You are here

Article Text

Article menu
Oncolytic and local immunotherapy
Original research
Treatment with oncolytic vaccinia virus infects tumor-infiltrating regulatory and exhausted T cells
  1. Kristin DePeaux1,2,
  2. William G Gunn1,2,
  3. Dayana B Rivadeneira1,2 and
  4. Greg M Delgoffe1,2
  1. 1Tumor Microenvironment Center, UPMC Hillman Cancer Center, Pittsburgh, Pennsylvania, USA
  2. 2Department of Immunology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, USA
  1. Correspondence to Dr Greg M Delgoffe; delgoffeg{at}upmc.edu

Abstract

Background Oncolytic viruses (OVs) are an attractive way to increase immune infiltration into an otherwise cold tumor. While OVs are engineered to selectively infect tumor cells, there is evidence that they can infect other non-malignant cells in the tumor. We sought to determine if oncolytic vaccinia virus (VV) can infect lymphocytes in the tumor and, if so, how this was linked to therapeutic efficacy.

Methods To investigate infection of lymphocytes by VV, we used a GFP reporting VV in a murine head and neck squamous cell carcinoma tumor model. We also performed in vitro infection studies to determine the mechanism and consequences of VV lymphocyte infection by VV.

Results Our findings show that VV carries the capacity to infect proportions of immune cells, most notably T cells, after intratumoral treatment. Notably, this infection is preferential to terminally differentiated T cells that tend to reside in hypoxia. Infection of T cells leads to both virus production by the T cells as well as the eventual death of these cells. Using a mouse model which overexpressed the antiapoptotic protein Bcl2 in all T cells, we found that reducing T cell death following VV infection in MEER tumors reduced the number of complete regressions and reduced survival time compared with littermate control mice.

Conclusions These findings suggest that OVs are capable of infecting more than just malignant cells after treatment, and that this infection may be an important part of the OV mechanism. We found that exhausted CD8+ T cells and regulatory CD4+ T cells were preferentially infected at early timepoints after treatment and subsequently died. When cell death in T cells was mitigated, mice responded poorly to VV treatment, suggesting that the deletion of these populations is critical to the therapeutic response to VV.

  • Oncolytic virus
  • Infection
  • Head and Neck Cancer
  • T cell

Data availability statement

Data are available upon reasonable request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/jitc-2024-009062

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Oncolytic viruses (OVs) like talimogene laherparepvec (T-VEC) can infect non-malignant cells in the tumor. What effect, if any, this has on therapeutic efficacy is unknown.

    WHAT THIS STUDY ADDS

    • We show that oncolytic vaccinia virus (VV) directly infects T cells within the tumor microenvironment, that this infection yields infectious VV, leads to T cell death, and is required for treatment efficacy.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • This study directly shows that T cells are productively infected by a strain of oncolytic vaccinia designed to only replicate in tumor cells. It also shows that infection and killing of T cells are required for treatment efficacy. This suggests that the role of OVs infecting other cells in the tumor may be an important step in the OV mechanism of action and should be investigated in other virus backbones, as well as considered in the design of OV therapy protocols.

    Background

    Oncolytic viruses (OVs) have gained interest for their ability to act as an immunotherapy. OVs can induce a de novo immune response in an otherwise immunologically inert or ‘cold’ tumor through their induction of an immunogenic cell death within tumors. While the majority of OVs have both natural and engineered tropisms for tumor selectivity, there is still the potential for infection of non-malignant cells. For example, in the case of T-VEC, the only Federal Drug Administration (FDA)-approved OV, 1 day post injection it was found that all analyzed cell types contained HSV1 transcripts.1 This included non-malignant leukocytes such as B cells, dendritic cells, CD8 and CD4 T cells. It is currently unknown what the effects of lymphocyte infection by an OV are, making this important to study. Others have shown that oncolytic vesicular stomatitis virus can infect chimeric antigen receptor (CAR)-T cells as a mechanism to deliver both therapies into the tumor simultaneously.2 3 This suggests that at least these two OVs are capable of infecting non-tumor cells. Data from the virology field suggests that other viruses used for OVs, Newcastle disease virus and adenovirus, can both infect lymphocytes.4 5 Similarly, the Western Reserve strain of vaccinia was found to infect human peripheral blood mononuclear cells (PBMCs) better than the attenuated vaccinia strain, modified vaccinia Ankara.6 However, none of these studies were specifically conducted in the context of oncolytic virus immunotherapy, or within the tumor microenvironment. As such, many questions still remain such as: is this universal to oncolytic viruses (OVs)? Does infection lead to virus production? What effect does infection have on treatment efficacy?

    Oncolytic VV has been studied in many clinical trials and is currently in ~10 active trials in the USA and EU.7 We set out to determine if the oncolytic strain of Western Reserve vaccinia virus can infect lymphocytes in the tumor and if so what was the effect on treatment efficacy.

    Methods

    Experimental models

    Mice

    C57/BL6 mice were obtained from Jackson Laboratories, bred in house, and bred off-site at Charles River. Foxp3-RFP reporter mice (C57BL/6-Foxp3tm1Flv/J) and Foxp3-ametrine reporter mice (C57BL/6.Foxp3flpo-mAmetrine) were bred in house and off-site at Charles River; Rosa26LSL.BCL2.IRES.GFP mice were obtained from Dr. H. Christian Reinhardt at the University of Cologne and crossed in house to CD4cre mice. Hif1αf/f mice were obtained from Jackson Labs and crossed to CD4cre mice in house. Mice used in experiments were males and females between 6 and 10 weeks old at the initiation of the study.

    Cell lines

    Tumor experiments were conducted with MEER tumor cells. The MEER tumor line is a murine tonsillar epithelial cell which stably expresses E6/E7 and H-ras and has been rendered resistant to anti-PD1 therapy through serial treatment in vivo.8 HeLa cells, an immortalized epithelial line, were a gift from Dr. Saumendra Sarkar at the University of Pittsburgh. HeLa cells were cultured in DMEM, supplemented with 10% FBS (v/v). MEER cells were cultured in DMEM/F12 supplemented with 10% FBS (v/v), epidermal growth factor, insulin, transferrin, hydrocortisone, cholera toxin and tri-iodo-tyronine (all from Sigma-Aldrich).9 All cells were maintained at 37 °C with 5% CO2. T cells were cultured in RPMI supplemented with 10% FBS (v/v) and IL2 at 50 U/mL.

    Method details

    Tumor models

    C57BL/6, Foxp3-RFP reporter mice (C57BL/6-Foxp3tm1Flv/J), Foxp3-ametrine reporter mice (C57BL/6.Foxp3flpo-mAmetrine), or CD4cre Rosa26LSL.BCL2.GFP mice were implanted intradermally with 250 000 MEER tumor cells. When tumors reached approximately 20 mm2, they were treated intratumorally (IT) with a 25 µL injection of PBS or vaccinia virus expressing GFP (VV-GFP). VV-GFP was dosed at 2.5×106 PFU/mouse. Tumor growth and survival were monitored until tumors reached 15 mm in any direction.

    For tumor-infiltrating lymphocytes (TIL) analysis, tumors were implanted and treated the same way. At 1, 4, or 7 days post treatment, tumors were harvested and digested as previously described10 before staining single-cell suspensions for flow cytometric analysis. In experiments where pimonidazole staining was performed, mice were injected with pimonidazole (Hypoxyprobe, 80 mg/kg) intraperitoneally 20 min prior to tumor harvest.

