CD30^−/− mice (22) generated on the 129 background and extensively backcrossed to C57BL/6 (B6), mice were kindly provided by Tak W. Mak (Ontario Cancer Institute, Toronto). These mice are now available from Jackson Laboratories (Bar Harbor, ME, USA). We analyzed the CD30^−/− mice by SNP analysis (performed by The Center for Phenogenomics, Toronto, ON, Canada) and found them to be 96% similar to Charles River B6 mice across 1,200 SNPs. The mice were further backcrossed to B6 mice purchased from Charles River (Wilmington, MA, USA) to generate F2 littermates for experiments. For influenza experiments, male mice (age 5–6 weeks) were immunized with 30 µL of influenza A/PR8 or A/HK-X31 at the indicated doses by intranasal (i.n.) infection while anesthetized with isofluorane. Initial influenza experiments were done in non-littermate mice with all experiments except those at day 100 post-influenza infection confirmed with littermate controls. For PR8 infections, mice were monitored closely with weights monitored daily and were euthanized when moribund. For the chronic infection model, female littermate mice were infected intravenously with 2 × 10⁶ ffu of LCMV clone 13, provided by Michael B.A. Oldstone (Scripps Research Institute, San Diego, CA, USA). All mice were housed in sterile micro-isolator cages under specific pathogen-free conditions. This study was carried out in accordance with the recommendations of the Canadian Council on Animal Care. All animal procedures were conducted as approved by the University of Toronto Animal care committee (animal protocol permit number 200111642).

Splenocytes from CD30^+/+ and CD30^−/− B6 mice were stimulated in vitro with 1 µg/mL of plate-bound anti-CD3 (145-2C11) and 10 µg/ml of soluble anti-CD28 (37.51), and expression of CD30 was assessed by flow cytometry after 24, 48, and 72 h of treatment.

Spleen, mediastinal lymph node (MLN), and lungs were harvested. Lungs were perfused with PBS, and lymphocytes were enriched by isolation over an 80/40% Percoll gradient. Single-cell suspensions were prepared from all organs through mechanical disruption of the tissue with a collagenase digestion step in some experiments and then stained for flow cytometry. MHC class I-peptide monomers were obtained from the National Institute for Allergy and Infectious Diseases tetramer facility (Emory University, Atlanta, GA, USA) and conjugated to streptavidin–allophycocyanin (Prozyme, purchased through Cedarlane, ON, Canada). For intracellular cytokine staining following influenza infection, lung cells and splenocytes were restimulated ex vivo with 1 µM of the MHC I-restricted NP_366-74 peptide for 6 h with GolgiStop (BD Biosciences) at 37°C. For the LCMV experiments, splenocytes were restimulated in the same way with either MHC I-restricted GP_33–41 or NP_396–404 peptides or MHC II-restricted GP_60–81 peptide. Cells were surfaced stained, fixed, permeabilized, and stained intracellularly for specific cytokines as indicated in the figures. Unstimulated samples (no peptide) were used as negative controls. Samples were analyzed using a FACScalibur (BD Biosciences) or LSRFortessa (BD Biosciences) and FlowJo (TreeStar Inc, Ashland, OR, USA) software.

The antibodies used in this study are as follows: anti-mouse CD30 (clone: mCD30.1) (BD Biosciences, San Jose, CA, USA), anti-mouse CD8α (clone: 53-6.7) (eBioscience, San Diego, CA, USA; BioLegend, San Diego, CA, USA), anti-mouse CD4 (clone: GK1.5) (eBioscience, BioLegend), anti-mouse CD3ε (clone: 145-2C11) (eBioscience, BioLegend), anti-mouse CD44 (clone: IM7) (eBioscience), anti-mouse T-bet (clone: 4B10) (eBioscience), anti-mouse Foxp3 (clone: FJK-16s) (eBioscience), anti-mouse Bcl6 (clone: BCL-DWN) (eBioscience), anti-mouse PD-1 (clone: J43) (eBioscience), anti-mouse CXCR5 (clone:SPRCL5) (eBioscience), anti-mouse CD62L (clone: MEL-14) (eBioscience), anti-mouse IFNγ (clone: XMG1.2) (eBioscience), anti-mouse Tim-3 (clone: RMT3-23) (eBioscience), and anti-mouse CD25 (clone: PC61.5) (eBioscience).

