Research ArticleINFECTIOUS DISEASES

Mitochondrial cyclophilin D regulates T cell metabolic responses and disease tolerance to tuberculosis

See allHide authors and affiliations

Science Immunology  11 May 2018:
Vol. 3, Issue 23, eaar4135
DOI: 10.1126/sciimmunol.aar4135

Mycobacteria and metabolism

Since the discovery of Mycobacterium tuberculosis (Mtb) over a century ago, great progress has been made in defining mechanisms of host resistance to tuberculosis (TB). By contrast, our understanding of how 90 to 95% of infected individuals live with chronic TB is extremely limited. Here, Tzelepis et al. examine the role of mitochondrial matrix protein cyclophilin D (CypD) in T cells using a mouse model of Mtb infection. CypD-deficient mice were more susceptible to Mtb infection in spite of enhanced Mtb-specific T cell responses, which has no impact on curbing bacterial loads but substantially increased lung immunopathology. Their findings indicate that CypD is a critical checkpoint of T cell metabolism for regulating disease tolerance in TB.

Abstract

Mycobacterium tuberculosis (Mtb) is one of the most ancient human pathogens, yet the exact mechanism(s) of host defense against Mtb remains unclear. Although one-third of the world’s population is chronically infected with Mtb, only 5 to 10% develop active disease. This indicates that, in addition to resistance mechanisms that control bacterial burden, the host has also evolved strategies to tolerate the presence of Mtb to limit disease severity. We identify mitochondrial cyclophilin D (CypD) as a critical checkpoint of T cell metabolism that controls the expansion of activated T cells. Although loss of CypD function in T cells led to enhanced Mtb antigen–specific T cell responses, this increased T cell response had no impact on bacterial burden. Rather, mice containing CypD-deficient T cells exhibited substantially compromised disease tolerance and succumbed to Mtb infection. This study establishes a mechanistic link between T cell–mediated immunity and disease tolerance during Mtb infection.

INTRODUCTION

Despite over a century since the discovery of Mycobacterium tuberculosis (Mtb), our understanding of protective immunity to Mtb remains elusive. Considering the ubiquitous infection of humans, the low conversion rate of chronic Mtb infection to active tuberculosis (TB) (1) indicates that Mtb has coevolved with humans to achieve an evolutionary trade-off, which is infrequently compromising host fitness for survival. Conventionally, this trade-off has been considered to be dependent almost exclusively on host resistance to control Mtb growth. However, disease tolerance is another host defense strategy that provides protection against infection by tolerating the presence of the pathogen and controlling tissue damage without affecting pathogen burden (24). Although the concept of disease tolerance is well established in plants (5), only recently has its contribution to Drosophila (68) and mammalian (911) host defense been appreciated. The relevance of this host defense mode during TB has not been carefully examined.

T lymphocytes are essential to control primary Mtb infection, because T cell–deficient animal models and lymphopenic HIV patients are extremely susceptible to Mtb infection (12). Although it is clear that T cell–mediated immunity is critical for preventing Mtb dissemination, whether enhancing the T cell response to Mtb improves protection against TB is less clear. Accumulating evidence suggests that dysregulated T cell responses during Mtb infection may actually increase bacterial burden and susceptibility. For instance, programmed cell death protein-1 (PD-1)–deficient mice were found to be extremely susceptible to Mtb infection despite markedly enhanced T cell responses. This increased susceptibility was associated with increased pulmonary bacterial growth (13, 14). Thus, a tightly controlled T cell response is fundamental for host survival to TB.

Specific cell surface receptors expressed on antigen-experienced T cells play a major role in regulation of T cell responses. Binding of the inhibitory receptor PD-1 to its ligands (PD-L1 or PD-L2) prevents T cell proliferation and function, whereas increased expression of the killer cell-lectin–like receptor subfamily G member 1 (KLRG1) enhances T cell differentiation (15). Although the surface expression of these markers on T cells appears to be critical for dictating their functional role during Mtb infection (13, 14, 16, 17), our understanding of intrinsic immunoregulatory mechanisms of T cells is limited.

One of the central programs initiated upon T cell activation is reprogramming of cellular metabolism to fuel T cell proliferation, differentiation, and cytokine production. Upon T cell receptor (TCR) activation, T cells increase their rate of nutrient uptake and initiate a shift in their metabolic program to aerobic glycolysis (18). High rates of glycolysis enable T cells to rapidly produce adenosine 5′-triphosphate (ATP) and biosynthetic precursors essential for cell growth and proliferation. This switch from predominantly oxidative phosphorylation (OXPHOS) (in the resting state) to aerobic glycolysis (in the proliferative state), despite the presence of abundant oxygen, is known as the “Warburg effect.” It has been recently shown that the Warburg effect occurs in the lungs of mice with TB and that it is necessary for the bioenergetic demands of host immunity to Mtb infection (19).

Mitochondria are central platforms for cellular bioenergetics and biosynthesis, rapidly switching from catabolic to anabolic states to provide the biosynthetic intermediates is pivotal for cellular function (20). In addition, mitochondrial metabolism influenced by the energy sensor adenosine 5′-monophosphate–activated protein kinase (AMPK) can buffer T cells from energetic stress and influences T cell responses to both viral and bacterial infections (21). However, the molecular mechanisms that regulate T cell metabolic demands during Mtb infection are poorly understood.

