C57BL/6 mice were purchased from Jax and bred in house (Catalog #00664). Mice were maintained at the University of Colorado Anschutz Medical Campus in accordance with IACUC protocol and were used for experiments at 6–12 weeks of age. Both male and female mice were used.

Wild-type (WT) MHV68 and MHV68.LANA:βlac viruses were grown and prepared as described previously (Diebel et al., 2015). For in vitro infections, mouse 3T12 fibroblasts and ex vivo peritoneal cells were infected overnight, with no removal of inoculum, at a multiplicity of infection (MOI) of 1 and 10 plaque forming units (PFU) per cell, respectively. Cells were cultured in DMEM with L-glutamine, 1% Penicillin/Streptomycin and 10% FBS. For in vivo infections, mice were inoculated intraperitoneally (i.p.) with 1 × 10⁶ PFU of virus diluted in sterile PBS.

Mice were sacrificed via CO₂ euthanasia and cervical dislocation. The PerC was exposed and filled with 10 mL of ice cold PBS containing 3% FBS using a 27 gage needle. After removal of the needle, the mouse was then agitated to encourage release of immune cells from the peritoneal wall, followed by fluid and cell extraction using an 18 gage needle. If necessary, cells were subjected to red blood cell lysis with ACK lysis buffer (Gibco Catalog #A1049201), washed and counted using the Beckman Coulter Vi-CELL BLU.

Following harvest and preparation, peritoneal cells were subjected to flow cytometric analysis. First, cells were assessed for viability using a 30 min RT incubation with Zombie NIR (BioLegend) diluted in PBS. CCF2-AM substrate (LiveBLAzer^TM FRET-B/G Loading Kit with CCF2-AM, Thermo Fisher Scientific) (Zlokarnik et al., 1998) was freshly diluted prior to substrate loading for 30–60 min RT in PBS. Cells were subjected to Fc receptor blocking using Fc Shield (Tonbo Biosciences) for 10 min, followed by antibody staining for 25 min. The following antibodies, with antibody clones designated in parenthesis, were used: CD19(1D3)-PECy7, CD11b(M1/70)-APC (eBiosciences/Thermo Fisher Scientific), CD11c(HL3)-BUV737, CD4(GK1.5)-BUV395, CD8a(53-6.7)-PerCPCy5.5 (BD), CD3e(145-2C11)-PECy5, F4/80(BM8)-BV785, MHC II(M5/14.15.2) PerCP, Gr1(RB6-8C5)-AlexaFluor700, Gal3(M3/38)-AlexaFluor647 (BioLegend). All staining reactions were done at RT in the dark in PBS for antibody stains and PBS containing 1% FBS for Fc receptor blockade. Samples were acquired on a Cytek Aurora Spectral Analyzer with five lasers (UV, violet, blue, yellow-green, and red) using SpectroFlo software. Gating for specific cell populations, as well as for CCF2, was determined using control samples for each run. Differences in gating between time points reflect variations in experimental and spectral analyzer conditions on different days. Detailed information concerning unmixing of CCF2 and effects of fixation and permeabilization of CCF2 on fluorescence can be provided upon request.

C57BL/6 mouse embryonic fibroblasts (MEFs) were cultured in DMEM with L-glutamine, 20% FBS, 1% Penicillin/Streptomycin and 1× Amphotericin B at 5,000 cells per well in 96-well flat bottom tissue culture-treated plates. Peritoneal cells (PerCs) were harvested from mice inoculated with PBS alone, MHV68 or MHV68.LANA:βlac at 3 days post infection (dpi). PerCs were washed, counted, and diluted to the indicated concentrations. In parallel, cells were subjected to mechanical disruption and plated in a comparable series of dilutions prior to plating on MEF monolayers (Niemeyer et al., 2018). Co-cultures were incubated at 37°C for 21 days prior to assessment of cytopathic effect (CPE). A dashed line at 63.2% was used to determine the frequency of virus-positive cells, defined by Poisson distribution. Analysis was performed via non-linear regression analysis in GraphPad Prism, with calculation of 95% confidence intervals. Data represents three experiments with two mice per condition per experiment.