Organs were immediately frozen at −80°C upon harvest. For viral load, organs were thawed and homogenized, and the supernatant collected to perform dilutions (10⁰- to10⁵-fold) for infection of an MC57 cell monolayer under a 2% methylcellulose-MEM overlay. MC57 monolayers were fixed with 4% paraformaldehyde 48 h later, permeabilized with 1% Triton X-100, and stained with Rat anti-LCMV mAb (VL-4). Secondary staining with Goat anti-Rat HRP and o-Phenylenediamine (Sigma-Aldrich, Oakville, ON, Canada) was used to induce a colorimetric reaction to label LCMV-infected foci.

Statistical analysis was performed using GraphPad Prism 6 (San Diego, CA, USA), with the specific test performed as indicated in the figure legends.

To examine the expression of CD30 on T cells, we stimulated splenocytes from CD30^+/+ and CD30^−/− littermate mice with anti-CD3 and anti-CD28 antibodies in vitro. CD30 was undetectable on resting T cells but induced on CD8 T cells by 48 h of stimulation and on both CD4 and CD8 T cells after 72 h (Figure 1A), consistent with the literature (35). CD30^−/− splenocytes were used as a staining control and showed only background levels of staining. Having demonstrated CD30 expression on activated T cells, we next asked whether CD30 could be detected in vivo, during a viral infection. Previous work has shown that the inducible TNFR family member 4-1BB is readily detectable on murine T cells in the lung at 6–8 days post-infection with a sublethal dose of influenza A/PR8 (PR8) (27). Therefore, we analyzed CD30 expression at similar time points post intranasal (i.n.) PR8 infection. CD30 was significantly expressed on Treg cells and marginally on Th1 cells but undetectable on antigen-specific CD8 T cells at day 9 pi (Figure 1B). CD30 was not detected at days 3, 5, or 7 in the lung and draining MLN on antigen-specific CD8 T cells, Th1, T follicular helper (Tfh), T follicular regulatory (Tfr), or Treg cells (data not shown, n = 2–3 mice per time point). These results confirm that CD30 is expressed during influenza virus infection, albeit on a limited subset of cells, prompting us to examine the effect of CD30 on immunity to influenza virus.

In pilot experiments, CD30 was not required for mouse to survive influenza A/PR8 infection and the CD8 T-cell responses to sublethal influenza A/PR8 at day 10 post-infection in CD30^+/+ and CD30^−/− littermates were comparable (data not shown). Therefore, we switched to a milder strain of influenza, Influenza A/HK/X31 (X31, an H3N2 virus) (36), with the idea that a weaker infection might be more costimulation-dependent. CD30^+/+ and CD30^−/− mice were infected i.n. with influenza X31 and, the antigen-specific CD8 T-cell response to the immunodominant epitope NP_366-74 was assessed using D^b/NP_366-74 MHC class I tetramers at day 10 (the peak of the primary CD8 T-cell response). Assessment of the frequency and absolute number of NP_366-74-specific CD8 T cells in the spleen, MLN, and lung showed comparable primary expansion of influenza-specific CD8 T cells in CD30^+/+ and CD30^−/− mice (Figure 2A). The tetramer⁺ cells were CD62L^lo in both groups, indicating an effector phenotype (Figure 2A). Therefore, CD30 is dispensable for primary CD8 T-cell expansion to influenza virus.

In vitro studies have shown that CD30 stimulation of T cells can enhance their production of IFN-γ, among other cytokines (6). Therefore, we assessed IFN-γ production by ex vivo restimulation with NP_366-74 peptide at day 10 and day 100 post-X31 infection in CD30^+/+ and CD30^−/− mice and found no significant difference in the proportion of IFN-γ producing CD8 T cells at day 10 or 100 (Figure 2B). Thus, CD30 does not play a discernable role in the function of effector and memory CD8 T cells during influenza infection.

