Article Text
PDF
Abstract
Background TP53, the most mutated gene in solid cancers, has a profound impact on most hallmarks of cancer. Somatic TP53 mutations occur in high frequencies in head and neck cancers, including oral squamous cell carcinoma (OSCC). Our study aims to understand the role of TP53 gain-of-function mutation in modulating the tumor immune microenvironment (TIME) in OSCC.
Methods Short hairpin RNA knockdown of mutant p53R172H in syngeneic oral tumors demonstrated changes in tumor growth between immunocompetent and immunodeficient mice. HTG EdgeSeq targeted messenger RNA sequencing was used to analyze cytokine and immune cell markers in tumors with inactivated mutant p53R172H. Flow cytometry and multiplex immunofluorescence (mIF) confirmed the role of mutant p53R172H in the TIME. The gene expression of patients with OSCC was analyzed by CIBERSORT and mIF was used to validate the immune landscape at the protein level.
Results Mutant p53R172H contributes to a cytokine transcriptome network that inhibits the infiltration of cytotoxic CD8+ T cells and promotes intratumoral recruitment of regulatory T cells and M2 macrophages. Moreover, p53R172H also regulates the spatial distribution of immunocyte populations, and their distribution between central and peripheral intratumoral locations. Interestingly, p53R172H-mutated tumors are infiltrated with CD8+ and CD4+ T cells expressing programmed cell death protein 1, and these tumors responded to immune checkpoint inhibitor and stimulator of interferon gene 1 agonist therapy. CIBERSORT analysis of human OSCC samples revealed associations between immune cell populations and the TP53R175H mutation, which paralleled the findings from our syngeneic mouse tumor model.
Conclusions These findings demonstrate that syngeneic tumors bearing the TP53R172H gain-of-function mutation modulate the TIME to evade tumor immunity, leading to tumor progression and decreased survival.
- immunotherapy
- cytokines
- tumor microenvironment
- head and neck neoplasms
- immune checkpoint inhibitors
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
p53 mutations, the most common genetic alterations in most cancers, occur in 75%–85% of non-HPV-associated head and neck squamous cell carcinoma (HNSCCs).
p53 mutations can disturb the infiltration of myeloid and T cells into the tumor microenvironment favoring immune suppression and promoting tumor development.
Although TP53R175H mutations have been studied in different tumor types, its role in shaping the tumor immune microenvironment (TIME) of HNSCC is poorly understood.
WHAT THIS STUDY ADDS
Mutant p53R172H contributes to elaboration of a cytokine transcriptome network that inhibits the infiltration of cytotoxic CD8+ T cells and promotes intratumoral recruitment of regulatory T cells and M2 macrophages.
Murine and human ortholog p53R172H/R175H mutation regulates the spatial distribution of immunocyte populations, and their distribution between central and peripheral intratumoral locations.
p53R172H-mutated tumors infiltrated with CD8+ and CD4+ T cells expressing programmed cell death protein 1, responded to immune checkpoint inhibitor and stimulator of interferon gene 1 agonist therapy.
Analysis of human oral squamous cell carcinoma (OSCC) revealed associations between immune cell populations and the TP53R175H mutation, which paralleled the findings from our syngeneic mouse tumor model.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study describes a functional role of the p53R172H/R175H gain-of-function mutation in shaping the TIME of OSCC using a syngeneic tumor model and analysis tissues from patients with OSCC.
This observation could be the basis of therapeutic approaches targeting p53 mutations or restoring p53 wild-type function to improve responsiveness to immunotherapy in head and neck tumors.
Introduction
Head and neck squamous cell carcinoma (HNSCC) is the most common malignancy of the upper aerodigestive tract. Oral squamous cell carcinoma (OSCC) is one of the most common forms of HNSCC and often presents in patients at an advanced stage with poor survival outcomes. Many OSCCs arise from oral premalignant lesions that have accumulated genomic alterations secondary to patient’s genetic predisposition to defective DNA damage repair and/or exposure to environmental carcinogens such as alcohol and tobacco.1–5 The 5-year survival rate of patients with advanced-stage OSCC has improved minimally over the last 30 years.6 Immune checkpoint inhibitors (ICIs) in advanced-stage OSCC have shown therapeutic benefit for a minority of patients7–10; however, the utility of ICI in OSCC may be limited, in part, to genomic alterations in tumor that contribute to an immunosuppressive tumor immune microenvironment (TIME). One such candidate alteration is TP53 mutation, which occurs in 75%–85% of HNSCC cases, including OSCC.11 While some p53 mutations lead to the loss of wild-type (WT) p53 function, many p53 mutations confer gain-of-function (GOF) activities, which promote tumor invasion, metastasis, genomic instability, cancer cell inflammation and cancer cell proliferation.11 Several recent studies have revealed that genomic alterations in p53 are also associated with immunosuppressive alterations in the TIME.12 13 p53 mutations in tumor cells can influence the TIME in different ways. For instance, p53 GOF mutations can modulate the activation of the cyclic GMP-AMP synthase (cGAS)-stimulator of interferon gene (STING) pathway, leading to impairment of the innate and adaptive immune responses.14 Furthermore, p53 missense mutations can inhibit the antigen presentation signaling pathway and modulate the expression of natural killer cell ligands and ICIs such as programmed death ligand 1. Moreover, p53 status can affect cytokine and chemokine secretion from cancer cells, thereby altering the TIME.15–17 Additionally, tumor cells expressing p53 GOF mutations can reprogram macrophages through exosome-based communication18 and decrease T-cell activation by activating the suppressive myeloid response.19 Interestingly, studies in mouse models of human pancreatic ductal adenocarcinoma have shown that p53 missense mutations are associated with increased tissue fibrosis and an immunosuppressive microenvironment.20 Remarkably, correlations between p53 mutation status and immune infiltration have been observed in different tumor types, including HNSCC.21–23 TP53 gene mutations are frequently found in exons 4 through 9, which encode the DNA-binding domain of the protein (residues R175, G245, R248, R249, R273 and R282) in many types of cancer.24 For example, in OSCC, TP53R175H is a hot-spot, high-risk, functional, acquired mutation that inhibits the function of WT p53 to a certain extent25–27; however, these correlations and functional roles of missense high-risk p53 mutations have not been experimentally addressed using syngeneic oral tumor models.
