NK cells are cytotoxic effector cells that play a vital role in the innate immune response to viral infections (
9,
12,
33). The critical role of NK cells in acute viral infections has been best characterized in acute murine cytomegalovirus (MCMV) infection (
14,
28). While several murine lab strains are resistant to MCMV infection, others are highly susceptible. Resistance to MCMV infection was mapped to a gene encoding an activating NK cell receptor, Ly49H, which has been shown to be critical in the early recognition and control of MCMV infection via the direct recognition of a viral product (M157) expressed on infected cells (
28). Remarkably, MCMV-infected mice exhibit a dramatic expansion of NK cells during acute infection, but this expansion is restricted to the specific accumulation of Ly49H
+ NK cells (
16). Data from these studies suggest that the antiviral activity of the Ly49H
+ NK cells is linked to their ability to expand early in infection, prior to the development of adaptive antiviral immunity.
While the critical role of Ly49H
+ NK cells in MCMV infection has been well established, very little is known about the clonal composition of NK cells that expand in human viral infections, and the NK cell receptors that mediate their antiviral activity. Unlike T cells and B cells, the specificity of NK cells is not determined by a single NK cell receptor (
8); rather, NK cells express an array of activating and inhibitory receptors that regulate their activity. While the expression of these receptors is stochastic, the random combinations of different receptors on the surface of a given NK cell clone determine its ability to respond to a specific target cell (
26,
27). It has been suggested that individual NK cell populations expressing a specific array of receptors may respond differentially to diverse viral infections (
7). This has been further supported by epidemiological studies associating the expression of individual activating or inhibitory NK cell receptors in combination with their HLA class I ligands with better or worse disease outcomes in viral infections such as hepatitis C virus (
22), human immunodeficiency virus (HIV) (
29,
30), human papillomavirus (
11), and CMV (
7). The functional basis for this protective immunity mediated by NK cells in human viral infections remains largely unknown.
Similar to MCMV infection, highly functional NK cells expand rapidly in acute HIV-1 infection, prior to the induction of adaptive immune responses (
2). One particular activating killer immunoglobulin-like NK cell receptor (KIR3DS1), in combination with its putative ligand, an HLA-B allele with isoleucine at position 80 (HLA-B Bw480I), has been shown to be associated with slower HIV-1 disease progression (
29). We have recently shown that KIR3DS1
+ NK cells can effectively suppress HIV-1 replication in HLA-B Bw480I
+ target cells in vitro (
1). Furthermore, a subset of inhibitory alleles from the same locus, KIR3DL1, that show high cell surface expression levels have similarly been associated with slower disease progression toward AIDS in the presence of their ligand, HLA-B Bw480I (
30). These data suggest that both KIR3DS1
+ and KIR3DL1
+ NK cells may play a critical role in the control of natural HIV-1 infection, depending on the interaction with their ligand on infected cells (
4). However, the mechanisms underlying their protective role are not understood.
DISCUSSION
Acute viral infections are typically characterized by a rapid expansion of NK cells. In MCMV infection, this expansion is preferentially restricted to a population of NK cells expressing the activating receptor Ly49H
+, which recognizes a ligand encoded by the virus (
5). While a dramatic expansion of total bulk NK cells occurs during acute HIV-1 infection (
2), little is known about the specific expansion of individual NK cell populations during this early window of infection. The delineation of clonal NK cell expansion in human viral infections is seriously hampered by the limitations of commercially available antibodies that only recognize half of the KIRs and do not differentiate between activating and inhibitory KIRs. We therefore adapted a quantitative RT-PCR approach (
13) to monitor changes in transcriptional expression of KIR mRNAs in NK cells, and demonstrate that acute HIV-1 infection is marked by a rapid nonspecific increase in activating KIR transcription, an upregulation of KIR2DL1 transcripts in chronic HIV-1 infection, and a clonal NK cell redistribution that is strongly regulated by the presence of HLA class I ligands.
HIV-1 peak viremia declines following acute infection, and the level of early containment of HIV replication correlates with the rate of disease progression (
20). While CD8
+ T cells have been implicated in this containment of viral replication (
25,
35), several additional factors may contribute to the initial decline of viremia. One proposed factor is the rapid loss of CD4
+ T cells, in particular in gut-associated lymphoid tissue, resulting from HIV-1 infection, limiting the number of available target cells (
34). Furthermore, the decline in HIV-1 replication and the resolution of acute infection symptoms frequently precede the development of significant virus-specific B- and T-cell responses, strongly suggesting an important role of the innate immune response in this early containment of infection while the adaptive immune response is still developing. Epidemiological evidence suggests that individuals that coexpress KIR3DS1 or some alleles of KIR3DL1 in conjunction with their putative HLA class I ligand (HLA-B Bw480I) exhibit significantly lower viral replication and slower progression toward AIDS (
29,
30); however, the underlying mechanism(s) for the protective effect of these KIR/HLA compound genotypes remains to be defined.
