TFAM Dependent Mitochondrial Fitness Limits CD8⁺ T Cell Immunopathology and Sustains Protective Immunity during Viral Pneumonia

Abstract

During respiratory virus infection, CD8⁺ T cells kill infected cells and establish antigen-specific memory, but mechanisms regulating these functions remain incompletely understood. Here, we identify mitochondrial transcription factor A (TFAM)-dependent mitochondrial fitness as a regulator of CD8⁺ T cell function during influenza infection. Human CD8⁺ T cells exhibited an age-associated decline in TFAM expression and mitochondrial function. To model this physiologically relevant decline and determine its consequences for antiviral immunity, we generated CD8⁺ T cell-specific TFAM-haploinsufficient mice. TFAM insufficiency disrupted mitochondrial integrity and bioenergetics and increased mitochondrial DNA and oxidative stress. During influenza infection, TFAM-insufficient CD8⁺ T cells exhibited increased cytotoxic and inflammatory activity associated with lung immunopathology without improved viral control. This early phenotype was followed by loss of effector function, diminished antigen-specific responses, reduced protection following adoptive transfer, and impaired heterosubtypic recall immunity. Thus, TFAM-dependent mitochondrial fitness is a cell-intrinsic regulator that limits immunopathology while sustaining recall immunity.

Introduction

CD8⁺ T cells are essential for controlling respiratory viral infections through the recognition and elimination of virus-infected cells and the establishment of long-term protective memory ^1,2. However, excessive or poorly regulated CD8⁺ T cell responses can also promote inflammatory tissue injury and contribute to pulmonary immunopathology². Thus, effective antiviral immunity requires coordinated regulation of effector activity, tissue protection, and the development of durable memory. The cell-intrinsic mechanisms that maintain this balance during acute respiratory viral infection remain incompletely understood.

Mitochondrial fitness is critical for supporting the energetic and biosynthetic demands of CD8⁺ T cell effector response and for sustaining long-lived memory T cells ^3,4. Conversely, mitochondrial stress and impaired bioenergetics have been associated with dysfunctional or exhaustion-associated T-cell states in chronic infection and cancer ^5,6. Mitochondrial dysfunction is a prominent feature of T-cell aging ^1,7, with particular relevance to respiratory viral infection because older individuals are disproportionately susceptible to severe disease and often experience prolonged recovery and impaired protective immunity ^8,9. CD8⁺ T cells from older individuals exhibit altered mitochondrial metabolism, reduced respiratory capacity, increased oxidative stress, and diminished bioenergetic capacity ^1,10. Aging additionally alters T-cell receptor signaling, epigenetic regulation, cellular differentiation, and tissue-derived signals ^1,11, making it difficult to distinguish the specific contribution of mitochondrial dysfunction from the broader cellular changes associated with aging.

Single-cell transcriptomic analysis of CD8⁺ T cells from young and older adults revealed an age-associated decline in mitochondrial transcriptional programs, including reduced expression of mitochondrial transcription factor A (TFAM) ¹², a finding recapitulated in CD8⁺ T cells from aged mice. TFAM is a nuclear-encoded mitochondrial protein required for mitochondrial DNA packaging, maintenance, transcription, and oxidative phosphorylation ¹³. Reduced TFAM abundance impairs mitochondrial genome integrity and can promote mitochondrial DNA stress, oxidative stress, and inflammatory signaling ^14–16. These observations raised the possibility that age-associated TFAM decline contributes to impaired mitochondrial fitness and altered CD8⁺ T-cell function. However, because aging affects numerous interacting cellular pathways, comparisons between young and older individuals cannot establish whether mitochondrial dysfunction alone is sufficient to alter CD8⁺ T cell responses during an infection.

To isolate the cell-intrinsic contribution of mitochondrial dysfunction, we generated CD8α-Cre Tfam+/− mice (Tfam-haploinsufficient), in which CD8⁺ T cells express a single functional Tfam allele (TFAM+/−). This approach produces a partial reduction in TFAM expression, rather than the severe mitochondrial dysfunction associated with complete TFAM deletion ^1,17, allowing us to selectively examine the consequences of diminished mitochondrial fitness in CD8⁺ T cells from young mice. Consistent with prior studies linking mitochondrial dysfunction to cellular senescence ¹, TFAM insufficiency impaired mitochondrial fitness and promoted senescence-associated features in CD8⁺ T cells. During IAV infection, TFAM-insufficient CD8⁺ T cells developed an early heightened cytotoxic effector phenotype with exhaustion-associated features and were associated with impaired viral control. This was followed by progressive loss of effector function, reduced antigen-specific response, and impaired heterosubtypic protective immunity. Together, these findings identify TFAM-dependent mitochondrial fitness as a cell-intrinsic regulator of antiviral CD8⁺ T cell function and provide a framework linking age-associated mitochondrial decline to impaired protective immunity during respiratory viral infection.

Results

TFAM insufficiency impairs mitochondrial fitness and promotes a senescence-like CD8⁺ T cell phenotype.

To identify molecular features associated with diminished mitochondrial fitness in human CD8⁺ T cells, we analyzed publicly available single-cell RNA-seq datasets from peripheral blood mononuclear cells (PBMCs) of young (age <65 years) and older adults (age ≥65 years) (Fig. S1A) ¹². Participants were stratified using a predetermined age threshold of 65 years, with individuals aged <65 years classified as younger adults and those aged ≥65 years classified as older adults, consistent with age classifications used in World Health Organization guidance ^18,19. CD8⁺ T cells from older individuals exhibited significantly reduced mitochondrial function scores (Fig. 1A). We next examined expression of TFAM, a master regulator of mitochondrial DNA maintenance and oxidative phosphorylation and found that TFAM expression was significantly reduced in CD8⁺ T cells from older adults (Fig. 1B), consistent with previous studies showing age-associated reductions in TFAM and linking reduced TFAM expression to mitochondrial dysfunction¹. Furthermore, TFAM expression positively correlated with mitochondrial function score (Fig. S1B), identifying TFAM as a candidate regulator of diminished mitochondrial fitness in aged human CD8⁺ T cells. We next examined expression of TFAM, a master regulator of mitochondrial DNA maintenance and oxidative phosphorylation and found that TFAM expression was significantly reduced in CD8⁺ T cells from older adults (Fig. 1B). Furthermore, TFAM expression positively correlated with mitochondrial function score (Fig. S1B), identifying TFAM as a candidate regulator associated with diminished mitochondrial fitness in aged human CD8⁺ T cells. Similarly, CD8⁺ T cells from aged (65-week-old) mice expressed significantly lower TFAM mRNA than those from young (12-week-old) mice (Fig. 1C), indicating that reduced TFAM expression is conserved across human and murine aging.

Fig. 1. TFAM insufficiency disrupts mitochondrial fitness and promotes senescence-associated features in CD8⁺ T cells.