    Oncolytic virus production

    VV-GFP (vvDD-GFP, Western Reserve strain11) was obtained as a gift from Dr. Saumendra Sarkar. VV-GFP contains an insertion of GFP into the viral thymidine kinase gene. HeLa cells were infected with VV-GFP at a multiplicity of infection (MOI) of 0.1 for 2 hours in DMEM supplemented with 10% FBS (v/v). Virus was then removed and fresh media added. Once cytopathic effect was visible (approximately 48 hours post infection (HPI)), cells were harvested into their supernatant by gentle rinsing. Cells were then pelleted, resuspended into 10 mL of 10 mM Tris–HCl (pH 7.0) and lysed by three cycles of freeze-thaw. The resulting supernatant was layered onto a sucrose cushion (36% sucrose in 10 mM Tris–HCl pH 7.0) and spun for 2 hours at 14 500 revolutions per minute at 4 °C. The viral pellet was resuspended in 200 µL of 10 mM Tris–HCl and stored at −80 °C. Purified virus was tittered on HeLa cells using a crystal violet plaque assay.

    Flow cytometry

    Cell surface staining was performed on ice in PBS for 20 min with surface antibodies and Zombie Viability Dye (Biolegend). Cells were then washed in PBS and either run for live panels or fixed with 4% paraformaldehyde for 5 min at room temperature and then washed in PBS. For nuclear staining, cells were fixed with the Foxp3 Transcription factor fixation kit (eBioscience) for 20 min at room temperature then stained overnight at 4°C. For tetramer staining, cells were stained in 1% BSA in PBS at 37°C for 30 min.

    Flow antibody information

    CD4 (GK1.5, PerCPCy5.5, APC, or BV786, Biolegend), CD8 (53–6.7, PacBlue, PECy7, or PE Biolegend), PD1 (29F.1A12, BV786, Biolegend), Tim3 (RMT3-23, APC or PE, Biolegend), CD45 (30-F11, PE, Biolegend), CD11b (M1/70, PacBlue, Biolegend), CD11c (N418, BV786, Biolegend), F4/80 (BM8, BV650, Biolegend), GR1 (R86-8C5, Biolegend), Ly6C (HK1.4, PECy7, Biolegend), CC3 (C92-605, BV605, BD Biosciences), Ly6G (1A8, PerCPCy5.5, Biolegend), Arginase1 (A1exF5, FITC, eBioscience), Hypoxyprobe (4.3.11.3, Biotin, Hypoxyprobe), Streptavidin Secondary (APCCy7, Biolegend), E7 tetramer (H2-Db RAHYNIVTF, PE, NIH Tetramer Facility), B8R tetramer (H2-Kb TSYKFESV, AF647, NIH Tetramer Facility), CD44 (IM7, BV421, Biolegend), and CD62L (MEL-14, BV786, Biolegend).

    In vitro T cell culturing and infection

    T cells were harvested from LN and spleen of C57Bl6, CD4creHifa1αf/f, or CD4cre Rosa26LSL.BCL2.GFP mice by dissociating the tissues followed by red blood cell lysis. For infection, experiments in figure 1 cell suspensions were stimulated with plate-bound αCD3 and soluble αCD28 with IL2 overnight and cultured for 5–7 days prior to infection. Cells were infected with VV-GFP overnight at various MOIs depending on the experiment as noted in the figure legends, washed, and given fresh media and IL2. 1.5×106 cells were plated per condition. For viral titers from T cell supernatants, media was harvested 48 hours post removal of virus. Supernatant was spun at 5000 revolutions per minute (RPM) for 5 min to pellet any cells and filtered with 0.45 µm pore size filter. Supernatant was then added to 85% confluent HeLa cells for 2 hours and washed off. HeLa cells were monitored for cytopathic activity and once observed plates were stained with crystal violet, washed, plaques counted, and titer calculated. For titers from infected HeLa cells, HeLa cells at 80% confluence were infected at an MOI of 10, 1.0, or 0.1 for 2 hours and then washed and media was replaced. After 48 hours, supernatant was harvested, spun, and filtered the same as above before tittering on additional HeLa cells. For T cell death over time, T cells were stimulated and infected as above and 48 HPI VV-GFP+ cells were sorted from infected samples and GFP− cells were sorted from mock-infected samples. Flow cytometry was run immediately post sorting and again 24 hours later (a total of 72 HPI).

    For infection experiments in figure 2, C57Bl/6 T cells were harvested using negative selection Mojo Sort magnetic separation kit (Biolegend) for all T cells and activated for 24 hours as described above prior to infection with VV-GFP overnight at an MOI of 10. Cells were infected overnight either in 20% O2, 5% CO2, at 37 °C or 1.5% O2, 5% CO2, at 37 °C. After overnight infection, cells were washed and given fresh media and IL2 and cultured for 48 hours in the same oxygen conditions. Cells were then stained for live flow cytometry. Cells were harvested from LN and spleens of CD4cre+ Hif1αf/f mice or CD4cre- Hif1αf/f wild-type littermate controls. Cells were activated with αCD3/αCD28 coated beads (10:1 beads to cells) and IL2 for 7 days in either 20% O2 or 1.5% O2. After 7 days, cells, stim was removed, cells were infected with VV-GFP at an MOI of 10 overnight. The next morning virus was washed off, fresh media with IL2 was replaced and cells remained in their O2 conditions. After 48 hours, cells were stained for live flow cytometry. For infection experiments in figure 3, cells were harvested using negative selection Mojo Sort magnetic separation kit (Biolegend) for all T cells from either CD4cre+Rosa26LSL.Bcl2.GFP mice or CD4cre-Rosa26LSL.Bcl2.GFP wild-type littermate controls. T cells were stimulated overnight with plate-bound αCD3, soluble αCD28, and IL2. Cells were then infected with VV-GFP at an MOI of 10 overnight in 1.5% O2, 5% CO2, at 37°C. Virus was washed off the next day and media replaced with fresh IL2. Cells were counted and plated in triplicate at 1×106 cells per condition and returned to the same culture conditions. Every 24 hours, cells were pulled from their conditions for counting by trypan blue staining and an extra 1 mL of media added to account for proliferation.

    In all experiments, ‘mock’ infected controls were included. These samples received identical conditions to infected cells with PBS added to their culture instead of virus during the viral incubations.

    Results

    Oncolytic vaccinia infects multiple immune populations in the tumor microenvironment

    As the strain of oncolytic VV used in our studies was not designed for cell-specific entry, it may be able to enter and replicate in other non-tumor cells within the TME. In fact, some studies have shown that OVs are capable of infecting immune cells.1 12–14 To test if our VV could infect lymphocytes in the TME, we employed MEER, a mouse tonsillar epithelial cell line transformed with HPV E6, E7, and h-Ras, an established model of HPV+ head and neck squamous cell carcinoma (HNSCC).15 MEER is an anti-PD1 resistant16 but VV-sensitive, HNSCC model.17 We treated mice harboring 4 mm diameter MEER tumors with an IT injection of VV-GFP at 2.5×106 PFU/mouse. VV-GFP is a Western Reserve strain of VV with a GFP reporter inserted into the thymidine kinase locus to produce GFP during the early stage of replication. Tumor (T) and draining lymph nodes (dLN) were analyzed at 1 or 4 days post treatment (figure 1A) for GFP expression in CD45+ cells. We found little GFP in the CD45+ cells at day 1 in either tissue (figure 1B–C). However, there was significantly more GFP in leukocytes at day 4 in the tumor than the dLN (figure 1B–C). We broke down the populations making up this GFP+ population at day 4 and found that the majority of non-tumor GFP+ cells were CD4+ and CD8+ T cells (figure 1D). CD8+ T cells had the highest MFI, or amount of viral GFP per cell, of the lymphoid populations, while the GFP expression in myeloid populations was highly variable (figure 1E). We also calculated the enrichment of each infected cell population within the total GFP+ population as a ratio of the percentage of that population within GFP+ cells to total CD45+ (% of GFP+: % of CD45+). We found that CD8+ T cells were significantly enriched in the GFP+ cells compared with all other analyzed populations (figure 1F).

