Murine Cytomegalovirus Encodes a Specific Cell Surface Marker That Can Trigger Attack by NK Cells

NK cells use a wide array of germline-encoded receptors that scan the surface of cells in the body to identify and spontaneously kill abnormal cells, especially malignant, virus-infected, or senescent cells. Inhibitory receptors that engage with MHC class I (MHC-I; H-2 in mice, HLA-A, B, and C in humans) provide NK cells with critical tolerance toward normal healthy cells, whereas numerous activating receptors detect expression of various stress-induced ligands on abnormal or infected cells. Encounter with abnormal cells that often downregulate MHC-I and/or upregulate stress-induced ligands shifts the NK cell’s intracellular signaling dynamic toward stimulation of targeted cytolysis and production of chemokines and cytokines, including IFN-γ and TNF-α. The 1990s was an exciting period in the history of NK cell research, when a plethora of their germline-encoded cell surface receptors were cloned and characterized. The rapid breakthroughs in receptor repertoire discovery created a multitude of subsequent exciting pursuits to identify their cognate ligands, which is still ongoing.

One major question within the field that remained unanswered in the early 2000s was, how NK cells from certain mice (notably C57BL/6) could specifically detect and promote clearance of murine CMV (MCMV), whereas NK cells from other strains are minimally responsive (such as BALB/c and 129/J), resulting in severe and often fatal infection from this prevalent herpesvirus? This puzzle originated in 1982, when Dr. Cornelis Melief’s laboratory discovered that MCMV infection stimulates potent IFN production and NK cell activation to provide protective immunity, but only in certain mouse strains, and the effect appeared independent of H-2 background (1). A decade later, Scalzo et al. (2, 3) showed that the protective immune response was mediated by NK cells and linked to a region on chromosome 6 known as the NK cell gene complex, specifically within a locus named Cmv-1. The quest to map the elusive gene product encoded in Cmv-1 required another decade of intensive research and culminated in publications from three groups in 2001 that identified Ly49H (Klra8) as the key receptor protecting the MCMV-resistant mouse strains (4–7). Ly49 receptors are C-type lectin-like proteins encoded by a polymorphic multigene family and consist of inhibitory and activating forms. Ly49H is an activating family member that physically associates in the plasma membrane with the signaling adaptor DAP12, which contains an ITAM to mediate potent stimulatory signaling, analogous to T and B cell Ag receptors. The identification of Ly49H as the requisite receptor mediating MCMV recognition was a key breakthrough, yet the mystery of how this receptor detects MCMV-infected cells remained unresolved.

In this Pillars of Immunology article, Dr. Hisashi Arase in Dr. Lewis Lanier’s laboratory, and their collaborators, advanced this discovery to the next level by providing compelling evidence that the MCMV-encoded protein m157 is the ligand for Ly49H (8). The specific recognition of a virus-encoded protein by an NK cell activating receptor was particularly astonishing, and this remarkable breakthrough was also reported soon thereafter by Dr. Wayne Yokoyama’s laboratory (9). m157 is a GPI-linked protein expressed on the surface of MCMV-infected target cells. The interaction of Ly49H with m157 was found to trigger potent cytotoxicity of the target cells and IFN-γ production by NK cells, thereby providing the basis for early viral clearance in MCMV-resistant mouse strains. Computer analysis of the sequence of m157 revealed MHC-I–like structural properties, consistent with the MHC-I ligand recognition features of most Ly49 family receptors.

Arase et al. (8) also screened for detection of m157 by other Ly49 family members, which revealed binding to the inhibitory Ly49I receptor from the MCMV-susceptible 129/J mouse strain. This finding suggested that some of the susceptibility to MCMV in certain mouse strains, such as 129/J, might result from m157 engaging with Ly49I to abrogate attack by a subset of NK cells. Based upon this result, Arase et al. (8) further speculated that m157 may have originally been beneficial in the MCMV genome by providing protection from NK cell killing through binding to inhibitory Ly49I, but pressure to survive this deadly virus may have resulted in evolution within the Ly49 gene locus to establish the activating Ly49H receptor for selective recognition of m157. The authors further suggested that some activating forms of the functionally related killer cell Ig-like receptor family in humans may have also evolved to recognize specific pathogen-encoded ligands. In support of these possibilities, evolutionary studies have provided evidence for dynamic recombination and gene duplication events within the Ly49 and KIR gene loci that were driven by selection pressures, including viral infections, to promote the evolution of beneficial activating receptor variants (10).

The discovery that the activating Ly49H receptor detects m157 on the surface of MCMV-infected cells revealed a highly specific stimulatory pathway on NK cells to protect mice. This model system has since been widely exploited as a valuable tool to study murine NK cell functional responses to viral infection through adoptive transfer, transgenesis, and knockout experiments. Most notably, m157 recognition by Ly49H has been instrumental in elucidating the concept of NK cell “memory” responses. The subset of Ly49H-expressing NK cells was found to selectively proliferate in MCMV-infected mice and become long-lived, self-renewing memory cells capable of mediating an m157-dependent secondary expansion and more potent recall response long after initial exposure to the virus (11). The secondary proliferative response of these memory NK cells was found to depend upon IL-12 (12), and a similar expansion of NKG2C-expressing “adaptive” or “memory-like” NK cells is observed in response to human CMV (13), although the mechanistic basis for potential involvement of NKG2C in human CMV responses remains unclear.

After the publication by Arase et al. (8), the MCMV-encoded m04 protein was also found to differentially modulate NK cell–mediated attack of infected cells in specific mouse strains. For example, the activating Ly49P in MA/MyJ mice recognizes m04 in association with H-2D^k alleles to stimulate NK cells (14, 15). In contrast, however, m04 was also found to promote surface expression of MHC-I on infected cells, allowing inhibitory Ly49G2 receptors to suppress NK cell attack in H-2^k+ BALB mice, which notably lack the activating Ly49P or Ly49H receptors (16). These and other observations have exemplified diverse impacts of MCMV-encoded proteins on NK cell responses in different genetic backgrounds of mice. In addition, many mouse strains lack expression of Ly49H, and wild-type MCMV strains may express m157 variants incapable of interacting with Ly49H (17). These findings reinforce the fact that Ly49H/m157 interaction does not provide universal resistance to MCMV, and more virulent viral strains may be evolving through evasion of this highly effective NK cell–mediated protective mechanism. Nonetheless, the Ly49H/m157 axis remains the most direct and specific recognition system yet identified for promoting effective MCMV protection by NK cells and has become a valuable model system for studying NK cell biological responses in vivo.