Cyclophilin D (CypD) is a mitochondrial matrix protein of the cyclophilin protein family (22) encoded by the Peptidyl-prolyl isomerase F (Ppif) gene, shown to play a key role in necrosis by regulating the mitochondrial permeability transition pore (MPTP), which allows the passage of solutes and water from the cytoplasm into the mitochondria (23, 24). Because necrosis in macrophages is an exit mechanism for Mtb (2527), Gan et al. (28) initially demonstrated that the pharmacological inhibition of CypD in human macrophages leads to the inhibition of necrosis and reduction of Mtb growth in vitro. Genetic blockade of CypD in the zebrafish model of TB prevented macrophage necrosis and enhanced their antimycobacterial capacity (29). Thus, from fish to human, CypD appears to play a critical role in antimycobacterial immunity of macrophages.

On the basis of the role of CypD in macrophage immunity to Mtb infection, here, we examined the function of CypD in host responses to Mtb infection. We found that Ppif−/− (CypD-deficient) mice were highly susceptible to Mtb infection compared with control animals, despite similar numbers of bacteria between groups. We established that susceptibility of Ppif−/− mice to Mtb infection was related to enhanced T cell responses that promoted lung immunopathology, independent of bacterial burden. Using loss- and gain-of-function experimental approaches, we have demonstrated that CypD intrinsically regulates T cell metabolism and critically controls disease tolerance in TB. These data indicate that, although host resistance is essential to Mtb infection, effector T cell responses must be tightly regulated to maintain disease tolerance and host survival.

RESULTS

CypD-deficient mice are highly susceptible to Mtb infection

Virulent Mtb inhibits apoptosis and triggers necrosis in macrophages to evade macrophage antibacterial responses (30), subvert adaptive immunity (31), and facilitate dissemination. Previous studies have shown that inhibition of CypD notably decreased necrosis in Mtb-infected macrophages and also prevented the growth of bacteria (26, 28, 29, 32). In line with these findings, we also found that after Mtb infection, CypD-deficient macrophages were significantly resistant to necrosis (fig. S1).

On the basis of these in vitro observations, we next investigated the potential contribution of CypD in vivo. We initially assessed bacterial burden and survival rate after intravascular (IV) infection in both Ppif−/− and wild-type (WT) mice with a high dose [~106 colony-forming units (CFU)] of virulent Mtb (H37Rv). Surprisingly, although there was no difference in the bacterial burden in the lungs, spleens, and livers at days 7, 14, and 21 after Mtb infection (Fig. 1A), Ppif−/− mice were highly susceptible to Mtb infection, and all succumbed to death within 100 days (Fig. 1B). We next assessed the kinetics of pulmonary bacterial burden after infection of WT and Ppif−/− mice with a low dose (~50 CFU) of aerosolized virulent Mtb. Similar to IV infection, there was no difference in the lung bacterial burden of Mtb-infected WT and Ppif−/− mice (Fig. 1C), whereas the survival of Ppif−/− mice was significantly reduced after pulmonary Mtb infection (Fig. 1D). However, lung gross and histology examinations revealed profound nodular lesions (Fig. 1E), and the affected alveolar spaces were significantly filled by increased lymphocytic inflammation (Fig. 1F) in the lungs of Mtb-infected Ppif−/− mice. Flow cytometry identified that most of the pulmonary lymphocytic infiltration was CD3+CD4+ and CD3+CD8+ T cells (Fig. 1G), and the expression levels of Ppif was also significantly up-regulated in T cells isolated from Mtb-infected lungs (fig. S2A). In line with this increased T cell–derived inflammation in Mtb-infected Ppif−/− mice, we observed enhanced immunopathology leading to severe damage in the lung architecture and alveolar septa (Fig. 1F, high magnification). Moreover, similar to another study (33), our data also indicate the impact of sex on CypD function was negligible (fig. S2, B to E). Collectively, these data suggest that in the absence of CypD, there is an increased pulmonary T cell–mediated immunopathology that may lead to enhanced mortality during Mtb infection.

Fig. 1 CypD-deficient mice are susceptible to Mtb infection.

(A) Bacterial burden in the lung, spleen, and liver of WT and Ppif−/− mice (n = 4 to 5 per group) at 7, 14, and 21 days after intravenous infection with 106 H37Rv (two-way ANOVA). (B) Survival of WT and Ppif−/− mice (n = 10 per group) after intravenous infection with 106 H37Rv (log-rank test). (C) Bacterial burden in the lungs after 14, 35, and 90 days of aerosol infection with ~50 H37Rv (n = 5 per group) (two-way ANOVA). (D) Survival of WT and Ppif−/− mice after aersolized infection with ~50 H37Rv (n = 10 per group). *P < 0.05 (log-rank test). (E and F) Representative gross pathology (E) and histology (H&E staining) and inflammatory scores (F) of the lungs of WT and Ppif−/− mice at day 90 after aerosol infection with ~50 H37Rv. Lymphocytes (black arrows) and macrophages (dashed arrows). *P < 0.05 (unpaired t test) (G) Total number of CD4+ (top) and CD8+ T cells (bottom) in the lungs of WT and Ppif−/− mice (n = 5 per group) at day 90 after aerosol infection with ~50 H37Rv. *P < 0.05 (unpaired t test). Data are representative of two independent experiments.

