Botulinum neurotoxin (available commercially as BOTOX) has been used successfully for treatment of several neuromuscular disorders, including blepharospasm, dystonia, spasticity, and cerebral palsy in children. Our data demonstrate that injection of Botox into the proximal intestinal wall of diet-induced obese (DIO) mice induces weight loss and reduces food intake. This was associated with amelioration of hyperglycemia, hyperlipidemia, and significant improvement of glucose tolerance without alteration of energy expenditure. We also observed accelerated gastrointestinal transit and significant reductions in glucose and lipid absorption, which may account, at least in part, for the observed weight loss and robust metabolic benefits, although possible systemic effects occurring as a consequence of central and/or peripheral signaling cannot be ignored. The observed metabolic benefits were found to be largely independent of weight loss, as demonstrated by pair-feeding experiments. Effects lasted ∼8 weeks, for as long as the half-life of Botox as reported in prior rodent studies. These results have valuable clinical implications. If the observed effects are translatable in humans, this approach could lay the foundation for therapeutic approaches geared toward robust and sustained weight loss, mimicking some of the benefits of bariatric operations without its cost and complications.
Introduction
Obesity currently affects more than one in three adults in the U.S. (1). An alarming consequence is the staggering increase in type 2 diabetes (T2D), which has emerged as the seventh leading cause of death (2–4). Lifestyle interventions have proven to be only modestly effective and rarely sustainable. Bariatric surgery, particularly Roux-en-Y gastric bypass (RYGB) is currently the most effective treatment for robust and sustained weight loss and results in remarkable improvements in T2D (5). RYGB reduces stomach volume and diverts nutrient flow from the proximal intestine, which is a major site for nutrient sensing and absorption. However, operations involve extensive and permanent rearrangement of the gastrointestinal (GI) tract, can be associated with serious complications, and are expensive. They are estimated to reach <2% of all eligible individuals.
The importance of the gut in nutrient metabolism and energy homeostasis is highlighted by the remarkable metabolic benefits of bariatric surgery (6,7). The ability to target the gut to treat obesity and diabetes without surgical intervention would represent a major therapeutic breakthrough. Botulinum neurotoxins are among the most potent anticholinergic agents (8–10). The most potent serotype A (available commercially as BOTOX) is approved by the U.S. Food and Drug Administration for treatment of several neuromuscular conditions, including blepharospasm, dystonia, spasticity, and cerebral palsy in children (11–14). Botox possesses several unique pharmacological properties—it is specific, very potent, has limited diffusion from the injection site (10,14), and is reversible. Its potential ability to impede intestinal wall contractions and consequently to affect nutrient absorption, food intake, glucose tolerance, and weight loss has never been explored.
Our study shows that Botox injected into the duodenal wall of diet-induced obese (DIO) mice induces weight loss, reduces food intake, ameliorates hyperglycemia and hyperlipidemia, and improves glucose tolerance, without altering energy expenditure (EE). Glucose and fat absorption were significantly reduced, and transit in the duodenum and proximal jejunum was found to be accelerated. We speculate that faster transit may possibly be contributing to reduced absorption due to less exposure time to luminal surface of enterocytes. These effects are sustained until ∼8 weeks after Botox administration, after which the phenotype and metabolic improvements are reversed. We demonstrate, for the first time, that botulinum administration in the proximal intestinal wall leads to weight loss and improves glucose tolerance. If the effects observed in our mouse model are translatable in humans, this approach could provide a new treatment paradigm for obesity with or without T2D.
Research Design and Methods
Mice and Surgical Procedures
All experiments and surgical preparations were performed according to protocols approved by the Vanderbilt University Medical Center Institutional Animal Care and Use Committee. The mice remained under the care of the Division of Animal Care in compliance with National Institutes of Health guidelines and the Association for Assessment and Accreditation of Laboratory Animal Care. Male C57BL/6J mice were purchased from The Jackson Laboratory and housed at 23°C on a 0700–1900-h light cycle and were fed a high-fat diet (HFD; 60% kcal from fat; Research Diets, New Brunswick, NJ), starting at 6 weeks of age for 12–14 weeks prior to surgery. Mice were anesthetized by isoflurane, and surgery was performed under a dissection microscope. A 2 cm incision was made at the upper abdomen to expose the intestine. Saline or Botox loaded into 25 μL Hamilton syringes (3.0 units/kg body wt) was injected into the medial side of the proximal small intestinal wall at three sites, within 3 cm distal to the pyloric sphincter. Following injection, the abdomen was sutured back, and mice were single housed until the end of the experiment.
