A noninvasive ¹³C-mannitol breath test to monitor semaglutide treatment-induced delayed oral–cecal transit in mice

Abstract

Glucagon-like peptide-1 receptor agonists (GLP-1RAs), such as semaglutide, are highly effective treatments for obesity and type 2 diabetes but are often associated with gastrointestinal side effects including delayed transit times through the gastrointestinal tract. We present a non-invasive breath test to measure oral-cecal transit time (OCTT) in mice that could be readily translated to clinical studies. Current methods for measuring OCTT are limited by invasiveness, terminal procedures, anesthesia, radioactive tracers, or stress-induced influence on the outcomes. Here we demonstrate that ¹³C-labeled mannitol is metabolized by murine cecal microbiota, producing measurable ¹³CO₂ in exhaled breath. Using indirect calorimetry combined with real-time measurement of the ¹³C/¹²C ratio, this approach provides a highly granular, non-invasive method to measure OCTT in mice. This technique is performed in unrestrained freely moving animals while enabling high-throughput measurements performed longitudinally. The method improves experimental rigor and reproducibility and provides a valuable tool for studying gastrointestinal physiology and the effects of pharmacologic therapies.

Graphical Abstract

New and Noteworthy:

A novel non-invasive assay uses orally administered ¹³C-mannitol, metabolized by cecal microbiota into exhaled ¹³CO₂, to enable continuous, longitudinal measurement of oral-cecal transit time in unanesthetized, freely moving mice via indirect calorimetry with isotope ratio analysis. Pharmacological validation confirmed sensitivity to delayed transit, including a pronounced semaglutide-induced delay in obese mice that models a clinically relevant GLP-1 receptor agonist adverse effect.

Introduction:

GLP1 receptor agonists (GLP1Ras), including semaglutide, are a major breakthrough in the treatment and management of obesity and type 2 diabetes (1). Despite their efficacy, GLP1Ras are often associated with side effects, with gastrointestinal symptoms being the most commonly reported (2). The ability to model efficacy and adverse effects is key to developing approaches to obviate or treat adverse effects. While the efficacy of GLP1Ra has been well studied in mice, the side effects are less well characterized (3).

One commonly observed GI side effect is delayed gastrointestinal transit time (4). However, methods for measuring oral-cecal transit time (OCTT) in murine models present significant challenges. In earlier studies of the effect of GI transit with GLP-1, methods typically used terminal studies or prolonged fasting prior to testing (5). Existing approaches, such as oral administration of fluorescently labeled polymers (e.g. FITC dextran) or metallic beads followed by dissection of the GI tract can offer high spatial resolution in determining tracer position along the gastrointestinal tract, but may confound results due to the use of anesthesia or physiological invasiveness (6, 7). These methods, including physical examination of dye fronts within the murine GI tract, can provide a high level of accuracy but are limited by the fact that they are terminal procedures. Available non-terminal whole gut transit measurements include the carmine red assay, which involves regular examination of animals to determine when the dye has been excreted (8). These examinations of the mice during the assay can drastically decrease transit time due to stress, and although automated methods have been developed to improve this assay, it still lacks the resolution of the aforementioned terminal procedures (9). Other non-terminal alternatives, such as Single-Photon Emission Computed Tomography (SPECT) after administration of 99mTc-DTPA-labeled activated charcoal, require anesthesia of murine models and use of short half-life radioisotopes (10). These constraints present an opportunity for the development of a non-invasive technique for accurately measuring OCTT.

The lactulose breath test is a widely used clinical approach for estimating oral-cecal transit time, most commonly employed in the diagnosis of small intestinal bacterial overgrowth (SIBO). Patients consume a lactulose solution and exhaled hydrogen (H₂) and methane (CH₄), produced by gut bacteria, are sampled at defined intervals (11). However, several considerations limit the utility of lactulose as a quantitative OCTT tracer in murine preclinical models. First, lactulose itself has been reported to accelerate gastrointestinal transit (12), introducing a potential confound in studies of already-disrupted motility—a limitation of particular relevance when investigating pharmacological agents that alter gut transit. In contrast, tracer amounts of sugar alcohols such as mannitol do not alter transit rate in murine models (13), making them more physiologically neutral probes for gastrointestinal transit dynamics. Second, it is not established that mice housed under specific pathogen-free (SPF) conditions reliably harbor the microbial populations required to ferment lactulose and generate detectable H₂ or CH₄. Third, uniformly ¹³C-labeled lactulose is less commercially available and considerably more expensive than ¹³C-mannitol, further limiting its practical utility for preclinical throughput studies. Taken together, ¹³C-mannitol represents a less perturbative, more technically accessible, and more cost-effective alternative for continuous, quantitative measurement of oral-cecal transit in mice.

