Key features and homing properties of NK cells in the liver are shaped by activated iNKT cells

The contribution of natural killer (NK) cells to the clearance of hepatic viral infections is well recognized. The recently discovered heterogeneity of NK cell populations renders them interesting targets for immune interventions. Invariant natural killer T (iNKT) cells represent a key interaction partner for hepatic NK cells. The present study addressed whether characteristics of NK cells in the liver can be shaped by targeting iNKT cells. For this, the CD1d-binding pegylated glycolipid αGalCerMPEG was assessed for its ability to modulate the features of NK cells permanently or transiently residing in the liver. In vivo administration resulted in enhanced functionality of educated and highly differentiated CD27+ Mac-1+ NK cells accompanied by an increased proliferation. Improved liver homing was supported by serum-derived and cellular factors. Reduced viral loads in a mCMV infection model confirmed the beneficial effect of NK cells located in the liver upon stimulation with αGalCerMPEG. Thus, targeting iNKT cell-mediated NK cell activation in the liver represents a promising approach for the establishment of liver-directed immune interventions.


Figure S1
Figure S1. Representative flow cytometry plots of αGalCerMPEG-activated NK cells. Hepatic lymphocytes were isolated from wt mice 72 h after administration of a single dose of αGalCerMPEG (10 µg) and the representative dot plots show the flow cytometry gating for the expression of IFNγ and CD107a detected in cells derived from wt mice assessed after 6 h co-incubation with YAC-1 target cells.

Figure S2
Figure S2. Functionality of NK cells derived from blood, lymph nodes, lung and adipose tissue. Lymphocytes were isolated from the blood, lymph nodes (LN), lung and adipose tissue (AT) of wt mice 72 h after administration of a single dose of αGalCerMPEG (10 µg). The expression of IFNγ and CD107a was assessed after 6 h co-incubation with YAC-1 target cells. (A) Absolute cell number of NK cells (out of 1x10 6 total cells). MFI and frequencies of (B) IFN secreting and (C) CD107a expressing blood-, LN-, lung-and AT-derived NK cells. The data are derived from one experiment with n = 6 mice per group. Columns represent the mean ± SEM. Asterisks denote significant values as calculated by unpaired, two-tailed Student's t-test. **** p ≤ 0.0001;*** p ≤ 0.001; ** p ≤ 0. 01; * p ≤ 0; 05.

Figure S3
Figure S3. αGalCerMPEG-mediated NK cell activation is dependent on NKT cells. Hepatic lymphocytes were isolated from wt and Jα281 -/mice 72 h after administration of a single dose of αGalCerMPEG (10 µg) and the expression of IFNγ and CD107a was assessed after 6 h co-incubation with YAC-1 target cells. (A) Frequencies of NK cells (NKp46 + CD3 -) detected in hepatic lymphocytes derived from wt and Jα281 -/mice (n=10-15). (B) MFI and frequencies of NK cells isolated from NKT cell-deficient Jα281 -/mice expressing IFN or CD107a. (C) MFI and frequencies of NK cells isolated from wt or NKT cell-deficient Jα281 -/mice expressing CD69. MFI: n=4 mice, one out of at least three independent experiments. Frequencies: n=7 mice; columns represent the mean ± SEM of data pooled from two independent experiments. Asterisks denote significant values as calculated by One-way ANOVA. **** p ≤ 0.0001; *** p ≤ 0.001; n.s. = not significant.

