NKG2A is a NK cell exhaustion checkpoint for HCV persistence

Abstract

Exhaustion of cytotoxic effector natural killer (NK) and CD8⁺ T cells have important functions in the establishment of persistent viral infections, but how exhaustion is induced during chronic hepatitis C virus (HCV) infection remains poorly defined. Here we show, using the humanized C/O^Tg mice permissive for persistent HCV infection, that NK and CD8⁺ T cells become sequentially exhausted shortly after their transient hepatic infiltration and activation in acute HCV infection. HCV infection upregulates Qa-1 expression in hepatocytes, which ligates NKG2A to induce NK cell exhaustion. Antibodies targeting NKG2A or Qa-1 prevents NK exhaustion and promotes NK-dependent HCV clearance. Moreover, reactivated NK cells provide sufficient IFN-γ that helps rejuvenate polyclonal HCV CD8⁺ T cell response and clearance of HCV. Our data thus show that NKG2A serves as a critical checkpoint for HCV-induced NK exhaustion, and that NKG2A blockade sequentially boosts interdependent NK and CD8⁺ T cell functions to prevent persistent HCV infection.

Fig. 1

CD8⁺ T cell exhaustion and PD-1 blockade. C/O^Tg (n = 55, n ≥ 6 for each time point) or wt mice (n = 32, n = 4 for each time point) were i.v. infused with 1 mL HCV J399EM (TCID₅₀ = 2 × 10⁷/mL). Mice were killed and blood, splenocytes, hepatocytes, and NPCs were collected at indicated time. a HCV RNA copies in hepatocytes were measured by qPCR. b The number of IFN-γ spot-forming cells (SFCs) by ELISpot assay determined 2.5 d after in vitro stimulation of splenocytes with respective epitope peptides, subtracted the SFC of OVA stimulation as the background. c FACS measurement of PD-1 expression in circulating or hepatic CD8⁺ T cells at indicated time post HCV infection. d FACS analysis of PD-L1 expression on hepatocytes isolated from naive (n = 4) or HCV-infected C/O^Tg mice (n = 6) 2 wpi. e C/O^Tg mice (n = 9 for each group) were i.p. injected with PD-1 blocking antibody or isotype Ig (200 μg/3 days) 1 day before HCV inoculation. f HCV RNA copies in serum and livers, and g HCV-specific T cells response to the indicated epitopes. h FACS analysis of Tim-3 expression on peripheral and liver CD8⁺ T cells. i qPCR measurement of HCV RNA copies in the liver or blood after C/O^Tg mice (n = 3 for each group) were i.p. injected with PD-1 (200 μg/3 days) plus Tim-3 (100 μg/2 days) blocking antibody or isotype IgG 1 day before HCV inoculation. Dash lines indicated limits of detection of related assays (qPCR: 100 copies/mg liver, 500 copies/mL serum; ELISpot: 50 spots/4 × 10⁵ splenocytes). #, below detection limit. Data were mean ± SD, Student t-test in b, d, f, h, and i. *P < 0.05; **P < 0.01; ***P < 0.001. ns Not significant. Source data are provided as a Source Data file

Fig. 2

Impaired hepatic NK cell response during HCV persistent infection in C/O^Tg mice. Mice were treated as in Fig. 1a. FACS analyses of intrahepatic NK cells for expression of a DX5, b CD11b, and CD27, c activation receptors, including Ly49D, Ly49H, and NKG2D, and d inhibitory receptors, including NKG2A, KLRG1, and TIGIT. e FACS analyses of peripheral NK cells for expression of NKG2A and KLRG1. f NK function assays were carried out at indicated time of HCV infection by Yac-1 cell stimulation. CD107a and IFN-γ expression were FACS analyzed. Data were mean ± SD, ANOVA test. *P < 0.05; **P < 0.01. Source data are provided as a Source Data file

Fig. 3

NKG2A blocking promoted HCV clearance. a C/O^Tg mice (n = 13) were infected with 1 mL HCV J399EM (TCID₅₀ = 2 × 10⁷/mL). b The self-limited infection in C/O^Tg mice exhibited an initial high peripheral viral load (<1 mpi) followed with RNA copies below the detection limit (≥1 mpi), whereas HCV persistent infection was characterized by sustained HCV viral loads in the serum beyond 1 mpi. c FACS analysis of NKG2A in peripheral NK cells and qPCR measurement of HCV RNA copies in the serum 1 m after C/O^Tg mice (n = 13) were infected. d C/O^Tg mice (n = 9 for each time point) were treated with anti-NKG2A or isotype Ig (100 μg/3 days, i.p.) 1 day before HCV infection. e qPCR measurement of HCV RNA copies in livers and sera; f FACS analysis of CD107a and granzyme B expression in NK cells; g FACS analysis of intracellular IFN-γ in isolated NK cells after co-cultured with Yac-1. h C/O^Tg mice (n = 6 for each time point) were treated with anti-NKG2A or isotype Ig (100 μg/3 days, i.p.) beginning at 2 wpi, i HCV RNA copies in livers and sera; j CD107a and granzyme B were measured in NK cells; and k HCV-specific CD8⁺ T cell response by IFN-γ ELISpot assays using splenocytes (4 × 10⁵ for each well) isolated from mice 1 mpi. Epitope peptides (final concentration 4 μg/mL) were indicated. Dash lines indicated limits of detection (qPCR, 500 copies/mL serum or 100 copies/mg liver; ELISpot, 50 spots/4 × 10⁵ splenocytes). Data were mean ± SD, Student t-test. *P < 0.05; **P < 0.01; ***P < 0.001. Source data are provided as a Source Data file

