Detailed analysis of inflammatory cell infiltration and the prognostic impact on nasopharyngeal carcinoma

Transcription

1 Received: 7 April 2017 Revised: 8 October 2017 Accepted: 18 January 2018 DOI: /hed ORIGINAL ARTICLE Detailed analysis of inflammatory cell infiltration and the prognostic impact on nasopharyngeal carcinoma Juan Lu MD 1 Xiao-Mei Chen MD 1 Hao-Ran Huang MD 1 Fei-Peng Zhao MD 1,2 Fan Wang MD 1 Xiong Liu MD 1 Xiang-Ping Li MD 1 1 Department of Otolaryngology - Head and Neck Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China 2 Department of Otolaryngology-Head and Neck Surgery, The Affiliated Hospital of Southwest Medical University, Southwest Medical University, Luzhou, China Correspondence Xiang-Ping Li, Department of Otolaryngology - Head and Neck Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China. li321162@qq.com Funding information This work was funded by National Natural Science Foundation of China ( to X.P.L.) Abstract Background: One of the most striking characteristics of nasopharyngeal carcinoma (NPC) is the presence of a very abundant immune cells infiltrate containing mainly T-lymphocytes. The purpose of this study was to present our analysis providing a comprehensive characterization of antitumor inflammatory response in NPC. Methods: The densities of 9 types of inflammatory cells were assessed in 197 patients with NPC, including CD3 1 T-lymphocytes, CD8 1 cytotoxic T-lymphocytes, CD20 1 B-lymphocytes, CD56 1 natural killer (NK) cells, FOXP3 1 regulatory T- lymphocytes, CD1a 1 immature dendritic cells, CD83 1 mature dendritic cells, neutrophil elastase 1 neutrophils, and tryptase 1 mast cells. We characterized the inflammatory infiltrate in relation to clinical stage and patient survival. The expression of programmed death-1 (PD-1) on tumor-infiltrating lymphocytes (TILs) was also detected. The correlations between PD-1 expression and clinical characteristics and posttreatment outcome were analyzed. Results: The patients with NPC with a low density of tumor-infiltrating FOXP31, CD8 1 T-lymphocytes, neutrophils, and mast cells showed a significantly longer overall survival (OS) and progression-free survival (PFS). However, patients with a high density of NK cells showed a better OS and PFS. The densities of NK cells and mast cells could be served as biomarkers for predicting recurrence or distant metastasis in patients with NPC. Moreover, PD-1 positivity predicted poor prognosis in patients with NPC. Conclusion: The densities of inflammatory cells are correlated with the prognosis of patients with NPC. KEYWORDS inflammatory cells, metastasis, nasopharyngeal carcinoma, prognosis, recurrence 1 INTRODUCTION Nasopharyngeal carcinoma (NPC) is an endemic carcinoma in Southeast Asia with a high rate of local invasion and locoregional lymphatic metastasis. 1 One of the most striking and consistent characteristics of NPC is the presence of a very abundant lymphocyte infiltrate. 2 Among the pathologists Juan Lu and Xiao-Mei Chen contributed equally to this work. who originally described NPC, several authors named this tumor lymphoepithelioma. 2 Many studies reported that the lymphocyte infiltrate often accounted for 40% or 50% of the tumor mass. Infiltrating lymphocytes are often located in clusters around cancer cells but sometimes are disseminated within epithelial cell nests. Most of them are CD3 1 Tcells with a similar morphology of small resting T-lymphocytes. 3 Additionally, several studies have reported the presence of monocytes, eosinophils, and dendritic cells (DCs) in NPC Head & Neck. 2018;40: wileyonlinelibrary.com/journal/hed VC 2018 Wiley Periodicals, Inc. 1245

