Abstract

   Metabolic dysregulation and altered metabolite concentrations are
   widely recognized as key characteristics of aging. Comprehensive
   exploration of endogenous metabolites that drive aging remains
   insufficient. Here, we conducted an untargeted metabolomics analysis of
   aging mice, revealing citrulline as a consistently down-regulated
   metabolite associated with aging. Systematic investigations
   demonstrated that citrulline exhibited antiaging effects by reducing
   cellular senescence, protecting against DNA damage, preventing cell
   cycle arrest, modulating macrophage metabolism, and mitigating
   inflammaging. Long-term citrulline supplementation in aged mice yielded
   beneficial effects and ameliorated age-associated phenotypes. We
   further elucidated that citrulline acts as an endogenous metabolite
   antagonist to inflammation, suppressing proinflammatory responses in
   macrophages. Mechanistically, citrulline served as a potential
   inhibitor of mammalian target of rapamycin (mTOR) activation in
   macrophage and regulated the mTOR–hypoxia-inducible factor
   1α–glycolysis signaling pathway to counter inflammation and aging.
   These findings underscore the significance of citrulline deficiency as
   a driver of aging, highlighting citrulline supplementation as a
   promising therapeutic intervention to counteract aging-related changes.
     __________________________________________________________________

   Citrulline acts as an endogenous metabolite antagonist to aging and
   inflammation by regulating mTOR-HIF1α-glycolysis pathway.

INTRODUCTION

   Aging, a prevalent risk factor for numerous chronic diseases in humans,
   presents a persistent and notable public health challenge ([38]1). The
   intricate relationship between metabolism and aging has garnered
   extensive attention and investigation, revealing metabolic
   dysregulation and altered metabolites as key hallmarks of aging ([39]2,
   [40]3). However, whether these metabolic changes are consequences of
   aging or drivers of the aging process remains largely unexplored.
   Aging-associated metabolic dysregulations often result in the
   accumulation or deficiency of certain metabolites ([41]4, [42]5). For
   instance, the metabolite trimethylamine N-oxide (TMAO) has been found
   to accumulate in human blood with aging and is associated with
   detrimental effects ([43]6). High concentrations of TMAO are linked to
   an increased risk of cardiovascular disease and all-cause mortality
   ([44]7, [45]8). In contrast, nicotinamide adenine dinucleotide (NAD)
   has emerged as a prominent example of an aging hallmark metabolite
   ([46]9–[47]11). Supplementation with NAD precursors, such as
   nicotinamide mononucleotide (NMN), has shown promise as a strategy for
   protecting against age-associated diseases and potentially extending
   health span and life span ([48]12, [49]13). Recently, Singh et al.
   ([50]14) highlighted taurine deficiency as a contributor to aging, and
   taurine supplementation has demonstrated potential in slowing down key
   markers of aging. Another noteworthy example is spermidine, extensively
   studied for its ability to enhance longevity in various animal models
   ([51]15). Dietary intake of spermidine from nutrients has been found to
   correlate with reduced cardiovascular and cancer-related mortality in
   human epidemiological studies ([52]16). Numerous studies have shown
   that changes in endogenous metabolites exhibit remarkable conservation
   across different species ([53]17). Therefore, understanding the role of
   metabolites in aging is crucial for unraveling the mechanisms
   underlying age-related diseases, developing effective interventions,
   and promoting healthy aging. However, the comprehensive exploration of
   endogenous metabolites that drive the aging process remains
   insufficient.

   Aging is characterized by chronic inflammation, commonly referred to as
   inflammaging, which manifests as a low-grade, persistent
   proinflammatory state ([54]18). Inflammaging plays a pivotal role in
   tissue dysfunction and the development of age-related pathologies, and
   it has emerged as a biomarker for accelerated aging ([55]19).
   Traditional inflammation intervention strategies, such as anticytokine
   therapy and immune cell depletion, may have serious safety concerns and
   limited applicability in extending human life span. Alternatively,
   metabolic intervention to suppress inflammaging has shown promising
   effects in extending both health span and life span. In particular,
   metabolite supplementation through daily nutrition has become prevalent
   intervention to modulate inflammaging and counter aging. For example,
   spermidine supplement has been shown to enhance immune function,
   resulting in the delay of age-related pathologies in various organisms
   ([56]15, [57]20). In addition, supplementation with the endogenous
   metabolite α-ketoglutarate has been demonstrated to extend lifespan and
   compress morbidity in aging mice by counteracting inflammatory
   processes ([58]21). These findings highlighted that targeting
   inflammaging serves as a potential strategy to counteract aging.

   The immune system undergoes a progressive decline as aging ([59]22).
   This decline is attributed to age-related changes in macrophages, which
   are major contributors to age-associated inflammation (i.e.,
   inflammaging) ([60]23). Macrophages play a crucial role in initiating
   inflammatory responses against pathogens and facilitating tissue
   repair, adapting their function based on local cytokine signals to
   promote either proinflammatory or anti-inflammatory responses ([61]24,
   [62]25). However, aging leads to phenotypic and functional changes in
   macrophages, affecting processes such as phagocytosis, wound healing,
   and polarization ([63]26, [64]27). Dysregulation of anti-inflammatory
   function of macrophages has been proposed as a contributing factor to
   the increased susceptibility of the elderly people to sepsis and
   inflammatory disorders ([65]28). Previous study has demonstrated that
   metabolic shifts during aging, including alterations in oxidative
   phosphorylation, glycolysis, and urea cycle, play regulatory roles in
   macrophage-mediated inflammatory responses ([66]23, [67]29). These
   aging-related metabolic shifts mediate pathogen clearance and tissue
   repair and increase susceptibility to infections and age-related
   inflammatory diseases ([68]30). Thus, the age-related impairment of
   macrophage metabolism, particularly the loss of anti-inflammatory
   capacity, serves as a potential target for countering aging.

   In this study, we used liquid chromatography–mass spectrometry
   (LC-MS)–based untargeted metabolomics to comprehensively analyze
   endogenous metabolites in various organs of aging mice. We observed the
   consistent down-regulation of citrulline with aging. Further results
   demonstrated that citrulline supplementation effectively reduced
   cellular senescence, protected against DNA damage, prevented cell cycle
   arrest, and mitigated inflammaging and other age-related phenotypes in
   various cell and mouse models. Notably, long-term citrulline
   supplementation in aged mice yielded beneficial effects, ameliorating
   age-associated phenotypes and extending health span. Comprehensive
   mechanistic investigations established that citrulline acts as an
   endogenous metabolite antagonist to inflammation and enables to
   suppress proinflammatory responses through mediating the mammalian
   target of rapamycin (mTOR)–hypoxia-inducible factor 1α
   (HIF1α)–glycolysis signaling pathway in macrophages to counter aging.
   These findings underscore the critical role of citrulline deficiency as
   a driver of the aging process and highlight citrulline supplementation
   as a potential therapeutic intervention to counteract aging-related
   changes.

RESULTS

Citrulline is significantly down-regulated during aging

   Metabolic dysregulation is widely acknowledged as one of the key
   characteristics of aging. To explore the role of metabolism in the
   aging process, we conducted a comprehensive analysis of metabolic
   changes in brain tissue, liver tissue, and serum samples from male mice
   at different ages (6, 24, 36, 52, and 78 weeks old). LC-MS–based
   untargeted metabolomics was used. By applying the Kruskal-Wallis test
   and Spearman correlation, we identified 75, 106, and 80 age-associated
   metabolites in the brain, liver, and serum, respectively ([69]Fig. 1A
   and fig. S1, A and B). Metabolic pathway enrichment analysis of
   age-associated metabolites revealed significant enrichment primarily in
   purine metabolism, pyrimidine metabolism, and five other pathways in
   mouse brain tissue, while purine metabolism, pyrimidine metabolism,
   amino sugar and nucleotide sugar metabolism, pyruvate metabolism, and
   nine other pathways were enriched in mouse liver tissue ([70]Fig. 1B).
   Furthermore, mouse serum exhibited enrichment in the cysteine and
   methionine metabolism pathway, phenylalanine metabolism, nicotinate and
   nicotinamide metabolism, and β-alanine metabolism pathway ([71]Fig.
   1B). Further analysis of all age-associated metabolites in mouse brain
   tissues, liver tissues, and serum identified seven metabolites
   consistently associated with aging across all three sample groups
   ([72]Fig. 1, C and D). These metabolites included 1,5-anhydroglucitol,
   citrulline, cyclamic acid, TMAO, ergothioneine, anserine, and
   homocitrulline ([73]Fig. 1E and fig. S1C). Previous studies have
   already linked 1,5-anhydroglucitol ([74]31), ergothioneine ([75]32),
   TMAO ([76]6, [77]8), and homocitrulline ([78]33) to aging. Notably, our
   findings revealed specific down-regulation of citrulline, along with
   1,5-anhydroglucitol, in all three sample types ([79]Fig. 1F). In
   particular, in mouse liver tissue, as early as 24 weeks, the level of
   citrulline decreased by ~50% compared to that at 6 weeks. Citrulline
   levels decreased by ~20% during aging in mouse brain and serum.
   Collectively, these findings highlight the identification of citrulline
   as a consistently down-regulated metabolite associated with aging in
   mice.

