Abstract
Background/Aim: The recurrence rate of head and neck squamous cell carcinoma (HNSCC) remains high; thus the control of recurrence is a clinical problem to be challenged. To clarify the precise mechanism, specific immunological biomarkers responsible for recurrence were investigated. Patients and Methods: The expression levels of immune response-associated and Shizuoka Cancer Center 820 cancer-associated genes, and genetic mutations from whole-exome sequencing were compared between HNSCC patients who developed recurrence (n=8) and HNSCC patients who did not develop recurrence (n=19) using a volcano plot analysis. Cytokine and epithelial-mesenchymal transition marker genes were analyzed using quantitative PCR. Tumor-infiltrating lymphocytes, immune checkpoint molecules, and human papilloma virus status were investigated using immunohistochemistry (IHC). Results: Twenty-seven evaluable patients with HNSCCs received radiation therapy after surgery. Recurrence was identified in 8 patients. TP53 mutations tended to be higher in patients who developed recurrence than in those who did not develop recurrence (75% vs. 31.6%). Gene expression profiling showed the down-regulation of T cell activation genes (ICOS, CD69 and CD83) and the upregulation of the ERBB4, EGFR, VEGF, HIF1A, TGFB1, TWIST1, IL-8, and PAX7 genes, which suggested the activation of the TP53 mutation-TGF-β1-PAX7 pathway and epithelial-mesenchymal transition. Additionally, IHC indicated a tendency toward a reduction in T cell accumulation and an increase in M2-type macrophage infiltration in tumors that recurred. Conclusion: A TP53 mutation-mediated immune-suppressive state in the tumor microenvironment and TGF-β1-PAX7-mediated EMT might contribute to the promotion of recurrence in patients with HNSCC after postoperative radiotherapy.
- Tumor microenvironment (TME)
- epithelial-mesenchymal transition
- tumor-infiltrating lymphocytes
- human papillomavirus
- TP53 mutation-TGF-β1-PAX7 pathway
Head and neck squamous cell carcinoma (HNSCC) is a common type of cancer; however, a reduction in the rate of its recurrence remains challenging; the 5-year survival rate of patients with HNSCCs is reported to be only 50-60% (1, 2). The recurrence rate is high, approximately 30%, which indicates not only high mortality and difficulty in treatment, but also the genetic complexity affecting cancer progression (3). These observations might suggest that the aspect shifted from a single cancer cell to a complex cancer tissue, namely, the tumor microenvironment (TME) (4).
The TME comprises stromal (cancer-associated) fibroblasts, immune cells (T cells, macrophages etc.) and other supporting cells. Recently, along with recent advances in next-generation sequencing (NGS)-based single-cell RNA sequencing technology, immune landscape studies have been performed according to genetic classification of tumor-infiltrating lymphocytes (TILs), enabling the evaluation of the immunological status in the TME (5-7). These studies have elucidated the importance of mutations in the tumor-suppressor gene TP53 and human papillomavirus (HPV)-positive status in the clinical prognosis of patients with HNSCCs (8, 9).
TP53 mutations are the most frequently identified mutations in non-HPV HNSCCs, and its frequency in HNSCCs is approximately 50%. Additionally, a higher rate of TP53 mutations was recognized in recurrent HNSCCs after chemoradiotherapy (CRT) compared to patients with HNSCC who did not recur (10-12). Meanwhile, an HPV-positive status has been demonstrated to be a good prognostic factor in patients with HNSCC (9, 13). Importantly, oncogenic HPV16 and other factors have recently been described as apparent etiologies for oropharyngeal cancer in young non-smoking patients (14). According to these observations, HPV status and smoking history define three groups with distinct survival outcomes: low risk (HPV-positive and <10 pack-years), intermediate risk (HPV-positive and >10 pack-years), and high risk (HPV-negative) (15).
In the present study, we focused on patients with HNSCCs who subsequently developed recurrence after postoperative radiotherapy, who showed a high frequency of TP53 mutation and an HPV-negative status and were categorized as the high-risk group. Eventually, we characterized genetic and immunological features using gene expression profiling (GEP) and whole exome sequencing (WES) analysis of these patients to identify novel prognostic biomarkers.
