were incubated with primary antibody [anti-acetylated α-Tubulin (Ac-Tub, 1:100, MilliporeSigma #T7451, US), anti-RGS12 (1:100, MilliporeSigma #GW21317, US), anti-KIF2A (1:200, Proteintech #13105-1-AP, US), or anti-MYCBP2 (1:100, MilliporeSigma #HPA058807, US)] and then incubated with the appropriate secondary antibody (1:500; Jackson ImmunoResearch Laboratory, US). A LAS-X (Leica) microscope was used to create maximum projections of z-stacks, and the fluorescence intensity, cilia length, and frequency were evaluated by ImageJ (NIH, US).38 All images were captured on a microscope with a comparable exposure time.

Cell culture

Bone marrow macrophages (BMMs) from 8-week-old LysM-Cre control and RGS12 cKO mice were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 20 ng·mL−1 recombinant mouse M-CSF, and 100 U/mL Penicillin-Streptomycin (Thermo Fisher Scientific, US). Tumor-associated macrophages (TAMs), T cells, and B cells were isolated from oral cancer tissues by indirect magnetic labeling according to the kit instruction (Thermo Fisher Scientific #11203D, US). Briefly, TAMs were incubated in phosphate-buffered saline (PBS) consisting of different antibodies (anti-F4/80 #60027, anti-CD4 #60017, anti-CD19 #6006, STEMCELL Technologies, Cambridge, MA, US) for 30 min. Then, the cells were washed with PBS 2 times and mixed with magnetic beads. Labeled cells were separated by magnetic rack and harvested by the addition of DMEM. Serum starvation (12 h) was performed for the cilia studies.

Plasmid construction and transfection

pCMV-RGS12 was created by cloning the mouse RGS12 open‐reading frame into p3×FLAG-Myc-CMV-26 (MilliporeSigma, US). shMYCBP2 (sc-155927) and shKif2A (sc-60884) were obtained from Santa Cruz (Dallas, TX, US). pCMV-KIF2A was obtained from Addgene (#52401, US). For cell transfection, TAMs from oral cancer tissues of WT mice were cultured in 6-well plates and transfected the indicated plasmids (3 µg) for 24 h using Lipofectamine 3000 reagent (Thermo Fisher Scientific, US). Subsequently, stably transfected cell clones were selected with appropriate antibiotics.

WST-1 proliferation assay

TAMs were cultured in 96-well plates and the cell proliferation assay was performed under kit instructions (MilliporeSigma, US). Briefly, after culturing the cells for the indicated time periods, the culture medium was changed to 90 µL growth medium supplemented with 10 µL WST-1 solution and incubated at 37 °C for 1 h. The absorbance was read at 450 nm and 490 nm using a microplate absorbance reader (Bio-Rad Laboratories, Hercules, CA, US).

Transwell assay

TAMs were harvested from LysM-Cre control and RGS12 cKO mice and cultured in the top chamber containing 8-µm-pore inserts in DMEM without FBS. DMEM supplemented with 10% FBS (750 μL) was added to each bottom chamber. The migrated TAMs were stained with 0.5% crystal violet (24 h) and counted by ImageJ software (NIH, US).

Immunoprecipitation (IP)

HEK293T cells or TAMs were lysed in NP40 buffer containing a protein inhibitor cocktail and phenylmethylsulfonyl fluoride (MilliporeSigma, US). In brief, lysates containing equal contents of proteins (1 mg) were incubated at room temperature with anti-RGS12 (Santa Cruz Biotechnologies #sc-514173, US) or anti-MYCBP2 antibody (MilliporeSigma #MABN2397, US) for 1 h and next with protein A/G magnetic beads (Bio-Rad Laboratories #161-4833, US) overnight. After washing with TBST buffer 3 times, the precipitated proteins were quantified by western blot assay.

Western blot (WB)

TAMs were lysed in radioimmunoprecipitation assay buffer with phenylmethylsulfonyl fluoride and protease inhibitor cocktail (MilliporeSigma, US). Proteins were separated by 10% SDS‐PAGE gel and transferred to 0.45 µm nitrocellulose (NC) membranes. The membranes were incubated with 5% skim milk for 1 h (room temperature) and next incubated with primary antibodies against RGS12 (1:1 000, MilliporeSigma #GW21317, US), KIF2A (1:1 000, Proteintech #13105-1-AP, US), MYCBP2 (1:1 000, Abcam #ab86078, UK), and β-Actin (1:5 000, Proteintech #66009-1-Ig, US) overnight (4 °C). The membranes were washed and incubated with secondary antibodies for 1 h. An enhanced chemiluminescent kit (Bio-Rad Laboratories, US) was used for imaging and quantitation by ImageJ software (NIH, US) after incubation with the secondary antibodies.

Real-time PCR

RNA extraction from TAMs or oral cancer tissues was implemented by using TRIzol reagent (Thermo Fisher Scientific, US). Reverse transcription was performed by using PrimeScript RT Master Mix (Takara Bio, US). Real-time PCR was implemented by CFX96 Touch System (Bio-Rad Laboratories, US) and 2× SYBR Green qPCR kit (Bimake, US). Relative expression was calculated by the 2−△△Ct method, with normalization to GAPDH expression. The primers were listed as follows: GAPDH (F): AGGTCGGTGTGAACGGATTTG, GAPDH (R): TGTAGACCATGTAGTTGAGGTCA; KIF2A (F): GCAGTGTTCCAGGAATCCAT, KIF2A (R): GCTTGCTTGCTTGCTCTTCT.

Data acquisition and processing

The RNA sequencing data of human healthy and oral cancer were extracted from the Gene Expression Omnibus database (GSE37991 or GSE39400). A total of 20 samples of RNA sequencing data were included, including 10 samples of oral cancer patients and 10 samples of healthy subjects. The data were normalized through standard correction and logarithmic operation and combined into a data matrix. The data were analyzed by the R package and the DEGs with P < 0.05 and fold change ≥1.5 were considered statistically significant.

Liquid chromatography–mass spectrometry (LC/MS)

The LC/MS experiment was introduced to compare the protein profiles of BMMs from LysM-Cre control and RGS12 cKO mice (8 weeks). A strict set of criteria was used for protein identification such as a low peptide and false discovery rate (FDR) of <0.05. Heat maps were created by the R package. The Database for Protein function (UniProtKB keywords), Annotation, Visualization, and Integrated Discovery (DAVID) was applied to analyze the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) enrichment.

Statistics

Statistical analyses were applied using GraphPad Prism 7. An unpaired 2-tailed t-test was used to compare differences between two groups. One-way ANOVA was applied to analyze changes among multiple groups. The correlation between the IF intensity counting was estimated by the coefficient of Person. The levels of Pearson correlations were described as no correlation (0), weakly (0.25), moderately (0.5), strongly (0.75), and perfectly positive correlation (1). Survival curves were evaluated with the Kaplan-Meier method from the Tumor Immune Estimation Resource (TIMER) platform. Data were presented as mean ± SEM, and P values less than 0.05 were considered significant.