Fig. 1. Citrulline is significantly down-regulated during aging.

   [80]Fig. 1.
   [81]Open in a new tab

   (A) Graphic illustration of the workflow to profile age-associated
   metabolites in brain tissue, liver tissue and serum from mice at
   different ages using LC-MS. Age-associated metabolites were screened by
   Kruskal-Wallis test (P < 0.05) and Spearman correlation (P < 0.05). w,
   weeks; m/z, mass/charge ratio. (B) Pathway enrichment analyses of
   age-associated metabolites from mouse brain, liver, and serum
   (hypergeometric test, ***P < 0.001, **P < 0.01, and *P < 0.05). TCA,
   tricarboxylic acid. (C) Age-associated metabolites were screened by
   Kruskal-Wallis test (P < 0.05) and Spearman correlation (P < 0.05). Red
   dots (brain, n = 26; liver, n = 46; serum, n = 21) and blue dots
   (brain, n = 49; liver, n = 60; serum, n = 59) represent metabolites
   that were increased and decreased over the age, respectively. Black
   dots represent unchanged metabolites. (D) Venn diagram for the overlap
   of age-associated metabolites in mouse brain, liver, and serum obtained
   from (C). (E) The spearman factors of seven common age-associated
   metabolites in mouse brain, liver, and serum datasets. Red and blue
   bars represent metabolites that were increased and decreased over the
   age, respectively. n = 3. Three dots for each metabolite represent
   their Spearman factors with age obtained from mouse brain, liver, and
   serum datasets, respectively. (F) Relative abundances of citrulline in
   mouse brain, liver, and serum at different ages. Bars represent
   means ± SEM. n = 8 to 10; all numbers are biologically independent
   samples. ***P < 0.001, **P < 0.01, and *P < 0.05; n.s., not significant
   (two-tailed Student’s t test).

Citrulline supplementation mitigates age-associated phenotypes in vitro and
ex vivo

   To investigate the biological function of citrulline in the context of
   aging, we used a reactive oxygen species–induced cell senescence model
   for preliminary exploration ([82]34). First, we demonstrated that
   levels of citrulline were dose-dependently decreased in senescent mouse
   embryonic fibroblast (MEF) cells upon the addition of tert-butyl
   hydroperoxide (tBHP), a reactive oxygen species inducer ([83]Fig. 2A).
   These results validated the citrulline deficiency observed in aging
   mice and confirmed the relevance of the senescent MEF cell model.
   Significantly, we further demonstrated that treatment with citrulline
   (500 μM) during passage of senescent MEF cells led to a notable
   reduction in the levels of γ–H2A histone family member X (γH2AX), a
   molecular marker associated with DNA damage and repair ([84]Fig. 2B).
   This finding indicated that elevated levels of citrulline contributed
   to a decrease in DNA damage in senescent cells. Furthermore, citrulline
   supplementation exhibited a significant inhibition of
   senescence-associated β-galactosidase (SA-β-Gal) activity ([85]Fig. 2C
   and fig. S2A). Aging is associated with a decline in cell division
   capacity, leading to cellular senescence characterized by irreversible
   growth arrest, including increased expression of senescence markers
   such as p21. Our study suggests that citrulline supplementation
   attenuates the senescence phenotype by decreasing the expression of p21
   ([86]Fig. 2D). This intervention could potentially delay the onset of
   senescence and improve cell division capacity in MEFs. In addition to
   cellular markers, we assessed the impact of citrulline supplementation
   on the senescence-associated secretory phenotype, characterized by the
   secretion of high levels of inflammatory cytokines. Our findings
   revealed that citrulline supplementation effectively reduced the mRNA
   levels of inflammatory cytokines, including Tnf, Il6, and Il1b, in
   senescent MEF cells, but not normal ones ([87]Fig. 2E).

Fig. 2. Citrulline supplement mitigates age-associated phenotypes in vitro
and ex vivo.

   [88]Fig. 2.
   [89]Open in a new tab

   (A) Schematic illustration of senescent immortalized MEF (iMEF)
   generation using tBHP. n = 6 to 7. h, hours; d, days. (B)
   Representative images illustrating γH2AX (green) levels responding to
   citrulline (Cit; 500 μM for 4 days) in iMEFs with
   4′,6-diamidino-2-phenylindole (DAPI)–stained nuclei (blue). n = 3; each
   point averaged from 3 images. Scale bar, 50 μm. (C) Representative
   SA-β-Gal activity staining images in iMEFs responding to citrulline
   (500 μM for 4 days). Scale bar, 50 μm. (D) p21 mRNA levels in iMEFs
   responding to citrulline (500 μM for 4 days). n = 6. (E) mRNA levels of
   Tnf, Il6, and Il1b in iMEFs responding to citrulline (500 μM for 4
   days). n = 5 to 6. (F) Graphical illustration of BMDM generation and
   citrulline levels in BMDMs. n = 5. (G) Representative images
   illustrating γH2AX (green) levels responding to citrulline (500 μM for
   24 hours) in BMDMs with DAPI-stained nuclei (blue). n = 3; each point
   averaged from three images. Scale bar, 100 μm. (H) Representative
   SA-β-Gal activity images in aged BMDM cells after citrulline treatment
   (500 μM for 24 hours). n = 3. Scale bar, 10 mm. (I) mRNA levels of Tnf,
   Il6, and Il1b in BMDMs after citrulline treatment (500 μM for 24
   hours). n = 3 to 4. (J) mRNA levels of TNF and IL1B in young (19 years
   old, male) and aged (56 years old, male) human macrophages after
   citrulline treatment (500 μM for 12 hours). n = 6 to 8. Bars represent
   means ± SEM. ***P < 0.001, **P < 0.01, and *P < 0.05 (two-tailed
   Student’s t test).

   Furthermore, we extended our investigation to primary bone
   marrow–derived macrophages (BMDMs) obtained from both young and aged
   mice (6 and 78 weeks old, male) to further explore the role of
   citrulline ex vivo. Consistently, intracellular levels of citrulline
   were found to be significantly down-regulated in aged BMDMs compared to
   young BMDMs ([90]Fig. 2F), providing additional validation of its
   relevance in the context of aging mice. Consistent with the results
   obtained from senescent MEF cells, we observed that citrulline
   supplementation in aged BMDM cells effectively attenuated
   age-associated phenotypes. This was demonstrated by the reduction in
   the DNA damage marker γH2AX, the inhibition of SA-β-Gal activity, the
   decrease in mRNA levels of inflammatory cytokines such as Tnf, Il6, and
   Il1b, and the cell cycle arrest marker p16, p19, and p21 ([91]Fig. 2, G
   to I, and fig. S2, B and C). Notably, these effects were not observed
   in young BMDM cells, further emphasizing the specific impact of
   citrulline in the context of aging. We also treated peripheral blood
   mononuclear cell (PBMC)–derived human macrophages with or without
   citrulline. Similar to BMDMs derived from mice, we observed that
   citrulline supplementation in aged human macrophages attenuated the
   mRNA levels of inflammatory cytokines, including TNF and IL1B ([92]Fig.
   2J). Collectively, citrulline supplementation demonstrated its ability
   to mitigate age-associated phenotypes, including DNA damage, senescence
   markers, and inflammatory cytokine levels, in both in vitro and ex vivo
   models. These findings underscore the significance of citrulline
   deficiency as a key driver of the aging process.