(A) Violin plot showing single-cell mitochondrial function scores in CD8⁺ T cells from PBMCs of young and older adults (p< 2 × 10⁻¹⁶). (B) Violin plot showing normalized single-cell TFAM expression in CD8⁺ T cells from PBMCs of young and older adults (p< 2 × 10⁻¹⁶). (C) Quantitative PCR analysis of Tfam mRNA expression in CD8⁺ T cells from wild-type (WT) young (12 weeks) vs old mice (>65 week), normalized to GAPDH (n= 4/group). (D) Schematic generation of CD8⁺ T cell-specific TFAM haploinsufficient mice (TFAM^+/−) by crossing CD8α Cre transgenic mice with floxed TFAM (TFAM^fl/fl) mice. (E) Representative immunofluorescence images and quantification of TFAM protein expression (MFI, green) in naïve splenic CD8⁺ T cells from WT and TFAM^+/− mice. (F) Representative immunofluorescence images and quantification of mitochondrial membrane potential (MFI) in naïve splenic CD8⁺ T cells using MitoSpy CMXRos. (G) Western blot analysis of pSTAT1, IRF3, STING in naïve CD8⁺ T cells, with β-actin as a loading control. (H) Flow cytometric analysis of mitochondrial ROS production in unstimulated and anti-CD3/CD28-stimulated CD8⁺ T cells. (I) ADP/ATP ratio in unstimulated and stimulated CD8⁺ T cells indicating bioenergetic status. (J) Western blot analysis of p16 (INK4a) in naïve (splenic) CD8⁺ T cells, with β-actin as a loading control. (K) Representative immunofluorescence images of lung sections assessing TFAM expression (red) and CD8⁺ T cells (green) at 7 dpi with IAV (PR8, 250 PFU) or mock control, with quantification of TFAM MFI within CD8⁺ T cells (right). MFI was quantified within multiple randomly selected regions of interest (ROIs) across tissue sections from each mouse (n=3/group). (L) Representative immunofluorescence staining and MFI quantification of b-galactosidase (SA-β-gal) activity in lung CD8⁺ T cells (n= 3/group). (M) Flow cytometric analysis and quantification of MitoSpy⁺ in CD8⁺ T cells harvested and sorted from PR8-infected lungs at 7 dpi (n= 3/group). Data represent mean ± SD. In vitro and in vivo experiments in (E-J) are representative of 2–3 independent experiments (n= 3–5 mice per group). For immunofluorescence analyses in (E, F, L and K), MFI was quantified across multiple randomly selected regions of interest/fields of view per sample under identical acquisition settings. Statistical significance was evaluated using Welch’s t-test (C, E, F, L, and M), Mann-Whitney t-test (I) or one-way ANOVA with Tukey’s post-hoc test (K): *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant.

To determine the cell-intrinsic consequences of a partial reduction in TFAM expression as is seen in aging, we generated a CD8⁺ T cell-specific TFAM-haploinsufficient mouse model using Cd8a-Cre-mediated deletion of a single Tfam allele (Fig. 1D). This approach resulted in partial reduction of TFAM protein (Fig. 1E) and mRNA (Fig. S1C), while TFAM expression remained unchanged in CD4⁺ T cells (Fig. S1D), confirming the lineage specificity of the E8I-Cre driver ²⁰. In contrast to complete TFAM deletion, which causes severe mitochondrial dysfunction ^1,17, TFAM haploinsufficient mice produced a partial mitochondrial defect without affecting viability or normal body weight gain through 30 weeks of age. TFAM-insufficient CD8⁺ T cells exhibited reduced mitochondrial membrane potential at baseline, as measured by MitoSpy CMXRos Red (immunofluorescence: Fig. 1F), together with increased activation of mitochondrial DNA stress pathways, including pSTAT1, STING, and IRF3 (Fig. 1G). Following CD3 and CD28 polyclonal activation, TFAM-insufficient CD8⁺ T cells generated substantially greater mitochondrial ROS (Fig. 1H) and displayed impaired bioenergetic function, reflected by an increased ADP/ATP ratio (Fig. 1I). Consistent with previous studies linking mitochondrial dysfunction to cellular senescence ¹, TFAM insufficient CD8⁺ T cells exhibited higher p16 protein expression (INK4a)²¹ (Fig. 1J), indicating acquisition of a senescence-like phenotype. Finally, we examined whether mitochondrial dysfunction and the senescence-like phenotype persisted during respiratory viral infection. Following IAV infection, immunofluorescent colocalization of TFAM and CD8α in lung sections confirmed reduced TFAM expression in lung-infiltrating CD8⁺ T cells from TFAM+/− mice (Fig. 1K). Consistent with these findings, lung CD8⁺ T cells from TFAM+/− mice exhibited increased senescence-associated β-galactosidase staining by immunofluorescence (Fig. 1L), whereas flow cytometric analysis revealed marked mitochondrial hyperpolarization (Fig. 1M), together indicating that TFAM+/− CD8⁺ T cells maintain senescence-associated features and mitochondrial stress during the antiviral response. These findings demonstrate that partial TFAM insufficiency disrupts mitochondrial function, activates mitochondrial stress signaling, and promotes a senescence-like phenotype in CD8⁺ T cells under both homeostatic and IAV infection conditions, establishing a genetic model to define how TFAM insufficiency and resulting mitochondrial dysfunction shape antiviral immunity.

TFAM insufficiency in CD8⁺ T cells exacerbates IAV-induced pulmonary injury and inflammation.

To determine how TFAM insufficiency in CD8⁺ T cells influences disease outcome during respiratory viral infection, littermate WT and TFAM⁺/⁻ mice were infected intranasally with 250 PFU of IAV and analyzed at 7 and 14 days post-infection, corresponding to the peak inflammatory and recovery phases of disease ², respectively (Fig. 2A). WT and TFAM^+/− mice exhibited similar weight-loss kinetics during the acute phase of infection, reaching maximal weight loss around 8–9 dpi. However, TFAM^+/− mice showed delayed weight regain during the recovery phase and remained significantly below their baseline body weight at 14 dpi, whereas WT mice exhibited more substantial recovery (Fig. 2B). TFAM^+/− mice exhibited higher viral load in the lungs at 7 days post-infection. However, by 14 days post-infection, both WT and TFAM^+/− mice had largely cleared the virus (Fig. 2C), indicating that TFAM insufficiency does not impact eventual viral clearance during primary infection. Despite viral clearance by day 14, TFAM^+/− mice exhibited persistent pulmonary pathology. Histological analysis showed persistent inflammatory infiltrates, alveolar wall thickening, increased areas of lung consolidation, and vascular injury in TFAM⁺/⁻ mice compared with WT mice, accompanied by significantly increased inflammation and vascular damage scores (Fig. 2D). TFAM^+/− mice had significantly higher LDH levels in BAL fluid at both 7 and 14 dpi, indicating acute and more sustained pulmonary cellular injury (Fig. 2E). Furthermore, TFAM^+/− mice exhibited significantly higher levels of inflammatory cytokines and chemokines, including IL-6, MCP-1, TNF-α, IFN-β, IL1β, IL-23, and IFN-γ, in the lungs at 7 dpi (Fig. 2F), indicating an increased inflammatory response. Although higher cytokine and LDH levels were evident during the acute phase, histological differences were most evident at 14 dpi, indicating persistent tissue injury during the recovery phase. Although cytokine and LDH levels were elevated during the acute phase, histological differences were most evident at 14 dpi, indicating persistent tissue injury during the recovery phase. Given the persistent lung injury and increased inflammatory response in TFAM⁺/⁻ mice, we next examined T1-alpha expression in the alveolar epithelium. Alveolar type 1 (AT1) cells form the primary gas-exchange surface of the lung ^22,23, and T1-alpha is commonly used as a marker of AT1 cells²². TFAM⁺/⁻ mice exhibited significantly reduced T1-alpha staining at 14 dpi compared with WT mice (Fig. 2G). These findings demonstrate that reduced TFAM expression in CD8⁺ T cells exacerbates pulmonary injury and promotes persistent lung inflammation despite viral clearance, accompanied by reduced T1-alpha expression in the alveolar epithelium during recovery from primary IAV infection.

