NK Cells Lose Their Inhibition

See allHide authors and affiliations

Science  06 Aug 2004:
Vol. 305, Issue 5685, pp. 786-787
DOI: 10.1126/science.1102025

Hepatitis C virus (HCV) is quietly colonizing the vast, plentiful archipelago that we call the human species (1, 2). For those infected, the immediate effects are mild and attract little attention. The evidence of infection is obtained only later, either in the form of virus-specific immunity or liver disease. About 20% of infected individuals clear the virus. For the rest—now numbering ∼180 million persons worldwide—the infection and its accompanying immune reactions persist for decades and can lead to liver cancer or liver failure. There is no vaccine for HCV, but for those on the chase, it helps to know what factors distinguish a successful immune response from one that just blunders around. One such factor is a forceful cytotoxic T cell response aimed at viral peptides presented by a range of HLA class I molecules expressed by immune cells and many other cell types (3, 4). On page 872 of this issue, Khakoo et al. (5) report that another function of HLA class I molecules—the regulation of a type of immune cell called a natural killer (NK) cell—may also influence the fate of HCV infection.

NK cells are the lymphocytes that secrete cytokines and kill infected cells at early stages of a primary viral infection. Such speed is possible because NK cells abound in blood as differentiated effectors that are ready to go. On entering infected tissue, NK cells are activated by antigen-presenting cells called dendritic cells through surface receptors that sense microbial products or cellular stress (6). In contrast, in healthy tissue the activation pathways are kept in check by signals coming from inhibitory receptors (see the figure). In NK cells, this control is mediated by one or more inhibitory receptors that recognize self HLA class I molecules. Such inhibitory receptors include CD94:NKG2A, which binds to HLA-E, and certain members of the killer-cell immunoglobulin-like receptor (KIR) family, which bind to HLA-A, -B, or -C. Whereas CD94:NKG2A and HLA-E are conserved, HLA-A, B, C, and their corresponding KIR receptors are highly polymorphic (7). This variability and their functional interactions imbue KIR and HLA class I molecules with considerable potential as biomarkers of disease progression.

Taking the brakes off NK cells.

(Top) The activity of NK cells is determined by the integration of signals coming from activating receptors and inhibitory receptors such as KIR2DL. (Middle) The three interactions between HLA-C1 and HLA-C2 and KIR2DL receptors that inhibit NK cells. The weak interaction between KIR2DL3 and HLA-C1 can be easily overcome such that individuals with this genotype more readily resolve infection with hepatitis C virus than infected individuals of other genotypes (5). (Bottom) The two common chromosomal arrangements of KIR2DL genes.


With this objective, Khakoo et al. determined the KIR and HLA genotypes of 1037 individuals, who were exposed to HCV and had raised a virus-specific immune response. Comparison of the 352 individuals who resolved infection to the 685 with persisting HCV revealed a difference in the distribution of HLA-C ligands and their inhibitory KIR2DL receptors. The HLA-C locus encodes two forms of KIR2DL ligand defined by the presence of either asparagine (HLA-C1) or lysine (HLA-C2) at position 80 in the protein's sequence. HLA-C1 allotypes are the ligands for the KIR2DL2 and KIR2DL3 receptors, whereas HLA-C2 allotypes are the ligands for the KIR2DL1 receptor. Khakoo et al. show that individuals who are homozygous for HLA-C1 and KIR2DL3 resolve HCV infection more frequently than individuals with other genotypes (see the figure).

Given that the KIR and HLA genes are on different chromosomes, the correlation itself is evidence that the underlying mechanism for the resolution of HCV infection involves HLA-C1 binding to KIR2DL3. At first this seems paradoxical; how does inhibiting an NK cell help in viral clearance? A consideration of the alternatives was the key to solving this puzzle. Khakoo et al. make the case that in HLA-C1:KIR2DL3 homozygotes there is a relatively weak inhibition of NK cells by HLA-C. Their brief goes like this: HLA-C1 homozygosity necessitates that the HLA-C2 ligand be absent, so its KIR2DL receptor (which most people have) is nonfunctional. Likewise, KIR2DL3 homozygosity necessitates that KIR2DL2 be absent, because KIR2DL2 and KIR2DL3 behave as alleles (see the figure). KIR2DL3 has a lower affinity for HLA-C1 than does KIR2DL2 (8), so HLA-C1-mediated inhibition of NK cells is inherently weaker in KIR2DL3 homozygotes than in either KIR2DL3/2DL2 heterozygotes or KIR2DL2 homozygotes. Hence, HLA-C1:KIR2DL3 homozygotes are predicted to be rich in NK cells using KIR2DL3 as their inhibitory receptor for self HLA class I. For these NK cells, the inhibitory control will be weaker than for other NK cells, so they will be more easily activated by viral infection. In short, Khakoo et al. propose that in responding to HCV infection, individuals homozygous for HLA-C1:KIR2DL3 more effectively activate NK cells than individuals with other genotypes.

Although HCV appears to be unique to humans, its origins and history are mysterious (9). Transmission occurs mainly through direct contact with infected blood—for example, through transfusion of contaminated blood or blood products, or injections or accidental wounds using contaminated syringe needles. These two modes of infection differ in the quantity of virus delivered: Blood transfusions contain “high doses” of virus, which likely exceed by orders of magnitude the “low doses” accompanying injection with a contaminated needle. The patient panel studied by Khakoo et al. comprised 490 individuals infected by blood infusion and 533 infected by injection. Although the frequency of resolved infection was the same in both groups, the dramatic association between HLA-C1:KIR2DL3 homozygosity and resolution of HCV infection was seen only in the group infected by injection.

The two modes of infection differ also in the anatomy of the infection site. Infection by injection is likely to begin as a local affair, like many normal infections, in which an inflammatory response involving NK cells starts at the site of injection and spreads to the draining lymphoid tissue to stimulate B and T cells. Later, as HCV reaches the liver, its preferred site of replication, further stimulation of the immune response could take place. In contrast, in a blood transfusion, massive quantities of HCV are introduced directly into the systemic circulation, from where the virus may quickly reach the liver and also the spleen. The amount of virus also may be sufficient to suppress NK cell functions by binding directly to these cells (10, 11). All of this is necessarily speculative, because HCV biology is poorly understood. The main point is that HCV and human immunity may well behave differently after the two modes of infection, as is implied by their differential association with KIR2DL3:HLA-C1. While in this cautionary vein, one should note that KIR receptors are expressed by some T cells (12), and they too might contribute to the effect of KIR2DL3:HLA-C1 on HCV infection.

HCV is periodically compared to HIV and judged a Cinderella (13). Both are rapidly evolving RNA viruses that have exploited modern mores and medicine to make their way around the world. Although HIV captures more attention, HCV captures more bodies: 180 million compared to 40 million for HIV. The rates of morbidity and mortality for HCV are lower than those for HIV and, unlike HIV, a substantial proportion of those infected get rid of the virus. This enhances the survival of the host species, but also of the virus, by maintaining a reservoir of subjects that can spawn future HCV infections. It also gives immunologists hope that vaccines will be developed to prevent and control HCV infection. In treating acute myelogenous leukemia with allogeneic stem cell transplantation, it is possible to improve the clinical outcome by temporarily releasing NK cells from the inhibition caused by interactions of KIR2DL receptors with HLA-C expressed by dendritic or other cells (14). Is it also possible that a similar strategy of inducing NK cells to lose their inhibition could encourage the resolution of HCV infection? What that would require is some sort of “lager” for lymphocytes that would not affect the liver.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
View Abstract

Stay Connected to Science

Navigate This Article