Cyclophilin D Regulates Antiviral CD8⁺ T Cell Survival in a Cell-Extrinsic Manner

Cyclophilin D (CypD) is a member of cyclophilin protein family, which is encoded by the peptidyl-prolyl isomerase F (Ppif) gene and has peptidyl-prolyl-cis-trans-isomerase activity (1–4). CypD is located within the mitochondrial matrix and plays a critical modulatory role in the mitochondria permeability transition pore (MPTP) (1, 2). Various stresses can trigger a conformational change of the MPTP, a transition that is facilitated by CypD, converting the pore to an open state (5). Upon pore opening, the mitochondria swell, causing membrane rupture and ultimately cell death via apoptosis or necrosis (5). CypD-mediated cell death has been heavily implicated in pathologies of the liver, heart, and lungs, and thus identification of drugs inhibiting CypD function, such as cyclosporine A, have drawn interest as potential therapies (5). Although CypD has recently been shown to be an intrinsic negative regulator of CD8⁺ T cell responses to intracellular Mycobacterium tuberculosis bacterial infection (6), its role in antiviral immunity is currently unknown.

Cytotoxic CD8⁺ T cells are critical for the control and elimination of intracellular pathogens such as viruses. To become fully activated, a naive CD8⁺ T cell must receive three signals. The first signal (signal 1), requires the interaction between an Ag-specific CD8⁺ T cell bearing the correct TCR and an APC such as a dendritic cell (DC) presenting cognate Ag on MHC class I (7–11). The second signal (signal 2) comprises the interaction of costimulatory molecules present on mature DCs such as CD80/CD86 and CD28 on the T cell (10, 12–15). Finally, inflammatory cytokines present at the time of activation, such as type I IFN (IFN-I), provide the third signal (signal 3) and are required for optimal CD8⁺ T cell proliferation, accumulation, and generation of an effective CD8⁺ T cell effector response (16–23). Upon receiving these three signals, the naive Ag-specific CD8⁺ T cell undergoes clonal expansion and acquires the ability to produce effector cytokines (such as IFN-γ and TNF-α) and cytolytic molecules to control the replicating pathogen (24–32).

Given the critical role of CypD in mitochondrial integrity and the recent demonstration of CypD’s role in T cell responses to M. tuberculosis bacterial infection (6) in which it acts in a cell-intrinsic manner to restrain T cell expansion, we asked whether deficiency in CypD impacts the CD8⁺ T cell response to acute virus infection. In this study, we demonstrate that CypD plays an important cell-extrinsic role in promoting Ag-specific CD8⁺ T cell survival. In addition, we observed that transfer of CypD-expressing macrophages in CypD-deficient hosts can partially rescue Ag-specific CD8⁺ T cell numbers. This establishes a novel T cell–extrinsic role for CypD in antiviral immunity and highlights the potential adverse effects of administering CypD inhibitory drugs, such as cyclosporine A, as host immunity to certain viruses may be compromised.

C57BL/6 wild-type (WT), CypD-deficient (Ppif^−/−), and WT P14 TCR–transgenic (specific for lymphocytic choriomeningitis virus [LCMV]–derived GP_33–41 epitope) (33) mice were bred at McGill University, and both male and female mice were used at 6–8 wk of age. LCMV Armstrong (LCMV-Arm) strain was kindly provided by John Harty (The University of Iowa) and propagated as described (19, 34, 35). Mice were infected with 2 × 10⁵ PFU (injection i.p. route) and housed at the appropriate biosafety level, and the morbidity was monitored at indicated time points.

Peptides were synthesized by either Bio Basic Canada (GP_276–286 [SGVENPGGYCL], NP_396–404 [FQPQNGQFI]) or Bio-Synthesis (GP_33–41 [KAVYNFATM]). H-2D^b MHC class I tetramers (D^b NP_396–404, D^b GP_33–41, or D^b GP_276–286) were prepared as previously described (36). The following Abs were used in an appropriate combination of fluorochromes: CD3 (clone 17A2; eBioscience), CD8α (clone 53-6.7; BioLegend), CD11a (clone M17/4; BioLegend), CD11b (clone M1/70; eBioscience), CD11c (clone N418; eBioscience), CD19 (clone MB19-1; eBioscience), CD40 (clone 1C10; eBioscience), CD80 (clone 16-10A1; eBioscience), CD86 (clone GL1; eBioscience), F4/80 (clone Bm8; eBioscience), IFN-γ (clone XMG1.2; eBioscience), MHC class II (MHCII) (clone M5/114.15.2; eBioscience), NK1.1 (clone PK136; eBioscience), TNF-α (clone MP6-XT22; BioLegend), and Ki67 (clone 16A8; BioLegend). Proliferation and cell death assays were performed according to the specifications of the manufacturer: Phase-Flow BrdU Cell Proliferation Kit (BioLegend), Vybrant FAM Caspase-3 and -7 Assay Kit (Molecular Probes), and FITC Annexin V apoptosis Detection Kit II (BD Biosciences).