2 1246 biopsies. 4 Although several studies have reported the different subsets of tumor-infiltrating lymphocytes (TILs) and their potential prognostic value on NPC, 5 7 there are only limited amount of data on the interrelationships between different inflammatory cell types. In NPC, subpopulations of tumorinfiltrating immune cells have only been studied in parts, with comparably small patient cohorts. It has been reported that lymphocyte infiltration was a prognostic factor in patients with NPC, 5 7 however, many immunosuppressive mechanisms in the cancer microenvironment can inactivate T cells, leading to distant metastasis or recurrence after conventional treatment. Programmed death-1 (PD-1) is a kind of immunosuppressive receptor mainly expressed by activated T cells. The PD-1 engagement by programmed death-ligand 1 (PD-L1) in cancer cells decreases T cell activation and induces tumor immune escape. 8 It has been generally recognized that PD-1/PD-L1 axis is associated with aggressive histology and poor prognosis in multiple cancer types. 9 However, until now, only a few studies on PD-1 in NPC are available, and the reported prognostic role of PD-L1 in patients with NPC remains controversial. 7,10 12 Therefore, the prognostic role of PD-1 in NPC needs to be clarified. In this study, we analyzed a retrospectively recruited group of 197 patients with NPC with a median follow-up time of 51 months (the mean follow-up time 49.9 months), and systematically evaluated the densities and distribution of 9 inflammatory cell types in tumor cell nests and tumor stroma. The main purposes of the study were to enlighten the mutual relationship between different inflammatory cell types in NPC, and clarify the contributions of individual cell types to patient survival in a large, well-characterized series of patients with NPC with long-time follow-up. 2 MATERIALS AND METHODS 2.1 Patients The formalin-fixed, paraffin-embedded NPC biopsy specimens (n 5 197) were collected from Nanfang Hospital between January 2007 and December All biopsy samples were pathologically reassessed. The clinical characteristics of the study participants are presented in Supporting Information Table S1. None of the patients had received radiotherapy or chemotherapy before sampling. The clinical classification for NPC was based on the seventh edition Union for International Cancer Control (UICC)/American Joint Committee on Cancer staging system. All patients were treated with a uniform protocol of image-guided intensitymodulated radiotherapy or/and cisplatin-based concurrent chemotherapy after induction chemotherapy, in accord with the National Comprehensive Cancer Network Guidelines. We had registered the research on ClinicalTrial.gov (NCT ). Informed consent was obtained from all individuals, and the research protocols were approved by the Ethics Committee of Nanfang Hospital. The median followup was 51 months. The 5-year overall survival (OS) rate and progression-free survival (PFS) rate were 65.8% and 58.6%, respectively. 2.2 Immunohistochemistry Sections (3 lm) cut from paraffin-embedded specimens were deparaffinized in xylene and rehydrated through graded alcohol series. Microwave antigen retrieval was performed for 20 minutes in Tris-EDTA buffer (ph 8.0) and all sections were treated with 3% hydrogen peroxide in methanol for 15 minutes to block endogenous peroxidase activity. After neutralizing endogenous peroxidase activity, the sections were incubated overnight at 4 8C with diluted primary antibodies. A monoclonal mouse immunoglobulin G (IgG) antibody directed against CD20 (1:100; DAKO, Glostrup, Denmark), a monoclonal mouse IgG antibody directed against CD3 (1:25; DAKO), a monoclonal mouse IgG antibody directed against CD8 (1:50; DAKO), a monoclonal mouse IgG antibody directed against CD56 (1:50; DAKO), a polyclonal rabbit IgG antibody directed against CD83 (1:50; Abcam, Cambridge, UK), a monoclonal mouse IgG antibody directed against CD1a (1:50; DAKO), a monoclonal mouse IgG antibody directed against neutrophil elastase (1:100; DAKO), a monoclonal mouse IgG antibody directed against FOXP3 (1:50; Abcam), a monoclonal rabbit IgG antibody directed against PD-1 (1:250; Abcam), and a monoclonal mouse IgG antibody directed against tryptase (1:100; DAKO) were used. The 3, 3 0 -diaminobenzidine was used as the chromogen and hematoxylin as the counterstain. For the negative controls, the primary antibody was replaced with phosphatebuffered saline. 