    Figure 1
    Figure 1

    Oncolytic vaccinia infects multiple immune populations in the TME. (A) Schematic of the figure. Foxp3-reporter mice (Foxp3-RFP or Foxp3-Ametrine) were implanted with 250 000 MEER tumor cells and when tumors reached 4 mm in diameter (7 days), were treated intratumorally with PBS or vaccinia virus expressing GFP (VV-GFP) at 2.5×106 PFU/mouse. 1 or 4 days post treatment tumors (T) and draining lymph nodes (dLN) were harvested for live flow cytometric analysis. (B) Representative flow plots of VV-GFP expression in CD45+ cells in PBS control or VV-GFP treated tumors. (C) Quantification of vaccinia expressed GFP in CD45+ cells in dLN and T of VV-treated mice. (D) Pie chart showing the makeup of immune populations within the GFP+ CD45+ gate from (B) in VV-GFP treated tumors and (E) MFI of VV-GFP in those populations. Gating strategy is shown in Suppl 1 and populations were gated as follows, all from live CD45+: B cells, CD4− CD8− CD19+. NK cells, CD4− CD8− NK1.1+. CD4, CD4+CD8−. Treg, CD4+ CD8− Foxp3+. Tconv, CD4+ CD8− Foxp3−. CD8, CD4− CD8+. Texh, CD4− CD8+ PD1+ Tim3+. Dendritic cells (DCs), CD11c+CD11blo/+ MHCII+ F4/80−. Macrophages (Macs), CD11c− CD11b+ Ly6G− Ly6C− F4/80+. Monocytes (Mono), CD11c− CD11b+ Ly6G− Ly6C− F4/80−. MDSCs, CD11c− CD11b+ MHCII F4/80lo. (F) Infection enrichment calculated as a ratio of the population as % of GFP over the population as % of CD45+. (G) Representative flow plots and (H) quantification of VV-GFP in PD1 and Tim3 expressing CD8+ populations in both dLN and T. (I) Representative flow plots and (J) quantification of VV-GFP expression in Tconv (Foxp3−) and Treg (Foxp3+) cells in dLN and T. Data represent three independent experiments. Each point represents an individual mouse. *P<0.05, **p<0.01, ***p<0.001, ****p<0.0001 by one-way analysis of variance (ANOVA) with Sidak’s multiple comparison test (C, E), one-way ANOVA compared with CD8+ (F). Mixed effects analysis (H-TIL), paired t-test (H-dLN, J). ns, non-significant. Error bars indicate standard error of mean (SEMs). D1, day 1; D4, day 4; MFI, mean fluorescent intensity; dLN, draining lymph node; T, tumor; TME, tumor microenvironment.

    We next looked more closely at what populations of CD4+ and CD8+ T cells were infected. While MDSC and NK cells also made up a notable percentage of the GFP+ cells (figure 1D), we chose to focus on the T cells as they were more enriched in the GFP+ cells compared with the total CD45+ population (figure 1F) and are the majority population in the CD45+ compartment (online supplemental figure S1B). MDSC and NK cells are rarer populations in these tumors. We also chose not to further investigate infection of B cells, macrophages, and DCs as their uptake of antigen could result in GFP expression due to engulfment and not true infection. Infection of T cells was compared with PBS control animals (online supplemental figure S1C–D). We found little infection in the dLN of both CD4+ and CD8+ T cells (figure 1G–J). In the treated tumor infiltrate, we found that as CD8+ T cells expressed more inhibitory receptors there was an increase in their infectivity, with PD1+ Tim3+ exhausted T (Texh) cells having the most VV-GFP (figure 1G–H). We also found that Foxp3+ regulatory T (Treg) cells were significantly more infected than Foxp3− conventional T (Tconv) cells (figure 1–J).

    Supplemental material

    Oncolytic vaccinia depletes exhausted T cells in treated tumors

    We next sought to understand if there were any changes to the CD4+ and CD8+ T cell populations at day 4 compared with earlier and later timepoints where infection is not directly observed. We performed time course analysis at days 1, 4, and 7 post treatment with an IT injection of VV-GFP or PBS (figure 2A). We found no significant change in the total counts or percentage of CD4+ and CD8+ T cells infiltrating at days 1 or 4 post treatment (figure 2B–D); however, by day 7 post treatment, there was a significant increase in both populations as a percentage of live cells and by total cell counts, as previously reported17 (figure 2B–D).

    Figure 2
    Figure 2

    Oncolytic vaccinia depletes Texh cells from the tumor microenvironment. (A) Experimental schema for B–L. Foxp3-reporter mice were implanted with 250 000 MEER cells and, when tumors reached 4 mm in diameter (7 days) were treated intratumorally with PBS or vaccinia virus expressing GFP (VV-GFP) (2.5×106 PFU/mouse). Tumor-infiltrating lymphocytes (TIL) analysis was performed 1, 4, or 7 days post treatment. Representative flow plots (B, E, G) and quantification (C, D, F, H) of CD4+ and CD8+ T cells (B–D), Foxp3− conventional CD4+ (Tconv) or Foxp3+ regulatory CD4+ (Treg) cells (E and F), and PD1 and Tim3 on CD8+ T cells (G and H). (I) Total cell counts in TIL of PD1-Tim3−, PD1+Tim3−, and PD1+Tim3+ CD8+ cells at 4 days post treatment. (J) Representative flow histogram and quantification of cleaved caspase 3 (CC3) in CD8+ T cells at 4 days post treatment. (K) Total cell counts in TIL of Tconv and Treg cells 4 days post treatment. (L) Representative flow histogram and quantification of CC3 in TIL Treg cells at 4 days post treatment. Data represents four (B–I, K) or two (J, L) independent experiments. Each point represents an individual mouse. *P<0.05, **p<0.01, ***p<0.001, ****p<0.0001 by two-way analysis of variance with Sidak’s multiple comparison test (C, D, F, H), or unpaired t-test (I–L). ns, non-significant. Error bars indicate SEMs.

    We next looked at subsets of CD4+ and CD8+ T cells. We found no changes in the percentage of total CD4+ cells of Treg or Tconv cells at days 1 or 4 post treatment (figure 2E–F). We observed a significant increase in Tconv and concomitant decrease in Treg cells at day 7 (figure 2E–F) consistent with previous reports.17 18 We also looked at PD1 and Tim3 expression on CD8+ T cells as these markers could identify populations of tumor-infiltrating progenitor and terminally exhausted T cells. On entering the TME, CD8+ T cells differentiate into progenitor-exhausted T cells (Tpex), marked by intermediate PD1 expression, and ultimately into terminal exhaustion (Texh), where they coexpress multiple inhibitory receptors such as PD1 and Tim3.10 We find no changes at day 1 post treatment in Texh; however, as early at day 4 in VV-treated tumors, we find a significant reduction in PD1+ Tim3+ Texh which continues to day 7 (figure 2G–H). At 4 days post treatment, there is not yet an increase in CD8+ T cell counts (figure 2C) and, as such, this decrease in Texh is likely not just a difference in the ratio of PD1+ Tim3+ to PD1+ Tim3 cells. At day 7, we observe a significant increase in PD1+ Tim3 T cells, likely newly activated T cells (figure 2G–H).