Fig. 4

NK cells instructed T cells in HCV persistence. a Mice were treated as in Fig. 3d. ELISpot assays using splenocytes and the indicated epitope peptides. C/O^Tg mice were i.p. pre-treated with mock IgG (n = 9) or b anti-AGM1 (20 μg/3 days, n = 9), c anti-CD8 (100 μg/3 days, n = 6), together with anti-NKG2A antibody 1 d before HCV inoculation. b HCV-specific CD8⁺ T cell response by IFN-γ ELISpot and c hepatic and serum HCV RNA copies were measured 2 wpi. Dash lines indicated limits of detection of related assays (qPCR, 500 copies/mL serum or 100 copies/mg liver; ELISpot, 50 spots/4 × 10⁵ splenocytes). Data were mean ± SD, Student t-test. *P < 0.05; **P < 0.01; ***P < 0.001. ns Not significant. Source data are provided as a Source Data file

Fig. 5

HCV infection of hepatocytes impaired hepatic NK cells function. a Flowchart of NK functional inhibition assays. PHTs (2 × 10⁵/mL) isolated from naive C/O^Tg mice were infected with HCVcc (MOI = 1) for 3 days before NPCs (1 × 10⁶) were added. Yac-1 cells (1 × 10⁵/well) or PMA (50 ng/mL) plus ionomycin (1 μM) were used to stimulate NK cells. b FACS analysis of intracellular IFN-γ and CD107a of NK cells 6 h after stimulation. c Indicated blocking antibodies and isotype IgG were added to the PHT/NK co-culture assay. IFN-γ expression of NK cells was detected 6 h after Yac-1 cell stimulation. d Measurement of intracellular IFN-γ in NK cells after supernatants of PHTs after HCV infection (HCV-CM) were added to NPCs. Yac-1 cells (1 × 10⁵/well) or PMA (50 ng/mL) plus ionomycin (1 μM) were used to stimulate NK cells. e NPCs or MACS purified intrahepatic NK cells were co-cultured with or transwell separated from HCV-infected PHTs. IFN-γ expression of NK cells was detected 6 h after Yac-1 cell stimulation. Data were mean ± SD, Student t-test. *P < 0.05; **P < 0.01. ns Not significant. Source data are provided as a Source Data file

Fig. 6

Upregulation of Qa-1 on hepatocytes accounted for NK cell dysfunction upon HCV infection. a Measurement of Qa-1 mRNA in liver isolated from C/O^Tg mice (n = 3) at the indicated time post HCV infection. b FACS analysis of Qa-1 expression on hepatocytes or NPC subsets as indicated before or 2 weeks post HCV infection. c C/O^Tg mice (n = 6 each group) were i.p with Qa-1 blocking antibody or isotype IgG (100 μg/3 days). d HCV RNA copies in serum and liver, e CD107a and IFN-γ expression in hepatic NK cells upon Yac-1 cell stimulation, f IFN-γ ELISpot assays of T cells stimulated by the indicated epitopes, were assayed 2 wpi. g C/O^Tg mice (n = 4 each group) were i.v. injected with liposomes/siRNA complex at a dose of 2.5 mg/kg every 3 days since day −1. Mice were killed at 2 wpi. h HCV RNA copies in serum and liver, i CD107a and IFN-γ expression in hepatic NK cells upon Yac-1 cell or P + I stimulation as indicated. Dash lines indicated limits of detection of related assays. Data were mean ± SD, Student t-test. *P < 0.05; **P < 0.01; ***P < 0.001. ns Not significant. Source data are provided as a Source Data file

Fig. 7

Anti-NKG2A restored T cell response through enhanced IFN-γ production. a FACS analysis of hepatic IFN-γ positive cells with or without NKG2A blockade 2 wpi. b, c C/O^Tg mice (n = 6 each group) were i.p. pre-treated with IFN-γ neutralization antibody or isotype IgG (100 μg/3 days) together with anti-NKG2A antibody 1 d before HCV inoculation. b HCV-specific CD8⁺ T cell response by IFN-γ ELISpot and c serum and hepatic HCV RNA copies were measured 2 wpi. d Diagram of Qa-1/NKG2A immune checkpoint during HCV infection. Dash lines indicated limits of detection of related assays. Data were mean ± SD, Student t-test. *P < 0.05; **P < 0.01. Source data are provided as a Source Data file