2.3 Evaluation Immune cells were identified as either T-lymphocytes (CD31), cytotoxic T-lymphocytes (CTLs; CD81), B- lymphocytes (CD201), natural killer (NK) cells (CD561), regulatory T-lymphocytes (Tregs, FOXP31), immature dendritic cells (idcs, CD1a1), mature dendritic cells (mdcs, CD831), neutrophils (neutrophil elastase1), and mast cells (tryptase1). We counted the inflammatory cells using an earlier described computer-based method based on their size, shape, and the intensity of positivity. 13 The cases were grouped by using the median as a cutoff. The percentages of PD-1-positive lymphocytes and staining intensity were evaluated by 2 pathologists who counted 20 sequential high-power fields (0.54-mm field diameter) judged to be representative of the sample, while remaining blinded to the clinical information. Staining intensity was

3 1247 scored as follows: 0 for negative or trace; 1 for weak; 2 for moderate; and 3 for high. The percentage of stained cells (0%-100%) was multiplied by the dominant intensity pattern of staining ranging from 0 to 3. Therefore, the overall semiquantitative score (H-score) ranged from 0 to 300 (maximum value of 300 corresponding to 100% of lymphocytes positive for PD-1 with an overall staining intensity score of 3). 2.4 Statistical analysis All statistical analyses were conducted using SPSS 19.0 software (SPSS, Chicago, IL). Spearman s rankcorrelation test was used in determining the correlation between different immune cell types. The associations between immune cells and clinicopathological characteristics were analyzed with the Mann-Whitney U test, the Kruskal- Wallis test, and the chi-square test. Survival curves were plotted by the Kaplan-Meier method and compared by the log-rank test. The significance of various variables for survival was analyzed by the Cox proportional hazards model in the multivariate analysis. Receiver operating characteristic analysis was used in the recurrence analysis. A 2-sided P value <.05 was considered statistically significant. 3 RESULTS 3.1 Characteristic distribution of inflammatory infiltrate in nasopharyngeal carcinoma Immunohistochemistry was used to characterize the inflammatory infiltrate and it was successful in all 197 cases (Supporting Information Figure S1). The densities of inflammatory cells were shown in Supporting Information Table S2. The CD3 1 T cells, CD8 1 T cells, and FOXP3 1 Tregs were observed both in the intratumoral stroma and tumor epithelium. The CD20 1 B cells, neutrophil elastase1, tryptase1 mast cells, CD56 1 NK cells, and CD83 1 mdcs were mainly distributed in the intratumoral stroma. The CD1a 1 idcs were only distributed in the tumor epithelium. The CD3 1 T cells were the most frequent in all the tumor locations, followed by CD8 1 T cells, CD20 1 B cells, and FOXP3 1 Tregs. In the intratumoral stroma, the median CD8 1 and FOXP3 1 T cell numbers were 59.1% and 25.2% of the amount of CD3 1 T cells, respectively. The B cells were quite frequently encountered in every slide as well as in the intratumoral stroma. The number of NK cells was very low. Of the infiltrating DCs, the majority was a subset of the CD1a 1 idcs (76.2%). 3.2 Interrelationships between different inflammatory cells in nasopharyngeal carcinoma We analyzed the interrelationships between different inflammatory cells by calculating Spearman s correlation coefficients (Supporting Information Table S3). Different inflammatory cells had high positive correlations with each other. However, the statistical strength of correlations varied. Although, for CD3 1 T-lymphocytes and CD8 1 T- lymphocytes, strong correlations showing r values well above 0.5 were observed, FOXP3 1 Tregs and CD3 1 T cell counts showed only a moderate association (r ). This correlation was even weaker for NK cells and mast cells (r and r , respectively). The densities of CD3 1 T-lymphocytes, CD8 1 T-lymphocytes, and CD20 1 B-lymphocytes were quite strong, positively linked to each other, whereas the density of CD1a 1 idcs seemed to be independent, with no positive or negative correlation with another 8 immune cell types. 