Blood metabolism and gut microbes suggest tryptophan-related compounds support healthy aging in centenarians

A total of 425 volunteers were recruited from Suixi county, Guangdong province, China, including 145 centenarians. These participants were divided into four groups: centenarians group (≥100 years old, CE), lineal relatives of the centenarians (24–86 years old, CE‐L), elderly control group (65–88 years old, Elderly), and young control group (23–54 years old, Young), as scheme shown in Figure 1. The medical history, questionnaire, and routine physical examination were conducted for all participants. Based on the previous results, the frailty index (FI) was calculated for the centenarian group, and then used to further distinguish between healthy centenarians (HC) and frail centenarians (FC). Clinical samples, containing blood, saliva, and feces were collected according to participants' preferences. Clinical tests were conducted on these samples, including complete blood count, liver function, renal function, C‐reactive protein, and routine fecal examinations. Then, laboratory examinations were performed for further study, including untargeted metabolomic techniques for serum metabolites, 16S rRNA sequencing for the gut microbiota in feces and the oral microbiota in saliva. The serum metabolites, gut and oral microbiota among various groups were analyzed to identify differential metabolites and bacterial genera. Finally, multi‐omics analyses were integrated to discover the serum metabolic characteristics of centenarians and their relationship with gut or oral microbiota.

Blood samples were collected from all 425 participants for clinical indicator assessment and untargeted metabolomic analysis. The data and clinical indicators for this study cohort are detailed in Table S1. We observed that most health indices of centenarians were significantly reduced compared to those of the other three groups, such as body mass index (BMI), albumin (ALB), white blood cell count (WBC), hemoglobin (HGB), and alanine aminotransferase (ALT). The levels of aspartate aminotransferase (AST) in centenarians were higher than those in the elderly and young group. Besides, the levels of urea were observed to be higher in centenarians compared to the other groups, while the levels of creatinine (Cr) in centenarians were slightly higher than those in the young group. However, there were no significant differences observed in total bilirubin (TB) and direct bilirubin (DB) between the centenarians and the other groups (Figure 2A). These indicators have changed and deviated in centenarians, but they remain within a healthy reference range in the clinic. Therefore, the degree of change in these clinical indicators may be considered for establishing the reference ranges of clinical indicators for the unique population of long‐lived individuals, especially centenarians.

Subsequently, serum metabolites were analyzed using ultrahigh‐performance liquid chromatography‐tandem mass spectrometry (UHPLC‐MS), identifying 12,809 and 12,137 metabolites in positive and negative ion modes, respectively. 808 metabolites at metabolomics standards initiative level 2 confidence were annotated and confirmed through databases, including human metabolome database (HMDB), MassBank, LipidMaps, and mzCloud (Table S2). Partial least squares discriminant analysis (PLS‐DA) was utilized to elucidate the serum metabolic characteristics of the study cohort, with results for the positive ion mode (Figure 2B) and for the negative ion mode (Figure S1A). The PLS‐DA plot revealed a trend in the distribution of metabolites that shifts progressively from the young to centenarians with advancing age. At the principal component 1 (PC1) level, centenarians exhibited significant differences compared to the other three groups, suggesting the presence of distinct metabolites that characterize their unique metabolic profile. At the principal component 2 (PC2) level, no statistical differences were observed between centenarians and both their lineal relatives and young groups, indicating similar characteristics within longevity families (i.e., centenarians and their lineal relatives) that may also resemble those of young individuals.

To identify differential metabolites for centenarians and lineal relatives of the centenarians, metabolites were characterized by three distinct signatures [9]: (Ⅰ) the centenarian signature, (Ⅱ) the longevity family signature, and (Ⅲ) the aging signature. The centenarian signature encompasses metabolites unique to centenarians, which show significant abundance differences for centenarians in comparison with lineal relatives, the elderly and young control groups, but not between the two control groups of the elderly and young group. The longevity family signature includes similar abundance of metabolites between centenarians and their lineal relatives, but with significant abundance differences between centenarians and the two control groups. The aging signature comprises metabolites that vary with age, exhibiting significant differences in abundance among centenarians and the two control groups, with a trend of gradual increase or decrease. Following the above concepts, three distinct categories of metabolites were identified with significantly elevated abundance in the centenarian cohort: (Ⅰ) 21 centenarian‐signature metabolites, (Ⅱ) 36 metabolites with the longevity family signature, and (Ⅲ) 21 aging‐signature metabolites, as illustrated in Figure 2C. Conversely, those metabolites displaying a marked decreasing trend within the centenarian group are delineated in Figure S1B–D; and the enrichment of KEGG pathways was further analyzed between CE and the other three groups, respectively (Figure S2). Specifically, among the top 10 elevated centenarian‐signature metabolites, chenodeoxycholic acid, lithocholic acid glycine conjugate, and 3‐oxo‐5beta‐cholanate belong to the bile acid family, typically secondary bile acids produced by gut microbiota metabolism. Among the top 10 elevated metabolites with the longevity family signature, l‐arginine and dimethylglycine are known metabolites associated with the gut microbiota.

