NK cell recognition of herpesviruses : mechanisms of viral immune escape

Abstract

Herpesviruses are ubiquitous pathogens which after an often asymptomatic infection establish lifelong latency in the host. Control of the viral infection depends on both innate and adaptive immune system. Clinical evidence and murine models have firmly established a decisive role of Natural Killer (NK) cells in the early control of herpesvirus infections. In order to escape adaptive T cell responses, herpesviruses target antigen presentation in the infected host cell. This usually leads to downregulation of MHC class I molecules which in turn should increase susceptibility to NK cells. However, a broad arsenal of viral gene products counteracts this danger and interferes with NK cell recognition on various levels. The studies in this thesis address three of these immunoevasins in human cytomegalovirus (HCMV) are discussed: UL16, UL18 and UL40.

In paper I we showed that productive HCMV infection induces ULBP1, ULBP2, ULBP3 and MICA, ligands for the NK cell activating receptor NKG2D. The upregulation of ULBP1 and ULBP2 was delayed on the cell surface in the presence of the viral protein UL16, known to bind to these two molecules. Using a HCMV mutant deficient for UL16, we observed higher levels of NKG2D ligands on infected target cells and this increase resulted in a higher susceptibility to NKG2D-mediated polyclonal NK cell responses. This is one of the first studies demonstrating the direct viral targeting of an NK cell activating pathway during productive infection.

In paper II we tried to investigate the molecular reasons for the previously known extraordinary high affinity of the viral UL18 protein, an MHC class I homolog, to the inhibitory cellular receptor LIR-1, that had been reported earlier. Based on a model of the UL18 structure, we introduced mutations in different sites of UL18 and assessed their influence on binding and on the functional consequences for the LIR-1/UL18 interaction. Substitutions of residues K42/A43 and Q202, located in the alpha1- and alpha3-domain respectively, reduced the binding affinity between UL18 and LIR-1 by about half. Disruption of an additional disulfide bridge, predicted by our model, completely abolished binding of UL18 to beta2m and the interaction with LIR-1. We propose that the high binding affinity of UL18 for LIR-1 is in part due to a more stable alpha3-domain and a larger binding interface when compared to HLA-A2/LIR-1

In paper III we extended the concept of herpesvirus interference with the NKG2D pathway to Herpes simplex virus (HSV) type I. We detected higher NKG2D levels on NK cells recovered from patients with acute blister formation. In vitro we could demonstrate that HSV downregulates the NKG2D ligand MICA by an as of yet undefined late viral gene. Surprisingly, the lysis of HSV infected target cells by polyclonal NK cells became more dependent on NKG2D, possibly indicating that other viral evasion mechanisms targeting NK cell recognition operate in parallel.

In paper IV we revisited the HCMV protein UL40, that has been described to upregulate HLA-E on infected cells and thereby transduce inhibitory signals via the NK cell receptor CD94/NKG2A.We hypothesized that UL40 may have additional functions and using an HCMV mutant deficient for UL40, observed decreased levels of cellular heat shock protein (Hsp) 60 and 70, when compared to cells infected with a wild type virus. Functional knockdown of these molecules resulted in higher susceptibility to NK cell lysis and apoptotic stimuli, suggesting that HCMV induces Hsps in order to prevent target cell death. Based on previous reports regarding the role of Hsp derived sequences in the context of CD94/NKG2A-HLA-E dependent recognition, we propose a model for UL40 action that can reconcile our findings with the existing literature. The multiple recognition strategies of NK cells and the multiple immune evasion molecules of herpesviruses offer a fascinating system to learn more not only about viruses but also about the immune system and the co-evolution driven by host-virus interactions.