Beneficial effects of long-term citrulline supplementation in aged mice

   To investigate the biological function of citrulline in vivo in
   counteracting aging, we conducted a long-term administration of
   citrulline to both young and aged mice (6 and 72 weeks old, male) via
   drinking water for ~9 weeks ([93]Fig. 3A). LC-MS analyses clearly
   demonstrated a significant increase in citrulline levels in aged mouse
   brain tissue, liver tissue, and BMDMs (fig. S3A). Longitudinal
   monitoring of mouse body weights revealed a significant decrease in
   body weights of old mice after ~5 to 6 weeks of citrulline
   supplementation, while the body weights of young mice remained
   unaffected ([94]Fig. 3B). To eliminate the contribution of body weight,
   we further selected and analyzed mice with overlapping body weights
   between the aged group (n = 3) and the aged + citrulline group (n = 6)
   (fig. S3B; P = 0.63, two-tailed Student’s t test). We then compared the
   mRNA levels of aging-associated inflammatory cytokines in brain and
   liver tissues. The results showed that citrulline supplementation
   reduced the mRNA levels of inflammatory cytokines (Tnf, Il6, and Il1b)
   in these tissues (fig. S3, C and D). Furthermore, organ weights such as
   the spleen and liver were also reduced in response to 9 weeks of
   citrulline supplementation exclusively in old mice ([95]Fig. 3C).
   During aging, the liver and spleen of mice increase in weight primarily
   due to factors such as fat accumulation ([96]35, [97]36) and
   inflammatory responses ([98]37). Obesity is often accompanied by a
   low-grade chronic inflammatory state ([99]38). In light of this,
   anti-inflammatory nutrition has emerged as a pharmacological approach
   to treating obesity ([100]39). We hypothesize that citrulline
   supplementation may prevent the increase in spleen and liver weight by
   inhibiting inflammation during aging. Moreover, we assessed the ability
   of long-term citrulline supplementation to alleviate age-associated
   phenotypes in mice. Citrulline supplementation significantly decreased
   the mRNA levels of aging-associated inflammatory cytokines (Tnf, Il6,
   and Il1b) in the whole-brain tissues, whole-liver tissues, and primary
   BMDMs of old mice ([101]Fig. 3, D to F). Microglia, which are
   tissue-specific macrophages in the brain, have been reported to be
   activated during aging, resulting in a notable increase in the
   expression of specific markers such as ionized calcium binding adaptor
   molecule 1 (IBA1). Our research findings suggest that citrulline
   inhibits the activation of microglia in aged brain tissues ([102]Fig.
   3, G and H). Immunofluorescence staining of γH2AX in brain tissues
   further confirmed a substantial reduction in age-induced DNA damage
   following citrulline administration ([103]Fig. 3, G and I). These
   findings were also supported by immunofluorescence staining of P21 and
   P16, markers of cell cycle arrest associated with aging ([104]Fig. 3, H
   and J, and fig. S3, E and F). Notably, these effects of citrulline
   supplementation were not observed in young mice, providing additional
   evidence that citrulline deficiency is a key driver of aging. Overall,
   long-term supplementation of citrulline to aged mice resulted in
   increased endogenous citrulline levels and effectively reduced body
   weights, as well as spleen and liver weights, along with other
   age-associated phenotypes. These results demonstrate the potential of
   citrulline as a therapeutic intervention to counteract aging-related
   changes.

Fig. 3. Beneficial effects of long-term citrulline supplementation in aging
mice.

   [105]Fig. 3.
   [106]Open in a new tab

   (A) The long-term supplementation scheme of young and aged mice (male)
   with drinking water only or drinking waters containing citrulline (1
   g/kg of body weight of mice) for about 9 weeks (6 and 72 weeks old,
   male). n = 8. (B) Recorded body weights of young and aged mice (male)
   supplemented with water or citrulline over 9 weeks. n = 8. (C) Weights
   of spleen and liver organs in young and aged mice (male) supplemented
   with water or citrulline over 9 weeks. n = 6 to 8. (D to F) Real-time
   polymerase chain reaction (PCR) analyses of senescence-associated
   secretory phenotype genes, including Tnf, Il6, and Il1b in mouse brain
   (D), liver (E) and primary BMDMs (F). n = 5 to 6. (G and H)
   Representative immunofluorescence images of γH2AX (G) and P21 (H)
   levels in brain tissues of young and aged mice (male) supplemented with
   water or citrulline over 9 weeks. The microglia in the brain were
   stained with anti-IBA1 antibody (green), and the nuclei were stained
   with DAPI (blue). Scale bars, 50 μm. (I and J) The averaged density of
   positive cells in images from (G) and (H). n = 3; each point represents
   the mean densities of three images in each sample. Bars represent
   means ± SEM. ***P < 0.001, **P < 0.01, and *P < 0.05 (two-tailed
   Student’s t test).

Citrulline is an endogenous metabolite antagonist to inflammation

   While previous studies have established the role of citrulline in
   modulating inflammation, its function in the context of aging has yet
   to be explored. To address this, we first conducted experiments to
   validate the anti-inflammatory function of citrulline. Our findings
   from LC-MS analysis revealed that lipopolysaccharide (LPS) treatment
   significantly increased endogenous citrulline levels in various
   macrophages, including immortalized BMDM (iBMDM), BV2, and RAW264.7
   cells (fig. S4, A to G). Notably, citrulline supplementation
   effectively reduced the mRNA expressions of inflammatory cytokines
   (Tnf, Il6, and Il1b) and secretory proteins [tumor necrosis factor–α
   (TNFα) and interleukin-6 (IL-6)] in LPS-treated iBMDMs ([107]Fig. 4, A
   and B). Similar results were observed in BV2 and RAW264.7 cells (fig.
   S4, H to K). These results strongly support the notion that citrulline
   acts as an endogenous metabolite antagonist to inflammation.

Fig. 4. Citrulline is an endogenous metabolite antagonist to inflammation.

   [108]Fig. 4.
   [109]Open in a new tab

   (A) Tnf, Il6, and Il1b levels in iBMDMs responding to citrulline (4.5
   hours) and LPS (100 ng/ml for 4 hours). n = 12, three samples with four
   technical replicates. (B) TNFα and IL-6 levels in iBMDM culture medium
   responding to citrulline (4.5 hours) and LPS (100 ng/ml for 4 hours).
   n = 5 to 8. (C) The scheme of mice administrated for 7 days. (D and E)
   Mouse survival after LPS challenge using log-rank (Mantel-Cox) test.
   (F) Citrulline levels in mice treated with citrulline and LPS (20
   mg/kg). n = 8. (G) Tnf, Il6, and Il1b levels in mouse liver treated
   with citrulline and LPS (20 mg/kg). n = 5 to 8. (H) TNFα, IL-6, and
   IL-1β levels in mouse serum treated with citrulline and LPS (20 mg/kg).
   n = 8 to 16. (I) Graphical illustration of BMDM generation and
   citrulline levels in BMDMs treated with LPS (100 ng/ml for 12 hours).
   n = 5 or 8. (J) Tnf, Il6, and Il1b levels in BMDMs responding to
   citrulline (500 μM for 12.5 hours) and LPS (100 ng/ml for 12 hours).
   n = 3 to 4. (K) Quantifications of SA-β-Gal activities in BMDMs
   responding to citrulline (500 μM for 12.5 hours) and/or LPS (100 ng/ml
   for 12 hours). n = 3. (L) Graphical illustration of human macrophages
   from PBMCs. (M) TNF, IL1B, and IL6 levels responding to citrulline (500
   μM for 12.5 hours) and LPS (100 ng/ml for 12 hours) in human
   macrophages. n = 6 to 8. Bars represent means ± SEM. ***P < 0.001,
   **P < 0.01, and *P < 0.05 (two-tailed Student’s t test).

   Furthermore, we sought to validate the anti-inflammatory effects of
   citrulline in mouse models ([110]Fig. 4C). In an LPS-induced acute
   liver injury model, mice administered LPS and galactosamine (GalN)
   experienced fatal liver injury within 8 hours ([111]Fig. 4D). However,
   citrulline administration significantly prolonged the survival time of
   mice with acute liver injury (P = 0.003). Similarly, in a sepsis model
   where mice were exposed to a lethal dose of LPS (35 mg/kg of body
   weight), citrulline administration also extended the survival time of
   septic mice (P = 0.0176; [112]Fig. 4E). To further confirm the
   antagonist role of citrulline in inflammation, we administered mice a
   nonlethal low dose of LPS (20 mg/kg of body weight) and measured the
   levels of inflammatory cytokines. Our results demonstrated that in the
   context of LPS-induced inflammation, citrulline levels were
   up-regulated in mouse liver tissues but down-regulated in mouse serum
   ([113]Fig. 4F). Supplementation of citrulline increased the endogenous
   levels of citrulline in both the liver and serum of mice. Increased
   endogenous citrulline inhibited the expressions of Il6 and Il1b in
   mouse liver and reduced the protein levels of IL-6 and IL-1β in mouse
   serum, while having no significant effects on TNFα ([114]Fig. 4, G and
   H, and fig. S4L). We found citrulline also influenced the inflammasome,
   indicating by inhibiting the IL-1β release to the medium in iBMDMs
   (fig. S4M). These in vitro findings align with the observed effects in
   the in vivo model, particularly on IL-1β release. These findings
   provide further evidence for the antagonist role of citrulline in
   inflammation.

