Observational Study of PD-L1, TGF-β, and Immune Cell Infiltrates in Hepatocellular Carcinoma - PubMed (original) (raw)
Observational Study of PD-L1, TGF-β, and Immune Cell Infiltrates in Hepatocellular Carcinoma
Christian Ihling et al. Front Med (Lausanne). 2019.
Abstract
Introduction: Hepatocellular carcinoma (HCC) typically develops in cirrhotic livers, with increased programed death ligand 1 (PD-L1) and transforming growth factor beta (TGF-β) activity implicated in immunosuppression. Methods: In an observational study of HCC liver samples, we determined the incidence of PD-L1 and immune cell (IC) infiltrates, and signs of TGF-β activity. HCCs were characterized by the incidence and distribution of PD-L1+ cells, and CD8+, CD68+, and FoxP3+ infiltrating ICs in HCC and surrounding liver. Gene expression signatures (GESs) associated with TGF-β activity and ICs were evaluated by RNAseq. Results: In non-neoplastic cirrhotic and non-cirrhotic liver, PD-L1 occurred on sinusoidal lining cells (mostly Kupffer cells), endothelial cells and ICs. In HCC, PD-L1+ tumor cells were rare. Most PD-L1+ cells were identified as ICs. CD8+, CD68+, and FoxP3+ ICs were associated with HCC, particularly in the invasive margin. CD8+ cell incidence correlated with PD-L1+ cells, consistent with PD-L1 being upregulated in response to pre-existing cytotoxic T-lymphocyte activity. TGFB1 mRNA levels and TGF-β activation GES correlated with the strength of the tumor-associated macrophage GES. Conclusion: Inhibition of PD-L1+ ICs and TGF-β activity and their respective immunomodulatory pathways may contribute to antitumor effects in HCC.
Keywords: CD8; HCC; PD-L1; checkpoint inhibitor; hepatocellular carcinoma; immune cell infiltrates.
Figures
Figure 1
Qualitative assessment of PD-L1 expression by IHC in tumor-free liver adjacent to HCC with reactive changes. (A) non-cirrhotic liver, PD-L1 (brown); (B) non-cirrhotic liver, PD-L1 (brown) and CD68 (blue); (C) cirrhotic liver, PD-L1 (brown). (A) PD-L1 staining was seen in sinusoidal lining cells, as well as in ECs of a CV. The extent of PD-L1 staining varied considerably regionally as well as from case to case. (B) Double staining of PD-L1 (brown) and CD68 (blue) revealed that some, but not all, sinusoidal macrophages (Kupffer cells) in the TFL showed PD-L1 IR; however, the extent of PD-L1 staining in CD68+ cells varied considerably regionally as well as from case to case (C) in the cirrhotic liver tissue, membrane and/or cytoplasmic PD-L1 staining was present in the sinusoidal lining cells, ECs of microvessels (short arrows) and ECs of CVs, and in ICs. Hepatocytes were PD-1 negative. In contrast, the epithelium of bile ductules showed cytoplasmic and membranous PD-L1 staining (arrows). cv, central vein.
Figure 2
Cellular composition of a typical HCC in serial sections related to PD-L1 expression in this area (A) CD68; (B) CD8; (C) FoxP3; (D) PD-L1. Long arrows label a blood vessel that is present on every slide indicating that consecutive sections were stained. Of note, there are only a few scattered FoxP3+ cells, which were greatly outnumbered by CD8+ and CD68+ cells.
Figure 3
(A) PD-L1/cytokeratin double labeling (PD-L1 = brown; cytokeratin [clone AE1/AE3] = blue) confirmed that only a few tumor cells showed PD-L1 staining (inset A) Most PD L1+ cells in HCC were cytokeratin-negative, and were therefore ICs. (B) Semiquantitative analysis of PD-L1 expression in HCC revealed that only 6 of 68 HCC specimens (8.8%) were positive for PD-L1 staining in tumor cells with regard to the ≥1% cut-off.
Figure 4
CD8 protein expression (IHC) and associated mRNA expression in HCC. (A) Correlation of CD8a mRNA expression with the number of CD8+ cells in IHC (P = 2.7 × 10−16); (B) correlation of CD8b mRNA expression with the number of CD8+ cells in IHC (P = 7.3 × 10−11); (C) association of CD8 T-cell GES with the number of CD8-positive cells in IHC; (D) association of T-effector-IFN-γ-associated GES with the number of CD8+ cells in IHC; (E) association of perforin mRNA with the number of CD8+ cells in IHC; (F) association of granzyme A mRNA with the number of CD8+ cells in IHC; (G) association of granzyme H mRNA with the number of CD8+ cells in IHC. Asterisk indicates P < 0.05. TPM, transcripts per million; PRF1, perforin 1; GZMA, granzyme A; GZMH, granzyme H.
Figure 5
TGF-β expression and associated GES activity in HCC. (A) Expression of TGFB mRNA isoforms determined by RNAseq; (B) association between Hoshida molecular HCC subtype and TGF-β response-associated GES; (C) correlation between TGFB1 mRNA expression and the TGF-β1 response-associated GES (P = 7.9 × 10−14); (D) association between Hoshida molecular HCC subtype and EMT-associated GES; (E) correlation between TGF-β response-associated GES and EMT-associated GES (P = 7.9 × 10−14). An asterisk indicates P < 0.05; †fold-difference 5.5, P = 6.4 × 10−16; ‡fold-difference 2.3, P = 0.008. EMT, epithelial/mesenchymal transition.
Figure 6
Relationships between PD-L1/PDL1/, CD8/CD8a expression, and TGF-β TGFB1/TGFb1 response/immune cell (IC) infiltrate-associated GES and TGF-β1 and tumor-associated ICs in HCC. (A) Association between PDL1 mRNA levels and the incidence of CD8+ cells. High and Low incidence was defined as above and below the median, respectively; (B) association between PD-L1+ and the incidence of PD-L1+ tumor cells and the incidence (Low or High) of CD8+ cells; (C) association of TGF-β response-associated GES between PD-L1+ ICs and the incidence (Low or High) of CD8+ cells; (D) correlation between TGFB1 mRNA levels and tumor associated macrophage (TAM) -associated GES; (E) correlation for the strength of GESs associated with the TGF-β1 activity response and TAMs (P = 7.0 × 10−14); (F) correlation between a CD8-associated GES and a TGF-β1-activation response-associated GES (P = 5.9 × 10−1). CD8 IHC High and Low defined as above and below the median, respectively. An asterisk indicates P < 0.05. TAM, tumor-associated macrophages.
References
- Chiang DY, Huang A. Co-expression networks identify distinct immune infiltrates in hepatocellular carcinoma. J Immunother Cancer (2015) 3(Suppl. 2):P213 10.1186/2051-1426-3-S2-P213 - DOI
LinkOut - more resources
Full Text Sources
Research Materials
Miscellaneous