Bidobacterium adolescentis improves lifespan and healthspan by regulating catalase activity and oxidative stress-associated metabolites

Microbiota-host interaction was involved in aging, while the specic bacterium was undetermined. To identify candidate bacterium with aging, we performed fecal microbiota sequencing. Less richness of gut microbial community, and a reduction of B.adolescentis abundance was observed in elderly individuals. B. adolescentis supplement improved osteoporosis and neurodegeneration in telomerase RNA component deletion (Terc −/− ) aged mice. B.adolescentis induced prolongevity and healthspan improvement in Drosophila melanogaster and C. elegans. Transgenic deletion of ctl-2 in C. elegans abolished the effect on lifespan and healthspan by B. adolescentis. The catalase activity was decreased in skeletal muscle and brain tissues of Terc −/− mice, as well as cellular senescence in mouse embryonic broblasts. B. adolescentis alleviated ROS accumulation by regulation of oxidative stress-associated metabolites. These results suggest a role for B. adolescentis in improving lifespan and healthspan by regulating catalase activity and host metabolism. Supplement with commensal bacteria is a promising strategy against age related diseases.


Main Manuscript
Aging is a process and status of holistic host organs, characterized as physiological function deterioration, cellular function decline underpins the development of pathological alterations 1 . Studies have been conducted to extend lifespan or induce healthy aging in different model organisms, ranging from yeast, zebra sh, worms, ies to mice [2][3][4] . As the largest component of host microbiota, gut microbes have been proved to function in a variety of physiological and pathological processes. Recent studies demonstrated that fecal microbiota transplantation could extend lifespan and healthspan in progeria mice 5 . Studies have been performed on the elderly or centenarian populations through 16S rRNA, metagenomic or metatranscriptomic shotgun sequencing to distinguish the compositional and functional alteration of gut microbiota with aging [6][7][8] . However, biological effect of speci c commensal bacteria on aging and the potential mechanisms were unclear.
Here, we observed that the abundance of Bi dobacterium adolescentis (B. adolescentis) exhibited a signi cant decline in elderly populations by fecal microbiota sequencing, which was con rmed in 3500 samples from GMrepo database. B. adolescentis supplement improved age-related osteoporosis and neurodegeneration in telomerase RNA component deletion (Terc −/− ) aged mice, as well as inducing lifespan prolongation and healthspan improvement in nonmammalian organisms Drosophila melanogaster and Caenorhabditis elegans. The anti-aging effect of B. adolescentis was regulated by the activity of catalase and oxidative stress-associated metabolites.

