CD4 T cell therapy counteracts inflammaging and senescence by preserving gut barrier integrity

Abstract

Healthy aging relies on a symbiotic host-microbiota relationship. The age-associated decline of the immune system can pose a threat to this delicate equilibrium. In this work, we investigated how the functional deterioration of T cells can affect host-microbiota symbiosis and gut barrier integrity and the implications of this deterioration for inflammaging, senescence, and health decline. Using the Tfam^fl/flCd4^Cre mouse model, we found that T cell failure compromised gut immunity leading to a decrease in T follicular cells and regulatory T cells (T_reg cells) and an accumulation of highly proinflammatory and cytotoxic T cells. These alterations were associated with intestinal barrier disruption and gut dysbiosis. Microbiota depletion or adoptive transfer of total CD4 T cells or a T_reg cell-enriched pool prevented gut barrier dysfunction and mitigated premature inflammaging and senescence, ultimately enhancing the health span in this mouse model. Thus, a competent CD4 T cell compartment is critical to ensure healthier aging by promoting host-microbiota mutualism and gut barrier integrity.

Fig. 1. Intestinal barrier integrity is compromised in Tfam^fl/flCd4^Cre micex.

(A) Longitudinal assessment of body weight in Tfam^fl/fl and Tfam^fl/flCd4^Cre mice, where m denotes the slope of the regression line (n = 3 to 8). (B and C) Concentration of FITC-dextran in the serum (n = 4 to 6) and nonmetric multidimensional scaling (NMDS) plots of β-diversity (θYC indexes) in fecal microbiota (n = 5 to 6) during (B) m₂ and (C) m₃ phases. (D) Quantification of colony-forming units (CFUs) per gram of liver (n = 2 to 3) and levels of LPS-binding protein (LBP) in the serum of 12-month-old mice (n = 3 to 5). (E) Percentage of mice displaying rectal prolapse or diarrhea (n = 3 to 8). (F) Representative hematoxylin and eosin (H&E)–stained sections of the small intestine (scale bar: 120 μm). (G) Quantification of villus height and crypt depth in the small intestine (n = 7). (H) Representative H&E-stained sections of the colon (scale bar: 100 μm). (I) Pathological scoring of colon histology (n = 4 to 7). (J) Chord diagrams of RNA-sequencing analysis representing up-regulated (red) or down-regulated (blue) genes in the colons of Tfam^fl/flCd4^Cre versus Tfam^fl/fl mice. (K) Representative image (scale bar: 20 μm) and quantification of occludin (Ocln) immunofluorescence staining in the colon (n = 4). Data are pooled from (A and E) N = 7 to 8 or (C and D) N = 2 to 4 independent experiments. Data are shown as means ± SEM, where each dot is a biological sample. P values were determined by (A) mixed-effects analysis with Šidák’s multiple comparison test, (D, G, and K) unpaired Student’s t test, (E) Fisher’s exact test, or (I) two-tailed Mann–Whitney U test. (B and C) P values were determined by unpaired Student’s t test (bar graphs) or permutational multivariate analysis of variance (PERMANOVA) (NMDS plots). *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; and ****P ≤ 0.0001.

Fig. 2. Tfam^fl/flCd4^Cre mice display an inflammation-biased gut microbiota with exacerbated production of SCFAs.

(A) Shannon index, and operational taxonomic unit (OTU) richness parameters of α-diversity in the ileum and colon-resident microbiota from 12-month-old Tfam^fl/fl and Tfam^fl/flCd4^Cre mice (n = 4 to 5). (B) Non-metric multidimensional scaling (NMDS) plots showing β-diversity values (θYC indexes) of microbiota in the ileum and colon (n = 4 to 5). (C) Left: differentially abundant OTUs depicted with linear discriminant analysis (LDA) values of linear discriminant effect size (LEfSe, P < 0.05; false discovery rate, q < 0.05; fold change > 5; maximal abundance > 0.001) comparing ileal and colonic microbiota in Tfam^fl/flCd4^Cre versus Tfam^fl/fl mice. Right: heatmap depicting abundance values. (D) Quantification of short-chain fatty acids (SCFAs) in the feces (n = 4 to 5). (A to D) Data are pooled from N = 2 independent experiments. Data are shown as means ± SEM, where each dot is a biological sample. P values were determined by (A) two-tailed Mann–Whitney U test, (B) permutational multivariate analysis of variance (PERMANOVA), or (D) unpaired Student’s t test. *P ≤ 0.05 and **P ≤ 0.01.