A noninvasive method to measure gastrointestinal transit times in mice would significantly advance our ability to study motility, nutrient absorption, complex interactions between diet, microbiota, and host physiology, as well as the impact of pharmacologic therapies such as semaglutide.

The use of real-time stable isotope measurement is well established in both clinical and preclinical research (14–16). One such example is the Gastric Emptying Breath Test where patients eat a meal containing a non-radioactive, ¹³C-labeled octanoic acid (OA), a medium-length fatty acid absorbed early in the duodenum. Once absorbed, OA is metabolized in the liver, and a fraction is exhaled as ¹³CO₂. The time course of ¹³CO₂ production reveals information about the rate at which solid food is emptied from the stomach and reaches the duodenum (17). However, detection of stable isotope tracer-metabolites more distally in the GI tract is not commonplace due to the lack of appropriate reagents.

Sugar alcohols, including mannitol, sorbitol, and xylitol are used as artificial sweeteners as they produce a sweet taste sensation and increase the palatability of foods but are poorly absorbed in the human gut (18). Mannitol occurs naturally in watermelon, peaches, and olives (19, 20). As a food additive, mannitol does not promote weight gain in animal models (21). At high concentrations, orally ingested mannitol can produce a laxative effect (22), but these effects are not commonly seen in doses used for sweeteners.

Mannitol is administered clinically for several indications. This includes use as an osmotic diuretic to reduce acute renal failure, to reduce intracranial pressure or intraocular pressure when given by intravenous routes (23, 24). Administration of high doses of mannitol in water will have a laxative effect in bowel preparations (25). ¹³C-labeled mannitol has been demonstrated as safe when administered to patients to detect altered intestinal permeability (26). These studies suggest that clinical applications to study GI transit will be well tolerated in patients and rodents.

Here we demonstrate ¹³C-labeled mannitol is metabolized by microbiota in the murine cecum. This study uses new indirect calorimetry technology to detect total CO₂ production combined with a real-time detector of ¹³C/¹²C ratio. This allows for rapid, sequential, measurement of multiple cages simultaneously. These capabilities can increase the throughput of preclinical studies while also limiting potential extraneous variables. The ability to conduct repeat studies with greater ease and consistency enhances reproducibility and robustness of findings. This advancement could represent a step forward in experimental design, aligning scientific rigor with ethical considerations in animal research.

Materials & Methods

Animals

All animal experiments were approved by the Beth Israel Deaconess Medical Center IACUC. Wildtype, male C57BL/6 mice (Jackson Laboratory, Strain No. 000664) were used in all animal experiments. Mice were maintained at 12 h/12 h light/dark cycles (0600/1800), 22 ± 2 °C room temperature and 30%–70% humidity with ad libitum access to food and water, in individually ventilated cages. Cages and bedding were changed once every two weeks, and mice were monitored regularly for their health status by animal technicians with the support of veterinary care and remained free of any adventitious infections for the entire duration of this study. Mice were fed a standard chow diet (13% kcal fat, LabDiet, no. 5008) or a high-fat diet (HFD, 60% kcal fat, Research Diets, no. D12492i) for the indicated durations. All experiments were performed in adult mice (12 to 24 weeks of age). For semaglutide diet comparison studies, mice were placed on HFD 4 weeks prior to the start of the studies.

Peanut butter acclimation

While octanoic acid tracer incorporation into cooked egg yolks has traditionally been used with patients and with rats(15, 27), we found inconsistent consumption of this test meal in C57Bl/6J mice possibly due to the odor. Therefore, to promote the voluntary consumption of our tracer compounds, peanut butter was selected as a palatable alternative. Mice were acclimated to peanut butter with either unlabeled mannitol or octanoic acid in a ceramic dish for 30 minutes on three successive days.

¹³C-octanoic acid gastric emptying assay

We used 1-¹³C-labeled octanoic acid (Cambridge Isotope Laboratories, CLM-293) to validate gastric emptying measurements. 160 μL of liquid octanoic acid was diluted in 1 mL DMSO and further diluted in sterile water to 5.4 mg·mL⁻¹, and 160 μL of this solution was mixed into 5 g of peanut butter (Jif, WB Mason, SMU5150024114) warmed to 60 °C, stirred, then cooled to room temperature. Mice were provided a 100-mg portion of the ¹³C-octanoic acid-enriched peanut butter in a clean ceramic dish and monitored to confirm full consumption within 30 minutes. Mice that did not fully consume the dose were excluded from analysis. Mice had free access to food and water for the duration of the experiment. Exhaled breath was sampled continuously for up to 200 minutes using an ABB Off-Axis Integrated Cavity Output Spectroscopy (OA-ICOS) gas analyzer connected to a Promethion indirect calorimetry system (Sable Systems). Data were exported from Promethion Live (Sable Systems, v24.0.12), converted to CalR-compatible files with Macro Interpreter, and analyzed using RStudio with R version 4.4.1. For pharmacological validation studies, a subset of mice received 10 mg·kg⁻¹ metoclopramide, intraperitoneally (Sigma Aldrich, M0763) or saline 20 minutes before administration of a 200-μL oral gavage of ¹³C-octanoic acid in water.

