Age-Associated Microbial Dysbiosis Promotes Intestinal Permeability, Systemic Inflammation, and Macrophage Dysfunction

Abstract

Levels of inflammatory mediators in circulation are known to increase with age, but the underlying cause of this age-associated inflammation is debated. We find that, when maintained under germ-free conditions, mice do not display an age-related increase in circulating pro-inflammatory cytokine levels. A higher proportion of germ-free mice live to 600 days than their conventional counterparts, and macrophages derived from aged germ-free mice maintain anti-microbial activity. Co-housing germ-free mice with old, but not young, conventionally raised mice increases pro-inflammatory cytokines in the blood. In tumor necrosis factor (TNF)-deficient mice, which are protected from age-associated inflammation, age-related microbiota changes are not observed. Furthermore, age-associated microbiota changes can be reversed by reducing TNF using anti-TNF therapy. These data suggest that aging-associated microbiota promote inflammation and that reversing these age-related microbiota changes represents a potential strategy for reducing age-associated inflammation and the accompanying morbidity.

Keywords: Streptococcus pneumoniae; elderly; host defense; immunosenescence; inflamm-aging; inflammation; macrophage; microbiome; microbiota.

Graphical abstract

Figure 1

Inflammatory Responses Increase with Age (A) Killing of S. pneumoniae by resident peritoneal macrophages isolated from young and old mice (n = 5). (B) Killing of S. pneumoniae by bone marrow-derived macrophages from young and old C57BL/6 mice (n = 6). (C) Intact tetramethylrhodamine (TRITC)-labeled S. pneumoniae was observed in macrophages derived from old mice, but not young mice, up to 4 hr post-infection. (D–F) Levels of TNF (D) and IL6 (E) were higher in the plasma of old mice as was IL6 in slices of whole-lung homogenates from old mice (F). (G) H&E stain of formalin-fixed histological sections from the lungs of young and old WT mice at 5× magnification. One representative image is at least five biological replicates. The degree of cellular infiltration within each image was measured by expressing the total area of the cellular infiltrate within the lung as a percentage of the total lung area. (H) Lung slices were processed from the lungs of young and old mice and cultured in media. IL6 production was subsequently measured in the supernatant at 4 hr following stimulation with heat-killed S. pneumoniae or PBS control (n = 3, representative of two independent experiments). (I) IL6 production in the whole blood of young and old WT mice following stimulation with LPS or a vehicle control (PBS) (n = 5–9). (J) IL6 production from macrophages derived from young and old mice following 24-hr stimulation with a vehicle control (PBS), LPS, or S. pneumoniae as measured by ELISA (n = 6). Results represent mean ± SEM. Statistical significance was determined using the Mann-Whitney test or two-way ANOVA with Fisher’s post-test where appropriate (^∗p < 0.05, ^∗∗p < 0.005, and ^∗∗∗p < 0.0005).

Figure 2

Chronic Exposure to TNF Contributes to Increased Inflammatory Responses and Tissue Damage that Occur with Age (A) Young and old murine bone marrow macrophage-mediated killing of S. pneumoniae is decreased in the presence of 10 ng/mL exogenous TNF (n = 5). (B) Unlike old WT mice, old TNF KO mice do not have increased levels of plasma IL6 (n = 3–10 mice per group, one of two independent experiments shown). (C) IL6 production in the whole blood of young and old TNF KO mice following stimulation with LPS or a vehicle control (PBS) demonstrates that old TNF KO mice do not have higher inflammatory responses to LPS compared to young mice (n = 5). (D) IL6 levels as detected by ELISA in whole-lung tissue homogenates were no higher in old TNF KO mice than in young TNF mice (n = 3). (E) H&E stain of formalin-fixed histological sections of lungs of young and old TNF KO mice (20× magnification, one representative of at least four). The degree of cellular infiltration within each image was measured by expressing the total area of the cellular infiltrate within the lung as a percentage of the total lung area. (F) Bone marrow-derived macrophages from young and old TNF KO mice do not differ in their ability to kill S. pneumoniae (n = 5). Results represent pooled data and are shown as mean ± SEM. Statistical significance was determined using the Mann-Whitney test or two-way ANOVA with Fisher’s post-test where appropriate (^∗p < 0.05, ^∗∗p < 0.005, and ^∗∗∗p < 0.0005).