WT congenically marked (Thy1.1) P14 TCR–transgenic CD8⁺ T cells were obtained from the peripheral blood of young naive P14 female mice, and 5000 cells were injected i.v. into either naive WT (Thy1.2) or naive Ppif^−/− (Thy1.2)–recipient mice. Quantification of endogenous response was done in the absence of adoptive transfer. Mice were infected with LCMV-Arm the following day (2 × 10⁵ PFU i.p.). Spleens were isolated on day 8 postinfection and processed into single-cell suspensions. RBCs were lysed with ACK buffer (ACK buffer: 150 mM NH₄Cl, 1 mM KHCO₃, 0.1 mM Na₂EDTA, dissolved in H₂O and adjusted to [pH 7.2–7.4]), and lymphocytes were stained with indicated Abs then fixed with IC Fixation Buffer (eBioscience). Total CD8⁺ T cells were identified with CD8α staining, total endogenous LCMV-specific responses were identified with CD8α CD11a staining (37, 38), endogenous LCMV-specific CD8⁺ T cells were identified by H-2D^b MHC class I tetramer (D^b NP_396–404, D^b GP_33–41, or D^b GP_276–286) CD8α staining or WT P14 TCR–transgenic CD8⁺ T cells were identified with Thy1.1 and CD8α staining. Samples were analyzed with a BD LSRFortessa Flow Cytometer (BD Biosciences) and FlowJo software (Tree Star).

On day 8 post–LCMV-Arm infection, spleens were harvested and processed into single-cell suspensions. RBCs were lysed with ACK buffer and samples were stimulated for 5.5 h at 37°C with 5% CO₂ in the presence or absence of peptide (NP_396–404, GP_33–41, or GP_276–286) and brefeldin A (BioLegend). The frequency of CD8⁺ T cells capable of producing cytokine after stimulation with indicated peptide was determined with intracellular staining for cytokines (IFN-γ, TNF-α) stained in Perm/Wash Buffer (eBioscience). Samples were analyzed with a BD LSRFortessa Flow Cytometer (BD Biosciences) and FlowJo software (Tree Star).

On day 3 postinfection, livers were harvested and homogenized, and the virus was quantitated using Vero cell plaque assay as previously described (19, 34, 35).

On day 3 postinfection, serum was collected via cardiac puncture from WT and Ppif^−/− mice and levels of total active IFN-I (both IFN-α and IFN-β) were assessed using cultured B16-blue IFN-α/-β reporter cell line (InvivoGen), according to the specifications of the manufacturer. Briefly, 25 μl of samples was incubated with 175 μl of B16 cells (850,000 cells/ml) at 37°C and 5% CO₂ overnight in a 96-well plate. Following, 25 μl of B16 cell supernatant was incubated at 37°C and 5% CO₂ with 175 μl of QUANTI-Blue medium (InvivoGen) and OD₆₂₀ readings were taken beginning at 1 h postincubation.

B16 cells were maintained in RPMI 1640 enriched with 10% FBS, 2 mM l-glutamine, and 100 U/ml of penicillin/streptomycin. One hundred micrograms per milliliter of Zeocin and 200 μg/ml of Normocin (InvivoGen) were added to the cell culture media for all subsequent passages and assays.

Bone marrow was extracted from the femur and tibia of WT and Ppif^−/− mice. Macrophages were differentiated from each mouse strain for 1 wk in bone marrow–derived macrophage (BMDM) media (BMDM media: RPMI 1640 supplemented with 10% FBS, 1% penicillin-streptomycin, 1% l-glutamine, 1% nonessential amino acids, 1% essential amino acids, 1% sodium pyruvate, 2% HEPES, 0.14% sodium hydroxide) containing 30% L929 cell culture supernatant (L929 supernatant contains M-CSF growth factor for macrophage differentiation). For rescue experiments, 1 × 10⁶ BMDM were injected i.v. into naive Ppif^−/−–recipient mice. Mice were infected with LCMV-Arm the following day (2 × 10⁵ PFU i.p.), and on day 8 postinfection, splenocytes were isolated and CD8⁺ T cell analysis was performed as described above.