Fig. 2. TFAM insufficiency in CD8⁺ T cells exacerbates IAV pulmonary injury and inflammation despite viral clearance.

(A) Experimental schematic: WT and TFAM⁺/⁻ mice were intranasally infected with IAV (PR8, 250 PFU) or mock-challenged with PBS and analyzed at 7 and 14 dpi. (B) Disease severity assessed by daily percentage change in body weight relative to baseline over 14 dpi. (C) Lung viral load quantified by TCID₅₀ assay at 7 and 14 dpi. (D) Representative H&E-stained lung sections (left) and semi-quantitative scoring of pulmonary inflammation and vascular injury (right) at 7 and 14 dpi. Scale bars, 150 μm. Arrows indicate areas of dense inflammatory infiltration and alveolar wall thickening. (E) Pulmonary cellular injury assessed by lactate dehydrogenase (LDH) activity in BAL fluid at 7 and 14 dpi. (F) Multiplex quantification of cytokines and chemokines (IL-1β, IL-6, MCP-1, TNF-α, IFN-β, IL-10, IL-23, and IFN-γ) in BAL fluid at 7 dpi. (G) Representative immunofluorescence images and quantification of T1α mean fluorescence intensity (MFI) per unit area in lung sections from mock and IAV-infected mice at 7 and 14 dpi, as a measure of alveolar type I (AT1) epithelial integrity. Scale bars, 400 μm. Data represent mean ± SD. Data in (B, D) are from 4 independent experiments (n=4–5 mice/group). Data in (C, E, F, G) are representative of 2 independent experiments (n= 4–5 mice/group). For immunofluorescence analysis in (G), MFI was quantified across multiple random high-power fields per lung section per mouse. Statistical significance was evaluated using unpaired Student’s t-test, one-way or two-way ANOVA with Tukey’s post-hoc test: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant.

Single-cell transcriptomic profiling identifies hyper-cytotoxic and exhaustion-associated states in TFAM-insufficient CD8⁺ T cells.

To define how TFAM insufficiency alters CD8⁺ T cell responses during the peak effector phase of influenza virus infection, we performed single-cell RNA sequencing of lung cells from WT and TFAM⁺/⁻ mice at 7 dpi. Unsupervised UMAP analysis identified the major immune and stromal populations in IAV-infected lungs, including naive and effector CD8⁺ T cells, proliferating T cells, inflammatory monocytes, alveolar and tissue-resident macrophages, endothelial and epithelial cells, fibroblasts, neutrophils, B cells, and plasma cells, with broadly comparable cellular composition between WT and TFAM^+/− mice (Fig. 3A, Fig. S2A-D). We next focused on the CD8⁺ T cell compartment to determine how TFAM insufficiency reshaped the effector response, which is essential for viral control but can also drive pulmonary immunopathology when excessive or dysregulated, as demonstrated by our previous work and others ^2,24–27. Compared with WT cells, TFAM⁺/⁻ effector CD8⁺ T cells showed enrichment of cytotoxic and inflammatory genes, including Gzmb, Prf1, Ifng, Ccl4, Il2ra, Klrc1, and Tnfrsf1b (Fig. 3B, Fig. S3A). Analysis of functionally defined gene programs further revealed higher transcriptional levels of effector-associated genes (Gzmb, Prf1, Ifng, and Nkg7), exhaustion-associated genes (Pdcd1, Havcr2, Lag3, Nr4a1, and Tox), and glycolytic and metabolic-stress genes (Hk2, Ldha, Slc2a1, and Hif1a). Several oxidative phosphorylation-associated genes, including Ndufs1, Sdha, Uqcrc2, and Cox5a, showed increased expression in TFAM⁺/⁻ CD8⁺ T cells, whereas Atp5f1a expression was reduced, suggesting a dysregulated mitochondrial transcriptional program rather than uniform suppression of OXPHOS genes (Fig. 3C). Consistent with these transcriptional patterns, TFAM^+/− CD8⁺ T cells displayed significantly higher metabolic dysfunction and exhaustion module scores (Fig. 3D). Gene-set enrichment analysis further identified enrichment of cytokine-cytokine receptor interaction, natural killer cell-mediated cytotoxicity, antigen processing and presentation, and cellular senescence pathways in TFAM^+/− effector CD8⁺ T cells (Fig. 3E). These findings demonstrate that TFAM insufficiency promotes a hyper-cytotoxic effector state characterized by a shift away from oxidative phosphorylation toward metabolic dysfunction and the early acquisition of exhaustion-associated transcriptional features.

Fig. 3. TFAM insufficiency promotes early cytotoxic and dysfunctional CD8⁺ T cell programming during influenza infection.

(A) UMAP visualization of single-cell RNA-seq data from total lung cells of WT and TFAM⁺/⁻ mice following IAV infection, identifying major immune and structural cell populations, including naïve, effector, and proliferating CD8⁺ T cell subsets. (B) Heatmap showing expression of genes associated with T cell activation and effector function across naïve, effector, and proliferating CD8⁺ T cell subsets from WT and TFAM⁺/⁻ mice at 7 days post-infection (dpi). (C) Dot plot showing the average expression and percentage of expressing cells for genes associated with oxidative phosphorylation (OXPHOS), effector function, exhaustion, and metabolic dysfunction in WT and TFAM⁺/⁻ CD8⁺ T cells at 7 dpi. (D) Violin plots showing metabolic dysfunction and exhaustion module scores in WT and TFAM⁺/⁻ CD8⁺ T cells at 7 dpi. (E) Gene set enrichment analysis (GSEA) comparing TFAM⁺/⁻ and WT CD8⁺ T cells at 7 dpi, demonstrating enrichment of pathways associated with antigen processing and presentation, cellular senescence, cytokine-cytokine receptor interactions, and natural killer cell-mediated cytotoxicity in TFAM⁺/⁻ CD8⁺ T cells. Mouse scRNA-seq data were generated from pooled lung single-cell suspensions (n= 4 mice/group). Statistical significance for module scores was evaluated using Wilcoxon rank-sum test, as appropriate: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant.