Subsequently, we conducted a Spearman's rank correlation analysis between the aforementioned top 10 metabolites significantly elevated in centenarians and the above‐mentioned clinical indicators shown in Figure 2A (Figure 2D). The analysis revealed that a significant number of metabolites increased in centenarians, such as 7‐methylguanine, hydantoin‐5‐propionic acid, mesobilirubinogen, N7‐methylguanosine, oxindole, and pseudouridine, exhibit positive correlations with urea and Cr levels. This finding implies a likely link between these centenarian‐specific metabolites and renal metabolic capacity, potentially underpinning the sustained renal function observed in this demographic. Similarly, Figure 2E illustrated that metabolites elevated in longevity families, notably l‐arginine and p‐hydroxyphenylacetic acid, also show significant positive correlations with urea and Cr levels, suggesting their potential involvement in renal metabolic processes. Although the exact mechanisms connecting these metabolites to renal metabolism in centenarians and longevity families remain to be fully elucidated, the observed associations highlighted the possible influence of these metabolites on overall health and longevity. This underscores the importance of further exploration into how these metabolic signatures might contribute to physiological resilience and sustained health in these unique populations.

To investigate the characteristics of gut and oral microbiota in centenarians, we conducted 16S rRNA analysis across different cohorts. The alpha diversity analysis of the gut microbiota at ASV level in centenarians revealed that both the Chao richness index and the Shannon diversity index were significantly higher than those in the older and younger groups. However, there was no significant difference in the index of Shannon diversity and Pielou's evenness between centenarians and their lineal relatives. These results indicated that the richness of gut microbiota in centenarians is significantly higher than that in control groups, and the composition of bacteria is similar within the longevity family (Figure 3A). The alpha diversity analysis of oral microbiota showed that the lineal relatives of centenarians possessed the highest Chao richness index, while the centenarians' richness index was similar to the elderly and significantly higher than that observed in the young (Figure 3B). We hypothesized that the common loss of teeth among most centenarians might result in a decreased abundance of certain anaerobic bacteria in the alveolar region, which could subsequently impact microbial diversity compared to their lineal relatives.

Principal coordinate analysis (PCoA) of the gut microbiota composition revealed notable similarities in the microbiota composition between centenarians and their lineal relatives along the PCoA2 axis, distinct from the other two control groups. This suggests a unique microbiota profile shared in the longevity family, significantly different from the unrelated control group (Figure 3C). PCoA‐based beta diversity in the oral microbiota revealed a high overlap in the composition of oral microbiota between centenarians and their lineal relatives (Figure 3D). This may be attributed to the shared living environment and dietary habits within families, resulting in a similar oral microbiota environment.

The linear discriminant analysis effect size (LEfSe) analysis was further employed to identify significantly different genera to investigate the characteristic changes in gut and oral microbiota within the study cohort. In reference to the above analysis of serum metabolite, 3 categories of signature genera were identified: (Ⅰ) characteristic genera with centenarian signature, (Ⅱ) characteristic genera with longevity family signature, and (Ⅲ) genera of the aging signature (Figure 3E). Figure 3E illustrated the top 5 characteristic genera of gut microbiota and the top 3 characteristic genera of oral microbiota across all three signature categories, with the left side representing gut microbiota and the right side representing oral microbiota. These gut bacterial genera that met the above three characteristics were displayed in Figure S3. In the gut microbiota, a total of 16 characteristic genera were identified for centenarian signature, and the top 5 characteristic genera were consecutively listed as Christensenellaceae_R‐7_group, Romboutsia, UCG‐002, Fusicatenibacter, and Anaerostipes. Then, there were 9 characteristic genera with longevity family signature, with the top 5 genera being Dorea, Phascolarctobacterium, Bacteroides, Erysipelotrichaceae_UCG‐003, and Eubacterium_hallii_group. Furthermore, 6 characteristic genera represented the aging signature, with Parabacteroides, Alistipes, Odoribacter, NK4A214_group, and Blautia being the top 5 genera. Notably, the genus of Christensenellaceae and Parabacteroides have been reported to be closely associated with longevity in multiple studies of longevity population in China [17, 18], Italy [19], and Korea [20]. Similarly, these oral bacterial genera that met the above three characteristics were displayed in Figure S4. In the oral microbiota, 4 characteristic genera of centenarian signature were identified, with the top 3 being Porphyromonas, Peptostreptococcus, and Eubacterium_nodatum_group, all of which exhibited reduced abundance in centenarians. Porphyromonas, which has been reported to be associated with an increased risk of periodontitis, may have reduced abundance correlating with tooth loss in centenarians, leading to an overall decline in periodontal pathogenic bacteria. Additionally, 7 characteristic genera with longevity family signature were displayed in Figure S4, and in the top 3 genera, the abundance of Neisseria and Haemophilus showed a decrease in centenarians, whereas the abundance of Achromobacter increased. In the analysis of characteristic genera of aging signature, no genus varied significantly with age. Altogether, we identified characteristic genera from feces and saliva with altered abundances that correlated with longevity, including an increase of potential probiotics in gut microbiota and a reduction of harmful bacteria in oral microbiota, which highlighted the potential novel biomarkers and probiotic candidates for promoting healthy aging.

