|Home | About | Journals | Submit | Contact Us | Français|
Herpes simplex virus 1 (HSV-1) enters mice via olfactory epithelial cells and then colonizes the trigeminal ganglia (TG). Most TG nerve endings are subepithelial, so this colonization implies subepithelial viral spread, where myeloid cells provide an important line of defense. The outcome of infection of myeloid cells by HSV-1 in vitro depends on their differentiation state; the outcome in vivo is unknown. Epithelial HSV-1 commonly infected myeloid cells, and Cre-Lox virus marking showed nose and lung infections passing through LysM-positive (LysM+) and CD11c+ cells. In contrast, subcapsular sinus macrophages (SSMs) exposed to lymph-borne HSV-1 were permissive only when type I interferon (IFN-I) signaling was blocked; normally, their infection was suppressed. Thus, the outcome of myeloid cell infection helped to determine the HSV-1 distribution: subepithelial myeloid cells provided a route of spread from the olfactory epithelium to TG neurons, while SSMs blocked systemic spread.
IMPORTANCE Herpes simplex virus 1 (HSV-1) infects most people and can cause severe disease. This reflects its persistence in nerve cells that connect to the mouth, nose, eye, and face. Established infection seems impossible to clear. Therefore, we must understand how it starts. This is difficult in humans, but mice show HSV-1 entry via the nose and then spread to its preferred nerve cells. We show that this spread proceeds in part via myeloid cells, which normally function in host defense. Myeloid infection was productive in some settings but was efficiently suppressed by interferon in others. Therefore, interferon acting on myeloid cells can stop HSV-1 spread, and enhancing this defense offers a way to improve infection control.
The alpha-, beta-, and gammaherpesviruses establish broadly neuro-, myelo-, and lymphotropic persistent infections (1). Less is known about acute infection, as sporadic transmission and late clinical presentation make it difficult to analyze. Acutely adaptive immunity exerts little restraint on viral tropism, so common themes are likely. The difficulty in clearing established infections makes these themes important to understand. Genomic comparisons indicate that herpesvirus infections long predate human speciation (2). Therefore, related mammalian herpesviruses are likely to share mechanisms of host colonization, allowing those of experimentally tractable hosts to provide new insights. Murid herpesviruses have particular value in this regard, as their hosts provide the main in vivo experimental model of mammalian biology.
Murid herpesvirus 4 (MuHV-4) (a gammaherpesvirus), murine cytomegalovirus (MCMV) (a betaherpesvirus), and herpes simplex virus 1 (HSV-1) (an alphaherpesvirus) all enter mice via olfactory neurons (3,–5). MuHV-4 and MCMV spread from there to lymph nodes (LNs) (4, 6), while HSV-1 spreads to trigeminal ganglia (TG) (5). Nonetheless, each virus penetrates the epithelium and so will encounter subepithelial myeloid cells. While these cells normally provide an early defense against invading pathogens, MCMV exploits them to spread (7) and persist (8), and MuHV-4 exploits them to reach B cells (9). How HSV-1 interacts with myeloid cells is less well understood.
Ex vivo human blood-derived monocytes resist productive HSV-1 infection but become susceptible after culture (10). Murine macrophages are similar (11, 12). Human monocyte-derived dendritic cells (DCs) support productive infection when they are immature and lose susceptibility with maturation (13). Again, murine DCs appear to be similar (14). MCMV (8) and human cytomegalovirus (HCMV) (15) establish latent infections of myeloid cells that are reactivated by maturation signals (8). MuHV-4 also establishes latency in myeloid cells (16) but with a strong tendency toward lytic reactivation. It inhibits myeloid cell functions extensively when lytic and minimally when latent (17). HSV-1 also impairs myeloid cell functions (18), causing host shutoff even when infection is abortive (19). Herpesvirus infections remain immunogenic because uninfected cells can engage in cross-priming. Therefore, the purpose of viral evasion in infected myeloid cells is probably to delay their recognition (20, 21). For MCMV and MuHV-4, this makes sense, as they use infection of myeloid cells to reach other cell types. The relevance for HSV-1 is less clear.
