The expression of PD-L1 in salivary gland carcinomas

Objective was to analyze the role of PD-L1 and its relation to demographic, patho-clinical and outcome parameters in salivary gland carcinoma (SGC) patients. Patients treated for salivary gland carcinomas between 1994 and 2010 were included. A retrospective chart review for baseline characteristics, pathohistological, clinical and outcome data was performed. Immunohistochemistry for PD-L1 was performed using tissue microarrays. PD-L1 expression was assessed in tumor cells and tumor-infiltrating immune cells (TIIC) and statistical analysis with regard to baseline and outcome data was performed. Expression of PD-L1 (by means ≥1% of the cells with PD-L1 positivity) was present in the salivary gland carcinoma cells of 17%, in the TIIC of 20% and in both tumor cells and TIIC of 10% the patients. PD-L1 expression in tumor cells and both tumor cells and TIIC was related to tumor grading (p = 0.035 and p = 0.031, respectively). A trend towards higher grading was also seen for PD-L1 expression in TIICs (p = 0.058). Patients with salivary duct carcinomas and PD-L1 expressing TIICs showed a significantly worse DFS and OS (p = 0.022 and p = 0.003, respectively), those with both tumor cells and TIIC expressing PD-L1 a significantly worse DFS (p = 0.030). PD-L1 expression is present in 17% and 20% of salivary gland carcinoma cells and TIIC. Ten percent of the patient showed a PD-L1 positivity in both tumor cells and TIIC. This is related to high tumor grading and therefore might be a negative prognostic factor.


Materials and Methods
Patients. The study protocol was approved by the ethics committee of the local authority (Kantonale Ethikkommission Zürich (KEK); KEK-ZH-Nr. 2010-0206/0). All experiments and analyses were performed in accordance with the relevant guidelines and regulations. Informed consent was obtained. Patient with salivary gland carcinomas, who underwent treatment between January 1 st , 1994 and December 31 st , 2010 (17 years) were eligible to be included in this study. The same cohort was used in a previous study focusing on the expression of cancer testis antigens in salivary gland carcinomas 15 . Patients with incomplete data sets or insufficient tissue quality or quantity were excluded. Retrospective chart review focused on age at initial diagnosis, sex, tumor entity and localization, TNM classification (UICC, 7 th ed.), tumor grading, resection margins, extranodal extension of lymph node metastasis, perineural invasion, blood and lymphatic vessel invasion, recurrence and survival. Given that there is no universally used grading system for the different SGC entities, the grading system according to Brandwein et al. 16 was used for mucoepidermoid carcinoma (MEC) and the grading system according to Szanto et al. 17 for acinic cell carcinoma (AcCC). Grading of all the other SGC entities was bases on similarity to normal tissue of origin, grade of pleomorphism/anaplasia, vascular and perineural invasion and mitotic activity.
All patients underwent resection of their SGC, frequently treated with postoperative radiation therapy. Radiation therapy was indicated in case of large primaries with infiltration of surrounding tissues, high-grade histology, perineural or lymphovascular invasion, close or positive resection margins, multiple lymph node metastasis or lymph node metastasis with extranodal extension. None of the patients underwent treatment with PD-1/PD-L1 inhibitors.
Disease free survival (DFS) is the period of time without evidence of disease (i.e. persistence, recurrence or metastasis) after primary treatment. Disease-specific survival (DSS) and overall survival (OS) were defined as the duration after primary treatment to death due to SGC or any other reason, respectively. Oncologic follow-up was performed at the Department of Otorhinolaryngology, Head and Neck Surgery, University Hospital of Zurich. It consisted of clinical examinations and routine use of ultrasonography as well as cross-sectional imaging (MRI and/or PET-CT). Minimal oncologic standard surveillance time in SGC patients with salivary gland carcinomas is 10 years.
Tissue microarray construction/immunohistochemistry. To perform immunohistochemistry, sections of a tissue microarray (TMA), which has been previously described, were used 15 . Briefly, two core biopsies (diameter, 0.6 mm; length, 3-4 mm) from a morphologically representative area of interest of the paraffin 'donor' blocks were precisely arrayed into a new 'recipient' paraffin block using a stereomicroscope and the Beecher TMA instrument (Beecher Instruments, Sun Prairie, WI) 18 . From these ("recipient") TMA paraffin blocks, 3.0 μm sections were freshly cut and used for further immunohistochemistry. The staining procedure was conducted according to the manufacturer's instructions on an automated staining system (Leica Bond-III, Leica Biosystems, Wetzlar Germany). Briefly for PD-L1: Pretreatment Buffer H1, rabbit monoclonal anti PD-L1 (SP142) antibody (Spring Bioscience, Pleasanton, CA, USA) 19 . A tonsillectomy specimen served as positive control. A dilution of 1:100 resulted in a strong and distinct membranous staining pattern without unspecific background signal in positive controls (tonsil, Fig. 1). Staining intensity was analyzed using the percentage of positive cells (PP). Positivity for PD-L1 was defined as any unequivocal membranous staining of at least 1% of the tumor cells or the tumor infiltrating immune cells (TIIC) 1,20 . Multiple core's PP were averaged and rounded to the nearest whole number. Two different authors (K.I. and D.V.), both blinded for the patient's data, analyzed the percentage of positive tumor cells and TIIC.
Statistics. Data was analyzed using descriptive statistics. Analysis of variance (ANOVA) and chi-square tests were performed to analyze baseline characteristics. The relationship between the PD-L1 expression and the clinico-pathologic variables such as TNM classification or tumor grading (which were used in its non-dichotomized form) were correlated using cross tabulation and chi square test. DFS, DSS and OS were compared with Kaplan-Meier survival analyses and log rank tests. SPSS statistics (IBM Corp; Armonk, NY), version 24, was used to assist with statistical analysis. A p-value < 0.05 was considered statistically significant.
As expected the well-established prognostic factors TNM classification, tumor grading, perineural, vascular and lympho-vascular invasion had a highly significant impact on DFS, DSS and OS (all p < 0.018).
Baseline characteristics of the patients among PD-L1 positivity and negativity are depicted in Table 1 Table 3. In patients with PD-L1 positive TIIC, extent of immune infiltrate was significantly higher (p < 0.001).
PD-L1 expression in tumor cells and TIIC of to the various tumor entities are depicted in Table 4. Outcome analysis with regard to tumor subtype showed a worse DFS (Fig. 3) and OS (Fig. 4) in patients with salivary duct carcinoma and PD-L1 expressing TIIC (p = 0.022 and p = 0.003, respectively). This was also true with regard to DFS in patients with both SDC cells and TIIC being positive for PD-L1 (p = 0.030, Fig. 5). Due to the small size of the sub-cohort of patients with salivary duct carcinoma (n = 10), a multivariate analysis was not performed. All other subtypes of SGC did not show a relation between PD-L1 expression in tumor cells and/or TIIC and outcome.