In the present study, we evaluated the effect of a clinically relevant missense p53R175H GOF mutation on the TIME of OSCC. Based on gene expression and multiplex immunofluorescence (mIF) studies on syngeneic murine carcinogen-associated oral cancers and human OSCC, we now report that missense p53 mutations (murine, R172H; human, R175H) are associated with increased infiltration of regulatory T cells (Tregs) and macrophage polarization toward the M2 (tumor-promoting) phenotype of tumor-associated macrophages (TAMs). In addition, mutant p53 modulates the expression of cytokine genes to promote immunosuppression, and affects the cGAS-STING pathway, impairing the innate immunity mediated by type I interferons (IFNs). Furthermore, these TP53 genomic alterations (R172H/R175H) increase the expression of immune checkpoint molecules and restrict the infiltration of cytotoxic CD8+ T cells. Therefore, targeting missense p53R175H mutations or restoring p53 WT function may improve responsiveness to immunotherapy in head and neck tumors.
Materials and methods
Cell culture
The murine oral cancer cell lines ROC2 and ROC328 were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 1 g/L D-glucose, supplemented with 10% fetal bovine serum, L-glutamine, sodium pyruvate, non-essential amino acids, vitamin solution and penicillin/streptomycin. Cell lines were routinely tested for Mycoplasma. The cell lines were maintained in a cell culture incubator at 37℃ with, 95% humidified air and 5% CO2 atmosphere.
Generation of p53 stable short hairpin RNA knockdown in ROC3 cells
We obtained p53 short hairpin RNA (shRNA) clones and non-targeting control (NTC) shRNA lentiviral glycerol stock clones from The University of Texas MD Anderson Cancer Center’s Functional Genomics Core Facility. The p53 shRNA clone (v3lhs_646511) had the corresponding nucleotide sense sequence CACTACAAGTACATGTGTA. Glycerol stocks were plated on ampicillin Luria broth agar and incubated at 37°C overnight. Single bacterial colonies were grown in Luria broth for plasmid DNA purification using a Miniprep kit (Qiagen). Briefly, lentiviral plasmids were mixed with Lipofectamine 2000 (ThermoFisher Scientific) and transfected into 293 FT cells for 8 hours in serum-free medium. At 48 hours after transfection, the medium containing the virus were collected and centrifuged at 1200 rpm to remove cellular debris. The ROC3 oral cancer cell line expressing the hotspot GOF mutation p53R172H as previously described28 were grown at 70% confluence and infected with the virus-containing medium supplemented with polybrene (1 μg/mL) for 24 hours. After 48 hours, cells were plated at 50% confluence and selected with puromycin (2 µg/mL) in complete DMEM (ThermoFisher Scientific) for 7 days or until all non-infected control cells were dead. Next, ROC3 cells stably expressing p53 shRNA or NTC shRNA were enriched for green fluorescent protein expression using flow cytometry under sterile conditions to improve p53 knockdown.
In vivo orthotopic mouse model
Male and female RAG-1 knockout (strain # 002216) and C57BL/6 (strain # 000664) mice, 8–10 weeks old, were purchased from The Jackson Laboratory and housed in a pathogen-free animal facility at the MD Anderson Cancer Center, Houston, Texas, USA. Briefly, 5×105 ROC3 cells were implanted in mouse tongues, as previously reported.28 Tumor size was measured twice a week after treatment. Tumor volume was determined using the following formula: tumor volume=(length×width2)/2. Tumor-bearing mice were euthanized using CO2 according to the protocol 00000950-RN03 approved by the Institutional Animal Care and Use Committee. Tissues were embedded in paraffin and cut into 5 μm thicknesses.
Opal multiplex immunofluorescence analysis
Slides were prepared from formalin-fixed, paraffin-embedded murine tongue tumors. We performed mIF consecutive staining of a single slide using an Opal 7-Color Kit for Multiplex Immunohistochemistry (IHC) (Akoya Biosciences, Cat# NEL811001KT) to study the TIME. Sequential cycles of staining, image scanning and destaining of chromogenic substrates were performed on formalin-fixed, paraffin-embedded tissue samples. Visualization of the Opal 7-Color slides was performed using the Mantra or Vectra quantitative pathology imaging systems (Perkin Elmer). The system uses multispectral imaging for quantitative separation of many fluorophores and autofluorescent tissues. The primary antibodies and dilutions used for mIF are shown in online supplemental material and methods. The image processing and analysis were performed as follows. First, cell segmentation was performed by identifying the cell nuclei in the 4′,6-diamidino-2-phenylindole (DAPI) channel. The gray value of each channel was then characterized, and cell subsets were divided according to specific cell markers. Next, we calculated the distance and number between single cells using Visiopharm software. In units of single cells, the number of each cell subset among the tumor cells within 50 μm was determined and the distance to the nearest cell within 1 mm was calculated. Finally, we merged the multiplex IHC data covering the entire tumor area. Dimensionality reduction was performed using the R package umap.