Myeloid cell depletions increase murine susceptibility to HSV-1-induced disease (22, 23), presumably because uninfected myeloid cells protect via immune priming and type I interferon (IFN-I) production (24,–26). Infected myeloid cells might also promote antiviral responses. However, how in vitro myeloid cell phenotypes relate to those encountered in vivo is difficult to know. A fundamental question is whether in vivo myeloid cell infection is productive. Key contexts are when incoming virions first encounter subepithelial myeloid cells and when infection spreads to the myeloid sentinels of LNs. We show by Cre-mediated genetic marking that HSV-1 can pass productively through subepithelial myeloid cells of infected mice. LN myeloid cells contrastingly restricted infection, unless IFN-I signaling was blocked.
C57BL/6J (Animal Resources Centre, Perth, Australia, or Harlan Ltd., Oxford, United Kingdom), CD11c-cre (27), and LysM-cre (28) mice were maintained at University of Queensland or University of Cambridge animal units and infected when they were 6 to 12 weeks old. Experiments were approved by the University of Queensland Animal Ethics Committee in accordance with Australian National Health and Medical Research Council guidelines (project 301/13) and by the University of Cambridge ethical review board and the United Kingdom Home Office under the 1986 Animal (Scientific Procedures) Act (project 80/2538). For nasal infections, virus (106 PFU in 5 μl) was pipetted onto the nares of mice held prone under light restraint without anesthesia and was spontaneously inhaled (29). For lung infections, mice were anesthetized with isoflurane, and virus (106 PFU) was inhaled in 30 μl. For whisker pad infections, mice were anesthetized with isoflurane, virus (106 PFU in 20 μl) was applied to each whisker pad, and 20 scratches were made through each drop with a 27-gauge needle. Ear pinna infections were performed similarly by scarification under anesthesia, applying virus (106 PFU in 20 μl) to the left ear pinna and making 20 scratches through the drop. Footpad infections (106 PFU in 50 μl) were done by injection under isoflurane anesthesia. To deplete NK cells, mice were injected intraperitoneally (i.p.) with 200 μg purified monoclonal antibody (MAb) PK136 (anti-NK1.1; Bio-X-Cell) 1 and 3 days before infection and every 2 days thereafter. Cell depletion was >90% effective, as measured by flow cytometry of spleen cells with an antibody to NKp46. To block IFN-I responses, we gave mice 200 μg purified MAb MAR1-5A3 (Bio-X-Cell) i.p. 1 day before infection and every 2 days thereafter. This MAb binds to the IFN-I receptor (IFNAR) and prevents IFN-I binding. Experimental groups were compared statistically by Student's two-tailed unpaired t test.
Macrophages were recovered by postmortem peritoneal lavage, followed by removal of nonadherent cells. They were >90% F4/80 positive (F4/80+) by immunostaining. Embryonic fibroblasts were harvested from day 13 to 14 mouse embryos by trypsin digestion and gentle tissue grinding. These cells, BHK-21 fibroblasts (ATCC CCL-10), NIH 3T3-cre fibroblasts (30), RAW-264 monocytes/macrophages (ATCC TIB-71), K562 myeloid leukemia cells (ATCC CCL-243), THP-1 monocytes (ATCC TIB-202), and U937 histiocytic lymphoma cells (ATCC CRL-1593) were grown in Dulbecco's modified Eagle's medium supplemented with 2 mM glutamine, 100 IU/ml penicillin, 100 μg/ml streptomycin, and 10% fetal calf serum (complete medium).
MHV-RG is a derivative of MuHV-4 with a viral M3 (lytic) promoter between the 3′ ends of open reading frames (ORFs) 57 and 58 driving LoxP-flanked mCherry upstream of green fluorescent protein (GFP) (9). MHV-RG expresses mCherry, but LoxP site recombination by Cre excises the mCherry coding sequence, switching the virus irreversibly to GFP expression from the same promoter (MHV-G). We used HSV-1 strain SC16 (31). The HSV-GFP derivative has an HCMV IE1 promoter transcribing GFP from the US5 locus (5). To make HSV-RG, the LoxP-mCherry-poly(A)-LoxP-GFP construct of MHV-RG was amplified with Pfu polymerase (Promega Corporation), adding HindIII and BamHI restriction sites to its respective 5′ and 3′ ends. The PCR product was cloned into the same sites of pcDNA3 (Invitrogen Corporation) and then subcloned into pHD5-CRE (32) using an SpeI restriction site in the HCMV IE1 promoter and XhoI sites in the pcDNA3 polylinker and downstream of the Cre coding sequence in pHD5-CRE. Thus, HCMV IE1-LoxP-mCherry-poly(A)-LoxP-GFP-poly(A) was inserted into US5 (genomic site 137945; GenBank accession number X14112). The plasmid was linearized with ScaI and cotransfected with HSV-1 SC16 viral DNA into BHK-21 cells by using Fugene-6 (Roche Diagnostics). mCherry+ virus was identified under UV illumination, enriched by flow cytometric sorting of infected cells, and plaque purified by limiting dilution. We derived switched HSV-RG (HSV-G) by passage in NIH 3T3-cre cells and limiting-dilution cloning in BHK-21 cells. All viruses were checked by sequencing across the US5 insertion site and by restriction enzyme mapping of viral DNA. Virus stocks were grown in BHK-21 cells (5). Virus was recovered from infected cells and supernatants by ultracentrifugation (38,000 × g for 90 min). The pelleted cells were sonicated to break up aggregates and then stored in aliquots at −80°C.