   Then, we extended our investigation to determine whether the
   anti-inflammatory effect of citrulline remains functional in the
   context of aging. To achieve this, we used primary BMDMs obtained from
   both young and aged mice (6 and 78 weeks old, male). Consistently, LPS
   treatment significantly increased the endogenous levels of citrulline
   in both young and aged BMDMs, albeit to a slightly lesser extent in
   aged BMDMs ([115]Fig. 4I). To assess the impact of citrulline on
   age-associated inflammation, we measured the levels of inflammatory
   cytokines. LPS-induced inflammation was significantly heightened in
   aged BMDMs; however, supplementation with citrulline effectively
   reduced the mRNA expressions of age-associated inflammatory cytokines
   (Tnf, Il6, and Il1b) exclusively in LPS-induced aged BMDMs, while no
   significant effect was observed in young BMDMs ([116]Fig. 4J).
   Consistently, citrulline supplementation successfully rescued the
   activity of SA-β-Gal in LPS-treated aged BMDMs ([117]Fig. 4K and fig.
   S4N). We also detected the mRNA expressions of age-associated
   inflammatory cytokines in PBMC-derived human macrophages. Similar to
   mouse macrophages, citrulline down-regulated the mRNA expressions of
   TNF, IL6, and IL1B exclusively in LPS-treated aged human macrophages,
   while no significant effect was detected in young human macrophages
   ([118]Fig. 4, L and M). These findings provide compelling evidence that
   citrulline acts as an endogenous metabolite antagonist to inflammation,
   thereby serving as a potential approach to counteract aging.

Citrulline modulates macrophage metabolism to counter aging and inflammation

   To further investigate the regulatory mechanism of citrulline in aging
   and inflammation, we treated young and aged primary BMDMs with or
   without citrulline and subjected to comprehensive untargeted
   metabolomics profiling ([119]Fig. 5A). A total of 192 metabolites were
   altered between young and aged BMDMs, and citrulline supplementation
   rescued 102 of these metabolites (fig. S5, A and B). Principal
   components analysis (PCA) revealed that age-associated metabolic
   alterations in BMDMs were partially rescued upon citrulline treatment,
   particularly in PC1 dimension ([120]Fig. 5A). Metabolic pathway
   enrichment analysis further showed that metabolic activity of aged
   BMDMs was up-regulated compared to young BMDMs, but this increase was
   attenuated in aged BMDMs with citrulline treatment ([121]Fig. 5B). In
   particular, age-induced up-regulation of glycolysis and gluconeogenesis
   pathway in BMDMs, which is an important energy metabolism pathway, was
   partially diminished by citrulline supplementation. Therefore, we
   propose that citrulline might exert its antiaging and anti-inflammation
   effects through influencing the glycolysis pathway. Furthermore, our
   data validated that citrulline supplementation rescued aging-induced
   expressions of glycolysis genes encoding Hk2 (hexokinase 2), Pkm2
   (pyruvate kinase muscle isozyme M2), Pgk1 (phosphoglycerate kinase 1),
   Pgam1 (phosphoglycerate mutase 1), Pfkl (phosphofructokinase, liver
   type), Eno1 (enolase 1), Gpi (glucose-6-phosphate isomerase), and Ldha
   (lactate dehydrogenase A), as well as glycolysis metabolites, including
   fructose 6-phosphate, pyruvate, and lactate in aged BMDMs ([122]Fig. 5C
   and fig. S5C). Since mTOR and HIF1α signaling pathway was well reported
   to regulate glycolysis and had close associations with aging and
   inflammation, we further demonstrated that increased mRNA and protein
   expressions of HIF1α in aged BMDMs were also rescued by citrulline
   treatment ([123]Fig. 5D). Consistently, we observed that the activation
   of mTOR in aged BMDMs was also rescued by citrulline supplemention
   through measuring phosphorylation of mTOR phosphorylation (mTOR;
   Ser^2448) and its downstream substrates p70S6 kinase (p70S6k; Thr^389)
   and 4E-binding protein 1 (4EBP1; Thr^37/46) ([124]Fig. 5E). These
   results suggested that citrulline may modulate macrophage metabolism
   and regulate mTOR-HIF1α-glycolysis signaling pathway to counter aging
   ([125]Fig. 5F).

Fig. 5. Citrulline modulates macrophage metabolism to counter aging and
inflammation.

   [126]Fig. 5.
   [127]Open in a new tab

   (A) BMDMs from mice were treated with citrulline (500 μM for 12 hours)
   for untargeted metabolomics. (B) Sixteen enriched metabolic pathways in
   aged BMDMs treated with citrulline (500 μM for 12 hours). CoA, coenzyme
   A. (C) Glycolysis gene levels in BMDMs treated with citrulline (500 μM
   for 12 hours). n = 3 to 4. (D) HIF1α expressions in BMDMs treated with
   citrulline (500 μM for 12 hours). n = 3 to 4. (E) p-mTOR, mTOR,
   p-p70S6k, p70S6k, p-4EBP1, and 4EBP1 expressions in BMDMs responding to
   citrulline (500 μM for 12 hours). (F) A proposed model for citrulline
   to counter aging and inflammation through modulating
   mTOR-HIF1α-glycolysis. (G) p-mTOR, mTOR, p-p70S6k, p70S6k, p-4EBP1, and
   4EBP1 expressions r4esponding to citrulline (500 μM for 4.5 hours) and
   LPS (100 ng/ml for 4 hours) in iBMDMs. (H) HIF1α expressions in iBMDMs
   treated with citrulline (500 μM for 4.5 hours) and LPS (100 ng/ml for 4
   hours). n = 3. (I) Glycolysis genes in iBMDMs treated with citrulline
   (500 μM for 4.5 hours) and LPS (100 ng/ml for 4 hours). n = 3. (J) ECAR
   of iBMDMs treated with citrulline (500 μM for 4.5 hours) and LPS (100
   ng/ml for 4 hours). n = 6 to 11. (K and L) Representative images of
   p-S6 (K) and HIF1α (L) in mouse brain. Microglia were stained with
   anti-IBA1 antibody (green) and DAPI (blue). Scale bars, 50 μm. (M and
   N) Averaged density of positive cells in (K) and (L). n = 3; each point
   represents the mean from three images. Bars represent means ± SEM.
   ***P < 0.001, **P < 0.01, and *P < 0.05 (two-tailed Student’s t test).

   We further demonstrated the regulatory role of citrulline in the
   context of inflammation by modulating mTOR-HIF1α-glycolysis pathway.
   Treatment with citrulline reduced the LPS-induced elevation of
   phosphorylated mTOR, p70S6k, and 4EBP1 in iBMDM cells ([128]Fig. 5G).
   Consequently, increased mRNA level and protein level of HIF1α
   responding to LPS induction were also rescued by citrulline
   supplementation in iBMDMs ([129]Fig. 5H). Measurements of mRNA
   expressions of glycolysis genes, including Pfkl, Pgk1, Tpi, Pkm2, Gpi,
   Pgam1, Ldha, Hk2, Gapdh, Eno1, and Aldoa, and extracellular
   acidification rates (ECARs) showed citrulline supplementation decreased
   LPS-induced up-regulation of glycolysis activity in iBMDM cells
   ([130]Fig. 5, I and J, and fig. S5D). Consistently, LPS-induced
   up-regulations of glycolysis metabolites, including dihydroxyacetone
   phosphate, phosphoenolpyruvate, fructose 1,6-bisphosphate, glucose
   1,6-bisphosphate, and lactate, were also rescued by citrulline
   treatment (fig. S5E). Oxygen consumption rate (OCR) analysis showed the
   addition of citrulline significantly increased LPS-induced adenosine
   5′-triphosphate (ATP) production in iBMDM cells (fig. S5F).

   We further confirmed that citrulline exerts its anti-inflammatory
   effects through the suppression of the mTOR-HIF1α-glycolysis pathway
   under both reduced and elevated citrulline conditions. We first used
   l-NG-nitroarginine methyl ester (L-NAME), a widely used inhibitor of
   nitric oxide synthase 2 (NOS2) ([131]40), to create a
   citrulline-depleted environment. Consistently, our results showed that
   L-NAME treatment reduced citrulline levels in a dose-dependent manner
   (fig. S5G). Accordingly, Western blot results revealed that the
   LPS-induced activation of the mTOR pathway was enhanced in response to
   L-NAME treatment (fig. S5H). Furthermore, citrulline supplementation
   effectively suppressed mTOR activation in L-NAME–inhibited iBMDMs (fig.
   S5I). Consistently, in L-NAME–inhibited iBMDMs, we observed that the
   mTOR inhibitor rapamycin significantly decreased the mRNA expression of
   LPS-induced proinflammatory cytokines, including Tnf, Il6, and Il1b
   (fig. S5J). In addition, we found that L-NAME treatment decreased
   citrulline levels without influencing arginine levels, indicating that
   the anti-inflammatory effect was independent of arginine (fig. S5K). We
   also used α-methyl-dl-aspartic acid, a specific argininosuccinate
   synthase 1 (ASS1) inhibitor ([132]41), to induce citrulline
   accumulation. Western blot results indicated that the LPS-induced
   activation of the mTOR pathway was inhibited in response to
   α-methyl-dl-aspartic acid treatment (fig. S5L). Furthermore, under
   citrulline supplementation condition, we used the mTOR agonist MHY1485
   or the HIF1α stabilizer dimethyloxalylglycine to promote mTOR
   activation or HIF1α stabilization, respectively. These treatments
   resulted in increased HIF1α protein levels (fig. S5M) and elevated mRNA
   and protein levels of the proinflammatory cytokine TNFα (fig. S5, N and
   O).