Data were analyzed using GraphPad Prism6 software. Unpaired Student t tests were performed and p values < 0.05 were considered statistically significant. The p values are represented as follows: *p < 0.05, **p < 0.01, ***p < 0.001.

All animal procedures were carried out in accordance with the Canadian Council on Animal Care and were approved by the McGill University Animal Care Committee (Protocol no. 7554).

Before investigating whether CypD deficiency has an impact on CD8⁺ T cell antiviral immunity, we first examined total spleen cellularity and CD8⁺ T cell numbers from naive C57BL/6 WT and Ppif^−/− (CypD-deficient) mice. We observed no differences in total spleen cellularity (Fig. 1A) and no reduction in frequencies (Fig. 1B) and total numbers (Fig. 1C) of CD8⁺ T cells in naive Ppif^−/− mice compared with WT mice. Together, these results established there is no reduction in baseline CD8⁺ T cell numbers in naive CypD-deficient mice, demonstrating that at steady-state, CypD deficiency does not affect total CD8⁺ T cell numbers. Thus, this model can be used to study the role of CypD in CD8⁺ T cell responses to viral infection.

To determine whether CypD deficiency impacts antiviral immune responses, we infected naive (WT) and Ppif^−/− mice with LCMV-Arm and assessed morbidity and viral load. LCMV-Arm, an enveloped ambisense RNA virus that is a prototypic virus in the Arenaviridae family, is a natural mouse pathogen that establishes an acute and typically asymptomatic viral infection in mice (39) that is controlled by a robust CD8⁺ T cell response (40). As expected following infection with LCMV-Arm, C57BL/6 WT mice exhibited minimal weight loss that returned to preinfection weight levels on day 5 postinfection (Fig. 2A). In contrast, Ppif^−/− mice exhibited a slight but significant increase in morbidity with significant weight loss that did not return to preinfection weight levels during the 8-d time course (Fig. 2A). Importantly, the increased morbidity correlated with compromised viral control, as demonstrated by a significant increase in viral titers in the livers from day 3 post–LCMV-Arm–infected Ppif^−/− mice compared with WT mice (Fig. 2B, 5.8-fold). Together, these results suggested that deficiency in CypD compromises host responses, leading to a higher viral burden at an early time-point and increased morbidity following challenge with a virus that typically establishes a relatively asymptomatic infection.

CD8⁺ T cells are important for limiting LCMV-Arm infection–associated morbidity and for controlling viral burden, so we next assessed whether deficiency in CypD impacts Ag-specific antiviral CD8⁺ T cell responses. One advantage of using LCMV-Arm as a model pathogen is that multiple H-2D^b–restricted virus-specific CD8⁺ T cell epitopes have been identified in C57BL/6 WT mice with NP_396–404, GP_33–41, and GP_276–286 shown to be the three immunodominant epitopes (41–44). Splenocytes isolated from day 8 LCMV-Arm–infected WT and Ppif^−/− mice were stimulated in the presence of LCMV-derived peptides (NP_396–404, GP_33–41, or GP_276–286) and brefeldin A, and the frequency of CD8⁺ T cells capable of producing IFN-γ was determined via intracellular staining (Fig. 2C). Following ex vivo stimulation with NP_396–404 peptide, we observed no significant change in the frequencies of CD8⁺ T cells capable of producing the cytokine IFN-γ between WT and Ppif^−/− mice (Fig. 2C, 2D). However, total number of IFN-γ–producing NP_396–404–specific CD8⁺ T cells were significantly lower in Ppif^−/− mice compared with WT mice (Fig. 2E, 3.9-fold). In addition, the amount of IFN-γ produced on a per-cell basis (as measured by geometric mean fluorescence intensity [gMFI] for IFN-γ) after ex vivo stimulation was also significantly reduced in cells isolated from Ppif^−/− mice compared with WT mice (Fig. 2F). Ex vivo stimulation of splenocytes with GP_33–41 peptide revealed that Ppif^−/− mice harbored slightly higher frequencies of CD8⁺ T cells capable of producing IFN-γ (Fig. 2G). Nevertheless, we observed a substantial decrease in the overall magnitude of the IFN-γ producing GP_33–41–specific CD8⁺ T cell response (Fig. 2H, 3.0-fold) in Ppif^−/− mice compared with WT mice, as well as a reduction in the amount of IFN-γ produced on a per-cell basis in GP_33–41–specific CD8⁺ T cells (Fig. 2I). Similarly, following ex vivo GP_276–286 peptide stimulation, we observed no significant reduction in the frequency of CD8⁺ T cells able to produce IFN-γ (Fig. 2J) but a significant reduction in the total numbers of IFN-γ–producing GP_276–286–specific CD8⁺ T cells (Fig. 2K, 5.7-fold) and a significant decrease in the expression of IFN-γ produced on a per-cell basis (Fig. 2L) in cells from Ppif^−/− mice compared with WT mice. Together, these results demonstrated that deficiency in CypD is associated with a reduction in the overall number of CD8⁺ T cells specific for the three LCMV immunodominant epitopes examined and a reduction in their capacity to express IFN-γ on a per-cell basis.