Figure S4
Figure S4. Phenotype and functionality of hepatic NKT cell populations. Lymphocytes were isolated from livers of mice treated with a single dose of GalCerMPEG (10 µg) at the indicated time points. (A) Absolute cell numbers of NKT cell subsets (out of 1x10 6 total cells). MFI of IFN-, IL-4-and IL-17-expressing NKT cell subsets detected as (B) CD3 + TCR + PLZF + , (C) CD3 + TCR + PLZF + IL-4 + T-bet + and (D) CD3 + TCR + PLZF + IL-17RB -. Violin plots represent the interquartile range, horizontal lines show the mean value and the width displays the distribution of data points. n=6 mice, data derived from one experiment. Asterisks denote significant values as calculated by unpaired, two-tailed Student's t-test. **** p ≤ 0.0001;*** p ≤ 0.001; ** p ≤ 0. 01; * p ≤ 0; 05.   Wild type mice were injected by s.c. route with a single dose of αGalCerMPEG (10 µg) and hepatic lymphocytes were collected 72 h later. NK cells (NKp46 + CD3 -) were stained for CD27, Mac-1 uneducated (Ly49C/Iand NKG2A -) and educated (Ly49C/I + or NKG2A + ) subsets and the expression of IFNγ and CD107a after 6 h co-incubation with YAC-1 target cells. (A) Frequency and MFI of IFNγ-and CD107a-expressing NK cells (MFI: n=4 mice, shown is one out of three independent experiments; Frequencies: n=11 mice). (B) Relative distribution of CD27 and Mac-1 expression within the subset of uneducated and educated NK cells (n=12 mice, columns represent the mean ± SEM of data pooled from three independent experiments). Asterisks denote significant values as calculated by Two-way ANOVA as compared to untreated controls. **** p ≤ 0.0001; *** p ≤ 0.001; ** p ≤ 0.01; * p ≤ 0.05; n.s. = not significant.  Wild type mice were injected by s.c. route with a single dose of αGalCerMPEG (10 µg) and hepatocytes were isolated at the indicated time points for the assessment of IP-10 expression by western blot analysis. (A) Representative western blot membranes depict the bands detected for IP-10 and α-tubulin (loading control) in untreated and treated samples. (B) The shown membranes were cropped from different parts of two gels run simultaneously. The membranes were cut for the incubation with the antibodies for α-tubulin (upper part of the gel) and IP-10 (lower part of the gel). (C) Representative light microscope pictures showing the isolated hepatocytes derived from untreated and treated samples and lymphocytes as reference (40x magnification). (D) Representative histograms for the flow cytometry analysis of CD95 expression on purified hepatocytes comparing untreated and treated samples (left) as well as before and after the purification steps (right). Wild type mice were injected by s.c. route with a single dose of αGalCerMPEG (10 µg) and hepatic cells were collected 12 h and 24 h later. The depicted hepatic cell populations were identified by flow cytometric analysis using the following cell surface markers: cDC1: CD45 + CD11c + MHC cl. II + PDCA1 -CD103 + CD11b -CX3CR1 -F4/80 -, cDC2: CD45 + CD11c + MHC cl. II + PDCA1 -CD103 -CD11b + CX3CR1 + Langerin -, classical macrophages (MΦ): CD45 + F4/80 + CD11b high Ly6C high CX3CR1 + Ly6G + , Kupffer cells: CD45 + CD11c -F4/80+CD11b intermediate , pDCs: CD45 + CD11c + PDCA1 + , pre-DCs: CD45 + CD11c + MHC cl. II -. Violin plots represent the interquartile range, horizontal lines show the mean value and the width displays the distribution of data points. n=6 mice, data derived from one experiment. Asterisks denote significant values as calculated by unpaired, two-tailed Student's t-test as compared to untreated controls. **** p ≤ 0.0001;*** p ≤ 0.001; ** p ≤ 0. 01; * p ≤ 0; 05; n.s. = not significant.

Figure S10
Figure S10. tSNE analysis of CD45 + hepatic cell populations. Wild type mice were injected by s.c. route with a single dose of αGalCerMPEG (10 µg) and hepatic cells were collected 12 and 24 h later. Cells were stained for CD45 and the indicated surface expression markers for subsequent analysis by flow cytometry. Shown is the tSNE analysis of CD45 + cells for the indicated markers (blue=low expression, red=high expression).

Figure S11
Figure S11. Serum IL-10 level upon αGalCerMPEG administration. Serum samples were taken from wt mice at several time points after the administration of a single dose of αGalCerMPEG (10 µg) and IL-10 levels were assessed by cytometric bead array (MFI, n=5 mice).