Healthy adult human donor buffy coats were obtained from Australian Red Cross Lifeblood service from which peripheral blood mononuclear cells (PBMCs) were isolated through density gradient centrifugation using Ficoll-Paque PLUS (GE Healthcare). There was no selection bias based on gender. Isolated PBMCs were cultured in complete RPMI medium (RPMI 1640 with L-Glutamine (Lonza) supplemented with 10% human serum (Sigma-Aldrich). In experiments which utilized purified MAIT cells, MAIT cells were FACS isolated from PBMCs after positive co-staining in FACS buffer (PBS supplemented with 1% FCS and 10 mM EDTA) with fluorochrome conjugated antibodies: anti-CD3, anti-TCR Vα7.2 and 5-OP-RU loaded MR1-Tetramer. MAIT cells were sorted to a >98% purity using BD Influx (BD Biosciences). For some experiments, where specified, MAIT cells were further FACS sorted on the basis of CD69 expression to a purity of >97% for both CD69^- and CD69⁺ populations using BD Influx (BD Biosciences).

ARPE-19 epithelial cells (ATCC) and ARPE-19-GFP cells that overexpress MR1 with GFP under the same promoter via a downstream internal ribosome entry sequence (38) were cultured in complete DMEM medium (DMEM with 4.5 g/L glucose and L-glutamine (Lonza), supplemented with 10% Foetal calf serum (FCS) (Sigma Aldrich) and 1% penicillin streptomycin (Gibco). A clinical VZV strain (VZV-S) and a recombinant VZV rOka-ORF10-GFP (VZV-GFP), which expresses GFP in fusion with ORF10 (39), were propagated in ARPE-19 cells in complete DMEM medium. All cells were cultured at 37 °C 5% CO₂.

PBMCs were either mock or VZV inoculated via co-culture with either uninfected or VZV infected ARPE-19 cells, respectively. The viral inoculum consisted of >75% VZV infected ARPE-19 cells demonstrating cytopathic effect (CPE). Inoculum was trypsinized, washed and resuspended in supplemented RPMI and added to PBMCs at a ratio of 1:2-5 ARPE-19: PBMC. For some experiments assessing viral infectivity, inoculum was added to PBMCs at various ratios of: 1:2, 1:5, 1:10, and 1:20 inoculum: PBMC. In parallel, inoculum was also added to ARPE-19 GFP expressing cells at identical ratios of 1:2, 1:5, 1:10 and 1:20 inoculum: ARPE-19-GFP cells for comparison. Kinetics of infection was investigated via a time-course experiment, inoculum was co-cultured with either PBMCs or ARPE-19-GFP cells at a ratio of 1:5 and harvested at various time-points of: 6, 24, 48 and 72 hours post inoculation. For experiments using total PBMCs, infections were performed in 12-well plates with 1-2 x 10⁶ PBMCs in 2 ml of complete RPMI medium per well. For experiments using FACS sorted MAIT cells, infections were performed in 24-well plates with 4 x 10⁵ cells in 600 ml complete RPMI medium per well. Following the addition of either mock or VZV infected cells to sorted MAIT cells and/or total PBMCs, cells were spinoculated in tissue culture plates for 15 minutes at 150 x g at 37°C. Plates were then incubated at 37°C 5% CO₂ for 2 days.

PBMCs for either surface staining flow cytometry or FACS sorting experiments were stained with the following fluorochrome conjugated antibodies: CD3-BUV395 (SK7), CD25-APC-H7 (M-A251), CD27-BUV661 (M-T271) (all BD Bioscience), CD8-SB780 (OKT8) (Thermo Fisher), CD4-PerCP/Cy5.5 (OKT4), CD56-BV605 (NCAM 16.2), TCR Vα7.2 (OF5A12), CCR4-BV421 (L291H4), CLA-AF647 (HECA-452), CCR2-APC (K036C2), CCR5-PE/Cy7 (J418F1), CCR6-BV421 (G034E3), CD69-BV421 (FN50), PD-1-PE/Dazzle (EH12.2H7), CD71-BV650 (CY1G4) (all Biolegend), VZV gE:gI (SG1-1, conjugated in house to Dy488), VZV gE:gI (SG1-1, conjugated in house to PE) (Meridian Life Sciences), 5-OP-RU loaded MR1 tetramer-PE, Ac-6-FP loaded MR1 Tetramer-PE. Matched isotype controls were used as negative controls.

Cells were collected and viability stained with Live/Dead Blue (Invitrogen) as per manufacturer’s protocol. Cells were then resuspended and washed in FACS buffer, before staining with antibodies on ice for 45 minutes. Cells were washed in FACS buffer then fixed in 4.2% formaldehyde (BD Biosciences) at 4°C for 15 minutes before acquiring on a LSR-II cytometer (BD Biosciences).

