Programmed death ligand 2 expression plays a limited role in adenocarcinomas of the gastroesophageal junction after preoperative chemotherapy

Gastroesophageal adenocarcinoma (AEG) is one of the ten most common causes of cancer deaths worldwide [1, 2]. Even though the use of multimodal therapies combining surgery, cytotoxic chemotherapy, radiotherapy, and targeted therapy has improved patientsʼ survival rates, 5‑year survival rates remain poor, at 10–15% [3‐7]. The majority of patients present with locally advanced AEG and are therefore eligible for preoperative treatment [8, 9]. Perioperative chemo(radio)therapy along with surgery is now the standard of care treatment for patients with locally advanced AEG [10, 11]. However, due to the development of resistance to cytotoxic chemotherapy, these conventional therapies are limited in efficacy. Recently, targeted therapy strategies have become a promising approach to overcome resistance to cytotoxic chemotherapies. Human epidermal growth factor receptor 2 (HER-2) has been confirmed to play an important role in the treatment of AEG. Cytotoxic chemotherapy in combination with trastuzumab, a monoclonal antibody against HER‑2, significantly improves the prognosis in patients with HER-2-positive AEG [12]. Despite these promising data, recent studies also point out emerging resistance mechanisms to anti-HER2 therapy [13]. Therefore, new approaches for molecular-targeted therapeutics are needed to further improve patients’ survival. Cancer immunotherapy has revolutionized the field of oncology and immune checkpoint therapy provides promising treatment results in various cancer entities, including esophageal cancer [14, 15]. The therapeutic modulation of the PD‑1 pathway has emerged as a promising target in the treatment of melanoma, renal cancer, non-small cell lung cancer, and urothelial cancer [16]. Programmed cell death 1 (PD-1) is one of the most potent immune-checkpoint molecules, expressed mainly on the surface of T‑cells during activation. Binding of this co-receptor results in limited activation and function of the T‑cell and is therefore of great importance in avoiding hyperactivation [17]. Cancer cells can take advantage of this mechanism and misuse it to hide from immunosurveillance [18]. Programmed death ligand 1 (PD-L1) and programmed death ligand 2 (PD-L2) are cell surface glycoproteins belonging to the B7 family and are the two known ligands for PD‑1 [19, 20].

In advanced esophageal squamous cell cancer (ESCC), pembrolizumab, a monoclonal antibody targeting PD-L1, has been approved by the United States Food and Drug Administration (FDA) and significantly improves patient survival rates [21]. Recently, several studies could show that cytotoxic chemotherapy can affect the expression of PD-L1. So far, our group has investigated the prognostic value of PD‑1, PD-L1, and PD-L2 regarding survival rates of patients with AEG [22, 23]. Since neoadjuvant chemotherapy has become indispensable in the treatment of patients with locally advanced AEG, it is necessary to understand its impact on immune characteristics. To the best of our knowledge, there are no data on the neoadjuvant chemotherapeutic effect on the expression of PD-L2 in AEG until today. Therefore, the aim of this study was to investigate the potential change in the expression level of PD-L2 in neoadjuvantly treated AEG patients.

A total of 40 paired sets of diagnostic biopsies and surgical specimens from patients with neoadjuvantly treated AEG were included in this study. The median age of the patients in the study population was 64 years (range: 35.0–80.5 years) and the ratio of female to male was 1:2.3. The most frequent tumor localization (28 patients, 70.0%) was AEG type I according to the Siewert classification [28]. Before neoadjuvant chemotherapy, the majority of patients were diagnosed with UICC stage III (30 patients, 75.0%). After surgical resection, UICC stage III (25 patients, 62.5%) remained the most frequent staging. 23 (57.5%) patients presented with a moderately differentiated tumor (G2). Regarding the response to administered chemotherapy, 18 patients (45.0%) showed partial response (TRG 2–3), 15 patients (37.5%) showed minor regression (TRG 4), and 7 patients (17.5%) had no signs of regression at all (TRG 5). 34 (85.0%) patients presented with an ECOG (Eastern Cooperative Oncology Group) score of 0 before neoadjuvant chemotherapy and 15 patients (37.5%) were classified as ASA (American Society of Anesthesiologists) 1 before surgery. Clinico-pathological data are presented in Tab. 1.