3.3 Associations between immune cell infiltration and clinicopathological parameters in nasopharyngeal carcinoma As shown in Supporting Information Table S4, age and sex were not associated with significant differences in the densities of 9 inflammatory cells (Supporting Information Tables S5 and S6). Higher UICC stages were significantly associated with a modest increase in mast cells and Tregs infiltrates (Supporting Information Table S4). In contrast, Tregs and NK cell counts increased by trend in locally more advanced tumors (Supporting Information Tables S4 and S5). Moreover, an increase of densities of mast cells, Tregs, and CD1a 1 idcs in N1 classification was observed, which were much higher than other N classifications (Supporting Information Tables S4-S6). The densities of CD20 1 B-lymphocytes, CD3 1 T- lymphocytes, as well as CD83 1 mdcs were significantly higher in undifferentiated carcinoma than in nonkeratinizing carcinoma and keratinizing squamous cell carcinoma (Supporting Information Tables S4-S6). No striking differences were noted between CD8 1 T cells as well as neutrophil counts and clinicopathological parameters in NPC. 3.4 Prognostic impact of tumor-infiltrating inflammatory cells in nasopharyngeal carcinoma Classical survival prognosticators, such as UICC stage and others, proved to have high prognostic impact on our NPC cohort (Supporting Information Table S7). Patients who had lower mast cell, Treg, CD8 1 T-lymphocyte, and neutrophil counts had a significantly better prognosis when compared with those patients whose tumor tissues showed higher than

4 1248 FIGURE 1 Kaplan-Meier survival curves of overall survival (OS) and progression-free survival (PFS) stratified for inflammatory cell infiltrates in patients with nasopharyngeal carcinoma. A, FOXP3 1 regulatory T-lymphocytes, B, CD8 1 cytotoxic T-lymphocytes, C, tryptase 1 mast cells, D, CD56 1 natural killer cells, and E, neutrophil elastase 1 neutrophils. The P values were calculated with a log-rank test these 4 immune cells counts (Figure 1A,C,E). The higher density of NK cells positively significantly correlated with a longer OS and PFS in patients with NPC (Figure 1D). The presence of CD20 1 B-lymphocytes, CD3 1 T-lymphocytes, and DCs either immature or mature was not associated with significant differences in OS and PFS (data not shown). Multivariate analysis under inclusion of UICC stage revealed the number of NK cells (OS: HR 0.456, P 5.004; PFS: HR 0.489, P 5.004), mast cells (OS: HR 2.004, P 5.009; PFS: HR 1.964, P 5.008), and neutrophils (OS: HR 2.052, P 5.007; PFS: HR 1.636, P 5.014) as significantly independent predictors of OS and PFS (Table 1). Moreover, the prognostic impacts of CD8 1 T cells (HR 1.700; P 5.036) and Tregs (HR 1.714; P 5.034) on OS were proved in our NPC cohort. 3.5 Natural killer cells combined with mast cells infiltration as a predictor of recurrence or metastasis of nasopharyngeal carcinoma Finally, we carried out a recurrence analysis for 149 of 197 patients (75.6%), of whom 58 (38.9%) had a recurrence or

5 1249 TABLE 1 Summary of multivariate Cox regression analysis of overall survival and progression-free survival duration OS PFS Variables P value HR 95% CI P value HR 95% CI CD FOXP CD Neutrophil elastase Tryptase CD1a Abbreviations: CI, confidence interval; HR, hazard ratio; OS, overall survival; PFS, progression-free survival. Figures in boldface indicate statistical significance. distant metastasis. Of the 197 patients, 48 (24.4%) were excluded because they had distant metastasis when diagnosed or had treatment failure after the completion of radiochemotherapy. We explored the potential of tumorinfiltrating inflammatory cells as biomarkers for predicting recurrence and distant metastasis in NPC. The receiver TABLE 2 Recurrence analysis Variables AUC 95% CI T classification N classification UICC stage FOXP CD CD CD CD Tryptase Neutrophil elastase CD1a CD CD56 1 CD CD56 1 FOXP CD56 1 NE CD56 1 tryptase Log-rank P value Abbreviations: AUC, area under curve; CI, confidence interval; UICC, Union for International Cancer Control. Figures in boldface indicate statistical significance. operating characteristic analysis was used to compare the discriminatory capacity of different variables. As shown in Table 2, NK cells combined with mast cells infiltration yielded the highest area under the curves, which was even higher than N classification and classical UICC stage. In addition, the log-rank test indicated significant associations with recurrence or metastasis for NK cells combined with several other cell types, including CD8 1 T cells, Tregs, and neutrophils. 