Unveiling the dynamic processes of dietary advanced glycation end-products (dAGEs) in absorption, accumulation, and gut microbiota metabolism

Abstract

This study delves into the dynamics of dietary advanced glycation end-products (dAGEs) on host health and gut microbiota. Using ¹³C-labeled carboxymethyllysine (CML) bound casein, we identify bound AGEs as the primary entry route, in contrast to free AGEs dominating urinary excretion. Specifically, our results show that the kidneys accumulate 1.5 times more dAGEs than the liver. A high AGE (HA) diet prompts rapid gut microbiota changes, with an initial stress-induced mutation phase, evidenced by a 20% increase in Bacteroides and Parabacteroides within the first week, followed by stabilization. These bacteria emerge as potential dAGE-utilizing bacteria, influencing the microbiota composition. Concurrent metabolic shifts affect lipid and carbohydrate pathways, with lipid metabolism alterations persisting over time, impacting host metabolic homeostasis. This study illuminates the intricate interplay between dietary AGEs, gut microbiota, and host health, offering insights into the health consequences of short- and long-term HA dietary patterns.

---

Introduction

The global population is undergoing a profound demographic transformation characterized by progressive aging, while simultaneous shifts in lifestyle have ushered in an era of chronic metabolic diseases.¹ These interwoven trends have imposed a substantial socioeconomic burden on societies worldwide. Among the diverse array of factors implicated in these health challenges, attention has increasingly turned to the intricate role of diet-derived substances, particularly advanced glycation end products (AGEs), which have been extensively studied due to their potential impact on metabolic processes.^2,3 AGEs form through reactions with highly reactive precursors such as α-dicarbonyls.⁴ Despite the extensive research, the mechanisms underlying the detrimental effects of dietary AGEs (dAGEs) have remained a subject of controversy over the past decade.⁵ Some researchers argue that absorbed dAGEs promote inflammation or disrupt energy homeostasis, while others posit that dAGEs face obstacles in digestion and are rapidly metabolized and excreted.^6–8 Thus, the extent to which dAGEs contribute to health decline hinges on their rates of bioavailability, 24-hour clearance rates, and accumulation within various organs.

Previous studies on dAGEs’ bioavailability have primarily focused on immediate absorption processes, exploring absorbable AGE types, absorption rates, and mechanisms such as competition with amino acids, dipeptides, and tripeptides for transporters or free diffusion.^9,10 However, the biological effects of dAGEs require their accumulation to sufficient concentrations and prolonged residence times within the body.^8,11 Consequently, in addition to identifying the primary absorbed dAGE types and their absorption rates, it becomes imperative to investigate their retention rates in the cycle over time and the accumulation patterns of retained AGEs within key metabolic organs.

Prior research on AGE accumulation and its biological effects has often conflated free and bound forms.^12,13 Recent investigations have revealed that the portion absorbed in the free form is the pivotal contributor to the adverse effects of dAGEs on health.¹⁴ While absorbed dAGEs are not solely present as free AGEs, some peptide-bound AGEs can enter circulation and partially convert into the free form within the body.⁹ Hence, it becomes crucial to discern and quantitate the accumulation of free and bound forms of dAGEs within the body.

In recent years, evidence has shown that dAGEs significantly impact the body by influencing gut microbiota composition and metabolism.^15,16 Both long-term (≥6 months) and short-term (<2 months) studies have yielded similar results, indicating that dAGEs rapidly modulate gut microbiota.^17,18 However, differences in experimental protocols necessitate comparative analysis with controlled, time-graded interventions to determine the precise processes and minimum duration required for dAGEs to affect gut microbiota. This study aims to investigate the effects of dAGEs on human health by examining their absorption, accumulation, and impact on gut microbiota. We focus on Nε-(carboxymethyl)lysine (CML) as it is a prevalent marker in dietary sources and strongly associated with chronic diseases.^19,20 The insights from this research could help in developing dietary guidelines to reduce the adverse effects of dAGEs and promote healthy aging.

Materials and methods

Materials

Beta-casein (C8654) and ¹³C₆-glucose (389374) were obtained from Sigma-Aldrich (St Louis, USA). Nε-(1-Carboxymethyl)-L-lysine (sc-212438D) was obtained from TRC (Toronto, Canada). Chromatographic grade methanol (M116129) and acetonitrile (A119010) were purchased from Aladdin (Shanghai, China). All other chemicals used in this study were of analytical grade.

Protein glycation

Protein glycation was carried out using β-casein and glucose (or ¹³C₆-glucose) as described previously.¹⁴ The aim is to prepare ¹³C-labeled and unlabeled caseins for short- and long-term dAGE tracing, respectively. Briefly, β-casein and glucose (or ¹³C₆-glucose) aqueous solution were incubated at a molar ratio of 1 : 36 (2 equivalent glucose with protein glycation sites) for 60 min at 115 °C. After incubation, the glycated protein mixtures were dialyzed at 4 °C for 3 days using dialysis bags with a 3 kDa MW cutoff.

Animals and treatment

Male C57BL/KsJ mice were procured from SLAC Laboratory (Shanghai SLAC Laboratory Animal Co., Ltd, China) and were maintained in a controlled environment. Animal housing was under a controlled temperature (25 ± 1 °C), humidity (40–60%), a 12 h light/dark cycle; water was provided ad libitum; and the animals were fed a standard basal diet. The composition of the experimental diets is shown in ESI Table S1.† All procedures involving animal subjects were conducted in accordance with the relevant laws and institutional guidelines for the ethical treatment of animals. The study protocol was approved by the Laboratory Animal Committee (LAC) of South China University of Technology under protocol number AEC-2021045.

