Characterizing diversity in the tumor-immune microenvironment of distinct subclasses of gastroesophageal adenocarcinomas - PubMed (original) (raw)
. 2020 Aug;31(8):1011-1020.
doi: 10.1016/j.annonc.2020.04.011. Epub 2020 May 6.
L K de Klerk 2, X Xu 3, T Fleitas 4, K X Liu 5, Y Liu 6, F Dietlein 7, C Margolis 7, A M Chiaravalli 8, A C Da Silva 9, S Ogino 10, F G Akarca 11, G J Freeman 11, S J Rodig 12, J L Hornick 9, E van Allen 7, B Li 13, S X Liu 13, V Thorsson 14, A J Bass 15
Affiliations
- PMID: 32387455
- PMCID: PMC7690253
- DOI: 10.1016/j.annonc.2020.04.011
Characterizing diversity in the tumor-immune microenvironment of distinct subclasses of gastroesophageal adenocarcinomas
S Derks et al. Ann Oncol. 2020 Aug.
Abstract
Background: Gastroesophageal adenocarcinomas (GEAs) are heterogeneous cancers where immune checkpoint inhibitors have robust efficacy in heavily inflamed microsatellite instability (MSI) or Epstein-Barr virus (EBV)-positive subtypes. Immune checkpoint inhibitor responses are markedly lower in diffuse/genome stable (GS) and chromosomal instable (CIN) GEAs. In contrast to EBV and MSI subtypes, the tumor microenvironment of CIN and GS GEAs have not been fully characterized to date, which limits our ability to improve immunotherapeutic strategies.
Patients and methods: Here we aimed to identify tumor-immune cell association across GEA subclasses using data from The Cancer Genome Atlas (N = 453 GEAs) and archival GEA resection specimen (N = 71). The Cancer Genome Atlas RNAseq data were used for computational inferences of immune cell subsets, which were correlated to tumor characteristics within and between subtypes. Archival tissues were used for more spatial immune characterization spanning immunohistochemistry and mRNA expression analyses.
Results: Our results confirmed substantial heterogeneity in the tumor microenvironment between distinct subtypes. While MSI-high and EBV+ GEAs harbored most intense T cell infiltrates, the GS group showed enrichment of CD4+ T cells, macrophages and B cells and, in ∼50% of cases, evidence for tertiary lymphoid structures. In contrast, CIN cancers possessed CD8+ T cells predominantly at the invasive margin while tumor-associated macrophages showed tumor infiltrating capacity. Relatively T cell-rich 'hot' CIN GEAs were often from Western patients, while immunological 'cold' CIN GEAs showed enrichment of MYC and cell cycle pathways, including amplification of CCNE1.
Conclusions: These results reveal the diversity of immune phenotypes of GEA. Half of GS gastric cancers have tertiary lymphoid structures and are therefore promising candidates for immunotherapy. The majority of CIN GEAs, however, exhibit T cell exclusion and infiltrating macrophages. Associations of immune-poor CIN GEAs with MYC activity and CCNE1 amplification may enable new studies to determine precise mechanisms of immune evasion, ultimately inspiring new therapeutic modalities.
Keywords: chromosomal instability; gastric cancer subtypes; immunotherapy; tertiary lymphoid structure (TLS).
Copyright © 2020 European Society for Medical Oncology. Published by Elsevier Ltd. All rights reserved.
Conflict of interest statement
Disclosure This study was supported by research funds from Merck to AJB. AJB has funding from Merck, Bayer and Novartis, is an advisor to Earli and Helix Nano and a co-founder of Signet Therapeutics. SJR receives research funding from Merck, Bristol-Myers-Squibb, Affimed and KITE/Gilead. All remaining authors have declared no conflict of interest.
Figures
Figure 1.