3.6 Prognostic value of programmed death-1 expression The PD-1 positive immune cells were presented in 96 of 197 patients (48.7%) in a scattered manner. Moreover, PD-1 staining was mainly located at the membrane or in the cytoplasm region (or both) of T cells. Representative staining of PD-1 in NPC are shown in Figure 2. The PD-1 positivity was defined as cases with PD-1 staining intensity 2 in > 5% of TILs. As shown in Table 3, PD-1 positivity was significantly correlated with clinical stage (P 5.031). In addition, we found that PD-1 positivity was significantly correlated with shorter OS and shorter PFS (see Figure 3). 4 DISCUSSION In this study, we aimed to enlighten the interrelationships between different inflammatory cell types and the prognostic impact of immune cell infiltration in a large and wellcharacterized NPC cohort. To our knowledge, this is so far the most extensive study on this subject, including markers of both adaptive (CD3, CD8, CD20, and FOXP3) and innate immunity (CD56, neutrophil elastase, and mast cell tryptase), as well as antigen presenting cells (CD1a and CD83). We found that the densities of inflammatory cells were correlated with the prognosis of patients with NPC.

6 1250 FIGURE 2 Representative immunohistochemical (IHC) staining of programmed death-1 (PD-1) in nasopharyngeal carcinoma samples. A and C, The PD-1 staining in a biopsy evaluated as PD-1-negative. The IHC pictures are shown at A, low magnification: 20X and C, high magnification: 40X. B and D, The PD-1 staining in a biopsy evaluated as PD-1-positive. The IHC photographs are shown at B, low magnification: 20X and D, high magnification: 40X It is well-known that the abundance of leukocyte infiltrate containing mainly T-lymphocytes is a major characteristic of NPC tumor stroma. Our results showed that the most prevalent inflammatory cell infiltration was CD3 1 T- lymphocytes. We also found that the presence of tumor infiltrating CD8 1 CTLs was associated with a rapid fatal outcome in NPC, similar to previous results. 6,7 An explanation might be the functional inactivation of CD8 1 CTLs in NPC. Li et al 14 have demonstrated that CD8 1 CTLs could increase PD-1 expression and reduce CD3f expression, resulting in an impaired tumor-specific immunity. In addition, CD8 1 CTLs may upregulate the expression of PD-L1 and indoleamine-2,3-dioxygenase in tumor cells, recruit Tregs in the tumor microenvironment, and then promote tumor immune escape by the production of CCL22 and interferon-g. 7,15,16 Some discrepancy exists for the prognostic significance of Treg cells. High Treg cell infiltration has been associated with poor survival in renal cell carcinoma 17 and breast cancer, 18 which is in accordance with the role of Treg cells in dampening immune response. However, some studies found that a higher infiltration of Treg cells was correlated with better prognosis in colorectal, head and neck, and bladder cancers. 19,20 The prognostic role of Treg cells is even controversial in NPC. Some studies have shown that Tregs in the peripheral blood of patients with NPC were higher than healthy people, and tumor-infiltrating Tregs were also increased, supporting the view that Tregs play an important role in suppressing antitumor immunity in NPC. 16,21 Another TABLE 3 Clinicopathologic variables and immune-activity status of programmed death-1 in patients with nasopharyngeal carcinoma Characteristics No. of patients PD-1 positive P value Sex.729 Male (49.3%) Female (47.1%) T classification.238 T (44.7%) T (51.3%) T (44.4%) T (53.0%) N classification.385 N (47.6%) N (48.0%) N (50.8%) N (47.4%) M classification.136 M (48.9%) M (47.1%) UICC stage.031 I 10 3 (30.0%) II (36.7%) III (43.5%) IV (57.9%) Abbreviations: PD-1, programmed death-1; UICC, Union for International Cancer Control.