CypD-deficient mice have enhanced T cell responses

Considering that the lungs of Ppif−/− mice showed increased inflammation, predominantly involving T cell infiltration, compared with control animals, we next investigated the quantity and quality of the antigen-specific T cell responses. At various time points after a low dose (~50 CFU) of aerosolized virulent Mtb infection, there was a significant increase in the frequency and the total cell number of Mtb antigen–specific (TB10.4) CD8+ T cells in the lungs of Ppif−/− mice compared to WT mice (Fig. 2, A to C). CypD-deficient CD8+ T cells were highly active, producing significantly higher levels of interferon-γ (IFN-γ) and tumor necrosis factor–α (TNFα; Fig. 2D and fig. S3A). The assessment of T cell–lineage transcription factors indicated that CypD-deficient T cells highly express T-bet, which is a critical transcription factor for T helper 1 cells (TH1; Fig. 2E and fig. S3B), whereas the levels of other transcriptional factors for TH17 [retinoic acid receptor–related orphan receptor gamma t (RORγt)], regulatory T cell (FOXP3), and TH2 [GATA binding protein 3 (GATA3)] cells were not significantly changed (fig. S3C). Similar data were also obtained in IV model of Mtb infection (Fig. 2, F to H, and fig. S3, D to F).

Fig. 2 CypD-deficient mice have exacerbated Mtb antigen–specific T cell responses to Mtb infection.

(A to C) Frequency (A and B) and total number (C) of TB10.4(4–11)-specific CD8+ T cells in the lungs of WT and Ppif−/− mice (n = 5 per group) after 14, 35, and 90 days of aerosol infection with ~50 H37Rv. Phycoerythrin (PE)–conjugated streptavidin was used as a control for the tetramer staining. Numbers above outlined areas (A) indicate the percentage of CD8+ T cells stained with H-2Kb- TB10.4(4–11). *P < 0.05, **P < 0.01, ****P < 0.0001 (two-way ANOVA). (D) Frequency of TNFα- and IFN-γ–producing antigen-specific CD8+ T cells was determined by intracellular staining (ICS) upon restimulation with TB10.4(4–11) peptide at day 35 after aerosol infection with ~50 H37Rv. ****P < 0.0001 (two-way ANOVA). (E) Frequency of T-bet expression in CD8+ T cells was determined by ICS at day 35 after aerosol infection with ~50 H37Rv. *P < 0.05 (unpaired t test). (F to H) Frequency (F to G) and total number (H) of TB10.4(4–11)-specific CD8+ T cells in the lungs of WT and Ppif −/− mice (n = 5 per group) after 10 and 15 days of intravenous infection with H37Rv (~106 CFU). PE-conjugated streptavidin was used as a control for the tetramer staining. Numbers above outlined areas (G) indicate the percentage of CD8+ T cells stained with H-2Kb-TB10.4(4–11). *P < 0.05, ***P < 0.001, ****P < 0.0001 (two-way ANOVA). (I) Schematic of adoptive T cell transfer into Rag1−/− mice followed by aerosol infection with ~50 H37Rv. (J) Bacterial burden in the lungs, spleens, and livers of Rag1-deficient mice reconstituted with either with WT T cells (T cellsWT) or CypD-deficient T cells (T cellsPpif−/−) after 7, 14, and 35 days of infection with H37Rv (n = 5 per group). (K) Frequency of TB10.4(4–11)-specific CD8+ T cells in the lungs of Rag1-deficient mice reconstituted with either with WT T cells (T cellsWT) or CypD-deficient T cells (T cellsPpif−/−) (n = 5 per group) after 7, 14, and 35 days of aerosol infection with H37Rv. **P < 0.01 (unpaired t test). Data in (A) to (H) are representative of three independent experiments. Data in (I) to (K) are representative of two independent experiments.

To determine the protective capacity of CypD-deficient T cells in host resistance to Mtb, we adoptively transferred equal numbers of purified naïve T cells from WT and Ppif−/− mice into Rag1-deficient mice that were then infected with a low dose (50 to 100 CFU) of virulent Mtb via aerosol route (Fig. 2I). In recipient Rag1−/− mice, we found no significant difference in bacterial burden in the lungs, spleens, or livers at days 7, 14, and 35 after Mtb infection (Fig. 2J); however, there was a significant increase in Mtb antigen–specific CD8+ T cells in Rag1−/− mice that received CypD-deficient T cells (Fig. 2K and fig. S3G). In addition, in either aerosolized or IV models of Mtb infection, the frequency of other immune cells, including alveolar macrophages, monocytes, neutrophils, and B cells, was not markedly altered between Mtb-infected WT and Ppif−/− mice (fig. S4, A to C). These results indicate that after Mtb infection, there is an increase in CypD-deficient T cell proliferation/activation, and the protective capacity of CypD-deficient T cells against Mtb is intact.

CypD negatively regulates T cell proliferation

The increased number of T cells observed in Ppif−/− mice could be explained by two nonmutually exclusive possibilities: (i) enhanced T cell proliferation and (ii) reduced T cell death. To determine whether CypD affects the ability of naïve T cells to proliferate and expand, we first evaluated homeostatic T cell proliferation in vivo. Naïve T cells purified from WT or Ppif−/− mice were labeled with carboxyfluorescein succinimidyl ester (CFSE) and adoptively transferred intravenously into Rag1-deficient mice (Fig. 3A). Three days after adoptive transfer, where control T cells had just begun to initiate proliferation (less than two cell divisions), we observed greater expansion of CypD-deficient T cells (two to three divisions) in recipient mice (Fig. 3B). We next investigated whether the differences in T cell proliferation could also be observed in vitro. CD8+ T cells purified from WT or Ppif−/− mice were labeled with CFSE and stimulated in vitro with different concentrations of plate-bound anti-CD3 and anti-CD28 antibodies, and cell proliferation was assessed 3 days later. CypD-deficient T cells displayed increased proliferation after TCR engagement compared to controls (Fig. 3C). In addition, CD8+ T cells demonstrated elevated TNFα and IFN-γ production (Fig. 3D and fig. S5A). Corroborating enhanced T cell proliferation, expression of early T cell activation markers (CD69 and CD25) was also substantially increased in CypD-deficient CD8+ T cells compared with WT CD8+ T cells at both frequency and mean fluorescence intensity (MFI; Fig. 3, E to F, and fig. S5B). Similarly, CypD-deficient CD4+ T cells also showed elevated TNFα and IFN-γ production after anti-CD3 and anti-CD28 stimulation (fig. S5C).