   Immunofluorescence staining of p-S6 (an in vivo marker for mTOR
   activity) and HIF1α in brain tissues from aged mice confirmed
   significant reductions in age-induced activations of mTOR and HIF1α
   after long-term administration of citrulline over 9 weeks ([133]Fig. 5,
   K to N).These results demonstrated that citrulline supplementation
   rescued age-associated metabolic alterations and modulated the
   mTOR-HIF1α-glycolysis pathway in BMDMs, indicating its regulatory role
   in macrophage metabolism to counter aging and inflammation.

Age-dependent down-regulation of Nos2 contributed to citrulline deficiency in
macrophages

   We identified the deficiency of citrulline during the aging process;
   however, the factors contributing to this deficiency remain unknown. To
   investigate this further, we used stable-isotope tracing metabolomics
   technology using [U-^13C]-arginine as a tracer ([134]Fig. 6A). We
   administered 0.4 mM [U-^13C]-arginine to both young and old primary
   BMDMs for 24 hours and measured the ^13C-labeled metabolites involved
   in citrulline metabolism, namely, arginine, ornithine,
   argininosuccinate, and citrulline. Unexpectedly, there were no
   significant changes observed in the total levels or ^13C-labeled forms
   of arginine, ornithine, and argininosuccinate between the young and old
   BMDMs ([135]Fig. 6, B to D). The results were consistent with our
   untargeted metabolomics data in mice (fig. S6A). However, the M6
   isotopologue of citrulline exhibited a significant decrease in old
   BMDMs ([136]Fig. 6E). Citrulline anabolism involves two pathways. The
   first pathway involves the conversion of ornithine to citrulline in the
   mitochondria. In this case, the M6 isotopologue of arginine gives rise
   to the M5 isotopologue of ornithine, which further generates the M5
   isotopologue of citrulline. However, as illustrated in [137]Fig. 6E, no
   M5 isotopologue of citrulline was detected, suggesting that the
   conversion of ornithine to citrulline in the mitochondria does not
   occur in BMDM metabolism. The second pathway entails the direct
   conversion of arginine to citrulline through NOS2. In this case, the M6
   isotopologue of arginine directly forms the M6 isotopologue of
   citrulline. Consequently, our experiment measured the M6 isotopologue
   of citrulline, which exhibited a significant decline in aged BMDMs.
   Furthermore, we assessed the mRNA expression of Nos2, which
   consistently showed a decrease in aged BMDMs ([138]Fig. 6F). We
   conducted similar experiments using tBHP-induced senescent MEF cells
   (fig. S6, B to F). In these cells, the direct conversion of arginine to
   citrulline through NOS2 was also significantly reduced, as evidenced by
   the M6 isotopologue of citrulline (fig. S6F) and Nos2 mRNA expression
   (fig. S6G). We observed that the conversion of ornithine to citrulline
   in the mitochondria occurred in MEF cells, but there were no
   significant changes observed in the M5 isotopologue of citrulline (fig.
   S6F). These findings highlight that the deficiency of citrulline during
   aging can be attributed to the decreased expression of the anabolic
   enzyme NOS2.

Fig. 6. Age-dependent down-regulation of Nos2 contributed to citrulline
deficiency in macrophages.

   [139]Fig. 6.
   [140]Open in a new tab

   (A) Schematic illustration of stable-isotope tracing metabolomics using
   [U-^13C]-arginine as a tracer. Young and aged BMDMs (male mice) were
   treated with 0.4 mM [U-^13C]-arginine for 24 hours. OTC, ornithine
   transcarbamylase; ASL, argininosuccinate lyase. Black dots represent
   ^12C and red dots represent ^13C. (B to E) Relative abundances of
   metabolites in citrulline metabolism, including arginine (B), ornithine
   (C), arginosuccinate (D), and citrulline (E), in young and aged BMDMs
   (male mice) after stable-isotope tracing. n = 7 biologically
   independent samples in each group. n.d., not detected. (F) Real-time
   PCR analyses of Nos2 mRNA expressions in citrulline metabolism in young
   and aged BMDMs (male mice) after stable-isotope tracing. Bar graphs
   represent means ± SEM. n = 5. (G and H) Relative citrulline levels in
   the small intestine (G) and kidney (H) of young and aged mice (male)
   determined by LC-MS. n = 3 to 4. (I to K) Relative citrulline levels in
   plasma and PBMC samples from healthy humans: plasma (I), PBMC (J), and
   NOS2 expression from PBMCs (K) in young and aged human. The NOS2
   expressions were derived from the PBMC transcriptome of healthy human
   donors using RNA sequencing ([141]42). Bars represent means ± SEM.
   n = 7 to 9. Bars represent means ± SEM. ***P < 0.001, **P < 0.01, and
   *P < 0.05 (two-tailed Student’s t test).

   Furthermore, to investigate whether the anti-inflammatory effect of
   citrulline in macrophages relies on nitric oxide, we used L-NAME to
   reduce the production of nitric oxide induced by LPS treatment. Our
   data revealed that the increase in citrulline induced by LPS was
   diminished following L-NAME treatment (fig. S6H). These results
   confirmed the successful inhibition of NOS2 by L-NAME. In addition, the
   analysis of mRNA levels of proinflammatory cytokines, including Tnf,
   Il6, and Il1b, in response to LPS treatment indicated that inhibiting
   NOS2 with L-NAME significantly increased inflammation, which could be
   effectively alleviated by citrulline supplementation (fig. S6I). Hence,
   the anti-inflammatory effect of citrulline is not contingent on the
   generation of nitric oxide.

   To elucidate the age-related decline in citrulline levels in mice, we
   also assessed citrulline levels in the small intestine and kidney of
   mice. Note that citrulline was not present in the mouse diet used in
   our experiment and citrulline is primarily synthesized by intestinal
   epithelial cells within the mouse small intestine and subsequently
   released into the bloodstream. In addition, citrulline can undergo
   metabolism in the kidney to produce arginine. Our findings in [142]Fig.
   6 (G and H) illustrate that, compared to young mice (12 weeks old,
   male), citrulline levels in the small intestine of aged mice (78 weeks
   old, male) were significantly diminished. However, citrulline levels in
   the kidney had no changes during aging. Hence, we deduce that the
   declined citrulline levels in aged mice primarily stem from reduced
   biosynthesis within the small intestine. In addition, citrulline levels
   in primary BMDMs were observed to decrease with aging ([143]Figs. 2F
   and [144]6E), possibly due to the reduced citrulline levels in the
   bloodstream. Furthermore, we provided evidence demonstrating that Nos2,
   specifically expressed mainly in immune cells upon stimulation by
   proinflammatory cytokines, is significantly down-regulated during
   aging, thereby contributing to the decreased levels of citrulline in
   primary BMDMs ([145]Fig. 6F). Therefore, we think that both the
   decreased circulatory citrulline and impaired biosynthesis of
   citrulline jointly contribute to citrulline deficiency in aged mouse
   macrophages.

   We also expanded our investigation to include plasma and PBMC samples
   from healthy individuals, measuring citrulline levels using LC-MS. We
   found that citrulline levels significantly increased during aging in
   human plasma ([146]Fig. 6I), and the physiological concentration of
   citrulline in human plasma is ~40 μM (fig. S6J). In contrast,
   citrulline levels significantly decreased in PBMC samples from aged
   individuals ([147]Fig. 6J), and NOS2 mRNA expressions were also
   decreased in aged human PBMCs ([148]Fig. 6K) ([149]42), aligning with
   our observations in aged mouse macrophages. Notably, more than ~80% of
   citrulline released by the intestine is metabolized within the kidney
   to synthesize arginine ([150]43). Previous reports have indicated that
   the decline in renal function during aging leads to reduced citrulline
   excretion and citrulline accumulation in the body, ultimately resulting
   in up-regulated citrulline levels in the elderly population ([151]44).
   In summary, our findings indicate that citrulline levels decreased in
   PBMCs in aged humans, mirroring observations in aged mouse macrophages.

DISCUSSION

   Metabolic dysregulation and altered metabolites are widely recognized
   as key characteristics of aging ([152]2, [153]4). Numerous studies have
   investigated the roles of endogenous metabolites, such as NAD ([154]9,
   [155]10), taurine ([156]14), spermidine ([157]15), and others ([158]8,
   [159]45, [160]46), as drivers of the aging process. In this study, we
   conducted a comprehensive analysis of metabolic changes in multiple
   organs of mice at various ages. Our findings provide the first evidence
   linking citrulline deficiency to aging. We identified multiple
   antiaging effects of citrulline, including the reduction of cellular
   senescence, protection against DNA damage, prevention of cell cycle
   arrest, modulation of macrophage metabolism, and mitigation of
   inflammaging. Notably, long-term supplementation of citrulline in aged
   mice demonstrated significant benefits by alleviating age-associated
   phenotypes and increasing health span. These findings underscore the
   critical role of citrulline deficiency as a key driver of the aging
   process and highlight the potential therapeutic intervention of
   citrulline supplementation to counteract age-related diseases.

