All animal protocols were approved by the Institutional Animal Care and Use Committee at Massachusetts General Hospital (Protocols #2009N000239 an #2013N000115). All methods were performed in accordance with relevant regulations. Females of C57BL/6J (Stock #007914) mice, both 2-3-month-old (Young) and 18-month-old (Old), were purchased from Jackson Laboratory (Bar Harbor, ME). 18-month-old mice represent approximately 56 years of age in humans (48,49). For generation of dual reporter mice in which enteric neurons express tdTomato and enteric glia/progenitors express GFP, Plp1^GFP mice (50) were kindly gifted by Dr. Wendy Macklin, University of Colorado, Denver. To obtain Plp1^GFP;Actl6b::Cre; ROSA26-tdTomato (annotated as Plp1^GFP; Baf53b-tdT) mice, Actl6b::Cre mice (Stock #027826) were crossed with Plp1^GFP mice, and their offspring were crossed with ROSA26-tdTomato mice (Stock #007914) (51). In order to isolate longitudinal muscle layer with myenteric plexus (LMMP) and enteric neural cells animals were euthanized by carbon dioxide overdose, which is displacement of chamber air with compressed carbon dioxide at 30-70% per min. Death of animals was confirmed by lack of chest movement and heartbeat, after which cervical dislocation was performed. Both male and female mice were used for the in vitro studies as the influence of factors such as gut sensation and reproductive hormons can be excluded.

AGE was prepared from cloves of garlic (Allium sativum L.) through a process of rinsing with purified water, slicing, soaking in ethanol 20-50% (v/v), and extracting/aging for more than 10 months (52). AGE powder was obtained from Wakunaga Pharmaceutical Co. ltd and sent to Bio-Serv (Flemington, NJ) to prepare 3% AGE-formulated diet using standard mouse chow (AIN-93G, Bio-Serv, NJ). The AIN-93G the purified rodent diet that containing 200 g Casein, 100 g Sucrose, 397.5 g Cornstarch, 132 g Dyetrose, 3 g L-cystine, 50 g Cellulose, 70 g Soybean oil, 0.014 g t-Butylhydroquinone, 35 g Mineral mix, 10 g Vitamin mix, and 2.5 g Choline bitartrate per kg of diet (53). The 3% AGE-formulated diet was given to female mice for 2 weeks after the mice were randomized based on body weight. Two weeks after the initiation of feeding the AGE-formulated diet, in vivo assays and analysis of oxidative stress in myenteric plexus were performed. For in vitro and ex vivo studies, AGE powder was dissolved in sterile PBS and the solution was passed through with a 0.22 µM filter.

Total gastrointestinal transit time was measured as previously described (54). Mice (n=6, each group) were acclimatized for 30 min individually in cages without bedding, and 0.15 ml of 6% (w/v) carmine red dye (Sigma, C1022) in 0.5% (w/v) methylcellulose was administered to each mouse by oral gavage. The time from gavage to the appearance of the first red pellet was recorded as total gastrointestinal transit time. Maximum observation time was 6 h.

Mice (n=6, each group) were fasted overnight before the test was performed as described previously (55). Mice were acclimatized for 30 min, then a 3-mm glass bead (Sigma, #1040150500) was inserted into the rectum of each mouse using a silicone pusher under anesthesia by isoflurane (Covetrus, #11695-6777-2). Isoflurane, at 3 and 2%, was used for induction and maintenance, respectively. After bead insertion, mice were placed in individual plastic cages. The time to evacuate the bead started after the mice recovered from the anesthesia.

Fecal pellet output was measured as reported previously (56) in individual metabolic cages for 24 h (n=6, each group). The weight of food consumed over 24 h was obtained. Pellet number was calculated using the average weights of dry feces (n=25) per mouse.

The previously described method was used (57). Each mouse (n=6, each group) was placed in a plastic cage individually and wet fecal pellets were collected for 2 h. The wet fecal pellets were dried in the oven at 60˚C for 48 h. Fecal water content was calculated according to the following equation: (wet fecal weight-dry fecal weight)/wet fecal weight x100.