Fig. 3 CypD regulates T cell activation and proliferation.

(A) Rag-deficient mice were used as recipients for splenic T cells (CD3+) purified from naïve Ppif−/− or WT mice and labeled with CFSE. (B) CD8+ T cell proliferation (B) and the percentage of undivided CD8+ T cells (C) after 3 days of adoptive transfer. ****P < 0.0001 (unpaired t test). (C) Proliferation of CFSE-labeled CD8+ T cells purified from WT (gray) and Ppif−/− (black line) after 3 days of stimulation with plate-bound anti-CD3 and anti-CD28. (D) CD8+ T cells were stimulated for 48 hours with anti-CD3 (5 μg/ml) and anti-CD28 (2 μg/ml) and analyzed for TNFα and IFN-γ production by ICS. **P < 0.01, ****P < 0.0001 (unpaired t test). (E to F) Frequency of CD8+ T cells expressing CD69 and CD25 after 24 hours of stimulation with plate-bound anti-CD3 (5 μg/ml) and anti-CD28 (2 μg/ml). MFI of CD69 and CD25 (F). ****P < 0.0001 (unpaired t test). (G) Frequency of proliferation of CFSE-labeled CD8+ T cells purified from WT OT-I or CypD-deficient OT-I after 3 days of coculture with BMDC loaded with OVA protein. Frequency of cells that undergo two or more cycles of proliferation (right). **P < 0.01 (unpaired t test). (H) Frequency of pulmonary Ki67+ CD8+ T cells at days 14 and 28 after aerosol infection with ~50 H37Rv. *P < 0.05 (unpaired t test). (I) MFI of Ki67, Bcl-2, and activated caspase 3 from TB10.4(4–11)-specific CD8+ T cell from lungs of WT and Ppif−/− mice at day 14 after infection with virulent with ~50 H37Rv via aerosol. *P < 0.05 (unpaired t test). (J to K) Frequency of pulmonary active caspase3+ CD8+ T cells (J) and 7AAD+ CD8+ T cells (K) at days 14 and 28 after aerosol infection with ~50 H37Rv. Data are representative of three independent experiments.

To further validate the intrinsic role of CypD in T cell proliferation, we generated CypD-deficient ovalbumin (OVA) TCR-transgenic CD8+ mice (OT-I mice). Purified CD8+ T cells from either WT OT-I or CypD-deficient OT-I mice were labeled with CFSE and cocultured with WT bone marrow (BM)–derived dendritic cells (BMDCs) loaded with OVA protein. Three days after coculture, we observed significantly increased proliferation of CypD-deficient OT-I CD8+ T cells specific to OVA protein compared to control OT-I CD8+ T cells (Fig. 3G). To evaluate the role of CypD in T cell proliferation during Mtb infection, we next aerosolized WT and Ppif−/− mice with a low dose of Mtb and assessed the proliferative capacity of T cells by using markers for proliferation (Ki67), anti-apoptotic protein (Bcl2), or apoptosis marker (active caspase 3). At days 14 and 28 after infection, we observed an increase in the frequency of Ki67+ CD8+ T cells from Ppif−/− mice compared to WT mice (Fig. 3H). Similarly, the MFI of Ki67 was significantly increased in Mtb antigen–specific CypD-deficient CD8+ T cells, whereas there were no significant changes in Bcl-2 or active caspase 3 (Fig. 3I).

To further assess whether the increased frequency of CypD-deficient T cells after infection is due to enhanced proliferation or reduced cell death, we measured the levels of active caspase 3, 7AAD, and annexin V in T cells from the lungs of Mtb-infected mice. At both 14 and 28 days after Mtb infection, no significant change was observed in active caspase 3, annexin V, or 7AAD expression in CD8+ T cells between genotypes (Fig. 3, J to K, and fig. S5D). In addition, the levels of 7ADD and annexin V was also similar in both WT and CypD-deficient CD4+ T cells (fig. S5E). Furthermore, after in vitro stimulation (anti-CD3/CD28) of WT or CypD-deficient T cells for 48 hours, there was no impact on cell viability (fig. S5F). Taking into consideration the role of PD-1 in limiting activation of CD4+ and CD8+ T cells, as well as its essential role in protection against Mtb (13, 14, 16), we next assessed the expression of PD-1 on both CD4+ and CD8+ T cells after Mtb infection. At both days 14 and 28 after pulmonary Mtb infection, we found no difference in the expression of PD-1 on CD4+ or CD8+ T cells isolated from the lungs of Ppif−/− mice compared to WT mice (fig. S5G). Collectively, these data indicate that CypD intrinsically regulates T cell proliferation independent of the cell death program.