Antibiotic Induced Commensal Depletion

To deplete commensal bacteria, mice were provided with an antibiotic cocktail in their drinking water. Antibiotics were dissolved in autoclaved, filtered facility water, adjusted to a neutral pH (7.4), and filter-sterilized prior to use. The medicated water was provided in light-protected bottles and replaced every 5–7 days to ensure stability and minimize microbial or fungal contamination. The antibiotic cocktail consisted of neomycin trisulfate (1 g·L⁻¹; Sigma-Aldrich, N5285–25G), metronidazole (0.25 g·L⁻¹; Patterson Vet Generics, 78912046), vancomycin hydrochloride (0.5 g·L⁻¹; Sigma-Aldrich, V8138–1G), and ampicillin trihydrate (1 g·L⁻¹; Sigma-Aldrich, A1593–25G). Neomycin and ampicillin were prepared from powder, while metronidazole was added as a liquid injectable formulation (500 mg/100 mL) at a dose of 0.25 g·L⁻¹. Depletion of commensal bacteria was verified using culture-based methods. Fecal pellets were collected from antibiotic-treated and untreated control mice, homogenized in sterile phosphate-buffered saline (PBS), serially diluted, and plated on LB agar.

¹³C-mannitol oral-cecal transit time assay

Oral-cecal transit time was assessed using ¹³C-labeled mannitol (Cambridge Isotope Laboratories, CLM-6733). The mannitol was received as a powder and dissolved to 100 mg·mL⁻¹ in sterile water to prepare it for administration. Two distinct administration routes were employed in this study, each selected to address different experimental objectives. For voluntary feeding studies, 160 μL of stock was mixed into 5 g of peanut butter and cooled after heating to 60 °C. For gavage, the solution was diluted to 0.8 mg·mL⁻¹ in sterile water. Mice received either a 100-mg portion of the prepared peanut butter or a 200-μL oral gavage of ¹³C-mannitol. Voluntary consumption of tracer incorporated into peanut butter was used for experiments prioritizing physiological and behavioral fidelity. Oral gavage is an inherently stressful procedure that can stimulate gastrointestinal motility and thereby alter baseline transit dynamics. In addition, the fasting that typically precedes gavage studies can significantly increase locomotor activity and metabolic variability in mice (28). Providing the tracer in a palatable vehicle with continuous ad libitum access to food and water minimizes handling-related stress and eliminates the requirement for food withdrawal, making this approach better suited to measuring baseline OCTT under naturalistic conditions. However, voluntary consumption studies were associated with inconsistent intake patterns in a subset of subjects, resulting in incomplete or variable dosing across groups. Oral gavage was therefore used for all pharmacological experiments involving loperamide and semaglutide, where precise and uniform dose delivery, a well-defined experimental time-zero, and rigorous between-group comparability are required. A direct comparison of the two administration routes is presented in Figure 2E–H, demonstrating that solid versus liquid vehicle delivery significantly influences apparent tracer oxidation kinetics. Accordingly, data obtained by the two methods were analyzed separately and were not combined in any comparative analysis. Mice had free access to food and water for the duration of the experiment. Breath samples were collected continuously for up to 6 hours with the Promethion and ABB OA-ICOS. Data processing and analysis were performed in RStudio following conversion to CalR-compatible format. In separate experiments, FD&C Blue No. 1 dye (200 μL of dyed water) was administered by gavage to map the dye front at specific time points (100 or 145 minutes) post-gavage. After sacrifice, the gastrointestinal tract was examined and photographed.

Figure 2: Determinants of [¹³C]-Mannitol Oxidation: Dose, Medium, Location.

