Cutting Edge: Killer Ig-Like Receptors Mediate “Missing Self” Recognition In Vivo¹

According to the “missing self” hypothesis, NK cells are able to detect the absence of autologous MHC from target cells by virtue of inhibitory receptors specific for MHC class I. This hypothesis was invoked in part to explain the hybrid resistance phenomenon; in contradiction to the rules of T cell alloreactivity, irradiated F₁ hybrid mice reject bone marrow cells donated by either parent (1). The selective rejection of H-2-deficient, but not proficient, lymphoma variants by C57BL/6 mice lent strong support to this concept (2). Subsequent experiments showed that the presence of inhibitory Ly49 receptors on NK cells, specific for particular MHC class I allotypes, could indeed explain NK cell reactivity (3) and identified multiple inhibitory Ly49 molecules with distinct MHC class I specificities (4). Together, these experiments provided formal proof that hybrid resistance was mediated by NK cells recognizing “missing self” based on Ly49-MHC interactions.

Humans do not possess functional Ly49 genes but do express inhibitory receptors that specifically recognize HLA class I allotypes. These inhibitory killer Ig-like receptors (KIR³; reviewed in Ref. 5) are structurally distinct from the lectin-like Ly49 molecules. Yet, both receptor families share a similar expression pattern, as their members are expressed by NK cells, γδ T cells, and memory αβ T cells in a stochastic and independent fashion (5). In addition, inhibitory KIR-ligand interactions, or the absence thereof, explain the in vitro alloreactivities of human NK cells (6). In stem cell transplantation, KIR-mediated “missing self” recognition may explain the beneficial effects of certain HLA class I mismatches (7). However, formal proof that KIR-HLA interactions can mediate “missing self” recognition in vivo is lacking.

To test whether KIR-HLA interactions can mediate “missing self” recognition in vivo, we developed a KIR and HLA transgenic mouse model. This mouse is transgenic for an almost intact and fully sequenced KIR B haplotype (8, 9). We previously showed that KIR expression patterns in the KIR transgenic mice closely mimicked the human situation, as KIR2DL2 was expressed selectively by NK cells and in a stochastic fashion, with expression levels indistinguishable from those of human NK cells (9). Crossing these KIR transgenic mice with mice transgenic for a genomic HLA-Cw*0304 (HLA-Cw3) construct (10) introduced a potential inhibitory interaction between KIR2DL2 and HLA-Cw3. In this study, we examined the effects of host KIR and HLA on the NK cell KIR and Ly49 repertoires, NK responsiveness, and the ability to respond to “missing self.”

Mice transgenic for a KIR B-haplotype and on a mixed (C57BL/6 and CBA) genetic background (9) were back-crossed eight times onto C57BL/6 (The Jackson Laboratory) mice. This KIR B haplotype has been sequenced in full (8) and contains the following intact genes: KIR3DL3*003, KIR2DS2*001, KIR2DL2*003, KIR2DL4*005, KIR3DS1*013, KIR2DL5A*001, KIR2DS5*002, and KIR2DS1*002. The presence and integrity of the KIR locus was checked after every backcross by KIR genotyping. C57BL/6 mice transgenic for the genomic HLA-Cw*0304 construct (10) were gifts from E. Vivier (Centre d’Immunologie Marseille-Luminy, Marseille, France). All experiments were approved by the Leiden University Medical Center animal experimental committee and performed according to local guidelines.

In vivo rejection of CFSE-labeled spleen cells was performed as described (11). Briefly, two different populations of spleen cells, one internal syngeneic control expressing HLA-Cw3 and one lacking HLA-Cw3, were labeled with 0.5 and 5 μM CFSE (Invitrogen), respectively, and mixed in a 1:1 ratio. Seven days after i.v. injection of this mixture, peripheral blood of recipient mice was collected and analyzed by FACS. The relative rejection of HLA-Cw3^−/− target cells was calculated as follows: (1 − (acquired number of CFSE^high cells in sample/acquired number of CFSE^low cells in sample)/(acquired number of CFSE^high cells in injection mix/acquired number of CFSE^low cells in injection mix)) × 100%. The HLA-C specific Ab WK4C11 (12), combined with PE-conjugated rabbit anti-human IgM (Dako), was used to verify HLA-Cw3 expression in donor and recipient mice.