Fig. 3. Microbiota depletion reinforces gut barrier integrity preventing inflammaging, senescence, and multimorbidity in Tfam^fl/flCd4^Cre mice.

(A) Experimental design of antibiotic-induced microbiota depletion. (B) Concentration of FITC-dextran in the serum of 12-month-old Tfam^fl/fl and Tfam^fl/flCd4^Cre mice treated with antibiotics (Abx) or vehicle (Veh) (n = 4 to 5). (C) Levels of LPS-binding protein (LBP) in the serum (n = 3 to 4). (D) Relative mRNA levels of genes associated with inflammation (Tnf, Infg, and s100a8) or with tight junctions (Ocln and Tjp1) in the colon (n = 3 to 5). (E) Heatmap depicting normalized concentration of inflammatory mediators in the serum (n = 5 to 7). (F) Relative mRNA levels of genes associated with inflammation (Tnf, Infg, Il6, and Stat1) in the liver (n = 3 to 6). (G) Quantification of β-galactosidase (β-Gal) activity in kidney lysates (n = 2 to 5). (H) Relative mRNA levels of the senescence-associated genes Cdkn1a and Tp53 in the liver (n = 3 to 5). (I) Body weight relative to the beginning of the treatment (n = 3 to 6). (J) Grip test (n = 4 to 5). (K) Clasping score (n = 4). (L) Glucose tolerance test and its area under the curve (AUC) (n = 3 to 4). (M) Kaplan–Meier survival curves (n = 4 to 5). (B to M) Data are pooled from N = 2 independent experiments. Data are shown as means ± SEM, where each dot is a biological sample. P values were determined by (B to H, and J) one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test, (I) two-way ANOVA with Šidák’s multiple comparison test, (K) Kruskal–Wallis H test with Dunn’s multiple comparison test, or (M) log-rank (Mantel–Cox) test. (L) P values were determined by one-way (curve) or two-way ANOVA (AUC) with Tukey’s multiple comparisons test. Tfam^fl/fl versus Tfam^fl/flCd4^Cre (*); Tfam^fl/flCd4^Cre versus Tfam^fl/flCd4^Cre Abx (^#). *^,#P ≤ 0.05; **^,##P ≤ 0.01; ***P ≤ 0.001; and ****P ≤ 0.0001.

Fig. 4. Gut mucosal immunity is impaired in Tfam^fl/flCd4^Cre mice.

(A) Percentage of CD4 T cells in Peyer’s patches (PPs) of 12-month-old Tfam^fl/fl and Tfam^fl/flCd4^Cre mice (n = 3 to 7). (B) Left: representative UMAP showing PP CD4 T cell clusters. Right: bar plots showing the percentage of clusters. (C) Percentage of PP clusters (n = 4 to 7). (D and E) Representative contour plots and quantification of (D) T follicular helper (T_FH) and (E) T follicular regulatory (T_FR) cells in PPs (n = 3 to 5). (F) Quantification of germinal center (GC) B cells in PPs (n = 3 to 5). (G) Absolute number of CD4 T cells in the colonic lamina propria (cLP) (n = 4 to 5). (H) Left: representative UMAP showing cLP CD4 T cell clusters. Right: bar plots showing the percentage of clusters. (I) Percentage of cLP clusters (n = 3 to 5). (J) Absolute number of cLP regulatory T (T_reg) cells (n = 4 to 6). (K) Representative contour plots and quantification of cLP T_reg cells (n = 4 to 6). (L) Quantification of resting (r), activated (a), and KLRG1⁺ (k)T_reg cells in the cLP (n = 3 to 6). (M) Heatmap depicting expression of genes related to IgA biosynthesis in the colon RNA-sequencing analysis. (N) Representative contour plot and quantification of cLP IgA⁺ plasma cells (PCs) (n = 7). (O) Concentration of serum IgA (n = 4 to 7). (P) Representative contour plot and quantification of IgA-coated fecal bacteria (n = 4 to 7). Data are (B, C, H, and I) representative of N = 2 to 3 independent experiments or (A, D to F, J to L, O, and P) pooled from N = 2 to 3 independent experiments. Data are shown as means ± SEM, where each dot is a biological sample. P values were determined by unpaired Student’s t test. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; and ****P ≤ 0.0001.