For the 24 h metabolic kinetics experiments, 30 mice were individually housed in wire-bottom cages without bedding, with a plastic receiver tray lined with plastic wrap placed beneath each cage. Prior to feeding, mice were subjected to a 12 h fasting period. Subsequently, they were provided with a specially-made AIN-93G semi-purified diet, where casein was replaced with ¹³C-labeled glycated casein (¹³C-HA diet) (Table S1†). The dietary regimen is summarized in Fig. 1c. Only labeled feed was given for 20 min during the 24 h observational period, while the mice had free access to water. It is noted as 0 h before feeding. At 0, 4, 8, 12, and 24 h after feeding, 6 mice were randomly selected and euthanized. Fecal and urine samples were collected from each mouse, followed by blood plasma, small intestinal contents, and cecum chyme, and stored at −80 °C until further analysis. The blood samples were centrifuged at 4 °C and 3500g for 10 min to collect plasma, which was stored at −80 °C until further analysis. To obtain small intestine contents, the entire small intestine was excised, and 1 mL of dd H₂O was gently injected into one end with a syringe. The contents were then gently pushed through to the other end and carefully collected and stored at −80 °C until further analysis, designated as small intestinal eluates.

Fig. 1 Dynamics of dAGE digestion over 24 h. (a) Schematic representation of the glycation of β-casein with ¹³C₆-glucose to form ¹³C-labeled AGEs. (b) Bar-graph plots illustrating the levels of ¹³C₂-CML in the HA diet. (c) Experimental design for the animal-feeding study. Eight-week-old male C57BL/KsJ mice were individually housed and fed ¹³C-labeled HA diets for 20 min after 12 h fasting, with sampling at 0, 4, 8, 12, and 24 h. (d) Dynamic changes in the levels of free and bound CML in plasma, small intestine eluates, cecal chyme, feces, and urine.

For subsequent 28 d animal experiments, 50 mice (7 weeks old) were randomly assigned to two dietary groups: the low AGE (LA) diet group, consisting of an ordinary sugar-free semi-purified diet as previously described by Pai et al.,²¹ processed with limited heat exposure, and the high AGE (HA) diet group, which had a nearly identical composition to the LA diet, except that the casein in the chow was replaced with unlabeled glycated casein. All the groups were followed up for 4 weeks. The dietary regimen is summarized in Fig. 2a. On days 1, 7, 14, 21, and 28 of the feeding periods, 6 mice from the HA group and 4 mice from the LA group were selected for euthanasia. The day before euthanasia, the mice were individually housed in metabolic cages, and urine and feces were collected over a 24 h period, rapidly frozen in liquid nitrogen, and stored at −80 °C. After the collection period, mice were fasted overnight (8 h), and the next morning, they were anesthetized with isoflurane, followed by cardiac puncture to obtain anticoagulant blood samples. The plasma samples and small intestine, colon, liver, and kidney tissues were collected from each mouse and weighed. A portion of the tissue was fixed in 4% neutral paraformaldehyde solution, while the remaining tissue was rapidly frozen in liquid nitrogen and stored at −80 °C for subsequent analyses.

Fig. 2 Dynamics of dAGE accumulation over 28 days. (a) Experimental design for the animal-feeding study. Eight-week-old male C57BL/KsJ mice were pair-fed either LA (blue) or HA (red) diets. Sampling was conducted at days 1, 7, 14, 21, and 28. (b) Dynamic changes in the levels of free and bound CML in the plasma, small intestine, colon, liver, kidneys, and urine. (c) Western blot analysis illustrating the accumulation of CML in the small intestine, colon, liver, and kidneys after 28 days of the trial. (d) Immunofluorescence analysis demonstrating the accumulation of CML in the small intestine, colon, liver, and kidneys after 28 days of the trial.

Mass spectrometry

The mass spectrometry procedures were conducted as previously described.¹⁴ In brief, to separate free and bound AGEs, tissues were cryogenically pulverized in a mortar following liquid nitrogen freezing. Subsequently, a 5-fold volume of dd H₂O was added, followed by 1 min of vigorous shaking. The mixture was then centrifuged at 4000g for 15 min at 4 °C. The resulting supernatant was collected, and as per the manufacturer's instructions, a HyperSep™ spin tip (Thermo Scientific) was employed for free and bound fractions’ separation to collect free AGEs, elute the bound fractions and set aside. For the precipitated tissue after centrifugation, a 5-fold volume of RIPA lysis buffer (P0013B, Beyotime, China) was added, and the mixture was incubated at room temperature until no visible solid particles were observed. And, a HyperSep™ spin tip (Thermo Scientific) was used for bound fractions’ collection. The bound fractions collected from two rounds of HyperSep™ spin tips were combined in a 2 mL pressure-resistant tube. An equal volume of protein solution was mixed with a 0.1 mol L⁻¹ sodium borohydride solution. Additionally, after thoroughly grinding the feed, 0.1 g was taken and placed in a 2 mL pressure-resistant tube, and 0.1 mL of 0.1 mol L⁻¹ sodium borohydride solution was then added. To this, a 10-fold volume of 6 N HCl was added, and the tube was placed in a preheated oil bath at 110 °C for 24 h. This process ensured the complete hydrolysis of peptide bonds, liberating AGEs bound to proteins and peptides. Subsequently, the samples were dried in a vacuum concentrator and dissolved in a 10% MeOH solution (200 μL). Plasma and urine samples underwent a similar procedure for the separation of free and bound AGEs using the HyperSep™ spin tip. Prior to LC-MS/MS analysis, all samples were filtered through a 0.22 μm filter.