Increased metabolic activity drives the enhanced proliferation and cytokine production of CypD-deficient T cells

T cell expansion and effector function (i.e., cytokine production) are tightly coupled to glycolytic and mitochondrial activity during infections (21, 34). Considering the location of CypD in the mitochondria and many cell survival pathways being linked to global cellular metabolism (35), we hypothesized that CypD may influence T cell proliferation and effector function via control of T cell metabolic activity. To test this, we performed bioenergetic flux analysis of in vitro–generated WT and CypD-deficient T effector cells using a Seahorse Bioanalyzer. CypD-deficient T cells displayed increased aerobic glycolysis compared to control T cells, as evidenced by elevated extracellular acidification rate (ECAR) and lactate production (Fig. 4, A and B). Mitochondrial OXPHOS was also elevated in CypD-deficient T cells, as seen by the overall increase of the basal oxygen consumption rate (OCR; Fig. 4C). CypD-deficient T cells appeared to operate at maximum mitochondrial capacity, with very little spare respiratory capacity compared to control cells (Fig. 4C and fig. S6A). The observed global increase in metabolism of CypD-deficient T cells (Fig. 4D) correlated with greater glucose and glutamine consumption in CypD-deficient CD8+ T cells (Fig. 4, E and F). These effects were not limited to CD8+ T cells, because similar changes in metabolism (fig. S6, B and C) and cytokine production (fig. S6D) were also observed in CD4+ T cells. These changes were also independent of the number of mitochondria in CD4+ and CD8+ T cells from WT or Ppif−/− mice (fig. S6E).

Fig. 4 CypD regulates T cell metabolism and proliferation via mROS.

(A and B) ECAR (A) and lactate (B) measurements of Ppif−/− and WT CD8+T cells activated for 48 hours with anti-CD3 (5 μm/ml) and anti-CD28 (2 μg/ml). **P < 0.01, ****P < 0.0001 (unpaired t test). (C) The oxygen consumption measured on the bioflux analyzer as a function of time (OCR). Mitochondrial function was analyzed through the addition of mitochondrial inhibitors in the following order: oligomycin A (O) (1.5 μM), FCCP (F) (2 μM), and antimycin A (A) (2 μM). (D) The energetic status of CD8+-activated T cells from WT (black square) and Ppif−/− (open circle) mice was assessed 48 hours after activation. The bioenergetics flux was represented as the change in ECAR on the x axis in milli-pH unit/minute (mpH/min) versus the OCR on the y axis. Each symbol represents one mouse per genotype, with n = 3. (E and F) Metabolite consumption on a per-cell basis between WT and CypD-deficient CD8+ T effector cells was assessed. Glucose (glc) consumption (E) and glutamine consumption (F) in the media was measured after 48 hours of in vitro activation. ***P < 0.001 (unpaired t test). (G) WT and CypD-deficient CD8+T cells were activated for 48 hours, and their percent of double-positive IFN-γ+/TNFα+ production was assessed by ICS and flow cytometry. Before restimulation for ICS, CD8+ T cells were either pretreated with 2 mM 2-DG, Na+ oxamate, or vehicle. ****P < 0.0001 (unpaired t test). (H) Nonmitochondrial OCR measured from bioenergetics flux analysis between WT and CypD-deficient CD8+ T effector cells. *P < 0.05 (unpaired t test). (I) mROS production measured as the MFI of MitoSOX dye in CD8+ T effector cells generated from WT and Ppif−/− mice. **P < 0.01 (unpaired t test). (J to M) ECAR (J), OCR (K), CFSE proliferation assay (L), and quantification (M) of WT (top) and CypD-deficient (bottom) CD8+ T cells in the presence of vehicle (gray) or 2.5 μM of CoQ (black line) stimulated for 3 days with plate-bound α-CD3 (1 μg/ml) and α-CD28 (0.5 μg/ml). **P < 0.01, ****P < 0.0001 (two-way ANOVA). Data are representative of at least three independent experiments.

Cytokine production is associated with rates of glycolytic metabolism (21, 34). Thus, we hypothesized that the hyperinflammatory phenotype of CypD-deficient CD8+ T cells may be due to the hypermetabolic state of these Tcells. To test this, we treated invitro–generated CD8+ WT and CypD-deficient T effector cells for 5 hours with the glycolytic inhibitor, 2-deoxyglucose (2-DG). Treatment of CypD-deficient T cells with 2-DG entirely ablated the increased IFN-γ response of these cells (Fig. 4G). Furthermore, inhibition of lactate dehydrogenase A (LDHA), which regulates the latter arm of glycolysis, using sodium oxamate led to markedly reduced IFN-γ responses in CypD-deficient T cells (Fig. 4G). These data suggest that increased cytokine production in CypD-deficient T cells is due, in part, to aberrant glucose metabolism in these cells. It appears that CypD regulates T cell metabolism independent of the classical mammalian target of rapamycin complex 1 (mTORC1), Akt, or hypoxia-inducible factor–1α (HIF-1α) pathways (fig. S6, F to I).

Another well-documented regulator of T cell proliferation is reactive oxygen species (ROS) (20). Mitochondrial ROS (mROS), produced at complex III of the electron transport chain (ETC), is required for T cell expansion (20). Nonmitochondrial oxygen consumption was markedly increased in CypD-deficient T cells (Fig. 4H), suggesting that residual oxygen consumption may occur from other sources such as ROS production. Consistent with this, mROS levels was significantly elevated in CypD-deficient CD8+ T cells (Fig. 4I). Quenching of mROS by coenzyme Q (CoQ; an electron carrier for the ETC that acts as an antioxidant) significantly reduced ECAR and OCR in CypD-deficient T cells (Fig. 4, J and K) and reversed the hyperproliferative phenotype of CypD-deficient CD8+ T cells while having no effect on WT T cell proliferation (Fig. 4, L and M). Thus, these data suggest that CypD regulates T cell metabolism in a ROS-dependent manner.

