Recognition of the nonclassical MHC class I molecule H2-M3 by the receptor Ly49A regulates the licensing and activation of NK cells

Abstract

The development and function of natural killer (NK) cells is regulated by the interaction of inhibitory receptors of the Ly49 family with distinct peptide-laden major histocompatibility complex (MHC) class I molecules, although whether the Ly49 family is able bind to other MHC class I-like molecules is unclear. Here we found that the prototypic inhibitory receptor Ly49A bound the highly conserved nonclassical MHC class I molecule H2-M3 with an affinity similar to its affinity for H-2D(d). The specific recognition of H2-M3 by Ly49A regulated the 'licensing' of NK cells and mediated 'missing-self' recognition of H2-M3-deficient bone marrow. Host peptide-H2-M3 was required for optimal NK cell activity against experimental metastases and carcinogenesis. Thus, nonclassical MHC class I molecules can act as cognate ligands for Ly49 molecules. Our results provide insight into the various mechanisms that lead to NK cell tolerance.

Figure 1. H2-M3 is recognised by C57BL/6 Ly49A

(a) Sequence alignment of H2-M3, H-2-D^d and H-2D^b demonstrates homology between the two ligands at the protein level. (b) Tetramer staining of cells expressing Ly49A, Ly49D and Ly49G2 molecules demonstrates that H2-M3 tetramers bind to Ly49A. Black solid histograms are staining with Ly49 specific antibody (left panel), fMIG loaded H2-M3 tetramer (middle panels) or ND1 loaded H2-M3 tetramer (right panel). Open histograms are isotype control antibody (left panel) or human HLA-B8 tetramer (middle and right panel). Results are representative of at least 4 independent experiments. (c). Antibody against Ly49A prevents binding of H2-M3 tetramers. CHO cells expressing Ly49A were pre-treated with isotype control antibodies or YE1/48 (Ly49A). Cells were then stained with H2-M3 tetramers loaded with fMIGWII or fMFFINIL. Black solid histograms are staining with H2-M3 tetramers on cells pre-treated with isotype control, grey histograms are cells pre-treated with Ly49 specific antibodies and open histograms are untreated cells stained with irrelevant human HLA-B8 tetramer. Histograms have been offset and angled by a factor of 5 to allow visualisation of all three populations. Results are representative of at least 3 independent experiments.

Figure 2. Ly49A binds H2-M3 with similar affinity to H-2D^d

H2-M3 refolded with f-MIGWII (a) or ND1 (b) peptides was serially diluted and injected over Ly49A immobilized by streptavidin on a BioRad ProteOn GLC chip. Serial dilutions of H-2D^d (c) and H-2K^b (d) served as positive and negative controls respectively. Sensorgrams show the binding (solid lines, response units, RU) of decreasing concentrations (16, 6.4, 2.6, 1 μM) of MHC following baseline subtraction. Binding affinities (K_D) were determined by kinetic analysis (lines show kinetic fits) using the ProteOn Manager 3.0.1 software. Dotted lines at 0 and 240 sec indicate injection start and injection stop respectively. Binding values are as follows: fMIGWII – H2-M3 K_a = 9.5 × 10² M⁻¹ s⁻¹, K_D = 2.75 μM, k_d = 2.62 × 10⁻³ s⁻¹ and Chi² = 3.1; fMFFINIL – H2-M3 K_a = 8.28 × 10² M⁻¹ s⁻¹, K_D = 4.06 μM, k_d = 3.36 × 10⁻³ s⁻¹ and Chi² = 2.83; AGPARAAAL – H-2D^d K_a = 1.65 × 10³ M⁻¹ s⁻¹, K_D = 2.05 μM, k_d = 3.38 × 10⁻³ s⁻¹ and Chi² = 2.14.

Figure 3. H2-M3 is not required for NK cell homeostasis

Analysis of NK cells from wild type and H2-M3-deficient mice demonstrated no difference in NK cell development and homeostasis. (a) Maturation of NK (NK1.1⁺CD3⁻) cells (assessed by CD27 and CD11b, CD43 and CD11b and NK1.1⁺ and CD122⁺ and CD3⁻ cells) in the spleen, liver, lung and bone marrow of wild type and H2-M3-deficient mice is presented. The density plots are representative of two independent experiments using 4 mice per time point (N=8). Frequencies of the populations are presented in the quadrants. Absolute numbers of NK cells in each organ were pooled from the two independent experiments (N=8) and are presented as mean ± SEM (lower panel). (b) The expression of Ly49, NKG2 and DNAM family receptors on splenic NK cells are presented. The density plots are representative of 3 independent experiments using 4 mice per time point (N=12). Frequencies of the populations are presented in the quadrants. (c) The expression of H-2K^b, H-2D^b and H2-M3 on bone marrow and spleen cells. Histograms are representative of 2 independent experiments using 4 mice per time point (N=8). The filled histogram is staining on H2-M3-deficient mice while the open histogram is wild type mice. Shaded histograms are isotype controls for H-2D^b and H-2K^b.