    To understand why there was a decrease in the percentage of PD1+ Tim3+ CD8+ cells at day 4, we did further analysis at this time point. We found that not only were these cells decreased in percentage, but they were also significantly decreased by total absolute numbers while PD1+ Tim3 and PD1− Tim3− CD8+T cells were unchanged (figure 2). CD8+ T cells at day 4 also expressed significantly higher cleaved caspase 3 (CC3), a marker of apoptosis (figure 2J), in VV-treated tumors compared with PBS. Together these data suggest that this population is dying after infection with VV in the TME. We do not observe significant changes in Tconv or Treg cell counts (figure 2K) or CC3 expression in Treg cells (figure 2L) at this timepoint. This is likely due to Treg cell replenishment from the periphery or expansion within the TME19 making it difficult to capture cell death.

    Depleted T cells are not HPV E7 or VV specific

    We next wanted to test if VV-induced depletion affects tumor or viral-specific T cells. As new infiltrate enters the tumor by day 7 post infection (figure 2B) and we wanted to observe changes in tumor and viral-specific T cells we chose this time point to analyze. As before, MEER tumors were implanted in C57Bl6 mice and treated with PBS control or VV-GFP. We found that 7 days post VV treatment there was a significant increase in cell counts of HPV E7 tetramer+ (online supplemental figure S2A) and counts and percentage of vaccinia B8R tetramer+ CD8 T cells (online supplemental figure S2B). HPV E7 tetramer+ T cells were slightly decreased by percentage of total CD8+ cells after VV treatment (online supplemental figure S2A). This may be due to tumor antigen spreading and increased T cell receptor (TCR) diversity and not virally induced death as the cell counts are significantly increased. We next looked at the PD1 Tim3 and effector populations within tetramer− and tetramer+ CD8+ T cells. We found increases in the total cell counts in the tetramer− population after VV treatment compared with PBS and the E7+ or B8R+ populations (online supplemental figure S2C–E). This suggests that while there is an increase in tetramer+ cells after VV treatment, the majority of T cells are not specific to the tetramers we investigated. This supports the idea that antigen spreading may be occurring; however, as TCR sequencing was not performed, we cannot definitively conclude this. If VV specifically targeted tumor or VV-specific T cells for deletion, we would expect to find that the PD1+ Tim3+ cells within the E7+ or B8R+ population would be depleted. This was not the case and in fact we observed very little differences in the PD1 and Tim3 expression after VV treatment between the tetramer− and tetramer+ populations by percentage (online supplemental figure S2C–D). While there are significantly fewer cells of all populations in the tetramer+ population compared with tetramer− by counts, there are significantly more compared with PBS control (online supplemental figure S2D). This suggests that this decrease from tetramer− to tetramer+ is not a depletion within the B8R or E7-specific population, but instead an expansion of the tetramer− cells. Similarly, we observed no depletion of effector (CD44+ CD62L−) CD8+ cells in the tetramer+ populations after VV treatment (online supplemental figure S2E). We did, however, observe a significant decrease in PD1+ Tim3+ CD8+ T cells by percentage from the tetramer− PBS group to the tetramer− VV group while there was an increase in the cell counts. This suggests that there is a shift in the populations not captured by the tetramers used here.

    Infection of T cells by VV is productive and leads to cell death

    We next determined if the infection we were observing by GFP+ signal could yield infectious virions. As VV must proceed through complete replication to yield infectious virions, either through lytic or endocytic release, we infected T cells and HeLa cells and harvested their culture supernatant to titer.20 To do this, we infected previously activated, in vitro expanded CD8+ T cells or HeLa cells, the prototypical cells used for virus production, with VV-GFP at an MOI of 10, 1.0, or 0.1 overnight, then washed off the virus and replaced the media (figure 3A). After 2 days, the supernatant was removed and flow cytometry was run to confirm infection of the T and HeLa cells by VV-GFP expression (figure 3B–C). The supernatant was spun and filtered to remove any cells and plated onto HeLa cells for titer by crystal violet staining. Supernatants were collected from 1.5×106 cells for both HeLa and T cells. We found that at an MOI of 10, infected T cells produced a similar titer of virus to HeLa cells (figure 3D). However, at an MOI of 1.0 HeLa cells made significantly more virus than T cells (figure 3D). T cells still produced lytic virus when infected at an MOI of 0.1 (figure 3D, online supplemental figure S3A). The lower viral production in HeLa cells at an MOI of 10 compared with 1.0 is likely due to rapid death at such a high MOI in a small culture dish. To confirm infection and lytic ability of the virus produced by T cells, we cultured HeLa cells in supernatants isolated from mock or VV-infected T cells, analyzing them by flow cytometry for GFP expression and cell death. We found that HeLa cells expressed VV-GFP 48 HPI and began to die by Zombie Viability stain (figure 3E). Together these data show that the virus produced from T cells is functional. We also sought to determine if we could still observe GFP in T cells infected at lower MOIs. We found that there appears to be a threshold where, above an MOI of 2.5, GFP can be observed in T cells but below that it is only slightly above background (online supplemental figure S3B). While we cannot see significant levels of VV-GFP at low MOIs, we do observe lytic virus being produced by T cells at an MOI as low as 0.1 (online supplemental figure S3A). We believe the lack of GFP in low MOI cells is due to the amount of viral GFP being below the limit of detection of the flow cytometer. However, since VV must proceed through late stages of replication to produce infective virus, we believe that even at low MOIs, T cells are being productively infected. Finally, to determine if infected T cells die, we again infected activated T cells in vitro, 48 HPI-sorted VV-GFP+ cells, then monitored by flow cytometry for survival both immediately after sorting and 24 hours later (figure 3F). We also sorted VV-GFP– cells from mock-infected cultures for comparison. We found at 72 HPI, that VV-GFP+ cells harbored significantly more CC3 than mock controls (figure 3G–H). We next looked at Zombie Viability dye versus Annexin V to observe dead and apoptotic cells (figure 3). Annexin V binds phosphotidlyserine which is exposed to the plasma membrane during apoptosis while Zombie Viability dye stains dead cells that are permeable to the dye. We found significantly decreased live cells in the VV-GFP+ cells compared with mock-infected control at both 48 and 72 HPI as well as increased dead cells (figure 3J–K). From these data, we can conclude that infection of lymphocytes in the tumor is both productive, leading to release of infectious VV-GFP, as well as cytolytic, leading to the death of the infected T cells.