Reversed-phase chromatography and mass spectrometry analysis were carried out using a SCIEX ExionLC system coupled with a SCIEX X500R Accurate Mass Q-TOF instrument. Specimens were introduced onto a Kinetex® C18 column (100 × 3 mm², Phenomenex, USA) and subjected to elution with a gradient comprising 0.1% formic acid in water (solvent A) and methanol (solvent B) as follows: 0–2 min, 3% B; 2–5 min, 3%–80% B; 5–6 min, 80%–98% B; 6–7 min, 98% B; 7–7.1 min, 98%–3% B; 7.1–8 min, 3% B. The mass spectrometer was operated in positive mode and featured a dual electrospray ionization (ESI) source. MS and MS/MS spectra were collected with the following settings: ESI spray voltage, 5500 V; gas temperature, 550 °C; ion source gas 1, 60 psi; ion source gas 2, 60 psi; curtain gas, 35 psi; CAD gas, 7; declustering potential, 80 V; collision energy, 10 V. Data acquisition occurred at a rate of 0.15 seconds per spectrum with a scan range spanning 40–1000 Da. The quantitative analysis of samples was performed using an external standard method. The parent and daughter ions of CML are 205.1 and 84.1, respectively. The peak area of the parent ion for the CML standard exhibited linearity within the concentration range of 0.1–250 ng mL⁻¹ (ESI Fig. S3†).

Western blotting

The tissue pre-processing and western blotting procedures were conducted as previously described.¹⁴ In brief, tissues were pulverized after snap-freezing with liquid nitrogen using a mortar and pestle. Total protein from the tissues was extracted using a ProteinExt® mammalian total protein extraction kit (TransGen, China). The extracted proteins were diluted to a concentration of 2.4 mg mL⁻¹ in sonication buffer containing 5% (v/v) 2-mercaptoethanol and heated to 98 °C for 10 min. Subsequently, the protein samples were loaded onto 12% SDS-PAGE gels and transferred onto a 0.45 μm PVDF membrane (Solarbio, China) through electrophoresis. Following this, the PVDF membranes were blocked at room temperature for 1 h with a skimmed milk solution and then incubated overnight at 4 °C with a mixture of primary antibodies: anti-CML (Abcam, ab125145, 1 : 1000) and anti-β-actin (TransGen, HC201-01, 1 : 3500). After incubation with the corresponding secondary antibody, the target proteins were visualized using an ECL kit (Solarbio, China) and captured using the ChemiDoc Imaging Systems/Image Lab software (Bio-Rad, USA).

Immunofluorescence

The tissue pre-processing and western blotting procedures were conducted as previously described.²² Briefly, tissue sections were fixed with 4% paraformaldehyde for 15 min at room temperature. After washing with phosphate-buffered saline (PBS), the samples were permeabilized with 0.2% Triton X-100 in PBS for 10 min. Next, the samples were blocked with 5% bovine serum albumin (BSA) for 1 h. After blocking, the sections were incubated with anti-CML antibody at the recommended dilution in 1% BSA overnight at 4 °C. Following primary antibody incubation, the samples were washed with PBS and then incubated with fluorescently labeled secondary antibodies at the appropriate dilution in 1% BSA for 1 h at room temperature, protected from light. After secondary antibody incubation, the samples were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) to visualize cell nuclei. Slices were observed under a positively placed Nikon fluorescence microscope. Quantitative analysis of immunofluorescence staining was performed using CaseViewer 2.2 (3DHISTECH, Budapest, Hungary) and Image-Pro Plus 6.0 (Media Cybernetics, MD).

Taxonomic microbiota analysis

The microbiotas of the fecal samples were deciphered by Novogene Co., Ltd (Beijing, China). In brief, the total genomic DNA was extracted from fecal samples by employing the CTAB/SDS method. Subsequently, the DNA concentration and purity were assessed and the DNA was then diluted to a concentration of 1 ng μL⁻¹ in sterile water. The distinct region of the 16S rDNA, known as 16S V4, was amplified using a specific primer set (515F-806R) equipped with barcodes. Subsequently, sequencing libraries were generated utilizing the ion plus fragment library kit 48 rxns (Thermo Scientific) following the manufacturer's recommendations. Sequence analysis was performed through the utilization of Uparse software (Uparse v7.0.1001, https://drive5.com/uparse/), with sequences showing ≥97% similarity being assigned to the same operational taxonomic units (OTUs).

Fecal metabolome analysis to determine metabolic changes induced by dAGEs

The metabolome of the fecal samples was deciphered by Novogene Co., Ltd (Beijing, China). In brief, fecal metabolites underwent profiling via LC-MS, involving the utilization of a Vanquish UHPLC system (Thermo Fisher Scientific, MA, USA) coupled with an Orbitrap Q Exactive series mass spectrometer (Thermo Fisher). Samples were introduced onto a Hyperil Gold column (100 × 2.1 mm², 1.9 μm) using a 16 min linear gradient at a flow rate of 0.2 mL min⁻¹. The Q Exactive mass spectrometer operated in positive/negative polarity mode with the following parameters: a spray voltage of 3.2 kV, a capillary temperature of 320 °C, a sheath gas flow rate of 35 arb, and an aux gas flow rate of 10 arb. Raw data obtained from UHPLC-MS/MS were subjected to processing via Compound Discoverer 3.0 (CD3.0, Thermo Fisher). This included peak alignment, peak selection, and quantitation for each metabolite. Molecular formulas were predicted based on additive ions, molecular ion peaks, and fragment ions. Peak matching was performed against the mzCloud (https://www.mzcloud.org/) and ChemSpider (https://www.chemspider.com/) databases to obtain precise qualitative and relatively quantitative results. Statistical analyses were carried out using the software programs R (R version R-3.4.3), Python (Python 2.7.6 version), and CentOS (CentOS release 6.6). In cases where data did not exhibit a normal distribution, normal transformations were attempted using the area normalization method.

