PD-L1 expression in angiomatoid fibrous histiocytoma

Angiomatoid fibrous histiocytoma (AFH) is a rare tumor of intermediate malignancy. Treatment options for unresectable and/or metastatic tumors are very limited. Immunotherapy with PD-1/PD-L1 inhibitors may be worth exploring. The aim of this study was to evaluate the expression of PD-L1 in AFHs. PD-L1 expression was assessed on 36 AFHs from 36 pediatric patients by immunohistochemical staining of PD-L1 (clone 22C3). Positivity was defined as membranous expression in ≥ 1% of either tumor or immune cells. The correlations between PD-L1 expression and clinicopathologic features were assessed. Two patients had lymph node metastasis. All patients underwent surgical resection; three of them also had systemic chemotherapy. Three patients had recurrence after initial resection; all patients were alive with a median follow-up of 2.5 years. Overall, twenty-two (61%) tumors were positively stained for PD-L1 and positivity was seen on both tumor and immune cells in eighteen of the 22 tumors. A positive correlation was found between tumor cell PD-L1 expression and CD8+ T-cell infiltration. There were no statistically significant differences between the status of PD-L1 expression and the clinicopathological features assessed. PD-L1 expression was identified in 61% of AFHs with a predominantly adaptive pattern. Our findings provide a rationale for future studies evaluating the potential of checkpoint immunotherapy for patients with unresectable and/or metastatic tumor.

www.nature.com/scientificreports/ EWSR1-CREB1 fusion and one case with an EWSR1-ATF1 fusion. No other clinically significant gene mutation or gene amplification events were identified in these four cases. CD99 positivity was seen in 17/17 cases tested. Most cases were also positive for CD68 (23/24) and desmin (16/21). Four patients had a paraneoplastic syndrome. All tumors underwent surgical resection. Three patients had recurrence after initial resection, and two of three patients underwent re-excision with no evidence of residual disease at last follow-up. Two patients had lymph node metastasis at the time of diagnosis. Three patients also received chemotherapy. All patients were alive with a median follow-up of 2.5 years (range: 1 month to 13 years). Two patients were alive with disease at the last follow-up. Twenty-four (65%) tumors showed typical peripheral lymphoid cuffing and all tumors demonstrated intratumoral lymphoplasmacytic infiltrates.

Association between PD-L1 expression and CD8+ T-cell infiltration. Positive CD8 immunostain-
ing highlights cytotoxic/suppressor T cells, which are mediators of adaptive immunity. We found that 81.8%  Table 1). Representative immunostaining of CD8 was shown in Fig. 2. A moderate to strong positive correlation was found between tumor cell PD-L1 expression and CD8+ T-cell infiltration (r = 0.727, p = 0.000). The correlation between immune cell PD-L1 expression and CD8+ T-cell infiltration was also significant (r = 0.486, p = 0.003). The findings further supported that PD-L1 was expressed in AFH in an adaptive pattern.

Association between PD-L1 expression and clinicopathologic features. The relationship between
the PD-L1 expression and clinicopathological parameters of AFH is summarized in Table 2. There were no statistically significant differences between the clinicopathological features assessed and the status of tumor or immune cell PD-L1 expression. However, there appeared to be a weak trend toward correlation. Both patients with metastasis showed positive PD-L1 expression in both tumor and immune cells. Most notably, the adrenal AFH, which had diffuse metastasis and an EWSR1-ATF1 fusion, showed epithelioid histology with strong and diffuse PD-L1 expression. (Fig. 1H,I). Three of four patients with paraneoplastic syndrome had positive immune cell expression of PD-L1.