Infectivity titers in virus stocks, cells, and organ homogenates were determined by a plaque assay. Virus dilutions were incubated with BHK-21 cell monolayers (37°C for 2 h), overlaid with complete medium plus 0.3% carboxymethylcellulose, and cultured at 37°C. After 2 days (HSV-1) or 4 days (MuHV-4), the monolayers were fixed in 4% formaldehyde and stained with 0.1% toluidine blue. Plaques were counted under a microscope at a ×30 magnification. To measure both preformed infectious and reactivatable MuHV-4 titers, freshly isolated popliteal LN (PLN) or spleen cell suspensions were cocultured with BHK-21 cells for 4 days, and plaques were detected as described above. To measure both preformed infectious and reactivatable HSV-1 titers, TG were disrupted gently and then incubated (37°C for 30 min) with Liberase TL (2 Wünsch units [WU]/ml) and DNase I (0.2 mg/ml) (Roche Diagnostics). The released cells were plated onto BHK-21 cell monolayers and cultured for 2 days before fixation, staining, and plaque counting. Viral fluorochrome switching was determined by a plaque or infectious-center assay at a limiting dilution in 96-well plates (16 wells per dilution). After 2 days (HSV-1) or 4 days (MuHV-4), wells were scored under UV illumination for green (GFP+, switched) and red (mCherry+, unswitched) fluorescence to derive titers for each form of the virus. We calculated percent switching as 100 × green titer/(red titer + green titer).
Organs were fixed in 1% formaldehyde–10 mM sodium periodate–75 mM l-lysine (18 h at 4°C). Noses were decalcified by gentle agitation in 150 mM NaCl–50 mM Tris-Cl (pH 7.2)–270 mM EDTA for 2 weeks at 23°C, changing the solution every 2 to 3 days. All tissues were then equilibrated in 30% sucrose (24 h at 4°C) and frozen in OCT. Sections (6 μm) were air dried (1 h at 23°C); washed three times in phosphate-buffered saline (PBS); blocked with 0.3% Triton X-100–5% normal donkey serum (1 h at 23°C); and then incubated (18 h at 4°C) with combinations of primary antibodies to GFP (rabbit polyclonal antibody [PAb] or goat PAb; Abcam), B220 (rat MAb RA3-6B2; Santa Cruz Biotechnology), CD68 (rat MAb FA-11; Abcam), α-tubulin (rat MAb YL1/2; Serotec), βIII-tubulin (mouse MAb TU-20; Abcam), CD31 (rat MAb ER-MP12; Serotec), F4/80 (rat MAb CI:A3-1; Santa Cruz Biotechnology), mCherry (rabbit PAb; Badrilla), CD169 (rat MAb 3D6.112; Serotec), and polyclonal rabbit sera to MuHV-4 (raised in-house by subcutaneous virus inoculation three times) and HSV-1 (rabbit PAb, either obtained from Sigma Chemical Co. or raised in-house by immunizing rabbits subcutaneously three times with HSV-1 SC16). After incubation, sections were washed three times in PBS; incubated (1 h at 23°C) with combinations of Alexa 568- or Alexa 647-conjugated donkey anti-rat IgG PAb, Alexa 488- or Alexa 568-conjugated donkey anti-rabbit IgG PAb (Life Technologies), and Alexa 488-conjugated donkey anti-goat PAb (Abcam); and then washed three times in PBS, counterstained with 4′,6-diamidino-2-phenylindole (DAPI), and mounted in Prolong Gold (Life Technologies). Fluorescence was visualized with a Zeiss LSM 510/710 or a Leica TCS SP2 confocal microscope or a Nikon epifluorescence microscope and analyzed with Zen imaging software or ImageJ.