To determine whether the transcriptional responses identified by single-cell analysis were accompanied by functional changes in vivo, we analyzed lung CD8⁺ T cell responses by flow cytometry at 7 dpi. Although the total number of CD8⁺ T cells were comparable between WT and TFAM^+/− mice, TFAM^+/− CD8⁺ T cells expressed significantly higher levels of CD44, granzyme B, and perforin (Fig. 4A). Following ex vivo stimulation with influenza NP_366–374 peptide, TFAM^+/− CD8⁺ T cells exhibited a significantly higher number of IFN-g⁺ cells than WT CD8⁺ T cells (Fig. 4B). We next examined whether the transcriptional signatures of exhaustion and metabolic dysfunction identified by single-cell analysis were accompanied by phenotypic evidence of early CD8⁺ T-cell exhaustion in vivo. Flow cytometric analysis demonstrated significantly increased expression of the exhaustion-associated markers PD-1, LAG-3, and CD39 in TFAM⁺/⁻ CD8⁺ T cells from lungs at 7 dpi (Fig. 4C)²⁸. Together, these findings demonstrate that TFAM insufficiency promotes a heightened cytotoxic phenotype accompanied by the early development of exhaustion-associated features during the peak CD8⁺ T cell response.

Fig. 4. TFAM insufficiency drives an exaggerated cytotoxic response and increased expression of inhibitory receptors on CD8⁺ T cells during acute IAV infection.

(A) Representative flow cytometry plots and quantification of total lung CD3⁺CD8⁺ T cells, activated CD44⁺CD8⁺ T cells, and GZMB⁺ and Perforin⁺ CD8⁺ T cells at 7 dpi with PR8 (250 PFU) or mock treatment. Cell numbers are presented as counts per 50,000 events. (B) Representative flow cytometry plots and quantification of IFN-γ⁺ CD8⁺ T cells following ex vivo recall stimulation with influenza NP_366–374 peptide at 7 dpi. Cell numbers are presented as counts per 50,000 events. (C) Representative flow cytometry plots and quantification of CD39⁺, LAG-3⁺, and PD-1⁺ lung CD8⁺ T cells at 7 dpi. Cell numbers are presented as counts per 100,000 events. Data for recruitment, activation, cytotoxicity, and recall IFNγ production (Panels A and B) are of 3–4 independent experiments (n-6–8 mice/group). Exhaustion marker profiling (Panel C) is of two independent experiments (n-6–8 mice/group). Statistical significance was evaluated using one-way ANOVA with Tukey’s post-hoc test: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant.

Early exhaustion-associated features in TFAM-insufficient CD8⁺ T cells precede impaired antigen-specific responses during the recovery phase.

To determine whether the early acquisition of exhaustion-associated features progressed to functional impairment, we next examined CD8⁺ T cell responses at 14 dpi. Single-cell transcriptomic analysis revealed a distinct transcriptional state in TFAM^+/− effector CD8⁺ T cells, characterized by reduced expression of mitochondrial and effector-associated genes compared with WT cells (Fig. 5A). TFAM^+/− CD8⁺ T cells exhibited reduced expression of genes associated with metabolic function, oxidative phosphorylation, and cytotoxic effector function, accompanied by significantly lower mitochondrial function and effector-function module scores (Fig. 5B, C). Relative to WT, gene set enrichment analysis revealed negative enrichment of cell adhesion molecule and ribosomal pathways in TFAM⁺/⁻ CD8⁺ T cells (Fig. 5D). These transcriptional changes indicate a transition from the heightened cytotoxic phenotype observed at 7 dpi toward impaired mitochondrial and effector response during the recovery phase of infection. Consistent with the scRNA-seq findings, flow cytometric analysis showed comparable frequencies of total CD8⁺ T cells between WT and TFAM^+/− mice, whereas the frequencies of CD44⁺ and granzyme B⁺ CD8⁺ T cells were significantly reduced in TFAM^+/− mice at 14 dpi (Fig. 5E). To directly assess antigen-specific CD8⁺ T-cell response, lung cells were restimulated ex vivo with influenza NP_366–374 peptide, and intracellular IFN-γ expression was measured by flow cytometry. TFAM^+/− mice exhibited a significantly reduced frequency of IFN-γ⁺ CD8⁺ T cells following IAV peptide stimulation (Fig. 5F). Furthermore, TFAM⁺/⁻ mice exhibited a significantly reduced frequency of influenza-specific H-2D^b/NP_366–374 (ASNENMETM) tetramer⁺ CD8⁺ T cells. CD8⁺ T cells at the 14 dpi timepoint (Fig. 5G), indicating diminished antigen-specific CD8⁺ T cell responses during the recovery phase. Together, these findings demonstrate that the heightened cytotoxic and exhaustion-associated phenotype observed at the peak CD8⁺ T cell response is followed by progressive loss of effector function and diminished antigen-specific CD8⁺ T cell responses during the recovery phase of IAV infection.

Fig. 5. TFAM insufficiency drives a transition from early hyper-cytotoxicity to impaired effector and antigen-specific CD8⁺ T cell responses during recovery from influenza infection.

(A) Heatmap showing differential gene expression profiles across CD8⁺ T cell subsets (Naïve, Effector, Proliferating) in WT and TFAM⁺/⁻ mice at 14 days post-infection (dpi) with PR8. (B) Dot plot showing average gene expression and percentage of expressing cells across key metabolic, exhaustion, effector, and oxidative phosphorylation (OXPHOS) gene signatures at 14 dpi. (C) Transcriptomic module scores for Effector Function (p< 0.0001) and Mitochondrial Function (p < 0.001) in CD8⁺ T cells at 14 dpi. (D) Gene Set Enrichment Analysis (GSEA) showing significant downregulation of Cell Adhesion Molecule and Ribosome pathways in TFAM⁺/⁻ CD8⁺ T cells compared with WT controls at 14 dpi. (E) Representative flow cytometry plots and frequency quantification of total lung CD3⁺CD8⁺ T cells, activated CD44⁺CD8⁺ T cells, and cytotoxic GZMB⁺CD8⁺ T cells at 14 dpi in WT-Mock, TFAM⁺/⁻-Mock, WT-PR8, and TFAM⁺/⁻-PR8 groups. (F) Representative flow cytometry plots and frequency quantification of intracellular IFN-γ production in lung CD8⁺ T cells following ex vivo recall stimulation with influenza NP_366–374 peptide at 14 dpi. (G) Representative flow cytometry plots and frequency quantification of influenza Tetramer⁺ lung CD8⁺ T cells at 14 dpi. Data represent mean ± SEM. Mouse single-cell transcriptomic analyses (A-D) were performed using lung single-cell suspensions (n= 5 mice/group). In vivo flow cytometry analyses (E-F) are representative of 2 independent experiments (n = 6–7 mice/group) and Panel G is representative of 3 independent experiment (n = 6–7 mice/group). Statistical significance for module scores was evaluated using the Wilcoxon rank-sum test, and flow cytometry data were analyzed using one-way ANOVA with Tukey’s post hoc test: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant.