Body Composition
Body composition (i.e., fat mass, lean body mass, and water content) was measured before and 1, 4, and 10 weeks after treatments by quantitative magnetic resonance using an Mq10 nuclear magnetic resonance analyzer (Bruker Optics, Billerica, MA). Fat and muscle mass were calculated in grams.
Oral Glucose Tolerance Test
Mice were fasted for 5 h prior to oral glucose tolerance test (OGTT) at 2, 4, or 8 weeks after surgery. Blood was sampled from the tail vein before and at 5, 15, 30, 45, 60, 75, 90, and 120 min after an oral gavage of 20% dextrose solution to achieve a dosage of 2 mg/g body wt. Blood glucose levels (mg/dL) were measured using a blood glucose meter (SureStep, LifeScan).
Lipid Tolerance Test
Mice were fasted overnight and then dosed with an oral gavage of lipid emulsion containing 20% soybean oil (Intralipid 20%; Sigma-Aldrich, St. Louis, MO) at 10 µL/g body wt. Blood was collected from tail vein bleeds at baseline (before gavage) and at hourly intervals between 1 and 5 h. Serum levels of nonesterified fatty acids (NEFAs) and triglycerides (TGs) were quantified using kits from Wako (Suwanee, GA) and Infinity Triglycerides from Fisher Diagnostics (Middletown, VA), respectively.
Intestinal Glucose and Lipid Uptake
Intestinal glucose uptake was determined using the radiolabeled glucose analog, 3-O-methyl-glucose (3-OMG), as described earlier (15). 3-OMG is transported by intestinal glucose transporters, sodium–glucose cotransporter 1 (SGLT1) and GLUT2, with an affinity like that of d-glucose but is nonmetabolizable. Fresh tracer, 3H was prepared for gavage each day. 3-[3H]OMG (∼5 μCi/mouse) was spiked into a 20% dextrose solution and provided via oral gavage. Blood was sampled from the tail vein at baseline and then at 5, 15, 30, 45, 60, 90, and 120 min after gavage. The rate of 3-OMG appearance in the blood was used as a surrogate for intestinal uptake. Plasma was isolated by centrifuging at 1,000g for 10 min at 4°C and deproteinized using a 1:1 solution of BaSO4 and Zn(OH)2. Tracer appearance (disintegrations per minute) was measured using scintillation counting (LS 6500; Beckman Coulter, Brea, CA) and expressed as a percentage of total disintegrations per minute administered to each mouse.
Intestinal lipid was determined using [9,10-3H(N)]oleic acid tracer ([3H]OA) (16). Mice were fasted overnight, and an intraperitoneal injection of Triton WR1339 (0.5 mg/kg) in saline (1:6 v/v) was delivered ∼30 min before gavage to inhibit plasma lipoprotein lipase. Olive oil spiked with [3H]OA (∼5 μCi/mouse) was provided via oral gavage. Blood was sampled from the tail vein at baseline, 30 min, and at hourly intervals for the next 4 h (0, 30, 60, 120, 180, and 240 min) after gavage. Plasma was isolated and deproteinized, and tracer appearance was measured using a scintillation counter, as described above.
Total Gut Transit Time and Intestinal Transit
Total gut transit time was measured using carmine red dye, which is not absorbed from the lumen. A solution of carmine red (10 µL/g body wt; 6%; Sigma-Aldrich) suspended in 0.5% methylcellulose (Sigma-Aldrich) was administered by oral gavage. The time at which the gavage was administered was recorded as t0. Fecal pellets were monitored for the presence of red color. Total gut transit time was considered as the interval between t0 and the time of first appearance of red color in the feces.