Data was analyzed using FlowJo software (versions 10.0.7 and 10.2; Tree Star). All PBMC data depicted was gated on live (as per the Live/Dead Blue viability dye staining) lymphocytes (as per distinct forward and side scatter morphology). MAIT cells were identified through positive co-staining with 5-OP-RU loaded MR1-Tetramer, anti-CD3 and anti-Vα7.2. All data observing viral infection of ARPE-19-GFP cells was gated on live (as per the Live/Dead Blue viability dye staining) GFP expressing cells.

PBMCs were co-cultured with mock or VZV-ORF10-GFP infected ARPE-19 cells for 2 days. Cells were collected, stained and FACS sorted for CD3⁺ MR1-Tetramer⁺ Vα7.2⁺ (MAIT) cells. To remove extracellular virus, MAIT cells were washed in citrate buffer (40 mM C6H5O7Na3, 135 mM NaCl, 10 mM KCl [pH 3]), at room temperature for 2 minutes before washing in PBS (35, 40–42). In duplicate, MAIT cells (2.5x10⁵) were resuspended in complete RPMI medium and then added to pre-seeded ARPE-19 monolayers (7.5x10⁵) on glass coverslips in 24-well plates. Co-cultures were spinoculated at 15 minutes at 150 x g 37°C, before being incubated at 37°C 5% CO₂ for 5 days to allow for formation of any CPE. Monolayers were fixed with 4.2% formaldehyde (BD Biosciences) at room temperature for 15 minutes, and then counterstained with DAPI. Imaging was performed using the Nikon Ti2-E Widefield fluorescence microscope.

Statistical analyses were performed using GraphPad Prism (version 9; GraphPad Software).

All blood work was performed in accordance with The University of Sydney Human Research Ethics Committee approval. All blood donations were obtained under agreement with the Australian Red Cross Lifeblood service.

Whilst the ability of VZV to infect human T cells is well documented (30, 31, 33, 34), the permissibility of MAIT cells to VZV is not known. To investigate this potential interaction, we assessed via flow cytometry the capacity of VZV clinical isolate (VZV-S) infected ARPE-19 epithelial cell-associated inoculum to infect MAIT cells in PBMCs. This cell-associated model of infection is commonly used in VZV studies (29, 30, 35) as VZV is very highly cell-associated in vitro (43). The mock and VZV infected inocula were excluded from analysis as per FSC-A SSC-A morphology gating and CD3 negative expression ( Figure 1 ). The detection of the surface VZV glycoprotein (g)E:gI complex, which is expressed late in the VZV replicative cycle, was utilized as a means to detect virally infected cells (33, 35–37, 44, 45). Therefore, all results presented within this study that refer to the “VZV^+” or “VZV infected” are expressing VZV gE:gI.

MAIT cells were identified via co-staining of 5-OP-RU loaded MR1 tetramer and CD3 ( Figure 1A ), with a vast majority of that population positively staining for the canonical MAIT TCR, Vα7.2 (range: 96.2-99.8%) ( Supplementary Figure 1A ). In line with previous studies (1, 2), we detected an average of 2.4% of live PBMCs as MAIT cells, with no difference in the frequency of live MAIT cells between mock and VZV inoculated samples ( Supplementary Figure 1B ). When examining the transfer of viral infection to the total pool of live PBMCs, a mean of 19.2% (range 11.4-23.2%) of these cells were VZV gE:gI⁺ ( Figure 1B ).

A mean of 17% of MAIT cells were gE:gI⁺ (range 6.2-23.2%). When examining gE:gI expression in non-MAIT cell (ie MR1 Tet^-) CD4⁺ cells (mean 15.8%, range 4.8-22%) and non-MAIT cell (ie MR1-Tet^-) CD8⁺ cells (mean 16%, range 5.7-24.1%) populations, we observed no significant difference in infection level when compared to MAIT cells across 14 different donors ( Figure 1B ). Furthermore, gE:gI expression on MAIT cells was detected as early as 6 hours post inoculation (mean 4.7%, range 3.6-5.2%) and peaked at 48 hours post inoculation (mean 23%, range 20.5-25%) ( Supplementary Figure 2A ). Additionally, gE:gI expression was detected on MAIT cells at various viral inoculum: PBMC ratios from 1:2 (mean 21.7%, range 20.5-22.6%) to 1:20 (mean 7.7%, range 5-12.1%) ( Supplementary Figure 2B ). In a comparison to epithelial (ARPE-19) cells, MAIT cells were less permissive to VZV infection ( Supplementary Figure 2 ). Together, these results identify MAIT cells as a T lymphocyte compartment that is permissive to VZV infection. Furthermore, VZV infection of MAIT cells was less than that of infection of ARPE-19 cells but comparable to infection of non-MAIT cell CD4⁺ and CD8⁺ T lymphocyte populations.