Intracellular cytokine staining analyses revealed reduced functional LCMV-specific CD8⁺ T cells, but these assays cannot distinguish whether this is due to a reduction in the capacity of T cells to produce cytokines (e.g., exhaustion) versus impaired T cell expansion/survival in Ppif^−/− mice. Thus, we further evaluated the magnitude of the LCMV-specific CD8⁺ T cell response in WT and Ppif^−/− mice using MHC class I tetramer reagents. Splenocytes were isolated from day 8 LCMV-Arm–infected WT and Ppif^−/− mice and LCMV-Arm–specific CD8⁺ T cells were enumerated with H-2D^b MHC class I tetramers (D^b NP_396–404, D^b GP_33–41, or D^b GP_276–286) (Fig. 3A). Although the frequencies of both D^b NP_396–404– and D^b GP_33–41–specific CD8⁺ T cells were significantly increased in Ppif^−/− mice (Fig. 3B, 3D, respectively), we observed a significant reduction in the total number of D^b NP_396–404– (Fig. 3C, 2-fold) and D^b GP_33–41 (Fig. 3E, 2.5-fold)–specific CD8⁺ T cells compared with WT mice. The frequencies of D^b GP_276–286–specific CD8⁺ T cells were similar in the WT and Ppif^−/− mice (Fig. 3F); however, the total numbers of D^b GP_276–286–specific CD8⁺ T cells were reduced in Ppif^−/− mice (Fig. 3G, 4.1-fold). These results confirmed CypD also plays an important role in regulating the magnitude of expansion of three immunodominant LCMV-specific CD8⁺ T cell populations.

Our data highlighted that although the numbers of immunodominant epitope-specific CD8⁺ T cells was decreased, the overall frequencies of Ag-specific cells within the CD8⁺ T cell compartment was unchanged following infection. Thus, our data suggest that CypD deficiency broadly affects the accumulation of all Ag-specific CD8⁺ T cells following viral infection, resulting in a decrease in number without affecting the frequency of cells within the CD8⁺ T cell compartment responding to each epitope. To evaluate whether the absence of CypD globally impacts CD8⁺ T cell responses to LCMV infection, regardless of their Ag specificity, we enumerated the total number of Ag-experienced CD8⁺ T cells responding to LCMV-Arm infection using a surrogate marker approach. Previous studies have established that surrogate markers allow for the analysis of the total number of all Ag-experienced T cells responding to infection without a priori knowledge of Ag specificity (37, 38, 45, 46). For CD8⁺ T cells, Ag-experienced cells can be identified by the downregulation of CD8α (CD8α^lo) and up regulation of CD11a (CD11a^hi), which enables evaluation of the total CD8⁺ T cell response induced by infection (37, 38). Splenocytes were isolated from day 8 LCMV-Arm–infected WT and Ppif^−/− mice, and total numbers of Ag-experienced LCMV-specific CD8⁺ T cells were enumerated with CD8α and CD11a staining (Fig. 4A). In the bulk CD8⁺ T cell population, we observed significant reductions in the frequencies (Fig. 4B) and total numbers (Fig. 4C, 3.5-fold) of CD8⁺ T cells in Ppif^−/− mice compared with WT mice following LCMV-Arm infection. Furthermore, based on CD8α^lo CD11a^hi expression patterns, we found that although the frequencies of Ag-experienced CD8⁺ T cells were minimally lower (Fig. 4D), the total number of Ag-experienced CD8α^lo CD11a^hi CD8⁺ T cells was significantly decreased (Fig. 4E, 3.6-fold) in Ppif^−/− mice compared with WT mice following LCMV-Arm infection. The decreased cell numbers observed was attributed by a decrease in total spleen cellularity (Fig. 4F, 2.2-fold) and the reduction in CD8⁺ T cell numbers was the primary contributor, representing ∼70% of the overall reduction. Together, this suggests that CypD deficiency globally affects the accumulation of all Ag-specific CD8⁺ T cells during the course of viral infection.