Statistical analysis

Statistical analysis of the data was conducted using GraphPad Prism 9.0 software (GraphPad, San Diego, USA). One-way analysis of variance (ANOVA) was employed to assess the overall significance of differences among multiple groups. Subsequently, the Tukey post hoc analysis was applied to identify specific pairwise differences between groups. All results are expressed as mean ± standard error of the mean (SEM) to represent the central tendency and variability within the data. Differences were noted as significant with * P ≤ 0.05, ** P ≤ 0.01, and *** P ≤ 0.001.

Results and discussion

Characterization of site-specific glycation of β-casein

To investigate the absorption and clearance of dietary advanced glycation end-products (dAGEs) while excluding interference from endogenous sources, β-casein was reacted with (¹³C₆-) glucose at 115 °C for 60 min to produce (¹³C-labeled) dAGEs (ESI Fig. S1†). Using orbitrap-MS/MS-based peptide mapping, we evaluated the glycation levels of β-casein in low-AGEs (LA) and (¹³C-labeled) high-AGEs (¹³C-) HA diets, identified specific products, and determined the glycation sites. Several glycation modifications were detected, including carboxymethylation (CM), carboxyethylation (CE), pyrrolization (Pyrr), arginine-derived pyrimidine (AP), glyoxal-derived hydroimidazolone (GO-H), and methylglyoxal-derived hydroimidazolone (MG-H) (ESI Fig. S2†). A complete list of modified sites is presented in ESI Table S2.† From the distribution of glycation modifications, it was found that the number of sites involved in the glycation of β-casein is 9 in the LA diet and 11 in the (¹³C-) HA diet, respectively. Carboxymethylation and carboxyethylation were the most frequent modifications observed. In the (¹³C-) HA diet, 6 sites participate in the generation of carboxymethyllysine (CML), and even in the LA diet, 5 sites participate in the generation of CML. Besides, CML is recognized as the most widespread and extensively studied representative dAGE.²³ Hence, in this study, CML was used as the representative of dAGEs (Fig. 1a).

Release rate of CML in the GI tract and elimination rate in the urine

The content of ¹³C₂-CML in the ¹³C-HA diet was measured as 34.2 mg kg⁻¹ protein using MS/MS (Fig. 1b, and ESI Fig. S3†). After fasting for 12 h, mice were fed the ¹³C-HA diet, and plasma and urine samples were collected at 4, 8, 12, and 24 h post-feeding to detect the levels of free and bound ¹³C₂-CML (Fig. 1c). As depicted in Fig. 1d, we observed peak absorption of free CML in plasma at 4 h, reaching 6.6 ng mL⁻¹, which then decreased to 1.4 ng mL⁻¹ after 24 h of metabolism. In contrast, the peak absorption of bound CML in plasma occurred at 8 h post-feeding, reaching 84.6 ng mL⁻¹, and subsequently decreased to 33.1 ng mL⁻¹ after 24 h of metabolism. These findings suggest that bound AGEs may represent the predominant form through which dietary AGEs enter the body (>90%). However, urinary analysis revealed that the excretion of CML primarily occurred in the free form, suggesting that bound AGEs undergo metabolic transformations in various organs, including the renal glomeruli,²⁴ before being excreted as the free form.

To further speculate on the primary absorption sites for dietary free and bound CML, we characterized the changes in the levels of free and bound CML in small intestine washings, cecal contents, and feces (representing post-small intestine digestion, post-digestion, and post-colon digestion, respectively) over the course of digestion (Fig. 1c and d). The results indicated that the peak release of free CML from small intestine contents occurred at 4 h post-feeding, while the peak presence of bound CML occurred at 8 h post-feeding, coinciding with the timing of peak free and bound CML levels in plasma. In cecal contents, the peak levels of both free and bound CML, as well as the peak level of free CML in feces, were observed at 12 h post-feeding, which did not align with the timing of peak free and bound CML levels in plasma. Collectively, these findings suggest that the small intestine is the main absorption site for both free and bound dietary AGEs. The colon (feces) harbors a substantial amount of AGEs, particularly free AGEs, but lacks an absorption pathway. This outcome indirectly supports the notion that the absorption of free CML occurs via active transport rather than passive diffusion.¹⁴

The dynamic accumulation process of dAGEs

To investigate the temporal retention of absorbed dAGEs in circulation and the accumulation of retained dAGEs in major metabolic organs, mice were fasted for 12 h before being fed a high AGE (HA) diet, with mice fed a normal low AGE (LA) diet serving as controls (Fig. 2a). The HA diet contained 58.9 mg kg⁻¹ protein of CML, while the LA diet contained 23.7 mg kg⁻¹ protein of CML, as determined by MS/MS (Fig. S1†). Plasma, urine, small intestine, colon, liver, and kidney samples were collected at days 1, 7, 14, 21, and 28 post-feeding to measure the levels of free and bound CML.

As depicted in Fig. 2b, plasma levels of free CML remained relatively constant over the 28-day culture period for both LA and HA mice, with HA mice generally exhibiting higher levels than LA mice. Conversely, plasma levels of bound CML in HA mice showed no significant difference compared to LA mice. Interestingly, the urinary excretion of free CML increased with longer exposure to the HA diet. This suggests that a portion of absorbed free or bound dAGEs accumulates in circulation, elevating total circulating AGE levels, while concomitantly increasing the metabolic excretion of AGEs.