    Figure 3
    Figure 3

    Oncolytic vaccinia virus (VV) infection of T cells is productive and leads to cell death. (A) Experimental schematic. T cells from spleen and lymph node were activated with aCD3, aCD28, and IL2, and after 7 days were infected overnight with VV-GFP at an MOI of 10 or mock infected. The next morning, cells were washed in 10% FBS RPMI and then given fresh media with IL2. HeLa cells were also infected at an MOI of 10 or mock infected for 2 hours then washed with 10% FBS+ DMEM complete media and given fresh media. Cells were then cultured for 48 hours. At 48 hours post infection (HPI), supernatant was harvested, spun to remove cells, then filtered through an 0.45 µm filter. Supernatants were plated onto HeLa cells for crystal violet titer and flow cytometric analysis. (B) VV-GFP expression in T cells as analyzed by flow cytometry at the time of supernatant harvest. (C) VV-GFP expression in HeLa cells as analyzed by flow cytometry at the time of supernatant harvest. (D) Crystal violet titer on HeLa cells of supernatants harvested from infected T and HeLa cells in (B) and (C). Cells infected at the MOI listed on the x-axis. (E) VV-GFP expression as analyzed by flow cytometry of HeLa cells cultured in T cell supernatant harvested as in (A). (F) Schematic of G–K. T cells were activated and infected as in (A). 48 hpi GFP+ cells were sorted from VV-infected cultures and GFP− from mock infected. Cells were placed back in culture and ran for live flow cytometry analysis. After another 24 hours cells were analyzed by flow cytometry again. (G) Representative flow plots and (H) quantification of cleaved caspase 3 (CC3) staining. (I) Gating scheme for (J and K). (J) Representative flow plots and (K) quantification of live, early apoptotic, apoptotic and dead cells. Data represent three (A–E) or four (F–J) independent experiments. Each point represents an individual mouse (A–E T cells), replicate (A–E HeLa cells), or four individual mice with technical replicates per mouse (F–K). *P<0.05, **p<0.01, ***p<0.001, ****p<0.0001 by unpaired t-test (B–E) or one-way analysis of variance (ANOVA) with sidaks multiple comparisons test (H, K). ns, non-significant. Error bars indicate SEMs. Apop, apoptotic; EA, early apoptotic; HPI, hours post infection; O/N, overnight; sup, supernatant.

    Lymphocytes in hypoxia are infected by VV due to HIF1α signaling

    As we observed a significantly higher frequency of infection in Texh and Treg cells compared with other cell subsets, we wanted to determine the cause of this bias. Hypoxia is an important factor acting on T cells in the tumor microenvironment. HIF1α (hypoxia-inducible factor 1α) has been shown to be important in immune cell fate and function in many tissue settings.21 22 It is required for effector t cells in the tumor23 and overexpression of HIF1α leads to improved resident memory T cell production and better anti-tumor immunity.24 VV also interacts with hypoxia as it expresses C16, a protein which stabilizes HIF1α by binding prolyl hydroxylase domain-2 which prevents it from targeting HIF1α for proteasomal degradation.25 VV uses this to alter host cell metabolism to produce increased glutamine and TCA intermediates required for its replication.25 26 Treg cells and Texh cells have been shown to reside in regions of hypoxia within the tumor.27 28 To confirm this in the MEER model, we gave mice MEER tumors and treated with an IT injection of VV-GFP as in figure 2. Four days post treatment, mice were given an IP injection of pimonidazole 20 minutes prior to tissue harvest (figure 4A). Pimonidazole is a reagent which covalently binds proteins under low oxygen tension, forming an irreversible adduct which can then be stained with a fluorescently tagged antibody, termed hypoxyprobe. As OVs can alter tumor vasculature, we wanted to confirm any differences in hypoxia experienced by lymphocytes in VV-treated compared with PBS-treated tumors. While hypoxia levels increase from the dLN to the T as expected, we saw no significant difference in the levels of hypoxia experienced by CD45+ lymphocytes in dLN or tumor when comparing PBS to VV-treated tumors (figure 4B) and as such VV treatment itself does not seem to alter tumor hypoxia by day 4. We show that Treg cells experience more hypoxia than Tconv cells (figure 4C) and PD1+ Tim3+ CD8+ cells have increased hypoxia exposure (figure 4D), confirming previous reports. Together, these data show that both Treg cells and Texh cells, which are the predominant infected populations in the tumor, also experience more hypoxia.

    Figure 4
    Figure 4

    Infection of lymphocytes by vaccinia virus (VV) is hypoxia mediated. (A) Schematic of B–D. MEER tumors were implanted in Foxp3-reporter mice (Foxp3-RFP or Foxp3-Ametrine) and when tumors reached 4 mm in diameter were treated with an IT dose of VV-GFP at 2.5×106 PFU/mouse. 20 min prior to tissue harvest mice were given IP pimonidazole, which forms irreversible adducts on proteins under low oxygen tension. Tumors and draining lymph nodes (dLN) were harvested 4 days post treatment. Representative flow plots and quantification of hypoxyprobe staining (which detects pimonidazole adducts) in (B) CD45+ lymphocytes in PBS (gray)-treated and VV (blue)-treated tumors (C) tumor-infiltrating lymphocytes (TIL) Tconv (TC) and Treg (TR) cells, and (D) TIL PD1 Tim3 expressing CD8+ populations. (E) Schematic of F–J. Spleen and lymph node T cells were harvested from C57Bl/6 (F) or HIF1αf/f CD4Cre−/+ (G–J) mice and stimulated with αCD3/αCD28 beads and IL2 overnight in 20% O2. The next day cells remained in stimulation to simulate chronic antigen found in the tumor but were placed in either 20% or 1.5% O2 for 7 days. After 7 days, stimulation was removed, and cells were infected with VV-GFP at an MOI of 10 or mock infected overnight in their oxygen conditions. Cells were then cultured for another 48 hpi until flow analysis of VV-GFP+CD8+ T cells was analyzed. (F) Representative flow plots and quantification of VV-GFP+CD8+ T cells. (G) Representative flow plots and (H) quantification of VV-GFP+CD8+ T cells in HIF1αf/f CD4Cre-/+ cells cultured with chronic stimulation. (I) Representative flow plots and (J) quantification of VV-GFP+CD8+ T cells in HIF1αf/f CD4Cre−/+ cells cultured with acute stimulation. Data represent three independent experiments. Each point represents an individual mouse. *P<0.05, **p<0.01, ***p<0.001, ****p<0.0001 by paired t test (C), one-way analysis of variance (ANOVA) with Sidak’s multiple comparisons (B, D), unpaired t test (F), or two-way ANOVA with Sidak’s multiple comparisons test (H, J). ns, non-significant. Error bars indicate SEMs.

    As we show that Texh and Treg cells reside more in hypoxia compared with other T cell subtypes, are more infected than other subtypes, and previous data have shown that VV induces HIF1α stabilization for efficient viral replication, we wondered if hypoxia exposure itself increased VV infection. We moved to an in vitro model to test the effects of hypoxia in isolation using an in vitro culturing method which replicates the chronic stimulation and hypoxia conditions T cells experience in the tumor to yield an in vitro cultured Texh cell.10 To do this, spleen and lymph node T cells are stimulated for 24 hours as normal before being moved to their culture conditions (figure 4E). Cells were either removed from stimulation (acute stimulation) and placed in normoxia (20% O2) or tumor hypoxia (1.5% O2), or maintained in stimulation (chronic stimulation) for 7 days under normoxia or tumor hypoxia. We confirmed that these conditions yielded in vitro exhausted T cells by expression of inhibitory receptors (online supplemental figure S4A–D) and reduction in cytokine expression (online supplemental figure S4E–H). After 7 days in these conditions, cells were removed from stimulation and infected overnight with VV-GFP at an MOI of 10 in their respective oxygen tension conditions. 48 HPI, flow cytometry analysis was performed to determine GFP expression. We found that culture in 1.5% O2 with chronic stimulation significantly increased infection compared with culture in 20% O2 (figure 4F).

