The role of GLUT-1 in upregulation of PD-L1 expression after radiotherapy and PD-L1 is associated with a favourable overall survival in hypopharyngeal cancer

Purpose: Although the alteration of tumor immunity after radiotherapy (RT) has been studied widely in recent years, how radiotherapy mediates tumor immunity and whether glycolysis is involved in the mediation in hypopharyngeal cancer are still unclear. This study aimes to determine whether radiotherapy regulates programmed cell death ligand 1( PD-L1) partly via glucose transporter 1(GLUT-1) expression and whether PD-L1 expression predicts overall survival (OS) in patients with hypopharyngeal cancer. Methods: PD-L1, Glut-1 expression and C D4+, CD8+ T cell were detected by immunohistochemistry analysis on 47 pre-RT and 25 post-RT specimens of hypopharyngeal cancer. The changes of these indicators before and after radiotherapy were compared, and their association with overall survival of patients were analyzed. Moreover, we used siRNA-GLUT-1 to inhibit GLUT-1 expression and determined whether GLUT-1 was a key factor involved in mediation of PD-L1 expression by RT in vitro. Results: In multivariate analysis, patients with higher PD-L1 expression (P=0.037), higher CD4+ T cell infiltration (P=0.016) and earlier clinical stage (P=0.019) had favourable OS. The PD-L1 expression and CD4+, CD8+ T cell increased significantly after RT. PD-L1 expression was correlated with Glut-1 in pre-RT (P=0.002), but not after-RT(P=0.051). The PD-L1 expression of FaDu cells was upregulated after RT, especially at 96h after RT in vitro. However, the PD-L1 expression of siRNA-GLUT-1 FaDu cells was significantly decreased at 96h after RT when compared with FaDu cells. Conclusion: The patients with high PD‑L1 expression and CD4+ T cell infiltration might have favourable OS in hypopharyngeal cancer. RT could increase PD-L1 expression and alter tumor immunity, the expression of PD-L1 was correlated with Glut-1, and inhibiting GLUT-1 expression might decrease the expression of PD-L1. GLUT-1 might participate in the alteration of tumor immunity after RT. and hexokinase 2(HK-2) expression in metabolic reprogramming after irradiation. GLUT-1 controls glucose uptake and HK-2 encodes the key kinase involved in glycolysis. They found that after radiation the two genes expression declined while the PD-1 expression elevated in both activated CD4 + and CD8 + populations relative to the unirradiated wild type C57BL/6 male mice[12]. So how irradiation mediate tumor immunity of hypopharyngeal cancer and whether glycolysis is involved in the mediation


Introduction
Head and neck squamous cell carcinoma (HNSCC) represents about 6% of all cancers, and there are approximately 644,000 new cases and 352,000 cancer deaths every year in the world [1]. Despite comprehensive treatments such as surgery, radiotherapy, and chemotherapy, the 5-year survival rate 3 of HNSCC has not significantly improved. Since 1975, 5-year survival of laryngeal cancer has fallen from 67-61%, while oral cavity and pharynx cancer has increased merely from 53-63% [2]. Therefore, there is an urgent need to find more effective treatments. Recently, targeted therapy for programmed cell death 1 (PD-1) and PD-L1 has shown enormous prospects for tumor treatment. Even after traditional therapies have failed, immune checkpoints blockade treatments could increased the progression free survival of advanced stage patients more than two years [3]. Despite encouraging results, many cancer patients have low respond rates to clinical checkpoint blockade, the response to anti-PD-1/PD-L1 therapy is only approximately 18%-25% response rate in HNSCC [4,5]. To overcome the unsatisfactory respond rates, combining other strategies, including radiotherapy, chemotherapy, other immunotherapy and targeted therapy, has become a commonly used strategy for HNSCC.
Preclinical studies have shown that RT could upregulate PD-L1 expression on tumor cells, RT and anti-PD-L1 therapy had synergistically antitumor effect [6]. In the clinical studies, patients who received RT before anti-PD-L1 therapy had a better prognosis than those who did not receive RT in Lung Cancer [7,8], mammary cancer [9]. However, the knowledge about PD-L1 upregulation after RT in hypopharyngeal cancer is scarce [10], this has led us to interest in studying the expression of PD-L1 after RT in hypopharyngeal cancer. Moreover, little is known about how RT-mediated immune responses alter the tumor immunity. The alteration of PD-1/PD-L1 expression after RT is dependent on multiple factors such as signaling cascades, general somatic mutation prevalence, individual genetic background, tumor environment, therefore it cannot be generalized. Previous studies have found that tumor glycolysis and tumor immune evasion are interdependent. Glycolytic activity was a stronger predictor for tumor immunity in a number of cancers, glycolysis could increase PD-L1 expression in tumors [11]. Li D et al assessed Glut-1 and hexokinase 2(HK-2) expression in metabolic reprogramming after irradiation. GLUT-1 controls glucose uptake and HK-2 encodes the key kinase involved in glycolysis. They found that after radiation the two genes expression declined while the PD-1 expression elevated in both activated CD4 + and CD8 + populations relative to the unirradiated wild type C57BL/6 male mice [12]. So how irradiation mediate tumor immunity of hypopharyngeal cancer and whether glycolysis is involved in the mediation are need to be studied further. 4 Moreover, PDL1 expression level has been associated with the therapeutic response and prognosis in diverse cancers. Jiang C et al demonstrated that positive PDL1 expression was associated with a higher survival in patients of esophageal squamous cell carcinoma (ESCC) who underwent RT, they found that the patients with high PDL1 expression had an increased infiltration of Tumor Infiltrating Lymphocytes (TILs), these patients had highly immunogenic tumors prior to RT, it was a independently predictor of favourable prognosis for patients with ESCC [13]. However, the data on the prognostic value of PD-L1 expression in hypopharyngeal cancer remains limited, so it is important to demonstrate the clinical significance of PD-L1 expression in patients with hypopharyngeal cancer.
The present study aimed to focus on the immune related changes of tumor micro-environment after RT in hypopharyngeal cancer and whether RT regulates PD-L1 expression partly via GLUT-1 expression, and to determine the association of PD-L1 expression with OS in patients of hypopharyngeal cancer.