The organ bath experiments with colon rings followed well-established protocols. We conducted pilot studies to optimize parameters such as intensity (40-50 volts), frequency (5 Hz), and pulse duration (0.3 ms) to ensure reproducible and physiologically relevant responses for assessing smooth muscle contractility. These parameters have been validated in our laboratory, and experiments were performed using sparameters previously described (58-61). Freshly excised segments of distal colon were immediately placed in oxygenated Krebs solution (118 mmol/l NaCl, 4.7 mmol/l KCl, 1.2 mmol/l MgSO₄·7H₂O, 1.2 mmol/l KH₂PO₄, 25 mmol/l NaHCO₃, 11.7 mmol/l glucose, and 1.25 mmol/l CaCl₂) at 37˚C. Tissue rings, approximately 5 mm in length, were mounted between two metal hooks attached to force displacement transducers in a muscle strip myograph bath (Model 820 MS; Danish Myo Technology, Aarhus, Denmark) containing 7 ml of oxygenated Krebs solution. The rings were gently stretched to establish a basal tension of 0.5 g and allowed to equilibrate for 30-45 min, with Krebs solution being replaced every 20 min. Spontaneous contractions were recorded in both the absence and presence of AGE (1% w/v). Afterwards, contractions were recorded again following the addition of the nitric oxide synthase inhibitor L-NAME (100 mM; Sigma-Aldrich, St. Louis, MO). Electrical field stimulation (EFS) was then applied to the tissue using a pulse train of 40-50 V (15-sec duration, 300 µs pulse width, 5 Hz frequency) via a CS4+ constant voltage stimulator controlled by MyoPulse software (Danish Myo Technology). The procedure was repeated after the addition of AGE (1% w/v). Force contraction data from the circular smooth muscle were recorded and analyzed using a Power Lab 16/35 data acquisition system and LabChart Pro Software v8.1.16 (ADInstruments, NSW, Australia). Tissue viability was confirmed by assessing the contraction response to 60 mM KCl at the conclusion of the experiment. Baseline spontaneous activity was quantified by measuring the area under the curve (AUC), from 60 sec of data collected 5 min before the addition of AGE or L-NAME This was compared to the response following AGE or L-NAME, which was similarly quantified by measuring the AUC, from 60 sec of data collected immediately after the addition of AGE or L-NAME. Baseline maximum values were determined by averaging 60 sec of data recorded 1 min before EFS application. Changes in contraction were measured from the first 60 sec after stimulation onset and expressed as absolute differences from baseline. EFS was applied three times at 5-min intervals, and the maximum response was calculated as the mean of the three trials.

Enteric neural cells, including enteric neural stem cells (ENSCs), were isolated from mice as previously reported (62-64). Briefly, LMMP was separated from large intestine of mice (young, old, and Plp1^GFP; Baf53b-tdT). Enzymatic dissociation was achieved using dispase (250 µg/ml, STEMCELL Technologies, Vancouver, Canada) and collagenase XI (1 mg/ml, Sigma Aldrich, St. Louis, Missouri) at 37˚C for 45 min. Single cells were isolated by filtration through a 40-µm filter and plated at 5x10⁵ cells/ml in a 25-cm² flask in a 1:1 mixture of DMEM (Thermo Fisher Scientific) and NeuroCult Mouse Basal Medium (StemCell Technologies) supplemented with 1% penicillin/streptomycin (Gibco, #15140122), 20 ng/ml insulin growth factor (StemCell Technologies), and 20 ng/ml basic fibroblast growth factor (StemCell Technologies), 2% B27 supplement (gibco), 1% N2 supplement (gibco), 50 µM beta-mercaptoethanol (gibco), and 75 ng/ml retinoic acid (Sigma Aldrich). After 7 days in culture, primary neurospheres were obtained.