As mentioned above, centenarian have several specific serum metabolites and genera of gut and oral microbiota. To better elucidate the biomarkers of serum metabolites or gut/oral microbiota to uniquely differentiate the centenarian, we singled out the centenarian group (CE) for comparison with non‐centenarian populations (groups CE‐L + Elderly + Young) (Figure 4). Significant differences in serum metabolite profiles between CE and the non‐centenarians (CE‐L + Elderly + Young) were demonstrated in Figure 4A–C. For initial screening of differential serum metabolites, metabolites with a p‐value less than 0.05 and a value of variable importance in the projection (VIP) above 1.5 were selected, resulting in 64 differential serum metabolites [8]. PLS‐DA score plot was displayed in Figure 4A, exhibiting a significant difference in the centenarian group compared to the group L + E + Y in both PC1 and PC2 axes. The top 30 metabolites ranked by the VIP values are exhibited in Figure 4B. Notably, metabolites such as isoniazid alpha‐ketoglutaric acid, quinolin‐2‐ol, xylulose 5‐phosphate, and d‐ribulose 5‐phosphate are classified as organic acids, primarily involved in energy metabolic processes such as the tricarboxylic acid (TCA) cycle. Furthermore, by intersecting these 64 differential metabolites with the previously identified 47 centenarian‐specific metabolites (Table S3), we identified 17 shared differential serum metabolites (Figure 4C). Of these, 10 metabolites were significantly elevated in centenarians: 2‐(methylamino) benzoic acid, lithocholic acid glycine conjugate, epiandrosterone, ecdysone, N7‐methylguanosine, beta‐tyrosine, mesobilirubinogen, oxindole, indolebutyric acid, and 7‐methylguanine. To evaluate the discriminative potential of these mentioned 10 candidate biomarkers between centenarians and non‐centenarians (group CE‐L + Elderly + Young), receiver operating characteristic (ROC) analysis was performed. The area under the curve (AUC) for these metabolites' logistic regression reached 0.9182, indicating that the 10 metabolites mentioned above can be used as biomarkers for predicting longevity (Figure 4D).

To further elucidate the characteristics of the gut microbiota in centenarians, PCoA was employed to reveal significant differences in the gut microbial composition between centenarians and the non‐centenarian group (Figure 4E). Subsequently, differential genera of centenarians were identified by utilizing LEfSe (LDA ≥ 3) analysis and by considering relative abundance in the cohort (>10%). Several characteristic potential probiotic genera of centenarians overlapped with previous finding, as shown in Figure 3E, including Christensenellaceae R‐7 group, Akkermansia, Pyramidobacter, Alistipes, and Parabacteroides, suggesting that centenarians may possess a favorable gut microbiota composition (Figure 4F). The analysis of a Venn plot in gut microbiota was conducted to compare these 34 differential genera with the previously recognized 27 genera of centenarian signature. The analysis revealed 17 overlapping genera, among which 13 were significantly elevated in centenarians (Figure 4G). Then, ROC analysis was performed on these 13 gut genera, and the AUC reached 0.861, indicating that this analysis had good efficacy for predictive biomarkers of gut microbiota in distinguishing centenarians from the non‐centenarian group (Figure 4H).

Similarly, the PCoA analysis of the oral microbiota between centenarians and non‐centenarians revealed significant differences along the PCoA1 axis (Figure 4I). Differential oral genera of centenarians were identified by utilizing LEfSe (LDA ≥ 3) analysis and by considering relative abundance in the cohort (>10%). The analysis revealed a significant enrichment of Lactobacillus in the oral microbiota of centenarians, a well‐known probiotic, suggesting its potential role in maintaining oral health of centenarians (Figure 4J). Furthermore, the Venn plot of oral microbiota displayed the shared genera between these 4 differential genera and the previously recognized 21 genera with centenarian signature. The analysis revealed three overlapping genera, all of which exhibited a significant decrease in centenarians (Figure 4K). This reduction may be attributed to tooth loss in centenarians, as hypothesized in earlier discussions. Finally, ROC analysis of these three oral genera showed an AUC of 0.7574 (Figure 4L).

Overall, the analysis of serum metabolomics in conjunction with gut and oral microbiomes between centenarians and non‐centenarians has highlighted distinct serum metabolic and microbial traits characteristic of the centenarian population. These common differential metabolites and bacteria genera obtained by different grouping methods are worthy of further study. These screened metabolites and genera serve as potential biomarkers indicative of longevity and possess substantial predictive value.