TFAM insufficiency compromises CD8⁺ T cell-mediated antiviral protection and heterosubtypic immunity.

To determine whether the reduced antigen-specific CD8⁺ T cell response observed in TFAM⁺/⁻ mice at 14 dpi translates into diminished antiviral protection in vivo, we adoptively transferred CD8⁺ T cells isolated from IAV-infected WT or TFAM⁺/⁻ mice into TCRβ-deficient recipient mice (Fig. 6A). A total of 1 × 10⁶ CD8⁺ T cells isolated from WT or TFAM^+/− mice at 14 dpi were transferred retro-orbitally into TCRβ deficient (TCRβ^−/−) recipient mice prior to IAV challenge (Fig. 6A). Because TCRβ^−/− mice lack endogenous αβ T cells, donor-derived CD8⁺ T cells constitute the primary adaptive T cell response following infection, enabling direct assessment of CD8⁺ T cell response without interference from recipient T cells²⁹. Following IAV infection, adoptive transfer of WT CD8⁺ T cells completely protected TCRβ⁻/⁻ mice from weight loss, with low viral load detected in the lung. In contrast, mice receiving TFAM⁺/⁻ CD8⁺ T cells exhibited greater weight loss and significantly higher lung viral load (Fig. 6B, C). Furthermore, compared with mice receiving WT CD8⁺ T cells, TCRβ⁻/⁻ mice receiving TFAM⁺/⁻ CD8⁺ T cells exhibited a significantly reduced frequency of IAV tetramer⁺ CD8⁺ T cells following viral challenge, indicating a diminished antigen-specific recall response (Fig. 6D). Thus, the adoptive-transfer experiments establish that the defects associated with TFAM insufficiency are intrinsic to CD8⁺ T cells and result in impaired antigen-specific CD8⁺ T cell responses.

Fig. 6. TFAM insufficiency compromises CD8⁺ T cell-mediated antiviral protection and heterosubtypic immunity.

(A) Experimental schematic for adoptive transfer: CD8⁺ T cells isolated from WT or TFAM⁺/⁻ donor mice at 14 dpi with PR8 (250 PFU) were adoptively transferred into TCRβ KO recipient mice, followed by challenge with PR8 (25 PFU) and analysis at day 6 post-challenge. (B) Disease severity in recipient TCRβ KO mice assessed by daily percentage body weight change relative to initial body weight. (C) Viral load quantified by TCID₅₀ assay in lung homogenates of recipient TCRβ KO mice at day 6 post-challenge. (D) Representative flow cytometry plots and frequency quantification of IAV tetramer⁺ lung CD8⁺ T cells in recipient TCRβ KO mice at day 6 post-challenge. (E) Experimental schematic for heterosubtypic challenge: WT and TFAM⁺/⁻ mice were infected with PR8 (250 PFU) or mock-challenged with PBS, rested until day 30, and subsequently challenged with heterosubtypic influenza A/X31, followed by analysis at day 6 post-secondary infection. (F) Disease severity assessed by daily percentage body weight change following secondary X31 challenge. (G) Lung viral load quantified by TCID₅₀ assay at day 6 post-X31 challenge. (H) Representative flow cytometry plots and frequency quantification of total CD3⁺CD8⁺ T cells, activated CD44⁺CD8⁺ T cells, and GZMB⁺CD8⁺ T cells in the lungs following secondary X31 challenge. (I) Representative flow cytometry plots and frequency quantification of IFN-γ⁺ lung CD8⁺ T cells following ex vivo recall stimulation with influenza NP_366–374 peptide at day 6 post-X31 challenge. (J) Representative flow cytometry plots and frequency quantification of IAV tetramer⁺ lung CD8⁺ T cells following secondary X31 challenge. Data represent mean ± SEM. Adoptive transfer in vivo experiments represent data from 2–3 independent experiments (3–4 mice/group). Heterosubtypic X31 challenge assays (Panels F–J) represent a single independent experiment (6–8 mice/group). Statistical significance was evaluated using unpaired Student’s t-test, one-way or two-way ANOVA with Tukey’s post-hoc test: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant.

We next determined whether TFAM insufficiency also compromised the development of durable heterosubtypic CD8⁺ T cell immunity against a distinct influenza subtype. WT and TFAM^+/− mice were initially infected with PR8 (250 PFUs) and, after recovery, rechallenged 30 days later with the heterosubtypic X31 (H3N2: 10000 PFUs) influenza (Fig. 6E), a model that assesses CD8⁺ T cell-mediated cross-protective immunity against conserved influenza antigens independent of neutralizing antibodies³⁰. Following X31 challenge, PR8-immune WT mice were largely protected from infection-associated weight loss, whereas TFAM^+/− mice exhibited significantly greater weight loss and higher viral load (Fig. 6F, G), indicating impaired heterosubtypic protection. The increased morbidity in TFAM⁺/⁻ mice was accompanied by reduced frequencies of activated CD44⁺ and granzyme B⁺ CD8⁺ T cells, although total CD8⁺ T cell frequencies were comparable between the two groups (Fig. 6H). Furthermore, ex vivo stimulation with influenza NP_366–374 peptide revealed a significantly reduced frequency of IFN-γ⁺ CD8⁺ T cells, together with a reduced frequency of influenza tetramer⁺ CD8⁺ T cells in TFAM⁺/⁻ mice (Fig. 6I, J). These findings demonstrate that TFAM insufficiency compromises the antigen-specific CD8⁺ T cell recall response following heterosubtypic challenge. Together, these findings indicate that the early transition from hyper-cytotoxic to exhaustion-associated dysfunction during primary infection compromises CD8⁺ T cell-mediated protection and heterosubtypic immunity. These complementary models further identify TFAM insufficiency and associated mitochondrial dysfunction as a cell-intrinsic mechanism that compromises the protective capacity of antigen-experienced CD8⁺ T cells and the development of long-term heterosubtypic immunity.