In addition to the magnitude (the number of cells responding) of the CD8⁺ T cell response, the quality (diversity of function) of the responding T cells is also important for an effective antiviral immunity. Specifically, the production of multiple cytokines (polyfunctionality; i.e., the ability of a single cell to produce two or more cytokines) is associated with protective immune responses against various pathogens (47–51). Thus, we next examined whether deficiency in CypD affected CD8⁺ T cell polyfunctionality. Splenocytes were isolated from day 8 LCMV-Arm–infected WT and Ppif^−/− mice and stimulated in the presence of peptide (NP_396–404, GP_33–41, or GP_276–286) and brefeldin A, and the frequencies of stimulated CD8⁺ T cells capable of producing both IFN-γ and TNF-α (i.e., double cytokine–producing cells) were determined by intracellular staining. Ppif^−/− mice harbored significantly lower frequencies (Fig. 5A, 5B) and total numbers (Fig. 5C, 5.2-fold) of NP_396–404–specific CD8⁺ T cells expressing both IFN-γ and TNF-α. In addition, we observed a significant decrease in the amount of TNF-α produced on a per-cell basis (as measured by TNF-α gMFI) from Ppif^−/− cells compared with WT cells (Fig. 5D). Similar results were observed after ex vivo stimulation of splenocytes with either GP_33–41 or GP_276–286 peptide (Fig. 5E–J). Collectively, these results demonstrated that deficiency in CypD not only reduced the number of IFN-γ TNF-α double–producing LCMV-specific CD8⁺ T cells but also impacts their polyfunctionality on a per-cell basis, irrespective of the targeted viral epitope. Thus, CypD deficiency impaired the quality of CD8⁺ T cells, demonstrating that both CD8⁺ T cell accumulation and function may be compromised in the absence of CypD.

Thus far, our results have demonstrated that CypD deficiency compromises CD8⁺ T cell numbers and function. A recent study demonstrated that CypD elicits T cell–intrinsic effects to negatively regulate T cell expansion during mycobacterial infection (6). Because our data point to a positive role for CypD during acute viral infection, we next asked whether CypD plays this role in a T cell–intrinsic or –extrinsic manner. To address this, we used an adoptive transfer approach. Congenically marked WT TCR-transgenic P14 CD8⁺ T cells (CD8⁺ T cells specific for the LCMV-derived GP_33–41 epitope), which express CypD, were injected into either WT or Ppif^−/−–recipient mice. One day posttransfer, recipient mice were infected with LCMV-Arm, and WT P14 CD8⁺ T cell expansion was examined on day 8 postinfection. We observed a significant reduction in the frequencies (Fig. 6A, 6B) and total numbers (Fig. 6C, 8.3-fold) of WT P14 CD8⁺ T cells recovered in the spleens of Ppif^−/− mice compared with WT mice on day 8 postinfection. Given that donor-derived P14 T cells express CypD, these data showed that the accumulation defect of LCMV-specific CD8⁺ T cells in CypD-deficient mice is due, at least in part, to CD8⁺ T cell–extrinsic factors.