The metabolomic and gut microbiota profiles of elderly individuals have been reported to be closely linked to their health status in previous studies [21]. During our field investigations, significant individual variability in the physiological and cognitive functions of centenarians was also observed (Table S4). To assess the physical condition and cognitive abilities of centenarians, the frailty index (FI) was evaluated for each centenarian participant according to the comprehensive assessment that included the activity of daily living for assessing behavioral capacity and the minimum mental state examination for cognitive ability. In evaluation assessment system of FI, centenarian with a lower FI score indicates that he or she have better independence and cognitive function. Based on the FI scores, centenarians were further categorized into healthy centenarians (HC, FI < 11) and frail centenarians (FC, FI ≥ 11) to explore the characteristics of individuals with healthy longevity (Table S5).

The distance‐based redundancy analysis (db‐RDA) was employed to evaluate the effects of age, gender, BMI, and FI on metabolomic profiles in CE, and the FI was observed to be the primary determinant influencing the distribution of serum metabolites (Figure 5A). Then, PLS‐DA was used to display the variance in metabolites between HC and FC, revealing significant differences in metabolite distribution between healthy centenarians and frail centenarians (Figure 5B). A total of 34 serum metabolites which were differentially expressed in HC were screened out, with criteria of VIP > 1 and p < 0.05 (Figure 5C). Among the characteristic serum metabolites of HC, 6‐phosphoglucono‐D‐lactone and Dopamine have attracted our attention. The 6‐phosphoglucono‐D‐lactone is known as an essential intermediate in the hexose monophosphate pathway, in which process NADPH is concomitantly generated — a critical component of the cellular antioxidant defense system — to protect cells from oxidative damage [22]. Dopamine, an important neurotransmitter, has well‐established roles in cognitive function and emotional regulation, as validated by numerous studies [23, 24]. Both mentioned metabolites have the ability to help centenarians maintain physical vitality and mental well‐being. In addition, various food‐source substances were enriched in HC, such as 8‐gingerol, baicalein, and sakuranetin [25]. The 8‐gingerol and baicalein are the anti‐inflammatory compounds derived from herbal medicine, whereas sakuranetin, a flavonoid compound, has been confirmed to exert protective effects on rat brain cells through antioxidant mechanisms. This indicates that healthy centenarians may maintain good physical condition through their dietary intake. Furthermore, the ROC curves were constructed using the top 20 metabolites ranked by VIP scores to assess their discriminative validity between HC and FC, and the AUC of the ROC curve reached 0.8578 (Figure 5D). Altogether, serum metabolites serve as a clear indicator for distinguishing the health status and frailty levels of centenarians. Moreover, individuals who achieve healthy longevity often demonstrate superior antioxidant capabilities and emotional regulation abilities.

Similarly, LEfSe analysis was employed to analyze the characteristics of the gut and oral microbiota in HC (Figure 5E). In the gut microbiota, the HC group showed an enrichment of genera such as genera Roseburia, Ruminococcus gnavus group, Negativibacillus, and Flavonifractor, along with a reduction in Senegalimassilia and Family XIII AD3011 group. Among them, the genus Roseburia is known as the cornerstone genus of the gut microbiota and a common producer of short‐chain fatty acids. The genus Flavonifractor contains the well‐known probiotic Flavonifractor plautii which is famous for its ability to reduce the inflammatory response in adipose tissue [26] and acute colitis [27], and to prevent arterial stiffness [28]. Whereas, in the oral microbiota, the HC group demonstrated an enrichment of the genera Acholeplasma, Pseudopropionibacterium, Oribacterium, and g_norank_o_Clostridia_UCG_014, along with a decrease in Propionibacterium. Previously, Oribacterium has been demonstrated to be a beneficial oral genus capable of significantly reducing the risk of plaque and gingival bleeding [29, 30]. The best‐known species of the genus Propionibacterium is Propionibacterium acidifaciens, an oral pathogen related to dentinal caries lesions [31]. A reduction of Propionibacterium suggests that healthy centenarians have a lower risk of suffering from oral diseases.

Finally, the correlation network was used to elucidate the relationships among these elevated serum metabolites and differential genera of gut and oral microbiota in HC (Figure 5F). Plant‐derived sakuranetin was found to be positively correlated with the gut genus Ruminococcus gnavus group, which implied that the gut microbiota was involved in sakuranetin metabolism. Additionally, food‐source metabolites, 8‐gingerol and baicalein were significantly positively correlated with the oral genus Oribacterium, suggesting that dietary intake of these substances may contribute to maintaining oral health by raising the levels of Oribacterium. Overall, within the centenarian between healthy and frail individuals, significant differences in metabolite profiles and microbiota compositions were observed and were associated with diet.

To gain a deeper understanding of the functions of serum metabolites and microbiota in centenarians, the enrichment of KEGG pathways was further analyzed. The significantly differential serum metabolites in CE were found to be enriched in 18 KEGG pathways, with the top 5 enriched pathways including tryptophan metabolism, vitamin B6 metabolism, interconversion of pentoses and glucuronates, the pentose phosphate pathway, and arginine and proline metabolism (Figure 6A). Likewise, we performed KEGG enrichment analysis on the differential serum metabolites between HC and FC groups. The 7 metabolic pathways were involved, with the top 3 being tryptophan metabolism, vitamin B6 metabolism, and terpenoid backbone biosynthesis (Figure 6B). Furthermore, functional prediction analysis by PICRUSt2 from 16S rRNA sequencing data of gut and oral microbiota was conducted separately. According to the centenarian signature, as mentioned above in Figure 2, the tryptophan metabolism pathway exhibited highest fold change compared to the other three groups in the top 20 enriched pathways (Figure 6C). Our results illustrated the potential importance of tryptophan as a link between serum metabolism and gut microbiota for maintaining health during the aging process.