Figure 4. H2-M3 is required for NK cell control of B16F10 metastases and MCA-induced fibrosarcoma

Two independent models of NK cell-dependent control of tumours were used to demonstrate the role of H2-M3 in NK cell licensing. (a) WT and H2-M3-deficient mice were challenged intravenously with different doses of B16F10 as indicated and tumour metastases counted 14 days post injection. NK cells and CD8 T cells were depleted using anti-asialoGM1 or anti-CD8β, respectively. Results are pooled from two independent experiments using 5 mice per time point (N=10) and show the mean ± SEM. (* P=0.0491, ** P= 0.0002 at 10⁵ and 5 × 10⁴ and P=0.0006 at 1 × 10⁴) (b). WT and H2-M3-deficient mice were challenged with MCA and the development of fibrosarcoma assessed for 300 days (left panel). The numbers above the panels define the amount of MCA administered in micrograms/mouse. The panels show the percentage of tumour free mice. Some groups of mice were depleted of NK cells using anti-asialoGM1 on day −1,0 and weekly until day 42 after MCA inoculation. (* P=0.0365)

Figure 5. Ly49A⁺ NK cells from H2-M3-deficient mice are not licensed

(a) Representative gating scheme for analysis of intracellular IFN-γ production by Ly49A⁺ NK cells (Ly49A⁺ NK1.1⁺ CD3⁻ CD19⁻ NKG2ACE⁻ Ly49C⁻ Ly49F⁻ Ly49I⁻). In the top left panel singlet cells are gated and then lymphocytes are segregated on the basis of morphology (top right panel). NK1.1⁺CD3-CD19-NKG2ACE-Ly49CEIFH negative cells are then gated (middle right panel) and Ly49G2 on these cells is determined (middle left panel). On these cells (which are inhibitory receptor negative except Ly49A) the production of IFN-γ is then determined in the Ly49A⁺ and Ly49A⁻ (includes Ly49D and NKG2D) as shown in the bottom panels. The numbers in the density plot represents the percentage of cells. The density plots are representative of 4 independent experiments using 3 mice per time point (N=12). (b) The frequencies of IFN-γ production by the Ly49A⁺ and Ly49C/I⁺ populations from the 4 experiments were pooled and show the mean ± SEM (N=12). (c) B16F10 cells were i.v. injected into C57BL6 mice (1 × 10⁵ cells/mouse). Lungs were harvested at 14 days post injection. Results for YE1/32 and 2A3 are pooled from 3 independent experiments using 5–6 mice per experiment (N=16–17) while anti-asialoGM1 is from two experiments using 5–6 mice per experiment (N=11–12). Results are presented as mean ± SEM. *** P<0.0001 comparing WT control Ig to WT YE1/32 and *** P<0.0001 comparing WT control Ig to H2-M3^−/− control Ig.

Figure 6. Rejection of H2-M3^−/− bone marrow demonstrates “missing self” recognition

Bone marrow from wild type or H2-M3-deficient mice was fluorescently labelled and injected into wild type recipients for analysis of rejection by NK cells. Rejection of H2-M3-deficient bone marrow occurs in wild type mice in an NK cell-dependent manner. Bone marrow from wild type (CFSE labelled) and H2-M3-deficient (Cell Trace Violet) mice were labelled and mixed prior to injection into wild type mice. After 24 h leukocytes were harvested from spleen and liver and analysed by flow cytometry. The density plots are electronically gated to exclude host cells and are representative of 2–3 independent experiments using 3–5 mice per experiment (N=8 to 12). The numbers in the quadrants represent the frequency of cells within the region. The bottom panel is pooled data from the 2 independent experiments and shows the mean ± SEM (N=12). * P<0.0001 (spleen and liver). α-L is the Ly49A depleting antibody YE1/32 and α-N is the NK cell depleting antibody PK136.