Mononuclear cells were isolated from spleen using a Ficoll-Hypaque gradient and incubated with fluorescently labeled Abs. Samples were acquired on a LSR II cytometer (BD Biosciences) and analyzed using FACSDiva software (BD Biosciences). NK cells were identified as CD3⁻NK1.1⁺ cells using a combination of CD3-Pacific Blue (clone 500A2; BD Biosciences) and NK1.1-PE-Cy7 (clone PK136; BD Biosciences). For staining of mouse NK receptors, the following FITC-conjugated Abs were purchased from BD Biosciences: Ly49A (clone A1), Ly49C/I (clone 5E6), Ly49D (clone 4E5), Ly49G2 (clone 4D11), and NKG2A/C/E (clone 20d5, BD). As C57BL/6 NK cells do not express appreciable levels of NKG2C and NKG2E (13), the specificity of the latter Ab was designated as NKG2A. A PE-conjugated Ab to KIR2DL2/KIR2DL3/KIR2DS2 (clone GL183; Coulter Immunotech) was used to detect KIR2DL2/KIR2DS2 because the mouse does not carry the KIR2DL3 gene.

Spleen NK cells were stimulated with plate-bound PK136 Ab (specific for NK1.1/Nkrp1c/Klrb1c) for 5 h, with the addition of brefeldin A after 1 h, and analyzed for intracellular accumulation of IFN-γ (using Ab clone XMG1.2; BD Biosciences) as described (14, 15).

To test whether KIR and HLA could mediate “missing self” recognition in vivo in the same way as Ly49 and mouse MHC class I, we analyzed the rejection of wild-type C57BL/6 spleen cells by KIR and HLA transgenic mice (Fig. 1). “Missing self” rejection was tested in those mice by using an in vivo assay based on differential labeling of donor cells with the CFSE dye (11). Mixed CFSE^high wild-type C57BL/6 and control CFSE^low HLA-Cw3^+/− spleen cells were injected i.v. into KIR^+/−HLA-Cw3^+/− or control KIR^−/−HLA-Cw3^+/− mice. KIR^+/−HLA-Cw3^+/− mice rejected on average 8% of injected wild-type spleen cells, whereas control KIR^−/−HLA-Cw3^+/− mice did not reject such targets (Fig. 1; p < 0.01). Therefore, KIR and HLA-Cw3 transgenic mice rejected wild-type cells, and this rejection was dependent on the presence of KIR in the host and on the absence of HLA-Cw3 from the target cells.

In both mice and humans, MHC class I ligands modulate the frequencies of NK cells expressing a cognate inhibitory NK receptor (16, 17), and it has been suggested that this optimizes the NK repertoire by ensuring that most NK cells express at least one inhibitory receptor specific for self-MHC class I (6). We therefore examined the frequency of Ly49 and KIR expression in mice transgenic for KIR and/or HLA-Cw3. The frequency of KIR2DL2/KIR2DS2 expression was consistent with observations in humans (17), but not significantly different between KIR^+/−HLA-Cw3^+/− and KIR^+/−HLA-Cw3^−/− mice (Fig. 2,A). The frequencies of NK cells expressing endogenous mouse receptors (Ly49A, Ly49C/I, Ly49D, Ly49G2, NKG2A) were also unaffected by the presence of KIR and/or HLA-Cw3 (Fig. 2 A). Therefore, there was no detectable effect of KIR and/or HLA-Cw3 on the endogenous mouse NK repertoire and no effect of HLA-Cw3 on the frequency of NK cells expressing KIR2DL2/KIR2DS2. In humans, HLA effects on KIR expression frequencies were detected only in the case of strong KIR-HLA interactions (17). The interaction between transgenic KIR2DL2*003 and HLA-Cw*0304, albeit strong enough to mediate rejection of “missing self” targets in vivo, may be too weak to detectably influence expression of KIR2DL2.