Ethics Statement
The study has been approved by the appropriate institutional committee and have been performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. The institutional review board of The First Affiliated Hospital, College of Medicine, Zhejiang University (Hangzhou, Zhejiang, China) approved the present study. Written informed consent was obtained from all individual participants included in the study.

Clinical data
Formalin-fixed paraffin-embedded tissue of 61 patients with locally advanced hypopharyngeal cancer treated with (chemo)radiotherapy and surgery at the First Affiliated hospital, College of Medicine, Zhejiang University between 2008 and 2017 was collected, among them, 14 patients were subsequently excluded due to the unavailable tumor samples or missing clinical data. Excluded criteria: Patients with distant metastasis at diagnosis, patients who received preoperative targeted therapy, immune therapy or chemotherapy, patients with more than one malignancy.
All patient datas were reviewed for the following baseline characteristics: gender, age, primary tumor histologic subtype, tumor location, clinical stages. RT parameters were also recorded, including total dose, dose per fraction, time between RT and sample resection.

Follow-up
OS data for patients were obtained by outpatient service or telephone. OS was determined from the date of diagnosis to the death of patients. A follow-up examination was performed every month during the first year, every three months during the second year, and every six months after the third year. In addition to routine physical examinations, patients underwent laryngoscopy, cervical CT or magnetic resonance imaging (MRI), or whole body PET/CT. The final reaction product was developed by exposure to 0.03% diaminobenzidine, and the nuclei were counterstained with hematoxylin.

Immunohistochemical analysis and evaluation
We scored the percentage of tumor cells with PD-L1 positive stained, increasing by 5% increments, and used a semiquantitative scoring method to evaluate PD-L1 expression, high PD-L1 expression was defined at a minimum of 10% of stained cells.
The staining intensity of Glut-1 was classified as no staining, weak, moderate, and strong intensity for 0, 1, 2, or 3 scores, respectively. The percentage of stained cells was classified as follows: 0-25% stained cells for 1 score, 26-50% stained cells for 2 score, 51-75% stained cells for 3 score, and 6 >75% stained cells for 4 score. The Glut-1 expression was assessed semi-quantitatively using the product of these scores (intensity × percentage of stained cells): 0-5 points = negative expression and 6-12 points = positive expression.
To evaluate CD4+, CD8+ T cells infiltration in tumor, we counted the numbers of CD4+ and CD8+ T cells in a selected hotspot under 400×magnification and selected the median number of CD4+, CD8+ T cells as the cut-off point for CD4+ and CD8+ T cells density.
Two experienced pathologists calculated the staining intensity and the percentage of stained cells independently.