Figure 3

The Microbiota Increase Intestinal Permeability, Age-Associated Inflammation, Macrophage Function, and Longevity (A) Intestinal permeability of aging mice (3, 12, 15, and 18 months old) was measured by FITC-dextran translocation to the circulation following oral gavage (n = 4–8), and it was found to increase significantly with age (p < 0.007, one-way ANOVA). (B) Colons of young and old mice do not have detectable changes in either epithelial architecture or inflammatory infiltrate when measured as described in the STAR Methods. (C) Mucosal-to-serosal flux of ⁵¹Cr-EDTA as measured in Ussing chambers was used to measure the paracellular permeability of ileums and colons from young and old WT mice (n = 6–12). (D) Circulating muramyl dipeptide (MDP) in the plasma of young and old WT mice as measured by nucleotide-binding oligomerization domain containing protein (NOD)-nuclear factor κB (NF-κB) promoter bioassay. Significant changes shown are relative to young WT mice. (E) Mucosal-to-serosal flux of ⁵¹Cr-EDTA as measured in Ussing chambers was used to measure the paracellular permeability of the colons from young and old GF mice (n = 5–8). There was no significant increase in permeability in old GF mice. (F) Survival analysis showing all-cause mortality of WT and GF mice up to 600 days of life. Differences in the survival curves were analyzed by log rank (Mantel-Cox) test. (G) Plasma cytokines are not higher than young GF mice and are lower than WT SPF mice (n = 3–5). (H) Histological analysis of lung sections stained with H&E from young and old GF mice does not indicate any increased leukocyte infiltration with age (20× magnification; one representative image of at least five mice). (I) IL6 levels in the lung homogenates of old GF mice are not higher than in young WT or young GF mice. (J) IL6 production in the whole blood of young and old WT SPF and GF mice following stimulation with LPS or a vehicle control (PBS). Old GF mice do not have higher levels than young WT SPF or young GF mice (n = 5–9). Significant changes shown are relative to young WT mice. (K) Macrophages from young and old GF mice do not differ in their ability to kill S. pneumoniae (n = 3). Results represent the mean ± SEM of three biological replicates. (L) Bone marrow-derived macrophages from old GF mice do no not have decreased killing or produce more IL6 following stimulation with LPS or a vehicle control (PBS) in macrophages from young GF mice (n = 5). Statistical significance was determined using the Mann-Whitney test or two-way ANOVA with Fisher’s post-test where appropriate (^∗p < 0.05, ^∗∗p < 0.005, and ^∗∗∗p < 0.0005).

Figure 4

Age-Associated Inflammation Is Dependent on the Composition of the Intestinal Microbial Community (A and B) Mice with minimal ASF-derived microbiota were aged, and their intestinal permeability was measured via FITC-dextran oral gavage assay (A; n = 5). Old (18 months) ASF mice had higher intestinal permeability than young (10–14 weeks) ASF mice in addition to higher levels of plasma IL6 (B). (C) IL6 production after LPS stimulation in whole blood was higher in old ASF mice (n = 3). (D) The taxa summary of microbiota of young and old ASF mice indicates that age-related microbial dysbiosis occurs. (E) Taxa summaries illustrate that the composition of the young and old microbiota is retained upon transfer to young or old GF mice (n = 5–16 mice/group over four independent colonization experiments). (F) Paracellular permeability was measured in GF mice colonized with young and old microbiota (n = 6–8 mice per group). There was a statistically significant increase in paracellular permeability in mice colonized with old microbiota (n = 23 total, n = 11 young mice and n = 12 old mice) compared to young microbiota (n = 13, n = 6 young mice and n = 7 old mice). This demonstrates that the age of the microbiota alters barrier function. (G) Colonization of young GF mice with old microbiota increased paracellular permeability compared to those colonized with young microbiota; however, old mice colonized with either young or old microbiota demonstrated increased permeability, indicating that age-related changes in the host increased susceptibility to the microbiota. (H) Circulating TNF was measured from all the young GF mice (n = 13 total, n = 5 colonized with young microbiota and n = 8 colonized with old microbiota) and old GF mice (n = 11 total, n = 5 colonized with the young microbiota and n = 6 colonized with the old microbiota). Old recipient mice had higher levels of circulating TNF than young recipient mice. (I) The microbiota contributed to the increased TNF in the circulation of young mice, since young GF mice colonized with old microbiota had higher circulating levels of TNF than those colonized with young microbiota. In contrast, colonization with either the young or old microbiota increased circulating TNF in old GF mice, indicating that the age of the host interacts with the age of the microbiota to induce systemic inflammation. Bars represent the mean ± SEM. Statistical significance was determined using the Mann-Whitney test or two-way ANOVA with Fisher’s post-test or unpaired t test where appropriate (^∗p < 0.05, ^∗∗p < 0.005, and ^∗∗∗p < 0.0005).

Figure 5

Microbial Dysbiosis Occurs with Age and Inflammation (A) Intestinal permeability, as measured by plasma FITC-dextran following oral gavage, was increased in old WT/SPF mice, but not old TNF KO or old GF mice. (B) Circulating MDP is increased in old SPF/WT mice. Old TNF KO and GF mice are not significantly different from young WT/SPF mice (n = 5–10). GF mice do not have any detectable MDP in the circulation. (C) Principal coordinate analysis based on Bray-Curtis demonstrates that the microbial communities of old WT mice diverge from young mice, but this is not the case in old TNF KO mice. Mice were sampled from multiple cages. Chi-square of the likelihood ratio test in DESeq2 shows old and young microbiota are significantly different (p < 0.001). (D) Anti-TNF (Adalimumab) or a human IgG isotype control was administered at a dose of 50 ng/g of body weight every other day for 2 weeks. Principal co-ordinate analysis was used to visualize differences in the microbial communities after 2 weeks of anti-TNF treatment. Anti-TNF treatment altered the composition of the fecal microbiota of old, but not young, mice. (E) Basal translocation of microbial products occurs throughout life; however, with age, these induce an inflammatory response, which contributes to microbial dysbiosis. Microbial dysbiosis increases intestinal permeability, which increases bacterial translocation. This feed-forward process increases with age.