7 1251 FIGURE 3 Programmed death-1 (PD-1) positivity predicts poor prognosis. Kaplan-Meier survival curves of patients with nasopharyngeal carcinoma shows that high expression of PD-1 predicts shorter A, overall survival (OS) and B, progression-free survival (PFS) study found that the density of FOXP3 1 Tregs was negatively correlated with T classification and clinical stage and positively correlated with OS and PFS in 106 cases of NPC. 5 Our results support the majority of the published results linking Treg cells with unfavorable outcome, 19 as low FOXP3 1 Treg counts had the highest association with significantly longer OS and PFS. Nevertheless, the mechanisms accounting for the role of Tregs in NPC and for the inconsistencies in their roles in different cancers remain unclear. Mast cells are identified as modulators of tumor microenvironment, which directly regulate tumor cell proliferation and invasion, or indirectly help tumors by promoting tumor angiogenesis, remodeling tumor microenvironment, and modulating immune responses. 22 The functional contribution of mast cells to the inflammatory response of solid tumors is not well understood. Until now, the distribution and role of mast cells have rarely been documented in NPC. Here, we showed that mast cells were common in the intratumoral stroma, and the density was positively correlated with N classification and clinical stage, suggesting that mast cells might play a key role in promoting NPC progression and metastasis. We further confirmed that high mast cell infiltration was an independent prognostic indicator of poor survival. More importantly, the data showed that mast cell infiltration might be served as a predictive biomarker for predicting recurrence and metastasis in patients with NPC. However, the sensitivity and specificity of these predictive biomarkers needs to be further confirmed with bigger samples sizes and a long-term follow-up. Our results were consistent with the findings in colorectal cancer, 23 pancreatic ductal adenocarcinoma, 24 and prostate cancer. 25 In addition to mast cells, NK cells are cytotoxic innate immune cells specialized in defense against tumor and virusinfected cells. The NK cells could quickly identify and dissolve mutant cells without antibody participation and sensitization in advance, by perforin/particle enzyme-mediated cytotoxic pathways or FAS/FASL-mediated apoptosis pathway. 26 However, the distribution and prognostic role of NK cells in NPC are still unknown. Zheng et al 27 reported that NK cell counts in the peripheral blood of patients with NPC was much higher than that of healthy subjects, and the cytolytic effect of the NK cells against NPC cells in the patient group was also higher than that of the donors. In this study, high NK cell infiltration was an independent favorable prognostic indicator of good survival. More importantly, the data supported that mast cells combined with NK cell infiltration might be served as a predictive biomarker for predicting recurrence and metastasis in patients with NPC. Our data were similar to the results in gastric cancer 28 and breast cancer, 29 supporting the antitumor role of NK cells in NPC. The PD-1 is a transmembrane receptor, which is mainly expressed on T cells. In tumor tissues, PD-1 is mainly expressed in TILs. High expression of PD-1 in TILs can lead to depletion and deactivation of T cells. 