Cells were seeded onto glass coverslips and then infected, fixed 18 h later in 2% paraformaldehyde–PBS, permeabilized in 0.1% Triton X-100, blocked with 5% goat serum, and incubated with rabbit anti-HSV-1 PAb followed by Alexa 488-conjugated goat anti-rabbit PAb (Invitrogen). Cellular actin was stained with tetramethyl rhodamine isocyanate (TRITC)-conjugated phalloidin (Sigma Chemical Co.). Nuclei were stained with DAPI. Cells were mounted in Prolong Gold (Invitrogen) and imaged on a Leica TCS SP2 confocal microscope.
Fibroblasts were trypsinized, washed in PBS, and analyzed on a FACSCalibur instrument (BD Biosciences). mCherry and GFP fluorescence was visualized directly. To identify NK cells, dissociated spleen cells were blocked with anti-CD16/32 (BD Biosciences), incubated with biotinylated anti-NKp46 MAb (BioLegend) and then with Alexa 488-conjugated streptavidin (Invitrogen), washed twice with PBS, and analyzed on a FACSCalibur instrument (BD Biosciences).
Cells were lysed (4°C for 30 min) in a solution containing 1% Triton X-100, 50 mM Tris-Cl (pH 7.4), and 150 mM NaCl with Complete protease inhibitors (Roche Diagnostics). Cell debris and nuclei were removed by centrifugation (13,000 × g for 15 min). Lysates were heated to 70°C in Laemmli buffer, followed by SDS-PAGE and electrophoretic transfer onto nitrocellulose membranes. Blots were probed with mouse MAbs CB24 to gB (33) and LP1 to VP16 (34) and developed with rabbit anti-mouse IgG PAb and Li-Cor imaging.
Most experimental HSV-1 infections are initiated by scarification. Natural infection is more likely to occur at an intact mucosal surface. HSV-1 fails to infect nonscarified mice orally but infects them nasally via olfactory neurons (5). Nasal infection showed extensive subepithelial spread (Fig. 1a). Epithelial infection was always present, and early infection was solely epithelial (5), but subepithelial infection evidently spread faster. Myeloid cells (CD68+) were abundantly recruited to subepithelial infection sites, whether from primary olfactory infection or from secondary spread to the respiratory epithelium (Fig. 1b). Many of these infiltrating cells expressed viral lytic antigens (Fig. 1c). Therefore, HSV-1 commonly infected subepithelial myeloid cells after host entry at a mucosal site.
HSV-1 strain SC16, a low-passage-number isolate derived in the 1970s, was used to establish antiviral chemotherapy (35) and a glycoprotein H-deficient vaccine (36). Its tropism for myeloid cells has not been tested. This strain replicated in RAW-264 monocytes/macrophages (Fig. 2a). Productive infection was validated by immunoblotting for virion gB and VP16 (Fig. 2b). However, RAW-264 cells produced fewer infectious virions than did BHK-21 fibroblasts (Fig. 2a), and after overnight infection (3 PFU/cell), 16.8% of RAW-264 cells expressed viral lytic antigens (Fig. 2c), whereas >99% of BHK-21 cells did so (data not shown). Three human myeloid cell lines, K562, THP-1, and U937, supported productive infection even less well than did RAW-264 cells (Fig. 2a). Therefore, SC16 was similar to other HSV-1 strains in showing a modest capacity to replicate in myeloid cells in vitro.
Viral lytic gene expression in CD68+ cells (Fig. 1) suggested productive myeloid infection. To track this functionally, we generated a floxed reporter virus, HSV-RG, by inserting a cassette comprising an HCMV IE1 promoter, a floxed mCherry coding sequence plus poly(A) site, and then a GFP coding sequence plus poly(A) site into the nonessential US5 locus (37) (Fig. 3a). HSV-RG expressed mCherry (red fluorescence) until mCherry excision by Cre irreversibly switched its fluorochrome expression to GFP (green) (Fig. 3b). Unswitched HSV-RG and switched HSV-G showed no difference in replication in Cre+ or Cre-negative (Cre−) cells in vitro (Fig. 3c). Both viruses showed minor in vivo attenuation after nasal inoculation relative to the parental HSV-1 SC16 wild-type strain, presumably due to US5 disruption, but no defect relative to each other (Fig. 3d). Therefore, HSV-RG provided a tool capable of unbiased virus tracking through Cre+ cells.
