There are two mechanisms by which NK cells may affect disease. As already discussed above, NK cells do not kill targets that express “self ” class I HLA molecules. Viral infection, like some cancers, may downregulate class I MHC, allowing for killing of virally infected cells by direct cytotoxic mechanisms. A second major role of NK cells is their production of cytokines, which may be critically important to eliminate infections. Understanding of the pathophysiology of viral infection and immune recognition by NK cells has been advanced by the discovery of NK cell-inhibitory receptors that recognize MHC molecules. For instance, NK cell clones were isolated from peripheral blood to study the lysis of HeLa cells before and after infection with herpes simplex virus or cytomegalovirus (CMV). Both viral infected targets downregulate MHC class I molecules, making them undetectable by T cells, but more susceptible to NK cell killing than uninfected control cells.34 However, the factors that influence MHC class I expression are complex and may underlie the mechanisms of escape from NK cell-mediated lysis of virus-infected targets.
Several studies have evaluated the mechanisms of CMV infection and specificity of MHC downregulation. Unlike many viral proteins that downregulate class I molecules, CMV US2 and CMV US11 are specific to HLA-A and -B alleles, but do not affect HLA-C molecules showing allelic specificity.35 CMV UL40 has been interesting because of the homology between the signal sequence (amino acids 3–11) of most HLA-C alleles and the UL40 peptide (amino acids 15–23). Two independent laboratories have recently found UL40 sequence homology to HLA-E binding peptides.36,37 The first study used fibroblasts that constitutively express low levels of HLA-E to test UL40 constructs. UL40 bound HLA-E and increased its expression, without affecting expression of HLA-C, and protected against NK cell attack through CD94 lectin inhibitory receptors.36 The second study co-transfected HLA-E and UL40 into K562 cells, which became resistant to killing by NKG2A+ NK cell clones.37 Blocking antibodies to CD94 or HLA-E restored lysis, showing the specificity of this interaction. Therefore, some CMV proteins (US2 and US11) promote NK cell attack by downregulation of MHC, but UL40 has an opposing effect by stabilizing and upregulating HLA-E. Other viral proteins may have similar allelic specificity, but not all downregulation of MHC is necessarily associated with increased susceptibility to NK killing. Kaposi’s sarcoma-associated herpes virus (KSHV) proteins K3 and K5 dramatically downregulate class I molecules. K3 downregulates HLA-A, -B, -C and -E, while K5 downregulates HLA-A and -B. Surprisingly, KSHV K5 transduced targets, which do not downregulate Cw4, were resistant to an NK cell line that lacked a Cw4-specific KIR.38 Thus, K3 induces target sensitivity by MHC downregulation, and K5 induces resistance independent of MHC expression. The mechanism of this K5-induced resistance was found to be by downregulation of target cell intracellular adhesion molecule-1 needed for NK recognition and downregulation of B7-2 needed for effector cell activation. This illustrates a novel mechanism by which KSHV downregulates class I molecules, which should make it more susceptible to NK lysis but, by other mechanisms, escapes this immune control.
Interferon (IFN)-γ is probably the most important cytokine produced by NK cells with anti-infection activity. Mice that lack IFN-γ receptors have an increased susceptibility to murine CMV, listeria monocytogenes and the protozoan parasite, Toxoplasma gondii.39 Detailed studies of murine CMV infection show that IFN-γ and tumor necrosis factor (TNF)- α levels are upregulated in mice infected with murine CMV.40 By selective depletion of NK cells, IFN-γ production is eliminated, suggesting that NK cells are the primary producers of this cytokine. The role of IFN-γ in the pathogenesis of murine CMV is suggested by studies using monoclonal antibodies to neutralize IFN-γ, resulting in an increase in viral replication and subsequent severity of viral hepatitis. Interestingly, neutralizing IFN-γ does not alter direct cytotoxicity, suggesting that direct killing of infected cells is not the mechanism of NK cell control of murine CMV. Further studies showed that IL-12 stimulates IFN-γ in NK cells to mediate a protective effect against murine CMV. This may be virus-specific since lymphocyte choriomeningitis virus does not induce IFN-γ and was not affected by IL-12 neutralization.41