Supplemental material
Flow cytometry of tumor-infiltrated immune cells
Murine tongue tumors were harvested after perfusion with phosphate-buffered saline. Collected tumors were digested with collagenase IV (1 mg/mL) in DMEM for 30 min at 37°C. Single-cell suspensions were passed through a 70 µm nylon mesh. Red blood cell lysis was performed for 1 min. The tumor cells were suspended in fluorescence-activated cell sorting buffer for surface staining. For lymphocyte sorting, tumor cells were layered on lymphocyte separation medium (Lonza). They were then centrifuged at room temperature for 20 min at 900× g, and mononuclear cells were harvested from the gradient interphase. Cell surface staining was conducted by incubating the cells with antibodies for 30 min on ice in the presence of mouse 2.4G2 monoclonal antibody (Tonbo, Cat# 70-0161 U100) to block the binding of Fc-gamma receptors. For FoxP3 and granzyme B staining, we used a Transcription Factor Staining Buffer Set (Cat# 00-5523-00, Invitrogen). To assess cytokine production, T cells were stimulated with phorbol 12-myristate 13-acetate (50 ng/mL; Sigma, Cat# P148), ionomycin (500 ng/mL; Sigma, Cat# 407950), GolgiStop (1 µL/mL; BD Biosciences, Cat# 554724) or GolgiPlug (1 µL/mL; BD Biosciences, Cat# 555029) for 4 hours at 37°C. Subsequently, the T cells were stained for cell surface markers prior to intracellular cytokine staining. All the antibodies used are listed in online supplemental table S1. All data were acquired using an LSRFortessa X-20 Cell Analyzer (BD Biosciences) and analyzed with FlowJo software (Tree Star). The gating strategies are shown in online supplemental figures 1 and 2.
Gene expression-based profiling of tumor-infiltrating immune cells with CIBERSORT
HTSeq-FPKM-UQ data for The Cancer Genome Atlas head and neck cancer cohort (including 433 human papillomavirus-negative patient samples) were downloaded from the Cancer Genome Atlas website (https://cancergenome.nih.gov/). In total, 141 patients with WT p53 and 11 with p53R175H mutations were included in the analysis. We used CIBERSORT, a gene expression-based deconvolution algorithm for quantifying cell fractions from gene expression profiles in bulk tissue,29 to infer the relative proportions of 22 human immune cell types from admixed expression data. The fragments per kilobase of exon-per-million mapped fragment data from all available head and neck cancer samples were uploaded to the CIBERSORT website (https://cibersort.stanford.edu/index.php) as the mixture expression. A validated leukocyte gene signature matrix (LM22) with the characteristics of the 547 marker genes was used as the reference signature gene file. We implemented 1000 permutations to quantify the relative proportions of the 22 immune cell types.29 30 After obtaining the proportions of each cell type, the total number of T cells was calculated as the sum of the numbers of CD8− T cells, CD4+-naïve T cells, CD4+ memory resting T cells, CD4+ memory-activated T cells, follicular helper T cells, Tregs and gamma/delta T cell fractions. The total macrophage fraction was calculated as the sum of the M0, M1 and M2 macrophage fractions. The total number of B cells was calculated as the sum of the numbers of memory and naïve B cells.
Patient samples
Nine patients with OSCC (five with WT TP53 and four with the TP53R175H mutation) were identified and profiled the TIME by opal mIF. All patients underwent primary surgical excision with curative intent in the Department of Head and Neck Surgery, The University of Texas MD Anderson Cancer Center, Houston Texas, USA. Patient sample information is shown in online supplemental table S2.
Statistical analysis
Statistical analysis was performed using SPSS V.19.0 and GraphPad Prism V.7.0 (GraphPad Software). All data are presented as mean±SE. Two-tailed Student’s t-tests and one-way analysis of variance tests were used to analyze the data. The log-rank test was used for the survival analysis. P values <0.05 were considered statistically significant.
Additional experimental procedures are shown in online supplemental material and methods.
Results
Mutant p53R172H modulates cytokines and chemokine-related gene expression in oral cancer progression
Recently, we generated syngeneic oral tumor cell lines using genetically engineered mouse models in combination with an oral carcinogenesis model. The ROC3 cell line was obtained from mice expressing the hotspot GOF mutation p53R172H.28 To investigate the function of mutant p53R172H in the TIME, we generated a stable p53-shRNA expression in ROC3 cells (ie, a syngeneic, warm-tumor model).28 Western blot analysis confirmed the knockdown of the mutant p53R172H in the ROC3 cell line (figure 1A). Next, parental, NTC and p53-shRNA ROC3 cells were implanted into the tongues of immunocompetent (C57BL/6J) and immunodeficient (RAG1-knockout) mice. We found that in C57BL/6J mice, the growth of p53-shRNA ROC3 tumor cells was impaired compared with that in parental and NTC-shRNA tumor cells (figure 1B). Similar results were obtained for ROC3 tumor cells implanted in the tongues of C57BL/6J mice, in which the p53R172H mutant was inactivated using CRISPR Cas9 technology (online supplemental figure 3). However, the same p53-shRNA ROC3 cells implanted in RAG1-knockout mice generated oral tumors, suggesting that the p53R172H mutation modulated lymphocyte function to promote tumor development (figure 1B). Next, to identify immune-related factors that impaired oral tumor growth, we selected the time point at which p53-shRNA ROC3 tumors showed reduced tumor volumes and then paired these tumors with NTC samples (figure 1B). Briefly, paraffin-embedded tumor samples from NTC-shRNA and p53-shRNA ROC3 tumors were profiled using the HTG EdgeSeq mouse messenger RNA (mRNA) tumor response panel. Differential gene expression analysis revealed significant fold changes in the expression of immune checkpoint regulatory molecules, cytokines and chemokines involved in the immune response mechanisms (figure 1C). p53-shRNA (p53-knockdown) ROC3 tumors showed downregulation of Bst2, Cd274, Lgals9, Pvr and Tigit ICIs. Moreover, we detected a marked reduction in the expression of the myeloid marker Cd33 but increased expression levels of Cd8a, Cd3e, Cd4 and Cd83 (figure 1C). In addition, we observed reduced mRNA levels of proinflammatory cytokines and chemokines, such as Cxcl5, Cxcl10, Cxcl2, Ccl2, Il6, Lif, Il1b, S100a8 and S100a9, but increased levels of Ccl5 and Ccl9 chemokines in p53-knockdown ROC3 tumors. A more detailed differential gene expression of cytokines/chemokines is described in online supplemental figure 4. Next, we performed quantitative PCR analysis, which confirmed that p53R172H mediated the upregulation of Ccl2 and Il6 and the downregulation of Ccl5 and Ccl9 in ROC3 tumors (figure 1D). To validate the relevance of Ccl2 in the ROC3 tumor microenvironment, we used a neutralizing anti-Ccl2 antibody in tumor-bearing mice. Interestingly, all tumors responded to anti-Ccl2 treatment compared with control IgG isotype control treated mice (figure 1E). These studies strongly support the relevance of Ccl2 in the tumor immune suppressive mechanisms in the ROC3 syngeneic oral tumor model.