The liver and kidneys are the major organs for dAGEs’ accumulation and metabolism.¹⁶ Therefore, assessing the cumulative changes in free and bound dAGEs in the liver and kidneys with increasing duration of HA diet intake is crucial for understanding the mechanisms underlying the short- and medium-term effects of dAGE ingestion on organismal health. Mass spectrometry results demonstrated that the levels of bound CML in the liver remained insignificantly different from the LA group over the course of the 28-day HA diet, a finding confirmed by western blot and immunofluorescence results (Fig. 2b–d and ESI Fig. S4†). This implies that, at least during the 28-day culture period, digested and absorbed dAGEs do not significantly accumulate in the liver. We speculate that the impact of dAGEs on liver energy homeostasis and inflammation may be acute. In contrast to the liver, the kidneys accumulated most of the free and bound AGEs, and with prolonged HA diet intake, the accumulated levels of both free and bound CML in the kidneys significantly increased (Fig. 2b). Western blot and immunofluorescence results after the 28-day feeding period also confirmed the accumulation of dAGEs in the kidneys (Fig. 2c and d). Uribarri et al.²⁵ align with our observation that the kidneys accumulate more dAGEs than the liver. These findings suggest that the kidneys, rather than the liver, are likely the primary organs responsible for the effects of absorbed dAGEs.

Furthermore, considering that dietary AGEs have been observed to accumulate in the intestines and potentially trigger intestinal inflammation,²⁶ we evaluated the accumulation capacity of free and bound AGEs separately in the small intestine and cecum. Our findings revealed that levels of free CML in the small intestine and both free and bound CML in the colon of mice fed a high AGE diet did not significantly surpass those of mice on the low AGE diet over the 28-day feeding period (Fig. 2b). This further reinforces that the colon is not a major absorption site for dAGEs. In contrast, we noted an enrichment of bound CML in the small intestine of mice fed a high AGE diet (Fig. 2b–d). Interestingly, the accumulation of bound CML in the small intestine reached a peak at day 7 (approximately 6 ng g⁻¹), which did not increase further with the subsequent 21 days of a high AGE diet (Fig. 2b). We hypothesize that this is related to the fact that carboxypeptidases enriched in small intestinal epithelial cells efficiently hydrolyze dAGEs in the peptide-bound state.²⁷

Dynamic process of the impact of unabsorbed dAGEs on gut microbiota composition

In addition to the extensively studied impact of absorbed dietary advanced glycation end-products (dAGEs) on the body,²⁸ we emphasize the role of unabsorbed dAGEs, which are extensively utilized by the gut microbiota in the colon and almost entirely converted into free forms (Fig. 1d). Given such high utilization rates of bound dAGEs, we hypothesized that dAGEs rapidly and significantly regulate the composition of the gut microbiota. While many researchers have noted this phenomenon,²⁹ the duration required for a high AGE (HA) diet to completely reshape the gut microbiota and the specific strains that are affected initially remain unknown. Investigating the temporal effects of dAGEs on the composition of the gut microbiota is crucial for uncovering dAGE-utilizing bacteria and defining dAGE-sensitive strains.

To address this, we employed microbiome analysis to assess the gut microbiota composition of HA-fed mice at days 1, 7, 14, 21, and 28 post-HA diet initiation, with mice on a low AGE (LA) diet serving as controls. Results showed significant changes in microbial composition within the first 7 days of HA diet intake, as evidenced by a notable increase in the phylum Firmicutes/phylum Bacteroidetes (F/B) ratio, rising significantly from 0.7 to 11.7, before returning to 0.3 by day 14 and stabilizing (Fig. 3a and ESI Fig. S5†). This indicates that the pylum Firmicutes and the pylum Bacteroidetes demonstrate high sensitivity to dietary AGEs in the initial stages of HA diet intake and rapidly adjust to reach a stable state within two weeks, which is evident from the Unweighted UniFrac PCoA plot (Fig. 3b). Notably, many studies have discussed how dAGEs alter gut microbiota composition and function, particularly the lower abundance of Firmicutes and increased abundance in Bacteroides over a long feeding period.³⁰ However, we are the first to report that the alteration of gut microbiota composition by dAGEs is phased, with a lower abundance of Firmicutes and an increased abundance of Bacteroides only occurring after more than 2 weeks of dAGE treatment. This finding is of great significance for understanding the impact of dAGEs on gut microbiota.

Fig. 3 Dynamics of gut microbiota changes influenced by HA diet. (a) Dynamic changes in the relative abundance of microbial taxa at the phylum level. (b) Principal component analysis (PCA) plot comparing the effects of the HA diet ingestion period on gut microbiota distribution. (c) Top changing microbial taxa at various taxonomic levels under different HA diet ingestion periods.

The distinct microbial changes were observed at various taxonomical levels (Fig. 3c). In comparison to LA-fed mice, within the first day of HA diet intake, microbial taxa that were significantly modulated included the order Enterobacterales and Clostridia UCG 014, the genus Leuconostoc, and the family Enterobacteriaceae, which can be defined as dAGE-sensitive bacteria. By the seventh day of HA diet intake, the most significant changes in microbial composition occurred, and by the fourteenth day, the microbial composition had essentially stabilized. This period suggests that the rapid proliferation of dAGE-utilizing bacteria might have balanced the changes in the gut microbiota caused by the substantial influx of dAGEs. Therefore, we consider the family Bacteroidaceae and the genus Bacteroides, significantly altered in the 14-day HA diet intake compared to LA-fed mice, as potential dAGE-utilizing bacteria that may contribute to stabilizing the gut microbiota during high AGE diet intake.