Analysis of flow cytometry data was performed using Cytek SpectroFlo version 2.1 and DeNovo FCS Express version 7. Fluorophore signatures with partial overlap of emission spectra were resolved by unmixing using SpectroFlo software without compensation. Statistical analysis was performed using student’s t-test in GraphPad Prism version 8.

We sought to harness the powerful technology of multiparametric flow cytometry with the beta lactamase reporter virus to track virally infected cells in vitro and in vivo. Previous use of a beta-lactamase expressing MHV68 virus detected virally infected cells utilizing conventional, band-pass flow cytometry (Coleman et al., 2010; Nealy et al., 2010; Diebel et al., 2015; Niemeyer et al., 2018). Due to the broad emission spectrum of the beta-lactamase substrate, this approach was limited in its utility for simultaneous measurement of multiple fluorescent markers. In this study, we utilized the power of full spectrum (herein referred to as spectral) flow cytometry. This technique characterizes fluorophores by their entire spectra, rather than peak emission, allowing for the detection of fluorophores relatively close in peak emission and thus vastly expanding the number of fluorophores that can be simultaneously analyzed. Full spectrum flow cytometry also allows for the detection of fluorescent reporters and proteins whose spectra are not well characterized or are unknown (Nolan and Condello, 2013; Nolan et al., 2013).

Studies with the MHV68.LANA:βlac reporter virus rely on detection of cleavage of a beta-lactamase substrate, such as CCF2-AM (Zlokarnik et al., 1998; Knapp et al., 2003), an esterified beta-lactamase substrate that readily crosses the cell membrane (Figure 1). Once inside the cell, endogenous cytoplasmic esterases cleave ester groups, resulting in the lipophobic molecule CCF2. CCF2 consists of a coumarin (identified as “A” in Figure 1A) linked to a fluorescein (identified as “B,” Figure 1A) via a cephalosporin group. The coumarin group absorbs light at a maximum wavelength of 408 nm and, via Fluorescence Resonance Energy Transfer (FRET), transfers the energy to the fluorescein group, which emits at a maximum wavelength of 520 nm. In the uncleaved state, CCF2 (Figure 1A, left) exhibits peak emission from the fluorescein group. In the presence of beta-lactamase enzymatic activity (Figure 1A, right), the cephalosporin group is cleaved to separate the coumarin and fluorescein groups, disrupting FRET and altering the resulting fluorescence to the coumarin group, which emits at a maximum wavelength of 447 nm. Thus, the presence of beta-lactamase (βlac) enzyme, such as the LANA:βlac fusion protein, can be detected by the unique fluorescence profile of the cleaved CCF2 substrate.

To examine the MHV68.LANA:βlac reporter system by spectral flow cytometry, we infected mouse 3T12 fibroblasts with either WT MHV68, which lacks βlac expression, or MHV68.LANA:βlac for 20 h at an MOI of 1 PFU/cell, and incubated cells with a final concentration of 10 μM CCF2-AM substrate according to manufacturer’s recommendations. When cells were analyzed for fluorescence on a Cytek Aurora spectral analyzer, both MHV68 and MHV68.LANA:βlac infected samples were characterized by an extremely bright fluorescent signal (Figure 1B), consistent with what had been observed in previous reports with conventional band pass flow cytometry (ETC, personal communication). Maximal fluorescence of the uncleaved CCF2 substrate, in MHV68-infected cultures, was detected in the V7 and B2 fluorescent detectors (Figure 1B, second row), consistent with fluorescein emission. In contrast, MHV68.LANA:βlac-infected cultures demonstrated an additional fluorescence maximum in the V3 fluorescent detector, consistent with LANA:βlac-dependent substrate cleavage. We propose that retention of fluorescence in the V7 and B2 detectors likely results from incomplete cleavage of CCF2 in LANA:βlac+ cells, a phenomenon also observed in other reports using this system (Zlokarnik et al., 1998; Nealy et al., 2010; Diebel et al., 2015). Despite the ability to detect a unique fluorescent signature in MHV68.LANA:βlac-infected cells, CCF2-associated fluorescence in these conditions was so great that the original Cytek assay gain settings for the violet and blue detectors had to be decreased by more than 90%. This reduction severely compromised the potential to analyze additional fluorophores excited primarily by these lasers.

