Significance: Strains of human cytomegalovirus (HCMV) that have been passaged in the laboratory fail to recapitulate the phenotypic characteristics of ex vivo isolates due to the rapid acquisition of genetic mutations. In particular, clinical viruses spread via direct cell–cell contact, whereas passaged strains spread via diffusion of cell-free virions. We developed a method that enables in vitro propagation of genetically intact HCMV. Using this system, we report that cell–cell transmission of wild-type HCMV not only is highly efficient, but also enables the virus to overcome key innate and adaptive immune defenses. These findings inform our understanding of viral persistence and validate a new approach to experimentation that may facilitate the design of improved prophylactic and therapeutic interventions against HCMV.
Bovine herpes virus-4 (BoHV-4) and bovine viral diarrhoea virus (BVDV) infect the uterus of cattle, often resulting in reduced fertility or abortion of the foetus, respectively. Here, exposure of primary bovine endometrial cells to BoHV-4 or BVDV modulated the production of inflammatory mediators. Viral pathogen-associated molecular patterns are detected via pattern-recognition receptors (PRRs). However, the relative contribution of specific PRRs to innate immunity, during viral infection of the uterus, is unclear. Endometrial epithelial and stromal cells constitutively express the PRR Toll-like receptor (TLR)-3, but, the status of RIG-I (retinoic acid-inducible gene I), a sensor of cytosolic nucleic acids, is unknown. Primary endometrial epithelial and stromal cells had low expression of RIG-I, which was increased in stromal cells after 12 h transfection with the TLR3 ligand Poly(I:C), a synthetic analogue of double-stranded (ds)RNA. Furthermore, short interfering RNA targeting TLR3, or IRF3 (interferon regulatory transcription factor 3), an inducer of type I interferon transcription, reduced Poly(I:C)-induced RIG-I protein expression and reduced inflammatory mediator secretion from stromal cells. We conclude that antiviral defence of endometrial stromal cells requires coordinated recognition of pathogen-associated molecular patterns, initially via TLR3 and later via inducible RIG-I.