Given the importance of the tryptophan metabolic pathway in serum metabolites and gut microbiota, we compared the relative abundances of all 30 tryptophan‐related metabolites in the serum metabolome (Figure 6D). To better understand the trends with age of tryptophan‐related metabolites, all individuals were reassigned into the age gradient group, ensuring that each cohort had no fewer than 20 individuals, and named as Y20 (20–29 yrs), Y40 (40–49 yrs), Y50 (50–59 yrs), Y60 (60–69 yrs), Y70 (70–79 yrs), Y80 (80–89 yrs), and Y100 groups (≥100 yrs), respectively. 9 metabolites were found to exhibit not only the highest abundance in the CE group but also the highest prevalence in Y100 group. These metabolites include 5‐MIAA, kynurenic acid, methyl (indol‐3‐yl) acetate, L‐kynurenine, hydroxyindole, oxindole, indolebutyric acid, indole‐3‐acetate and 5‐hydroxyindoleacetic acid. As previously described in Figure 2C, we found that oxindole and indolebutyric acid were significantly enriched in the CE group, while 5‐MIAA was notably higher in the HC group, as shown in Figure 5C. Additionally, seven metabolites appeared similar abundance between CE and CE‐L group, and further analysis showed that three metabolites have no significant difference between CE and CE‐L, but there is a significant difference in CE compared to Elderly and Young, specifically for serotonin, indolelactic acid and 6‐hydroxymelatonin (Figure S5). Serotonin [32], a well‐known neurotransmitter, plays a critical role in regulating mood and cognitive function, and its elevation can promote the breakdown of fat in adipocytes, thereby aiding in maintaining blood sugar levels and controlling obesity [33]. Indolelactic acid [34] is noted for its anti‐inflammatory and potential antiviral activities, indicating that these metabolites may contribute to health and longevity. The 6‐Hydroxymelatonin is a primary metabolic of melatonin, metabolized by CYP1A2, which has been reported [35] to protect against oxidative stress and iron‐induced neurotoxicity [36]. The mentioned metabolites of serotonin, indolelactic acid, and 6‐hydroxymelatonin exhibit distinct enrichment characteristics within longevity families, suggesting that the metabolism of these longevity families is more beneficial for maintaining health and promoting longevity.

Moreover, Spearman's rank correlation was employed to analyze the involvement of gut and oral microbiota in tryptophan metabolism pathway across different groups (CE, CE‐L, Elderly, Young, non‐centenarians, HC, and FC). It was observed that four genera in the gut microbiota were positively associated with the tryptophan pathway in both the CE and HC groups, including Christensenellaceae R‐7 group, Eisenbergiella, norank_f_norank_o_Oscillospirales and norank_f_UCG‐010. Meanwhile, we found seven oral genera significantly were positively correlated with the tryptophan pathway in both the CE and HC groups, including Achromobacter, Sphingomonas, Enhydrobacter, and Escherichia Shigella (Figure 6E).

Following these findings on the serum metabolites and microbiota related to the tryptophan pathway, we next focus on exploring the interconnections among tryptophan metabolites, gut and oral microbiota to identify the key metabolites.

To identify the key metabolites of tryptophan pathway that promote healthy longevity, correlation network and pathway reconstruction were further conducted to analyze. Firstly, we performed a correlation analysis among nine tryptophan metabolites enriched in both CE and Y100 group, four genera in gut microbiota, and seven genera in oral microbiota. Significant positive correlations were observed between genus Christensenellaceae R‐7 group and 5‐MIAA (r = 0.37, p = 5.62E‐08), L‐Kynurenine (r = 0.27, p = 9.20E‐05), and oxindole (r = 0.39, p = 5.32E‐09), respectively (Figure 7A). The genus Christensenellaceae R‐7 group has been previously reported as a potential probiotic, and these findings suggest that it may participate in the tryptophan metabolism by influencing the level of 5‐MIAA, L‐Kynurenine, and oxindole. We attempted to recover the identified tryptophan metabolites to the tryptophan pathway and found that 19 of 30 metabolites were present in the tryptophan pathway. Among them, 5 serum metabolites, such as 5‐MIAA, 3‐Methoxyanthranilate, serotonin, L‐Kynurenine and indolepyruvate, showed significant variation in either the CE or HC group, which were marked in Figure 7B, along with their upstream metabolites: 5‐Hydroxy‐L‐tryptophan, L‐Formylkynurenine and L‐Tryptophan. 5‐MIAA was found to be the singular metabolite which was not only enriched in the CE group but also in the HC group (Figure 7B). 5‐MIAA is a downstream product of L‐tryptophan metabolism, and its increase may be attributed to the elevated levels of its precursor — serotonin, a neurotransmitter known for its various beneficial effects. Studies have reported that plasma 5‐MIAA levels in healthy individuals, determined using UHPLC‐MS/MS, range from 118 to 578 pM [37]. It is speculated that 5‐MIAA is not only an important gut microbiota‐derived tryptophan metabolite but may also have significant biological functions in promoting healthy aging.