To address this limitation, we analyzed the impact of CCF2-AM substrate concentration and incubation time on fluorescent emission. Notably, use of 0.1 μM CCF2-AM, a 100-fold dilution in substrate, produced a signal that was readily detectable and on-scale yet did not require altering laser voltages on the cytometer (Figure 1B, bottom two rows). We further determined that incubation with the CCF2-AM substrate for half the recommended time (30 min instead of 60 min) still produced a robust signal (Figure 1B). These optimized substrate conditions afforded the opportunity to discriminate between uncleaved substrate, resulting in maximal fluorescence in the V7 and B2 detectors, and cleaved CCF2 substrate, uniquely fluorescing in the V3 detector while retaining fluorescence in the V7 and B2 detectors. Significantly, these settings did not require compromised gains for either the blue or violet laser, allowing detection of other fluorophores. In our optimization studies, we also determine the best method for unmixing cleaved and uncleaved substrate, as well as the effect of fixation and permeabilization on the CCF2 fluorescent signal. Detailed information on these methods will be provided by the authors upon request.

To determine if our optimized CCF2-AM staining conditions worked in primary cells, we infected mouse peritoneal cells (PerCs) directly ex vivo. The PerC is rich in macrophages and B cells, two cell types known to be latently infected following MHV68 infection (Weck et al., 1999; Rekow et al., 2016). PerCs were isolated and incubated overnight in vitro with either WT MHV68 or MHV68.LANA:βlac (MOI = 10 PFUs/cell), followed by use of optimized CCF2-AM incubation conditions. WT MHV68 infected cultures had minimal fluorescence in the CCF2 cleaved substrate detection channel (Figure 1C, top row). In contrast, ∼12% of MHV68.LANA:βlac infected PerCs demonstrated CCF2 cleaved substrate-dependent fluorescence (Figure 1C, bottom row). Furthermore, these conditions allowed us to use additional fluorophores excited by the violet and blue lasers, a result unachievable using the manufacturer’s recommended staining conditions. The use of multiple antibodies to detect a panel of cell surface markers allowed us greater resolution to immunophenotype virus-positive cells. MHV68.LANA:βlac was detected in a high frequency of F4/80+ macrophages, with a lower frequency of cleaved substrate detected in CD11c+ dendritic cells and CD19+ B cells, and no detectable fluorescence in CD3+ T cells (Figure 1C). Next, we tested our ability to detect MHV68.LANA:βlac infection in PerCs harvested from mice that were subjected to intraperitoneal infection and harvested at 3 days post-infection (dpi). Again, a robust cleaved CCF2 signal was observed, particularly among F4/80+ macrophages (Figure 1D).

With optimization established, we sought to apply this to analyze primary MHV68 infection using MHV68.LANA:βlac infection and spectral flow cytometry.

The viral lifecycle of γHVs is dynamic, with significant changes in viral load and cell and tissue tropism over time. To identify the consequence of acute MHV68 infection on cell types and targets of virus infection, mice were infected intraperitoneally (i.p.) with either WT MHV68 or MHV68.LANA:βlac, and analyzed over 14 days of infection. Though i.p. infection is likely not the natural route of infection for this virus, i.p. infections are a common strategy for virus delivery, including MHV68 (Barton et al., 2011). Additionally, MHV68 has a known tropism for peritoneal cells, and i.p. infection establishes robust and temporally synchronized infection within this tissue, providing an opportunity to investigate detailed kinetics of viral infection.