Discussion

In this study, we used a CD8⁺ T cell-specific TFAM-haploinsufficient model to selectively disrupt mitochondrial fitness in a young CD8⁺ T cell compartment, thereby isolating the consequences of mitochondrial dysfunction from the broader cellular and molecular changes that accompany aging. Notably, selective TFAM reduction was itself sufficient to promote mitochondrial stress and senescence-associated features, indicating that features associated with CD8⁺ T cell aging can emerge downstream of impaired mitochondrial fitness even in a young cellular background. During IAV infection, TFAM insufficiency produced a distinct temporal shift in CD8⁺ T cell function, characterized by an early increase in cytotoxic effector markers that failed to improve viral control and was associated with persistent lung injury, followed by progressive loss of antigen-specific responses and protective capacity. These defects ultimately compromised heterosubtypic immunity following secondary IAV challenge. Collectively, our findings identify TFAM-dependent mitochondrial fitness as a central regulator of the balance between protective and pathogenic CD8⁺ T cell immunity and provide a mechanistic framework linking age-associated mitochondrial decline to CD8⁺ T cell dysfunction during respiratory viral infection.

Although mitochondrial fitness is recognized as a critical regulator of T cell metabolism and function ^3,31, the cell-intrinsic contribution of mitochondrial dysfunction to CD8⁺ T cell responses during acute respiratory viral infection remains incompletely understood. Previous studies have linked mitochondrial dysfunction to T cell senescence, exhaustion, and impaired memory persistence in the settings of aging, chronic infection, and cancer ^1,5,6,10. Our findings expand these observations by demonstrating that impaired mitochondrial fitness alters CD8⁺ T cell responses during acute respiratory virus infection. Importantly, the phenotypic alterations observed during IAV infection were preceded by defects in baseline mitochondrial homeostasis. TFAM insufficiency increased oxidative stress, mitochondrial stress signaling, and markers of cellular senescence, consistent with disruption of the bioenergetic response required to sustain effector CD8⁺ T cell responses. Single-cell transcriptomic analysis at the peak effector phase of infection (7 dpi) revealed that TFAM insufficiency promoted a heightened cytotoxic effector phenotype accompanied by enrichment of metabolic dysfunction and exhaustion-associated gene signatures. These transcriptional changes were supported by increased expression of granzyme B, perforin, and IFN-γ, together with the exhaustion-associated markers PD-1, LAG-3, and CD39. These findings indicate that impaired mitochondrial fitness promotes a heightened cytotoxic effector phenotype with exhaustion-associated features. Notably, this phenotype occurred in the setting of impaired viral control, indicating that increased expression of cytotoxic and effector molecules did not translate into more effective antiviral immunity. We previously identified hyper-cytotoxic CD8⁺ T cell subsets as key mediators of IAV-induced pulmonary immunopathology ². The enhanced lung injury observed in TFAM⁺/⁻ mice further supports a link between reduced TFAM expression, dysregulated CD8⁺ T cell effector responses, and pulmonary immunopathology.

Longitudinal analysis demonstrated that the early heightened cytotoxic effector phenotype was followed by broad suppression of mitochondrial and effector-associated responses during the recovery phase (14 dpi), accompanied by reduced antigen-specific CD8⁺ T cell responses. Functionally, these changes translated into reduced CD8⁺ T cell-intrinsic antiviral protection, as adoptive transfer of TFAM-insufficient CD8⁺ T cells conferred significantly less protection to TCRβ^−/− recipient mice and was associated with reduced frequencies of IAV tetramer-specific CD8⁺ T cells. TFAM insufficiency also markedly compromised heterosubtypic immunity, resulting in greater morbidity following secondary X31 challenge, accompanied by reduced frequencies of tetramer-specific CD8⁺ T cells and diminished granzyme B and IFN-γ responses. These findings suggest that the dysfunctional state established during the primary immune response compromises the protective capacity of antigen-experienced CD8⁺ T cells and durable antiviral immunity. This is consistent with previous studies demonstrating that mitochondrial stress shortens the effector phase and promotes premature T cell dysfunction ^5,32. Similar findings have been reported during respiratory viral infection in aged hosts, where antigen-specific CD8⁺ T cell effector function and memory formation are impaired ^8,31,33. These findings implicate impaired mitochondrial fitness as a cell-intrinsic contributor to progressive CD8⁺ T cell dysfunction and loss of protective antiviral immunity following IAV infection.

Our TFAM haploinsufficiency model allowed mitochondrial dysfunction to be examined independently of the numerous biological changes that accompany aging. Aging affects multiple aspects of CD8⁺ T cell immunity, including thymic output ^11,34, inflammatory milieu^34,35, metabolic function ³⁶, and antigen presentation ^37,38, making it challenging to define the cell-intrinsic mechanisms responsible for T cell dysfunction during respiratory viral infection. By selectively reducing TFAM expression in CD8⁺ T cells, we show that impaired mitochondrial fitness is sufficient to recapitulate several features associated with immune aging, including a heightened cytotoxic phenotype with exhaustion-associated features, reduced antigen-specific response, and compromised heterosubtypic recall immunity. Although mitochondrial dysfunction is unlikely to account for the full complexity of immune aging, our findings support a role for age-associated TFAM decline in reduced mitochondrial fitness and dysregulated CD8⁺ T cell responses during respiratory viral infection. Collectively, our findings establish TFAM-dependent mitochondrial fitness as a key cell-intrinsic regulator of CD8⁺ T cell immunity and provide a mechanistic framework for understanding its role in balancing protective antiviral responses with CD8⁺ T cell-mediated immunopathology during respiratory viral infection.

Material and methods

Mice, viral strains, and influenza models.

Influenza A virus (IAV) strains A/Puerto Rico/8/1934 (H1N1; PR8) and A/X-31 (H3N2; X31) were purchased from Charles River Laboratories (Norwich, CT, USA), and viral titers were determined prior to experimental use ^2,39. Wild-type (WT) C57BL/6, TFAM^fl/fl, CD8α-Cre (JAX# 008766) and TCRβ−/− (TCRβ KO) mice were purchased from the Jackson Laboratory and bred in-house. In CD8a Cre-expressing strain, unlike the proximal Cd8a promoter, the E8I enhancer is not active during the double-positive (DP) stage of thymocyte development. Its activity is restricted until the cells have committed to the CD8 single-positive (SP) lineage, allowing Cre-mediated deletion to occur only in mature CD8⁺ T cells and avoids any ‘off-target’ recombination in the CD4 lineage ²⁰ (Fig. S1D). An equal number of male and female mice, aged 8–10 weeks, were included in the study. All animal experiments were conducted under a protocol approved by the University of Florida Institutional Animal Care and Use Committee (IACUC; protocol #IACUC202100000057). Mice were provided ad libitum access to food and water. CD8α-Cre mice were crossed with TFAM^fl/fl mice to achieve a heterozygous TFAM deletion in CD8α-expressing cells, resulting in TFAM^+/− mice. For mouse infection studies, the influenza A virus (IAV) strains A/Puerto Rico/8/1934 (H1N1; PR8) and A/X-31 (H3N2) were used. Mice were intranasally infected with 250 plaque-forming units (PFU) of PR8 or 10,000 PFU of X31 in 50 μL PBS under 4% (v/v) isoflurane anesthesia. Mock-infected mice received 50 μL PBS. Following euthanasia by CO₂ exposure and cervical dislocation (6, 7 or 14 dpi), the lungs were perfused, bronchoalveolar lavage fluid (BALF) was collected by instillation and aspiration of 1 mL sterile ice-cold PBS through a 20-gauge tracheal catheter (BD Biosciences), and lungs were aseptically harvested for downstream analyses.