At the genus level, the relative abundance of the genera Staphylococcus, Lactobacillus, and Atopostipes significantly increased after 7 days of HA diet intake, while the genera Alistipes, Bacteroides, Parabacteroides, Lachnospiraceae NK4A136 group, Terrisporobacter, and Odoribacter decreased significantly (Fig. 4a, c and ESI Fig. S6†). Using the 28-day HA diet intake as a representative timepoint for a stable gut microbiota adaptation to dAGEs, we found that the genera Bacteroides, Parabacteroides, Parasutterella, and Faecalibaculum were significantly higher in relative abundance compared to LA-fed mice at this time, while the genera Alistipes, Staphylococcus, Lachnospiraceae NK4A136 group, Terrisporobacter, Odoribacter, Alloprevotella, and Akkermansia were significantly lower (Fig. 4b, c and ESI Fig. S6†). Intriguingly, many microbial genera displayed completely opposite trends in relative abundance at days 7 and 28 of HA diet intake, such as the genera Bacteroides, Staphylococcus, and Parabacteroides. The genera Bacteroides and Parabacteroides showed a decrease followed by an increase, while the genus Staphylococcus exhibited an increase followed by a decrease in relative abundance, with the transition period occurring between days 7 and 14.

Fig. 4 Genus-level gut microbiota affected by HA diet. (a and b) Major regulated genera at day 7 (a) and day 28 (b) after HA diet supplementation. (c) Relative abundance (%) of the genera Alistipes, Bacteroides, Staphylococcus, Parabacteroides, Lachnospiraceae_NK4A136_group, Terrisporobacter, Odoribacter, and Atopostipes. Significance determined using one-way ANOVA with Tukey's post hoc analysis and expressed as mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001.

Dynamic process of the impact of unabsorbed dAGEs on gut microbiota metabolism

The metabolic homeostasis of the gut microbiota is a crucial component of host health.³¹ Over 2000 metabolites were identified in feces using metabolomics methods. Principal Component Analysis (PCA) identified metabolites enriched in PC1 (53.14% of the total variance) and/or PC2 (11.07% of the total variance) in feces (Fig. 5a). Multiple fecal metabolites were common among groups with more than 14 days of HA diet feeding but were distinctly separated from the 0 d, 1 d, and 7 d groups (Fig. 5a). Additionally, pairwise comparisons of the groups revealed clear separation between the 0 d, 1 d, and 7 d groups.

Fig. 5 Fecal metabolomics changes with varied HA diet intake periods. (a) Principal component analysis (PCA) plot comparing the effects of the HA diet intake periods on fecal metabolite distribution. (b) Volcano plot depicting the significance of differences between fecal metabolomes of HA diet intake days 1, 7, and 28 compared to 0 d. (c) Metabolites that significantly change in the same direction (blue) or opposite direction (red) at HA diet intake days 7 and 28 compared to 0 d are plotted as loading coefficients, as they contribute to PLSR scores of the fecal metabolome. Each metabolite is represented by a dot. (d) Predicting potential host functional changes resulting from differential microbiota–metabolite interactions induced by the HA diet, top changes after 28 days of HA diet consumption compared to 0 d.

The volcano plot demonstrated that as the duration of HA diet intake increased, the number of differential metabolites and their differences continued to rise (Fig. 5b). ANOVA indicated specific metabolites that showed dynamic changes compared to the LA diet group during HA diet intake (Fig. 5c and ESI Fig. S8†). Statistically significantly associated metabolites were colored to indicate whether they changed in the same or opposite direction between days 7 and 28 of HA diet intake (Fig. 5c). As shown, over 25% of the differential metabolites were fatty acids and their metabolites, and all metabolites significantly changed relative to the LA group in both days 7 and 28 of HA diet intake (Fig. 5c and ESI Fig. S8†). Together, these findings suggest that HA diet intake leads to a sustained deviation in the gut microbiota's metabolism from normal levels, potentially impacting host health.

Predicting potential host functional changes resulting from differential microbiota–metabolite interactions induced by a HA diet, pathway enrichment analysis indicated that regulated host functional pathways primarily involved amino acid metabolism, lipid metabolism, and carbohydrate metabolism (Fig. 5d and ESI Fig. S8†). Microbial cometabolites whose levels were modulated by microbiota abundance,³² such as hippurate, trimethylamine, 4-hydroxyphenylacetate, 3-indoxylsulfate, tyrosine, serotonin, and tryptophan, were significantly reduced among the metabolites significantly changed by the HA diet. Conversely, long-term HA diet intake was associated with elevated levels of lipids and lipid metabolites, such as FAHFA (18:3/18:2), LPI 16:0, PG (12:0/16:0), PS (15:0/15:0), PS (15:0/18:2), PG (14:0/18:1), and FAHFA (2:0/18:1).

To gain further insights into the mechanisms of microbiota–metabolome interactions, we performed semi-quantitative correlation analysis using differential plasma metabolites and differential gut microbiota as horizontal and vertical coordinates, respectively (Fig. 6). We found no significant associations between gut microbiota composition and dietary preference at the phylum level but significant associations at the genus level. The HA diet led to a significant increase in the relative abundance of the microbial genera Bacteroides, Parabacteroides, Atopostipes, Akkermansia, and Faecalibaculum, which may have resulted in elevated levels of metabolites like PE (18:2/18:2), LPC 16:0, FAHFA (18:3/18:2), and PG (14:0/18:1). Importantly, these metabolites’ substrates are not directly derived from or linked to dAGEs. Furthermore, dAGEs promoted an increase in lipid metabolism while inhibiting the levels of sugar metabolism products or hydrophilic metabolites such as hippurate, cAMP, D-glucarate, and hydroxyphenylpyruvate. The impact of dAGEs on microbial metabolism resembles their previously reported effects on host metabolism, primarily due to AGEs interfering with glucose and lipid metabolism pathways.³³

Fig. 6 Correlation analysis revealing the interplay among diet, gut microbes, and fecal metabolites. The heat map illustrates the correlation r-value between differential gut microbes and differential fecal metabolites (inner), with the p-value representing the relative association of microbes/metabolites with dietary comparisons (outer), as indicated in the legend.