Cell counts revealed that i.p. MHV68 infection triggers a robust inflammatory response in the PerC, peaking at 8 dpi (Figure 2A). This is accompanied by appreciable splenomegaly (Supplementary Figure 1). Leukocyte cell counts began to subside after 8 dpi, but remain elevated compared to mock-infected mice. Despite host cell counts in response to MHV68.LANA:βlac infection appearing more variable than those following MHV68 infection, no significant differences were observed between these two experimental groups. MHV68 infection induced a prominent increase in T cell frequency and number, dominated by CD8+ T cells (Figures 2B,C). Notably, CD8 T cells increased by 3 dpi, indicating rapid engagement of CD8 T cells. Immune cell populations were defined as according to the gating strategy detailed in Supplementary Figure 2.

Murine gammaherpesvirus 68 infection was also associated with pronounced changes in myeloid subsets. In naïve mice, there are two dominant macrophage populations, LPMs, defined as F4/80^High CD11b^High MHC II^Low, and SPM, defined as F4/80^Mid CD11b^Mid MHC II^High (Ghosn et al., 2010). MHV68 infection reduced the frequency of LPMs relative to mock-infected mice, from day 1 through day 14 post-infection, with no LPMs detected at 8 dpi and an apparent reappearance by day 14 (Figures 3A,B). This dramatic loss of LPMs was accompanied by a pronounced increase in the frequency and number of F4/80^Mid CD11b^Mid cells, consistent with an SPM or monocyte phenotype, between days 3 and 14 post-infection (Figures 3A,B). The phenotype of F4/80^Mid CD11b^Mid cells changed over time. In mock-infected mice, F4/80^Mid CD11b^Mid cells were predominantly a Gr1^Low MHC II^High phenotype consistent with SPMs (Ghosn et al., 2010). Following MHV68 infection, however, F4/80^Mid CD11b^Mid cells demonstrated a transient increase in the frequency of Gr1^High MHC II^Low cells, consistent with monocyte infiltration into the PerC (Figure 3C). By day 8 and 14 post-infection, F4/80^Mid CD11b^Mid cells were primarily Gr1^Low MHC II^High cells, consistent with SPMs (Figure 3C).

The disappearance and reappearance of the LPM population by 14 dpi suggested self-renewal and/or replenishment. Despite prior reports detailing LPMs as self-renewing (Cassado Ados et al., 2015), the complete absence of LPMs at 8 dpi suggests replenishment from another population. Monocytes from the circulation are a likely candidate, as they have been shown to enter the PerC and acquire LPM characteristics (Bain et al., 2016). Notably, as infection progressed, SPM/monocyte and LPM populations became less distinct based on CD11b and F4/80 expression (e.g., see 14 dpi, Figure 3A). SPM and LPM protein expression was further altered by infection. In uninfected mice, SPMs have been characterized as MHC II^High CD86^Low, in contrast to LPMs which are MHC II^Low CD86^High (Ghosn et al., 2010); we also find these distinct phenotypes in mock-infected mice (Figure 3D). In contrast, by 14 and 21 dpi, F4/80^Mid CD11b^Mid SPM/monocytes upregulated CD86 expression compared to mock-infected conditions (Figure 3D). Conversely, F4/80^High CD11b^High LPMs increased their expression of MHC II, with LPMs demonstrating an MHC II^High phenotype by day 14 and 21 dpi (Figure 3D). These phenotypic changes resulted in SPM/monocyte and LPM populations both expressing an MHC II^High CD86^High phenotype by 14 dpi, blurring the phenotypic distinction between these cell subsets. In addition to phenotypic changes in macrophage subsets in the PerC, MHV68 infected mice further demonstrated an increase in the frequency of neutrophils in the PerC by 14 dpi (Figure 3E). These data demonstrate multiple time-dependent changes in leukocyte composition and phenotype following primary MHV68 infection.