Results
Gut microbiota diversity and B. adolescentis abundance decreased with aging To investigate the change of gut microbiota with aging, we performed 16S rRNA sequencing of fecal samples. 166 participants were divided into three groups according to chronological age (Supplementary Table 1). We found that alpha-diversity estimators shannon index and heip index of human gut microbiota signi cantly decreased from the younger to the elderly populations (Fig. 1A,B), indicating less richness of gut microbiota with aging. A remarkable reduction of Firmicutes, Actinobacteria, Saccharibacteria, together with an increment of Fusobacteria, were detected with age at phylum level ( Fig. 1C). The composition of gut microbiota didn't change with age in a linear association 6 . In addition, an increment of Bacteroidetes:Firmicutes (B:F) ratio and a shift of butyrate-producing bacteria were observed in centenarians 6,7 . We observed that 45 KEGG pathways indispensable for maintaining lifespan, including 28 metabolism-associated pathways were differentially enriched with aging. Most metabolismassociated pathways were predicted to be more active in elderly individuals, such as lipopolysaccharide biosynthesis, N-glycan biosynthesis, glutathione metabolism and amino acid metabolism ( Supplementary Fig. 1). Collectively, gut microbiota pro les altered with aging and might participate in the regulation of host metabolism.
We subsequently performed linear discriminant analysis (LDA) coupled with effect size measurements to screen out candidate bacteria (Fig. 1D). Notably, the B. adolescentis distribution was the most signi cant difference between two groups (Supplementary Table 2). Short chain fatty acids (SCFAs) were recognized as bene cial bacterial metabolites which were important for host metabolism. We found that relative abundance of bacterial taxa associated with SCFA production, such as B. adolescentis, Ruminococcaceae, Faecalibacterium prausnitzi and Eubacterium ractale, was signi cantly decreased in elderly individuals while B. adolescentis exhibited most prominent reduction (Fig. 1E). Furthermore,relative abundance of B. adolescentis was con rmed to be signi cant higher in younger individuals by qPCR assay (Fig. 1F). Similar result was obtained by analyzing GMrepo sequencing database between young and old ages (n = 2821 vs. n = 679 samples) (Fig. 1G). As reported in previous study, microbiota in aged people has displayed reduction in the abundance of several bacteria with antiin ammatory and immunomodulatory properties, including Bi dobacterium, Akkermansia, Lactobacillus and Christensenellaceae 9 . Taken together, we concluded that the abundance of B. adolescentis was decreased with aging and B. adolescentis might play pivotal role in the regulation of healthy aging .
Oral gavage with B. adolescentis alleviated aged related osteoporosis and neurodegeneration in Terc −/− progeroid mice Telomerase, consisting of three main components, is essential for maintaining telomere length and plays an important role in tissue renewal and organism lifespan 10 . Telomerase RNA component deletion (Terc −/ − ) mice with C57BL/6 background show progressive telomere shortening from rst generation (G1) until the third (G3) generation, which exhibits signi cant phenotype of premature aging [11][12][13] . To verify the effect of B. adolescentis on age, Terc −/− G3 progeroid mice of 6-8 weeks age were oral gavaged with B.
The body weight of Terc −/− mice was signi cantly lower than those of wild-type Terc +/+ mice, whereas B. adolescentis supplement signi cantly increased body weight as compared to control group (Fig. 2B).
Aging has been characterized by multiple organic dysfunction, including musculoskeletal and neurodegenerative diseases 15,16 . To assess the biological effect of B. adolescentis on bone density, we performed micro-CT scan and three-dimensional reconstruction of femora in mice. The bone volume/total volume (BV/TV) and trabecular thickness (Tb.Th) of Terc −/− mice were signi cantly decreased as compared to wild-type mice. These indices in Terc −/− mice were increased after gavaged with B. adolescentis (Fig. 2D), which suggested B. adolescentis supplement could improve osteoporosis in aged Terc −/− mice. Previous study showed that neurodegenerative changes in brain was linked with gut microbiota 17 . We then assessed senescence status by comparing the morphological changes and surviving number of neurons in the CA3 region of hippocampus in mice (Fig. 2E). Nuclear deviation, cytoplasm condensed and nuclear fragmented of neurons were more prominent in Terc −/− mice than those of control group, and the surviving number of neurons in mice gavaged with B. adolescentis also showed a signi cant increase compared to controls. Collectively, these results indicated that supplement with B. adolescentis improved healthspan in Terc −/− progeroid mice.

B. adolescentis supplement improved lifespan and healthspan in D. melanogaster and C. elegans
Invertebrate organisms such as the fruit y D. melanogaster and the nematode C. elegans with a relative short lifespan, availability of different genetic mutants and morphological and functional similarities of gut were broadly used as proof-of-concept models on microbiome-aging studies [18][19][20] . Some bacterial strains, such as Comamonas DA1877 and Lactobacillus gasseri SBT2055, have been identi ed to in uence the lifespan and reproduction of C. elegans by regulating series of signaling pathways [21][22][23] . We administrated B. adolescentis and conventional food to D. melanogaster and C. elegans, and comprehensively verify the effect of B. adolescentis on the lifespan and healthspan of the two organisms. Wild-type D. melanogaster strains (both w 1118 and Canton-S) supplemented with B. adolescentis both showed a signi cant increase of lifespan (Fig. 3A,B), approximately 20% increment was observed (Supplementary Table 3). Then we tested healthspan parameters of ies on day 30. The climbing ability of female w 1118 supplemented with B. adolescentis was improved as compared to control group (Fig. 3C). Canton-S ies supplemented with B. adolescentis exhibited stronger creep ability than controls as well (Fig. 3D).
The lifespan of C. elegans was also signi cantly improved when supplied with B. adolescentis with different mixture ratios (Fig. 4A). 1:1 mixture with B. adolescentis and E. coli OP50 can signi cantly improve the mean maximum lifespan of C. elegans (Fig. 4B). In addition, the locomotion ability of aged worms was signi cantly enhanced (Fig. 4C,D). We observed that the heat stress resistance was obviously changed with B. adolescentis supplement (Fig. 4E). Healthspan was further evaluated by auto uorescence quanti cation of intestinal lipofuscin, which accumulated with age ( Fig. 4F) The auto uorescence intensity of worms in B. adolescentis intervention group was signi cantly lower than that of control group (Fig. 4G). Collectively, B. adolescentis supplement could improve lifespan and healthspan in both D. melanogaster and C. elegans. ctl-2 is essential for B. adolescentis-induced lifespan extension and healthspan improvement in C.

