Herpesviruses have coevolved with their hosts over hundreds of millions of years and exploit fundamental features of their biology. Cytomegaloviruses (CMVs) colonize blood-borne myeloid cells, and it has been hypothesized that systemic dissemination arises from infected stem cells in bone marrow. However, poor CMV transfer by stem cell transplantation argues against this being the main reservoir. To identify alternative pathways for CMV spread, we tracked murine CMV (MCMV) colonization after mucosal entry. We show that following intranasal MCMV infection, lung CD11c ⁺ dendritic cells (DC) migrated sequentially to lymph nodes (LN), blood, and then salivary glands. Replication-deficient virus followed the same route, and thus, DC infected peripherally traversed LN to enter the blood. Given that DC are thought to die locally following their arrival and integration into LN, recirculation into blood represents a new pathway. We examined host and viral factors that facilitated this LN traverse. We show that MCMV-infected DC exited LN by a distinct route to lymphocytes, entering high endothelial venules and bypassing the efferent lymph. LN exit required CD44 and the viral M33 chemokine receptor, without which infected DC accumulated in LN and systemic spread was greatly reduced. Taken together, our studies provide the first demonstration of virus-driven DC recirculation. As viruses follow host-defined pathways, high endothelial venules may normally allow DC to pass from LN back into blood. IMPORTANCE Human cytomegalovirus (HCMV) causes devastating disease in the unborn fetus and in the immunocompromised. There is no licensed vaccine, and preventive measures are impeded by our poor understanding of early events in host colonization. HCMV and murine CMV (MCMV) both infect blood-borne myeloid cells. HCMV-infected blood cells are thought to derive from infected bone marrow stem cells. However, infected stem cells have not been visualized in vivo nor shown to produce virus ex vivo , and hematopoietic transplants poorly transfer infection. We show that MCMV-infected dendritic cells in the lungs reach the blood via lymph nodes, surprisingly migrating into high endothelial venules. Dissemination did not require viral replication. It depended on the constitutively active viral chemokine receptor M33 and on the host hyaluronan receptor CD44. Thus, viral chemokine receptors are a possible target to limit systemic CMV infections.

Cytomegaloviruses (CMVs) infect the lungs and cause pathological damage there in immunocompromised hosts. How lung infection starts is unknown. Inhaled murine CMV (MCMV) directly infected alveolar macrophages (AMs) and type 2 alveolar epithelial cells (AEC2s) but not type 1 alveolar epithelial cells (AEC1s). In contrast, herpes simplex virus 1 infected AEC1s and murid herpesvirus 4 (MuHV-4) infected AEC1s via AMs. MCMV-infected AMs prominently expressed viral reporter genes from a human CMV IE1 promoter; but most IE1-positive cells were AEC2s, and CD11c-cre mice, which express cre in AMs, switched the fluorochrome expression of 90% of floxed MuHV-4 in the lungs and 50% of floxed MCMV in the blood. AM depletion increased MCMV titers in the lung during the acute phase of infection. Thus, the influence of AMs was more restrictive than permissive. Circulating monocytes entered infected lungs in large numbers and became infected, but not directly; infection occurred mainly via AEC2s. Mice infected with an MCMV mutant lacking its m131/m129 chemokine homolog, which promotes macrophage infection, showed levels of lung infection equivalent to those of wild-type MCMV-infected mice. The level of lung infiltration by Gr-1-positive cells infected with the MCMV m131/m129-null mutant was modestly different from that for wild-type MCMV-infected lungs. These results are consistent with myeloid cells mainly disseminating MCMV from the lungs, whereas AEC2s provide local amplification. IMPORTANCE Cytomegaloviruses (CMVs) chronically and systemically infect most mammals. Human CMV infection is usually asymptomatic but causes lung disease in people with poor immune function. As human infection is hard to analyze, studies with related animal viruses provide important insights. We show that murine CMV has two targets in the lungs: macrophages and surfactant-secreting epithelial cells. Acute virus replication occurred largely in epithelial cells. Macrophages had an important defensive role, as their removal increased the level of infection. These results establish the dual nature of lung infection, with local virus replication occurring in epithelial cells and spread occurring via quiescently infected macrophages. Distinct therapies may be needed to target these contrasting events.

Establishment of latent infection within specific tissues in the host is a common biological feature of the herpesviruses. In the case of bovine herpesvirus 2 (BoHV-2), latency is established in neuronal tissues, while bovine herpesvirus 4 (BoHV-4) and ovine herpesvirus 2 (OvHV-2) latent virus targets on cells of the monocytic lineage. This study was conducted in quest of BoHV-2, BoHV-4 and OvHV-2 DNA in two hundred trigeminal ganglia (TG) specimens, derived from one hundred clinically healthy cattle, majority of them naturally infected with bovine herpesvirus 1 (BoHV-1) and bovine herpesvirus 5 (BoHV-5). Total DNA extracted from ganglia was analyzed by polymerase chain reaction (PCR) designed to amplify part of the genes coding for BoHV-2, and BoHV-4 glycoprotein B and, for OvHV-2, the gene coding for phosphoribosylformylglycinamidine synthase-like protein. BoHV-2 DNA was detected in TG samples of two (2%) and BoHV-4 DNA in nine (9%) of the animals, whereas OvHV-2 DNA could not be detected in any of the TG DNA. The two animals in which BoHV-2 DNA was identified were also co-infected with BoHV-1 and BoHV-5. Within the nine animals in which BoHV-4 DNA was detected, six were also co-infected with BoHV-1 and BoHV-5. This report provides for the first time evidence that viral DNA from BoHV-2 and BoHV-4 can be occasionally detected in TG of naturally infected cattle. Likewise, in this report we provided for the first time evidence that the co-infection of cattle with three distinct bovine herpesviruses might be a naturally occurring phenomenon.