We next analyzed the cellular distribution of MHV68.LANA:βlac infection in the PerC over time. The frequency of LANA:βlac+ cells was relatively constant between 1 and 3 dpi, with at least 5% of total peritoneal cells demonstrating CCF2-AM substrate cleavage, a single-cell marker for LANA:βlac expression (Figure 4). The frequency of LANA:βlac+ cells decreased significantly (∼ 10-fold) by 8 dpi, with a further decrease in the frequency of LANA:βlac+ PerCs by 14 dpi (Figure 4). Despite the decreased frequency of LANA:βlac+ cells over time, these cells were readily detected above the limit of detection defined by coincident analysis of mock and WT MHV68 infected samples (Figure 4).

We next immunophenotyped PerCs from infected mice using a panel of 12 cell surface markers. As predicted from our initial in vivo experiments (Figure 1D), we detected the virus in multiple cell types, including macrophages, dendritic cells, and B cells, at varying frequencies (Figure 5). Of these, the majority (>75%) of LANA:βlac+ cells were identified within LPMs early after infection (1 and 3 dpi) (Figure 5). By 8 dpi, however, F4/80^Mid CD11b^Mid SPMs were the major LANA:βlac+ population (Figure 5), a change in tropism coincident with a nearly complete disappearance of LPMs from the PerC (as shown in Figure 3). B cells and dendritic cells accounted for less than 5% of LANA:βlac+ cells through 14 dpi (Figure 5). Despite an increase in number over time of infection (Figure 3), neutrophils were not identified as a target of early peritoneal infection (data not shown). Thus, we conclude that macrophages are the major target of MHV68 infection in the PerC during the first 2 weeks of infection.

In parallel, we assessed the percentage of each cell population that was virally infected. This analysis identified a hierarchy of infection 1 day post infection, with ∼25% of LPMs, 7% of dendritic cells, 5% of SPM/monocytes and 0.2% of B cells which were LANA:βlac+ (Figure 6). LPMs remained a predominant target of infection at 3 dpi, with >40% of LPMs characterized by LANA:βlac expression, in contrast to SPM/monocytes, dendritic cells and B cells which contained <3% LANA:βlac+ cells out to 14 dpi (Figure 6). LANA:βlac expression was not detected in T cells at any time analyzed. These data identify multiple targets of primary MHV68 infection in the PerC.

Given the impact of MHV68 infection on myeloid cells, and the partial infection of LPMs and SPMs, we next analyzed how MHV68 infection affected macrophage phenotype in a cell-intrinsic and cell-extrinsic manner. To do this, we analyzed MHC II and CD86 expression, two proteins differentially expressed in LPMs and SPMs at baseline (Ghosn et al., 2010), comparing LPMs and SPMs in mock-infected mice with LANA:βlac+ (virus-infected) and LANA:βlac− (uninfected) cell subsets from infected mice. Timepoints of 1, 3, and 8 dpi were examined; the 14 dpi time point was excluded due to the very low frequency of LANA:βlac+ events observed in these samples. As anticipated, mock-infected mice contained LPMs with a predominant MHC II^Low CD86^High phenotype (Figure 7, “Mock LPM”). After 1 dpi, >94% of LANA:βlac+ and LANA:βlac− LPMs were MHC II^Low CD86^High, comparable to mock-infected LPMs (Figure 7, left panel). By 3 dpi, however, MHV68 infected mice demonstrated an increased frequency of MHC II^High LPMs compared to mock-infected mice. MHC II^High CD86^High LPMs were most prominent among virally infected (LANA:βlac+) LPMs (49.2% of events), with an intermediate frequency (23.6%) in LANA:βlac− LPMs (Figure 7, middle panel). CD86 expression remained high in LPMs in all conditions (Figure 7, bottom panel). These data demonstrate that MHV68 infection is associated with an increased frequency of MHC II^High LPMs, a process that occurs in both virus-infected and -uninfected cells during acute infection.