    We then asked if we removed the cell’s ability to respond to hypoxia could we alter infection. To do this, we isolated T cells from CD4creHif1af/f animals, either CD4Cre+ bearing the T cell-specific HIF1α deletion or CD4Cre- as littermate controls, and infected the T cells under the same conditions. We found in littermate controls under both chronic (figure 4G–H) and acute stimulation (figure 4I–J) that culture in 1.5% O2 increased infection as previously observed in wild-type animals. In cells lacking HIF1α however, infection was significantly reduced in the 1.5% O2 condition compared with littermate controls (figure 4G–J). Interestingly, this effect is more robust in the chronic stimulation setting than the acute, suggesting that while HIF1α and the hypoxia response are critical for the infection of T cells by VV-GFP, it is not the only factor. Increasing TCR stimulation has been shown to increase HIF1α expression in T cells,29 and as such the increased infection with chronic stimulation and hypoxia compared with acute may be a result of more HIF1α induction. Overall, these data suggest that hypoxia increases infectivity of VV.

    While we have previously infected at a range of MOIs, we chose to use the highest MOI tested (10) for these experiments as it leads to similar percentages (figure 1G–I) and MFI of VV-GFP (online supplemental figure S4I) in vitro culture as observed in the tumor. We do not know what MOI T cells are experiencing in the tumor but we believe that the MOI of 10 used in vitro may be physiologically relevant based on these data.

    VV infection of T cells in the tumor is required for treatment efficacy

    Finally, we wanted to determine whether T cell infection and death within the TME would be important for the therapeutic efficacy of OV treatment. To do this, we employed a mouse model which overexpresses the antiapoptotic protein Bcl2 in all T cells (CD4cre+/- Rosa26 LSL.BCL2.IRES.GFP).30 Bcl2 has been shown to protect from VV-induced cell death in macrophages31 but not T cells. To confirm that Bcl2 overexpression prevented VV-induced cell death, we harvested T cells from either CD4cre+ or CD4cre- littermate controls and infected with VV-GFP overnight at an MOI of 10 in 1.5% oxygen. We next counted live and dead cells by trypan blue over time and found that while there was a similar percentage of viability in the mock-infected controls (figure 5A), the VV-infected Bcl2 expressing T cells were significantly more viable at all timepoints compared with the wild-type littermate control cells (figure 5A). By 72 HPI 80% of wild-type cells were dead following VV infection compared with only 50% of Bcl2 overexpressing cells. While this was not full protection from VV-induced cell death, as vaccinia can also induce other cell death pathways, this was a significant increase that allowed for the study of the importance of OV-induced T cell death.

    Figure 5
    Figure 5

    Reducing vaccinia virus (VV)-mediated T cell death in the tumor reduces tumor clearances. (A) T cells isolated from CD4Cre+/- Rosa26LSL.Bcl2.IRES.GFP mice, which overexpress the antiapoptotic marker Bcl2 in all T cells, were stimulated overnight with αCD3, αCD28, and IL2. 24 hours post-stimulation cells were either mock infected or infected with VV at MOI of 10 in 1.5% oxygen overnight. The next morning virus was removed, cells were plated at 1×106 per condition, and cultured at 20% oxygen with IL2. Every 24 hours live and dead cells were counted with trypan blue. Percent viability from these counts is displayed. (B) Tumor growth curve of CD4cre+ Rosa26LSL.Bcl2.IRES.GFP mice or CD4Cre- Rosa26LSL.Bcl2.IRES.GFP littermate controls implanted with MEER tumors and treated IT with PBS or VV-GFP at 2.5×106 PFU/mouse (black arrow). Tumor size (B) and survival (C) were monitored until tumors reached 20 mm in any direction. Representative cytograms and quantifications of percentages and cell counts (D) CD4 and CD8 T cells, (E) Foxp3− Tconv and Foxp3+ Treg cells and (F) PD1 and Tim3 expression on CD8+ cells on day 4 post treatment with PBS or VV-GFP (2.6×106 PFU/mouse) in CD4cre+ Rosa26LSL.Bcl2.IRES.GFP mice. Data represent 3 (A–C) or 2 (D–F) independent experiments. Each point or line represents an individual mouse. *P<0.05, **p<0.01, ***p<0.001, ****p<0.0001 by two-way analysis of variance (ANOVA) with Sidak’s multiple comparisons test (A), Mantel-Cox analysis (C), unpaired t-test (D–F), or one-way ANOVA with Sidak’s multiple comparisons test (F). ns, non-significant. Error bars indicate SEMs. WT=CD4Cre- Rosa26LSL.Bcl2.IRES.GFP; Bcl2=CD4cre+ Rosa26LSL.Bcl2.IRES.GFP; HPI, hours post infection; CR, complete response.

    We next implanted CD4cre+ mice or CD4cre- littermate controls with MEER tumors and treated with a single IT dose of PBS control or VV-GFP (2.5×106 PFU/mouse) and monitored for tumor growth and survival. Previous literature exploring the role of Bcl2 in tumor-infiltrating T cells suggests that it may improve response to immunotherapy through increasing survival in the tumor, increasing anti-tumor efficiency, and maintaining a progenitor-like phenotype which is crucial for anti-tumor responses to immunotherapy.32–35 However, while we observed that 64% of CD4Cre- littermate controls (WT) treated with virus completely cleared their tumors as previously described,17 mice overexpressing Bcl2 in their T cells performed worse, only clearing tumors 25% of the time (figure 5B). This also led to a significant reduction in survival in the Bcl2 overexpressing animals treated with VV compared with littermate controls (figure 5C). We observe no significant changes in total CD4+ or CD8+ counts and percentages in the tumor with VV treatment (figure 5D). We find significantly higher counts and percentages of Treg cells (figure 5E) and a trend towards an increase in counts of Texh cells (figure 5F) after treatment with VV in the Bcl2 overexpressing animals suggesting that Bcl2 overexpression protects these cells from VV-induced cell death in vivo as well. These data suggest that the loss of exhausted or regulatory T cells from the TME early after VV treatment is an important part of the VV mechanism and that sustaining these cells is detrimental to the anti-tumor efficacy of the virus.

    Discussion

    Many OVs are targeted for tumor specificity through mechanisms that reduce viral proliferation in non-proliferating cells; however, they are not targeted through specific entry or binding mechanisms. This means that they may still be able to enter and replicate in other cells within the TME. In fact, many studies have shown that OVs are capable of infecting immune cells.1 12–14 We have shown that oncolytic vaccinia virus can infect multiple populations in the tumor, predominantly within the T cell fraction, which is highly proliferative and expresses thymidine kinase. Within the T cell fraction, we observed that most infected cells were Texh and Treg cells, both populations that reside in hypoxia. We showed through in vitro experiments that T cells lacking HIF1α had significantly lower levels of infection compared with wild-type cells. VV produces a protein, C16, which stabilizes HIF1α, to alter metabolism and improve intracellular conditions for its replication.25 26 We show that Treg and Texh cells are experiencing greater hypoxia than other T cells in the tumor and as such would likely have greater HIF1α induction. In this way, these cells are likely primed for viral replication and allow for faster and greater viral replication than others. We found that infection of T cells was productive by tittering supernatant from in vitro cultured infected cells. At a high MOI, this was similar to the amount of virus produced by HeLa cells which suggests that depending on the amount of virus T cells are exposed to in the tumor they may serve as an important viral reservoir in the tumor and may contribute to the viral spread within the TME.