Therefore, the potential of 5‐MIAA for life extension was further validated through in vitro and in vivo experiments. First, WI‐38 cells, a well‐known model of replicative senescence, were used to assess whether 5‐MIAA could delay senescence. The effects of different 5‐MIAA concentrations on WI‐38 cell viability (PD = 38) were examined to determine the maximum tolerance concentration. The CCK‐8 assay showed that within a range of 0–1000 μM, 5‐MIAA was nontoxic at concentrations below 100 μM. Additionally, compared to 24 h, the cell viability at 72 h of treatment increased by 52.30%, 15.97%, and 6.70% at 5‐MIAA concentrations with 50, 75, and 100 μM, respectively. Notably, WI‐38 cells exhibited sustained proliferation after 72 h of treatment with 50–100 μM 5‐MIAA (Figure 7C, Figure S6). Subsequently, WI‐38 cells cultured with 100 μM of 5‐MIAA for 72 h were subjected to senescence‐associated β‐galactosidase (SA‐β‐Gal) staining. Compared to controls, these cells displayed significantly reduced staining, with the average integrated density decreasing by 51.23% (p = 0.023), suggesting that 5‐MIAA may delay senescence (Figure 7C).

To further validate these findings in vivo, lifespan assays were conducted using Caenorhabditis elegans (C. elegans). Animals treated with 0.5% DMSO had a mean lifespan of 19.95 ± 0.74 days, whereas those treated by 1 mM 5‐MIAA had a mean lifespan of 22.93 ± 0.96 days (Figure 7D). The average lifespan of C. elegans was extended by 14.94% (p = 0.003), indicating that 5‐MIAA promotes life extension.

Given the strong association between 5‐MIAA and gut microbiota, we evaluated its effects on intestinal mucosal cells (Caco‐2 cell) to further elucidate its mechanism in the intestinal epithelium. To assess the protective potential of 5‐MIAA, we first determined its impact on Caco‐2 cell viability under dextran sulfate sodium (DSS)‐induced stress. The half‐maximal inhibitory concentration (IC50) of DSS for Caco‐2 cells was 60 mg/mL, while the maximum safe dose of 5‐MIAA was 100 μM (Figure S7). Subsequently, Caco‐2 cells were co‐stimulated with DSS and 5‐MIAA to evaluate the changes in cell viability (Figure 7E). The results showed that cell viability in the 5‐MIAA‐treated group was significantly higher than that in the DSS‐model group (p = 1.96E‐04), suggesting that 5‐MIAA effectively protects Caco‐2 cells from DSS‐induced viability loss. Additionally, we examined the mRNA expression levels of tight junction‐associated proteins and inflammatory cytokines in DSS‐treated Caco‐2 cells co‐stimulated with 5‐MIAA (Figure 7E). Compared to the model group, co‐treatment with 5‐MIAA significantly increased OCLN mRNA levels (p = 0.0094), while reducing the expression of the inflammatory cytokine TNF (p = 0.0215). These findings indicated that 5‐MIAA effectively mitigates DSS‐induced damage in Caco‐2 cells, enhancing the barrier‐protective function of intestinal epithelial cells.

Collectively, we have identified 5‐MIAA, a key tryptophan metabolite, as strongly correlated with genus Christensenellaceae R‐7 group in centenarians, particularly healthy individuals. It plays a pivotal role in delaying cell senescence, extending lifespan, and mitigating inflammation in intestinal cells.

The cohort for this study was recruited from Suixi county, Zhanjiang city, located in the western region of Guangdong province, China, which has been elected as the “International Longevity and Health Care Base.” A total of 425 participants were enrolled, including 145 centenarians, as well as 126 lineal relatives of the centenarians, 92 elderly individuals and 62 young individuals.

All participants underwent medical history interviews, questionnaire surveys, and routine physical examinations. Participants were included in this study when met these criteria: (1) aged ≥ 20; (2) absence of major diseases, such as cancer, autoimmune diseases, or Alzheimer's disease; and (3) no history of using hormones, immunosuppressants, or anticoagulant medications. In addition, the elderly and young control group have been investigated the history of family longevity, and participants were recruited when he/she has no blood relatives over 80 years old within three generations. Exclusion criteria for all participants were as follows: (1) cognitive impairment, severe organic diseases, or history of major trauma or surgery within the past 2 years; (2) history of Clostridium difficile infection or use of medications that may affect gut microbiota (e.g., antibiotics or hormones) within the 3 months before sample collection; (3) use of antimicrobial agents or anti‐inflammatory drugs (such as NSAIDs) within the past 3 months; and (4) a history of gastrointestinal diseases, including gastrointestinal cancer, chronic or infectious gastroenteritis, colitis, gastritis, gastrointestinal ulcers, or gastrointestinal bleeding. For the control group, additional exclusion criteria included a history of gastrointestinal diseases, diabetes, hypertension, hyperlipidemia, coronary heart disease, cerebrovascular disease, autoimmune diseases, inflammatory bowel disease, severe organ damage, or systemic diseases.