Histopathology.

Following euthanasia and lung perfusion, the left lung lobe was fixed in 10% neutral-buffered formalin (pH 7.4) for 24 hours at room temperature and then transferred to 70% ethanol. Lung tissues were embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin and eosin (H&E). H&E staining was performed at the University of Florida Histology Core. Lung sections were evaluated independently by three blinded pathologists using a semiquantitative scoring system ranging from 0 to 5 in 0.5-point increments, where 0 indicated no pathological changes and 5 represented the highest degree of pathology ². Histopathological assessment included inflammatory cell infiltration and vascular injury, including endothelial vacuolation, as previously described ⁴⁰.

Flow Cytometry.

The spleens and lungs were aseptically collected, minced, and processed into single-cell suspensions. Lung tissue was enzymatically digested with collagenase D as previously described ⁴⁰. For flow cytometry, one million single cells per sample were stained with Zombie Red/Aqua live-dead dye in phosphate-buffered saline (PBS). Cell surface staining was performed by incubating cells for 30 minutes at room temperature in staining buffer (2% FBS) with fluorescently conjugated anti-mouse antibodies against CD3, CD4, CD8, CD44, CD62L, PD-1, CD39, LAG3 and APC-Conjugated Tetramer. Intracellular staining was performed using two distinct protocols. For the detection of cytokines and cytotoxic molecules, cells were fixed and permeabilized using the BD Cytofix/Cytoperm^™ kit and stained with anti-mouse Granzyme B, Perforin, and IFN-γ. For TFAM detection, cells were fixed and permeabilized using the FoxP3 Transcription Factor Staining Buffer. The cells were then stained with unconjugated rabbit anti-mouse TFAM primary antibody, followed by an AF594-conjugated anti-rabbit IgG secondary antibody. To assess antigen-specific response (IFN-γ), lung single cells from mock and IAV-infected mice were stimulated ex vivo with 1 μM IAV nucleoprotein peptide (NP_366–374) (NIH) for 6 hours in the presence of Brefeldin A in complete RPMI-1640 medium supplemented with 10% FBS and penicillin/streptomycin. The sample acquisition was performed on a Sony ID7000, Cytek Aurora, or an Attune NxT acoustic-focusing flow cytometer. Data was analyzed using FlowJo software (version 10). The list of the reagents used in this study, along with their catalog number, is provided in the Dataset S1.

Cytokine Analysis.

Cytokine protein levels in BALF from mock and IAV-infected mice were quantified using the LEGENDplex Mouse Inflammation Panel (BioLegend) according to the manufacturer’s instructions. Samples were acquired on an Attune NxT Acoustic Focusing Cytometer (Thermo Fisher Scientific), and data were analyzed using the LEGENDplex Data Analysis Software (BioLegend).

Lactate Dehydrogenase (LDH) Activity Assay.

LDH levels in BALF from mock and IAV-infected mice were quantified using a colorimetric LDH Cytotoxicity Assay Kit according to the manufacturer’s instructions. Absorbance was measured at 490 nm using a Varioskan^™ LUX multimode microplate reader ².

Mitochondrial Membrane Potential Assessment.

Splenic CD8⁺ T cells from naive mice and lung CD8⁺ T cells from day 7 post-infection mice were isolated using magnetic cell sorting (Miltenyi Biotec). For mitochondrial membrane potential assessment, isolated CD8⁺ T cells (1 × 10⁶ cells) were stained with MitoSpy Orange CMTMRos or MitoSpy Red CMXRos (BioLegend) according to the manufacturer’s instructions.

Intracellular ROS Quantification.

For intracellular reactive oxygen species (ROS) measurement, magnetically sorted splenic CD8⁺ T cells (1 × 10⁶ cells/well) were seeded into tissue culture plates pre-coated with anti-CD3 antibody and stimulated in the presence of soluble anti-CD28 antibody (Invitrogen) for 24 hours. Following activation, cells were stained with the CellROX Deep Red Flow Cytometry Assay Kit (Thermo Fisher Scientific) following the manufacturer’s protocol.

Immunofluorescence.

Formalin-fixed, paraffin-embedded lung sections from mock and IAV-infected mice were prepared as previously described ⁴⁰ and stained with primary anti-mouse antibodies against CD8α and TFAM for 2 hours. Sections were then incubated with the fluorophore-conjugated secondary antibodies diluted in 5% donkey serum in PBS for 1 hour at room temperature, followed by nuclear counterstaining with DAPI. Colocalization analysis was performed using NIH FIJI (ImageJ); intensity profiles of defined regions of interest (ROIs) were generated with the Plot Profile tool, and co-localization mean fluorescence intensity (MFI) was quantified using the Coloc2 plugin. Immunocytochemistry was performed on glass chambered slide as previously described ^41,42 with magnetically sorted CD8⁺ T cells. Cells were stained with TFAM (Abcam), MitoSpy CMXROS (Biolegend), and B-gal (SPiDER β-Gal), followed by counterstaining with DAPI (Invitrogen). Alveolar type 1 (AT1) cells were identified by immunofluorescence staining for anti-T1 alpha (DSHB) on lung sections. Images were acquired using a Zeiss LSM 710 confocal microscope (University of Florida, Center for Immunology and Transplantation) and analyzed using NIH FIJI (ImageJ) software. Mean fluorescence intensity (MFI) and expression area were quantified from 10 predefined unbiased ROIs/sample.

Western Blot.

Magnetically sorted splenic CD8⁺ T cells were lysed in RIPA or M-PER Mammalian Protein Extraction Reagent (Thermo Scientific) supplemented with protease and phosphatase inhibitor cocktails. Protein concentration was determined, and 20 ug of protein lysate per lane was resolved by SDS-PAGE before being transferred onto Nitrocellulose membranes. Membranes were blocked and incubated overnight at 4°C with primary antibodies targeting pSTAT1 (Cell Signaling Technology), STING (Cell Signaling Technology), IRF3 (Abcam), p16 (Abcam), and β-actin (Abcam) as a loading control. Blots were subsequently incubated with appropriate secondary-conjugated antibodies and visualized using enhanced chemiluminescence ^42,43.

Quantitative PCR (qPCR).

Total RNA was extracted from magnetically sorted splenic CD8⁺ T cells isolated from naïve 8-week-old (young) and 65-week-old (aged) WT mice using the Quick-RNA MiniPrep Kit (Zymo Research) according to the manufacturer’s protocol. First-strand cDNA synthesis was performed using qScript Ultra SuperMix (Quantabio). qPCR was conducted to measure Tfam mRNA expression using PerfeCTa SYBR Green FastMix (Quantabio) on a CFX96 Touch Real-Time PCR Detection System equipped with a C1000 Touch Thermal Cycler (Bio-Rad). Relative gene expression levels were normalized to housekeeping gene control (Gapdh) using the 2^{− ΔΔCT} Method ⁴⁰.