Tumor cell line irradiation
To determine whether GLUT-1 is a key factor involved in the mediation of tumor immune mircoenvironment by radiotherapy, we inhibited GLUT-1 expression using GLUT-1 siRNA. GLUT-1 siRNA was purchased from GenePharma Co. Ltd. (Shanghai, China). The sequences were: sense, 5'-GGAAUUCAAUGCUGAUGAUTT-3'; antisense, 5'-AUCAUCAGCAUUGAAUUCCTT-3'. We performed the GLUT-1-siRNA transfection when the cells reached 50% confluence. The FaDu cell and siRNA-GLUT-1 FaDu cell were both seeded at a density of 10,000-20,000 per 25 cm 2 and were divided into 4 goups respectively: control group, 24h, 48h, 96h group. Tumor cells were subjected to radiation after resting overnight except the control group. RT was taken using an X-ray generator (22.7 mA, 120 kV, variable time; GE Inspection Technologies, Germany) with a single dose of 10 Gray on day 1. The tumor cells of 24h, 48h, 96h group were harvested at 24h, 48h and 96h after the radiotherapy respectively, PD-L1 and GLUT-1 expression on tumor cells were analyzed. The control group tumor cells were harvested and analyzed on day2.

Western blotting
Tumor cells were lysed in Radio Immunoprecipitation Assay (RIPA) lysis solution and were separated by gel electrophoresis and transferred to membranes. We blocked the membranes with 5% non-fat dry milk in TBST and soaked in the primary antibody buffer at 4°C overnight (PD-L1 1:800 dilution(Proteintech, Chicago, IL, USA, Art No: 66248-1-1g), (GLUT-1 1:800 dilution (Proteintech, Chicago, IL, USA, Art no: 20960-1-AP). We soaked the membranes in secondary antibody buffer and incubated at room temperature for for 2h. Enhanced chemiluminescence was used to visualize the proteins and then the proteins were exposed to X-ray film. Protein expression was analyzed semiquantitatively using the ChemiDoc XRS+ System (Bio-RAD, USA).

Statistical analysis
SPSS software (ver. 22.0; SPSS, Inc., Chicago, IL, USA) was used for statistical analyses. Categorical variables were assessed by χ2 or Fisher's exact tests. Correlation analyses were performed using 8 Spearman's rank analysis. Changes in PD-L1, Glut-1 expression and CD4+, CD8+ T cells before and after RT were tested with Student's t-test. The Kaplan-Meier method and log-rank test were used to calculate survival curves and compare the results. The Cox proportional hazards regression model was used for multivariate analysis. P values <0.05 were considered to indicate statistical significance.
PD-L1 expressed mainly on membranes (cell surface) and partially in the cytoplasm of tumor cells, and PD-L1 expression exhibited heterogeneity in the specimens, a cutoff of 10% was used to define "low" vs. "high" expression. The percentage of stained tumor cells in the patient population ranged from 3.7-23.1% before RT. 53.2% (25/47) patients were considered with high PD-L1 expression. Glut-1 was expressed on cell membranes in tumor cells, and its expression exhibited positive in 51.1% (24/47) patients. Based on the median number of CD4 + and CD8 + cells respectively, a cutoff of 8% and 25% were used to define "low" vs "high" CD4 + and CD8 + T cells infiltration respectively. The  Table 1.    (Fig. 1B, Fig. 3A).
Glut-1 expression, CD4 + and CD8 + T cell infiltration were also studied in pre-RT and post-RT specimens. CD4 + T cell infiltration in tumor tissues increased obviously after RT with statistical signifificance. for the pre-RT specimens, the proportion of CD4 + T cells was only 8.6%, it increased remarkably after RT reaching 21.0%(p < 0.001) (Fig. 3B). Meanwhile, CD8 + T cell infiltration in tumor tissues also increased obviously after RT with statistical signifificance. The proportion of CD8 + T cells in pre-RT specimens was only 24.9%, it increased clearly after RT reaching 35.7%(p = 0.006) (Fig. 3C).
However, Glut-1 expression only had an increased tendency in post-RT specimens compared with pre-RT specimens(p = 0.097).
However, CD4 + and CD8 + T cell infiltration status were not correlated with the expression of Glut-1 neither in pre-RT nor after-RT.
RT induced PD-L1 upregulation on FaDu cells in vitro, and GLUT-1 might be a key factor of the mechanism To study the influence of RT on PD-L1 expression, FaDu cells were treated with a single fraction of 10 Gy in vitro. PD-L1 mRNA expression of FaDu cells was upregulated after 10 Gy of RT compared with the control group, the upregulation was most significant at 96 h after RT (p < 0.001) (Fig. 4A).
The PD-L1 protein expression of FaDu cells was also upregulated after RT, and at 96 h after RT it was significantly higher than the control group(p < 0.001) (Fig. 4B).
After RT the GLUT-1 mRNA expression of FaDu cells decreased firstly, and then increased. It decreased at 48 h after RT compared with the control group(p < 0.01), and it increased significantly at 96 h after RT compared with 48 h after RT (p < 0.001) (Fig. 4C). The alteration of Glut-1 protein expression was mainly the same as GLUT-1 mRNA expression. Glut-1 protein expression of FaDu cells decreased at 48 h after RT compared with the control group and the 24 h after RT group (p < 0.01, p < 0.05 respectively), and at 96 h after RT it increased significantly compared with 48 h after RT(p < 0.001) (Fig. 4D).
we used GLUT-1 siRNA to inhibit GLUT-1 expression (Fig. 5A, 5B). At the siRNA-GLUT-1 FaDu cells, the PD-L1 mRNA and protein expression were also upregulated significantly at 96 h after RT compared with the control group (p < 0.01, P < 0.001 respectively). More importantly, at 96 h after RT the PD-L1 mRNA and protein expression of siRNA-GLUT-1 FaDu cells was significantly decreased when compared with that of FaDu cells as presented by the two-way analysis of variance (p < 0.001, P < 0.001 respectively) (Fig. 5C, 5D). This result demonstrated that inhibiting the expression of GLUT-1 could interfere the increasing of PD-L1 expression after RT.