Conclusion

In summary, unabsorbed dAGEs are efficiently utilized in the colon, significantly influencing gut microbiota and its metabolic functions. Changes in microbial composition induced by a high-AGE (HA) diet exhibit a stress-induced mutation period during days 1–7, followed by gradual recovery and stabilization after day 14. At the microbial metabolic level, HA diet-induced changes persistently and stably affect fatty acid metabolism through the regulation of the microbial composition and metabolic pathways, potentially impacting host metabolic homeostasis. However, the mechanism through which dAGEs influence health is multifaceted. The interaction between dAGEs and gut microbiota is pivotal, as gut bacteria can metabolize these compounds, potentially leading to the production of bioactive metabolites. Additionally, dAGEs can alter gut microbiota composition, promoting dysbiosis and affecting metabolic pathways.

This study highlights the importance of considering both direct and indirect effects of dAGEs on host health. It contributes to the growing body of literature on the impact of dietary AGEs on health, underscoring the crucial role of gut microbiota. The findings align with previous research indicating that the consumption of AGEs influences metabolic pathways and microbiota composition. The gut microbiota's ability to metabolize AGEs significantly affects host metabolic health. Moreover, changes in lipid and carbohydrate metabolism associated with AGE intake corroborate earlier studies linking AGEs to metabolic dysregulation. Additionally, the long-term health implications of AGE accumulation emphasize the importance of dietary interventions to mitigate adverse effects. Future research should explore the potential for targeting gut microbiota to develop therapeutic strategies aimed at reducing AGE-related health risks.

Abbreviations

AGEs	Advanced glycation end-products
CML	Nε-(carboxymethyl)-L-lysine
dAGEs	Dietary AGEs
HA	High AGE
LA	Lower AGE

Author contributions

Yi Wu performed the experiments and analyzed the data. Yi Wu, Yongtai Wu, and Z. Z. designed the study. Yi Wu and W. W. wrote the paper. Yi Wu, Y. Z., Z. Y., and B. L. prepared the figures. All authors discussed the results and commented on the paper.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request. Data sharing is subject to ethical and legal considerations, including the protection of participant confidentiality. Specific datasets analyzed during the current study are not publicly available due to these constraints but can be made accessible through a formal data sharing agreement.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

This work was supported by the Zhejiang University of Science and Technology Youth Science Fund (no. 2023QN100), the Zhejiang Provincial Natural Sciences Foundation of China (grant no. LZ22C200006), the Top Young Talent of the Ten Thousand Talents Program of Zhejiang Province (no. ZJWR0308016), and the Zhejiang Basic Public Welfare Research Project (no. LGN22C200024 & LGN21C200006).