elegans
To elucidate the mechanisms of B. adolescentis-induced lifespan improvement, genes expression involved in lifespan were evaluated in C. elegans and D. melanogaster. Expression of sod-3 and ctl-2 was signi cantly higher in C. elegans supplemented with B. adolescentis than control group (Fig. 5A). Similar result was observed in ies considering corresponding homologous gene sod-3 and cat (Supplementary Fig. 2A,B). To clarify the gene involved in lifespan, experiments of corresponding mutants were performed in C. elegans. Interestingly, B. adolescentis supplement could still extend the lifespan in C. elegans carried with sod-3 mutant, while the lifespan prolongation was abolished in ctl-2 mutant (Fig. 5B-D). To validate this nding, we constructed transgenic worms and detected the expression of ctl-2 directly with mCherry uorescence. The ctl-2 expression was signi cantly increased with B. adolescentis supplement in aged worms (Fig. 5E,F). We then detected healthspan indicators aforementioned in ctl-2 mutant C. elegans. The enhancement of locomotion ability by B. adolescentis was blocked in ctl-2 mutant (Fig. 5G). Similarly, no signi cant improvement of survival time was observed in ctl-2 mutant with B. adolescentis supplement (Fig. 5H). Moreover, the auto uorescence intensity was increased in ctl-2 mutants group (Fig. 5I,J). In conclusion, B. adolescentis supplement could prolong lifespan and improve healthspan of C. elegans through the regulation of ctl-2.
B. adolescentis suppressed aged Terc −/− G3 mice by regulating the activity of catalase and oxidative stress-associated metabolites The activity of catalase (CAT), which was homologous to C. elegans ctl-2 gene, was subsequently detected in muscle and brain tissue of mice. Terc −/− aged mice exhibited decreased activity of CAT compared to wild-type mice, and B. adolescentis supplement signi cantly enhanced the activity of CAT (Fig. 6A). In addition, B. adolescentis-gavaged Terc −/− aged mice showed prominent increased protein expression of CAT in muscle and brain tissues (Fig. 6B,C and Supplementary Fig. 3A).
Immunohistochemistry staining revealed that B. adolescentis supplement exhibited downregulation of p53, while upregulation of CAT, in cortex and hippocampus regions ( Fig. 6D and Supplementary Fig. 3B). These results demonstrated that B. adolescentis supplement suppressed aged related phenotype in Terc −/− G3 mice by regulating CAT.
To verify the effect of B. adolescentis in vitro, B. adolescentis was then administrated to culture medium in both replicative and DOX-induced senescent MEFs. B. adolescentis supplement signi cantly suppressed cellular senescence as shown by senescence-associated β-galactosidase staining (Fig. 6E,F and Supplementary Fig. 4A,B). In line with observation in mice, the mRNA and protein expression level of CAT was upregulated by B. adolescentis in senescent MEFs (Fig. 6G,H and Supplementary Fig. 4C).
Finally, we performed metabolomics analysis of mice feces to evaluate the effect of B. adolescentis on oxidative stress-associated metabolites as CAT was an important ROS scavenger 24 . In accordance with our ndings in brain and muscle tissues, apiin and erucic acid, which could increase the activity of