Participants were divided into four groups: the centenarian group (CE), the lineal relatives of centenarians group (CE‐L), the elderly control group (Elderly), and the young control group (Young). Blood, fecal, and saliva samples were collected from volunteers who were willingness to provide them, resulting in a total of 425 serum samples, 347 saliva samples, and 291 fecal samples. Blood and fecal samples were sent to the Clinical Laboratory of Affiliated Hospital of Guangdong Medical University for biochemical analysis and routine fecal testing. Additionally, a frailty index (FI) was calculated based on medical history and questionnaire results for the centenarian group. Centenarians were further divided into a healthy group and a frail group based on the top half of FI scores. Detailed information on the frailty index is provided in Table. S4

Fasting venous blood samples were collected from participants and then left to clot at room temperature for 30 min. The serum was obtained after being centrifuged at 3500 rpm for 10 min and frozen at −80°C until use for untargeted metabolomics analysis by liquid chromatograph mass spectrometer (LC‐MS). In brief, the serum samples were mixed with methanol, vortexed and centrifuged at 12,000 rpm for 10 min at 4°C. Then, the supernatant was collected, concentrated and dried for detection.

The FastPure DNA Extraction Kit (TransGen Biotech, Beijing, China) was used to extract the microbial genomic DNA from fecal and saliva samples, according to the manufacturer's instructions. Then, the V3‐V4 regions of 16S rRNA gene of the bacteria were amplified through the primer pair 338F (5′‐ACTCCTACGGGAGGCAGCAG‐3') and 806R (5′‐GGACTACHVGGGTWTCTAAT‐3′) on a T100 Thermal Cycler (BIO‐RAD, USA). The amplification for PCR was conducted under the following processes: initial denaturation at 95°C for 3 min, followed by 27 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 45 s, with a final extension at 72°C for 10 min, and held at 4°C. The PCR products were sequenced on Illumina NextSeq 2000 PE300 platform at Majorbio Bio‐Pharm Technology Co., Ltd.

Metabolomics analysis: the raw data was converted into the mzXML format by ProteoWizard (v3.0.8789), and then the XCMS (R package, v3.1.3) was used to identify, filtrate and align the peaks. The metabolites identified by accuracy mass (<30 ppm) and MS/MS data were matched with HMDB (https://www.hmdb.ca/↗), massbank (https://massbank.jp/↗), LipidMaps (http://www.lipidmaps.org↗), mzcloud (https://www.mzcloud.org↗) and KEGG (http://www.genome.jp/kegg/↗). After scaling data, models were established on partial least‐square discriminant analysis (PLS‐DA). Moreover, all the models were evaluated by overfitting with methods of permutation tests. The discriminating metabolites were determinated by the VIP. The p‐value, VIP produced by PLS‐DA, fold change (FC) was applied to discover the contributable‐variable.

16S rRNA sequencing: after quality control, the raw paired‐end sequences were performed using the fastp software (https://github.com/OpenGene/fastp↗, v0.19.6), and paired reads were merged using FLASH (http://www.cbcb.umd.edu/software/flash↗, v1.2.11). Processed sequences were denoised using DADA2 (or Deblur) plugins in the Qiime2 pipeline under default parameters, resulting in amplicon sequence variants (ASVs). Taxonomic classification of ASVs was conducted with the Naive Bayes (or Vsearch or Blast) classifier in Qiime2, based on the SILVA 16S rRNA gene database (v138). Functional predictions of microbial communities were performed through PICRUSt2 software (v2.2.0).

CCK8 assay: the cell viability of WI‐38 was tested by the CCK8 method. WI‐38 cells (Abbkine, Wuhan, China) at a concentration of 5.0 × 10³ cells per 100 μL of MEM medium (Gibco, USA) were added to a 96‐well plate. The assay was conducted using a commercial kit from Beyotime Biotechnology (C0038, Shanghai, China). A solution of 10 μL CCK8 was infused into each well after culturing with different concentrations of 5‐MIAA. The absorbance at 450 nm was measured by MicroplateReader (ThermoFisher, USA). Cell viability was calculated as: Cell viability (%) = [A (metabolite) − A (blank)]/[A (control) − A (blank)] × 100%.