Discussion
In 2016, the FDA approved the application of anti-PD-1 monoclonal antibodies pembrolizumab and nivolumab to the treatment of HNSCC, which opened a new page in HNSCC treatment. Now FDA has approved anti-PD-1/PD-L1 drugs for head and neck cancers and there are many clinical trials worldwide. However, the dark side of this therapy has emerged with the time going, including drug resistance. Therefore, combination therapy have been adopted to improve the efficacy [15].
Diverse studies have examined the expression of the PD-L1 in HNSCC and found that the expression levels of PD-L1 was between 46% and 100% depending on the fixation, staining method and 13 site [16][17][18], however, there is very little research on hypopharyngeal cancer. In our study, we found that the PD-L1 expressed in 100% patients of hypopharyngeal cancer, the percentage of stained tumor cells in the patient population ranged from 3.7-23.1% before RT.  [20] . Liu YJ et al showed that low PD-L1 expression on tumor cells significantly correlates with local recurrence in EBV-positive nasopharyngeal carcinoma patients after RT [21].

Combined low PD-L1 expression on inflammatory cells and tumor cells is an independent negative
prognostic factor for OS in rectal adenocarcinoma [22]. Adversely, Lin YM et al found that high PD-L1 expression was associated with poor outcome and metastasis in Oral Squamous Cell Carcinoma [23].
Furthermore, several clinical reports suggested that a high density of tumor-infiltrating lymphocytes was associated with favorable prognosis in patients with locally advanced non-small cell lung cancer [24], and brease cancer [25]. In addition to cytotoxic effect, RT could induce anti-tumor immune effect [26]. The effects of RT on tumor microenvironment and its interaction with tumor immunity is a complex balance of suppressing and activating signals [27]. Sufficient evidence has shown that RT might enhanced the therapeutic effects of anti-PD-1/PD-L1 agents in HNSCC [28], however, many questions remain regarding how RT affects tumor immunity in hypopharyngeal cancer.
In fact, RT plays a role in the recruitment of T cells in the tumor microenvironment [29] and increases PD-L1 expression in tumors [30]. However, the knowledge about upregulation of PD-L1 expression and TILs recruitment in tumor cell after exposure to RT in HNSCC especially in hypopharyngeal cancer is scarce. Our study demonstrated that PD-L1 expression was increased after RT in hypopharyngeal cancer patients, and in the vitro experiment we found PD-L1 expression was increased after RT at colon adenocarcinoma and TUBO mammary carcinoma [30,31]. Herter-Sprie GS observed that the CD8+/Treg ratio increased 96 hours after RT in Kras-mutant lung cancer [32]. In clinical study, some data also demonstrated that radiotherapy is correlated with an increased PD-L1 expression in rectal adenocarcinoma [22] and locally advanced esophageal adenocarcinoma [33]. Similar to our results, Keung EZ found that the PD-L1 expression in undifferentiated pleomorphic sarcoma of the extremity and trunk (ET-UPS) increased after radiation and an increase in median number of CD4+, CD8 + T cells after RT was also recognized, moreover, although PD-L1 was not expressed at baseline, positive PD-L1 expression was observed in 21% (3/14) of ET-UPS tumor cells after RT [34].
Oweida A et al demonstrated that RT could transform the immune landscape of tumors and render poorly immunogenic murine orthotopic HNSCC sensitive to anti-PD-L1 drugs [35]. There were indeed a number of reports in the preclinical and clinical studies to show that radiotherapy and immunotherapy have the synergistic anti-tumor effect in other tumors [6,30,36].
Therefore, the changes in the immune microenvironment after radiotherapy make the tumor more sensitive to immunotherapy, but the mechanism of how radiotherapy interferes with the tumor immune microenvironment is still unclear and needs further studies.
The alteration of PD-1/PD-L1 expression after RT is dependent on multiple factors such as the signaling cascades, individual genetic background, general somatic mutation prevalence, tumor mircoenvironment and it cannot be generalized. Radiotherapy kills tumor cells by free radical-induced DNA damage, and also promotes metabolic changes in tumor cells, ie, metabolic reprogramming.