Because IFN-Ιs (i.e., IFN-α and IFN-β) are important for CD8⁺ T cell–mediated antiviral immunity, we reasoned that these inflammatory cytokines (i.e., signal 3) may be reduced in mice lacking CypD, potentially explaining the cell-extrinsic impairment in CD8⁺ T cell accumulation following LCMV-Arm infection. To address this, we collected serum from WT and Ppif^−/− mice at day 3 postinfection with LCMV-Arm and examined bioactive IFΝ-Ι levels with a reporter cell line. We observed a significant decrease in detectable bioactive IFN-Is in the serum of LCMV-Arm infected Ppif^−/− mice compared with WT mice (Fig. 6D, 2.0-fold). IFN-I has been demonstrated to play a role in the accumulation of CD8⁺ T cells in vivo by increasing cellular proliferation (20). Thus, using the same adoptive transfer approach, we examined cell proliferation of WT P14 CD8⁺ T cells recovered in the spleens of WT and Ppif^−/− mice 8 d post–LCMV-Arm infection. We assessed cell proliferation in two ways: first, we examined Ki67 expression, which is a marker of proliferating cells only expressed when cells enter the cell cycle (post–G0-phase), and we observed no differences in the frequency of Ki67⁺ WT P14 CD8⁺ T cells from spleens of WT and Ppif^−/− mice 8 d postinfection (Fig. 6E, 6F). Second, we pulsed mice with BrdU on day 0 and maintained them on BrdU drinking water for 8 d. BrdU incorporates into cells with newly synthesized DNA during the S-phase of mitosis, and as such, more BrdU incorporation indicates increased cell proliferation. We observed no differences in the frequency of WT P14 CD8⁺ T cells incorporating BrdU in spleens of WT and Ppif^−/− mice 8 d post–LCMV-Arm infection (Fig. 6G, 6H). Together, these results demonstrated that the ability of CD8⁺ T cells to enter cell cycle and to proliferate following infection are not regulated by CypD, although CypD deficiency does lead to a reduction in IFN-I production.

Given that our results showed no defect in cell proliferation, we next examined whether increased cell death was contributing to the reduction in CD8⁺ T cell numbers observed. First, we examined cell death via Annexin V staining, which is used in conjunction with viability dye staining to determine the percentage of cells that are undergoing apoptosis. We observed significant increases in the frequency of WT P14 CD8⁺ T cells undergoing apoptosis (Annexin V⁺ Zombie⁻) from Ppif^−/− mice compared with WT mice 8 d post–LCMV-Arm infection (Fig. 6I, 6J). Next, we examined the activation of caspase enzymes using a Caspase-3/7 assay. The fluorescent inhibitor of caspases (FLICA) interacts with the enzymatic reactive center of activated caspase, thus directly measuring the amount of active caspase present. We observed a significant increase in the frequency of WT P14 CD8⁺ T cells positive for FLICA from Ppif^−/− mice compared with WT mice 8 d post–LCMV-Arm infection (Fig. 6K, 6L), indicating that the cells are expressing active caspases. Together, these results demonstrated that the reduction in WT P14 CD8⁺ T cell numbers observed post–LCMV-Arm infection is attributed to an increase in apoptotic cells in Ppif^−/− mice.

To assess which CD8⁺ T cell–extrinsic factors are contributing to the CD8⁺ T cell impairment in the absence of CypD, we first examined frequencies and numbers of DCs in LCMV-Arm–infected WT and Ppif^−/− mice on day 2 postinfection (DCs identified as CD3⁻, CD19⁻, NK1.1⁻ CD11c⁺, MHCII⁺ cells). We observed reductions in both the frequencies (Fig. 7A, 7B) and total numbers (Fig. 7C, 2.1-fold) of DCs in Ppif^−/− mice compared with WT mice on day 2 postinfection. Although these data show that DC numbers are reduced following LCMV-Arm infection from mice lacking CypD, it has been demonstrated that overall DC numbers are not critical for the induction of an optimal CD8⁺ T cell response following LCMV-Arm infection (52). Thus, this partial reduction in DC numbers is unlikely to be the key factor in the CD8⁺ T cell impairment we observed in Ppif^−/− mice. Moreover, the maturation and activation status of DCs, rather than total number, are the major drivers of T cell activation by DCs. Thus, we next evaluated whether DC activation and/or maturation is altered in LCMV-Arm–infected Ppif^−/− mice compared with WT mice by assessing the expression of CD40, CD80, and CD86 on day 2 postinfection. CD40 is a costimulatory transmembrane protein that is expressed at low levels on resting DCs and CD80 and CD86 are costimulatory ligands expressed by activated APCs (53–58). We observed similar per-cell expression of CD40 by DCs recovered from day 2 post–LCMV-Arm–infected Ppif^−/− mice compared with WT mice (Fig. 7D, 7E). In contrast, DCs in Ppif^−/− mice expressed significantly more CD80 and CD86 protein on a per-cell basis (Fig. 7F–I). These results suggested that DC activation and maturation are unaffected (or enhanced) in Ppif^−/− mice following LCMV-Arm infection, suggesting that other extrinsic factors may be contributing to the CD8⁺ T cell impairment in mice deficient in CypD. These data are in line with our observations that the proportions of CD8⁺ T cells that encounter Ag (as analyzed by the expression of surrogate markers of Ag-experience) or enter the cycle (Fig. 6E–H) are unaffected in Ppif^−/− mice, which suggests that priming of CD8⁺ T cells is unaffected.