CypD intrinsically regulates T cell–mediated immunopathology

To directly test the potential role of CypD in hyperproliferation and activation of T cells, we next generated chimeric mice in which only T cells lack CypD (T cellPpif−/− chimeric mice). As shown in Fig. 5A, irradiated α/β TCR–deficient mice (Tcra−/−) were reconstituted with BM cells from Tcra−/− and Ppif−/− mice mixed in a 4:1 ratio, respectively. Thus, T cells in reconstituted T cellPpif−/− mice are only derived from Ppif−/− precursors, whereas the non–T cells are predominantly Ppif+/+. Control chimeric mice were generated in an equivalent manner by using a BM mixture from Tcra−/− mice and WT littermates (T cellsPpif+/+ chimeric mice). Twelve weeks after reconstitution, we confirmed the absence of CypD in T cells purified from T cellPpif−/− mice (Fig. 5B). To evaluate T cell responses, we aerosolized chimeric mice with virulent Mtb. Five weeks after Mtb infection, we observed a significant increase in antigen specific CD8+ T cells in the lungs of T cellPpif−/− mice (Fig. 5, C and D). In parallel, the levels of IFN-γ and TNFα were significantly higher in CD8+ T cells isolated from the lungs of T cellPpif−/− mice (Fig. 5, E and F). These results indicate that CypD intrinsically regulates T cell responses after Mtb infection.

Fig. 5 CypD-deficient T cells are pathological.

(A) Schematic diagram for generation of mixed BM chimeric mice. Twelve weeks after reconstitution, mice (n = 4 to 5 per group) were infected via aerosol route with ~50 H37Rv. (B) Immunoblot of CypD expression in purified T cells from T cellWT or T cellPpif−/− chimeric mice after 12 weeks of reconstitution. (C to D) Frequency of TB10.4(4–11)-specific CD8+ T cell in the lung of T cellWT or T cellPpif−/− chimeric mice after 35 days of H35Rv infection. PE-conjugated streptavidin was used as a control for the tetramer staining. Numbers above outlined areas indicate the percentage of CD8+ T cells stained with H-2Kb–TB10.4(4–11). *P < 0.05 (unpaired t test). (E and F) Frequency of TNFα- and IFN-γ–producing antigen-specific CD8+ T cells was determined by ICS upon restimulation with TB10.4(4–11) peptide at day 35 after aerosol infection with H37Rv. ***P < 0.001 (unpaired t test). (G) Survival of WT and Ppif−/− mice (n = 9 to 10 mice per group) temporarily depleted of T cells with anti-CD8 (150 μg per mouse) and anti-CD4 (150 μg per mouse) or IgG-control antibody at days 7 and 14 after intravenous infection with H37Rv (~5 × 105 CFU). Survival was monitored by log-rank test. (H) Schematic of adoptive T cell transfer from WT or Ppif−/− into Rag1-deficient mice followed by aerosol infection with H37Rv. (I) Survival of Rag-deficient mice (n = 7 per group) after adoptive transfer of T cells from Ppif−/−, WT mice, or equal number of Ppif−/− T cells and WT T cells (1:1) following aerosol infection with ~50 H37Rv. Survival was monitored by log-rank test. *P < 0.05, **P < 0.01, ***P < 0.001. (J) Representative lung histology (H&E staining) of Rag-deficient mice that received T cells from Ppif−/− mice before death after aerosol infection with ~50 H37Rv. Dense polymorphous inflammatory composed of macrophages (green arrows), lymphocytes (dashed arrows), and neutrophils (wide arrows). Data are representative of two independent experiments.

Considering in the absence of CypD, the capacity of T cells in host resistance to Mtb infection was not impaired (Figs. 1, A and C, and 2J), we next investigated the potential contribution of CypD in T cell–mediated disease tolerance using both loss- and gain-of-function approaches. First, we depleted T cells in Ppif−/− and WT mice at days 7 and 14 after IV infection with the low dose of Mtb (fig. S6J). As expected, Ppif−/− mice treated with rat immunoglobulin G (IgG) succumbed earlier to infection than WT mice. Temporary depletion of T cells significantly prolonged survival of Ppif−/− mice (Fig. 5G), supporting the contribution of an exacerbated T cell response to susceptibility. Second, to further support the contribution of CypD-deficient T cells in disease tolerance to Mtb infection, we used an adoptive transfer model where Rag1-deficient mice were reconstituted with naïve WT or CypD-deficient T cells or equal numbers of CypD-deficient T cells and WT T cells (1:1) before aerosolized infection with low dose of Mtb (Fig. 5H). Without T cells, mice quickly succumbed to death, whereas reconstitution of WT T cells significantly enhanced survival of mice from Mtb infection. By contrast, Ppif−/− T cells failed to enhance survival, and the lung histology examination showed severe parenchymal inflammatory infiltrate and tissue damages (Fig. 5, I and J). However, reconstitution of Rag1−/− mice with equal numbers of CypD-deficient T cells and WT T cells significantly increased the survival compared with mice that received only CypD-deficient T cells (Fig. 5I). Together, these results indicate that, although the protective capacity of CypD-deficient T cells in host defense is integral, their heightened proliferative capacity and activation state compromised disease tolerance in TB.

DISCUSSION

Although control of TB requires T cells to prevent disease progression and delaying T cell priming is one strategy that Mtb uses to subvert the host immune response (36), the direct role of T cells in protection against Mtb is still controversial. Here, we identify CypD as a critical regulator of T cell activation that controls the magnitude of the T cell response to Mtb infection.