Discussion
Our retrospective analysis of PD-L1 expression in salivary gland carcinoma cells and in its tumor infiltrating immune cells reports the following major findings:  www.nature.com/scientificreports www.nature.com/scientificreports/ in 51.1% of the patients with salivary gland carcinomas 22 . Other groups, which analyzed the immunoprofile of adenoid-cystic carcinoma, did not find any expression of PD-L1 6 or PD-1 positive SGC infiltrating T cells 23 . In our data, PD-L1 positivity of TIIC was furthermore associated with more extensive immune infiltrate. Consistent with the findings of Mukaigawa et al. 1 , PD-L1 expression in tumor cells and TIIC was associated with higher histological grading of SGC. However, a correlation with AJCC/UICC stage as reported by Harada et al. for membranous PD-L1 positivity 22 or a correlation with age, sex, tumor localization, T and N stage as found by Mukaigawa et al. 1 could not be confirmed with our data.
There is an ongoing debate about the prognostic role of PD-L1 (i.e. the prognostic implication of PD-L1 expression apart from the treatment with PD-1/PD-L1 blockers) with both favorable and unfavorable outcomes being reported in various malignancies [24][25][26][27][28] . Most of these authors conclude a negative impact of PD-L1 on outcome, which seems, looking to the data of Mukaigawa et al. 1 and Harada et al. 22 , also to be true for salivary gland carcinomas. This effect is explained by its role as an immunosuppressive molecule, which interacts with the PD-1 receptor and leads to tumor protection and immunotolerance [29][30][31] . Indeed, the link between PD-L1 expression in both tumor and immune cells can roughly be explained by interferon gamma, which is produced by the tumor infiltrating immune cells and is one of the molecular mechanisms inducing PD-L1 expression 25,32,33 . The www.nature.com/scientificreports www.nature.com/scientificreports/ prognostic relevance of PD-L1 in our study is limited to patients with salivary duct carcinomas, demonstrating an adverse effect on outcome in patient with PD-L1 positivity of TIIC.

PD-L1 in tumor cells PD-L1 in tumor-infiltrating immune cells PD-L1 in tumor and tumor-infiltrating immune cells
Salivary duct carcinoma (SDC) is a relatively rare, highly aggressive subtype of SGC, which most likely arises from the ductal epithelium and accounts for about 1-3% of all salivary gland malignancies 5,34-36 . It has one of the www.nature.com/scientificreports www.nature.com/scientificreports/ worst outcomes within the group of SGC. Patients with SDC show rapid progression of disease and a high incidence of metastasis (distant metastatic disease in 40-70%), which leads to tumor-related death within 3 years in the majority of the patients [37][38][39] . Due to its very low incidence, literature is limited to small retrospective studies and the optimal treatment remains still unclear, especially given the fact that current adjuvant treatments do not have a detectable impact on survival and the effect of chemotherapies for unresectable or metastatic tumors remains minimal 1,[5][6][7][8] .
Looking at the paucity of literature on SDC, it is difficult to discuss our findings by reviewing of the contemporary literature. To our knowledge, there is only one article considering PD-L1 receptor expression as a prognostic factor of SDC, which studied 67 patients with SDC or SDC ex pleomorphic adenoma. However, the authors Haderlein et al. failed to demonstrate a significant relationship between PD-L1 receptor expression and outcome of patients with SDC 21 . A more recent study by Sato et al. analyzed a small collective of 12 patients with SDC and found an unfavorable prognosis in patients with PD-L1 positive tumor cells 40 . Invasive ductal carcinoma of the breast, furthermore, resembles SDC and shows similar immunohistochemical staining patterns 37,41,42 . It is well known that PD-L1 expression in ductal carcinoma of the breast is correlated with high-risk features such as tumor grading and therefore could be an indicator of advanced stage and poor prognosis 30,43,44 . The adverse effect of PD-L1 expression in TIIC on outcome in the context of invasive ductal breast cancer can be explained by the following effects: An inhibition of T lymphocyte clonal expansion either by reverse signaling processes or by inhibition of other T lymphocytes by binding to its receptors (T cell-T cell interaction) 30,45,46 .