Host immunosenescence compromises Mycobacterium tuberculosis clearance

Lung Mtb burden in C57BL/6 mice remains unaffected by age or sex through 6 weeks post-infection

The influence of host age and sex on susceptibility to Mtb infection remains an area of interest, particularly at early time points. Thus, female and male C57BL/6 mice of wider age groups were aerosol infected with a low dose of H37Rv (Figure 1A). Mice were categorized as young (2-4M), middle-aged (9M females, 12M males) and old (17-19M males). Prior to Mtb infection, body weight was higher in middle-aged and older mice than in younger controls (Supplementary Figure 1). During the study period, old mice lost ~3% weight, whereas young mice gained ~8% weight (Figure 1B). Young female mice gained more body weight (~20%) than males (~10%) (Figure 1B). At 6 weeks post infection (w.p.i.), middle-aged male mice showed significant body weight loss (Figure 1B). Despite age-related variations in body weight, lung Mtb burden progressively increased up to 4 w.p.i. before plateauing by 6 w.p.i., with no significant differences attributable to age or sex (Figure 1C). Thus, immunosenescence manifests primarily during therapeutic clearance rather than initial containment phase, rationalizing our focus on age-associated differences in treatment response.

Figure 1. Lung Mtb burden remains similar till 6 weeks post-infection across age groups and sexes in C57BL/6 mice. (A) Schematic of the experimental design used for aerosol Mtb H37Rv infection of C57BL/6 mice [female (♀) and male (♂)] across age groups. (B) Changes in body weight of mice groups (healthy: H, ^….; infected: I, ^___) are presented, n=5-10/age group/condition; dashed vertical line represents day of infection. (C) Lung bacterial burden (in log₁₀cfu/mL) in mice post Mtb infection; n=5-10/timepoint/age group/condition; dashed horizontal line represents limit of detection (LOD). Young (2-4 months) in blue, middle-aged (9-12 months) in green and old (17-19 months) mice in brown; cfu= colony forming unit; w.p.i.= weeks post infection; p-value: ** <0.005 at 95% confidence interval by Mann-Whitney test. Data shown as mean ± SD. See also Supplementary Figures 1, 2.

Inflammaging and immune defects drive delayed TB clearance in aging despite hepatic resilience

To evaluate the impact of age on treatment response, a subset of mice received RIF-INH following Mtb infection. Remarkably, older mice retained a significantly higher ~3 log₁₀ cfu lung Mtb load at 2 weeks post-treatment (w.p.t.) compared to the young group (~1.5 log₁₀ cfu). And by 4 weeks, both mouse age groups achieved complete bacterial clearance, i.e. resulting below the limit of detection (LOD) (Figure 2A, 2B). Histologically, old mice developed granulomatous lung lesions with prominent foamy cells later (6 w.p.i.) compared to young mice (4 w.p.i.) (Supplementary Figure 1). Lung inflammation was resolved in young mice by 4 w.p.t. but persisted in old mice (Supplementary Figure 1). Additionally, middle-aged mice were able to resolve the inflammation by 4 w.p.t., irrespective of sex (Supplementary Figure 2). In extrapulmonary sites, such as spleen and liver, mycobacterial clearance was observed by 2 w.p.t. in both age groups (Figure 2B).

Figure 2. Old C57BL/6 mice exhibit delayed lung Mtb clearance following two weeks of RIF-INH treatment. (A) Lung bacterial burden (in log₁₀cfu/mL) in male C57BL/6 mice from infected (I; at 2, 4 and 6 w.p.i.) and treated (T; at 2 and 4 w.p.t.; treatment started at 2 w.p.i.) groups; n=10/timepoint/age group/condition; dashed horizontal line represents limit of detection (LOD). A set of two independent experiments is depicted. (B–F) showing data of infected groups at 4 w.p.i. and treated groups at 2 w.p.t.. (B) Bacterial load (in log₁₀cfu/mL) in lung, spleen and liver; individual data points at 4 w.p.i./2 w.p.t. presented in histograms (n=10/age group/condition). (C) Serum alanine transaminase (ALT) and aspartate aminotransferase (AST) activities in mice of different groups (n=5/age group/condition) at 4 w.p.i./2 w.p.t.. (D) Area under the curve (AUC) for Mtb H37Rv lysate-specific serum IgG in healthy (H; n=15/age group), infected (I; n=27-28/age group) and treated (T; n= 16-17/age group) mice at 4 w.p.i./2 w.p.t.. (E) Serum cytokines (in picogram per milliliter): IL-6, IFN-γ in different mice groups (n=6-7/age group/condition) at 4 w.p.i./2 w.p.t.. (F) Lung IFN-γ in the different mice groups (n=4-5/age group/condition) at 4 w.p.i./2 w.p.t.. Young (2-4 months) in blue and old (17-19 months) mice in brown; cfu= colony forming unit; w.p.i.= weeks post infection; w.p.t.= weeks post treatment; p-values: * ≤0.05, ** <0.005, *** <0.0005 and **** <0.0001 at 95% confidence interval by Mann-Whitney test. Data shown as mean ± SD. See also Supplementary Figures 1, 4.