Despite the lack of an obvious role for CD30 in CD8 T-cell responses to influenza A X31, it was possible that CD30 could influence protective memory against a more severe influenza infection, such as induced by influenza A/PR8. To this end, we infected mice with influenza X31 and allowed them to naturally clear the virus and develop a memory response, then challenged the mice at day 30 with a sublethal dosage of influenza A/PR8, which typically causes a 20–25% weight loss in naive mice. PR8 shares the same NP epitope as the initial priming X31 strain but contains different HA and NA proteins (PR8 is H1N1, whereas X31 is H3N2), thereby allowing us to assess protection from CD8 T-cell immunity without the interference of neutralizing antibody responses. CD30^+/+ and CD30^−/− mice exhibited protection to the challenge, losing only about 5% of their body weight before making a full recovery, and exhibited comparable weight loss (Figure 2C). CD30^+/+ and CD30^−/− mice also showed equivalent secondary expansion of NP-specific CD8 T cells in the spleen, lung, and MLN (Figure 2D). IFN-γ production by CD8 T cells upon ex vivo peptide restimulation was assessed, and although mean fluorescent intensities (MFIs) were not different, CD30^−/− mice had a greater frequency of IFN-γ-producing CD8 T cells than their CD30^+/+ littermates (Figure 2E), despite disease outcome being comparable. Taken together, these data show that CD30 is dispensable for the establishment and programing of secondary recall responses to influenza virus and does not play a role in the disease outcome of a secondary infection.

A previous study by Nishimura et al. reported that early antigen-specific CD8 T-cell responses to Listeria monocytogenes infection in CD30L^−/− mice are intact but the memory pool assessed at 84 days post-infection was defective compared to wild-type controls (20). This implies a role for the CD30-CD30L pathway in the generation or maintenance of CD8 T-cell memory. Therefore, we investigated whether CD30 is necessary for the establishment of a long-lived CD8 memory T-cell pool following influenza infection. As the bone marrow is a known reservoir of memory T cells and a preferential organ for CD8 memory T-cell homeostasis, we analyzed this organ as well. Mice were infected with influenza X31 and examined for NP_366–374-specific CD8 T cells 100 days after infection. There were no significant differences observed between CD30^+/+ and CD30^−/− mice in the spleen and bone marrow at this much later time point (Figure 2F). Therefore, CD30 is dispensable for the generation and maintenance of CD8 memory T cells to influenza virus.

Given the expression of CD30 on lung Treg cells during peak influenza response (Figure 1B), we evaluated the frequency and absolute numbers of Treg cells in CD30^+/+ and CD30^−/− mice at day 9 post-PR8 infection and found no difference in the lung (Figure 3A) despite detecting expression at this time point. Th1 cells, which have only marginal expression of CD30 in the lung (Figure 1B), were also evaluated and comparable populations found as well (Figure 3B). Similarly, no differences in Tfh or Tfr frequencies and absolute numbers were found when the draining lymph node was examined at peak response (Figures 3C,D). Taken together, the data show that CD30 is dispensable for both CD8 and CD4 T-cell expansion during acute influenza infection.