    We also found that subsets of other cell populations were infected, specifically NK cells, MDSCs and monocytes. While we observed GFP signal in dendritic cells and macrophages as well, these cells undergo significant phagocytosis and as such it is uncertain if we are observing infection or these cells picking up viral GFP protein from other lysing cells. However, none of these cell types were as enriched in the GFP+ population as CD8+ T cells. The importance of the infection of these other cells is currently unknown and further studies are needed to determine the effects of viral infection on innate immune cells as well as within other tumor models where these cell types are more prevalent.

    We also show that T cells die after VV infection. Numerous studies have shown reduced Texh and Treg cells after treatment with an OV.1 17 18 36 Our data suggest a possible mechanism for this loss as, at least in the case of VV, death through infection-induced lysis. However, further studies are necessary to determine if all OVs infect the same cells in the tumor and if that infection leads to cell death. When cell death from VV infection in T cells was reduced through Bcl2 overexpression, we saw a significant reduction in tumor clearances and survival compared with littermate control mice after VV treatment. We observed that VV-treated Bcl2 overexpressing animals also had more Treg and Texh cells in their tumors by cell count, the opposite of what we observed in wild-type animals. In wild-type animals, superior tumor control may be a result of fewer suppressive or dysfunctional cells residing in the TME when a newly stimulated immune infiltrate enters the tumor. In this way, the oncolytic virus may be working similarly to the use of drugs like cyclophosphamide, which preferentially reduces cycling T cells like Treg or Texh cells.37 Another possible mechanism is that preventing the death of these cells reduces the amount of virus released into the tumor and reduces viral titers within the tumor overall. The focus of future studies will be to determine the exact mechanism of viral infection of T cells and improved therapeutic response.

    Another remaining question is whether other OVs infect immune cells within the tumor and what effect that has on therapeutic outcomes. It is critical to determine how common this ‘off-target’ effect may be, and if future OV studies should further engineer viruses towards or away from this characteristic. Most clinical studies with OVs use a sequential dosing regimen where patients receive multiple doses over multiple weeks. What effect, if any, does viral infection of lymphocytes have in the context of a dosing regimen like this? Would subsequent doses infect the newly stimulated immune response and lead to worse therapeutic responses as these cells are killed by VV infection? These data reveal a potentially important phenotype when treating with oncolytic vaccinia that requires follow-up studies to determine its importance to clinical studies.

    Supplemental material

    Data availability statement

    Data are available upon reasonable request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    All animal work and protocols in this study were approved by the University of Pittsburgh Institutional Animal Care and Use Committee, accredited by the AAALAC (Protocol # 23073380).