Liver enzymes are pivotal for drug metabolism, with metal ions exerting substantial influence over their activity. Age-related changes may influence this variability in liver function and micronutrient status [11]. Thus, we monitored the liver micronutrient levels of these mouse groups using inductively coupled plasma mass spectrometry (ICP-MS) (Supplementary Figure 3). Old mice had higher hepatic Cu (4.28×), low Zn (0.14×) levels with similar Fe and Se levels compared to controls (Supplementary Figure 3). Additionally, old mice receiving treatment showed higher hepatic Cu levels (Supplementary Figure 3). Liver micronutrient status impacts activities of hepatic enzymes like alanine aminotransferase (ALT) and aspartate aminotransferase (AST), commonly used markers of liver function and injury. Old mice exhibited comparable serum ALT and AST activities to the young mice (Figure 2C). However, ALT activity showed a non-significant upward (p=0.06) trend in old mice (Figure 2C). Mtb infection did not impact serum AST or ALT activity in old mice, but ALT activity was higher in Mtb-infected younger mice (Figure 2C). Strikingly, young mice receiving treatment had higher serum AST activity reflecting drug-induced liver injury, while old mice remained unaffected (Figure 2C).

To assess age-related differences in humoral immunity against Mtb, if any, serum anti-H37Rv specific IgG titers were measured. The overall IgG response was significantly higher in old mice compared to young mice (Figure 2D). Importantly, old mice receiving treatment mounted higher IgG responses at a later time point i.e. 4 w.p.t., while young mice showed it at 2 w.p.t. compared to their respective infected controls (Figure 2D and Supplementary Figure 4). While robust humoral responses are beneficial for pathogen control, the delayed yet heightened antibody response in old mice could indicate persistent antigenic stimulation and likely reflect altered immune regulation with aging. Circulatory IL-6 levels were higher in old mice, consistent with a state of chronic low-grade inflammation or inflammaging (Figure 2E). Higher circulatory IFN-γ levels were observed in Mtb-infected young and old mice, which resolved partially in treated groups (Figure 2E). The cytokine response in old mice appears skewed toward a T_H1-type response rather than a broad systemic inflammatory state, which may explain selective upregulation of IFN-γ without a parallel increase in IL-6. However, at the site of infection, IFN-γ levels trended higher in young mice compared to old mice (Figure 2F). Critically, despite RIF-INH treatment selectively reducing lung IFN-γ levels in old mice, yet bacterial clearance remained impaired (Figure 2F). This reveals a central paradox showing T_H1 contraction coupled with effector dysfunction, where diminished IFN-γ, rather than unresolved inflammation, impairs bacterial killing during early therapy.

Age-related T cell dysregulation impairs immune Mtb control

CD4⁺ T cells control Mtb primarily by producing cytokines and licensing infected phagocytes, but aging drives progressive dysfunction marked by exhaustion-like phenotypes. T cells were profiled using activation markers (CD44, PD1), homing receptor (CXCR5), senescence (KLRG1, CD153) and lineage-defining transcription factors (Tbet for T_H1 and FoxP3 for T_reg) and a significant change in their distribution was observed in old mice (Supplementary Figure 5). CD44hi marks antigen-experienced CD4⁺ T cells, while PD-1 captures chronically stimulated cells that can be either functional (protective, CXCR5⁺ T_FH-like) or dysfunctional depending on expression of other markers. CXCR5⁺CD4⁺ T cells support B cell help and are associated with better Mtb control when present in organized structures [12, 13]. KLRG1 is enriched on highly differentiated, senescent-like CD4⁺ T cells that accumulate with age. KLRG1⁺PD1⁺ T represents late-differentiated, senescent-like cells with constrained effector function. Notably, high expression of CD44, CXCR5, PD1 and CD153 in CD4⁺ and CD8⁺ T cells of old mice were observed (Supplementary Figure 5, 6). Increased circulatory CD4⁺KLRG1⁺PD1⁺ cells were observed in old mice, indicative of a higher frequency of exhausted T cell phenotype (Supplementary Figure 5).