Intestinal transit was measured using FITC-conjugated dextran (70-kDa FITC-dextran; Sigma-Aldrich), which does not cross the intestinal epithelium (17). Saline- or Botox-treated mice (n = 5–8/group) were fasted overnight, and 5 mmol/L FITC-dextran was administered by oral gavage. Mice were euthanized at 0, 15, 30, 45, 60, 90, 120, 180, 240, and 360 min later, and the small intestines were collected for analyses of FITC-dextran measurements by fluorometry (excitation, 490 nm; emission, 530 nm). Serial dilutions of FITC-dextran in PBS were used for generation of standard curve. Experiments were performed within 3 weeks after saline or Botox injections.
EE by Indirect calorimetry
EE was determined by indirect calorimetry using a system consisting of 16 separate metabolic cages equipped for the continual monitoring of ambulatory activity and ad libitum access to the HFD and water (Promethion; Sable Systems, Las Vegas, NV). Oxygen, carbon dioxide, and water vapor levels were constantly monitored, and temperature and humidity were tightly regulated (GA3, Sable Systems). The incurrent airflow rate was set at 3,000 mL ⋅ min−1 (FR8, Sable Systems). VO2 and VCO2 were measured for each mouse at 10-min intervals for 1 min. Respiratory quotients were calculated as the ratio of VCO2 over VO2. EE was calculated using the Weir equation.
Data acquisition and instrument control were coordinated by MetaScreen 2.2.49, and the obtained raw data were processed using ExpeData 1.7.30 (Sable Systems) using an analysis script detailing all aspects of data transformation. The EE ANCOVA analysis done for this work was provided by the National Institute of Diabetes and Digestive and Kidney Diseases Mouse Metabolic Phenotyping Centers (www.mmpc.org) using their Energy Expenditure Analysis page (https://www.mmpc.org/shared/regression.aspx).
Fecal Lipid Analyses
Feces were collected over a 48-h period. Feces (100 mg) were extracted with 3:1 chloroform/methanol, and the resulting lipid extract was evaporated to dryness and weighed (18). Free fatty acids (FFAs) were analyzed by the NEFAs analysis kit (Wako Life Sciences, Richmond, VA). TG and cholesterol content were analyzed using Infinity reagents (Thermo Scientific, Middletown, VA).
Gut Hormone Measurements
Gut hormones were measured in Botox- or saline-treated mice after 4–6 weeks of injections (n = 6–8 per group per time point), corresponding with maximal weight loss and reductions in food intake. Mice were fasted overnight. and a liquid mixed meal (Ensure Nutritional Shake, 10 μL/g body wt; 220 kCal, 6 g fat, 32 g carbohydrates, 16 g glucose, 9 g protein per 8 oz) was administered via oral gavage. Blood was collected via tail bleeding at baseline (before gavage) and 15, 30, 60, 90, and 120 min after oral gavage (n = 6–8 mice per group per time point). The plasma hormones gastric inhibitory polypeptide (GIP), active glucagon-like peptide 1 (GLP-1), leptin, and peptide tyrosine tyrosine (PYY) were measured using the Multiplex Assay (MILLIPLEX Mouse Metabolic Hormone Magnetic Bead Panel, MilliporeSigma). Cholecystokinin (CCK) was measured using the RayBio Mouse Cholecystokinin EIA Kit (RayBiotech, Peachtree Corners, GA).
Immunoblot Analyses
Tissues were homogenized in lysis buffer and pelleted. Protein concentrations of the supernatant were analyzed, and equivalent amounts of protein were loaded onto a polyacrylamide gel. Primary antibodies (synaptic vesicle-associated protein [SNAP] 25, 1:1,000; Thermo Fisher, no. 7895) were diluted in Li-Cor (Lincoln, NE) blocking buffer and incubated overnight at 4°C. Appropriate fluorescently labeled secondary antibodies (Li-Cor) were diluted 1:10,000 and incubated in blocking buffer at room temperature for 1 h with shaking. After three rinses with PBS solution, the membrane was scanned using the Odyssey Infrared Imaging System (Li-Cor).
Statistics
All data are expressed as mean ± SEM. Unless otherwise indicated, one-way ANOVA with the Dunn posttest was used to compare three or more groups, while the Student t test (paired, two tailed) was used for pair-fed experiments. Repeated-measures ANOVA was used for OGTT and lipid tolerance tests that involved repeated sampling from the same mouse. All statistical analyses were performed using Prism 8.0 software (GraphPad Software, San Diego, CA). The threshold of statistical significance was set at P < 0.05.
