Cmv5s Drives Inflammation and Macrophage Loss, Promoting Viral Infection-Induced Tissue Damage and Corrupting Protective NK Cell Immunity

MHC class I (MHC-I) molecule H-2Dk conveys resistance to murine cytomegalovirus (MCMV) infection in both C57L and MA/My mice, mediated by a subset of NK cells bearing the H-2Dk-binding Ly49G2 receptor. M.H2k/b mice are MA/My congenic aside from a C57L-derived region surrounding the MHC, which dominantly drives susceptibility and tissue damage during MCMV infection, despite expression of H-2Dk and presence of Ly49G2+ NK cells. The interval for which the M.H2k/b mice are heterozygous is denoted Cmv5, while the C57L-derived phenotype-associated haplotype is referred to as Cmv5s. The originally described Cmv5s interval encompassed the entire MHC region and many other genes that could be driving pathology. To determine the role of the MHC and associated sub-regions, we generated and tested several Cmv5-recombinant strains, substantially narrowing the phenotype-associated interval and cementing the contribution of at least 2 independent loci to Cmv5s pathology during infection. Cmv5s corrupts the ability of otherwise protective Ly49G2+ NK cells to control MCMV infection. To determine how this may be occurring, we asked at what stage of the immune response Cmv5s inhibits NK cell function. We found no differences in NK cell phenotype or function prior to infection, and no Cmv5s-driven defects in the early activation or expansion of the Ly49G2+ subset. By 4 days post infection (dpi), Cmv5s NK cells had increased expression of activation markers and checkpoint receptors, however there was no loss of functionality when stimulated in vitro. Marginal zone (MZ) histopathology consistent with necrosis was the first apparent Cmv5s phenotype, occurring prior to divergence in weight loss, viral load, or NK cell phenotype, and progressing to a widespread loss of spleen cellularity by 4dpi. Following this earliest Cmv5s-driven phenotype, we found increased neutrophil infiltration, death, and oxidative stress at the MZ, along with increased global IL-6 and TGF-β1 in Cmv5s spleens by 2dpi. Neutrophils were required for increased cell death and oxidative stress, but not necrotic MZ histopathology, leading us to ask about the state of other MZ populations at this timepoint. Marginal zone macrophages (MZMs), but not similarly localized stromal cells, were lost in both strains by 2dpi, and analysis at 36 hours post infection (hpi) revealed infection of SIGNR1+ MZMs and recruitment of neutrophils to areas of increased MCMV positivity and spread. Depletion confirmed that macrophages were required for MZ neutrophil recruitment and the differences between strains in spleen IL-6 and TGF-β1. By 4dpi, Cmv5s spleens showed buildup of MZM debris and substantial red pulp macrophage loss with widespread oxidative stress. These data identify macrophage loss and inflammation mirroring the spatial and temporal appearance of Cmv5s-driven tissue damage in the spleen during acute MCMV infection. Bringing our data together into a model of how Cmv5s may drive pathology during MCMV infection, we conclude that macrophage activation and loss underlie spleen histopathology, and are likely the primary targets of Cmv5s. This Cmv5s-modified tissue environment marked by increased cell death, oxidative stress, and inappropriate inflammation further serves to inhibit NK cell function, preventing effective viral control.