Next, we monitored the immune cell distribution in lungs and spleen of Mtb-infected and RIF-INH treated mice groups (Figure 3A, 3B and Supplementary Figure 7). Old mice showed a lower proportion of splenic CD4⁺, higher CD8⁺ T cells, along with higher CXCR5 and PD1 expression compared to young mice (Figure 3C and Supplementary Figure 7). Interestingly, lungs of old mice showed accumulation of CD4⁺CD44⁺CXCR5⁺ (T_FH)-like cells, similar to young mice post Mtb infection (Supplementary Figure 8). Additionally, CD8⁺CD44⁺CXCR5⁺ (T_FC)-like cells were observed to be more abundant in the lungs and spleens of Mtb-infected old mice at 4 w.p.i. (Figure 3C and Supplementary Figures 7, 8). T_FC-like cells are a follicular, B-cell interacting CD8 subset, which compete with T_FH cells for germinal centre reactions, disrupting antibody maturation.

Figure 3. Deregulated splenic CD4⁺CD44⁺ T cell proteome in old C57BL/6 mice upon Mtb infection and RIF-INH treatment. (A) Schematic showing Mtb H37Rv aerosol infection in male C57BL/6 mice of two age groups (young: 2-4 and old: 17-19 in months) followed by treatment with RIF-INH starting at 2 w.p.i.. (B) Workflow for immune cell isolation from the spleen, flow cytometry data acquisition and data analysis. (C) Frequencies of CD4⁺CD44⁺CXCR5⁺, CD4⁺CD44⁺CXCR5⁺PD1⁺ and CD4⁺CD44⁺CXCR5⁺PD1⁺ICOS⁺ and CD8⁺CD44⁺CXCR5⁺ cells in the spleen of C57BL/6 mice at 4 w.p.i./2 w.p.t.. Healthy (H; n=3/age group), infected (I; n=6-10/age group) and treated (T; n=3/age group). (D) TMT10plex workflow for CD4⁺CD44⁺ T cell proteome analysis (for TMT set-1) by liquid chromatography- mass spectrometry (LC-MS/MS). (E–H) showing volcano plots of splenic CD4⁺CD44⁺ T cell proteome showing significantly deregulated proteins (in -log₁₀p-value ≥ 1; log₂fold change > |0.58|): (E) Young infected versus healthy, (F) Old infected versus healthy, (G) Young treated versus infected and (H) Old treated versus infected; ▲/▲: up-/down-regulation; HOGA: 4-hydroxy-2-oxoglutarate aldolase; AST: aspartate aminotransferase. Young (2-4 months) mice in blue and old (17-19 months) mice in brown; cfu= colony forming unit; w.p.i.= weeks post infection; p-values: * ≤0.05, ** <0.005, *** <0.0005 and **** <0.0001 at 95% confidence interval by Mann-Whitney test. Data shown as mean ± SD. See also Supplementary Figures 7–10 and Supplementary Tables 1–4.

In a separate experimental set, at 4 w.p.i., lung Mtb load was found to be similar between young (2M) and aged (31M) male C57BL/6 mice groups (Supplementary Figure 9). Old and aged mice had overall higher splenic CD4⁺CD44⁺CXCR5⁺PD1⁺, CD4⁺CD44⁺CXCR5⁺PD1⁺ICOS⁺ and CD8⁺CD44⁺CXCR5⁺PD1⁺ cells compared to young mice, irrespective of infection and treatment status (Figure 3C and Supplementary Figure 9). The detailed distribution of immune cells observed in multiple tissues of 2-, 4-, 17-, 19-, and 31M old mice groups is tabulated in Supplementary Table 1. Immunosenescence creates pathological T_FH-T_FC competition, diverting adaptive immunity toward dysregulated CD8⁺ follicular responses that compromise TB therapeutic efficacy.