Previous studies have shown that NK cells expand in numbers during the first days to weeks of acute HIV-1 infection (
2); however, it is uncertain whether specific populations of antiviral NK cells expand during this early window. Using a quantitative RT-PCR approach (
13), we studied changes in transcriptional levels of individual KIR mRNAs and KIR
+ NK cell populations over the course of HIV-1 infection. Similar to acute MCMV infection (
16), we demonstrate that acute HIV-1 infection is also associated with a nonspecific expansion of activating KIR (KIR2DS1, KIR2DS2, and KIR3DS1)-expressing NK cells. However, following this nonspecific expansion, a specific accumulation of protective KIR3DS1- and KIR3DL1-expressing NK cells appears to occur, but only in the presence of their putative ligand, HLA-Bw480I. Taken together, these data along with previously published work by Martin et al. (
29,
30) suggest that part of the protective activity of these KIR/HLA compound genotypes may be the result of a proliferative signal elicited by HLA-B Bw480I to KIR3DS1 and KIR3DL1
+ NK cells, allowing them to expand sufficiently to potentially help contain early viral replication in acute HIV-1 infection.
What might account for the activation and expansion of KIR3DS1
+ and KIR3DL1
+ NK cells in acute HIV-1 infection? The coexpression of KIR and its HLA class I ligand has been shown to result in a preferential accumulation of NK cells bearing the binding KIR in healthy donors compared to individuals that do not express the KIR ligand (
37). Here we show that HIV-1 infection exaggerates these differences in such a way that the proportion of NK cells expressing KIR3DS1 and KIR3DL1 is significantly higher in subjects that coexpress the putative HLA class I ligand family for these receptors. Inhibitory KIR ligands have recently been shown to play a critical role in modulating NK cell function during NK cell development (
6,
17,
23). Engagement of an inhibitory KIR by its ligand during NK cell development serves as a critical checkpoint that results in the generation of highly functional NK cells that are easily inhibited by self HLA class I molecules expressed on normal cells in the periphery (
6,
17,
23). Given this intimate interaction between the NK cell receptor and its ligand during early NK cell development, it is plausible that HLA-B Bw480I may deliver critical signals to KIR3DL1
+ NK cells allowing them to develop into highly functional killers that are able to respond aggressively to target cells that do not express their respective self ligands at a sufficient level. This model supports epidemiological data demonstrating that KIR3DL1 allele subtypes encoding for KIR3DL1 receptors expressed at higher levels on the surface of NK cells are highly protective against HIV-1 disease progression when expressed in conjunction with their ligand HLA-B Bw480I (
30). KIR3DL1 molecules expressed at high levels may deliver stronger licensing signals to the NK cell during development when engaged with HLA-B Bw480I compared to KIR3DL1 molecules expressed at lower levels, resulting in more potent NK cell populations (
4). This is further supported by recent data suggesting a dose-dependent effect of ligand on KIR-mediated NK cell licensing, as increased expression levels of inhibitory NK cell receptor ligands correlate with elevated NK cell effector functions (
24). Thus, overall the presence of HLA-B Bw480I may deliver increased licensing signals through inhibitory KIR3DL1 to NK cells, resulting in higher activation of these NK cells in response to target cells expressing no or reduced levels of HLA class I. Additionally we observed a slight increase in the level of both KIR3DS1 and KIR3DL1 transcripts and cell numbers even in HIV-uninfected controls that possessed HLA-B Bw480I. These results suggest that, even in the absence of infection, the presence of HLA-B Bw480I may have an impact on the persistence of populations bearing these receptors in the peripheral circulation, albeit at significantly lower levels than those observed in acute HIV-1 infection.
While NK cell licensing provides a model to explain the preferential expansion of KIR3DL1
+ NK cells in HLA-B Bw480I
+ subjects, thus far licensing appears to function through inhibitory KIR only. Thus, a different mechanism might account for the observed changes in KIR3DS1
+ NK cells following exposure to HIV-1. Previously we have shown that KIR3DS1
+ NK cells become activated by CD4
+ T cells expressing HLA-B Bw480I when they become infected with HIV, resulting in potent inhibition of viral replication (
1). These data suggest that KIR3DS1 may sense HIV infection specifically through HLA-B Bw480I. While the precise ligand for KIR3DS1 has not been identified, epidemiological and functional data suggest a crucial role of HLA-B Bw480I molecules in modulating the function of KIR3DS1
+ NK cells during HIV-1 infection (
1,
30). In addition, here we observed an early expansion of KIR3S1
+ NK cells in the presence of HLA-B Bw480I, suggesting that in addition to delivering a potent antiviral signal to KIR3DS1
+ NK cells, HLA-B Bw480I may also provide a proliferative signal during acute HIV-1 infection to drive the expansion of this potentially antiviral subset of NK cells. Despite the fact that there is no concrete evidence thus far to demonstrate a physical interaction between KIR3DS1 and HLA-B Bw480I, an activating NK cell receptor in mice, which appears to be involved in the control of MCMV infection, was recently shown to analogously interact with its major histocompatibility complex class I ligand H-2D
k only in collaboration with another yet-undefined protein (
15). Thus, it is equally possible that the interaction between KIR3DS1 and HLA-B Bw480I, which may present specific HIV-1-derived epitopes or epitopes derived from self stress proteins, might require an additional cellular protein expressed during HIV-1 infection.
Overall we demonstrate the HLA-B Bw480I-dependent expansion of KIR3DS1+ and KIR3DL1+ NK cells in acute HIV-1 infection. These results demonstrate for the first time a ligand-specific expansion of KIR+ NK cell populations in an acute human viral infection. The early accumulation of highly activated NK cells may provide a potent first-line defense allowing for the initial control of acute HIV-1 replication while adaptive immune responses are still developing.