Metabolic reprogramming means the activity and expression of enzymes and their regulators involved in the metabolic activities of tumor cells altered, involving multiple metabolic pathways, the most important is the glycolysis pathway. Aerobic glycolysis and immune escape are two major features of tumors [37]. The dependence of immune cell proliferation and activity on cell metabolism has received increasing attention [38].
GLUT is an important energy transporter that mediates the Warburg effect. The glucose transporter is a protein that mediates the transmembrane transport of glucose, which is a major reason for the in both CD8 + and CD4 + populations. They also studied the metabolic reprogramming parameters and found that the expression of Glut1 and HK-2 decreased in activated T cells after RT compared with unirradiated controls [12]. Li HH et al also found that RT influence T cell activation via metabolic reprogramming [39].
Previous studies showed that tumor immune evasion and tumor glycolysis are interdependent [40,41]. Metabolic reprogramming of tumor cells is an important biochemical basis for tumor immune escape. Enhanced tumor glycolysis attenuates the clearance of tumor cells by immune cells [41]. It is increasingly appreciated that immune cell proliferation and function is dependent on cellular metabolism [38]. A recent study reported that PD-L1 expression is regulated by GLUT-1 in clear cell renal cell carcinoma [42], in pulmonary pleomorphic carcinoma [43]. Chang CH et al found that the metabolic competition between tumor cells and immune cells may induce to tumor immunosuppression, the competition for glucose in tumor microenvironment could drive cancer progression, it would occur when tumors surpass T-cells for glucose supply, impeded their IFN-γ production, which is critical for anti-tumor activity [40].
Glycolytic activity was a more consistent and stronger predictor for immune signatures in a number of cancers. Glycolytic activity increases anti-PD-1/PD-L1 immunotherapy effects via enhancing PD-L1 expression on tumor cells. Thus, glycolytic activity of the tumor cells could be a predictive factor for immunotherapy response in diverse cancers [11]. Many metabolic mechanisms are thought to be related to tumor immune escape and could serve as co-targets in immunotherapy [44].
Therefore, we hypothesized that RT can change the tumor immunity, increase PD-L1 expression and content of CD4+, CD8 + T cells in hypopharyngeal cancer, the mechanism may be that RT interferes with cell glycolysis, alleviates competition for glucose with T cells, metabolic reprogramming after radiotherapy is one of the causes of tumor immune changes. But this still needs further research.
The present study had some limitations. Firstly, this was a single-institution retrospective study with a small sample size, not a trial-based correlative study, there was bias due to the small sample size and its retrospective design. The second limitation is that PD-L1 immunohistochemistry was conducted using only one cut-off value and one antibody, the cutoffs used for high PDL1expression in the present study was inconsistent with some other studies, and there was no unified standard for PDL1 expression positivity. Finally, as no standardized PDL1 immunohistochemistry assay is available currently, caution should be taken in the interpretation of these results.

Conclusion
The present study demonstrates for the first time that high PDL1 expression and CD4 + T cell

Consent for publication
Not applicable.
Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no conflict of interest.

Funding
This study was supported by the National Natural Science Foundation of China (No. 81372903 and 81172562).

Authors' contributions
Shui-Hong Zhou and Yu Guo analyzed and interpreted the patient data regarding hypopharyngeal cancer. Li-Fang Shen performed the histological examination of the hypopharyngeal cancer, and was a major contributor in writing the manuscript. All authors read and approved the final manuscript.  FaDu cells at 96h after RT was significantly decreased when compared with that of FaDu cells as presented by the two-way analysis of variance (p<0.001, P<0.001 respectively). (***p<0.001).