Similar to other innate-like T cell populations, the MAIT cell compartment consists of a heterogenous mix of subpopulations with distinct functional attributes (1, 46, 47). Using flow cytometry, we sought to determine the extent to which VZV infection of MAIT cells was associated with different subsets of MAIT cells. MAIT cells were split into distinct subpopulations based on the expression of the following cell surface markers: co-receptor (CD8 and CD4), co-stimulatory (CD27), immune checkpoint marker (PD-1), and CD56. Flow cytometric analysis revealed the frequencies of MAIT cell sub-populations ( Figure 2A ), and these were consistent with previous literature (1, 46–48). Furthermore, no change of frequencies within MAIT cell subpopulations between mock and VZV infected samples was observed ( Supplementary Figure 1C ).

When examining the co-receptor subpopulations, there was a significantly higher level of infection of CD4⁺ (mean 23.6%, range 12-32.5%) and double positive (CD4⁺/CD8⁺) (mean 22.8%, range 9.6-35.1%) MAIT cells compared to CD8⁺ (average 16.8%, range 6.2-22.7%) and double negative (CD4^-/CD8^-) (mean 16%, range 5.4- 31.4%) MAIT cells ( Figure 2B ). There was no significant difference in the level of infection between CD56⁺ (mean 18.1%, range 6.6-30.6%) and CD56^- (mean 17.1%, range 6.1-23.4%) MAIT cells ( Figure 2B ). Furthermore, there was no significant difference in the level of infection between CD27⁺ (mean 7.9%, range 6.2-10.4%) and CD27^- (mean 9.4%, range 6.3-15%) MAIT cells, and both PD-1⁺ (mean 9.3%, range 6.5-14.5%) and PD-1^- (mean 7.7%, range 6.1-10.2%) MAIT cells demonstrated similar levels of infection ( Figure 2B ). Overall, all subpopulations examined in this study were comparably infected, with the exception of the proportion of CD4⁺ and CD8⁺/CD4⁺ double positive MAIT cells being associated most with VZV infection.

We sought to determine whether VZV infection of MAIT cells was associated with an altered expression profile of activation (CD69) and proliferation (CD71) markers, as determined by flow cytometry of mock and VZV infected MAIT cells. As an additional comparison we also examined mock and VZV infection of non-MAIT (ie MR1-Tet^-) CD3⁺ T cells. Mock inoculated MAIT cells endogenously expressed higher CD71 (mean 1%, SEM +/- 0.15%) compared to non-MAIT CD3⁺ T cells (mean 2.18%, SEM +/- 0.19%). Whilst a significantly greater proportion of mock MAIT cells endogenously expressed higher CD69 (mean 1.95%, SEM +/- 0.35%) compared to non-MAIT CD3⁺ T cells (mean 19.95%, SEM +/- 3.56%) ( Figure 3A ). Furthermore, our analysis revealed a significantly higher proportion of VZV infected (gE:gI⁺) MAIT cells expressed CD71 (mean 40.62, SEM +/- 2.01) compared to mock infected (mean 2.18%, SEM +/- 0.19%) ( Figure 3B ). This was also the case when observing CD69 expression in VZV infected MAIT cells (mean 36.15%, SEM +/- 6.32%) compared to mock (mean 19.95%, SEM +/- 3.56%) ( Figure 3B ). Indeed, a significantly greater number of CD71/CD69 double positive MAIT cells was correspondingly observed in the VZV infected condition (mean 15.84%, SEM +/- 2.59%) compared to mock (mean 0.99%, SEM +/- 0.17%) ( Figure 3C ). Furthermore, a higher proportion of CD69 and CD71 expressing cells were similarly detected in VZV infected non-MAIT CD3⁺ T cells compared to mock infected counterparts ( Figure 3B ). When analysing the VZV bystander (gE:gI^-) MAIT cell subpopulation, only a conservative albeit significant increase of CD71 expression compared to mock MAIT cells was observed (mean 4.58%, SEM +/- 0.1%), whilst no significant change of CD69 expression compared to mock was detected ( Supplementary Figure 4A ).