Mice voluntarily ate a [¹³C]-mannitol labeled test meal and the production of ¹³CO₂ was detected (mean with SEM ribbon) (n = 7–8), plotted by dose. A) Rate of ¹³CO₂ production, plotted in nmol·min⁻¹. B) Cumulative production of ¹³CO₂ (mean with individual subject lines plotted). C) Cumulative production of ¹³CO₂. D) T₅₀%, in minutes. E-H) [¹³C]-mannitol was gavaged or voluntarily eaten in a test meal and the production of ¹³CO₂ was detected (mean with SEM ribbon) (N = 7,9). E) ¹³CO₂ detected (mean with SEM ribbon) after consumption of [¹³C]-mannitol incorporated into peanut butter or after oral gavage of [¹³C]-mannitol (n = 8), grouped by administration strategy. F) The percentage of administered mannitol that was oxidized to ¹³CO₂ during the course of the assay. G) The total amount of tracer oxidized (mean with individual subject lines plotted) during the course of the assay. H) T_1/2. A statistically significant difference was found between administration strategies. I) ¹³CO₂ detected (mean with SEM ribbon) after oral gavage of [¹³C]-mannitol (n = 7) J) 100 and 145 minutes after gavage of a solution of FD&C1 blue dye, mice were sacrificed and the location of the dye front in the gastrointestinal tract was located. J) graphical representation of the location of the dye after oral gavage of dye. Statistical significance determined by two-tailed unpaired t-test; p ≥ 0.05 (NS), p < 0.001 (***).

Functional modulation of oral-cecal transit time

To test whether the system could detect pharmacologically induced changes in OCTT, mice received agents known to alter gastrointestinal motility. Loperamide hydrochloride (1 mg·kg⁻¹; Sigma, L4762) was prepared in 100% ethanol and saline and administered intraperitoneally 1 hour before ¹³C-mannitol gavage. Semaglutide sodium salt (0.06 mg·kg⁻¹·day⁻¹; BOC Sciences, B0084–007194) was prepared in saline and administered subcutaneously daily for one week before and during ¹³C-mannitol testing in mice on a 60% high-fat diet (Research Diets, D12492i). Breath sampling and data analysis were conducted as previously described. For dye transit studies, dye was administered by gavage to mice, and their gastrointestinal tracts were examined at predetermined time points to assess the location of the dye front. Measurements of metabolic rate were collected in the Promethion High Definition Multiplexed Respirometry System (Sable Systems) for the semaglutide study as well. Following a 7-day injection schedule, mice were monitored with this system for 36 hours. Measurement of VO₂, VCO₂, RER, locomotor and ambulatory activity, food intake was done at 23 ± 0.1°C. Energy expenditure was calculated with the Weir equation (Weir, 1949).

Semaglutide Metabolic Parameters Assay

To show the effect of semaglutide and the rapid changes in metabolism after semaglutide treatment, 16-week-old male mice on chow diet were injected daily with either PBS or Semaglutide sodium salt subcutaneously for 5 days. Measurements of metabolic rate were collected in the Promethion High Definition Multiplexed Respirometry System (Sable Systems, Las Vegas, Nevada, USA) during this injection schedule, mice were monitored within this system. Measurement of VO₂, VCO₂, RER, locomotor and ambulatory activity, food intake were done at 23 ± 0.1°C. Three out of 16 mice in the PBS group were excluded for poor quality data. Two out of 16 mice in the semaglutide group were excluded for poor quality data.

Data Analysis

Data were processed and analyzed using R (version 2024.12.1). Data Preparation: Raw data were exported from CalR (29) and imported into R with units of VCO₂ as mL*hr⁻¹ and δ¹³C in per mil (‰) relative to a reference standard (30). Feeding start times were aligned across subjects using a pre-recorded offset. An adjusted experimental time variable was calculated by subtracting each subject's offset from their recorded time points. Baseline Correction: Baseline CO₂ production and ¹³C-enrichment (C13) were calculated over a defined pre-treatment window for each subject. Baseline-subtracted (adjusted) values were computed for both variables. Filtering and Quality Control: Data were filtered to focus on the experimental period of interest (e.g., 0–600 minutes post-feeding). Visual inspection of ¹³C × VCO₂ plots for individual subjects was used to identify and exclude data artifacts or outliers.

Normalization and Calculations:

$${\mathrm{R}}_{\text{Standard}}=0.0112372,\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{p}\mathrm{o}\mathrm{n}\mathrm{d}\mathrm{s}\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{h}\mathrm{e}{}^{13}\mathrm{C}/{}^{12}\mathrm{C}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{o}\mathrm{f}\mathrm{n}\mathrm{a}\mathrm{t}\mathrm{u}\mathrm{r}\mathrm{a}\mathrm{l}\mathrm{l}\mathrm{y}\mathrm{o}\mathrm{c}\mathrm{c}\mathrm{u}\mathrm{r}\mathrm{r}\mathrm{i}\mathrm{n}\mathrm{g}\mathrm{P}\mathrm{e}\mathrm{e}\mathrm{D}\mathrm{e}\mathrm{e}\mathrm{B}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{m}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{e}\left(30\right)$$