A cellular path connects each recovered virion to host entry. Virus tagging tells us what proportion of productive paths traversed a Cre+ cell. LysM-cre mice express Cre mainly in neutrophils, mature macrophages (28, 38, 39), and type 2 alveolar epithelial cells (40). HSV-RG accordingly showed fluorochrome switching in peritoneal macrophages but not embryonic fibroblasts of LysM-cre mice (Fig. 4a). We compared lung infection by HSV-RG with that by MuHV-4 carrying a similar switching cassette (MHV-RG) (Fig. 4b and andc).c). Gr1+ neutrophils and inflammatory monocyte cells entering the lungs do not express LysM (40), and neither MuHV-4 nor HSV-1 infects type 2 alveolar epithelial cells (41). Thus, at least acutely, viral fluorochrome switching could be interpreted as replication in alveolar macrophages. MuHV-4 enters the lungs via alveolar macrophages (40), and MHV-RG accordingly showed substantial switching after 1 day. HSV-1 infects mainly type 1 alveolar epithelial cells (40) but also showed substantial switching at days 1 and 2 postinoculation.
Herpesviruses given intranasally (i.n.) also infect the upper respiratory tract (29). Therefore, we also assayed the fluorochrome expression of HSV-RG and MHV-RG recovered from noses (Fig. 4c). Upper respiratory tract infection proceeds more slowly than lung infection, so we sampled the mice at day 3. HSV-RG and MHV-RG were both less switched in noses than in lungs, so fewer virions followed paths through LysM+ cells, but switching was detectable nonetheless.
Nasal HSV-1 spreads to TG after 2 to 3 days and reemerges in the facial skin after 4 to 5 days (5). Thus, virions in the TG should be at least as switched as those in noses, and those in the skin should be at least as switched as those in the TG. After lung and nose infections (large inoculation volume with anesthesia), the HSV-RG recovered from TG or skin at day 4 was similarly switched to that from noses. Selective upper respiratory tract infection (low-volume inoculation without anesthesia) also showed no significant differences in switching between HSV-RG from noses, TG, and skin (Fig. 4e). Therefore, productive myeloid cell infection occurred early, en route from the olfactory epithelium to the TG. Immunostaining of tissue sections identified mCherry+ and GFP+ infected cells in the TG (Fig. 4f) and in the superficial layers of the skin (Fig. 4g). Thus, virus reemerging from the TG possibly avoids exposure to LysM+ cells because it reemerges in the epidermis (5).
Inoculation of the whisker pad or the ear pinna by scarification, routes commonly used for experimental HSV-1 infection, gave less switching (Fig. 4h). This possibly reflected that scarification provides direct access to subepithelial nerve endings, bypassing the normal myeloid cell defenses of intact epithelia.
Myeloid cells are highly diverse. No single promoter identifies them all or defines exclusive subpopulations (42). Thus, to back up the results with LysM-cre mice, we tracked HSV-RG fluorochrome switching in CD11c-cre mice. Immunostaining shows CD11c expression in DCs and some macrophage populations, including LN subcapsular sinus macrophages (SSMs) (43). Few DCs express LysM (28). Thus, CD11c and LysM expressions identify partly overlapping populations, with CD11c-cre mice measuring HSV-1 passage through more DC-type myeloid cells.
As in LysM-cre mice, HSV-RG was more switched in CD11c-cre lungs than in noses, although switching was detectable at both sites (Fig. 5a). A direct comparison of i.n. infections of CD11c-cre and LysM-cre mice at day 4 (Fig. 5b) showed somewhat more switching in CD11c-cre mice for both noses and TG. Each transgenic mouse showed more switching in noses than in TG. This was statistically significant for CD11c-cre mice but not for LysM-cre mice. Analysis of larger numbers of i.n. infected CD11c-cre mice at day 5 (Fig. 5c) confirmed greater switching in noses than in TG.