Splenic T cell proteome reveals age-specific mitochondrial collapse

To further elucidate the molecular drivers of delayed Mtb clearance, we conducted global proteomic profiling of flow-sorted CD4⁺CD44⁺ T cells (Figure 3D and Supplementary Table 2). In set-I, 109 proteins were identified, and Mtb-infected young and old mice showed 23 and 30 deregulated (p ≤ 0.1, log₂fold change (log₂fc): I/H > |0.58|) proteins, respectively (Figure 3E, 3F and Supplementary Figure 10 and Supplementary Table 3). In set-II, out of 41 identified proteins, ~20 were deregulated in the mouse groups receiving RIF-INH treatment (Figure 3G, 3H and Supplementary Table 4). Proteins involved in immune response (e.g., complement, serotransferrin) were suppressed in old mice, potentially reflecting age-related immune decline. Young mice exhibited few changes, with upregulation of stress response proteins like galectin-1 and superoxide dismutase, possibly reflecting a more robust acute response (Supplementary Table 3). Both mouse age groups showed downregulation of 4-hydroxy-2-oxoglutarate aldolase (HOGA), suggesting disrupted hydroxyproline degradation [14] (Supplementary Table 3). Upon Mtb infection, old mice showed an upregulation of prostaglandin E synthase 3 (PTGES3, log₂fc: I/H = 1.15), indicating an age-biased induction of prostaglandin synthase E2 (PGE2) machinery (Supplementary Table 3). PGE2 disrupts alveolar macrophage metabolic fitness through mitochondrial dysfunction, suppressing electron transport chain gene expression [15].

Additionally, old mice showed elevated levels of mitochondrial AST after treatment (log₂fc: T/I = 3.15), indicative of unresolved mitochondrial stress, consistent with increased serum AST and altered hepatic micronutrient status observed earlier (Supplementary Table 4). Old mice showed lower HOGA abundance after treatment, implying prolonged mitochondrial stress in T cells of old mice (Supplementary Table 4). Together, these proteomic signatures reveal age-specific mitochondrial defects, providing a molecular basis for T cell dysfunction, inflammaging, and impaired mycobacterial clearance in older hosts, linking cellular metabolic stress in T cells during Mtb infection and therapy.

Experimental groups, Mtb infection and treatment

The procedures adopted in this study were performed following the recommended Guidelines for Care and Use of Laboratory Animals and approved by the Animal Ethics Committee of ICGEB- New Delhi (Ref No. ICGEB/IAEC/07032020/TH-10). C57BL/6 mice of different age groups (2-, 4-, 9-, 12-, 17-, 19- and 31 months, M) were procured from the ICGEB bio-experimentation facility. The mice age equivalents to humans, used in this study, is as follows: young (2-4M), mid-aged (9-12M; 38-44 human years), old (17-19M; ~54-60 human years), and aged (31M; >80 human years) [7]. These mice were transferred to the BSL-III Tuberculosis Aerosol Challenge Facility (TACF) for experiments involving Mtb infection. Mice were fed a chow diet, water ad libitum and maintained at 20-22° C; 45-60% humidity with a 12h light/dark cycle. Mice were weight-adjusted, randomly grouped and after 1 week of acclimatization in TACF, infected with a low dose (100-120 cfu) of Mtb H37Rv strain in the Madison aerosol exposure chamber. After 2 w.p.i., a subset of mice received RIF and INH (100 mg/L each) in drinking water and fresh drug solutions were replaced every 48 hours. Two 17M mice (one bearing an abscess, the other died before the completion of 4 w.p.t.) were excluded from the study.

Cfu assay

Tissues (lung, spleen and liver) were homogenized in PBS (1 mL) and plated on 7H11 agar supplemented with Middlebrook 10% OADC. These plates were incubated at 37° C, and bacterial colonies were enumerated 3 weeks later.

Histopathology

Formalin-fixed lungs were paraffin-embedded, sectioned and stained with hematoxylin and eosin, followed by capturing images using Nikon Eclipse TS2. The slides were scored based on the severity of inflammation, granuloma formation and necrosis.

Liver micronutrient profiling

Liver (~50 mg) was subjected to microwave digestion post addition of H₂O₂ and HNO₃ (1:4). These were diluted and introduced to iCAP™-TQ ICP-MS using argon as carrier. Data acquisition was performed using Qtegra (Thermo Fisher Scientific, version 2.10.4345.39) and involved running different concentrations of a multielement standard mix (Supelco, #92091).

ALT and AST assay

Serum ALT and AST activities were measured using kits (Cayman Chemical, #700260 and #701640) according to the manufacturer’s instructions.