$${}^{13}\mathrm{C}\mathrm{A}\mathrm{t}\mathrm{o}\mathrm{m}\mathrm{P}\mathrm{e}\mathrm{r}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\left(\mathrm{A}\mathrm{P}\right)=100/\left(1/\left(\left(\left({\delta }^{13}\mathrm{C}/1000\right)+1\right)\mathrm{*}{\mathrm{R}}_{\text{Standard}}\right)+1\right)$$

$$\mathrm{A}\mathrm{t}\mathrm{o}\mathrm{m}\mathrm{F}\mathrm{r}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{a}\mathrm{l}\mathrm{E}\mathrm{x}\mathrm{c}\mathrm{e}\mathrm{s}\mathrm{s}\left(\mathrm{A}\mathrm{F}\mathrm{E}\right)=(\mathrm{A}\mathrm{t}\mathrm{o}\mathrm{m}\mathrm{\%}\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}-\mathrm{A}\mathrm{t}\mathrm{o}\mathrm{m}\mathrm{\%}\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l})/100$$

$${\mathrm{C}\mathrm{O}}_2\mu \mathrm{m}\mathrm{o}\mathrm{l}={\mathrm{C}\mathrm{O}}_2\mathrm{m}\mathrm{L}\mathrm{*}{\mathrm{h}\mathrm{r}}^{-1}\mathrm{*}0.6945$$

$$\mathrm{L}\mathrm{a}\mathrm{b}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{d}\mathrm{C}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n}\mathrm{o}\mathrm{x}\mathrm{i}\mathrm{d}\mathrm{i}\mathrm{z}\mathrm{e}\mathrm{d}({\mathrm{n}\mathrm{m}\mathrm{o}\mathrm{l}}^{\mathrm{*}}{\mathrm{m}\mathrm{i}\mathrm{n}}^{-1})={\mathrm{C}\mathrm{O}}_2\mu \mathrm{m}\mathrm{o}\mathrm{l}\mathrm{*}{\mathrm{m}\mathrm{i}\mathrm{n}}^{-1}\mathrm{*}\mathrm{A}\mathrm{F}\mathrm{E}\mathrm{*}1000$$

Cumulative sums of CO₂ production and tracer oxidation were calculated. Total tracer oxidation was determined based on dosing information. The percentage of tracer oxidized over time was calculated per subject.

Data Cleaning: strongly negative ratios during cage adjustments (such as during tracer administration) were removed prior to plotting and statistical analysis. Data Aggregation: Separate datasets for each condition (e.g., low vs. high mannitol dose) were processed and combined for comparative analysis.

Statistical Analysis:

The time to 50% tracer oxidation was calculated per subject. Group comparisons were performed using Welch’s t-test (unequal variances). Sample sizes, means, standard deviations, standard errors, t-statistics, degrees of freedom, p-values (two-tailed), 95% confidence intervals, pooled standard deviations, and Cohen’s d effect sizes were computed. Boxplots were generated to visualize group differences in 50% oxidation times, annotated with significance markers.

Results:

Detection of gastric emptying through ¹³C-octanoic acid oxidation.

We sought to validate our experimental system using the established gastric emptying breath test in mice to measure transit time to the duodenum. We fed ¹³C-octanoic acid-containing food to individually housed, unanesthetized, freely moving mice with ad libitum access to food and water to measure the rate of gastric emptying. After consumption of a meal containing ¹³C-octanoic acid, rates of ¹³C-CO₂ detection began to peak 50 minutes post-administration (Figure 1A, n = 3). Roughly 2.5% of the consumed tracer was oxidized to exhaled ¹³CO₂ by the end of the observation period (Figure 1B). We next evaluated the timing of gastric emptying following administration of a prokinetic agent, metoclopramide (MCP), used to treat gastroparesis and as an antiemetic. In patients, MCP acts as an enteric dopamine receptor (D2) antagonist to accelerate movement of food through the stomach and proximal small intestine. We examined the effect of MCP versus saline (n = 4 per group) on kinetics of ¹³C-octanoic acid oxidation. After administration of the tracer, we measured ¹³CO₂ production for the next 200 minutes. We observed a sharp peak of ¹³CO₂ occurring within 30 minutes, which is indicative of accelerated gastric emptying (Figure 1D). In this interval, similar amounts of total tracer were oxidized to ¹³CO₂ in both control and MCP treated groups (Figure 1E). To quantify the rate of gastric emptying, we measured the fraction of tracer oxidized within the 200-minute experimental monitoring period. The time at which 50% of tracer was oxidized, T₅₀_% was determined for mice in both groups. The T₅₀_% of MCP-treated mice was significantly lower (p < 0.01) than the T₅₀_% of saline-treated mice (Figure 1F–G). As we can effectively monitor ¹³C-octanoic acid oxidation, in a multiplexed, non-invasive system we next sought to identify a tracer that would be absorbed or metabolized at a more distal site in the gastrointestinal tract. Statistical significance determined by two-tailed unpaired t-test.