The gain-of-function (GOF) p53R172H mutation controls the immune-related transcriptome in oral cancer progression. (A) Western blot analysis of p53R172H expression in parental, non-targeting control (NTC) and p53-short hairpin RNA (shRNA) ROC3 tumor cells. (B) ROC3 tumor cells from the parental, NTC and p53-shRNA lines (500 000 cells per line) were implanted into the tongues of C57BL/6J mice (n=5) and immunodeficient RAG1-knockout (KO) mice (n=5). Error bars show the SD. Arrows show tumor time point for HTG EdgeSeq studies. (C) Gene expression analysis of immune-related genes performed using an HTG EdgeSeq mouse messenger RNA (mRNA) tumor response panel. The heatmaps represent high (red) and low (green) RNA expression by arbitrary units (range 5.5–14 units). Every column represents an NTC or p53-knockdown (KD) tumor sample, and every row represents an immune-related gene. (D) Validation of chemokine RNA expression using quantitative PCR. Bars show the means of triplicate experiments. Error bars show the SEMs. Unpaired, two-tailed Student’s t-test p values are shown for the comparison of control versus p53-KD ROC3 tumor cells. **P<0.005; ***p<0.0005; ****p<0.0001. (E) ROC3 tumors responded to anti-CCL2 antibody treatment. (Top panel) Process diagram of the orthotopic injection of tumor cells (500 000 cells implanted: n=5 mice) and the treatment of ROC3-tumor-model mice with anti-C-C motif chemokine ligand 2 (CCL2) antibody. (Middle graph) ROC3 tumor-bearing mice that received single-antibody immunotherapy showed tumor response. (Bottom graph) Anti-CCL2 treatment significantly improved overall survival (log-rank test; p<0.0001).
We also confirmed the reduced levels of the protein interleukin-6 in the supernatant of p53-knockdown ROC3 cells (data not shown), suggesting an associated disruption of the cGAS-STING pathway by the mutant p53R172H. Recent studies have demonstrated that the GOF p53 mutation disrupts the cGAS-STING pathway, leading to an impaired innate immune signaling response14; therefore, we analyzed the role of mutant p53R172H in the modulation of cGAS-STING signaling molecules. Western blot analysis showed increased phosphorylation of STING, TBK and IRF3 in p53-knockdown ROC3 cells (online supplemental figure 5A). Since TBK and IRF3 both help to activate the IFN-signaling pathway, we then evaluated the transcriptional activation of type I IFN in the p53-knockdown ROC3 cells. Quantitative real-time PCR demonstrated significant mRNA expression of IFN-α and IFN-β in ROC3 cells with reduced levels of mutant p53R172H (online supplemental figure 5B). Next, we used a p53-null ROC2 cell line28 to overexpress mutant p53R172H and to evaluate the cGAS-STING signaling molecules. Interestingly, ROC2 cells stably overexpressing mutant p53R172H contained reduced levels of phosphorylated STING, TBK and IRF3 (online supplemental figure 5A). Next, qPCR analysis identified a modest gene expression upregulation of several chemokine and cytokine genes as Ccl2, Ccl4, Ccl5, Ccl9, Ccl20, Cxcl3, Cxcl5, Cxcl10, Cxcl16 and Il6 in the ROC2 cell line overexpressing the mutant p53R172H. Furthermore, Il1b, Infa1 and Ifnb1 were downregulated, and no gene expression changes were observed in Cxcl12, Lif, S100a8 and S100a9 (online supplemental figure 6).
While we observed a subtle change in the STING protein levels in the engineered ROC cell lines, immunohistochemical studies of p53-knockdown ROC3 tumors showed significantly higher levels of optical density staining for STING compared with NTC-shRNA. Similar results were observed in the p53R172H-CRISPR ROC3 tumors (online supplemental figure 5C). Furthermore, we injected ROC2-pBABE vector control and p53R172H cells into the tongue of C57BL/6J mice. By 8 days postinjection, the ROC2 p53R172H tumors grew much faster than the vector control (online supplemental figure 7). Overall, our results indicate that mutant p53R172H regulates the cytokines and chemokine-related immunosuppressive transcriptomes that favor tumor growth in the syngeneic tumor model.
Inactivation of mutant p53R172H reduces the infiltration of immune suppressor cells in oral tumors
Because the inactivation of mutant p53R172H in ROC3 tumors led to changes in cluster differentiation molecules, we next sought to quantify immune cell infiltration using flow cytometry. Thus, we prepared NTC and p53-knockdown ROC3 tumors for flow cytometry studies. In p53-knockdown ROC3 tumors, we found an increased infiltration of CD8+ IFNγ+ cells (figure 2A), whereas we found a significant reduction of CD4+ FoxP3+ Treg, CD8+ PD-1+ T cells, CD4+ PD-1+ T cells, CD4+ IL-17+ T cells and LyG+ LyClo, int G-MDSCs (figure 2B–F). After global evaluation of the tumor immune cell composition, we next explored the localization of the immune cells within the TIME using Opal mIF. After scanning and analyzing the digital slides, we determined that p53-knockdown ROC3 tumors contained a higher percentage of activated cytotoxic T lymphocytes (CTLs; CD8a+ granzyme B+/CD8a+) and a significantly lower percentage of infiltrated Tregs (CD4+ FoxP3+/CD4+) than p53-NTC ROC3 tumors (figure 3A). Interestingly, the percentage of active CTLs was higher on the tumor edges than in the tumor cores in both NTC-shRNA and p53-knockdown tumors (figure 3A). Next, we evaluated myeloid cells within the TIME. We found that CD206+ M2 TAMs were most abundant in the tumor core and were significantly reduced in p53-knockdown ROC3 tumors; however, we identified an increased infiltration of CD11c dendritic cells on the edges and in the cores of the tumors (figure 3B). Similarly, increased numbers of M2 TAMs were detected in tumor edge areas (figure 3B). Then, we evaluated the PD-1 levels expressed in tumor-infiltrated T cells. In NTC-shRNA ROC3 tumors, the PD-1 protein level was elevated in both CD4+ and CD8a+ T cells, but it was markedly reduced in p53-shRNA ROC3 tumors (figure 3C). These results agreed with the HTG EdgeSeq platform and flow cytometry studies. In addition, multiplex staining confirmed that mutant p53R172H regulates STING expression in ROC3 tumor cells (figure 3D).