Individual KIR and Ly49 genes are expressed stochastically and largely independently of other KIR and Ly49 genes. However, there appears to be selection against NK cells expressing more than one inhibitory receptor specific for self-MHC class I (17, 18). In the KIR and HLA transgenic (KIR^+/−HLA-Cw3^+/−) mice, KIR2DL2/KIR2DS2 expression was relatively frequent on NK cells expressing Ly49D, Ly49G2, or Ly49A, none of which binds C57BL/6 MHC class I molecules (Fig. 2,B). In contrast, expression of KIR2DL2/KIR2DS2 was relatively infrequent on NK cells expressing Ly49C/I or NKG2A, inhibitory receptors that bind H-2K^b (Ly49C and Ly49I) and Q-a1 (CD94/NKG2A), respectively (Fig. 2,B). This suggested that NK cells expressing two functional inhibitory receptors were indeed underrepresented in the NK repertoire. Yet, removing HLA-Cw3 from the equation by using KIR^+/−HLA-Cw3^−/− mice did not alter the KIR and Ly49 coexpression patterns (Fig. 2 B). Therefore, coexpression of transgenic human and endogenous mouse NK receptors was not random, but this effect was independent of the presence of HLA-Cw3. This would be consistent with a sequential model of NK receptor acquisition (19, 20), in which acquisition of KIR2DL2/KIR2DS2 occurs later in development and is less likely to occur in cells that have already expressed a Ly49 with a ligand present.

NK cells expressing inhibitory receptors that bind endogenous MHC class I are more responsive to activating stimuli than NK cells that do not express such “useful” inhibitory receptors (14, 15, 21). Hence, the presence of HLA-Cw3 might influence the potency rather than the frequency of NK cells expressing a HLA-Cw3-specific inhibitory receptor. In line with previous reports (14, 15), NK cells expressing Ly49C, Ly49I, and/or NKG2A, all self-specific inhibitory receptors in C57BL/6 mice, produced considerably more IFN-γ in response to NK1.1 crosslinking than mice lacking these receptors (Fig. 3). In the latter NK cell subset, expression of KIR2DL2/KIR2DS2 was associated with a slightly increased IFN-γ production, but this effect was not detectably influenced by the presence of HLA-Cw3 (Fig. 3). Thus, the educational impact of interactions between mouse NK receptors and their ligands was considerably greater than the effect of the KIR2DL2-HLA-Cw3 interaction in these mice.

These results provide formal proof that KIR and HLA can mediate “missing self” recognition in vivo. KIR-mediated rejection of “missing self” targets must be mediated by NK cells, as these are the only cells expressing KIR in these mice (data not shown). Rejection is most likely mediated by KIR2DL2, as this is the only known inhibitory receptor binding HLA-Cw3 in this mouse model, whereas the activating KIR2DS2 does not bind HLA-Cw3 (22). Indeed, approximately one-third of KIR2DL2/KIR2DS2⁺ NK cells, but no KIR2DL2/KIR2DS2⁻ NK cells, bound the HLA-Cw3 tetramer and therefore expressed KIR2DL2, whereas all tetramer-negative KIR2DL2/KIR2DS2⁺ cells bound a KIR2DS2/KIR2DL3-specific Ab and therefore expressed KIR2DS2 (data not shown).

It is not entirely surprising that rejection (Fig. 1) occurred in the absence of detectable HLA effects on selection (Fig. 2) and education (Fig. 3). In vitro experiments with human NK cells classically show a clear-cut effect of KIR/HLA interactions on lysis of “missing self” targets (23), whereas effects on NK receptor repertoire (17) and responsiveness (21) tend to be more subtle. Furthermore, it is possible that the strong interaction between Ly49C and H-2K^b dominates the NK repertoire in our mice, obscuring the effects of KIR-HLA interactions (24).

Due to the extreme genetic diversity of KIR and HLA in man, it has been difficult to examine KIR function in a controlled in vivo setting. This humanized mouse model will be useful for examining KIR-HLA effects on disease, transplantation, and reproduction. Furthermore, it provides a useful preclinical platform for “proof of principle” studies for KIR-based cancer therapies such as haplo-identical stem cell transplantation, NK cell-adoptive transfer, and blocking of inhibitory KIR.

We are grateful to Geert Westerhuis and to Dr. Melissa van Pel for guidance in mouse experiments, to the National Institutes of Health tetramer facility for providing HLA-Cw3 tetramers, to Dr. A. Mulder for the WK4C11 Ab, and to Dr. E. Vivier for providing HLA-Cw3 transgenic mice.

The authors have no financial conflict of interest.