Viral fluorochrome switching is irreversible and so should increase cumulatively between infection sites. Thus, the decrease in HSV-RG switching from noses to TG indicated that although CD11c+ cells generated new virions, they passed infection to neurons less well than did CD11c− cells. Switched and unswitched viruses had equal fitness, so this result suggested that replication in CD11c+ cells carried an extra cost, for example, due to innate immune stimulation. CD11c+ cells readily pass MuHV-4 to LNs (6, 9). In contrast, HSV-1 lung and nose infections gave <50 PFU per LN (data not shown). HSV-1 may lack the capacity to exploit DC migration. However, virions should still reach LNs via the lymph. Therefore, we considered that CD11c+ cell infection might inhibit HSV-1 propagation by local immune activation, for example, by eliciting IFN-I, which has anti-HSV-1 activity in both humans and mice (44,–46). The LN subcapsular sinus is a prominent site of IFN-I responses (47), and herpesvirus virions inoculated into footpads (intrafootpad [i.f.]) directly reach CD11c+ SSMs (39, 43). Therefore, to test in vivo how IFN-I affected HSV-1 myeloid cell infection, we gave mice anti-IFNAR antibody i.p. or not and then gave mice GFP+ HSV-1 i.f. (Fig. 6).
At 1 day postinoculation, IFNAR blockade had no significant effect on HSV-1 titers in footpads but increased titers substantially in PLNs (Fig. 6a). By day 3, IFNAR blockade had increased footpad titers, PLN titers remained elevated, and infection had spread to the liver and spleen (Fig. 6b), implying passage from the PLNs to the blood. Immunostaining of PLN sections at day 1 (Fig. 6c and andd)d) showed significantly more viral GFP+ and viral antigen-positive cells around the subcapsular sinus after IFNAR blockade. Both viral markers colocalized with CD68 and CD169, indicating SSM infection. The few infected cells of control mice also included examples of colocalization with CD68 and CD169. By day 3, a substantial inflammatory infiltrate into the PLNs of IFNAR-blocked mice was evident by CD68+ staining (compare Fig. 6e to tod).d). CD169 expression was largely lost, but the expression level of the tissue macrophage marker F4/80, which SSMs lack (48), was increased, and both CD68+ and F4/80+ cells were HSV-1 antigen positive. B220+ B cells and CD31+ vascular endothelial cells showed no infection. Therefore, IFNAR blockade increased HSV-1 infection specifically in SSMs and other myeloid cells.
One action of IFN-I at the subcapsular sinus is NK cell recruitment (49). To test whether this could account for the protection of SSMs against HSV-1 by IFN-I, we compared IFNAR blockade with NK1.1+ cell depletion in C57BL/6 mice. NK cell depletion significantly increased day 1 PLN virus titers (Fig. 6f). However, IFNAR blockade increased them more, and while IFNAR blockade increased viral GFP+ cell numbers on PLN sections, NK cell depletion did not have a significant effect. Therefore, NK cells contributed to anti-HSV-1 defense at the subcapsular sinus but could not account for most IFN-I-dependent protection. The strong antiviral efficacy of IFN-I at day 1, with inhibition of both viral reporter and lytic gene expression, suggested that it acted directly on SSMs to block infection at a very early stage.
We next tested whether IFNAR blockade increased HSV-1 production by myeloid cells, as measured by fluorochrome switching (Fig. 7). IFNAR blockade increased day 3 lung virus titers in LysM-cre mice (Fig. 7a). However, the recovered virus showed no increase in switching. As virus titers were higher, more switched virus was produced, but IFNAR blockade evidently also increased LysM− cell virus production. Total virus titers also increased in LysM-cre noses and footpads without increasing the fraction switched (Fig. 7b and andcc).
In TG (Fig. 7b), both total titers and switching increased. Therefore, IFN-I limited macrophage-dependent passage to the TG more than macrophage-independent passage, although the proportion of virus that was switched remained low, so passage through LysM+ cells to the TG was still an accessory route. In contrast, LN virus showed abundant switching in IFNAR-blocked mice, comparable to that in lungs. The level of LN infection was too low to assess switching in control mice, but the substantial rise in virus titers and the high level of switching after IFNAR blockade implied copious virus production by LysM+ cells, most likely SSMs (Fig. 6).