Mutant p53R172H promotes the infiltration of immune suppressor cells in the tumor microenvironment. Comparison of the frequencies of (A) IFNγ+CD8+ T cells; (B) FoxP3+CD4+ (Treg) cells; (C) CD4+ PD-1+ T cells; (D) CD8+ PD-1+ T cells; (E) CD4+ IL-17+ T cells and (F) LyG+LyClow G-MDSC. Dots represent individual samples (n=4–5). Error bars show the SD. Unpaired, two-tailed Student’s t-test p values are shown. ****P<0.0001. G-MDSC, granulocyte myeloid-derived suppressor cell; IFN, interferon; IL, interleukin; KD, knockdown; NTC, non-targeting control; PD-1, programmed cell death protein 1; Treg, regulatory T cells.
Spatial evaluation of immune cell populations associated with mutant p53R172H. Tumor tissues were stained using fluorescent multiplex immunohistochemical analysis to detect different (A) T-cell markers (CD4, CD8, granzyme B (GzmB) and FoxP3); (B) myeloid cell markers (CD68, CD206 and CD11c); (C) the immune checkpoint marker programmed cell death protein 1 (PD-1) and (D) the tumor immunity regulation marker stimulator of interferon gene (STING). The numbers in each cell subset among the tumor cells within 50 μm was determined using Visiopharm software, and multiplex immunohistochemical quantification was defined as the number of cells per mm2. Three tumors were stained, and every sample had three tumor core views and three tumor edge views analyzed. *P<0.05; **p<0.01; ***p<0.001; ****p<0.0001. CK, cytokeratin; KD, knockdown; NTC, non-targeting control; shRNA, short hairpin RNA. Bars represent SE.
Next, we performed a uniform manifold approximation and projection clustering analysis (UMAP) using the opal mIF data. We observed that the CD11c dendritic cells and the other macrophage cell population—CD68+ cells—had a close spatial relationship with the STING− tumor cells in the NTC-shRNA ROC3 tumors. In contrast, these two cell types had a closer interaction with STING+ tumor cells in the p53-knockdown tumors. Furthermore, the M2 TAMs population (CD68+, CD206+) was closely associated with active CTLs in the NTC tumors but not in the p53-knockdown ROC3 tumors (figure 4A). Interestingly, the distribution of immune cells in the tumor cores and on the edges of the p53-knockdown tumors was similar, in contrast to the spatial distributions in the tumor cores and on the edges of the control ROC3 tumors (figure 4A). Next, we analyzed the distances between different immune cell populations within the tumor microenvironment. We found that active CTLs and CD11c dendritic cells were more spatially distant in p53-knockdown ROC3 tumors than in NTC-shRNA ROC3 tumors. Tregs and M2-TAMs are located close to active CTLs and tumor cells in NTC-shRNA ROC3 tumors. However, in p53-knockdown ROC3 tumors, CTLs and dendritic cells are closely situated to tumor cells located on carcinoma edges and cores (figure 4B). Interestingly, the spatial distances between active CTLs, Tregs and CD11c dendritic cells are different between the tumor cores and the edges in the NTC ROC3 tumors (figure 4B). Next, we analyzed NTC-shRNA tumors by selecting STING+ and STING− cells. This analysis showed that more active CTLs and CD11c dendritic cells were located around STING+ tumor cells. Interestingly, the infiltrations of active CTLs and M2-TAMs varied between the tumor cores and edges (online supplemental figure 8A,B). Finally, the percentages of Tregs (CD4+ FoxP3+/CD4+), M2-TAMs (CD68+ CD206+/CD68+) and exhausted CD8 T cells (CD8+ PD-1+/CD8+) were dramatically reduced around STING+ tumor cells (online supplemental figure 8A-C). These results strongly suggest that tumor cells with high STING protein levels attract CTLs and CD11c+ cells to activate the innate and adaptive immune responses.
Mutant p53R172H modulates the immune cell distribution within a warm tumor immune microenvironment that responds to immunotherapy treatment. (A) Uniform manifold approximation and projection clustering analysis (UMAP) of tumor tissues stained with Opal multiplex immunofluorescence. (B) Distance measurements of different immune cell populations within tumor tissues. The calculated distance was to the nearest cell within an area 1 mm in diameter. (C) Process diagram of the orthotopic injection of tumor cells (500 000 cells implanted; n=10 mice) and the treatment of ROC3 tumor model mice with antiprogrammed cell death protein 1 (PD-1) antibody, anti T-cell immunoreceptor with immunoglobulin and immunoreceptor tyrosine-based inhibitory motif domains (TIGIT) antibody and stimulator of interferon gene (STING). (D) ROC3 tumor-bearing mice that received single-antibody immunotherapy showed tumor response. (E) Anti-PD-1 and anti-TIGIT significantly improved overall survival (log-rank test; p<0.0001). (F) ROC3 tumors responded to STING-agonist monotherapy (c-di-GMP). (G) STING-agonist treatment improved tumor-free survival. Three tumors were stained, each sample had three tumor core views and three tumor edge views were analyzed. *P<0.05; **p<0.01; ****p<0.0001. Bars represent the SE. CK, cytokeratin; GzmB, granzyme B; KD, knockdown; NTC, non-targeting control; PBS, phosphate-buffered saline.
