NK cells force cytomegalovirus to use hematopoietic cells and immune evasion for dissemination after mucosal infection

Cytomegalovirus (CMV) infects most people in the world and causes clinically important disease in immune compromised and immune immature individuals. How the virus disseminates from the initial site of infection is poorly understood. We used an innovative approach, involving insertion of target sites for the haematopoietic specific miRNA miR-142-3p into an essential viral gene in murine cytomegalovirus. This virus was unable to disseminate to the salivary gland following intranasal infection, demonstrating a strict need for hematopoietic cells for dissemination from the natural site of infection. Viral immune evasion genes that modulate MHC-I expression and NKG2D activation were also required in this setting, as MCMV lacking these genes exhibited impaired dissemination of the viral genome to the salivary gland, and there was no detectable viral replication in the salivary gland. Depletion of T cells rescued the replication of this evasion-deficient virus in the salivary gland. Surprisingly however, the early dissemination to the salivary gland of this evasion-deficient virus, could be rescued by depletion of NK cells, but not T cells. These data are the first to show a profound loss of MCMV fitness in the absence of its MHC-I evasion genes and suggest that they protect the virus from NK cells during hematopoietic dissemination to the salivary gland, where they continued to need the three evasion genes to avoid T cell responses. Remarkably, we found that depletion of NK cells also freed the virus from the need to infect hematopoietic cells in order to reach the salivary gland. Thus, our data show that MCMV adapts to NK cell pressure after intranasal infection by using hematopoietic cells for dissemination while immune evasion genes protect the virus from NK cells during dissemination and from T cells within mucosal tissues.


Introduction 51
Cytomegalovirus (CMV), is the most common infectious cause of birth defects in the 52 developed world, leading to hearing loss, vision impairment and cognitive/motor deficits 53 and is estimated to affect 0.5% to 5% of children globally [1][2][3]. The greatest risk for the 54 most devastating outcomes of congenital CMV infection occur when a pregnant mother 55 Recent work has shown that the nasal mucosa is a natural site of MCMV entry [4]. Thus, 165 we infected C57BL/6 mice by the intranasal route with either MCMV-IE3-142 or control 166 viruses (either parental wild-type BAC MCMV, or MCMV-IE3-015). All three viruses 167 replicated at the sites of entry (nasal mucosa and lungs). However, two weeks after 168 infection, only control viruses were found to be replicating in the SG ( Figure 1D). Similar 169 results were obtained after footpad inoculation, which is considered reflective of infection 170 via licking skin abrasions or biting, another possible natural route of infection for MCMV 171 ( Figure 1E). In contrast, infection by the i.p. route, which allows hematogenous spread of 172 cell-free virus [32], enabled MCMV-IE3-142 to replicate robustly in the SG ( Figure 1E). 173 Taken together, these data show that MCMV must utilize infected hematopoietic cells to 174 efficiently spread to SG after intranasal infection 175 176

Evasion of MHC-I antigen-presentation and CD8 + T cells is critical for viral 177 persistence at sites of entry and replication in the salivary gland 178
Since MCMV had to infect hematopoietic cells to spread to SG after i.n. inoculation, we 179 considered whether these infected hematopoietic cells must evade immune control in  Consistent with a role for CD4 + T cell help for CD8 + T cells, CD4 + T cell depletion prior 256 to i.n. infection significantly reduced the frequency and number of MCMV-specific CD8 + 257 T cells in the blood ( Figure 2D & Figure S1A). Representative gating strategies for these 258 and subsequent data is shown in Figure S2. For precise quantitation of CD8 + T cell 259 function per cell we used OT-I T cells stimulated by i.n. infection with MCMV-Ova. 260 Depletion of CD4 + T cells prior to infection resulted in impaired cytokine production and 261 degranulation of OT-Is ( Figure S1B). In contrast, delaying depletion of CD4 + T cells 262 until day 7 after infection significantly increased the frequency and number of functional 263 CD8 + T cells ( Figure 2E). Thus, CD4 + T cell help was critical for the development of 264 functional CD8 + T cell responses after i.n. inoculation, which were able to completely 265 prevent TKO-MCMV from replicating in the salivary gland. 266 267

Early dissemination of TKO-MCMV is restored by NK cell depletion 268
It was possible that MHC-I evasion genes protected MCMV during dissemination to the 269 SG or after it arrived. If MHC-I evasion genes were required during dissemination, we 270 reasoned that we would detect reduced quantities of TKO-MCMV DNA in the SG, which 271 should be rescued by T cell depletion. To specifically assess viral dissemination rather 272 than replication after dissemination, we assessed viral DNA load in the SG 4 days after 273 i.n. inoculation, a time point at which virus-specific CD8 + T cells can be detected in 274 draining lymph nodes, but not yet in the SG ( Figure 3A). Indeed, the TKO-MCMV DNA 275 load was approximately 10-fold reduced compared to WT-MCMV in unmanipulated 276 C57BL/6 mice ( Figure 3B). However, when CD8 + T cells, CD4 + T cells or both CD4 + 277 and CD8 + T cells were depleted, TKO DNA load was only marginally increased (about 278 2-fold) and this did not reach significance ( Figure 3B). Moreover, TKO-MCMV DNA 279 was present in similar amounts in draining lymph nodes (mandibular LNs, deep cervical 280 LNs and mediastinal LNs [40,41]) with or without CD8 + T cells ( Figure 3C). These data 281 show that dissemination of TKO-MCMV is markedly impaired, but suggest that T cells

Virus titration 553
Nasal mucosa, lungs and salivary glands (SG) were collected at indicated time points post 554 infection and frozen. Nasal mucosa was collected as previously described [40]. Twenty 555 percent homogenates (w/v) were prepared from each collected tissue for virus 556 quantification by plaque assay [40]. Briefly, tissues were weighed and homogenized 557 using a pestle with a small amount of sterile sand in a 1.5 ml centrifuge tube, then 558 suspended with RPMI supplemented with 10% FBS, 100 Units/mL penicillin, and 100 559 µg/mL streptomycin. Supernatants from the homogenate were collected after 560 centrifugation (2400 xg, 10 min) and viral plaque assay was performed on M2-10B4 cells. 561 562

Adoptive transfer of OT-I T cells 563
For adoptive transfer of OT-I T cells we used OT-I transgenic mice expressing CD45.1 as 564 donor cells. Splenocytes containing 5000 OT-I cells from naïve transgenic mice were 565 injected i.v. into sex-matched congenic recipients via the retro-orbital sinus suspended in 566 100 µl PBS. The following day, recipients were i.n. infected with 10 6 PFU MCMV-Ova. 567 568 Lymphocytes isolation, antibodies, tetramer staining, intracellular cytokine 569

DNA extraction and quantitative real-time PCR (qPCR) 592
For extracting DNA from the SG, 50 µl from a twenty percent homogenate (w/v) was 593 used. For mandibular lymph nodes (manLNs), deep cervical lymph nodes (DCLNs) and 594