Although in macrophages, CypD prevents necrosis that enhances macrophage resistance to Mtb infection (26, 28, 29), Ppif−/− mice were extremely susceptible to Mtb infection despite similar bacterial burden and enhanced T cell–mediated immunity. This enhanced susceptibility to Mtb infection was due to an intrinsic function of CypD in T cells, because depletion of CypD-deficient T cells significantly increased survival, and reconstitution of CypD-deficient T cells significantly increased mortality to Mtb infection. Our data suggest that CypD regulates a metabolic checkpoint in T cells—partly through the regulation of mROS—to promote enhanced T cell proliferation and effector function. These results also indicate that CypD regulates mitochondrial function in a cell-dependent manner that, under stress, may control either cell death program or survival and proliferation. One unique difference between fully differentiated macrophages, structural cells (e.g., cardiomyocytes and fibroblasts), and T cells is their capacity to proliferate. Thus, as T cells are intrinsically programmed to proliferate, the functional role of CypD appears to be mainly wired to regulate metabolism and proliferation rather than MPTP and cell death.

T cell metabolic programs adapt based on their activation state. Whereas the generation of ATP for quiescent naïve T cells depends on OXPHOS, activated T cells use glycolysis to adequately match production to their increased bioenergetic and biosynthetic needs (37). Although the role of CypD in the cell death program is well established, little is known about its influence on other mitochondrial functions such as metabolism. Here, we demonstrate that CypD is a key checkpoint for T cell metabolism. CypD-deficient T cells display increases in both glycolysis and OXPHOS, which support both the mitochondrial biosynthetic and ATP production programs needed for enhanced proliferation after Mtb infection. Blocking glycolysis with 2-DG or sodium oxamate ablated cytokine production by CypD-deficient T cells, indicating a direct link between increased glycolytic activity and increased effector function in CypD-deficient T cells. The metabolic reprogramming observed in activated CypD-deficient T cells was not driven by classical mTORC1/HIF-1α signaling but rather was dependent on ROS signaling. CypD-deficient T cells operated at maximal mitochondrial capacity and produced substantially higher levels of mROS compared with control T cells. Quenching excess ROS using antioxidants was sufficient to reverse the glycolytic phenotype of CypD-deficient T cells and return T cell proliferation to control levels. mROS production can drive glycolytic metabolism and is essential for T cell activation and proliferation (20). Thus, the enhanced proliferation may be a secondary response to metabolic dysfunction induced by heightened ROS production in CypD-deficient T cells. CypD-deficient cells also display increased pyruvate dehydrogenase and α-ketoglutarate dehydrogenase activity (38), which may help drive the increased OCR in these T cells.

Although we have demonstrated the regulatory impact of CypD on T cell metabolism and disease tolerance to Mtb infection, it will be important to investigate the impact of CypD in whole tissue or even systemic metabolic responses. TB was initially known as “consumption” due to its ability to cause cachexia (wasting). Consistent with this idea, a previous study examining Mycobacterium marinum infection in the fruit fly Drosophila melanogaster showed an increased mortality that was independent of bacterial load but was associated with altered host metabolisms and increased body wasting (7). Therefore, unraveling the cellular and molecular mechanisms involved in metabolic regulation of disease tolerance is essential for a complete understanding of TB pathogenesis.

Exposure to Mtb leads to two broad outcomes: elimination of the bacteria or its persistence (1). If Mtb is not eliminated (most likely within the airway), then the pathogen translocates into the lung parenchyma and leads to granuloma development. It appears that granuloma formation is a double-edged sword that, on the one hand, prevents the dissemination of Mtb but, on the other hand, allows persistence of Mtb by recruiting monocytes/macrophages, which are required for the growth of bacteria (39). T cells are a dominant constituent of granuloma formation that has long been thought as a critical component of host protective immunity to Mtb infection. The prevailing paradigm that T cells provide protection against Mtb infection stems from studies using major histocompatibility complex class I (MHC-I) knockout (KO; lacking CD8+ T cells), MHC-II KO (lacking CD4+ T cells), or TCRαβ KO (lacking both CD4+ and CD8+ T cells) mice, all of which are impaired in effective granuloma development, are unable to control Mtb growth and are profoundly susceptible to infection (40, 41). In contrast to the setting of T cell deficiency, an increased number of T cells is also not beneficial to the host. For instance, adoptive transfer of high numbers of Mtb antigen–specific T cells before aerosol Mtb infection fails to eliminate the bacteria (42). Several studies have also demonstrated that impaired PD-1 signaling notably enhances host susceptibility to Mtb infection due to excessive CD4+ T cell activation that drives both the growth of bacteria and immunopathology (13, 14). Although PD-1 regulates T cell functions that affect host resistance and immunopathology in TB, our study demonstrates that CypD intrinsically regulates T cell functions in disease tolerance against TB. Thus, one could speculate that, during Mtb infection, the initiation of granuloma formation is a transition of host defense mechanisms from resistance to tolerance, and T cells are the gatekeeper within granulomas. Validation of this concept requires further investigation.

In 90 to 95% of Mtb-infected individuals, bacteria stay in a quiescent or latent state with infrequent progression to active TB disease (43). This epidemiological data indicate that humans have coevolved with Mtb to tolerate its presence with minimum immunopathology. Here, we show that CypD regulates T cell metabolism and directly affects disease tolerance but not host resistance to Mtb infection. Our study furthermore demonstrates that the cellular and molecular mechanisms of disease tolerance to TB differ from host resistance with the former being a key component of host defense and survival. In addition, our data indicate that caution must be exercised in developing a TB-vaccination strategy to boost conventional T cell responses, which inadvertently may compromise disease tolerance to Mtb infection.

MATERIALS AND METHODS

Study design

This study used in vivo and in vitro models of Mtb infection to investigate disease tolerance in pulmonary TB. The original objective of this study was to determine the role of mitochondrial CypD in host resistance to Mtb infection. However, the experimental data revealed that CypD had no impact on host resistance (bacterial burden) and rather regulates T cell metabolism and disease tolerance to Mtb infection. All experiments were performed at least twice. Every group consisted of at least three mice. No outliers were excluded from the data analysis. This study was not randomized and was not blinded.