Following on previous reports that demonstrate a preferential infection of CD69 expressing T lymphocytes (30) we FACS sorted MAIT cells by CD69 expression into two populations: CD69^- and CD69⁺ MAIT cells ( Supplementary Figure 3A ). We observed no significant difference in VZV infection when comparing CD69^- and CD69⁺ sorted MAIT cells ( Supplementary Figure 3B ). We also examined CD69 expression on these MAIT cells sorted on the basis of CD69 expression. In comparison to mock infected counterparts, we did not observe an increase of CD69 expression by CD69-sorted MAIT cells following VZV infection, however, there was a significant increase in CD69 expression following VZV infection of CD69⁺ sorted MAIT cells ( Supplementary Figure 3C ). Collectively, these data indicate that VZV infection of MAIT cells and non-MAIT T cells is associated with the upregulation of CD71⁺ and CD69⁺ expression.

We next sought to determine the impact of VZV infection on the natively high expression of extravasation and skin homing markers on MAIT cells (9, 10). Initially, expression of key extravasation markers CCR2, CCR5 and CCR6 on mock inoculated non-MAIT (ie MR1-Tet^-) CD3⁺ cells was compared to mock infected MAIT cells. This analysis revealed that a greater proportion of MAIT cells endogenously expressed CCR5, CCR6 and CCR2 ( Figure 4A ), which is consistent with MAIT cells possessing a potent program for extravasastion (9). In the context of VZV infection of MAIT cells, we observed no significant difference in proportion of infected cells expressing CCR2, CCR5 or CCR6 in comparison to mock infection ( Figure 4B ). The proportion of non-MAIT CD3⁺ T cells expressing CCR2 and CCR6 were also not different between mock and VZV infection, whereas there was a significant increase in the proportion of non-MAIT CD3⁺ T cells expressing CCR5 in VZV infection (mean 21.3%, SEM +/- 4.94%) when compared to mock infected counterparts (mean 12.5, SEM +/- 3.56%) ( Figure 4B ).

The upregulation of skin homing markers such as CLA and CCR4 has been reported on VZV infected tonsillar T cells (30, 33, 34), providing impetus to examine the expression of these skin homing markers on VZV infected MAIT cells. A high proportion of mock infected MAIT cells expressed CLA compared to non-MAIT CD3⁺ cells ( Figure 5A ), which is consistent with previous literature (10). In contrast, the proportion of both uninfected MAIT cells and non-MAIT CD3⁺ cells that expressed CCR4 was relatively low, and comparable to each other ( Figure 5A ).

There was no significant difference between mock and VZV infected in the proportion of CLA⁺ MAIT cells, whilst VZV infected non-MAIT CD3⁺ cells demonstrated a significant upregulation of CLA compared to mock infected non-MAIT CD3⁺ cells ( Figure 5B ). Furthermore, there was a trend to an increased proportion of CCR4⁺ expression in VZV infected MAIT cells compared to mock ( Figure 5B ), whilst no change was observed in VZV bystander MAIT cells compared to mock ( Supplementary Figure 4B ). In addition, a greater proportion of CCR4⁺ non-MAIT CD3⁺ cells was observed during VZV infection compared to mock ( Figure 5B ). These data demonstrate that VZV infection of MAIT cells does not impair CLA or CCR4 expression and that VZV infection of non-MAIT CD3⁺ cells increases the expression of these skin homing markers.

Together, these results indicate that VZV infection of MAIT cells does not impair, but rather maintains expression of cell-surface cellular proteins associated with extravasation and skin homing programs.

Having established that MAIT cells were infected with VZV, we sought to determine whether VZV infected MAIT cells were capable transmitting infectious virus to other cells. We performed an infectious center assay which has been previously utilized to demonstrate productive infection and new infectious virion production in VZV infected T cells, NK cells and dendritic cells (29, 30, 35). MAIT cells were isolated by FACS sorting from PBMCs that had been exposed to a GFP-tagged VZV (VZV-ORF10-GFP) or had been mock infected ( Figure 6A ). The sorted MAIT cells were washed with citrate buffer to inactivate and detach any surface bound virions (41, 42, 49), before being co-cultured with uninfected ARPE-19 epithelial cell monolayers. After five days in culture, the presence of virus-induced cytopathic effect (CPE) in the ARPE-19 monolayer was determined via detection of GFP signal by fluorescence microscopy ( Figures 6C, D ). ARPE-19 cell monolayers co-cultured with Mock infected MAIT cells yielded no GFP fluorescence ( Figure 6B ) whereas distinct GFP⁺ infectious centers were readily detected in ARPE-19 monolayers co-cultured with MAIT cells infected with VZV ( Figures 6C, D ). These data demonstrate that that human MAIT cells infected with VZV are capable of transmitting infectious virus to epithelial cells.