The incidence of adenocarcinoma of the esophagogastric junction (AEG) has increased markedly over the past years in the western world. Its aggressive nature leads to early local invasion and systemic spreading [1, 2]. In recent years, cancer immunotherapy has become another pillar in cancer therapy. The clinical use of monoclonal antibodies targeting the PD‑1 pathway has achieved major improvements in patient survival in multiple malignancies. The immune suppressive proteins PD-L1 and PD-L2 limit T‑cell activation and cytokine production, which contributes to immunosurveillance [14, 29‐32]. Compared to PD-L1, the relevance of PD-L2 tumor-expression has not been explored intensively. Regarding AEG, only little information is available on its immune characteristics. Derks et al. were able to detect PD-L2 expression by cancer cells in 52% of patients with esophageal adenocarcinoma [33]. Recently, our group investigated the role of PD1, PD-L1, and PD-L2 expression as a prognostic factor in patients with resectable AEG. In that study, PD-L2 expression was only occasionally found on TILs and cancer cells (1.8 and 3.5%, respectively) [22, 23].

The standard approach for the treatment of patients with locally advanced AEG is cytotoxic chemotherapy or chemoradiotherapy followed by surgery. The use of cytotoxic chemotherapy is inevitable for treating advanced cancers. These drugs target nucleic acid metabolism or interfere with the microtubule network of proliferating tumor cells. In recent years, various studies reported that PD-L1 expression can be either decreased or increased by neoadjuvant chemotherapy according to the cancer type. These results suggested that the immune response, altered by the chemotherapeutical drugs, could be inhibited by upregulated PD-L1, which may play a role in chemoresistance [30, 31, 34‐40]. The potential change in PD-L2 expression has hardly been researched; regarding AEG, no data were available until now. With chemoresistance being a growing problem, it is important to further investigate the effects of chemotherapy on PD-L2 expression to gain more understanding of this issue in anti-cancer therapy.

In this study, we aimed to investigate the potential effect of oxaliplatin/capecitabine- and cisplatin/5-fluorouracil-based neoadjuvant therapy on the expression of PD-L2 in cancer cells and TILs in neoadjuvantly treated AEG using immunohistochemistry.

In our study, PD-L2 expression by AEG cancer cells and TILs was detected in only one out of 33 specimens of diagnostic biopsies. All the 40 surgical (post neoadjuvant therapy) specimens were rated as negative regarding PD-L2 expression. Therefore, we were not able to further investigate/illustrate a potential correlation between neoadjuvant treatment and the expression pattern of PD-L2. The positive expression by 6% of the specimens of diagnostic biopsies is a result that needs to be validated in a larger study population.

We have to consider certain limitations of our study. There is a potential selection bias, caused by only partial availability of tumor tissue, especially of diagnostic biopsies. The small amount of tissue obtained from a biopsy might not be sufficient to represent the PD-L2 expression rate of the tumor. Further, the retrospective nature of this single-center study represents another limitation. In general, the immunohistochemical methods for PD-L2 are not yet well established compared to those for PD-L1. Dhupar et al. have proposed that different criteria for positive staining of PD-L2 are a major reason for the generally inconsistent results [41]. Another aspect would be that patient cohorts from different studies vary regarding their demographic and pathologic factors, which can also affect the results of PD‑L expression analysis. Regarding the evaluation of the immunohistochemical expression grade, there is always a chance of interobserver variability, especially in borderline cases. Last but not least, an important limitation to mention, is the small study population.

Surprisingly, we found no PD-L2 expression in postsurgical specimens in this study. These findings are in contrast with our previously published data on the expression of PD-L2 as a potentially prognostic factor in AEG, describing 3.5% positive PD-L2 cases in patients with primarily resected and neoadjuvantly treated AEG [22]. The different expression rate may result from the inconsistent study populations, but it could also be interpreted as a small indication for a negative effect of the chemotherapy on PD-L2 expression. On the other hand, the non-existent PD-L2 expression in the present study is possibly related to the used antibodies. The issue of controversies attributed to different antibody specificity that leads to both underestimating and overestimating the results has been reported in various studies investigating PD-L1 and PD-L2 expression [41‐43].

To the best of our knowledge, this is the first study investigating the expression of PD-L2 in neoadjuvantly treated AEG patients.

In conclusion, we were not able to demonstrate a significant change in the expression of PD-L2 in AEG patients given neoadjuvant treatment. Our results can be interpreted as a tendency and definitely need further investigation. It is of great importance to gain better understanding into how cytotoxic chemotherapy affects PD-L2 expression and function to potentially contribute to develop new and highly effective anti-cancer therapies for patients suffering from AEG.