ROC3 oral tumors infiltrated with PD-1+ and TIGIT+ T-cell populations respond to immunotherapy treatment
ROC3 cells implanted in the tongue generate warm tumors with a high infiltration of effector and immunosuppressor immune cells expressing different ICIs to evade immunosurveillance.28 TIGIT and PD-1 ICIs are markers of exhausted cytotoxic CD8+ T cells, which help these cells to maintain a dormant state. Thus, we determined whether ICI-blocking antibodies and STING agonists could cause tumor regression by activating exhausted T cells (figure 4C). ROC3 tumors responded to treatment with anti-TIGIT and anti-PD-1; both immunotherapies exhibited single-agent activity in the therapeutic setting. This finding confirmed that ICI blocking is required to elicit optimal T-cell immune activity in tumors (figure 4C–E). Furthermore, PD-1 and TIGIT blockade systemically elicited a robust antitumor effect, with 100% and 80% of the treated mice, exhibiting a complete response, respectively (figure 4C). In addition, tumor-free mice efficiently resisted engraftment when rechallenged with fresh ROC3 tumor cells (online supplemental figure 9). We observed a similar tumor response in ROC tumors compared with control ROC3 tumors after intratumoral administration of the c-di-GMP STING agonist (figure 4F–G). Interestingly, the ROC3 tumors expressed mutant p53R172H, which generated a TIME to disrupt tumor immunity; however, immunotherapy treatments promoted a complete response that provided immune memory.
Mutant TP53R175H regulates immune cell populations and its associated prognosis in patients with OSCC
Because the p53R172H GOF mutation modulates the TIME in mouse oral tumors, we performed CIBERSORT analysis to determine the immune cell composition based on tumor gene expression in patients with OSCC with the TP53R175H mutation (equivalent to murine p53R172H) and compared it with patients with WT TP53. First, head and neck cancer cohort data (including 433 human papillomavirus-negative patient samples) were downloaded from The Cancer Genome Atlas website. We selected 141 patients with OSCC with WT TP53 and 11 patients with the TP53R175H mutation. The analysis revealed that compared with patients with OSCC with WT TP53, those with the TP53R175H mutation had a lower infiltration of total lymphocytes, CD8+ T cells and follicular helper T cells, but a significantly higher infiltration of M2 TAMs (figure 5A,B). Furthermore, we used the TCGA database to explore the cytokine and chemokine expression of the patients with OSCC containing WT TP53 TP53R175H mutation, however we did not reveal any significant gene expression difference (online supplemental figure 10). Next, to validate our computational predictions, we selected nine tissue samples of patients with OSCC (five patients with WT TP53 and four with the TP53R175H mutation) and profiled the TIME with opal mIF. Briefly, digital imaging analysis revealed that patients with the TP53R175H mutation had a lower infiltration of CTLs (CD3+ CD8a+) and helper T lymphocytes (CD3+ CD4+) than those patients with WT TP53. Furthermore, the percentage of activated CTLs (CD8a+ granzyme B+/CD8a+) was lower in patients with the TP53R175H mutation than in those within the tumor core and edge containing WT TP53 (figure 6A).
Human TP53R175H mutation regulates immune cell infiltration as demonstrated with CIBERSORT. (A) The relative proportions of 22 human immune cell types in patients with OSCC. (B) Mean values and SD for specially listed cell subsets, such as total lymphocytes, CD8 cells, follicular helper T cells, M1 macrophages, M2 macrophages and neutrophils were calculated for each patient group. CIBERSORT immune cell fractions were determined for each patient. Each dot represents one patient. **P<0.01; ***p<0.001. Bars represent SE. NK, natural killer; ns, not significant; WT, wild type.
However, the percentage of Tregs in the population of helper T cells (CD4+ FoxP3+/CD4+) was significantly higher within all tumor areas (edge/core) expressing the TP53R175H mutation (figure 6A). Of interest, the infiltration of active CTLs and Treg cells were different between tumor core and edge when tumors had the TP53R175H mutation (figure 6A). Infiltration of Tregs was similar between the tumor core and edge when tumors with TP53 WT (figure 6A). Furthermore, we evaluated PD-1 and TIGIT expression levels in tumor-infiltrated T cells from OSCC samples. Interestingly, the percentages of CD3+PD-1+ and CD3+TIGIT+ T cells were significantly higher in patients with OSCC harboring the TP53R175H mutation (figure 6B). As previously demonstrated in the CIBERSORT analysis, a mIF analysis confirmed a higher number of M2 TAMs (CD68+ CD206+) in patients with OSCC with the TP53R175H mutation than in those with WT TP53 (figure 6C). We also found that CD206+ M2 TAMs were most abundant in the tumor core with TP53R175H mutation, while there was a difference between tumor core and edge with TP53 WT (figure 6C). Furthermore, we observed only a higher infiltration of myeloid CD11c+ cells in OSCC cells containing WT TP53 (figure 6C). A higher infiltration of myeloid CD11c+ cells was observed at the tumor edge both in tumors with TP53R175H mutation and TP53 WT (figure 6C). Remarkably, when we analyzed the STING protein levels in the tissue samples, we found significantly reduced levels of STING protein in the samples with the TP53R175H GOF mutation (figure 6D). Next, we evaluated patients with TP53 WT and R175H mutant and performed immune cell composition using UMAP clustering analysis with mIF data. First, cell clustering analysis revealed a significant alteration in STING+ tumor cells in patients with TP53R175H. In addition, myeloid and lymphoid cells produced dramatically different cluster shapes within the tumor edge and core containing different TP53 statuses (figure 7A, online supplemental figure 11). The analysis revealed that CD11c+ dendritic cells and other CD68+ macrophage populations had a close spatial relationship with STING+ tumor cells containing TP53 WT. Interestingly, TP53R175H mutant tumor cells had a closer interaction with M2 macrophages (CD68+CD206+) and Treg cells (CD4+Foxp3+), while TP53 WT tumor cells had a closer spatial distance with active CTLs and myeloid CD11c+ cells (figure 7B). Notably, our patient sample cohort used for mIF studies showed that TP53R175H mutation correlates with reduced disease-free and overall survival compared with patients containing WT TP53 (online supplemental table S4). These studies demonstrate that TP53R175H mutation status modulates the spatial distribution of immune cells within the tumor microenvironment. Remarkably, these results are consistent with the syngeneic murine tobacco-associated oral tumor model ROC3 summarized in our working model (figure 7C). Overall, opal mIF studies validated our in silico prediction regarding the impact of mutant TP53R175H GOF on the TIME.