Mice

Six- to 10-week-old C57BL/6, Rag1−/− (Jax: 002216), and Tcra−/− (Jax: 002116) mice were from the Jackson Laboratory. OT-I T cell antigen receptor-transgenic mice (specific for the OVA257–264, peptide SIINFEKL, restricted by H-2kb) were from Taconic Farm (4175-M), Peptidyl-prolyl isomerase F (Ppif) KO mice (CypD-deficient mice) were obtained from M. Forte (Oregon Health and Science University, Portland, OR, USA) and were bred at McGill University. All animal studies were conducted in accordance with the guidelines of and approved by the Animal Research Ethics Board of McGill University.

Histopathology and bacterial staining

Lung tissues were fixed in 10% formalin for hematoxylin and eosin (H&E) staining. Scoring of histologic parameters was performed by a thoracic pathologist blinded to experimental data and included following parameters: bronchial/endobronchial inflammation, peribronchial inflammation, perivascular inflammation, interstitial inflammation, pleural inflammation, intra-alveolar inflammation, and overall severity (4448). Microphotographs were taken using a DXM1200F digital camera (Nikon).

Adoptive transfer

CD3+ T cells were purified from spleens of naïve WT and Ppif−/− mice using negative selection (Pan T cell kit, Stemcell Technologies). The purity was more than 95%. One million CD3+ T cells were transferred intravenously to Rag1−/− recipient mice. One day later, Rag1−/− mice were infected via aerosol route with 50 to 100 H37Rv. At 1, 2, 4, and 5 weeks after Mtb infection, CFU of lungs, spleens and livers were determined as described above. CD3+ T cells purified from the spleen and lymph nodes of naïve WT and Ppif−/− mice were stained with 5 μM CFSE in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin at room temperature. Staining was stopped after 8 min of incubation by addition of RPMI medium containing 10% fetal bovine serum. Cells (5 × 105) were transferred intravenously (via the retro-orbital vein) to Rag1−/− mice. Three days after adoptive transfer, CD8+ T cell proliferation was evaluated in the spleen of recipient mice by CFSE reduction.

Generation of mixed BM chimeric mice

BM from femurs and tibiae was flushed with RPMI medium. BM cells from Tcra-deficient mice and Ppif−/− mice were mixed in a proportion of 4 to 1, respectively. This mixture was used to generate T cellPpif−/− chimeric mice. BM cells from Tcra-deficient mice and WT mice were combined in the same way to generate T cellWT control mice. As recipient mice, we used Tcra-deficient mice that were irradiated with 9 gray (Gy). These mice were kept under antibiotic treatment [128 mg of polymixin B and 1.1 g of neomycin trisulfate (Sigma-Aldrich) per liter of drinking water] for 10 days after irradiation. Sixteen hours after irradiation, 4 × 106 mixed BM cells were intravenously transferred to recipient mice. Chimeric mice were used after 12 weeks, at which time the hematopoietic compartment was reconstituted as determined by flow cytometry.

Bioenergetic flux analyser (Seahorse) assay

In vitro–activated T cells were collected, washed in PBS, and counted. They were plated on 96-well plates from Seahorse that were previously coated with poly-d-lysine. Cells were plated at 200,000 cells per well and spun down at 1400 rpm for 5 min. A Seahorse Bioenergetic flux analyzer was used to measure the ECAR and the OCRs. Mitochondrial inhibitors were added every 10 min in the following order: oligomycin A (1.5 μM), carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP; 2 μM), and antimycin A (2 μM).

Statistics

Data are means ± SD and analyzed by two-tailed Student’s t test or two-way analysis of variance (ANOVA) as appropriate using the GraphPad Prism program (version 6). For survival analyses, log-rank test was performed. P values of less than 0.05 were considered significant.

SUPPLEMENTARY MATERIALS

immunology.sciencemag.org/cgi/content/full/3/23/eaar4135/DC1

Material and Methods

Fig. S1. CyD-deficient macrophages are more resistance to necrosis after infection with virulent strain of Mtb.

Fig. S2. CypD regulates T cell responses independent of sex during Mtb infection.

Fig. S3. CypD-deficient mice presents enhanced T cell response after infection with Mtb.

Fig. S4. CypD minimally affects innate immune cells and B cell responses after Mtb infection.

Fig. S5. CypD regulates T cells proliferation and activation.

Fig. S6. CypD regulates T cell metabolism.

Table S1. Raw data sets.

REFERENCES AND NOTES

Acknowledgments: Funding: This work was supported by Canadian Institutes of Health Research (CIHR) Operating (MOP-142259 to R.G.J.) and Foundation (FDN-143273 to M.D.) Grants. M.D. holds a Fonds de recherche du Québec–Santé (FRQS) Award and the Strauss Chair in Respiratory Diseases. N.K. holds a FRQS-Fellowship. J.B. and J.G. held studentships from the CIHR. E.H.M. and D.G.R. held fellowships from the FRQS. Author contributions: F.T., J.B., N.K., and M.D. designed the research. F.T., J.B., N.K., E.H.M., D.G.R., and J.G. performed the research. F.T., J.B., and N.K. performed the statistical analysis. R.G.J. and P.J. provided novel reagents/analytical tools. All authors analyzed data. F.T., N.K., and M.D. wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: Ppif −/− KO mice were obtained from M. Forte under a material transfer agreement with Oregon Health and Science University. This mouse strain is now commercially available from the Jackson Laboratory (stock # 022308).
View Abstract

Navigate This Article