Patients and Methods
Patient registration. Shizuoka Cancer Center (SCC) has been conducting the comprehensive human genome analysis Project HOPE (High-tech Omics-based Patient Evaluation) using WES and GEP (16, 17). Informed consent was obtained from all registered cancer patients. The study was approved by the Institutional Review Board of Shizuoka Cancer Center, Japan (Authorization Number: 25–33). All experimental protocols using clinical samples were carried out in accordance with the Helsinki Declaration and the Ethical Guidelines for Human Genome and Genetic Analysis Research. Thirty patients with HNSCC were registered in Project HOPE from 2014 to 2015 and received radiotherapy after surgical resection. Three patients were not evaluable due to insufficient tumor cell content in formalin-fixed, paraffin-embedded (FFPE) samples and were not included in the analysis. The median follow-up time for overall survival was 47.8 months (range, 7.5-63.1 months).
DNA microarray-based GEP and WES using next-generation sequencing and Sanger sequencing. The method used to perform GEP and WES analyses was described previously (17). Briefly, in GFP analysis, the ratio of the expression levels in the tumor tissue versus the normal tissue was calculated from the normalized values. In WES analysis, the variants called by the variant caller were filtered and those mutations that were identified in tumor samples only were identified as somatic mutations.
Determinations of tumor mutation burden (TMB) and copy number variation (CNV) number. Tumor mutation burden (TMB) represents the number of somatic mutations per megabase and the methods for determining of TMB and copy number variation (CNV) number were described previously (18).
Immune response-associated genes and SCC820 panel gene expression profiling. The lists of genes in the 204 immune response-associated gene panel and the SCC820 cancer-associated panel were shown previously (17, 19). Briefly, the expression levels of 204 immune response-associated genes and the SCC820 cancer-associated genes of HNSCCs were compared between patients who developed recurrence and patients who did not develop recurrence with a volcano plot analysis. The expression-altered (more than 2-fold) genes among the 204 immune response-associated genes and the SCC820 cancer-associated gene panel were identified.
Real-time PCR analysis. Real-time PCR analysis of the apoptosis-related, cytokine and EMT-associated genes was performed using the QuantStudio 12K Flex Real Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA) as described previously (20). Additionally, in the cancer metabolomic study, real-time PCR using primers targeting glycolytic enzyme genes, such as HK2, PDK1, PFKM, PGM1, and PKM, was performed. Total RNA was isolated from formalin-fixed paraffin-embedded (FFPE) head and neck cancer specimens using the High Pure RNA Paraffin Kit (Roche Diagnostics, Indianapolis, IN, USA).
Immunohistochemistry. For the tumor-infiltrating immune cells, antibodies against CD4 and CD8 (Thermo Fisher Scientific), FoxP3 (Abcam, Cambridge, UK), CD204 (Transgenic, Kobe, Japan), TGF-β1 (R and D Systems, Minneapolis, MN, USA), cytokeratin (AE1 and AE3, Nichirei Bio., Tokyo, Japan), IL-8 (Abcam), Granzyme B (Dako, Glostrup, Denmark), PD-L1 (Abcam) and PAX7 (Novus Biologicals, Centennial, CO, USA) were purchased and used for immunohistochemistry and immunofluorescence analyses. For the evaluation of HPV infection, an anti-human p16ink4a mouse monoclonal antibody (clone E6H4; Roche Diagnostics, Basel, Switzerland) was purchased. Positive staining of more than 70% of the area was diagnosed as positive according to Schache et al.’s report (21). To measure TIL numbers, 10 areas of tumor at a high magnification (×200) in each section stained with various antibodies were used and analyzed by two highly experienced pathologists.
For immunofluorescence staining, TGF-β1, cytokeratin, PAX7, IL-8 and CD204 staining was conducted using the Opal 4-color IHC kit (Perkin-Elmer, Waltham, MA, USA) and evaluated on a fluorescent Zeiss imager Z1 microscope (Carl Zeiss, Oberkochen, Germany).
T cell receptor gene repertoire analysis in head and neck tumors using human TCRα and TCRβ profiling kit. Total RNAs were isolated from 27 head and neck squamous tumors and applied to Switching Mechanisms at 5′ End of RNA Template (SMARTer™) human TCRα and TCRβ profiling kit (Clontech Laboratories, Mountain View, CA, USA) as described previously (22). TCR repertoire analysis was performed using MiTCR, software for T cell receptor sequencing data analysis (23).