Cellular Senescence‐associated β‐galactosidase (SA‐β‐gal) staining: WI‐38 cells were cultured in MEM medium to reach replicative senescence (PD = 38). SA‐β‐gal staining was conducted using the commercial kit (No.C0602, Beyotime, Shanghai, China). The cells were treated by MEM medium with and without 100 μM 5‐MIAA, respectively. After incubation for 72 h, the WI‐38 was washed one time with PBS and fixed with the fixative solution for 15 min at room temperature. Subsequently, the cells were stained according to the kit's manual and incubated at 35°C without CO₂ overnight. The cells were washed three times with PBS for 3 min each and then photographed using the MF52‐M microscope (Mshot). Pictures were taken for over three visions per well, and the ImageJ software was used to calculate the percentage of cells staining positively with SA‐β‐gal.

The strain and maintenance of Caenorhabditis elegans (C. elegans): the wild‐type N2 strain of C. elegans (SunyBiotech, Fuzhou, China) was used. All C. elegans were cultivated on the standard plates of nematode growth medium (NGM) at 20°C with Escherichia coli OP50 as a food source.

Lifespan assay: Synchronized L4 larvae of worms were transferred to three plates of NGM (nearly 30 per plate, diameter 3.5 cm) containing inactivated OP50 and 20 μM of FUDR (F0503, Sigma, Germany). Then, worms were fed with 5‐MIAA in 0.5% DMSO at concentrations of 0.05, 0.1, 0.5, 1, 2, and 10 mM, with 0.5% DMSO in M9 buffer as the control. The day of the transfer to the plate was recorded as Day 0. To ensure the potency of 5‐MIAA and sufficient food, worms were transferred to fresh plates every other day. The survival data were plotted with the Kaplan–Meier curves in GraphPad Prism (v10). Statistical significance was tested by the log‐rank (Mantel–Cox) method.

Caco‐2 cells were cultured in DMEM (Gibco, USA) with 10% FBS (NZ500, NEWZERUM), 1% NEAA (Gibco, USA), and 0.5% penicillin‐streptomycin (Gibco, USA). Upon reaching 85% confluence, they were digested with 0.25% trypsin (Gibco, USA), planted in 96‐well plates at 6000 cells per well, and incubated for 24 h. After PBS washing, metabolites were added at final concentrations of 10, 100, 250, 500, and 1000 μM, and cells were incubated for 24 and 48 h. Then, 100 μL of 10% CCK‐8 solution was added, and the OD value at 450 nm was measured after 2 h.

DSS (60316ES25, Yeasen Biotechnology, Shanghai, China) was suspended in sterile water, filtered through a 0.22 μm membrane, and diluted in culture medium to various concentrations (0–100 mg/mL). Cell viability was assessed using the CCK‐8 after treating cells in 96‐well plates with 60 mg/mL DSS and different metabolites for 24 or 48 h. Total RNA was extracted using the Trizol (Invitrogen, USA) method, and cDNA was synthesized with PrimeScript™ RT Master Mix (RR036A, Takara, Japan). RT‐qPCR was performed in a 10 µL reaction volume with TB Green Premix Ex Taq II (RR820A, Takara, Japan) under the following processes: 95°C for 30 s, followed by 40 cycles of 95°C for 5 s and 60°C for 30 s, and melting curve analysis at 95°C for 10 s, 65°C for 60 s, and 95°C for 1 s. Relative mRNA levels of OCLN and TNF were calculated by the ^△△Ct method with β‐actin as a reference. The primer sequences were: β‐actin (Forward: 5′‐CCTTCCCTCCTCAGATCATTGC‐3′, Reverse: 5′‐ATACTCCTGCTTGCTGATCCAC‐3′); OCLN (Forward: 5′‐CACGCTTGCCTGGGACAGAG‐3′, Reverse: 5′‐TCTGTATAGCCTCCGTAGCC‐3′); TNF (Forward: 5′‐CTCATCTACTCCCAGGTCCTCTTC‐3′, Reverse: 5′‐CGATGCGGCTGATGGTGTG‐3′).

Statistical data was analyzed by R (v4.3.3) and GraphPad Prism (v10). All the experiment data are presented as the Mean ± SD. Wilcoxon rank‐sum test or T test was performed for comparisons between two groups. Significance was achieved when false discovery rate (FDR) < 0.05 or p < 0.05, where appropriate. Bioinformatic analysis was partially performed using the OmicStudio tools at https://www.omicstudio.cn/tool↗.

The raw sequence data of the serum metabolome and 16S rRNA reported in this paper have been deposited in the Genome Sequence Archive [66] at the National Genomics Data Center [67], China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (serum metabolome: OMIX009339, gut microbiota: CRA021552 and oral microbiota: CRA021555). These data are publicly accessible at https://ngdc.cncb.ac.cn/omix/release/OMIX009339↗, https://ngdc.cncb.ac.cn/gsa/browse/CRA021552↗ and https://ngdc.cncb.ac.cn/gsa/browse/CRA021555↗, respectively. The data and scripts used are saved in GitHub https://github.com/XuemengLi6758/Qiu2024iMeta↗. Supplementary materials (figures, tables, graphical abstract, slides, videos, Chinese translated version, and update materials) may be found in the online DOI or iMeta Science http://www.imeta.science/↗.