   To explore the biological mechanism by which citrulline counteracts
   aging, we demonstrated that citrulline acts as an endogenous metabolite
   antagonist to inflammation. Macrophages are primary contributors to
   age-associated inflammation. Our findings unveil that the decline in
   endogenous citrulline levels impairs the anti-inflammatory function of
   macrophages, thereby enhancing susceptibility to inflammatory responses
   during aging. The anti-inflammatory effect of citrulline has been
   validated in various mouse models, and its efficacy remains intact even
   in the context of aging. This observation suggests that the
   age-dependent deficiency of citrulline, acting as an endogenous
   antagonist to inflammation, triggers inflammaging and accelerates the
   aging process. In-depth mechanistic investigations have revealed that
   citrulline supplementation rescues age-associated metabolic alterations
   in macrophage metabolism. Specifically, we have demonstrated that
   citrulline modulates the inflammatory responses by regulating the
   activities of the mTOR-HIF1α-glycolysis signaling pathway in
   macrophages. Collectively, these results establish that citrulline
   governs macrophage metabolism and inflammation as a means to counteract
   aging.

   Our study also underscores the remarkable potential of citrulline as an
   endogenous metabolite inhibitor of the mTOR pathway in the context of
   inflammation and aging. mTOR serves as a nutrient sensor that regulates
   cellular metabolism and is linked to cell proliferation, growth, and
   survival ([161]47, [162]48). Extensive research has established the
   mTOR pathway as a negative regulator of lifespan and aging ([163]49,
   [164]50). Pharmacological inhibition of mTOR using small-molecule
   compounds such as rapamycin has been shown to effectively extend
   longevity in various animal models, including Caenorhabditis elegans
   ([165]51), Drosophila melanogaster ([166]52), and Mus musculus
   ([167]53, [168]54). However, the discovery of an endogenous metabolite
   that inhibits mTOR to counter aging has remained elusive. Arginine,
   leucine, and S-adenosylmethionine (SAM; downstream metabolite of
   methionine) are the known metabolites that are directly sensed by mTOR
   components [CASTOR ([169]55, [170]56), Sestrin ([171]57, [172]58), and
   SAMTOR ([173]59), respectively]. Restriction of methionine or the three
   branched-chain amino acids—leucine, isoleucine, and valine—extends
   lifespan in mice ([174]60, [175]61), but the roles of these metabolites
   in regulating aging and mTOR are complicated. Our metabolomics data
   confirmed that arginine, leucine, SAM, and methionine levels remained
   unchanged during aging in mice. In our study, we extensively
   demonstrated that citrulline inhibits the activation of the mTOR
   pathway in macrophages in both inflammatory and aging contexts. This
   finding highlights citrulline as a promising endogenous metabolite with
   the potential to inhibit mTOR signaling. While further mechanistic
   exploration is needed to fully comprehend how citrulline regulates mTOR
   activation, our findings present an exciting avenue for investigating
   the interplay between metabolites, mTOR, and their associations with
   aging and longevity. In our study, we found that in the context of an
   inflammatory response, citrulline prominently inhibits the activation
   of the mTOR pathway. It has been reported that citrulline triggers the
   mTOR pathway in muscle cells and citrulline supplementation is now
   under investigation for the treatment of sarcopenia ([176]62). However,
   note that sarcopenia of muscle tissue occurs within a malnutrition
   context rather than an inflammatory environment. Under these differing
   physiological circumstances, citrulline may exhibit contrasting
   effects.

   Citrulline, a nonessential amino acid, is naturally synthesized in the
   liver and intestines of mammals ([177]63). It is also present in
   various foods, including watermelon, cucumber, pumpkin, melons,
   squashes, and gourds ([178]64). Citrulline plays a crucial role in the
   urea cycle. Its supplementation increases plasma levels of arginine,
   improving the ammonia recycling process and nitric oxide metabolism
   ([179]65). As a result, citrulline supplementation may be beneficial in
   situations where nitric oxide is relevant, such as enhancing athletic
   performance ([180]66), promoting vascular health ([181]67), and
   addressing erectile dysfunction ([182]68). In our study, we have
   revealed a function of citrulline, independent of nitric oxide, in
   mediating mTOR activation and acting as an endogenous metabolite
   antagonist to inflammation and aging. Therefore, citrulline
   supplementation holds potential therapeutic value in counteracting
   age-related diseases. In terms of clinical applications, citrulline
   supplementation has demonstrated potential benefits in relation to
   inflammation. Citrulline supplementation decreased inflammatory
   C-reactive protein and IL-6 in serum of patients from critically ill
   patients ([183]69). In addition, citrulline supplementation also
   improved glycemic status, TNF-α, and high-sensitivity C-reactive
   protein in patients with type 2 diabetes ([184]70). These studies,
   combined with our findings, suggest that citrulline supplementation may
   serve as a promising therapeutic intervention for age-related
   pathologies.

   In our study, we think both the decreased circulatory citrulline and
   impaired biosynthesis of citrulline jointly contribute to macrophage
   inflammatory profile. Under inflammatory conditions, reduced citrulline
   levels in the blood and decreased endogenous citrulline synthesis both
   lead to diminished inflammation suppression, resulting in a
   proinflammatory phenotype. Under physiological conditions, macrophages
   maintain intracellular citrulline levels through their own endogenous
   biosynthesis. Therefore, a reduction in circulating citrulline levels
   may not significantly affect the anti-inflammatory function of
   macrophages. During aging, the down-regulation of citrulline levels due
   to reduced endogenous synthesis, driven by lower expression of the
   enzyme NOS2, impairs the anti-inflammatory function of macrophages.
   Thus, the decline in citrulline levels from endogenous synthesis
   disrupts normal macrophage function. We propose that the synthesis of
   intracellular endogenous citrulline is more crucial for the
   anti-inflammatory function of macrophages under physiological
   conditions.

   In our study, we observed a significant increase in the citrulline
   level in macrophages upon LPS stimulation, contrasting the findings of
   Mao et al. ([185]71). To validate our results, we conducted additional
   experiments under various conditions (fig. S4, A to G) and conducted a
   thorough review of existing literature ([186]72–[187]74). Our findings
   align with numerous other publications consistently reporting similar
   conclusions, all of which demonstrate a significant increase in
   citrulline levels upon LPS stimulation. One of the limitations of our
   study is exclusive use of male mice as the model to screen age-related
   metabolites, overlooking the potential contribution of female mice.
   This decision was primarily driven by the need to minimize the inherent
   variability associated with cyclical reproductive hormones in female
   mice. However, focusing solely on male mice, we may have missed
   valuable insights into how sex-specific factors, such as reproductive
   hormones, could influence the observed metabolic changes. This
   limitation underscores the importance of future research endeavors to
   investigate the metabolic profiles in both male and female mice,
   allowing for a more comprehensive understanding of age-related
   metabolic alterations across sexes.

MATERIALS AND METHODS

Cell culture

   MEF, iBMDM, BV2, and RAW264.7 cells were cultured in Dulbecco’s
   modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS)
   and antibiotics [penicillin (100 U/ml) and streptomycin (100 μg/ml)].
   Mouse BMDMs were isolated from C57BL/6J strain of young and aged mice
   (6 and 78 weeks old, male) and maintained in 70% fresh DMEM
   [supplemented with 10% FBS, penicillin (100 U/ml), and streptomycin
   (100 μg/ml)] and 30% L929 cell culture medium. After 6 days of culture,
   adherent cells were harvested and plated into new wells for further
   experiments. PBMCs were isolated from healthy individuals and
   maintained in fresh RPMI 1640 [supplemented with 10% FBS, penicillin
   (100 U/ml), and streptomycin (100 μg/ml)] with macrophage
   colony-stimulating factor (50 ng/ml). After 7 days of culture, adherent
   cells were harvested and plated into new wells for further experiments.

   All cells were cultured at 37°C with 5% CO[2]. The iBMDMs exhibit
   comparable characteristics and functions to primary BMDMs, rendering
   them convenient for cell biology research.

