Time-restricted eating reveals a "younger" immune system and reshapes the intestinal microbiome in human

Abstract

Time-restricted eating (TRE) has been shown to extent lifespans in drosophila and mouse models by affecting metabolic and anti-inflammatory activities. However, the effect of TRE on the human immune system, especially on immunosenescence, intestinal microbiome, and metabolism remains unclear. We conducted a 30-day 16:8 TRE single-arm clinical trial with 49 participants. Participants consumed daily meals from 9 a.m. to 5 p.m., provided by a nutrition canteen with a balanced, calorie-appropriate nutrition, which is designed by clinical nutritionists (ChiCTR2200058137). We monitored weight changes and weight-related parameters and focused on changes in the frequency of CD4⁺ senescent T cells, immune repertoire from peripheral blood, as well as serum metabolites and gut microbiota. We found that up to 95.9 % of subjects experienced sustained weight loss after TRE. The frequency of circulating senescent CD4⁺ T cells was decreased, while the frequency of Th1, Treg, Tfh-like, and B cells was increased. Regarding the immune repertoire, the proportions of T cell receptor alpha and beta chains were increased, whereas B cell receptor kappa and lambda chains were reduced. In addition, a reduced class switch recombination from immunoglobulin M (IgM) to immunoglobulin A (IgA) was observed. TRE upregulated the levels of anti-inflammatory and anti-aging serum metabolites named sphingosine-1-phosphate and prostaglandin-1. Additionally, several anti-inflammatory bacteria and probiotics were increased, such as Akkermansia and Rikenellaceae, and the composition of the gut microbiota tended to be "younger". Overall, TRE showed multiple anti-aging effects, which may help humans maintain a healthy lifestyle to stay "young". Clinical Trial Registration URL: https://www.chictr.org.cn/showproj.html?proj=159876.

Keywords: Immune cellular senescence; Immune repertoires; Intestinal microbiome; Metabolomic; Time-restricted eating; Young.

Graphical abstract

Fig. 1

Experimental design and changes in body weight and related indexes before and after TRE. (a) Eating window, sampling timeline and specimen collection during TRE; (b) Proportion of participants who experienced weight loss after 30 days of TRE; (c) Statistics of weight loss based on different percentage ranges; (d) Continued weight loss observed in participants during 30-day TRE. (e) Statistical analysis of the change of body fat rate, body mass index (BMI), muscle mass, body moisture, visceral fat, basal metabolism, fat weight and skeletal muscle rate before and after a 30-day TRE. n = 45. Statistical analysis was performed using Mauchly's Test of Sphericity in Fig. 1d and paired t-test at two different time points in Fig. 1e and f. Data are mean ± s.e.m in Fig. 1d.

Fig. 2

Frequencies of CD4⁺ T subsets from PBMCs before and after TRE. Statistical analysis and cytometry gates of (a) CD4⁺Th₁, Th2 and Th17 cells; (b) CD4⁺Tfh-like and T_reg cells for each T-cell subset from PBMCs of the TRE Day 0 (n = 41), Day 14 (n = 41), Day 30 (n = 40) and Day 90 after TRE groups (n = 19). (c) Statistical analysis and cytometry gates of CD4⁺ senescent cells (CD27⁻CD28⁻) in the TRE Day 0 (n = 22), Day 14 (n = 38), Day 30 (n = 37) and Day 90 after TRE groups (n = 21). Statistical analysis was performed using paired t-test at two different time points in Fig. 2a, b and c. Data are mean ± s.e.m.

Fig. 3

Immune repertoire changes before and after 30 days of TRE. (a) Proportion of 7 chains in volunteer PBMCs before and after TRE by counting unique CDR3; (b) Statistical analysis of the percentage changes in the proportions of BCR and TCR chains across time points; (c) The representative treemaps showing the 7-chain repertoire in PBMCs before and after the 30-day TRE. (d) The proportion of the 5 Ig isotypes in >30 y TRE, <30 y TRE and Ctrl groups by counting reads; (e) The CSR index values of five Ig before and after the 30-day TRE for the >30 y TRE, <30 y TRE groups and in total. The arrow indicates the direction of class switching, and the thickness of the arrow connecting lines indicates the displacement of the CSR index. The spheres represent different Ig isotypes. The graph displays IgM and IgD as a whole. > 30 y TRE, n = 5, <30 y TRE, n = 5, Ctrl, n = 5. Statistical analysis was performed using a paired t-test shown in Fig. 3b. Data are mean ± s.e.m in Fig. 3b.

Fig. 4

Metabolic differences before and after 30-day TRE. (a) Score scatter plot of the OPLS-DA model for metabolites on Day 0 and Day 30 of TRE; (b) Volcano plot of differential metabolites on Day 0 and Day 30 (VIP >1 and P < 0.05 in paired t-test); (c) Heatmap showing the metabolite abundances in participants before and after TRE. The 110 differential metabolites (d) and 71 subclasses of lipid and lipid-like molecules (e) after TRE. (f) KEGG metabolic pathway analysis in participants before and after TRE; (g) Analysis of pathway-based differential metabolites after TRE; (h) and (i): correlation analysis between CD4⁺CD27⁻CD28⁻ cell frequency and metabolite abundances; Abbreviations: #PC(20:6/22:5): PC(20:6(4Z,7Z,10Z,13Z,16Z,19Z)/22:5(4Z,7Z,10Z,13Z,16Z)); #PS(22:6/22:2): PS(22:6(4Z,7Z,10Z,13Z,16Z,19Z)/22:2(13Z,16Z)); TRE Day 0 group, n = 20; TRE Day 30 group, n = 20, these two groups used the same subject's samples before and after TRE. Statistical analysis was performed using Spearman's correlation coefficients in Fig. 4i.

Fig. 5

Gut microbiome differences before and after 30-day TRE. (a) The Chao1 community richness index, (b) Shannon index and (c) the total number of amplicon sequence variants (ASVs) observed before and after TRE; (d) Changes in the phylum distribution of gut microbiomes before and after TRE; (e) Changes in abundance of Akkermansia and Rikenellaceae before and after TRE; (f) Changes in the genus distribution of gut microbiomes before and after TRE, only the top 30 predominant taxa are shown. (g) Correlation analysis between genus and differential metabolites on Day 0 and Day 30. (h) Correlation analysis between genus and the frequencies of Sene, Th1, Th2, Th17 and Treg cells. TRE Day 0 group, n = 12; TRE Day 30 group, n = 12. Abbreviations: ∗un: unclassified; Clostridiales_Family_XIV: ∗Clostridiales_Family_XIV._Incertae_Sedis_unclassified. Statistical analysis was performed using the paired Wilcoxon test at two different time points, as shown in Fig. 5a, and the paired t-test was performed in Fig. 5a,b&e. Data are the mean ± s.e.m.