In parallel, we analyzed the impact of MHV68 infection on the phenotype of F4/80^Mid CD11b^Mid cells, a phenotype containing a mixed population of SPM and monocytes (referred to collectively as SPM/monocytes). Mock-infected SPM/monocytes were primarily characterized by an MHC II^High CD86^Low phenotype (Figure 7). After overnight infection and through 3 dpi, LANA:βlac+ SPM/monocytes had a relatively comparable frequency of MHC II^High events comparable to mock-infected mice. In contrast, LANA:βlac− SPM/monocytes at these early times had a reduced frequency of MHC II^High events compared to mock-infected cells (Figure 7). By 8 dpi, however, both LANA:βlac+ and LANA:βlac− SPM/monocytes had an increased frequency of MHC II^High events relative to mock-infected mice. While CD86 expression was comparable in all measured SPM/monocyte populations after overnight infection, LANA:βlac+ demonstrated a transient upregulation at 3 dpi relative to mock-infected and LANA:βlac− cells, with both LANA:βlac+ and LANA:βlac− cells characterized by a modest upregulation of CD86 by 8 dpi (Figure 7). In total, these data identify that MHV68 infection is associated with enhanced MHC II and CD86 expression in virally infected and -uninfected macrophage subsets in the PerC.

Our analysis of the cellular distribution of LANA:βlac identified peritoneal macrophages as a prominent infected cell type at early time points post-infection. However, the state of virus infection in these cells remained in question. Though MHV68 has been reported to latently infect macrophages at late times post-infection (Weck et al., 1999), MHV68 has also been reported to lytically replicate in macrophages in vitro and in vivo (Tarakanova et al., 2007; Lawler et al., 2015). Additionally, it remained possible that the detection of LANA:βlac in peritoneal macrophages may result from phagocytosis of other lytically infected cells rather than bona fide infection. To address this issue, we quantified the frequency of cells capable of producing virus using a limiting dilution-based assay, comparing virus production from intact and mechanically disrupted PerCs (Weck et al., 1996). In this assay, virus production from intact cells can result from either preformed, cell-associated virions or reactivation from latency. In contrast, mechanical disruption of cells prevents reactivation from latency, a process that requires viable cells, while having minimal impact on the detection of preformed virus (Weck et al., 1996). PerCs from mock, MHV68, or MHV68.LANA:βlac-infected mice were isolated at 3 dpi and plated in a series of dilutions on a MEF monolayer, a sensitive and permissive indicator for virus infection. Cocultures were observed for virus-induced CPE on MEF monolayers, quantifying preformed virus in mechanically disrupted PerCs and the sum of preformed virus and reactivation from latency in intact PerC co-cultures.

Peritoneal cavities isolated from MHV68- or MHV-68.LANA:βlac infected mice demonstrated virus-induced CPE in both intact and mechanically disrupted cells (Figure 8). In contrast, mock-infected PerCs demonstrated no detectable CPE in either condition (Figure 8). When we quantified the frequency of intact cells capable of producing virus, we found that ∼1 in 9 PerCs (∼11, 95% confidence interval: 8.4–14%) were associated with virus production in both MHV68 and MHV-68.LANA:βlac-infected mice. In contrast, limiting dilution analysis of mechanically disrupted cells identified a frequency of ∼1 in 100 (1%) of cells containing preformed virus (Figure 8). These data indicate that PerCs from MHV68-infected mice contain a mixture of latent and preformed virus, with latent infection comprising the dominant fraction of infected cells. Of note, the quantification of virus-infected cells determined via this functional assay was greater than that observed by flow cytometry at the same time point post infection (∼11 versus ∼5 ± 0.6%). Thus, consistent with previous reports (Nealy et al., 2010) assessment of infection frequency by LANA:βlac cleavage and CCF2 fluorescence is likely an underestimate of total primary infection frequency.