8 The PD-1 also interacts with PD-L1 to inhibit T cell proliferation and promote the immune escape of cancer cells, playing an important role in immune suppression and cancer progression. 9 Recent studies have shown that NPC has high levels of PD- L1 and PD-1, indicating that NPC may be a candidate for PD-1/PD-L1-targeted therapies However, the reported prognostic role of PD-L1 in patients with NPC remains controversial. Hsu et al 7 found that PD-1 expression on CD8 1 T cells predicted a poor prognosis in a small size of patients with NPC, whereas PD-1 expression might not be a prognostic factor in patients with NPC with a larger sample size. 11 In our study, we found that PD-1 positivity was significantly correlated with shorter OS and shorter PFS, suggesting PD-1 as a potential prognostic biomarker for patients with NPC. The differences might be illustrated by heterogeneous applications of different experimental procedures, scoring criteria, sample size, and survival endpoints. In conclusion, we report that the presence of CTLs, Treg cells, neutrophils, and mast cells is correlated with a poor prognosis, and high numbers of tumor-infiltrating NK cells are positively correlated with a good prognosis. In addition, we found that the number of NK cells combined with mast cells could be served as biomarkers for predicting recurrence or distant metastasis in patients with NPC. These data provide strong support for the biological role of inflammatory infiltrate in NPC and encourage future further preclinical and

8 1252 clinical studies exploring immunomodulation for the benefit of patients with NPC. AUTHOR CONTRIBUTIONS Conducted most of the experiments and analyzed data: Lu, Chen Wrote the manuscript: Wang Provided clinical samples: Huang Edited the manuscript: Zhao Supervised the study: Liu, Li ORCID Xiang-Ping Li, MD REFERENCES [1] Wei WI, Sham JS. Nasopharyngeal carcinoma. Lancet. 2005; 365: [2] Gourzones C, Barjon C, Busson P. Host-tumor interactions in nasopharyngeal carcinomas. Semin Cancer Biol. 2012;22: [3] Herait P, Ganem G, Lipinski M, et al. Lymphocyte subsets in tumour of patients with undifferentiated nasopharyngeal carcinoma: presence of lymphocytes with the phenotype of activated T cells. Br J Cancer. 1987;55: [4] Leighton SE, Teo JG, Leung SF, Cheung AY, Lee JC, van Hasselt CA. Prevalence and prognostic significance of tumor-associated tissue eosinophilia in nasopharyngeal carcinoma. Cancer. 1996;77: [5] Zhang YL, Li J, Mo HY, et al. Different subsets of tumor infiltrating lymphocytes correlate with NPC progression in different ways. Mol Cancer. 2010;9:4. [6] Oudejans JJ, Harijadi H, Kummer JA, et al. High numbers of granzyme B/CD8-positive tumour-infiltrating lymphocytes in nasopharyngeal carcinoma biopsies predict rapid fatal outcome in patients treated with curative intent. J Pathol. 2002;198: [7] Hsu MC, Hsiao JR, Chang KC, et al. Increase of programmed death-1-expressing intratumoral CD8 T cells predicts a poor prognosis for nasopharyngeal carcinoma. Mod Pathol. 2010;23: [8] Boussiotis VA. Molecular and biochemical aspects of the PD-1 checkpoint pathway. N Engl J Med. 2016;375: [9] Gibney GT, Weiner LM, Atkins MB. Predictive biomarkers for checkpoint inhibitor-based immunotherapy. Lancet Oncol. 2016; 17:e542-e551. [10] Zhou Y, Shi D, Miao J, et al. PD-L1 predicts poor prognosis for nasopharyngeal carcinoma irrespective of PD-1 and EBV-DNA load. Sci Rep. 2017;7: [11] Zhang J, Fang W, Qin T, et al. Co-expression of PD-1 and PD- L1 predicts poor outcome in nasopharyngeal carcinoma. Med Oncol. 