Evaluation of the tumor immune microenvironment (TIME) in patients with oral squamous cell carcinoma (OSCC) with TP53R175H mutation. Patient tumor tissues were stained using Opal multiplex immunofluorescence (mIF) to detect (A) T-cell markers (CD3, CD4, CD8, granzyme B (GzmB) and FoxP3) and (B) immune-checkpoint markers (programmed cell death protein 1 (PD-1) and T-cell immunoreceptor with immunoglobulin and immunoreceptor tyrosine-based inhibitory motif domains (TIGIT)). (C) Myeloid cell markers (CD68, CD11c, CD206) and (D) stimulator of interferon gene (STING) expression. Scale bar, 50 μm. The Opal multiplex immunohistochemical quantification was scored as the cell density, defined as the number of positive cells per view (20×). Five TP53 wild-type (WT) and four TP53 R175H patient samples were used. Five fields per sample were analyzed. *P<0.05; **p<0.01; ***p<0.001; ****p<0.0001.
Mutant p53R175H controls the immune cell distribution within tumor microenvironment and is associated with reduced survival in patients with oral squamous cell carcinoma (OSCC). (A) Representative uniform manifold approximation and projection (UMAP) clustering analysis of tumor tissues stained with Opal multiplex immunofluorescence of patients 1 and 4. (B) Distance measurements of different immune cell populations within tumor tissues. The calculated distance was to the nearest cell within an area 1 mm in diameter. (C) Working model of mutant p53R172/TP53R175H role in shaping the immune microenvironment of OSCC. (Left panel) Gain-of-function mutant p53R172H drives the infiltration of cytotoxic CD8+ T cells, and promotes intratumoral recruitment of Tregs, and M2 macrophages, thereby promoting tumor progression. (Right panel) Mouse and human tumor cells expressing p53R172H/TP53R175H gain-of-function mutations promote a spatial association with immunosuppressive cells (M2 TAM, Tregs) to reduce the number prevent the activation of dendritic and cytotoxic T cells in different tumor zones (core and edge). Knockdown (KD) mutant Trp53 (mouse) or expression of TP53 WT (human) in STING (+) tumor cells, contains more dendritic cells, active cytotoxic T cells and reduces the numbers of Tregs and M2 TAMs within the tumor microenvironment. Bars in the graphs represent SE with the following p values: *p<0.05; ****p<0.0001. M2 TAM, tumor-associated macrophage type 2; Tregs, regulatory T cells; STING, stimulator of interferon response cGAMP interactor.
Discussion
In this study, we explored the functional role of the p53R172H/R175H GOF mutation in the TIME using a syngeneic tumor model and correlated our findings with those in tissues from patients with OSCC. First, we established that inactivation of the p53R172H GOF mutation impairs tumor growth in a syngeneic model. We identified a change in the expression of cytokines and chemokines which a previous study identified as being involved in tumor-promoting inflammation, one of the hallmarks of cancer.31 Reports have shown that genetic alterations in p53 can influence cytokine signaling by activating nuclear factor kappa B (NF-κB) signaling and increasing p65 nuclear localization, making a molecular complex that increases NF-κB’s transcriptional activity.32 33 A recent study demonstrated that p53R273H mutant and NF-κB drive alterations in enhancers and gene activation in response to chronic tumor necrosis factor-α signaling, inducing chronic inflammation.34 The same study determined that mutant p53R273H regulates the recruitment of RNA polymerase II to activate transcription of the CCL2 chemokine gene34 consistent with our finding that inactivation of mutant p53R172H in ROC3 tumors significantly downregulated Ccl2 mRNA expression. Moreover, Ccl2 drives intratumoral recruitment of inflammatory monocytes and promotes metastasis in head and neck cancers.35 36 These data support the role of mutant p53 in regulating proinflammatory cytokines and promoting the immunosuppressive tumor microenvironment. Other cytokines such as Cxcl5 are downregulated when mutant p53R172H is inactivated in the tumor microenvironment. Interestingly, Cxcl5 is involved in angiogenesis in lung cancers containing p53 missense mutations, and recruitment of immunosuppressive cells in pancreatic cancer.37 38 In ROC3 oral tumors, the inactivation of p53R172H GOF mutation reduces the levels of Ccl2 and Cxcl5 cytokines, both of which are important in establishing and sustaining an immunosuppressive tumor microenvironment. Interestingly, ROC3 tumors treated with neutralizing anti-Ccl2 antibodies showed a complete response even after antibody treatment ended; these results confirm the importance of Ccl2 in establishing an immune suppressive tumor microenvironment.
Furthermore, p53 missense mutations disrupt the cGAS-STING signaling pathway, which is important for the activation of the innate immune response.14 Remarkably, we found that the mutant p53R172H altered the activation of the cGAS-STING pathway, with reduced levels of phosphorylated IRF3 and reduced expression of type I IFN. Furthermore, we demonstrated a significant reduction in STING protein levels in murine syngeneic oral tumors and found similar results in patients with OSCC tumors expressing TP53R175H (equivalent to the murine mutation). These findings indicate that p53R172H/R175H GOF mutations are functionally conserved and prevent activation of innate immunity via IFN signaling. Because interleukin-6 is an inflammatory cytokine involved in chronic inflammation, immune escape and the acceleration of tumor progression,39 downregulation of interleukin-6 in tumors with the inactivated GOF p53R172H mutation is consistent with our hypothesis that interleukin-6 is involved in shaping the TIME in p53 mutant tumors. Other cytokines, such as Ccl5 and Ccl9, also play an important role in tumorigenesis and in the immune system. Follow-up studies using single-cell RNA sequencing or spatial transcriptomics will provide more comprehensive insights into the cell populations responsible for altered cytokine and chemokine expression within the tumor microenvironment.