Data and Resource Availability
All data generated or analyzed during this study are included in the published article and in its online supplementary files.
Results
Botulinum Injection Into the Proximal Intestinal Wall of DIO Mice Induces Weight Loss and Reduces Food Intake
Botox was injected into the proximal intestinal wall of DIO mice at three sites (∼1 cm apart) within 3 cm distal to the pyloric sphincter (Supplementary Fig. 1A). Sites injected with Botox showed significantly lower SNAP-25 expression compared with saline controls (Fig. 1A). SNAP-25 is one of the intracellular targets of Botox, and previous reports have shown proteolysis of SNAP-25 in the skeletal muscles of rodents treated with Botox (8,10,19–23). Although SNAP-25 proteolysis suggests possible blockade of cholinergic transmission, contribution of other plausible mechanisms and signaling pathways regulated by SNAP-25, such as voltage-gated calcium signaling, calcium-dependent neurotransmitter release, and presynaptic calcium concentrations cannot be precluded.
A dose-response experiment over a period of 8 weeks (Fig. 1B) demonstrated that a Botox dose of 3.0 /kg body wt was optimal and was subsequently used for all studies. All mice lost ∼10% of their body weight within a week after surgery. However, saline-injected controls and those injected with <3.0 units/kg of Botox regained the lost weight in a week and reverted to their presurgical weight at the end of 4 weeks (Fig. 1B). Mice injected with ≥3 unit/kg body wt lost 17–20% of their body weight until 7 weeks after surgery. After 7 weeks, mice started regaining weight and reverted to their presurgical weight at the end of 10 weeks (Fig. 1B). Actions of Botox are reversible, and the half-life of Botox in skeletal muscle of mice has been reported to be ∼3 months (22,23). Our study shows that when injected in the proximal small intestinal smooth muscle, Botox effects last for ∼8 weeks. Body composition analyses showed that ∼99.5% of the total weight loss was attributable to a decrease in fat mass (Fig. 1C), as no changes were observed in lean mass or water content (data not shown). Lean and fat mass contents were comparable in saline- and Botox-injected mice after 1 week of injections. After 4 weeks, fat mass in Botox-injected mice was 48.6% lower than in the saline controls (Fig. 1C). It did not differ between the groups following the reversal of the effects of Botox at 10 weeks.
Weight loss was associated with reductions in food intake (Fig. 1D). Botox-injected mice had ∼25–30% lower food intake than saline controls for up to 8 weeks after injections and were comparable to controls by the end of 10 weeks (Fig. 1D). To determine whether Botox-induced weight loss is entirely attributable to reduced food intake, we performed pair-feeding experiments, wherein saline- and Botox-injected mice were pair-fed the HFD for 6 weeks. Saline-treated controls in the pair-feeding were staggered by 1 day and consumed the same amount of food as the Botox-treated partner the day prior. As expected, saline controls lost ∼10% of their body weight during 6 weeks of pair-feeding (Fig. 1E). However, body weights of Botox-injected partners were significantly lower throughout the 6 weeks, demonstrating that Botox-induced weight loss is dependent on mechanisms including, but not limited to, reduced food intake.
Botox-Induced Weight Loss Is Not Due to Alterations in EE
Average EE of saline- and Botox-treated mice was measured after 1, 4, and 6 weeks of treatment (Fig. 2). EE of Botox-treated mice after 1 week was not significantly different from saline-treated controls (Fig. 2A). Despite ∼20% decrease in body weight of Botox-treated mice after 4 weeks, EE was comparable to that of saline-treated controls (Fig. 2B). To determine whether body weight differences affected EE, multiple linear regression analyses was performed using Energy Expenditure Analysis (https://www.mmpc.org/shared/regression.aspx) provided by the National Mouse Metabolic Phenotyping Center. Regression analyses demonstrated that interaction of body weight in EE outcome was not significant (P = 0.505) (Fig. 2C). Similar results were observed in EE computed after 6 weeks of treatment. Although EE (kcal/day) of Botox-treated mice was slightly lower (although nonsignificant) than saline controls after 6 weeks (Fig. 2D), regression analyses showed that differences in body weight at 6 weeks also did not affect the outcome (P = 0.459) (Fig. 2E). Frequency of locomotor activity measured by pedestrian meters, proportion of time walked, and all other movements of saline- and Botox-treated groups were comparable in both light and dark cycles (data not shown). These data demonstrate that Botox-mediated weight loss is not attributable to EE changes.