Discussion
Gut microbial community was changed with aging, and the abundance of B. adolescentis elicited a dramatic decline in elderly, which was supported by sequencing results in 3500 fecal samples from GMrepo database. Several bacteria have been previously identi ed to be related with aging and played critical role in aging prediction, such as family Ruminococcaceae, genera Alistipes, Bacteroides, Bi dobacterium, Faecalibacterium, Akkermansia, Roseburia and Eubacterium 9,39,40 . Transplantation with the gut microbiota of old donor mice to young germ-free mice exhibited age-sensitive enrichment in butyrate-producing microbes 41 . We observed that oral gavaged with B. adolescentis induced lifespan extension and healthspan improvement in both Terc −/− mice, and nonmammalian model organisms such as worms and ies. Transplantation with the Akkermansia muciniphila was su cient to exert bene cial healthspan effects in two progeroid mouse models of Hutchinson-Gilford progeria syndrome 5 . These results demonstrate the association between aging and the gut microbiota, and provide experimental evidence with commensal bacteria against age-related diseases.
Gut microbiota communicates with host organs through bacterial structural product, nutrient-metabolites and complex pathways. Development of series organic dysfunction was demonstrated to be associated with gut microbiota, including neurodegenerative diseases 16 49 . High-throughput lifespanassociated screening on C. elegans uncovered that 29 E. coli mutants and bacterial metabolite polysaccharide colonic acid were involved in the prolongevity process 50 . Transplant of Akkermansia muciniphila had bene cial effects in progeroid mice by reestablishing healthy microbiome through restoring secondary bile acids 5 . In present study, we observed that the accumulation of intestine lipofuscin, a lipid peroxidation product, was signi cantly reduced in B. adolescentis-treated C. elegans.
Furthermore, B. adolescentis supplement in Terc −/− mice regulated lipid metabolism and oxidative stressassociated metabolites. In the prediction of human gut microbiota function, endocytosis and phagocytosis were predicted to be more active in younger populations. Hence, we can hypothesize that probiotics may regulate age-related disorders through gut in ammation and host metabolism.
To gain an insight into the underlying mechanisms, we identi ed candidate ctl-2 gene function in Terc −/− mice, C. elegans and senescent MEFs supplemented with B. adolescentis. Studies have shown that ctl-2 accounts for the majority of catalase activity in nematodes, and lack of ctl-2 can lead to the premature aging of nematodes 51 . As an important substance in living organisms, catalase is participating in the process of active oxygen metabolism. Under environmental stress, ROS resulted in cell membrane damage, DNA damage and subsequent cell senescence 24 . Pathogenic microbes have been reported to impair intestinal cell repair and shorten host lifespan by generation of ROS in ies 52 . A recent study demonstrated that preventing ROS accumulation in gut allows survival without sleep in ies 53 . We found that the catalase activity in muscle and brain was upregulated by supplementation with B. adolescentis in Terc −/− mice, which was supported by the metabolomic analysis of mice feces. Metabolites facilitating the activity of CAT together with other antioxidants were enriched in B. adolescentis-gavaged Terc −/− mice, which could remarkably reduce the intracellular ROS level. In addition, erucic acid, cosmosiin and 2hydroxycinnamic acid have been proposed as potential therapeutic agents for Alzheimer disease by ameliorating nueroin ammation and blocking neural cell death 27,29,37 . Daidzin could inhibit LPS-induced bone loss by suppressing the osteoclast differentiation 54 . The enrichment of these metabolites may underlie the improvement of phenotype in Terc −/− mice supplemented with B. adolescentis.
In conclusion, we showed that B. adolescentis exerted bene cial effects on lifespan and healthspan by the regulating of catalase activity and host metabolism. The underlying molecular mechanisms should be further explored.