Figure 1: [13C]-octanoic acid breath test in mice.

Mice voluntarily ate a meal with trace amounts of [¹³C] octanoic acid and the production of ¹³CO₂ was detected (mean with SEM ribbon) (n = 3). A) Rate of ¹³CO₂ production, plotted in nmol·min⁻¹. B) Cumulative production of ¹³CO₂ (mean with individual subject lines plotted). C) The anatomical representation of the location of octanoic acid oxidation in a murine model. D-G) Intraperitoneal injection of saline or metoclopramide 20 minutes before oral gavage of [¹³C] octanoic acid. D) Rate of ¹³CO₂ production (mean with SEM ribbon, n = 4). E) Cumulative production of ¹³CO₂. F) The percentage of total tracer oxidized (mean with individual subject lines plotted) during the 200-minute monitoring period. G) The time at which 50 percent of the tracer was oxidized (T_1/2). Tracer 50-percent oxidation time was defined as the time at which half of the total tracer oxidized during the course of the assay, was oxidized. Significant differences were found in the 50-percent oxidation times between the saline and MCP treatment groups. Statistical significance determined by two-tailed unpaired t-test; p < 0.01 (**).

Mannitol for monitoring oral-cecal transit time (OCTT):

We assessed the kinetics of ¹³C-mannitol tracer oxidation at two sub-milligram doses by measuring rates of ¹³CO₂ production. Group sizes were n = 7 and n = 8 for the 0.31 mg and 0.46 mg groups, respectively. We aimed to characterize both the magnitude and temporal profile of ¹³CO₂ appearance. Notably, ¹³CO₂ generated from mannitol appeared later than that produced from octanoic acid (Figure 2A). Despite administering two different starting doses of mannitol, approximately 15% of the tracer from both groups was oxidized within 6 hours (Figure 2B). As expected, the higher mannitol dose resulted in greater total ¹³CO₂ output (Figure 2C). However, no significant difference in T₅₀% was observed between the two doses (Figure 2D).

Mannitol is widely used as a sugar substitute in processed foods and drinks and is generally well tolerated, indicating its suitability for administration in both food and water. To determine whether the route of administration influences tracer kinetics, we compared oral gavage of mannitol dissolved in water to voluntary consumption mixed with palatable food. For food consumption, we designed the experiment to introduce the tracer to mice with ad libitum access to food and water, as short-term fasting can significantly increase physical activity and metabolic rate in mice (28). Trace amounts of mannitol were mixed with palatable food or administered by oral gavage in water. Group sizes were n = 7 and n = 9 for the palatable food and gavage groups, respectively. We observed that liquid gavage of ¹³C-mannitol produced a faster time to peak oxidation when compared to food consumption (Figure 2E). Ultimately, the total fraction of oxidized tracer (~15%) was similar between administration routes (Figure 2F). The accelerated kinetics observed with gavage were also significant (p, 0.001) as is shown through the reduced T₅₀_% in this group (Figure 2H).

To further explore the sites of mannitol absorption, we tracked gastrointestinal transit using a dye administered by oral gavage (Figure 2I, n = 7). In our tracer studies, maximal ¹³CO₂ production occurred at 120 minutes post-administration. Accordingly, mice given the dye were sacrificed at 100 minutes and 145 minutes. Group sizes were n = 6 and n = 4 for the 100 minute and 145-minute groups, respectively. At 100 minutes, the dye had passed into the small intestine, but not the cecum, while at 145 minutes, it had passed the cecum and into the large intestine. These findings indicate that peak ¹³CO₂ production temporally coincides with tracer arrival in the cecum (Figure 2J). Statistical significance was determined by two-tailed unpaired t-test.

Microbial Metabolism of Mannitol:

While previous studies have shown that mannitol is not absorbed by mammals, we observed that approximately 20% of the administered mannitol tracer was converted to CO₂. To determine if this oxidation was due to microbial metabolism, we compared tracer oxidation in conventional mice and in mice depleted of commensal microbes using antibiotics, following administration of either ¹³C-mannitol or ¹³C-glucose (n = 4 per group). Oxidation of ¹³C-glucose was similar in both groups (Figure 3A–B). In contrast, antibiotic-treated mice exhibited no detectable ¹³CO₂ production following ¹³C-mannitol administration (Figure 3C–D), indicating that mannitol oxidation depends on the presence of intestinal microbes.