DCs are considered the major APC responsible for priming CD8⁺ T cells. However, other APCs, including macrophages, have also been implicated in T cell responses (52, 56). In addition, macrophages have been suggested to contribute to memory CD8⁺ T cell survival (59). Thus, we further analyzed the impact of CypD deficiency on macrophage populations. We first examined frequencies and numbers of macrophages in LCMV-Arm–infected Ppif^−/− and WT mice on day 2 postinfection (macrophages were identified as CD11b⁺, F4/80⁺ cells). Although macrophage frequencies were similar (Fig. 7J, 7K), total numbers of macrophages were reduced in Ppif^−/− mice compared with WT mice on day 2 post–LCMV-Arm infection (Fig. 7L, 1.9-fold). These results demonstrate that macrophage populations in the spleen are also impacted in Ppif^−/− mice following LCMV-Arm infection.

Because our data suggest that DCs are activated normally in the absence of CypD and that the T cell defect is unlikely to be caused by a defect in T cell priming, we determined whether WT macrophages can provide signals to rescue optimal T cell responses in Ppif^−/−mice. Thus, we asked whether BMDM from WT mice were sufficient to rescue CD8⁺ T cell numbers in LCMV-Arm–infected Ppif^−/− mice. To test this, we derived and differentiated BMDM from Ppif^−/− and WT mice and adoptively transferred them into Ppif^−/−–recipient mice. One day after transfer, recipient mice were infected with LCMV-Arm, and CD8⁺ T cell numbers were examined on day 8 postinfection (Fig. 8A). In Ppif^−/−–recipient mice that received WT BMDM, we observed increased total spleen cellularity (Fig. 8B, 1.4-fold), increased bulk CD8⁺ T cell populations (Fig. 8C, 1.7-fold), and increased Ag-experienced CD8α^lo CD11a^hi CD8⁺ T cells (Fig. 8D, 1.7-fold) compared with Ppif^−/−–recipient mice that received Ppif^−/− BMDM. Further, we observed increased LCMV-specific D^b GP_33–41– and D^b GP_276–286–specific MHC class I tetramer⁺ CD8⁺ T cell numbers (Fig. 8E, 1.8-fold and Fig. 8F, 1.5-fold, respectively) and increased total numbers of IFN-γ and IFN-γ–TNF-α polyfunctional GP_33–41– and GP_276–286–specific CD8⁺ T cells (Fig. 8G–J). Interestingly, although transfer of macrophages was sufficient to partially restore T cell numbers, it did not restore the expression of IFN-I (data not shown). Together, these results supported that CypD plays an extrinsic and critical role in regulating CD8⁺ T cell responses to LCMV-Arm infection and that expression of CypD by macrophages can provide signals that are important to, at least partially, restore CD8⁺ T cell responses in mice lacking CypD. However, this is independent of the role of CypD in regulating IFN-I expression.

Mitochondria play an important role in antiviral immunity, and activation of IFN-I signaling pathways are required for optimal antiviral CD8⁺ T cell responses. In this study, we showed that CypD, a modulatory component of the MPTP, plays a role in the induction of a CD8⁺ T cell–mediated antiviral immune response by promoting CD8⁺ T cell survival in a T cell–extrinsic manner. In addition, our results suggest that CypD plays a role in the induction of IFN-I but that this is likely independent of its role in maintaining CD8⁺ T cell survival. Reduced IFN-I or reduction in the number of innate cells such as macrophages and DCs in the absence of CypD may play a role in reducing the early capacity of the host to control pathogen burden, as we observed increased viral load at an early time-point following infection. Thus, we have identified a novel role for CypD as an important player in antiviral immunity, which can influence both the production of IFN-I and the induction of an optimal CD8⁺ T cell response.