    References

    1. ↵
      1. Ramelyte E,
      2. Tastanova A,
      3. Balázs Z, et al
      . Oncolytic virotherapy-mediated anti-tumor response: a single-cell perspective. Cancer Cell 2021;39:394–406. doi:10.1016/j.ccell.2020.12.022
      OpenUrlCrossRef
    2. ↵
      1. VanSeggelen H,
      2. Tantalo DG,
      3. Afsahi A, et al
      . Chimeric antigen receptor-engineered T cells as oncolytic virus carriers. Mol Ther Oncolytics 2015;2:15014. doi:10.1038/mto.2015.14
      OpenUrlPubMed
    3. ↵
      1. Evgin L,
      2. Kottke T,
      3. Tonne J, et al
      . Oncolytic virus-mediated expansion of dual-specific CAR T cells improves efficacy against solid tumors in mice. Sci Transl Med 2022;14:eabn2231. doi:10.1126/scitranslmed.abn2231
    4. ↵
      1. Burke S,
      2. Shergold A,
      3. Elder MJ, et al
      . Oncolytic Newcastle disease virus activation of the innate immune response and priming of antitumor adaptive responses in vitro. Cancer Immunol Immunother 2020;69:1015–27. doi:10.1007/s00262-020-02495-x
      OpenUrl
    5. ↵
      1. Zhang Y,
      2. Huang W,
      3. Ornelles DA, et al
      . Modeling adenovirus latency in human lymphocyte cell lines. J Virol 2010;84:8799–810. doi:10.1128/JVI.00562-10
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Sánchez-Puig JM,
      2. Sánchez L,
      3. Roy G, et al
      . Susceptibility of different leukocyte cell types to Vaccinia virus infection. Virol J 2004;1:10. doi:10.1186/1743-422X-1-10
    7. ↵
      1. Torres-Domínguez LE,
      2. McFadden G
      . Poxvirus oncolytic virotherapy. Expert Opin Biol Ther 2019;19:561–73. doi:10.1080/14712598.2019.1600669
      OpenUrlCrossRefPubMed
    8. ↵
      1. Zandberg DP,
      2. Menk AV,
      3. Velez M, et al
      . Tumor hypoxia is associated with resistance to PD-1 blockade in squamous cell carcinoma of the head and neck. J Immunother Cancer 2021;9:e002088. doi:10.1136/jitc-2020-002088
    9. ↵
      1. Williams R,
      2. Lee DW,
      3. Elzey BD, et al
      . Preclinical models of HPV+ and HPV- HNSCC in mice: an immune clearance of HPV+ HNSCC. Head Neck 2009;31:911–8. doi:10.1002/hed.21040
      OpenUrlCrossRefPubMed
    10. ↵
      1. Scharping NE,
      2. Rivadeneira DB,
      3. Menk AV, et al
      . Mitochondrial stress induced by continuous stimulation under hypoxia rapidly drives T cell exhaustion. Nat Immunol 2021;22:205–15. doi:10.1038/s41590-020-00834-9
      OpenUrlCrossRefPubMed
    11. ↵
      1. McCart JA,
      2. Ward JM,
      3. Lee J, et al
      . Systemic cancer therapy with a tumor-selective vaccinia virus mutant lacking thymidine kinase and vaccinia growth factor genes. Cancer Res 2001;61:8751–7.
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Villa NY,
      2. Wasserfall CH,
      3. Meacham AM, et al
      . Myxoma virus suppresses proliferation of activated T lymphocytes yet permits oncolytic virus transfer to cancer cells. Blood 2015;125:3778–88. doi:10.1182/blood-2014-07-587329
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Ong HT,
      2. Hasegawa K,
      3. Dietz AB, et al
      . Evaluation of T cells as carriers for systemic measles virotherapy in the presence of antiviral antibodies. Gene Ther 2007;14:324–33. doi:10.1038/sj.gt.3302880
      OpenUrlCrossRefPubMed
    14. ↵
      1. Pipperger L,
      2. Riepler L,
      3. Kimpel J, et al
      . Differential infection of murine and human dendritic cell subsets by oncolytic vesicular stomatitis virus variants. Oncoimmunology 2021;10:1959140. doi:10.1080/2162402X.2021.1959140
    15. ↵
      1. Hoover AC,
      2. Spanos WC,
      3. Harris GF, et al
      . The role of human papillomavirus 16 E6 in anchorage-independent and invasive growth of mouse tonsil epithelium. Arch Otolaryngol Head Neck Surg 2007;133:495–502. doi:10.1001/archotol.133.5.495
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Najjar YG,
      2. Menk AV,
      3. Sander C, et al
      . Tumor cell oxidative metabolism as a barrier to PD-1 blockade immunotherapy in melanoma. JCI Insight 2019;4:e124989. doi:10.1172/jci.insight.124989
    17. ↵
      1. DePeaux K,
      2. Rivadeneira DB,
      3. Lontos K, et al
      . An oncolytic virus-delivered TGFβ inhibitor overcomes the immunosuppressive tumor microenvironment. J Exp Med 2023;220:e20230053. doi:10.1084/jem.20230053
    18. ↵
      1. Rivadeneira DB,
      2. DePeaux K,
      3. Wang Y, et al
      . Oncolytic Viruses Engineered to Enforce Leptin Expression Reprogram Tumor-Infiltrating T Cell Metabolism and Promote Tumor Clearance. Immunity 2019;51:548–60. doi:10.1016/j.immuni.2019.07.003
      OpenUrlCrossRefPubMed
    19. ↵
      1. Takahashi M,
      2. So TY,
      3. Chamberlain-Evans V, et al
      . Intratumoral antigen signaling traps CD8+ T cells to confine exhaustion to the tumor site. Sci Immunol 2024;9:eade2094. doi:10.1126/sciimmunol.ade2094
    20. ↵
      1. Greseth MD,
      2. Traktman P
      . The Life Cycle of the Vaccinia Virus Genome. Annu Rev Virol 2022;9:239–59. doi:10.1146/annurev-virology-091919-104752
      OpenUrl
    21. ↵
      1. Taylor CT,
      2. Scholz CC
      . The effect of HIF on metabolism and immunity. Nat Rev Nephrol 2022;18:573–87. doi:10.1038/s41581-022-00587-8
      OpenUrlCrossRef
    22. ↵
      1. Tao JH,
      2. Barbi J,
      3. Pan F
      . Hypoxia-inducible factors in T lymphocyte differentiation and function. A Review in the Theme: cellular Responses to Hypoxia. Am J Physiol Cell Physiol 2015;309:C580–9. doi:10.1152/ajpcell.00204.2015
      OpenUrlCrossRefPubMed
    23. ↵
      1. Palazon A,
      2. Tyrakis PA,
      3. Macias D, et al
      . An HIF-1α/VEGF-A Axis in Cytotoxic T Cells Regulates Tumor Progression. Cancer Cell 2017;32:669–83. doi:10.1016/j.ccell.2017.10.003
      OpenUrlCrossRefPubMed
    24. ↵
      1. Liikanen I,
      2. Lauhan C,
      3. Quon S, et al
      . Hypoxia-inducible factor activity promotes antitumor effector function and tissue residency by CD8+ T cells. J Clin Invest 2021;131:e143729. doi:10.1172/JCI143729
    25. ↵
      1. Mazzon M,
      2. Peters NE,
      3. Loenarz C, et al
      . A mechanism for induction of A hypoxic response by vaccinia virus. Proc Natl Acad Sci U S A 2013;110:12444–9. doi:10.1073/pnas.1302140110
      OpenUrlAbstract/FREE Full Text
    26. ↵
      1. Mazzon M,
      2. Castro C,
      3. Roberts LD, et al
      . A role for vaccinia virus protein C16 in reprogramming cellular energy metabolism. J Gen Virol 2015;96:395–407. doi:10.1099/vir.0.069591-0
      OpenUrlCrossRefPubMed
    27. ↵
      1. Vignali PDA,
      2. DePeaux K,
      3. Watson MJ, et al
      . Hypoxia drives CD39-dependent suppressor function in exhausted T cells to limit antitumor immunity. Nat Immunol 2023;24:267–79. doi:10.1038/s41590-022-01379-9
      OpenUrl
    28. ↵
      1. Wang B,
      2. Zhao Q,
      3. Zhang Y, et al
      . Targeting hypoxia in the tumor microenvironment: a potential strategy to improve cancer immunotherapy. J Exp Clin Cancer Res 2021;40:24. doi:10.1186/s13046-020-01820-7
    29. ↵
      1. Nakamura H,
      2. Makino Y,
      3. Okamoto K, et al
      . TCR engagement increases hypoxia-inducible factor-1 alpha protein synthesis via rapamycin-sensitive pathway under hypoxic conditions in human peripheral T cells. J Immunol 2005;174:7592–9. doi:10.4049/jimmunol.174.12.7592
      OpenUrlAbstract/FREE Full Text
    30. ↵
      1. Knittel G,
      2. Liedgens P,
      3. Korovkina D, et al
      . B-cell-specific conditional expression of Myd88p.L252P leads to the development of diffuse large B-cell lymphoma in mice. Blood 2016;127:2732–41. doi:10.1182/blood-2015-11-684183
      OpenUrlAbstract/FREE Full Text
    31. ↵
      1. Humlová Z,
      2. Vokurka M,
      3. Esteban M, et al
      . Vaccinia virus induces apoptosis of infected macrophages. J Gen Virol 2002;83:2821–32. doi:10.1099/0022-1317-83-11-2821
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Lee YG,
      2. Guruprasad P,
      3. Ghilardi G, et al
      . Modulation of BCL-2 in Both T Cells and Tumor Cells to Enhance Chimeric Antigen Receptor T-cell Immunotherapy against Cancer. Cancer Discov 2022;12:2372–91. doi:10.1158/2159-8290.CD-21-1026
      OpenUrl
    33. ↵
      1. Charo J,
      2. Finkelstein SE,
      3. Grewal N, et al
      . Bcl-2 overexpression enhances tumor-specific T-cell survival. Cancer Res 2005;65:2001–8. doi:10.1158/0008-5472.CAN-04-2006
      OpenUrlAbstract/FREE Full Text
    34. ↵
      1. Grayson JM,
      2. Zajac AJ,
      3. Altman JD, et al
      . Cutting Edge: Increased Expression of Bcl-2 in Antigen-Specific Memory CD8+ T Cells. J Immunol 2000;164:3950–4. doi:10.4049/jimmunol.164.8.3950
      OpenUrlAbstract/FREE Full Text
    35. ↵
      1. Chen Z,
      2. Ji Z,
      3. Ngiow SF, et al
      . TCF-1-Centered Transcriptional Network Drives an Effector versus Exhausted CD8T Cell-Fate Decision. Immunity 2019;51:840–55. doi:10.1016/j.immuni.2019.09.013
      OpenUrlCrossRefPubMed
    36. ↵
      1. Feist M,
      2. Zhu Z,
      3. Dai E, et al
      . Oncolytic virus promotes tumor-reactive infiltrating lymphocytes for adoptive cell therapy. Cancer Gene Ther 2021;28:98–111. doi:10.1038/s41417-020-0189-4
      OpenUrlPubMed
    37. ↵
      1. Amini L,
      2. Silbert SK,
      3. Maude SL, et al
      . Preparing for CAR T cell therapy: patient selection, bridging therapies and lymphodepletion. Nat Rev Clin Oncol 2022;19:342–55. doi:10.1038/s41571-022-00607-3
      OpenUrlCrossRefPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • X @DelgoffeLab

    • Contributors KD performed experiments unless otherwise noted, analyzed data, and wrote the manuscript. WGG performed in vitro titer experiments. DBR assisted with experiments and edited the manuscript. GMD conceived of the study, oversaw the research, aided in data interpretation, obtained study funding, and edited the manuscript. GMD is the guarantor.

    • Competing interests KD, DBR, and GMD are inventors on various patent applications around the use of oncolytic viruses for cancer therapy and are thus entitled to licensing fees based on success of such agents. GMD is on the scientific advisory board of Kalivir Immunotherapeutics.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.