ADP/ATP quantification.

Magnetically sorted splenic CD8⁺ T cells from naïve WT and TFAM^+/− mice were stimulated on plates pre-coated overnight with anti-CD3 antibody in the presence of soluble anti-CD28 antibody for 24 hours. Cellular ATP and ADP levels were then measured using a luminescence-based ADP/ATP ratio assay kit according to the manufacturer’s instructions (Sigma). Luminescence was measured using a Varioskan^™ LUX multimode microplate reader.

Viral load.

Lung viral load was quantified by 50% tissue culture infectious dose TCID₅₀ endpoint dilution assay. The right lung lobes were homogenized in infection medium, normalized to tissue weight, and stored at −80°C until titration. Madin-Darby canine kidney (MDCK) cells were seeded in 96-well plates at 3×10³ cells/well and grown to 80% to 90% confluence. Clarified lung homogenate supernatants were serially diluted in DMEM supplemented with 0.3% bovine serum albumin (BSA), 0.0002% TPCK-treated trypsin (Worthington Biochemical), and penicillin-streptomycin. Serial dilutions were added in triplicate to MDCK monolayers. Following incubation for 4–6 days at 37°C with 5% CO2, cells were fixed with formaldehyde and stained with 0.1% crystal violet solution to visualize cytopathic effects (CPE). Viral titers were calculated according to the Spearman-Kärber method and expressed as TCID₅₀/mL of lung homogenate normalized to tissue weight ^2,43.

Adoptive Transfer CD8⁺ T Cells and Secondary Challenge.

WT and TFAM^+/− donor mice were infected intranasally with 250 PFU of PR8 in sterile PBS. At 14 days post-infection, donor mice were euthanized, and lungs were harvested following systemic perfusion with sterile PBS. Single-cell suspensions were prepared by enzymatic digestion of lung tissue, and CD8⁺ T cells were isolated by magnetic cell sorting according to the manufacturer’s instructions. Recipient TCRβ^−/− mice were infected intranasally with 25 PFU of PR8 or mock infected with PBS. Twenty-four hours later, 1×10⁶ purified lung CD8⁺ T cells isolated from WT or TFAM^+/− donor mice were adoptively transferred into recipient TCRβ^−/− mice by retro-orbital injection ². Control TCRβ^−/− mice received an equal volume of sterile PBS. Recipient mice were euthanized at 6 days post-infection, and lungs were collected for downstream analyses, including viral load and immune response.

Single-cell RNA-seq data processing and analysis.

Single-cell RNA sequencing was performed on lung single-cells using the Parse Biosciences Evercode^™ WT v3 platform, following the manufacturer’s instructions and as previously described⁴⁴. Briefly, lungs were harvested from mock and IAV-infected WT and TFAM^+/− mice (n=4–5 mice per group: male and female) and processed into single-cell suspensions. Viable cells were isolated and approximately 25k cells per sample were used for library preparation. Single-cell libraries were generated using Parse Biosciences split-pool combinatorial barcoding chemistry, which enables cell fixation, barcoding, and library construction through successive rounds of molecular indexing. Cell fixation, reverse transcription, barcoding, and library amplification were performed according to the Parse protocol. Barcoded libraries were pooled and sequenced on an Illumina platform to a depth of approximately 56,891 read per cell. scRNA-seq data were generated using a combinatorial indexing-based protocol and processed with split-pipe (v1.4.2). Raw sequencing reads were aligned to the Mus musculus GRCm39 reference genome using STAR (v2.7.11b)⁴⁵, and gene-cell expression matrices were generated following barcode correction and filtering. Filtered count matrices were imported into R and analyzed using Seurat (v 5.0.2)⁴⁶. Cells with fewer than 200 detected genes and genes detected in fewer than three cells were excluded. Low-quality cells were further removed based on mitochondrial and ribosomal transcript content, library size, and gene complexity, retaining cells with 200–6,000 detected genes and <25% mitochondrial transcripts⁴⁷ (SI Appendix, Fig. S2A). Putative doublets were identified and removed on a per-sample basis using Doublet Finder (v 2.22.0) ⁴⁸ after standard preprocessing. Each sample was log-normalized and processed independently to identify highly variable genes, followed by scaling and principal component analysis. Samples were merged into a single dataset, and batch effects across samples were corrected using Harmony ⁴⁹, with sample identity used as the integration variable. Clustering resolution was evaluated across a range of values, and a final resolution of 0.6 was selected for downstream analyses. Two-dimensional visualization was performed using UMAP based on the Harmony-corrected embedding. Cluster marker genes were identified using FindAllMarkers, and clusters were annotated based on canonical lineage marker expression ². Differentially expressed genes were defined as those with an adjusted P value <0.01 and an absolute log2 fold change > 0.25. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed using richR package (https://github.com/hurlab/richR/).

Module Score analysis.

Four module scores, mitochondrial function, metabolic dysfunction, CD8⁺ T cell exhaustion, and effector function, were computed for each cell using AddModuleScore with curated gene sets in the Seurat v4 pipeline ⁵⁰. To compare the expression of key genes spanning metabolic (Hk2, Ldha, Slc2a1, Hif1a), effector/cytotoxic (Gzmb, Prf1, Nkg7, Ifng), inhibitory receptor/exhaustion transcription factor (Pdcd1, Havcr2, Lag3, Tox, Nr4a1), and OXPHOS complex (Ndufs1, Sdha, Uqcrc2, Cox5a, Atp5f1a) categories between WT and TF genotypes, dot plots were generated in which dot size represents the percentage of cells expressing each gene and color intensity represents the average normalized expression level. All analyses were conducted using the Seurat v4 pipeline ⁵⁰ in R.

Human single-cell RNA-seq data acquisition and processing (GSE136184).

Single-cell RNA-seq dataset GSE136184 ¹², derived from peripheral blood mononuclear cells (PBMCs) of healthy human donors, was obtained from the Gene Expression Omnibus (GEO) database ⁵¹. The dataset comprises 18 samples spanning a broad age range (26–85 years). Cells were processed through the Seurat pipeline (v4;) ⁵⁰ with default parameters. For each cell, the mitochondrial function module score was computed using AddModuleScore with a predefined gene set, and TFAM expression was extracted as normalized counts. Group-level comparisons between Young and Old donors were performed via the Wilcoxon rank-sum test. In the correlation analysis, TFAM expression was regressed against the mitochondrial function score at the single-cell level and visualized with a regression line and 95% confidence interval.

Data Availability

All data supporting the findings of this study are available within the article. The scRNA-seq data reported in this study have been deposited in the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO; www.ncbi.nlm.nih.gov/geo/) under the accession number GSE317544.