2015;32:86. [12] Tang Y, He Y, Shi L, et al. Co-expression of AFAP1-AS1 and PD-1 predicts poor prognosis in nasopharyngeal carcinoma. Oncotarget. 2017;8: [13] Väyrynen JP, Vornanen JO, Sajanti S, B ohm JP, Tuomisto A, Mäkinen MJ. An improved image analysis method for cell counting lends credibility to the prognostic significance of T cells in colorectal cancer. Virchows Arch. 2012;460: [14] Li J, Zeng XH, Mo HY, et al. Functional inactivation of EBVspecific T-lymphocytes in nasopharyngeal carcinoma: implications for tumor immunotherapy. PLoS One. 2007;2:e1122. [15] Spranger S, Spaapen RM, Zha Y, et al. Up-regulation of PD- L1, IDO, and T(regs) in the melanoma tumor microenvironment is driven by CD8(1) T cells. Sci Transl Med. 2013;5: 200ra116. [16] Yip WK, Abdullah MA, Yusoff SM, Seow HF. Increase in tumour-infiltrating lymphocytes with regulatory T cell immunophenotypes and reduced zeta-chain expression in nasopharyngeal carcinoma patients. Clin Exp Immunol. 2009;155: [17] Siddiqui SA, Frigola X, Bonne-Annee S, et al. Tumor-infiltrating Foxp3-CD41CD25 1 T cells predict poor survival in renal cell carcinoma. Clin Cancer Res. 2007;13: [18] Bates GJ, Fox SB, Han C, et al. Quantification of regulatory T cells enables the identification of high-risk breast cancer patients and those at risk of late relapse. J Clin Oncol. 2006;24: [19] Zou W. Regulatory T cells, tumour immunity and immunotherapy. Nat Rev Immunol. 2006;6: [20] Fridman WH, Pagès F, Sautès-Fridman C, Galon J. The immune contexture in human tumours: impact on clinical outcome. Nat Rev Cancer. 2012;12: [21] Lau KM, Cheng SH, Lo KW, et al. Increase in circulating Foxp31CD41CD25(high) regulatory T cells in nasopharyngeal carcinoma patients. Br J Cancer. 2007;96: [22] Liu J, Zhang Y, Zhao J, et al. Mast cell: insight into remodeling a tumor microenvironment. Cancer Metastasis Rev. 2011;30: [23] Gulubova M, Vlaykova T. Prognostic significance of mast cell number and microvascular density for the survival of patients with primary colorectal cancer. J Gastroenterol Hepatol. 2009; 24: [24] Chang DZ, Ma Y, Ji B, et al. Mast cells in tumor microenvironment promotes the in vivo growth of pancreatic ductal adenocarcinoma. Clin Cancer Res. 2011;17: [25] Nonomura N, Takayama H, Nishimura K, et al. Decreased number of mast cells infiltrating into needle biopsy specimens leads to a better prognosis of prostate cancer. Br J Cancer. 2007;97: [26] Waldhauer I, Steinle A. NK cells and cancer immunosurveillance. Oncogene. 2008;27: [27] Zheng Y, Cao KY, Ng SP, et al. Complementary activation of peripheral natural killer cell immunity in nasopharyngeal carcinoma. Cancer Sci. 2006;97: [28] Ishigami S, Natsugoe S, Tokuda K, et al. Prognostic value of intratumoral natural killer cells in gastric carcinoma. Cancer. 2000;88:

9 1253 [29] Ascierto ML, Idowu MO, Zhao Y, et al. Molecular signatures mostly associated with NK cells are predictive of relapse free survival in breast cancer patients. J Transl Med. 2013;11:145. SUPPORTING INFORMATION Additional Supporting Information may be found online in the supporting information tab for this article. How to cite this article: Lu J, Chen X-M, Huang H-R, et al. Detailed analysis of inflammatory cell infiltration and the prognostic impact on nasopharyngeal carcinoma. Head & Neck. 2018;40: org/ /hed.25104