The tumor immune microenvironment (TME) of GEA subclasses. A) TIMER analysis of TCGA mRNA sequencing data shows heterogeneity of immune infiltration between GEA subtypes. B) IHC/digital image analysis using CD8 antibodies confirms lower CD8+ T cells in CIN and GS subclass. C) CD8 T cell have a tumor infiltrating character in EBV+ and MSI cancers (ratio cell density at TC tumor center compared to invasive margin (IM) >1) while the opposite is observed for CIN GEAs (ratio<1). D) Unsupervised hierarchical clustering of Nanostring cell type annotation data identfies 2 large clusters. CIN GEAs form cluster 1 together with half of diffuse/GS cases and half of MSI GEAs. EBV and the other 50% of MSI group in cluster 2, together with diffuse/GS cases of which the majority has tertiary lymphoid structures (TLS). E) Stromal PD-L1 expression (IHC), but not tumoral PD-L1 expression, was found to be associated with intratumoral CD8+ T cells. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001
Figure 2.. A)
A) unsupervised hierarchical clustering of cell type annotation data from Nanistring mRNA analysis of the CIN archival FFPE cohort reveals a ‘CIN-hot’ (cluster 1) and ‘CIN-cold’ (cluster 2) cluster. B) The ‘CIN-hot’ (left) and ‘CIN-cold’ (right) clusters were further analyzed by IHC and a genomic GEA targeted sequencing panel, presented in order of decreasing CD45 count (left to right). (C) Examples of IHC and the digital analysis process are shown. Green: tumor area, blue, invasive margin D) Cell density ratio’s (cell density at tumor center: cell density at invasive margin) show accumulation of most immune cell subsets at the invasive margin, except CD68+ macrophages and FOXP3+ regulatory T cells which are distributed more evenly over tumor margin and center E) MYC amplification (amp) is associated with a lower number of CD8+ T cells, CD4+ T cells, and T-bet+ (Th1 proinflammatory) T cells compared to MYC wildtype (WT) tumors. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001
Figure 3.
Differential analyses of ”CIN-hot” and “CIN-cold” phenotypes using TCGA data. A CIN samples were divided into ‘CIN-hot’ and ‘CIN-cold’ groups based on CD4 and CD8 enrichment scores (B), which was not significantly associated with overall survival Immunological cold tumors occurred more often patient from Asian, Western-European and Brazilian descent and were more often in the distal stomach. C) Gene set enrichment analysis of TCGA mRNA expression data comparing CIN-hot GEAs and CIN-cold GEAs. Red: pathways enriched in the CIN-hot. Blue: pathways enriched in the CIN-cold GEAs. D) Cyclin E1 (CCNE1) was one of the most differentially amplified genes between hot (red) and cold (blue) CIN tumors, and associated with a lower CD8 T cells abundance score. * p ≤ 0.05, ** p ≤ 0.01
Figure 4. A)
A) 50% diffuse type/GS GEA lymphoid aggregated which we further characterized as tertiary lymphoid structures (TLS) by additional IHC (A; PNAd, peripheral node addressin, marker for high endothelial venules; DC-LAMP, dendritic cell lysosomal associated membrane glycoprotein, marker for mature DCs). B) Presence of TLSs in diffuse type/GS GEA was associated with a higher T cell and B cell function score. C) Vulcano plot showing the most differentially expressed genes in diffuse type/GS GEA with and without TLS. D) The 12-chemokine gene expression signature is enriched in some but not all diffuse/GS GEA with TLS formation.
References
- Cunningham D, Starling N, Rao S et al. Capecitabine and oxaliplatin for advanced esophagogastric cancer. N. Engl. J. Med. 2008; 3581:36–46. -PubMed
- Akyala AI, Verhaar AP, Peppelenbosch MP. Immune checkpoint inhibition in gastric cancer: A systematic review. J. Cell. Immunother. 2018. doi: 10.1016/j.jocit.2018.05.001. -DOI
Publication types
MeSH terms
LinkOut - more resources
Full Text Sources
Medical
Research Materials