CypD plays a role in macrophage immunity to M. tuberculosis infection, in which CypD inhibition with cyclosporine A results in decreased in vitro pathogen growth and necrosis (60). Despite this, it has been recently shown that following M. tuberculosis infection, CypD-deficient mice have enhanced T cell responses that contribute to significant lung tissue damage and increased susceptibility to infection, without modulating in vivo pathogen loads. Furthermore, it was demonstrated that CypD intrinsically regulates T cell responses to M. tuberculosis infection (6). In this study, we showed that lack of CypD resulted in decreased CD8⁺ T cell responses to LCMV-Arm infection that is attributed, at least in part, to T cell–extrinsic factors. The fact that CypD intrinsically inhibits detrimental, overly exuberant T cell responses to M. tuberculosis while extrinsically promoting protective T cell responses to LCMV infection may have important underlying biological mechanisms. M. tuberculosis is a chronic bacterial infection that induces a delayed adaptive response (weeks postinfection) that persists, such that regulation of T cell proliferation versus contraction is essential for the maintenance of granuloma and prevention of M. tuberculosis dissemination. In contrast, LCMV-Arm infects acutely and promotes a robust T cell response (days postinfection) that contributes directly to viral killing. Thus, it may be that CypD has differential roles in response to bacterial versus viral infections. Furthermore, the role of CypD may differ between macrophages and T cells. Of note, although our study clearly identified a novel T cell–extrinsic role for CypD in inducing optimal T cell responses, our data do not rule out a potential T cell–intrinsic role for CypD in the context of viral infection. These interesting questions will be the subject of future investigations.

Importantly, our findings suggested that macrophages may be one of the major cell types responsible for the reduced CD8⁺ T cell response in Ppif^−/− mice, as adoptive transfer of WT macrophages into LCMV-Arm–infected CypD-deficient hosts enhanced Ag-specific CD8⁺ T cell numbers. Interestingly, WT macrophages were able to increase CD8⁺ T cell numbers without restoring the IFN-I response in Ppif^−/− mice, which is an important signal to maintain CD8⁺ T cell proliferation (16–21, 23). This is in line with our data demonstrating that the reduced number of Ag-specific CD8⁺ T cells in Ppif^−/− mice is not because of a defect in cell cycle entry or proliferation but rather to an increase in CD8⁺ T cell death. This suggests the intriguing concept that macrophages may be important to maintain effector CD8⁺ T cell viability in a CypD-dependent manner. Previous studies have suggested that macrophages provide key signals to maintain viability of memory CD8⁺ T cells (59). As such, it is tempting to speculate that macrophages may be providing key survival signals during the course of effector CD8⁺ T cell responses as well and that this is dependent on the expression of CypD. Thus, CypD expression by macrophages may be important for the induction of optimal CD8⁺ T cell responses, whereas CypD expression by other cell types (potentially including T cells) is important for the regulation of IFN-I. However, the adoptive transfer approach used in this study does not necessarily demonstrate that CypD expression by macrophages is absolutely required for optimal CD8⁺ T cell responses. Rather, it suggests that WT macrophages are at least able to provide some of the signals that are lacking in CypD-deficient mice to sustain CD8⁺ T cell survival. The macrophage-intrinsic role of CypD in maintaining CD8⁺ T cell survival will be the focus of future studies. In addition, it is important to note that transfer of macrophages was not sufficient to fully restore T cell numbers to WT levels. This may simply be a result of the low number of macrophages transferred that yields an incomplete rescue but may also suggest that CypD expression by other cell types is important for optimal CD8⁺ T cell accumulation.

In summary, we have shown a novel role for CypD in CD8⁺ T cell–mediated responses to viral infection. CypD plays a role in the MPTP and cell death, and drugs that inhibit its function such as cyclosporine A have been used in patients with liver, heart, and brain injuries (5). In this study, we demonstrated that CypD deficiency affected the innate antiviral response, which in turn is associated with marked impairments in CD8⁺ T cell survival and accumulation following acute viral infection in mice. Thus, caution is warranted when administering CypD-inhibitory drugs, as host immunity to certain viruses may be compromised.

We thank Dr. Noah S. Butler and members of the Richer Lab for the critical review of this manuscript.

The authors have no financial conflicts of interest.