Significant Amelioration of Hyperglycemia and Improvement of Glucose Tolerance in DIO Mice
Saline-injected HFD mice were obese and hyperglycemic (Fig. 3A). Fasting plasma glucose levels in Botox-injected mice were significantly lower than in saline-injected controls at 1 week (−32.4%), 4 weeks (−39.8%), and 7 weeks (−40.1%) after injections (Fig. 3A). Importantly, fasting glucose levels were significantly lower before weight loss when body weights of control- and Botox-injected mice were similar (week 1). Thus, the improvement in glucose levels is not entirely dependent on weight loss, although it is potentiated by continued weight loss as observed during 4 and 7 weeks after injections. Fasting insulin levels followed the same trend (Fig. 3B). Botox-injected mice had significantly lower insulin levels than saline controls (week 1: −30.4%; week 4: −32.4%; week 7: −24.3%). Insulin resistance (Fig. 3C), as calculated by the HOMA of insulin resistance (HOMA-IR) equation was significantly lower during 1–7 weeks after injections (week 1: −50%; week 4: −64%; week 7: −61%). Fasting glucose and insulin levels and HOMA-IR at 10 weeks after injections were comparable to those of saline-injected controls, which corroborates with effects on weight loss and food intake.
As noted, food intake in Botox-injected mice was lower than that of saline controls, with the decrease becoming significant starting at week 2 postinjections. To dissect the role of reduced food intake on Botox-mediated improvements in hyperglycemia, we assessed glucose levels in pair-fed saline- or Botox-treated mice. Plasma glucose (Fig. 3D), insulin (Fig. 3E), and HOMA-IR scores (Fig. 3F) of Botox-treated partners were significantly lower than those of saline controls after 1 week of pair-feeding (and treatment), despite similar body weights and food intake. Further loss of body weight observed after 4 weeks of Botox only led to mild improvements (Fig. 3D–F). Thus, Botox-mediated improvements in hyperglycemia and hyperinsulinemia are not entirely dependent on weight loss and/or reduced food intake.
A similar trend was observed in glucose tolerance (Fig. 4). Saline- and Botox-treated mice, with or without pair-feeding, were gavaged with 2 mg/kg body wt of dextrose, and blood glucose levels were measured at baseline and 15, 30, 60, 90, 120, and 180 min later. All DIO mice tested before saline/Botox injections and before the pair-feeding regimen showed impaired glucose tolerance (week 0) (Fig. 4A and B). Saline-injected mice that were not subjected to pair-feeding continued to show impaired glucose tolerance after 1 week of injections (Fig. 4C and D) that progressively worsened over 4 weeks (Fig. 4D and E). We speculate that this may be in part, attributable to continued weight gain observed in this group. On the other hand, pair-fed saline controls showed modest improvement in glucose tolerance after 4 weeks (Fig. 4E and F). In contrast, robust improvement in glucose tolerance was observed in Botox-treated mice (Fig. 4E and F) that far outweighed those observed in their saline-injected partners. At the end of 4 weeks of pair-feeding, glucose area under the curve (AUC) in Botox-injected partners was ∼22% lower than their saline counterparts. Additionally, despite the continued weight loss observed in Botox-treated mice after 4 weeks, glucose tolerance did not improve further (Fig. 4E and F). Thus, taken together, these data demonstrate that the Botox-mediated effects on glucose tolerance are not entirely dependent on weight loss. Weight loss observed in pair-fed saline counterparts did not achieve as robust improvements in glucose tolerance as that observed in Botox-treated mice (with or without pair-feeding).