Primary neurospheres from wildtype C57BL/6J (n=2, male and female) and PLP1^GFP; BAF53b-tdT mice (n=2, male and female) were dissociated by Accutase (StemCell technologies, # 7920). 5,000 cells/well were plated into 96-well plate (CORNING, #3474) and secondary neurospheres treated with AGE at 0.25 to 1 mg/ml in culture media for 7 days (n=3, each group). The samples were dissociated with dispase and collagenase XI to generate single cell suspension and fixed with 4% PFA for 15 min. Random images of secondary neurospheres from wildtype C57BL/6J were taken using a Keyence BZX-700 All-In-One Microscope (Keyence America Itasca, IL) and the number of neurospheres from wildtype C57BL/6J quantified by ImageJ software (NIH). A cell viability assay was performed using the CellTiter-Glo Luminescent Cell Viability Assay (Promega, Madison, WI, USA, PAG7570). Secondary neurospheres from Plp1^GFP; Baf53b-tdT mice were dissociated to single cells and tdT+ and GFP+ positive cell numbers were counted using ImageJ software (NIH).

Primary neurospheres from wildtype C57BL/6J mice (n=2, male and female) were generated as described previously. AGE (1 mg/ml) was added to the media for secondary neurospheres in the presence of 10 µM of 5-ethynyl-2'-deoxyuridine (EdU) for 2 days (n=3, each group). After removing the media, secondary neurospheres were incubated in new AGE-containing media for 5 days. The samples were dissociated with dispase and collagenase XI to generate single cell suspension fixed with 4% PFA for 15 min and Click-iT EdU Cell Proliferation Kit for Imaging (Fisher Scientific, C10340) was performed. For immunohistochemical staining, 10% donkey serum and 1% Triton X-100 in phosphate-buffered saline (PBS) was used for blocking. Primary antibodies, including human anti-HuC/D (Anna1, 1:20, kindly gifted by Lennon lab) and rabbit anti-P75 (1;400, Promega, G3231), were incubated overnight at 4˚C, followed by secondary antibodies for 1 h. Secondary antibodies included anti-human IgG (1:200, Alexa Fluor 594, Jackson ImmunoResearch) and anti-rabbit IgG (1:200; Alexa Fluor 488, Invitrogen). Random Images were taken from each group using a Keyence BZX-700 All-In-One Microscope (Keyence America Itasca, IL) and the number of double-labelled Hu+EdU+ or P75+EdU+ cells counted using ImageJ software (NIH).

MitoSOX (Thermo Fisher Scientific, M36008) was used to identify mitochondrial-derived production of superoxide in the myenteric ganglia of the ENS (65). Fresh colonic LMMP preparations were collected from the mice (n=4, each group). The samples were incubated in Hanks' Balanced Salt Solution (HBSS, Thermo Fisher Scientific) containing 5 µM of MitoSOX at 37˚C for 30 min. Tissues were washed with PBS and fixed with 4% PFA overnight at 4˚C. Images were captured, converted into binary format, and area of fluorescence within ganglia was measured in arbitrary units using ImageJ software (NIH).

LMMPs were separated from colons dissected from 2 months old Plp1^GFP; Baf53b-tdT mice (n=2, male and female) and dissociated enzymatically using dispase (250 µg/ml, STEMCELL Technologies) and collagenase XI (1 mg/ml, Sigma-Aldrich) at 37˚C for 45 min. Counter filtration was performed using a 20-µm cell strainer (pluriSelect, #43-50020-01) as previously (66). Samples were centrifuged at 350 G for 5 min and resuspended in NeuroCult Mouse Basal Medium (StemCell Technologies) containing 10% FBS and 1% penicillin/streptomycin (Gibco, #15140122). Isolated enteric ganglia in the media were plated into fibronectin (Sigma-Aldrich, #F1141)-coated 48 well plates and cultured for 24 h. Pre-treatment of AGE without hydrogen peroxide (H₂O₂) was performed for 48 h. Then, PBS treatment as control, 100 µM of H₂O₂ (Sigma-Aldrich, H1009) alone, and co-treatment of AGE and 100 µM of H₂O₂ were carried out for 24 h (n=3, each group). The samples were fixed with 4% PFA for 15 min at RT. Random images were taken using a Keyence BZX-700 All-In-One Microscope (Keyence America Itasca, IL). The number of neurons and glial cells, and neurite length, were quantified using ImageJ software (NIH).