Study design
The objectives of this study was to screen out candidate bene cial bacterium and elucidate its functional characteristics on aging and prolongevity. Fecal samples were collected from healthy volunteers and 16S rRNA sequencing was performed. Microbiota data obtained from GMrepo metagenomes database were used to con rm the B. adolescentis abundance in different age. Next, agerelated phenotype was conducted in Terc -/aged mice, D. melanogaster and C. elegans with B. adolescentis supplement. Micro-CT bone scan and staining in hippocampal CA3 region was evaluated in Terc -/aged mice gavaged with B. adolescentis. Lifespan was recorded and series healthspan indices were measured, which including mean survival time, frailty index score, locomotion ability and intestinal lipofuscin accumulation. For further mechanism exploration, lifespan assay on C. elegans carried with serial gene mutants were performed with B. adolescentis supplement. Lastly, expression of ctl-2 homologous gene catalase (CAT) was detected in muscle and brain tissues from Terc -/aged mice, as well as in senescent MEFs cells by qRT-PCR, western blot and immunohistochemistry assays.
Metabolomics of mice feces were performed to analyze effect of B. adolescentis on gut microbial metabolites. Researchers were blinded to group allocation, and mice, ies and worms were randomized to groups. All procedures were conducted in compliance with institutional guidelines and were approved by the Animal Ethical Committee of Zhejiang University prior to initiating the study.
Healthy subjects and fecal samples collection Online database resources Microbiota online data were obtained from GMrepo, an metagenomes database of human gut 56 . Using python software through RESTful APIs, we obtained relative abundance of B. adolescentis in samples with healthy phenotype of in all age groups from GMrepo. 3500 samples were included for nal analysis.

Micro-CT imaging
The left bones of mice were isolated and the attachments of the left femora muscle were removed. Three femora samples of each group were randomly selected and scanned by micro-CT (InspeXio SMX-225 CT FPD HR; Shimadzu Co. Ltd., Kyoto, Japan) after xation in 4% paraformaldehyde overnight. Each sample was reconstructed using micro-CT software (VGStudio MAX; Volume Graphics, Heidelberg, Germany) under the same conditions. A region of intersect (ROI) cube was taken underneath epiphyseal growth plate. The reconstruction parameters of bone volume/total volume (BV/TV) and trabecular thickness (Tb.Th) were analyzed.

Lifespan assay
For lifespan assay, ies were allowed to develop and mate for 4~6 days after eclosion. Then they were starved for 2h in an empty vial, sexually segregated and randomly assigned to intervention or control vials. Every 20 ies were ipped to a vial and transferred to new vials every 2 or 3 days. The number of dead ies were counted simultaneously. About 100 ies (5 vials) were used each assay.
For C. elegans, L4 stage (day 0) worms were cultured in OP50 plates until maturation and then transferred to OP50 or mixture plates.150 worms were distributed in 10 plates (15 worms/plate), and incubated at 20 °C. The worms were transferred to new freshly seeded plates every other day. The number of alive or dead worms were counted on every transfer day.
Survival rates were calculated as the percentage of surviving ies/worms versus the total number of ies/worms. Flies that lost and worms died as a result of getting stuck to the wall of the plate were excluded from the analysis. The evaluation of lifespan was performed at least three times.
C. eleganselegans uorescence microscopy Worms were synchronized and cultured as described in online supplementary materials and methods. On day 10 and 14, the lipofuscin accumulation in worm intestine and ctl-2 expression were quanti ed with auto uorescence and mCherry red uorescence, respectively. Worms were randomly selected and wash twice with M9 buffer. Then they were mounted on a slide coated with 200mM sodium azide to induce anesthesia and photos were taken under blue excitation light (405~488 nm) or red excitation light (559-585 nm) with inverted laser scanning confocal microscope (Olympus IX81-FV1000, Tokyo, Japan). Fluorescence was quanti ed with ImageJ (National Institutes of Health, Bethesda, MD, USA). Three independent experiments were conducted with >20 worms.

Metabolomics analysis
Metabolites from mice fecal samples were extracted according to the standard protocols by Biotree Biological Technology Co. Ltd. (Shanghai, China). LC-MS/MS analyses were performed using an UHPLC system (Vanquish, Thermo Fisher Scienti c) with a UPLC BEH Amide column (2.1 mm × 100 mm, 1.7 μm) coupled to Q Exactive HFX mass spectrometer (Orbitrap MS, Thermo). The raw data were converted to the mzXML format using ProteoWizard and processed with an in-house program, which was developed using R and based on XCMS, for peak detection, extraction, alignment, and integration. Then an in-house MS2 database (BiotreeDB) was applied in metabolite annotation. The cutoff for annotation was set at 0.3. Details were described in online supplementary materials and methods.

Statistical analysis
Differences of three groups were analyzed by one-way analysis of variance (ANOVA) for data with normal distribution or Kruskal-Wallis test for data with non-normal distribution.