Immune cell isolation and flow cytometry

Harvested tissues (lung and spleen) were incubated with collagenase D (1 μg/mL) and DNase I (0.5 mg/mL). Peripheral blood mononuclear cells (PBMCs) were isolated using Histopaque-1077 following standard procedures. Isolated cells were stained with recommended dilutions by the manufacturer: CD3-BV510 (BioLegend, #100234), CD4-BV785 (BioLegend, #100552), CD8a-AF700 (BioLegend, #100730), CD44-BV605 (BioLegend, #103047), CXCR5-PE/Dazzle 594 BioLegend, #145522), PD1-PE-Cy7 (BioLegend, #109110), KLRG1-BV421 (BioLegend, #138414), CD153-PE (BioLegend, #106405), CD25-AF488 (BioLegend, #102017) and CD127-BV650 (BioLegend, #135043) and LIVE/DEAD™ Fixable Near IR stain (Invitrogen, #L34994) on ice for 45 min. After washing with buffer (0.2% FBS in 1× PBS), cells were treated with fixation (Invitrogen, #00-5123) and permeabilization buffer (Invitrogen, #00-8333) and kept on ice for 30 min. Intracellular staining was done with FoxP3-AF647 (BioLegend, #126408) and Tbet-BV711 (BioLegend, #644820) and incubated at RT for 45 min, followed by washing. Data (~1 million events/sample) was acquired using BD LSRFortessa X-20 and analyzed using FCS Express (DeNovo software, version 6).

Another set of immune cells were stained with: CD3-FITC (Invitrogen, #11-0032-82), CD4-SB600 (Invitrogen, #63-0042-82), CD8a-BV421 (BioLegend, #100753), CD44-BV510 (BioLegend, #103044), CXCR5-PE, ICOS-PerCP/Cy5.5 (BioLegend, #145504), PD1-PE-Cy7 (BioLegend, #109110) and LIVE/DEAD™ Fixable Near IR stain (Invitrogen, #L34994) on ice for 45 min. After washing with buffer (3% FBS in 1× PBS), splenic CD4⁺CD44⁺ T cells were sorted using BD FACS Aria III Fusion. Data was analyzed using FlowJo (BD Biosciences, version 10.8.1).

Lung IFN-γ estimation: IFN-γ was estimated in lung homogenates by using the ELISA MAX™ Standard Set Mouse IFN-γ kit (BioLegend, #430801) following the manufacturer’s instructions.

Serum cytokine estimation

Two times diluted serum samples were used to estimate cytokines using capture bead-based assay LEGENDplex (BioLegend) along with the standard (BioLegend, #740371, lot no. B330381). The samples were acquired on BD LSRFortressa X-20 using PE and APC as the reporter and bead classification channel, respectively. Data was analyzed using Qognit software.

Serum IgG antibody quantification

ELISA plate was coated with Mtb H37Rv whole cell lysate (10 μg/mL) and incubated overnight at 4° C. Then, wash buffer (0.05% Tween-20 in 1× PBS) was added, followed by blocking in 3% BSA in 1× PBS containing 0.05% Tween-20 and incubated at 37° C for 2h at 180 rpm. After blotting the plate dry, sera were serially diluted fourfold in 0.1% BSA in 1× PBS containing 0.05% Tween-20, and 100 μL was added per well. The plate was then incubated overnight at 4° C. The plate was washed three times, followed by incubation with anti-mouse IgG (Southern Biotech, #1031-05) at 37° C for 2h at 180 rpm. The plate was washed six times and developed by adding o-phenylenediamine dihydrochloride, incubated at RT for 5 min. The reaction was terminated by adding 1N HCl, and the absorbance was measured at 490 nm.

Splenic CD4⁺CD44⁺ cell proteome analysis

Proteins were isolated using TRIzol reagent from sorted splenic CD4⁺CD44⁺ cells and quantified using BCA Protein Assay. An equal amount (~10 μg) of protein from each sample in a set (n = 10) was used for TMT10plex (Thermo Fisher Scientific, #90113). TMT-labelled peptides from all samples were pooled, dried and subjected to strong cation exchange chromatography and eluted using ammonium formate. Later, these fractions (n = 5) were cleaned up using C18 spin columns and dried, resuspended in formic acid (0.1%, 20 μL) and injected into an Orbitrap Fusion Lumos Tribrid Mass Spectrometer connected to a nano-LC 1200. Data was analysed with Proteome Discoverer version 2.3 using Mus musculus (ID: UP000000589; 55,087 protein count). Proteins qualifying the parameters [log₂ fold change ≥ ± 0.58; log₁₀p-value ≥1; false discovery rate<0.05] were selected as important deregulated molecules.

Statistical analysis

Data are presented as mean ± standard deviation. Details on the sample size and statistical analysis are provided in the figure legends.