The co-inhibitory receptor TIGIT regulates NK cell function and is upregulated in human intrahepatic CD56bright NK cells

Abstract

The crosstalk between NK cells and their surrounding environment is enabled through activating and inhibitory receptors, which tightly control NK cell activity. The co-inhibitory receptor TIGIT decreases NK cell cytotoxicity and is involved in NK cell exhaustion, but has also been associated with liver regeneration, highlighting that the contribution of human intrahepatic CD56^bright NK cells in regulating tissue homeostasis remains incompletely understood. A targeted single-cell mRNA analysis revealed distinct transcriptional differences between matched human peripheral blood and intrahepatic CD56^bright NK cells. Multiparameter flow cytometry identified a cluster of intrahepatic NK cells with overlapping high expression of CD56, CD69, CXCR6, TIGIT and CD96. Intrahepatic CD56^bright NK cells also expressed significantly higher protein surface levels of TIGIT, and significantly lower levels of DNAM-1 compared to matched peripheral blood CD56^bright NK cells. TIGIT⁺ CD56^bright NK cells showed diminished degranulation and TNF-α production following stimulation. Co-incubation of peripheral blood CD56^bright NK cells with human hepatoma cells or primary human hepatocyte organoids resulted in migration of NK cells into hepatocyte organoids and upregulation of TIGIT and downregulation of DNAM-1 expression, in line with the phenotype of intrahepatic CD56^bright NK cells. Intrahepatic CD56^bright NK cells represent a transcriptionally, phenotypically, and functionally distinct population of NK cells that expresses higher levels of TIGIT and lower levels of DNAM-1 than matched peripheral blood CD56^bright NK cells. Increased expression of inhibitory receptors by NK cells within the liver environment can contribute to tissue homeostasis and reduction of liver inflammation.

Keywords: DNAM-1; PVR/CD155; TIGIT; immune tolerance; intrahepatic NK cells; liver organoids; single-cell mRNA analysis; tissue homeostasis.

Figure 1

Single-cell mRNA expression. (A) Principal component analysis showing clustering of CD56^bright peripheral blood (pb)NK cells from healthy control individuals (blue triangles), CD56^bright pbNK cells from liver resection patients (orange squares) and CD56^bright intrahepatic (ih)NK cells from liver resection patients (green dots). (B) Quantitative analysis of CD56^bright ihNK cells and matched CD56^bright pbNK cells from liver resection patients (n=4), and CD56^bright pbNK cells from healthy control individuals (n=2) showing single-cell mRNA expression of liver-residency markers CD69, CXCR6, CCR5, Eomes and T-bet (TBX21). (C) Quantitative analysis of CD56^bright ihNK cells and matched CD56^bright pbNK cells from liver resection patients (n=4), and CD56^bright pbNK cells from healthy control individuals (n=2) showing single-cell mRNA expression of selected genes with highly significant differential gene expression comparing ih and pbNK cells. Green dots representing 178 ihNK cells, orange squares representing 173 pbNK cells and blue triangles representing 89 healthy control pbNK cells. A mixed-effects model with random intercept, considering intra-sample correlations, was used to compare the single-cell mRNA expression data from different donors (or groups). Black line indicates median.

Figure 2

Immunophenotyping of NK cells. (A) Representative plots of flow cytometry data of one donor showing TIGIT, DNAM-1 and CD96 expression on peripheral blood (pb) and intrahepatic (ih) bulk NK cells. Histogram showing TIGIT (left), DNAM-1 (middle) and CD96 (right) expression on bulk NK cells in the liver (green), the blood (orange) and a FMO control (grey). (B) Representative viSNE plots showing density and the expression of CD56, CXCR6, CD69, TIGIT, DNAM-1 and CD96 on ih and pbNK cells from one patient undergoing liver transplantation. Color coding indicates the density and marker expression. viSNE analysis was performed using Cytobank. (C) Quantitative analysis of flow cytometry data of matched ih and pbNK cells from liver resection patients (n=15 for DNAM1- and CD96, n=19 for TIGIT), and control pbNK cells from healthy individuals (n=5) showing TIGIT, DNAM-1 and CD96 expression on bulk, CD56^dim and CD56^bright NK cells. Green dots representing ihNK cells, orange squares representing pbNK cells and blue triangles representing healthy control pbNK cells. Wilcoxon matched-pairs sign rank test was used to determine statistical differences between ih and pbNKcells in all scatter plots, Mann-Whitney test was used to determine statistical differences between pb and pb control NK cells in all scatter plots. Black line indicates median.