Statistical analysis. The genes from the immune response-associated gene panel and SCC820 cancer-associated gene panel that were upregulated in patients who developed recurrence compared to patients who did not develop recurrence were identified using the volcano plot method. The statistical significance of the difference in the proportion of variables between recurrent patients and non-recurrent patients was calculated using an unpaired two-tailed t-test. Values with p<0.05 were considered significant. Data analysis was performed using GeneSpring GX software version 13.1.1 (Agilent Technologies, Santa Clara, CA, USA). For survival analysis, the association of PAX7 gene expression, driver mutations, TP53 mutation, disease recurrence and HPV status with overall survival was analyzed using Excel 2016. Specifically, PAX7 gene expression and the clinical data of HNSCC patients were downloaded from TCGA database. The overall survival was analyzed using the Kaplan–Meier method and the statistical significance of each parameter was evaluated using the log-rank test and a generalized Wilcoxon test.
Results
Clinical characteristics of HNSCC patients. The characteristics of 27 evaluable HNSCC patients are shown in Table I. There were 20 male and 7 female patients, and the mean age was 61 years (range, 26-83 years). The site of the primary tumor was the oral cavity in 12 patients, the hypopharynx in 6 patients, the oropharynx in 7 patients, and the larynx in 2 patients. According to the tumor-node-metastasis (TNM) 7th classification, one patient had stage II disease, two patients had stage III disease, and 24 patients had stage IV disease. Among 27 patients with HNSCC, eight patients developed recurrence (local and distal recurrence in 5 patients, distal recurrence only in 3 patients). HPV statuses were positive in seven patients. Radiotherapy or chemoradiotherapy was performed postoperatively. Concurrent chemoradiotherapy was administered to patients with high-risk features such as positive surgical margin or extranodal extension, and with good performance status. Chemotherapy regimens included weekly cisplatin (CDDP) at 40 mg/m2, triweekly CDDP at 80-100 mg/m2, or CDDP at 70 mg/m2 and 5-fluorouracil at 700 mg/m2. The median radiation dose was 60 Gy (range, 40-70 Gy).
Characteristics of patients with head and neck squamous cell carcinoma.
Characterization of genetic features using WES and GEP analysis. The features of genetic mutations were compared between patients who developed recurrence and patients who did not develop recurrence (Table II). The total number of single-nucleotide variants (SNVs) was not different between the groups. Regarding driver mutations suggested by Vogelstein, the frequency of TP53 mutation was significantly higher in patients who developed recurrence than in patients who did not develop recurrence. However, the frequencies of other mutations were not different. In contrast, patients with HPV-positive tumors developed no recurrence and had no driver mutations, including TP53 mutations (Table II).
Effect of driver mutation frequency and HPV status on the recurrence of head and neck cancer.
Expression profiling of immune response-associated and SCC820 panel genes in recurrent HNSCC. Volcano plot analysis showed one up-regulated and 10 down-regulated genes from the immune response-associated panel of genes (Figure 1A, Table III), and 18 up-regulated and five downregulated genes from the SCC820 cancer-associated panel of genes (Figure 1B, Table IV) in patients who developed recurrence. Briefly, T cell and dendritic cell-activating genes (CCL5, ICOS, CD69, and CD83) and the cyclin-dependent kinase inhibitor (CDKN)2A and CDKN2B genes were downregulated, while cancer-associated genes, such as PAX7, MAGEA1, TLX1 and ERBB4, were up-regulated in patients who developed recurrence.
Comparison of gene expression between the recurrence group and the no recurrence group after postoperative radiotherapy. (A) Immune response-associated genes and (B) cancer-associated SCC820 genes. Up-regulated or down-regulated genes with changes greater than 2-fold were identified using volcano plots with the Benjamini-Hochberg correction. The horizontal gray line represents a p-value of 0.05. The vertical gray lines show 2- and 0.5-fold changes in gene expression. The black and white squares represent the down-regulated and up-regulated genes, respectively.
Expression-altered genes of immune response-associated gene in recurrent head and neck cancer patients.