References

1. W. Rungratanawanich, Y. Qu, X. Wang, M. M. Essa and B.-J. Song, Advanced glycation end products (AGEs) and other adducts in aging-related diseases and alcohol-mediated tissue injury, Exp. Mol. Med., 2021, 53, 168–188.
2. C.-Y. Shen, C.-H. Lu, C.-H. Wu, K.-J. Li, Y.-M. Kuo, S.-C. Hsieh and C.-L. Yu, The development of maillard reaction, and advanced glycation end product (AGE)-receptor for AGE (RAGE) signaling inhibitors as novel therapeutic strategies for patients with AGE-related diseases, Molecules, 2020, 25, 5591.
3. D. Sergi, H. Boulestin, F. M. Campbell and L. M. Williams, The role of dietary advanced glycation end products in metabolic dysfunction, Mol. Nutr. Food Res., 2021, 65, 1900934.
4. M. Hellwig and T. Henle, Baking, ageing, diabetes: a short history of the Maillard reaction, Angew. Chem., Int. Ed., 2014, 53, 10316–10329.
5. C. W. Heizmann, The mechanism by which dietary AGEs are a risk to human health is via their interaction with RAGE: arguing against the motion, Mol. Nutr. Food Res., 2007, 51, 1116–1119.
6. A. Twarda-Clapa, A. Olczak, A. M. Białkowska and M. Koziołkiewicz, Advanced glycation end-products (AGEs): Formation, chemistry, classification, receptors, and diseases related to AGEs, Cells, 2022, 11, 1312.
7. Q. Zhang, Y. Wang and L. Fu, Dietary advanced glycation end–products: Perspectives linking food processing with health implications, Compr. Rev. Food Sci. Food Saf., 2020, 19, 2559–2587.
8. T. Henle, Protein-bound advanced glycation endproducts (AGEs) as bioactive amino acid derivatives in foods, Amino Acids, 2005, 29, 313–322.
9. F. J. Tessier, E. Boulanger and M. Howsam, Metabolic transit of dietary advanced glycation end-products-the case of N ε-carboxymethyllysine, Glycoconjugate J., 2021, 38, 311–317.
10. Z. Liang, X. Chen, L. Li, B. Li and Z. Yang, The fate of dietary advanced glycation end products in the body: From oral intake to excretion, Crit. Rev. Food Sci. Nutr., 2020, 60, 3475–3491.
11. C. Delgado-Andrade and V. Fogliano, Dietary advanced glycosylation end-products (dAGEs) and melanoidins formed through the Maillard reaction: physiological consequences of their intake, Annu. Rev. Food Sci. Technol., 2018, 9, 271–291.
12. G. Chen, Dietary N-epsilon-carboxymethyllysine as for a major glycotoxin in foods: A review, Compr. Rev. Food Sci. Food Saf., 2021, 20, 4931–4949.
13. K. Nowotny, D. Schröter, M. Schreiner and T. Grune, Dietary advanced glycation end products and their relevance for human health, Ageing Res. Rev., 2018, 47, 55–66.
14. Y. Wu, M. Zong, Z. Zhang, Y. Wu, L. Li, X. Zhang, H. Wu and B. Li, Selective transportation and energy homeostasis regulation of dietary advanced glycation end-products in human intestinal Caco-2 cells, Food Chem., 2022, 391, 133284.
15. R. Mastrocola, D. Collotta, G. Gaudioso, M. Le Berre, A. S. Cento, G. F. Alves, F. Chiazza, R. Verta, I. Bertocchi and F. Manig, Effects of exogenous dietary advanced glycation end products on the cross-talk mechanisms linking microbiota to metabolic inflammation, Nutrients, 2020, 12, 2497.
16. X. Yuan, C. Nie, H. Liu, Q. Ma, B. Peng, M. Zhang, Z. Chen and J. Li, Comparison of metabolic fate, target organs, and microbiota interactions of free and bound dietary advanced glycation end products, Crit. Rev. Food Sci. Nutr., 2023, 63, 3612–3633.
17. J. Wang, W. Cai, J. Yu, H. Liu, S. He, L. Zhu and J. Xu, Dietary advanced glycation end products shift the gut microbiota composition and induce insulin resistance in mice, Diabetes, Metab. Syndr. Obes.: Targets Ther., 2022, 427–437.
18. Y. Wu, M. Zong, H. Wu, D. He, L. Li, X. Zhang, D. Zhao and B. Li, Dietary Advanced Glycation End–Products Affects the Progression of Early Diabetes by Intervening in Carbohydrate and Lipid Metabolism, Mol. Nutr. Food Res., 2022, 66, 2200046.
19. K. Nowotny, D. Schroter, M. Schreiner and T. Grune, Dietary advanced glycation end products and their relevance for human health, Ageing Res. Rev., 2018, 47, 55–66.
20. Q. Wei, T. Liu and D.-W. Sun, Advanced glycation end-products (AGEs) in foods and their detecting techniques and methods: A review, Trends Food Sci. Technol., 2018, 82, 32–45.
21. M.-H. Pai, K.-H. Huang, C.-H. Wu and S.-L. Yeh, Effects of dietary arginine on inflammatory mediator and receptor of advanced glycation endproducts (RAGE) expression in rats with streptozotocin-induced type 2 diabetes, Br. J. Nutr., 2010, 104, 686–692.
22. Y. Wu, Y. Li, Z. Ruan, J. Li, L. Zhang, H. Lu and Z. Xu, Puerarin rebuilding the mucus layer and regulating mucin-utilizing bacteria to relieve ulcerative colitis, J. Agric. Food Chem., 2020, 68, 11402–11411.
23. D. Zhao, B. Sheng, Y. Wu, H. Li, D. Xu, Y. Nian, S. Mao, C. Li, X. Xu and G. Zhou, Comparison of free and bound advanced glycation end products in food: A review on the possible influence on human health, J. Agric. Food Chem., 2019, 67, 14007–14018.
24. A. Bettiga, F. Fiorio, F. Di Marco, F. Trevisani, A. Romani, E. Porrini, A. Salonia, F. Montorsi and R. Vago, The modern western diet rich in advanced glycation end-products (AGEs): An overview of its impact on obesity and early progression of renal pathology, Nutrients, 2019, 11, 1748.
25. J. Uribarri, S. Woodruff, S. Goodman, W. Cai, X. Chen, R. Pyzik, A. Yong, G. E. Striker and H. Vlassara, Advanced glycation end products in foods and a practical guide to their reduction in the diet, J. Am. Diet. Assoc., 2010, 110, 911–916.
26. K. Phuong-Nguyen, B. A. McNeill, K. Aston-Mourney and L. R. Rivera, Advanced glycation End-products and their effects on gut health, Nutrients, 2023, 15, 405.
27. F. A. Jansen, V. Fogliano, J. Rubert and T. Hoppenbrouwers, Dietary Advanced Glycation End products interacting with the intestinal epithelium: What do we really know?, Mol. Metab., 2023, 101734.
28. V. P. Reddy, P. Aryal and E. K. Darkwah, Advanced glycation end products in health and disease, Microorganisms, 2022, 10, 1848.
29. C. Nie, Y. Li, H. Qian, H. Ying and L. Wang, Advanced glycation end products in food and their effects on intestinal tract, Crit. Rev. Food Sci. Nutr., 2022, 62, 3103–3115.
30. W. Qu, X. Yuan, J. Zhao, Y. Zhang, J. Hu, J. Wang and J. Li, Dietary advanced glycation end products modify gut microbial composition and partially increase colon permeability in rats, Mol. Nutr. Food Res., 2017, 61, 1700118.
31. K. Oliphant and E. Allen-Vercoe, Macronutrient metabolism by the human gut microbiome: major fermentation by-products and their impact on host health, Microbiome, 2019, 7, 1–15.
32. L. S. Zhang and S. S. Davies, Microbial metabolism of dietary components to bioactive metabolites: opportunities for new therapeutic interventions, Genome Med., 2016, 8, 1–18.
33. Y. Wu, Y. Wu, Z. Zhang, J. Cai, M. Zong, L. Li, X. Zhang, H. Wu and B. Li, N ε-(Carboxymethyl) lysine (CML) Promotes Disorders of Glucose and Lipid Metabolism in Human Hepatic LO2 Cells, ACS Food Sci. Technol., 2023, 3(10), 1635–1645.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4fo01545b

---