Western blotting analysis

   Protein samples were isolated from respective cells by lysis in
   radioimmunoprecipitation assay buffer (#89900, Thermo Fisher
   Scientific). The protein samples were then denatured by boiling in SDS
   loading buffer (5×). Equal amounts of proteins were resolved by
   SDS–polyacrylamide gel electrophoresis and then transferred onto a
   polyvinylidene fluoride membrane. The blots were blocked in 5% bovine
   serum albumin for 1 hour at room temperature and incubated overnight at
   4°C with the respective primary antibodies. The following primary
   antibodies were used: mouse β-actin (Proteintech, #66009-1-Ig), rabbit
   p-S6(Ser^240/244) [Cell Signaling Technology (CST), #5364S], rabbit
   HIF-1α (CST, #36169T), rabbit p16 (Abcam, #ab211542), rabbit p21
   (Abcam, ab188224), mouse Iba1 (Abcam, #ab283342), mouse γH2AX (CST,
   #80312), rabbit p-mTOR(Ser^2448) (Proteintech, #67778-1-lg), mouse mTOR
   (Proteintech, #66888-1-lg), rabbit 4EBP1 (Proteintech, #60246-1-lg),
   rabbit p-4EBP1(Thr^37/46) (CST, #2855T), rabbit p70(S6K) (Proteintech,
   #14485-1-AP), rabbit p-p70(S6K)(Thr^389) (CST, #9234T), rabbit TNFα
   (Proteintech, #17590-1-AP), and rabbit IL-1β (CST, 12242S). Horseradish
   peroxidase (HRP)–conjugated secondary antibodies, including goat
   HRP-goat anti-rabbit immunoglobulin G (GenScript, A00098) and goat
   HRP-goat anti-mouse immunoglobulin G (GenScript, A00160), were
   incubated with blot membrane at room temperature for 1 hour, followed
   by detecting with ECL Western Blotting Detection Reagent (Thermo Fisher
   Scientific, 32132).

Mice

   Mice (C57BL/6) were group housed in a barrier facility at room
   temperatures of 20° to 26°C with 40 to 70% humidity and 12-hour
   light/12-hour dark cycles. All mice received a regular chow diet ad
   libitum. All animal care and experiments were conducted in accordance
   with the guidelines and approval of the Institutional Animal Care and
   Use Committees of Interdisciplinary Research Center on Biology and
   Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of
   Sciences. The number of experimental protocols approved by the ethical
   committee of IRCBC (Interdisciplinary Research Center of Biology and
   Chemistry) is ECSIOC_2023-14. Mice matched for age and sex were used
   and randomly assigned to experimental groups. Aged mice (6, 24, 36, 52,
   and 78 weeks old, male) were obtained from animal facility, College of
   Life Sciences, Xiamen University. Mice were first anesthetized with
   isoflurane and sacrificed. Then, mouse tissues were dissected and
   immediately frozen using liquid nitrogen.

Long-term citrulline supplementation

   Citrulline solutions were prepared in phosphate-buffered saline (PBS)
   and sterilized by filtration through a 0.22-μm filter (Millipore). Mice
   (6 and 72 weeks old, male, C57BL/6) received the drink water with or
   without citrulline (1 g/kg of body weight) for ~9 weeks. To enhance
   citrulline supplementation, mice were also orally delivered with a dose
   of water or water with citrulline (1.5 g/kg of body weight) once a
   week. Citrulline was purchased from Sigma-Aldrich (C7629). After
   acclimatization for 1 week, the mice were weighed twice a week and
   euthanized on day 66. Liver and spleen tissues were weighed. BMDMs were
   isolated from hind legs. Left brain tissues were harvested for
   immunohistochemistry, and right brain and liver tissues were used for
   metabolomics.

LPS challenge experiment

   Mice (8 weeks old, female, C57BL/6) were fed with water or water with
   citrulline (1 g/kg of body weight) for 7 days. To enhance citrulline
   uptake, these mice were also orally delivered with a dose of water or
   water with citrulline (1.5 g/kg of body weight) once a day during days
   3 to 7. All mice were exposed to LPS challenge on day 7. In the acute
   liver injury model, mice were treated with LPS (10 μg/kg of body
   weight) and GalN (500 mg/kg of body weight) through intraperitoneal
   injection on day 7. In the sepsis model, mice were treated with a
   lethal dose of LPS (35 mg/kg of body weight) through intraperitoneal
   injection on day 7. In the third model, mice were treated with a
   nonlethal dose of LPS (20 mg/kg of body weight) through intraperitoneal
   injection on day 7. For the survival experiment, mouse survival was
   observed and recorded every half hour. For low-dose LPS challenge
   experiment, after 6 hours of LPS treatment, mice were first
   anesthetized with isoflurane and euthanized. Then, mouse tissues were
   dissected and immediately frozen using liquid nitrogen.

Immunohistochemistry

   Serial brain tissue sections (30 μm) were used for immunohistochemical
   staining with the following modifications. Briefly, the sections were
   blocked in tris-buffered saline combined with 1% Triton X-100 and 10%
   donkey serum for 2 hours. All primary antibodies were used and
   incubated for 2 days at 4°C. Secondary antibodies were used and
   incubated for 4 hours at room temperature. Fluorescently stained
   sections were counterstained with 4′,6-diamidino-2-phenylindole (DAPI;
   200 ng/ml; Sigma-Aldrich) for 2 to 5 min and then coverslipped with
   Gel/Mount (Biomeda).

Cell imaging and qualification

   The DNA Damage Assay Kit by γH2AX immunofluorescence (Beyotime
   Biotechnology, #C2035S) was used to detect the phosphorylation of H2AX
   following the manufacturer’s instructions. Cells were photographed at
   ×20 magnification using Leica microscope (Germany). Image data
   qualification of γH2AX was determined by ImageJ for DAPI-labeled nuclei
   and proteins.

Enzyme-linked immunosorbent assays

   Proinflammatory cytokine secretions were quantified using mouse TNFα
   enzyme-linked immunosorbent assay (ELISA) kit (Proteintech, #EK1222),
   mouse IL-1β ELISA kit (Proteintech, #EK10002), and mouse IL-6 ELISA kit
   (Proteintech, #EK10007) according to the manufacturer’s instructions.

Induction of senescent immortalized MEFs

   We followed the previously reported protocol to prepare senescent
   immortalized MEFs (iMEFs) ([188]34). In detail, iMEFs were subjected to
   treatment with tBHP at various concentrations for 2 hours, followed by
   recovery in complete cell culture medium for 3 days. Subsequently,
   these cells underwent passaging for 4 days to induce cellular
   senescence.

SA-β-Gal assay

   SA-β-Gal activity was measured using SA-β-Gal staining kit (CST,
   #9860). Briefly, cells were washed with PBS and fixed for 10 min at
   room temperature in a fixing solution of the SA-β-Gal kit. After
   washing twice with PBS, cells were incubated in senescence detection
   solution overnight at 37°C. Subsequently, cells were photographed at
   ×10 magnification using Olympus CKX53 optical light microscope (Tokyo,
   Japan). The percentage of SA-β-Gal-positive cells was determined by
   counting blue-stained cells from three fields of each sample.

Metabolites extraction

   For extraction of cellular metabolites, the cell culture medium was
   quickly removed, and the cells were washed with PBS twice. The cell
   dishes were placed on dry ice, and the metabolite extraction solution
   (acetonitrile/methanol/water = 2:2:1, v/v/v; 1 ml) was added to the
   dishes to quench the metabolism. The extraction solution was precooled
   at −80°C for 1 hour before the extraction. The plates were then
   incubated at −80°C for at least 40 min. The cell contents were scraped
   and transferred to a 1.5-ml Eppendorf tube. The samples were vortexed
   for 1 min and centrifuged for 10 min at 17,000g and 4°C to precipitate
   the insoluble materials. The supernatant was taken to a new 1.5-ml
   Eppendorf tube and evaporated to dryness at 4°C using a vacuum
   concentrator. The dried extracts were kept in −80°C until LC-MS
   analysis. Before the LC-MS analysis, the dried extracts were
   reconstituted in 100 μl of acetonitrile/H[2]O (1:1, v/v), sonicated for
   10 min, and centrifuged for 15 min at 17,000g and 4°C to remove
   insoluble debris. The supernatant was then transferred to
   high-performance liquid chromatography (HPLC) vials for LC-MS analysis.

   For metabolite extraction from tissues, each tissue sample was
   homogenized by grinding on liquid nitrogen and aliquoted every ~20 mg.
   Two hundred microliters of water was added into 20 mg of tissue sample
   for further homogenization. Homogenized tissue solution (200 μl) was
   extracted using 800 μl of acetonitrile/methanol (1:1, v/v). The samples
   were vortexed for 30 s and sonicated for 10 min in an ice bath. To
   precipitate proteins, samples were incubated for 1 hour at −20°C,
   followed by 15 min of centrifugation at 17,000g and 4°C. The dried
   extracts were kept in −80°C as previously described.

   For metabolite extraction from serum, 50 μl of serum sample was
   extracted using 200 μl acetonitrile/methanol (1:1, v/v). The samples
   were vortexed for 30 s and sonicated 10 min in an ice bath. To
   precipitate proteins, the samples were incubated for 1 hour at −20°C,
   followed by 15 min of centrifugation at 17,500g and 4°C. The dried
   extracts were kept in −80°C as previously described.