Data analysis was performed using GraphPad Prism v10 (GraphPad Software, Inc., San Diego, CA). Two-tailed t-tests were performed for pairwise comparisons. A one-way analysis of variance (ANOVA) was performed with a post hoc Dunnett's test and Tukey's test for multiple comparisons. For all analyses, P<0.05 was considered significant. All data are presented as mean ± SEM, unless otherwise stated.

To characterize gastrointestinal (GI) motility in old mice, we performed multiple in vivo functional assays in young (2-3-month-old) and old (18-month-old) mice (Fig. 1A). Both total GI transit time (Fig. 1B, 2.98±0.2 h in young vs. 5.01±0.1 h in old, P<0.0001) and rectal bead expulsion time (Fig. 1C, 2.78±0.2 min in young vs. 6.47±0.3 min in old, P<0.0001) were significantly delayed in old mice. Furthermore, significantly decreased fecal pellet output (Fig. 1D, 127±7.8 in young vs. 35.7±8.9 in old, P<0.0001) and fecal water content (Fig. 1E, 42.7±2.2% in young vs. 33.2±1.9% in old, P<0.01) were observed in old mice despite no significant difference in food intake (Fig. 1F, 3.10±0.2 g in young vs. 2.37±0.3 g in old, ns).

AGE-formulated diet was given to old mice for 2 weeks and the same GI functional analyses were performed (Fig. 1A). Interestingly, all GI functional parameters, including total GI transit time (Fig. 1B, 5.01±0.1 h in old vs. 4.29±0.2 h in old+AGE, P<0.05), rectal bead expulsion time (Fig. 1C, 6.47±0.3 min in old vs. 3.87±0.4 min in old+AGE, P<0.0001), fecal pellet output (Fig. 1D, 35.7±8.9 in old vs. 120±14 in old+AGE, P<0.001), and fecal water content (Fig. 1E, 33.2±1.9% in old vs. 43.2±1.0% in old+AGE, P<0.01) were significantly improved by the AGE-diet. Food intake was not significantly reduced by AGE-diet (Fig. 1F, 2.37±0.3 g in old vs. 2.87±0.4 g in old+AGE, ns). These findings suggest that AGE has the potential to restore aging-related GI dysmotility in mice.

To evaluate the effect of AGE on colonic smooth muscle contractility, we performed organ bath studies on colon from 18-month-old mice. The representative tracings of each group during the baseline recording (Fig. 2A) and in response to electrical field stimulation (EFS) (Fig. 2B) are shown. In the presence of AGE, both baseline contractile activity (Fig. 2A) and EFS induced responses (Fig. 2B) were reduced compared to non-AGE-treated tissues. Quantitative analysis, determined by measuring area under the curve (AUC) and the amplitude of EFS responses, were significantly reduced in the presence of AGE (Fig. 2C, 16.4±1.8 g.s in absence of AGE vs. 13.4±1.7 g.s in presence of AGE, P<0.05; Fig. 2D, 2.41±0.4 g in absence of AGE vs. 1.75±0.3 g in presence of AGE, P<0.05). These changes in response to AGE were attenuated in the presence of L-NAME (N(ω)-nitro-L-arginine methyl ester), an inhibitor of neuronal nitric oxide synthase (nNOS) (Fig. 2E, 20.0±5.1 g.s in absence of AGE vs. 19.1±4.8 g.s in presence of AGE, ns), suggesting that the effect of AGE on colonic contractility is mediated, at least in part, by nNOS.

Recent evidence has demonstrated active regeneration and remodeling of the ENS postnatally (67). To evaluate the effect of AGE on ENS homeostasis in old mice, we isolated enteric neuronal stem/progenitor cells (ENSCs) from the colon of young and old mice.