Cytokines and chemokines regulate the activation of immune cells and their migration into and out of tissues, and they also guide their spatial organization within tissues. We found that disruption of the GOF p53R172H mutation was associated with altered cytokine and chemokine expression levels as well as maturation and infiltration of different immune subsets, including cytotoxic CD8+ T cells, helper CD4+ T cells, Tregs, G-MDSCs, TAMs and tumor-associated neutrophils in syngeneic ROC3 oral tumors. In addition, our histological spatial expression analysis revealed that ROC3 tumors expressing the p53R172H mutation promoted the infiltration of Tregs, M2 macrophages (CD206+), exhausted CD8+ T cells and excluded immune-activated effectors (granzyme B+) and CD11+ cells in both the tumor cores and on the tumor edges. However, inactivation of p53R172H in ROC3 tumors led to a significant infiltration of activated CTLs and CD11c+ cells, as well as a marked reduction of Tregs and M2 macrophages in the tumor cores and on the tumor edges. Interestingly, the spatial distance between the tumor cells and the activated CTLs and CD11c cells was reduced in p53-knockdown tumor cells (STING+). This association suggests that the cGAS-STING pathway might induce the production of IFNs to activate immune cells to migrate toward and kill tumor cells. Our findings support the recently reported finding that the GOF p53 mutation disrupts the cGAS-STING pathway.14 In our study, we found that murine p53R172H (human p53R175H) mutation disrupted the phosphorylation of the STING pathway, and decreased the absolute amount of STING protein.
Warm tumors are characterized by the accumulation of proinflammatory cytokines and T-cell infiltration, and have a better response rate to immune checkpoint blockade treatment than cold immune-excluded tumors.40 ROC3 oral tumors, which are warm tumors, express high levels of proinflammatory cytokines and are infiltrated by different immune cell subsets, including CD8+ and CD4+ T cells. In addition, ROC3 oral tumors respond to PD-1 and TIGIT blockage, and studies have shown that these tumor-free mice develop immune-cell memory, suggesting that ROC3 oral tumors express neoantigens. Interestingly, whole-exome sequencing studies of the ROC3 cell line have revealed that these cells have a larger mutational landscape than other murine cell lines,28 which is characteristic of hot tumors. These studies suggest that ROC3 tumors are infiltrated with antigen-specific CD8 T cells but are exhausted by Tregs and M2-TAMs; however, on ICI therapy, the immunosuppression mechanisms are disrupted, leading to tumor response and immune cell memory.
Using CIBERSORT and data from The Cancer Genome Atlas database, we compared the relative proportions of 22 human immune cell types in patients with OSCC expressing mutant GOF TP53R175H with those in patients expressing WT TP53. We identified a lower infiltration of CD8+ T cells and follicular helper T cells but a significantly higher infiltration of M2 TAMs in patients expressing the TP53R175H mutation. Remarkably, these results correlated with the results of our Opal mIF studies of patients with the TP53R175H mutation, suggesting that our in silico analyses were highly reliable and provided evidence of the functional role of the mutant TP53R172H in shaping the TIME in OSCC. Interestingly, we discovered that STING1 protein levels are associated with the TP53 mutational status, which led us to speculate that disruption of cGAS-STING signaling is relevant to the development of OSCC and might be associated with poor patient prognosis.
In conclusion, our findings demonstrate that the p53R172H/ TP53R175H GOF mutation modulates the tumor immune landscape of OSCC. The generation and characterization of syngeneic tumor models for p53R172H missense GOF mutations in oral cancer will allow us to understand the cell-intrinsic factors that modulate immunosuppression mechanisms in the TIME. Finally, the ROC3 syngeneic murine oral tumor model used in this study can be used as a tool to develop rational combinations of drugs, radiation and immunotherapies to treat patients with p53 GOF mutant OSCC.
Data availability statement
Data are available on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
All animal experiments were approved by the Institutional Animal Care and Use Committee of the University of Texas MD Anderson Cancer Center.
Acknowledgments
The authors thank Dr Jared Burks for help with the multiplex immunohistochemistry imaging and cell sorting at the Flow Cytometry and Cell Imaging Core Facility, and the Functional Genomics and Advance Technology Genomics Core Facilities at MD Anderson Cancer Center; for supporting reagents and services supported by the NIH/NCI under award number P30CA016672.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
YS and XR contributed equally.
Contributors YS: formal analysis, validation, investigation, methodology, writing—original draft and writing—review and editing. XR: investigation and data analysis. SC: investigation, and data analysis. XC: investigation and data analysis. BY: methodology and data analysis. FHBdC: data analysis, review and editing. AERR: data analysis, review and editing. AC: methodology. CM: investigation, methodology, review and editing. RV: investigation, methodology, review and editing. AAO: investigation, methodology, review and editing. TX: investigation, and methodology. WW: investigation, data analysis and editing. AS: conceptualization, resources, formal analysis, supervision, investigation, methodology, data analysis, funding acquisition, writing—review and editing. JNM: conceptualization, resources, formal analysis, supervision, funding acquisition, writing—review and editing, and project administration. RR: conceptualization, formal analysis, supervision, validation, investigation, methodology, writing—original draft, project administration, writing—review, editing and acts as guarantor.
Funding This work was supported by The Nancy L De Anda Research Foundation, Mary K Chapman Foundation (to JNM), NIH 5U01DE028233-03 (to AGS), NIH 5R01DE030875-02 (to AS and JNM).
Competing interests No, there are no competing interests.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.