Fig. 5. Adoptive transfer of CD4 T cells restores gut mucosal immunity in Tfam^fl/flCd4^Cre mice.

(A) Experimental design of CD4 T cell adoptive transfer (AT). (B) Donor cells in the peripheral blood of transferred and non-transferred 10-month-old Tfam^fl/flCd4^Cre mice. (C) Mean chimeric ratio in the peripheral blood, Peyer’s patches (PPs), and the colonic lamina propria (cLP). (D) Percentage of PP CD4 T cells in Tfam^fl/fl and Tfam^fl/flCd4^Cre mice after the transfer of CD4 T cells (n = 5 to 7). (E) Left: representative UMAP showing PP CD4 T cells clusters. Right: bar plots showing the percentage of clusters. (F) Absolute number of cLP CD4 T cells (n = 5 to 7). (G) Left: representative UMAP showing cLP CD4 T cell clusters. Right: bar plots showing the percentage of clusters. (H) Percentage of cLP regulatory T (T_reg) cells (n = 5 to 7). (I) Quantification of T follicular helper (T_FH) and regulatory (T_FR) cells and germinal center (GC) B cells in PPs (n = 5 to 7). (J) Concentration of serum IgA (n = 4 to 7). (K) Longitudinal quantification of IgA-coated fecal bacteria (n = 3 to 7). (L) qPCR quantification of colon-resident microbiota (n = 3 to 7). Data are (B to I) representative of N = 2 or (J to L) pooled from N = 2 independent experiments. Data are shown as means ± SEM, where each dot is a biological sample. P values were determined by (D, F, H, and I) one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test, (J and L) Kruskal–Wallis H test with Dunn’s multiple comparison test, and (K) mixed-effects analysis with Tukey’s multiple comparisons test. Tfam^fl/fl versus Tfam^fl/flCd4^Cre (*); Tfam^fl/flCd4^Cre versus Tfam^fl/flCd4^Cre AT (^#); Tfam^fl/fl versus Tfam^fl/flCd4^Cre AT (^‡). *^,#,‡P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; and ****P ≤ 0.0001.

Fig. 6. CD4 T cell therapy prevents inflammaging, senescence, and multimorbidity by restoring gut barrier integrity in Tfam^fl/flCd4^Cre mice.

(A) Left: principal component analysis (PCA) of colon RNA-sequencing data. Right: hierarchical clustering of differentially expressed genes. (B) Heatmaps depicting expression of genes in the colon RNA-sequencing analysis. (C) Concentration of FITC-dextran in the serum (n = 3 to 5). (D) Quantification of LPS-binding protein (LBP) in the serum (n = 5 to 7). (E) Heatmap depicting normalized concentration of inflammatory mediators in the serum (n = 5 to 7). (F) Quantification of β-galactosidase (β-Gal) activity in kidney lysates (n = 3 to 7). (G) Bar plots showing relative mRNA levels of the senescence-associated genes Cdkn1a and Tp53 in the liver (n = 3 to 5). (H) Body weight relative to the beginning of the treatment (n = 3 to 6). (I) Grip test (n = 3 to 5). (J) Clasping score (n = 3 to 4). (K) Glucose tolerance test and area under the curve (AUC) quantification (n = 3 to 5). Data are (C, E, and I to K) representative of N = 2 or (D, F to H) pooled from N = 2 independent experiments. Data are shown as means ± SEM, where each dot is a biological sample. P values were determined by (C to G, and I) one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test, (H) mixed-effects analysis with Tukey’s multiple comparisons test, (J) Kruskal–Wallis H test with Dunn’s multiple comparison test. (K) P values were determined by one-way (AUC) or two-way ANOVA (curve) with Tukey’s multiple comparisons test. Tfam^fl/fl versus Tfam^fl/flCd4^Cre (*); Tfam^fl/flCd4^Cre versus Tfam^fl/flCd4^Cre AT (^#); Tfam^fl/fl versus Tfam^fl/flCd4^Cre AT (^‡). *^,#P ≤ 0.05; **^,##,‡‡P ≤ 0.01; ***P ≤ 0.001; and ****^,####P ≤ 0.0001.