Expression-altered genes of SCC820 cancer-associated genes in recurrent head and neck cancer patients.
Apoptosis and EMT marker gene expression measured using quantitative PCR. Quantitative PCR demonstrated that the expression of EMT-associated genes (COL1A2, HIF1A, IL-8, SNAIL2, TWIST1, TGFB1, and TGFB3) and glycolytic enzyme genes (HK2 and PDK1) was up-regulated in patients who developed recurrence (Figure 2). In contrast, the expression of CDKN2A was down-regulated. Other cancer signaling pathway genes (EGFR, VEGF, and CCND1) were also up-regulated. Based on the gene expression profiling data from the volcano plot analysis and quantitative PCR data, a scheme of differentially expressed genes in patients who developed recurrence was created and is shown in Figure 3; this schematic indicates specific mechanisms or pathways that might be associated with head and neck cancer metastasis and progression after postoperative radiotherapy.
Apoptosis-related, cytokine and EMT-associated gene expression profiling using real-time PCR. Total RNA was isolated from formalin-fixed paraffin-embedded (FFPE) head and neck cancer specimens, and real-time PCR was performed using the QuantStudio 12K Flex Real Time PCR System. The open column shows nonrecurrent cases; the closed column shows recurrent cases. Each column represents the mean value of triplicate experiments.
Schematic of differentially expressed genes that could be associated with cancer metastasis in patients who developed recurrence. The up-regulated (EGFR, ERBB4, VEGF, TGFB1, PAX7, IL-8, SNAIL2, TWIST1, COL1A2, CLEC12B, etc.) and down-regulated genes (CDKN2A, CDKN2B, CD69, ICOS, CD83, etc.) might be associated with cancer metastasis, as determined from the gene profiling of immune response-associated and SCC820 panel genes, and real time PCR.
Association of PAX7-positive tumor cells with IL-8-producing tumor-associated macrophages using immunofluorescence staining. Immuno-fluorescence staining showed that TGF-β1 might be positive in tumor cells with PAX-7 expression and IL-8 was positively stained in CD204+ tumor-associated macrophages (Figure 4).
Tumor cell and macrophage characterization using immunofluorescence staining (IFS). (A) IFS for cytokeratin (blue), TGF-β1 (green) and PAX7 (red) proteins and (B) IFS for cytokeratin (blue), IL-8 (green) and CD204 (red) using the Opal 4-color kit. The tumor sections were derived from the recurrent group (No. 21). The arrows in panel A show TGF-beta1+PAX7+ cancer cells and the arrows in panel B show CD204+IL-8+ macrophages. Magnification: ×400.
Tumor-infiltrating immune cell characterization and HPV status determination in HNSCCs using IHC. The number of CD8+ T cells and granzyme B+ cells tended to decrease in patients who subsequently developed recurrence (Figure 5). Specifically, the granzyme B+ cell number was significantly smaller in patients who developed recurrence than in those who did not. Additionally, CD204+ macrophage infiltration tended to be higher in patients who subsequently developed recurrence than in those who did not; however, it was not a significant difference (Figure 5B). In contrast, none of the patients with HPV-positive tumors developed recurrence, and they had a significantly greater number of infiltrating CD8+ T cells than patients with HPV-negative tumors.
Tumor-infiltrating immune cell characterization using immunohistochemical (IHC) staining. (A) IHC staining against CD8, Granzyme B, Foxp3 and CD204 in the no recurrence group (No. 25) and the recurrence group (No. 17). Magnification; ×400. (B) Comparison of TIL numbers between the groups that developed or did not develop recurrence. Ten areas from the tumor section were stained with various antibodies and analyzed at high magnification (×200) using image analysis software. Positive cell counts per field were compared between the recurrence group and the no recurrence group. Each value shows the mean±SD. **p<0.01, *p<0.05, statistically significant.
