1645P - Analysis of programmed death-ligand 1 (PD-L1) expression, transforming growth factor (TGF)-β gene expression signatures (GES) and tumor-infiltratin...

Date 11 September 2017
Event ESMO 2017 Congress
Session Poster display session
Topics Hepatobiliary Cancers
Translational Research
Basic Principles in the Management and Treatment (of cancer)
Presenter Yue Zhang
Citation Annals of Oncology (2017) 28 (suppl_5): v573-v594. 10.1093/annonc/mdx390
Authors Y. Zhang1, B.J. Naughton2, P..A. Rolfe3, E. Frick-Krieger4, I. Dussault5, L. Terracciano6, C. Ihling7
  • 1R&d Global Early Development, EMD Serono Research & Development Institute, 01821 - Billerica/US
  • 2Bioinformatics, EMD Serono Research & Development Institute, Billerica/US
  • 3R&d Global Early Development, EMD Serono Research & Development Institute, Billerica/US
  • 4Rd-ebp-h, Merck KGaA, Darmstadt/DE
  • 5Clinical Biomarkers And Companion Diagnostics, EMD Serono Research & Development Institute, Billerica/US
  • 6Pathology, Institute of Pathology, University Hospital Basel, Basel/CH
  • 7Cbd, Merck KGaA, Darmstadt/DE

Abstract

Background

HCC evades antitumor immune responses via multiple mechanisms, including the PD-L1 and TGF-β pathways. PD-L1 expression correlates with tumor aggressiveness and recurrence. Increased TGF-β activity corresponds with poor clinical outcomes. Using immunohistochemistry (IHC), we previously showed that PD-L1 expression in HCC stems primarily from IC. To further assess the HCC immune milieu, we measured IC, TGF-β-associated GES, and PD-L1 expression using IHC/RNAseq.

Methods

We assessed protein expression in 50 resected HCC specimens by quantitative (Q) IHC (primary antibodies: PD-L1, CD8, CD68) using standard techniques and automated software. For RNAseq, we prepared strand-specific libraries from extracted RNA, which were sequenced and compared to GES from published papers, CIBERSORT and Ingenuity Pathway Analysis.

Results

All cases had typical morphology (low- to high-grade trabecular, pseudoglandular, or solid with common cytoplasmic features). Q CD8 IHC significantly correlated with CD8 mRNA expression and CD8 T cell GES, supporting the utility of RNAseq to evaluate the role of CD8+ T cells in HCC. RNAseq identified TGF-β1 as the main TGF-β isoform in HCC. Predefined TGF-β GES correlated strongly with EMT GES. There was a trend toward increased TGF-β1 activity and EMT marker expression in the S1 molecular subtype, which has previously been associated with TGF-β-driven aberrant Wnt signaling. Q CD8 IHC correlated with PD-L1 mRNA and protein levels in IC. In samples with high CD8, there was a trend of increased tumor-associated macrophages (TAMs); the presence of TAMs strongly correlated with TGF-β GES. Interestingly, few tumor cells displayed membranous PD-L1 staining as confirmed by PD-L1/pan-cytokeratin double labeling.

Conclusions

We used RNAseq and IHC to better understand the immunosuppressive environment in HCC driven by TGF-β and PD-L1, which may mediate different mechanisms to inhibit preexisting CD8+ IC.

Clinical trial identification

N/A

Legal entity responsible for the study

Funding was provided by Merck KGaA, Darmstadt, Germany.

Funding

Funding was provided by Merck KGaA, Darmstadt, Germany. Disclosure: Y. Zhang, B.J. Naughton, P.A. Rolfe, I. Dussault: Employee of EMD Serono Research & Development Institute, Billerica, MA, USA. E. Frick-Krieger, C. Ihling: Employee of Merck KGaA, Darmstadt, Germany. L. Terracciano: Consulting/advisory role to Merck AG