Figure 3: differential [¹³C]-mannitol oxidation in conventional and commensally depleted mice.

Mice were gavaged a solution containing [¹³C]-mannitol, [¹³C]-glucose, or water and the production of ¹³CO₂ was detected (n = 4). A) ¹³CO₂ detected (mean with SEM ribbon) after oral gavage of [¹³C]-mannitol (N = 4). B) percent of administered mannitol that was oxidized to ¹³CO₂ during the course of the assay. C) rate of ¹³CO₂ exhalation (mean with SEM ribbon) after oral gavage of [¹³C] glucose (N=4). D) percent of administered glucose that was oxidized to ¹³CO₂ during the course of the assay. E) Glycemia measured by implanted continuous glucose monitor after oral gavage of unlabeled-mannitol (N = 3).

Consistent with the literature, mannitol, like other sugar alcohols, is reported to neither increase nor decrease blood glucose levels, supporting their use as sweeteners for individuals with diabetes (31). Using continuous glucose monitoring, we observed that oral administration of mannitol at a standard dose used in a glucose tolerance test (2 mg·kg⁻¹) did not raise blood glucose levels over a period corresponding to peak ¹³CO₂ production (Figure 3E, n = 3 per group).

Functional measurements: Loperamide.

To evaluate the utility of non-invasive OCTT measurements, we performed several functional studies. In the first, we assessed the effect of loperamide, an FDA-approved agent known to reduce gut peristalsis on OCTT (32). Administration of loperamide resulted in a right shift in ¹³CO₂ production curves without affecting overall change in tracer absorption (Figure 4A–B, n = 8 per group). This corresponds to a 30-minute delay in time to T_1/2% (p < 0.001, one-tailed t-test; Figure 4C). In a separate cohort, we tracked the transit of dye through the GI tract following loperamide treatment or saline treatment (n = 3 per group). In control mice, the dye had reached the colon by 145 minutes, whereas in loperamide-treated mice the dye remained proximal to the cecum (Figure 4D).

Figure 4: Loperamide’s Effect on [13C]-Mannitol Oxidation.

Mice were treated with loperamide 1-hour prior to gavage of [¹³C]-mannitol, production of ¹³CO₂ was detected (n = 4). A) Rate of ¹³CO₂ production, plotted in nmol·min⁻¹. B) Cumulative production of ¹³CO₂ (mean with individual subject lines plotted). C-D) 1-hour after loperamide treatment, 145 minutes after gavage of a solution of FD&C1 blue dye, mice were sacrificed and the location of the dye front in the gastrointestinal tract was located. C) T_1/2, a significant increase in T_1/2 was found in the loperamide-treated group relative to control. D) Mice were injected with saline or loperamide, and 1 hour later gavaged with FD&C1 blue dye. After 145 minutes, mice were sacrificed and the position of the dye front in the gastrointestinal tract was determined. Graphical representation of the location of the dye front after GI-tract removal of mice gavaged with dye. Statistical significance determined by one-tailed unpaired t-test; p < 0.001 (***).

Functional measurements: Semaglutide.

Finally, we assessed the effect of the GLP1 receptor agonist (GLP1Ra) semaglutide on OCTT which is known to increase gut transit time in humans. In high-fat diet (HFD) fed mice, daily semaglutide treatment resulted in ~10% reduction in body weight (Figure 5A). Semaglutide strongly delayed OCTT and produced a right shift in the ¹³CO₂ production curve (Figure 5B). Semaglutide treatment reduced overall tracer absorption and delayed T_1/2% times by approximately 100 minutes, a two-thirds increase relative to controls (Figure 5C–D). Group sizes were n = 7 and n = 15 for the saline and semaglutide treated groups, respectively. Dye motility assays confirmed this delay with semaglutide significantly slowing dye progression through the GI tract (Figure 5E). RER, energy expenditure, total food consumption, and cumulative energy balance were significantly reduced immediately after the start of semaglutide treatment in comparison with PBS treated mice (Figure 5F–5I). Statistical significance was determined using two-tailed unpaired t-tests for comparisons between two groups or two-way ANOVA for comparisons across multiple time points and conditions. p > 0.001 (***)

Figure 5: semaglutide induced slow-down of mouse gut motility.