Association of PAX7 gene expression or other parameters with cancer prognosis. PAX7 is a transcription factor and plays critical roles during fetal development and cancer growth. Interestingly, PAX7 is known to form a fusion gene with the FKHR gene in rhabdomyosarcomas associated with TP53 mutation (24). Therefore, we investigated the prognostic potential of PAX7 gene expression. Using survival data from TCGA database, the PAX7-high group showed a worse prognosis than the PAX7-low group (Figure 6A). Meanwhile, among our 27 HNSCC patients, the association of driver mutations, TP53 mutation, recurrence and HPV status with overall survival was analyzed. Patients who developed recurrence and who had driver gene-mutations showed worse prognosis than those who did not develop recurrence and who did not have driver gene mutations, and those who had TP53-mutations tended to have shorter survival times than patients without mutations. Meanwhile patients with HPV-positive tumors showed no recurrence, had no driver mutations, and exhibited a tendency toward better prognosis (Figure 6B).
The effect of PAX7 gene expression and other parameters on the overall survival of patients with head and neck cancer. (A) PAX7 gene expression and clinical data of TCGA tumor samples downloaded from the UCSC Cancer Browser. Patients with head and neck cancer registered in the TCGA study were divided into high PAX7 expression (blue line) and low PAX7 expression (red line) groups by the median value. (B) The association of driver mutations, TP53 mutations, recurrence and HPV status with overall survival was analyzed. Recurrence (+) eight cases vs. recurrence (−) 19 cases, driver mutation (+) 15 cases vs. driver mutation (−) 12 cases, TP53 mutation (+) 12 cases vs. TP53 mutation (−) 15 cases, and HPV (+) seven cases vs. HPV (−) 20 cases. The prognostic significance of each parameter was evaluated using a Kaplan–Meier plot (log-rank test or generalized Wilcoxon test). **p<0.01, *p<0.05, statistically significant.
T cell receptor gene repertoire analysis in head and neck tumors. TCR repertoire profiling data did not show any association with the recurrence of HNSCC (Table V). However, HPV-positive cases exhibited a tendency toward higher number of total repertoires or higher DE50 (diversity evenness score) values compared to HPV-negative or recurrent HNSCC cases.
Association of TCRα and TCRβ repertoire profiling with tumor recurrence and HPV-status.
Discussion
Postoperative radiotherapy for locally advanced HNSCC with pathological high-risk features is a standard treatment strategy (25), but recurrence at an early follow-up time is still a clinical problem and achieving long-term local control and preventing early metastasis is a crucial issue.
The TME comprises stromal (cancer-associated) fibroblasts, immune cells (T cells, macrophages, etc.) and other supporting cells. Recently, advances in NGS-based clinical sequencing have widely extended the research targets from cancer cells to the entire TME. Genetic and immunological characterization studies of the TME based on NGS technology have demonstrated that tumors can be classified into several immune-types ranging from immune-suppressive to highly immunogenic, and some of them have been reported to be closely associated with the prognosis of cancer patients (7, 26, 27). Our group also reported that 1,734 tumors could be classified into 4 immune types based on PD-L1 and CD8B gene expression levels and that tumors with EGFR mutations showed immune-suppressive features (27). Regarding the TME status in HNSCC, many studies have demonstrated that specific genetic changes, such as TP53 mutation, contribute to the dysregulation of TME cells and cancer-associated fibroblasts (CAFs), the overproduction of cytokines and EMT, resulting in cancer invasion and metastasis (28, 29).
In the present study, we similarly demonstrated that patients who developed recurrence had a higher frequency of TP53 mutations than patients who did not develop recurrence, and down-regulation of the CDKN2 gene and the TP53 mutation/TGF-β/PAX7 pathway contributed to EMT induction (Figure 3). Additionally, patients who had TP53 mutations and subsequently developed recurrence exhibited cancer-related ERBB (EGFR) and VEGF pathway activation, and an immune-suppressive TME, which was achieved through a reduction in TIL number and function. Loss of function TP53 mutations are commonly identified in patients with advanced HNSCC and are closely associated with poor prognosis in patients with HNSCC (30, 31). Interestingly, loss of function TP53 mutation play an important role in cancer cell metabolism, where they can activate the glycolytic pathway and block the TriCarboxylic Acid cycle (Warburg effect) (32, 33). Our real-time PCR study showed the up-regulation of glycolytic enzyme gene (HK2 and PDK1) expression in patients who developed recurrence (Figure 2B), which may be mediated by the up-regulation of HIF1A gene expression, which contributes to radioresistance (34). Additionally, another metabolic factor which might be involved in the survival of HNSCC patients, the adenosine monophosphate deaminase 3 (AMPD3) gene, was reported to be significantly downregulated in head and neck cancer tissues (35).