Interestingly, viability of ENSCs isolated from old mice was significantly reduced in comparison to young-derived ENSCs (Fig. 3A, 100±5.0% in young ENSC vs. 14.8±0.3% in old ENSC, P<0.0001), and the number of old-mouse-derived ENSCs was also significantly lower (Fig. S1A, 4.9±0.4 in young ENSC vs. 2.5±0.4 in old ENSC, P<0.05). Addition of AGE to the culture media increased cell viability in and the number of both young- and old-mouse-derived ENSCs (Figs. 3B and S1B). We isolated ENSCs from Plp1^GFP; Baf53b-tdT mice in which enteric neurons express tdTomato and enteric glia/progenitors express GFP (51), and cultured them in the presence or absence of AGE (Fig. 3D-I). Quantitative analysis demonstrated that AGE significantly increased the Plp1-GFP population in a dose dependent manner (Fig. 3F and G, yellow arrows; Fig. 3J). AGE also expanded the neuronal (Baf53b-tdT positive, Fig. 3D and E, white arrows; Fig. 3K) population, but only at the 1 mg/ml concentration (Fig. 3K, 7.41±0.8 in control vs. 12.3±0.6 in AGE 1 mg/ml, P<0.01).

We tested whether AGE activates proliferation in ENS cells using the thymidine analogue, EdU (5-Ethynyl-2'-deoxyuridine). Isolated ENSCs were cultured in the absence (Control) or presence of AGE (1 mg/ml) for 2 days in medium containing EdU. We replaced the culture medium on day 3 and maintained the culture for an additional 5 days. Immunofluorescent staining was performed using anti-Hu and anti-P75 antibodies (Fig. 4A-H) to label neurons and glia/ENS progenitors, respectively. AGE promoted cell proliferation, as shown by an increase in neurons that are double positive for EdU+/Hu+ (Fig. 4I, 1.87±0.2 in control vs. 5.87±0.5 in AGE 1 mg/ml, P<0.01) and an increase in EdU+/P75+ as glia/progenitors (Fig. 4J, 1.43±0.2 in control vs. 3.41±0.2 in AGE 1 mg/ml, P<0.01).

Neurons are susceptible to oxidative stress (68), which plays a role in aging-related enteric neuronal damage (69). Therefore, we evaluated reactive oxygen species (ROS) within the myenteric plexus of the muscular layer of mouse colon using MitoSOX labeling. The ROS visualized by MitoSOX in old mice (Fig. 5B and E) (Fig. 5E, 48.7±1.5 in young vs. 117±13 in old, P<0.001) was significantly more prominent compared to that in young mice (Fig. 5A and E), supporting the idea that aging increases oxidative stress in the ENS. Old mice fed an AGE-formulated diet for 2 weeks (Fig. 1A) demonstrated significantly reduced ROS in the myenteric plexus (Fig. 5C and E) (Fig. 5E, 117±13 in old vs. 74.7±6.1 in old+AGE, P<0.05). These findings suggest that AGE ameliorates oxidative stress in the ENS of old mice.

Several recent studies have shown that prevention of oxidative stress could be a novel therapeutic strategy for neurodegenerative disorders (70-72). Here, we tested whether AGE has neuroprotective effects on cultured ENS cells. Hydrogen peroxide (H₂O₂) is commonly used to induce neuronal damage via oxidative stress (73,74). We isolated ENS cells from Plp1^GFP; Baf53b-tdT mice and cultured them in the absence (Fig. 6A-C) or presence of H₂O₂ (Fig. 6D-F). In the presence of H₂O₂, there was a significant reduction in the number of tdT+ neurons (Fig. 6M, 37.3±3.2 in control vs. 5.00±0.6 in H₂O₂, P<0.0001) and GFP+ glia/progenitors (Fig. 6N, 156±9.4 in control vs. 63.2±4.6 in H₂O₂, P<0.0001). We also found a significant reduction in the length of neurites (Fig. 6O, 50.0±2.7 in control vs. 4.67±0.9 in H₂O₂, P<0.0001) and decreased cell viability (Fig. 6P, 100±11 in control vs. 46.3±7.9 in H₂O₂, P<0.01) in H₂O₂-treated ENS cells, confirming that oxidative stress elicits damage to ENS cells in culture. When AGE was added to the H₂O₂-treated ENS cells at 0.5 mg/ml (Fig. 6G-I) and 1 mg/ml (Fig. 6J-L), there was significant improvement in the survival and health of neurons and glia. These findings suggest that AGE possesses neuroprotective effects on ENS cells against oxidative stress.