Fig. 7. Adoptive therapy of a T_reg cell–enriched pool rebalances gut immune homeostasis in Tfam^fl/flCd4^Cre mice.

(A) Experimental design of regulatory T (T_reg) cell adoptive transfer (AT_reg). (B) Donor cells in the peripheral blood of transferred and non-transferred 10-month-old Tfam^fl/flCd4^Cre mice. (C) Mean chimeric ratio and percentage of FoxP3⁺ donor cells in Peyer’s patches (PPs) and the colonic lamina propria (cLP). (D) Percentage of PP CD4 T cells in Tfam^fl/fl and Tfam^fl/flCd4^Cre mice after the transfer of a T_reg cell–enriched pool (n = 4 to 6). (E) Left: representative UMAP showing PP CD4 T cells. On Right: bar plots showing the percentage of clusters. (F) Absolute number of cLP CD4 T cells (n = 3 to 6). (G) Left: representative UMAP showing cLP CD4 T cell clusters. Right: bar plots showing the percentage of clusters. (H) Representative contour plots and quantification of cLP T_reg cells (n = 3 to 6). (I) Longitudinal quantification of IgA-coated fecal bacteria (n = 3 to 6). (J) qPCR quantification of colon-resident microbiota (n = 3 to 5). Data are (B, C, E, and G) representative of N = 2 or (D, F, and H to J) pooled from N = 2 independent experiments. Data are shown as means ± SEM, where each dot is a biological sample. P values were determined by (D, F, H, and J) one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test or (I) mixed-effects analysis with Tukey’s multiple comparisons test. Tfam^fl/fl versus Tfam^fl/flCd4^Cre (*); Tfam^fl/flCd4^Cre versus Tfam^fl/flCd4^Cre AT_reg (^#); Tfam^fl/fl versus Tfam^fl/flCd4^Cre AT_reg (^‡). *^,‡P ≤ 0.05; **^,##P ≤ 0.01; ***P ≤ 0.001; and ****P ≤ 0.0001.

Fig. 8. Transfer of a T_reg cell–enriched pool restores gut barrier integrity preventing inflammaging, senescence, and multimorbidity in Tfam^fl/flCd4^Cre mice.

(A) Left: principal component analysis (PCA) of colon RNA-sequencing data. Right: hierarchical clustering of differentially expressed genes. (B) Heatmap depicting expression of genes in the colon RNA-sequencing analysis. (C) Quantification of LPS-binding protein in the serum (n = 3 to 6). (D) Heatmap depicting normalized concentration of inflammatory mediators in the serum (n = 3 to 5). (E) Quantification of β-galactosidase (β-Gal) activity in kidney lysates (n = 2 to 4). (F) Bar plots showing relative mRNA levels of the senescence-associated genes Cdkn1a and Tp53 in the liver (n = 3 to 6). (G) Body weight relative to the beginning of the treatment (n = 3 to 6). (H) Grip test (n = 3 to 6). (I) Clasping score (n = 3 to 6). (J) Glucose tolerance test and area under the curve (AUC) quantification (n = 3 to 5). Data are (C, and E to I) pooled from N = 2 or (A, B, D, and J) representative of N = 2 independent experiments. Data are shown as means ± SEM, where each dot is a biological sample. P values were determined by (C to F, and H) one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test, (G) mixed-effects analysis with Tukey’s multiple comparisons test, or (I) Kruskal-Wallis H test with Dunn’s multiple comparison test. (J) P values were determined by (AUC) one-way ANOVA or (curve) two-way ANOVA with Tukey’s multiple comparisons test. Tfam^fl/fl versus Tfam^fl/flCd4^Cre (*); and Tfam^fl/flCd4^Cre versus Tfam^fl/flCd4^Cre AT_reg (^#). *P ≤ 0.05; **^,##P ≤ 0.01; ***P ≤ 0.001; and ****P ≤ 0.0001.