Mice were treated with semaglutide (0.06 mg·kg⁻¹·day⁻¹) for 7 days prior to gavage of [¹³C]-mannitol, production of ¹³CO₂ was detected (saline n = 7, semaglutide n = 15). A) body weights of mice acutely treated with semaglutide over a 4-day period, with day 0 being baseline (mean with SEM bar, saline n = 7, semaglutide n = 8). B) Rate of ¹³CO₂ production, plotted in nmol·min⁻¹, the dotted line represents the 100-minute mark during the assay. C) Cumulative production of ¹³CO₂ (mean with individual subject lines plotted). D) T_1/2, a significant increase in T_1/2 was found in the semaglutide-treated group relative to control. E) After daily injection of saline or semaglutide for 7 days, mice were gavaged with FD&C1 blue dye. After 100 minutes, mice were sacrificed and the position of the dye front in the gastrointestinal tract was determined. graphical representation of the location of the dye front after GI-tract removal of mice gavaged with dye. F) Respiratory exchange ratio (RER) of mice treated with either semaglutide or PBS over 4 days G) Energy expenditure (EE) of mice treated with either semaglutide or PBS treatment over 4 days H.) Cumulative energy intake (kcal) in the same group of mice I) Cumulative energy balance for mice treated with either semaglutide or PBS. Statistical significance was determined using two-tailed unpaired t-tests for comparisons between two groups or two-way ANOVA for comparisons across multiple time points and/or conditions. p > 0.001 (***)

Discussion

Here, we establish a non-invasive, high-resolution method to quantify OCTT in mice by monitoring microbially mediated oxidation of ¹³C-labeled mannitol to ¹³CO₂. This approach offers several critical advantages over existing methods, including the ability to perform continuous, longitudinal measurements in freely moving, unanesthetized mice without the need for restraint, terminal procedures, or radioisotopes. The ability to monitor stable isotope oxidation via breath in real time yields highly granular data on gastrointestinal transit dynamics, with applications for preclinical studies assessing genetic, dietary, microbial, and pharmacological influences on gut motility.

We first validated our system by replicating the well-characterized kinetics of the gastric emptying breath test via ¹³C-octanoic acid oxidation. Our data align with prior reports on the utility of the ¹³C-octanoic acid breath test in humans and rodents, supporting the reliability of our multiplexed detection system. Notably, our system detected accelerated gastric emptying with metoclopramide, confirming the assay’s sensitivity to pharmacologically induced changes in motility.

Building on this validation, we identified ¹³C-mannitol as a novel tracer that enables OCTT assessment. Unlike octanoic acid, mannitol is not absorbed by the host and requires microbial metabolism. Our data show that microbial depletion via antibiotics abolishes mannitol oxidation, confirming that the observed ¹³CO₂ reflects microbial activity rather than host metabolism. These findings are consistent with clinical observations that mannitol does not elevate blood glucose and is a largely non-metabolizable sugar alcohol by mammalian tissues. Our studies do not identify which microbial taxa are capable of mannitol metabolism, an important but unanswered question.

We further demonstrated the functional relevance of this method by pharmacologically altering gut motility. Loperamide treatment delayed the appearance of ¹³CO₂, consistent with delayed transit to the cecum, as corroborated by dye motility studies. Similarly, semaglutide caused a pronounced delay in OCTT in mice on a high-fat diet. These pharmacological challenges validate the sensitivity of our assay to physiologically meaningful changes in gut motility.

Importantly, the route of tracer administration, gavage versus voluntary consumption, altered the kinetics of mannitol oxidation without affecting the total amount oxidized. This flexibility allows the method to be adapted to a variety of experimental designs, including ad libitum administration strategies that minimize stress-related artifacts. It also enables the exploration of how different dietary compositions, such as high-fat, high-protein, high-carbohydrate, or fiber-rich diets, affect gut motility and transit dynamics.

Our approach addresses several limitations inherent to current OCTT methodologies. Terminal methods, while precise, require animal sacrifice at multiple time points and are thus resource-intensive and cross-sectional by nature. Other non-terminal approaches, like Carmine red assays, suffer from limited precision and the confounding effects of repeated handling. Breath-based stable isotope analysis circumvents these issues, offering continuous, non-invasive data collection that minimizes stress and variability.

There are several limitations to consider with this method. While mannitol oxidation provides an approximate marker of cecal arrival, its reliance on microbial metabolism may introduce variability in microbiota-depleted or dysbiotic animals. Alterations in gut permeability may also affect the results. Moreover, while we demonstrated equivalence in total oxidation between doses, higher resolution studies may be needed to fully understand dose-dependent kinetics, especially under pathological conditions.

Our work establishes a scalable, non-invasive method to quantify OCTT with high temporal resolution in mice, opening new avenues to investigate and visualize gut motility and host-microbe interactions. This method enhances the translational potential of murine models by reducing animal use, limiting experimental confounds, and enabling dynamic assessment of gastrointestinal function in response to genetic, environmental, and pharmacologic factors.