   For sample normalization, the protein precipitate was reconstituted
   with 200 μl of lysis buffer [100 mM tris-HCl with 4% (w/v) SDS (pH
   7.6)], vortexed for 30 s, and sonicated until the protein solution was
   generated. Bicinchoninic acid kit was used to determine the protein
   concentrations in each sample and used as a reference to adjust the
   volumes of reconstitution buffer for sample reconstitution. Before the
   LC-MS analysis, the dried extracts were reconstituted with
   acetonitrile/H[2]O (1:1, v/v), sonicated for 10 min, and centrifuged
   for 15 min at 17,000g and 4°C to remove insoluble debris. The
   supernatant was then transferred to HPLC vials for LC-MS analysis.

Stable-isotope tracing metabolomics

   MEF cells or primary BMDM cells were plated in 6-cm dishes at 2,000,000
   cells per dish and cultured in arginine-free DMEM containing dialyzed
   FBS (10%), penicillin/streptomycin (1%), and 400 μM [U-^13C]-arginine.
   For MEFs, cells were first grown to 75 to 80% confluence in the log
   phase in cell culture plates with normal culture medium. Then, the
   culture medium was changed to fresh medium with 400 μM
   [U-^13C]-arginine for 24 hours. Fast quench-based extraction of
   cellular metabolites was performed as previously described. For primary
   BMDMs, on day 6 of culture, adherent BMDM cells were counted and plated
   into new wells containing fresh medium with 400 μM [U-^13C]-arginine
   for 24 hours. After that, fast quench-based extraction of cellular
   metabolites was also performed as previously described.

LC-MS analysis

   The untargeted metabolomics analyses of aging mouse brain tissues,
   liver tissues, and serum followed our previous publication ([189]75).
   Briefly, an ultra-HPLC (UHPLC) system (1290 series, Agilent
   Technologies, USA) coupled to a quadruple time-of-flight mass
   spectrometer (TripleTOF 6600, SCIEX, USA) was used for LC-MS analyses.
   A Waters Ethylene Bridged Hybrid (BEH) amide column was used for LC
   separation. Mobile phases, linear gradient elution, and electrospray
   ionization source parameters followed the previous publication
   ([190]75). Other metabolomics analyses followed an updated method in
   our recent publication ([191]76). Briefly, LC-MS analyses were
   performed using a Vanquish UHPLC coupled to Orbitrap Exploris 480
   (Thermo Fisher Scientific, USA). The raw data were acquired using
   Xcalibur (version 4.4.16.14). A Waters BEH amide column and Phenomenex
   Kinetex C18 column were used for LC separation for hydrophilic
   interaction liquid chromatography (HILIC) mode and reversed phase
   liquid chromatography (RPLC) mode. Mobile phases, linear gradient
   elution, and electrospray ionization source parameters followed the
   previous publication ([192]76). Metabolomics data analysis and
   metabolite annotation were performed using MetDNA
   ([193]http://metdna.zhulab.cn/) ([194]75, [195]77). Metabolite
   annotations with confidence level 1 were reported in our study, which
   were obtained through matching of MS1, retention time (RT), and MS/MS
   spectra with the experimental metabolite spectral library. Metabolic
   pathway enrichment analysis was performed via hypergeometric test and
   visualized in R (v 4.3.0). The pathway database was Kyoto Encyclopedia
   of Genes and Genomes ([196]www.genome.jp/kegg/). Stable-isotope tracing
   metabolomics data analysis was performed using MetTracer ([197]76).

Measurement of cellular ECAR and OCR

   Cellular ECAR and OCR values were measured using the XFe96
   Extracellular Flux Analyzer (Agilent Technologies, USA) following the
   manufacturer’s protocols. Briefly, iBMDMs were seeded in XFe96-well
   microplate with 1.0 × 10^4 cells per well, pretreated with citrulline
   (500 μM) for 0.5 hours, and followed by LPS treatment (100 ng/ml) for 4
   hours before XF assay. ECAR was measured in XF medium under basal
   conditions. Glucose (10 mM), oligomycin (5 μM), and 2-deoxyglucose
   (2-DG; 100 mM; glycolysis inhibitor) were sequentially added to measure
   glycolysis, glycolysis capacity, and glycolysis reverse, respectively.
   OCR was measured in XF medium under basal conditions. Oligomycin (1.5
   μM), FCCP (carbonyl cyanide p-trifluoromethoxyphenylhydrazone; 0.5 μM),
   and rotenone/antimycin A (0.5 μM) were sequentially added to measure
   nonmitochondrial oxygen consumption, basal respiration, maximal
   respiration, proton leak, ATP production, and spare respiratory
   capacity, respectively.

   For normalization of the ECAR and OCR experiments, we counted the cells
   before seeding to ensure the same cell numbers in each well of a
   96-well plate (10,000 iBMDM cells). The reported ECAR values were
   normalized to the cell numbers. The Seahorse measurements were all
   performed with the following assay conditions: 3 min of mixing, 3 min
   of waiting, and 3 min of measurements.

   Nonglycolytic acidification, maximum glycolytic capacity, and
   glycolytic reserve were calculated in the ECAR experiment. Below is the
   detailed information: nonglycolytic acidification = last rate
   measurement before glucose injection; maximum glycolytic capacity =
   (maximum rate measurement before oligomycin injection) − (last rate
   measurement before glucose injection); glycolytic capacity = (maximum
   rate measurement after oligomycin injection) − (last rate measurement
   before glucose injection). Nonmitochondrial oxygen consumption, basal
   respiration, maximal respiration, proton leak, ATP production, and
   spare respiratory capacity were calculated in the OCR experiments.
   Below is the detailed information: nonmitochondrial oxygen
   consumption = minimum rate measurement after rotenone/antimycin A
   injection; basal respiration = (last rate measurement before first
   injection) − (nonmitochondrial respiration rate); maximal
   respiration = (maximum rate measurement after FCCP injection) −
   (nonmitochondrial respiration); proton leak = (minimum rate measurement
   after oligomycin injection) − (nonmitochondrial respiration); ATP
   production = (last rate measurement before oligomycin injection) −
   (minimum rate measurement after oligomycin injection); spare
   respiratory capacity = (maximal respiration) − (basal respiration). The
   results were analyzed and exported from software Seahorse Wave (version
   2.4.0.60).

Real-time polymerase chain reaction

   RNA was isolated using TRIzol reagent (Thermo Fisher Scientific, USA)
   in accordance with manufacturer’s instructions. RNA was resuspended in
   diethyl pyrocarbonate–treated ribonuclease-free water (Thermo Fisher
   Scientific, USA). Reverse transcription was catalyzed using the
   SuperScript III First-Strand Synthesis System (Thermo Fisher
   Scientific, USA). Real-time polymerase chain reaction (PCR) analysis
   was performed using the QuantStudio 6 Flex real-time PCR system with
   SYBR Select Master Mix (Thermo Fisher Scientific, USA). We extensively
   used the 2^−ΔΔCT method as a relative quantification strategy for
   quantitative real-time PCR data ([198]78). The relative abundance of
   RNA were calculated based on the 2^−ΔΔCT. ΔCT = CT (target) − CT
   (reference, β-actin), ΔΔCT = CT (case) − CT (control), relative
   abundance of RNA = 2^−ΔΔCT.

PBMC collection

   PBMC collection has received ethical approval from the ethical
   committee of Shanghai Yangpu Hospital, Tongji University School of
   Medicine with the approval number LL-2022-LW-016. In EDTA-K
   anticoagulant tubes, 3 ml of whole blood was centrifuged at 4000 rpm
   for 10 min, yielding the separation of three layers: yellow plasma,
   white blood cells/platelets, and red blood cells. Plasma was stored at
   −80°C. Next, 3 ml of PBS was added to the tubes to dilute the blood
   sample. In a separate tube, 3 ml of Ficoll-Paque reagent (Cytiva,
   #17544203-1) was equilibrated to room temperature. The diluted blood
   was carefully layered onto the Ficoll-Paque, avoiding mixing the two
   solutions, and then centrifuged at 400g for 30 min at 25°C, resulting
   in four layers: diluted plasma, PBMC cell layer, white separating
   liquid, and red blood cells. The diluted plasma was discarded, and PBMC
   cell layer was transferred to a new tube. PBMCs were washed twice with
   3 ml of PBS. Then, PBMC pellet was washed with 1 ml of PBS, transferred
   to 2-ml Eppendorf tube, and stored at −80°C.

Statistical analysis

   All quantifications of fluorescence images were represented as
   means ± SEM from three replicates. Quantifications of immunoblots and
   immunostaining were performed using ImageJ (version 1.8.0). All
   statistical analyses were performed using GraphPad Prism (v 8.0 and v
   9.0), Microsoft Excel 2010, and R (version 3.4.3). All Western blots
   from cell samples and mouse tissues were repeated at least three times
   with similar results. Differences were considered to be significant if
   P < 0.05.

Acknowledgments