It was possible that CD30 is more important in a persistent as compared to an acute rapidly cleared infection. Therefore, we examined the response of CD30^+/+ and CD30^−/− littermates to LCMV clone 13, an infection in which viral clearance takes several months. This chronic infection is characterized by the hierarchical loss of effector functions and peripheral exhaustion of virus-specific T cells (37). We assessed T-cell responses at day 21 post-infection, a time point at which T cells clearly exhibit signs of exhaustion. Frequencies and absolute numbers of antigen-specific CD8 T cells in the spleen, as measured by MHC tetramers containing GP_33–41 and NP_396–404 peptides, were comparable between CD30^+/+ and CD30^−/− littermate mice (Figure 4A). Furthermore, the absence of CD30 did not significantly affect the levels of Tim-3 (Figure 4B) and PD-1 (Figure 4C) on the antigen-specific CD8 T cells, two markers whose persistent expression is typically associated with T-cell exhaustion. The overall frequency of PD-1-expressing tetramer⁺ cells was also comparable between CD30^+/+ and CD30^−/− littermates (Figure 4D). Ex vivo restimulation of splenic CD8 T cells with GP_33–41 or NP_396–404 peptides did not reveal functional differences in their ability to produce IFN-γ, both in terms of total frequencies of IFN-γ-producing CD8 T cells and per cell IFN-γ production as measured by the MFI (Figure 4E). Viral loads assessed in the spleen, lung, and kidney at day 21 also did not differ between CD30^+/+ and CD30^−/− littermate mice (Figure 4F). These results suggest that CD30 is dispensable for the CD8 T-cell response and control of LCMV clone 13 infection.

We next assessed the role of CD30 in the CD4 T-cell response during LCMV clone 13 infection. It was previously reported that mice depleted of CD4 T cells fail to clear the virus, implicating CD4 T cells in control of LCMV clone 13 the infection (38, 39). Analysis of Tfh, Tfr, and Treg populations revealed no difference in frequency or total number of cells (Figures 5A–D). Moreover, ex vivo restimulation with the GP_61–88 peptide did not reveal differences in CD4 T-cell production of IFN-γ, thus suggesting comparable Th1 responses (Figure 5E). Therefore, CD30 is dispensable for CD4 and CD8 T-cell responses in chronic LCMV clone 13 infection.

As CD30 is expressed at its highest levels on Treg cells during influenza infection and also expressed on T cells in human lymphoproliferative disorders, we asked whether CD30 contributed to the number of steady state T cells, including Treg cells over time. Strikingly, by 8 months of age, wild-type unmanipulated CD30^+/+ mice had significantly more T cells than their young counterparts (age 4–6 weeks) (Figures 6A–C). While naive young CD30^+/+ and CD30^−/− littermate mice did not show differences in T-cell populations in the spleen, aged CD30^−/− mice exhibited significantly fewer CD8 T cells (Figure 6A), CD4 T cells (Figure 6B), and Treg cells (Figure 6C) compared to their wild-type littermate controls. The lower numbers of Treg cells in older CD30^−/− mice likely reflects the lower number of CD4 T cells overall, as the frequency of CD25⁺Foxp3⁺ was comparable between CD30^+/+ and CD30^−/− littermates (Figures 6D,E). The MFI for CD25 also did not differ between CD30^+/+ and CD30^−/− littermates (Figure 6F). This age-associated expansion of the T-cell compartment seen in wild-type mice was completely mitigated in CD30^−/− littermates, independent of the size and cellularity of the spleens themselves, as the T-cell numbers between young and aged CD30^−/− mice were comparable.

To determine which T cells were expanded in aged CD30^+/+ compared to CD30^−/− mice, we conducted additional analysis of aged littermate mice to differentiate naive and memory T-cell subsets. We assessed CD62L⁺CD44^lo, CD62L⁺CD44^mid, CD62L⁺CD44^hi, CD62L⁻CD44^lo, CD62L⁻CD44^mid-hi subsets of CD4 and CD8 T cells by flow cytometry (Figures 7A–D). CD30^+/+ and CD30^−/− mice had comparable numbers of naive CD62L⁺CD44^lo CD8 and CD4 T cells. In contrast, the CD62L⁻CD44^mid-hi antigen-experienced effector or effector memory cells were significantly reduced in the knockout mice compared to wild-type littermate counterparts. As KLRG1 was absent from the majority of these T cells (data not shown), the increased T-cell expansions are largely due to memory T cells. This trend was also seen in the CD62L⁺CD44^hi central memory T-cell compartment, although it did not reach statistical significance. Thus, CD30 contributes to the expansion of memory T cells in older mice.