Discussion
To our knowledge, this is the first study of PD-L1 expression in AFH by IHC. We found that PD-L1 was expressed in more than half of AFHs with a predominantly adaptive pattern. The majority of AFHs are curable by complete surgical resection. However, in rare instances, tumors may present with unresectable disease and or develop metastasis. Due to the rarity of unresectable and metastatic disease, there is a limited number of case reports available on the treatment of advanced disease. In a case with unresectable and metastatic AFH, chemotherapy with vincristine, doxorubicin, dactinomycin, and cyclophosphamide achieved a complete response 16 . Six cycles of ifosfamide and doxorubicin followed by three courses of ifosfamide monotherapy were tried with a good response in a AFH with local recurrence and metastases 17 . More recently, Potter et al. 23 described a clinical response to tocilizumab, an IL-6 receptor inhibitor antibody, in a child with metastatic AFH, whose disease progressed during treatment with traditional chemotherapeutic agents (vincristine, dactinomycin, and cyclophosphamide). In this cohort, all patients underwent primary surgical resection, and three patients recurred locally at follow-up. Two patients had metastatic disease, one of them was an unresectable adrenal gland AFH with diffuse lymph node metastases. Overall, three patients went on to receive systemic chemotherapy. At last follow-up, all patients were alive and two had residual disease.
It is known that AFH is associated with three characteristic gene fusions-EWSR1-CREB1, EWSR1-ATF1, and rarely FUS-ATF1. In this cohort, EWSR1 rearrangement was documented in 16 of 17 cases analyzed and there were more cases with EWSR1-CREB1 fusion than with EWSR1-ATF1 fusion. Unfortunately, so far, no therapies targeting either EWSR1 or FUS fusion proteins are available. Therefore, other novel treatment options for unresectable and metastatic tumors are needed.
PD-L1 is an immune modulator whose interaction with the receptor PD-1 on primed cytotoxic T cells inhibits T cell-mediated tissue damage [24][25][26] . A number of tumor types utilize this regulatory mechanism to evade immune surveillance by overexpressing PD-L1 26 . Previous studies reported that PD-L1 expression level predicts which patients are more likely to respond to anti-PD-1 or anti-PD-L1 therapy 27,28 . In this study, the correlation between positive PD-L1 expression, by either tumor cells or infiltrating immune cells, and patient's gender, age, tumor location, tumor size, or status of metastasis did not reach statistical significance. It might be due to relatively small number of samples in each groups as there is a weak trend toward correlation seen in Table 2. Notably, both instances of metastatic disease in our series showed positive PD-L1 expression in both tumor and immune cells, one of them with strong and diffuse tumor cell PD-L1expression and high CD8+ T-cell infiltration. Additionally, three of four patients with paraneoplastic syndrome had positive PD-L1 expression. The clinical significance of PD-L1 expression in AFH awaits further research, but theoretically, PD-L1 expression by www.nature.com/scientificreports/ any cell type exerts local immunosuppressive effects leading to evasion of immune surveillance. Therefore, we speculate that immunotherapy with a PD-L1 inhibitor may be a potential option for a subset of AFH patients. Several anti-PD-L1 antibodies are currently utilized for tumor staining. In this study, we used anti-PD-L1 antibody clone 22C3, which is an FDA approved companion diagnostic antibody for pembrolizumab and a complementary diagnostic antibody for other PD-1/PD-L1 inhibitors 29 . However, our immunostaining was performed by use of Leica bond Autostainer instead of pharmDx assay. Cutoff values (percentage of positive cells) were established on clinical trials for various different carcinomas, ranging from 1 to 50% 29,30 . In this study, the majority of AFHs showed positive staining in > 1% but < 10% of tumor cells (15 cases). Therefore, we arbitrarily used 1% as the initial cutoff. Additional work will be necessary to validate this cutoff value for AFH.
In conclusion, PD-L1 expression was identified in 61% of pediatric AFH cases with a predominantly adaptive pattern of expression. Our findings provide a rationale for future studies evaluating the potential of checkpoint immunotherapy for patients with unresectable and/or metastatic tumor.