Figure 3

Functional activity of TIGIT^+/– and DNAM-1^+/– cells. Functional responses of TIGIT^+/– and DNAM-1^+/– CD56^bright ihNK cells, following co-incubation of liver-derived cells with K562 (E:T ratio of 10:1) for 5h at 37°C. Representative flow plots showing distribution of CD107a (top) and TNF-α (bottom) expression on CD56^bright ihNK cells depending on their TIGIT or DNAM-1 expression. Graphs on top showing percentage of CD107a⁺ TIGIT^+/– or DNAM-1^+/–CD56^bright ihNK cells and graphs below showing percentage of TNF-α⁺ TIGIT^+/– or DNAM-1^+/–CD56^bright ihNK cells. Wilcoxon matched-pairs sign rank test was used to determine statistical differences. Each dot represents one donor (n=8).

Figure 4

Co-culture of NK cells and Huh7 cells. (A) Co-culture experiments using isolated NK cells from peripheral blood from healthy control individuals and Huh7 cells. NK cells were either cultured in direct contact with Huh7 cells, in a transwell insert with the size of 1 µm to ensure indirect contact with Huh7 cells or with medium only. (B) Graphs are displaying quantitative analysis of CD96, DNAM-1 and TIGIT expression on CD56^bright NK cells after 6, 12, 24 and 48 hours. At the timepoints 6, 12, and 24 hours a total of 5 healthy control individuals was used for comparison. At 48 hours, a total of 4 out of the 5 mentioned healthy control individuals were assessed for CD96, DNAM-1 and TIGIT. Black dots representing direct contact, pink squares the transwell insert and blue triangles medium only. Paired t-test was used to determine the displayed statistical differences between direct contact of NK cells and medium control. Each dot represents the median of the individuals; error bars represent the data range. (C) Representative histograms comparing PVR staining (light grey) and unstained control (dark grey) demonstrating PVR expression on Huh7 cells. (D) Graph displays the percentage of TIGIT⁺ NK cells in different conditions: NK cells only, NK cells in direct contact with Huh7 cells, isotype control and NK cells in direct contact with Huh7 cells including blocking of PVR with an anti-PVR-antibody. Data are shown for 5 paired experiments using NK cells derived from 3 different donors. Statistical analysis for significance was performed using a two-tailed paired T test.

Figure 5

Co-culture of NK cells and hepatocyte organoids. (A) Representative histograms comparing PVR staining (light grey) and unstained control (dark grey) of hepatocyte organoids demonstrating PVR expression on hepatocytes. (B) NK cells isolated from buffy coats were co-cultured with empty BME2 droplets or hepatocyte organoids in BME2 droplets for up to 48h. Graph displaying total cell count of CD56^bright NK cells that migrated into BME2 droplets at 24 hours of co-culture with empty droplets (blue triangles) or hepatocyte organoids (black dots). Wilcoxon matched-pairs sign rank test was used for statistical analysis. Black line indicates median. (n=3, triplicates) (C) Chemokine concentrations in hepatocyte organoid culture supernatants are shown in ascending order on a logarithmic scale. Bars represent mean values of three experiments in triplicates for each chemokine and error bars show standard deviations (n=3, triplicates). (D) Migrated CD56^bright NK cells (black dots) were phenotypically compared to non-migrated cells (pink squares) and to separately cultured NK cells that did not have any contact to hepatocyte organoids (blue rectangles). Percentage surface expression of TIGIT and DNAM-1 is shown in CD56^bright NK cells after 24 hours and 48 hours. For TIGIT, statistical differences represent comparison of migrated NK cells as well as non-migrated NK cells vs. NK cells in medium and for DNAM-1 statistical differences represent migrated NK cells vs. NK cells in medium (top value) or non-migrated cells (lower value). Wilcoxon matched-pairs sign rank test was used for statistical analysis. Each dot represents the median of nine experiments; error bars represent the data range.