IFNAR blockade in CD11c-cre mice gave similar results: it increased HSV-RG titers in lungs, noses, and footpads without significantly increasing switching; it increased both titers and switching in TG; and it increased LN titers with abundant switching. Thus, IFN-I regulated HSV-1 spread to a degree determined by myeloid cell involvement. In LNs, it protected SSMs and so blocked viremic spread. In subepithelial tissues, it also moderated myeloid infection but did not prevent myeloid cell-independent virus passage from the olfactory epithelium to the TG. Figure 8 outlines our understanding of how myeloid cell infection fits into the HSV-1 life cycle.
Sentinel macrophages and DCs monitor tissues for normal senescence and for pathogen invasion. They are particularly numerous below epithelial surfaces and where extracellular fluid enters LNs. Thus, despite the anatomical restriction of HSV-1 persistence to local neuronal ganglia, subepithelial spread after mucosal entry led it to myeloid cells. Comparison with other herpesviruses that enter via the olfactory epithelium (3, 4) reveals myeloid infection as a common theme, providing access to diverse latency reservoirs.
The different outcomes of HSV-1, MuHV-4, and MCMV myeloid infections can be explained in part by the tendency of each virus to initiate lytic or latent gene expression. MCMV must remain latent in monocytes to reach secondary infection sites such as the salivary glands, MuHV-4 must remain latent in DCs until they contact B cells, and each virus reactivates presumably in response to microenvironmental signals reaching the myeloid cell nucleus. HSV-1 has a superficially simple host colonization strategy of replicating lytically until it enters a neuron. Its tendency for lytic replication and its capacity to infect many cell types make innate immune defenses key to preventing acute disease. IFNAR blockade greatly increased virus titers, consistent with what is observed in IFNAR−/− mice (45). The fluorochrome switching of TG virus in LysM-cre and CD11c-cre mice indicated that myeloid cells intercept some of the HSV-1 penetrating the olfactory epithelium and through IFN-I hinder its spread to trigeminal neurons, but the key role of IFN-I was in LNs, where its protection of SSMs prevented systemic infection.
SSMs do not form a physical barrier to lymph-borne virus spread, as they merely stud the subcapsular sinus wall (50). Rather, they adsorb viruses from the lymph (51). This sampling allows SSMs to initiate early innate and adaptive immune responses. Cumulative virion adsorption along the tortuous lymphatic channels of serial LNs also stops lymph-borne virions from reaching the blood. Lymph cleansing depends on SSMs not supporting the replication of viruses that they adsorb or at least slowing their replication sufficiently for immune responses to become effective. The importance here of IFN-I was evident from the IFNAR blockade allowing HSV-1 to replicate in SSMs and to reach the liver and spleen, consistent with viremic spread.
SSMs also limit the spread of MuHV-4 and MCMV (39, 43). MuHV-4 bypasses this restriction by entering LNs in DCs. The route that MCMV takes is yet to be defined, but it is clearly more permissive than SSMs. HSV-1 was able to pass through upper and lower respiratory tract myeloid cells but not through SSMs. As vesicular stomatitis virus (VSV) can replicate in SSMs (51), it is unclear why herpesviruses have not evolved to do so. HSV-1 IFN-I evasion (52) may be more complete in humans than in mice. However, the restriction of clinical HSV-1 lesions to a trigeminal distribution argues that human LNs are also an effective barrier to spread. As HSV-1 still reached neurons when IFN-I signaling was intact, there may be limited selective pressure for more complete evasion. TG infection increased without IFN-I, but viral evolution is driven by transmission efficiency; whether greater initial HSV-1 delivery to neurons increases long-term shedding is uncertain. Viral reemergence from TG neurons is more directly relevant to transmission. Thus, it was of interest that HSV-1 passage from TG to skin seemed to avoid LysM+ cells, perhaps because most skin infection is epidermal (5). Excessive IFN-I evasion may have downsides: IFN-I contributes to the homeostasis of immune cells (53) and possibly also neurons (54), so a complete blockade might compromise persistence in these cell types. Such compromises forced on persistent viruses provide potential means of improving infection control.
We thank Stacey Efstathiou and laboratory members for advice, Colin Crump for MAb CB24, and Susanne Bell for MAb LP1.
M.S. was supported by a Medical Research Council UK studentship. The work was also supported by the BBSRC (BB/J014419/1), the Australian Research Council (FT130100138), the Australian National Health and Medical Research Council (project grants 1064015, 1060138, and 1079180), and Belspo (collaborative grant BelVir).