Materials and methods
Clinical data. Thirty six well-characterized AFH cases from 36 patients with available formalin-fixed paraffin embedded tissues were retrospectively collected from the following institutions: Children's Hospital Los Angeles, Children's Healthcare of Atlanta, Washington University in Saint Louis, University of Oklahoma Health Sciences Center, Cincinnati Children's Hospital and Rhode Island Hospital. The diagnosis of all cases was confirmed by two pathologists (JB and SZ) based on the morphologic features reviewed, and the immunohistochemical and molecular profiles, where applicable, from the submitting institutions.
Ewing sarcoma breakpoint region 1 (EWSR1) gene rearrangements were assessed by fluorescence in-situ hybridization (FISH) using an EWSR1 break-apart probe in 11 cases, and an EWSR1 and CREB1 dual fusion probe in 2 cases. Four cases were analyzed by the use of the OncoKids SM cancer panel 31 . Multiple immunohistochemical stainings such as CD68, CD99 and desmin were performed in most of cases.
PD-L1 and CD8 immunohistochemical staining and scoring. Representative sections from resection specimen of each primary tumor were stained for both PD-L1 and CD8 immunohistochemically along with appropriate controls by use of a Leica Bond Max Instrument (Leica, Buffalo Grove, IL). Tissue sections (4 μm) were deparaffinized and rehydrated using the Leica Bond Max De-Wax solution. For PD-L1 staining, antigen www.nature.com/scientificreports/ retrieval was performed with Leica Bond ER2 solution (pH 9.0, EDTA based buffer) for 30 min, then the slides were incubated with anti-PD-L1 antibody (Dako, clone 22C3, dilution 1:50) at ambient temperature for 60 min. Next the antibody was detected using the BOND Polymer Refine Detection kit (Leica), which contains a peroxide block, post primary, polymer reagent, DAB chromogen and hematoxylin counterstain. For CD8 staining, antigen retrieval was performed with Leica Bond ER1 solution (pH 6.0, citrate based buffer) for 20 min, then the slides were incubated with anti-CD8 antibody (Leica, ready-to-use) at ambient temperature for 15 min. Next the antibody was detected using the BOND Polymer Refine Detection kit (Leica), which contains a peroxide block, post primary, polymer reagent, DAB chromogen and hematoxylin counterstain. PD-L1 was scored based on a semi-quantitative "eyeballing" evaluation by two pathologists (JB and SZ), and cases with discordance were adjudicated by consensus. In this study, PD-L1 positivity was defined as membranous expression in ≥ 1% of either tumor and or immune cells. Tumor cells were distinguished from immune cells on IHC by histologic pattern and size. Tumor cells were typically arranged in syncytial clusters and sheets, and PD-L1( +) tumor cells were much larger and had more abundant cytoplasm than PD-L1 ( +) immune cells.
To get a better idea of PD-L1 expression pattern, the percentage of tumor and intratumoral immune cells were evaluated separately, and further scored as 1 + (1% to 10%), 2 + (11% to 30%) and 3 + (> 30%) based on percentages of positive cells. Adaptive PD-L1 expression pattern was defined as PD-L1 positivity seen in both tumor and infiltrating immune cells 32  Statistical analysis. The categorical parameters between the PD-L1( +) and PD-L1(−) groups were compared with the chi-square test or the Fisher exact test, where appropriate. The association between PD-L1 expression and CD8+ T-cell infiltration was evaluated using Spearman's correlation analysis. Statistical analysis was performed using IBM SPSS statistics software, version 17 (IBM Corp, Armonk, NY, USA). All tests were twosided and p values < 0.05 were considered significant.
Ethics approval. This study was undertaken at Children's Hospital Los Angeles (CHLA) with the approval of the Institutional Review Board (IRB) (CHLA-18-00278), and independently approved or exempted by the IRB of the participating institutions. All methods were carried out in accordance with relevant guidelines and regulations. This study was granted a waiver of informed consent/assent/permission and a waiver of HIPAA authorization per the Privacy Rule from the parent and/or legal guardian of all patients for publication of identifying information in an online open-access publication by the CHLA IRB.