Article Text


Response to: ‘Patterns of PD-L1 expression and CD8 T cell infiltration in gastric adenocarcinomas and associated immune stroma’
  1. Sascha Rahn1,
  2. Sandra Krüger2,
  3. Christoph Röcken2,
  4. Ole Helm1,
  5. Susanne Sebens1
  1. 1 Institute for Experimental Cancer Research, Christian-Albrechts-University Kiel (CAU) and University Medical Center Schleswig-Holstein (UK-SH), Kiel, Schleswig-Holstein, Germany
  2. 2 Department of Pathology, Christian-Albrechts-University Kiel (CAU) and University Medical Center Schleswig-Holstein (UK-SH), Kiel, Schleswig-Holstein, Germany
  1. Correspondence to Professor Susanne Sebens, Institute for Experimental Cancer Research, Christian-Albrechts-University Kiel (CAU) and University Medical Center Schleswig-Holstein (UK-SH), Kiel, 24105 SH, Germany; susanne.sebens{at}

Statistics from

We read with interest the manuscript by Thompson et al recently published in Gut 1 highlighting the expression of programmed death ligand 1 (PD-L1) within the stromal compartment of gastric adenocarcinomas (GC). This study showed that PD-L1 is rather expressed by stromal than neoplastic cells and primarily at the tumour margin of PD-L1positive GC. Moreover, they identified a high proportion of CD8+ T cells within these areas that, in contrast to previous studies, correlated with worse progression-free survival and overall survival.1 We fully agree with their findings because of very similar results obtained regarding PD-L1 expression in a well-characterised cohort of pancreatic ductal adenocarcinomas (PDAC).2 Immunohistochemical PD-L1 staining revealed marked PD-L1 expression in 17 of 59 cases (30.4%) (figure 1A). Furthermore, pan-cytokeratin/PD-L1 costainings validated that in 76.5% of the PD-L1positive subgroup PD-L1 is predominantly or exclusively expressed by stromal cells (figure 1B). In line with the findings in GC tissues, PD-L1high areas in PDAC tissues were mainly located at the tumour margin and showed high CD3+ T cell infiltration (online supplementary figure 1). Notably, immunostaining of serial tissue sections demonstrated that CD8+ T cells and CD68+ tumour-associated macrophages (TAM) represent the major immune cell populations within these areas while CD4+ T cells were almost absent (online supplementary figure 1). Moreover, the prominent coincidence of CD68 and PD-L1 in situ as well as preliminary results with in vitro differentiated macrophages indicates that PD-L1 is predominantly expressed by TAM in dependence on T cell secreted cytokines (unpublished data). Based on our findings, we agree with the conclusion by Thompson et al and others that in certain tumours PD-L1 expression is associated with a ‘T cell inflamed type’1 3–5 which is, however, compromised in its function. Thus, in these cases PD-1/PD-L1 blockade might be more beneficial to restore T cell reactivity than in tumours in which PD-L1 expression is restricted to neoplastic cells in the absence of effector T cell infiltration. Therefore, we agree with Thompson et al that more precise studies on PD-L1 expression in neoplastic and stromal cells considering intratumoural localisation and stromal composition are needed to better understand its tumour biological relevance and, thereby, develop diagnostically conclusive approaches for improved patient stratification. In line with the findings by Thompson et al our results indicate that the different clinical outcomes of the patients within our cohort (87% staged as T3N1M0) might be explained by differences in PD-L1 expression and the extent of immune cell infiltrates.

Supplementary file 1

Figure 1

Programmed death ligand 1 (PD-L1) is predominantly expressed by stromal cells in pancreatic ductal adenocarcinoma (PDAC). (A) Proportion of PD-L1negative/low and PD-L1high classified PDAC tissues in 59 patients. (B) Cellular origin of PD-L1 expression within the PD-L1high subgroup was determined by pan-cytokeratin (PanCK)/PD-L1 immunohistochemical costaining and categorised into epithelial (PanCK+) and stromal (PanCK) PD-L1 expression. Representative images of PanCK (red)/PD-L1 (brown) stainings show PD-L1+ neoplastic (left, white arrowheads) and stromal cells (right, black arrowheads). Scale bar: 50 µm.

In general, the assessment of PD-L1 status in solid tumours by current methods is controversially discussed.6–9 Several preclinical studies proposed PD-L1 expression as a prognostic marker but most clinical trials revealed a rather poor or absent correlation of treatment responses with PD-L1 status classified by established methods. Thus, clinical indication for application of PD-1/PD-L1 blocking antibodies has been mainly defined by the failure of standard oncological treatment regimens, yet. Encouraged by the findings of Thompson et al and our own data, indicating that tumour-associated PD-L1 expression highly varies in terms of cellular origin, intratumoural localisation and expression level (unpublished data), we are currently developing a novel scoring system that comprehensively considers all these facets. Thereby, we hope to provide a tool for improved PD-L1 classification in PDAC and better prediction of treatment responses to PD-1/PD-L1 targeted approaches in future clinical trials.


View Abstract


  • Contributors SR, SK and OH performed the experiments. SR and CR evaluated immunohistochemical stainings. SR, OH and SS analysed the data. SR and SS wrote the manuscript. All authors have read and corrected the manuscript.

  • Funding This work was supported by the Stiftung für Krebsentstehung & Immunologie.

  • Competing interests None declared.

  • Ethics approval Ethics Committee of the University Hospital Schleswig-Holstein (reference number: D430/09).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.