The PAX7 gene is reported to be commonly up-regulated as is TGF-β1 in radiation-treated muscle tissues (36) and to affect the TP53 pathway by way of fusion gene construction (24). Importantly, PAX7 gene up-regulation in patients with TP53 mutations who subsequently developed recurrence has been suggested to be the first sign of recurrence, and high expression levels of PAX7 mRNA in HNSCC could be a possible poor prognostic marker according to TCGA dataset analysis (Figure 6A).
Moreover, HPV-infection status is another important clinical factor because HPV-positive HNSCCs have a better prognosis than HPV-negative HNSCCs (13). Chen et al. demonstrated, using immune-related gene expression profiling based on TCGA genome data, that the active immune signature group was associated with good prognosis and high HPV infection in head and neck cancers (37). In our study, seven patients with HPV-positive tumors did not develop recurrence and had no driver mutations or increased CD8+ TIL numbers. In contrast, patients who developed recurrence were characterized by HPV-negative tumors and a history of smoking. In summary, patients with HNSCCs who developed recurrence tended to have a higher frequency of driver and TP53 mutations, negative-HPV status, EMT induction and an immune-suppressive state, including a decrease in T cells and an increase in M2-type tumor-associated macrophages (TAMs) in the TME, than patients who did not develop recurrence.
Additionally, TCR profiling analysis using next generation sequencing (NGS) did not show any association with HNSCC recurrence. However, HPV-positive cases exhibited a tendency toward high number of total repertoires and high DE50 score (Table V), which might contribute to no recurrence and good prognosis in HPV-positive HNSCC patients.
In the future, to improve the survival of patients with locally advanced HNSCCs by achieving long-term local control and preventing metastases, novel therapeutic approaches that enable the reversal of TME conditions to antitumorigenic should include the following: 1) immune-checkpoint blockade (e.g., combinations of anti-PD-1/PD-L1 and anti-CTLA4 antibodies), 2) agents targeting M2-type TAMs or TGF-beta signaling pathway (38), and 3) strategies aimed at restoring the metabolic skew with metformin (39). A clinical trial using metformin was reported by Amin et al.; metformin treatment decreased intratumoral FOXP3+ T cell numbers and increased stromal CD8+ T cell numbers in patients with head and neck cancer (40). Hopefully, novel promising regimens following RT will be developed in upcoming clinical trials aiming to overcome EMT and immune-suppression and achieve long-term disease control.
There are several limitations in the present study. Bulk tumor samples collected at the time of tumor resection were used for GEP and WES analyses, which limited the investigation of intratumoral heterogeneity. Moreover, this study was conducted in a single institution with a limited number of patients; thus, selection bias is a significant concern in the present study. Nevertheless, patients with high PAX7 gene expression exhibited poorer prognosis in TCGA dataset than those with low PAX7 expression. To validate our findings, we need prospective observational studies of the correlation between PAX7 gene expression and HNSCC treatment response and tumor recurrence.
In the present study, patients with HNSCCs who developed recurrence were genetically and immunologically characterized and compared to patients who did not develop recurrence. The present study demonstrated that PAX7 gene expression might serve as a prognostic marker in HNSCCs.
Acknowledgements
The Authors thank the staff of the Shizuoka Cancer Center Hospital for assistance in sample preparation and the members of the Shizuoka Cancer Center Research Institute for discussions. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Footnotes
Authors’ Contributions
KY and YA participated in the design of the study and drafting of the manuscript and were responsible for completing the study. HM, AI and TA performed immunological in vitro experiments. TN, KO and KU performed genetic analysis using NGS and gene microarray. RK was responsible for the statistical analysis. KM and TS contributed to the preparation of pathological specimens. HO, TO, HH, HA, SM, TN, SG, SO, TM, SH and TY participated in collecting clinical samples and clinical data. YO and KY reviewed the manuscript. All Authors read and approved the final draft.
Conflicts of Interest
The Authors declare that they have no conflicts of interest in relation to this study.
- Received June 1, 2022.
- Revision received July 5, 2022.
- Accepted July 6, 2022.
- Copyright © 2022 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