Amelioration of Hyperlipidemia and Improvement of Lipid Tolerance in DIO Mice
Botox-injected mice showed significant improvement in plasma lipid profile (Fig. 5). Fasting plasma TG levels were 22–28% lower than in saline-injected controls at 1, 4, and 7 weeks after injections (Fig. 5A). Fasting FFA (−47% and −31%) (Fig. 5B) and cholesterol levels (−42% and −48%) (Fig. 5C) were significantly lower than in saline controls at 4 and 7 weeks after injections, respectively. Like the trend observed in glucose tolerance, Botox-mediated improvements in plasma lipid levels were reversed after 10 weeks of injections. Next, we performed the lipid tolerance test to determine effects of Botox on an acute lipid challenge. As shown in Fig. 5D, fasting plasma TG as well as the lipid-induced surge in plasma TG levels was reduced in the Botox-treated mice throughout the 5 h of the study. The TG AUC of Botox-treated mice was significantly lower (−42.6%) (Fig. 5E), indicating improved lipid tolerance.
Plasma TG, FFA, and cholesterol levels of Botox-treated partners in the pair-feeding experiment (Fig. 6) followed similar trends to those described above for glucose and insulin. Plasma TG levels in Botox-treated partners were ∼30% lower than saline controls at 1 and 4 weeks of pair-feeding (Fig. 6A). Plasma FFAs were 45% lower after 4 weeks of pair-feeding (Fig. 6B). Similarly, plasma cholesterol levels were significantly lower than in saline controls (Fig. 6C). Thus, at comparable body weights and food intake to the saline controls, the Botox-injected mice at 1 week showed much more significant improvement in glucose and lipid tolerance, hyperglycemia, and hyperlipidemia. Additional loss of body weight did not enhance Botox-mediated improvements.
Metabolic Improvements Are Associated With Reduced Glucose and Fat Absorption
Given the evidence of improved glucose and lipid tolerance, we speculated that intestinal nutrient absorption might be reduced in Botox-injected mice. Effects of Botox on glucose absorption were tested by adding [3H]3-OMG tracer to the glucose administered during the OGTT. 3-OMG is absorbed across intestinal brush borders like glucose. The percentage of [3H]3-OMG appearing in the circulation increased with time in both groups with the increase in intestinal absorption. Levels of circulating [3H]3-OMG in Botox-treated mice were significantly lower than in the saline controls beginning at 30 min after gavage until the end of the experiment at 120 min (Fig. 7A) and the [3H]3-OMG AUC was 54% lower (Fig. 7B). These data showed that Botox injection into the duodenum reduces glucose absorption from the intestine.
We next determined whether Botox limits fat absorption. We first measured fecal total lipid content by chloroform-methanol extraction. We found that Botox-treated mice had significantly higher fecal lipid output at 1, 4, and 7 weeks after injections (Fig. 7C). Next, we identified the lipid components most impacted by Botox. As shown in Fig. 7D–F, fecal TG, FFA, and cholesterol levels from Botox-treated mice were all significantly higher than in saline-treated controls. Notably, fecal TG content of Botox-treated mice was approximately threefold higher at week 1 and twofold higher at weeks 4 and 7 (Fig. 7D). Fecal FFA content was 2.5-fold higher in Botox-treated mice at week 1, and ∼75% higher than that of saline controls at 4 and 7 weeks after treatment (Fig. 7E). Fetal cholesterol content increased by approximately twofold at 1 and 4 weeks and was ∼30% higher at 7 weeks after Botox injections (Fig. 7F). Thus, Botox injection into the proximal intestinal wall increases fat excretion, which suggests possibly reduced fat absorption. To directly confirm this, we measured [3H]OA counts in the peripheral circulation after gavaging olive oil supplemented with [3H]OA tracer. As shown in Fig. 7G, the percentage of [3H]OA recovered in the plasma of Botox-treated mice was significantly lower than in saline controls after 2 h of gavage through the end of the experiment. The [3H]OA AUC of Botox-treated mice was 62% lower than in controls (Fig. 7H). Thus, these data demonstrate that proximal intestinal Botox injections reduce glucose and fat absorption—mechanisms known to set in a negative energy balance leading to weight loss and improved tolerance to glucose and fat. A similar trend was observed in pair-fed saline- or Botox-treated groups (Supplementary Fig. 2). Fecal lipid output in the Botox pair-fed partner was more than twofold higher than in the saline control after 1 and 4 weeks of injections (Supplementary Fig. 2A), with significant increase noted in TG (Supplementary Fig. 2B), FFA (Supplementary Fig. 2C), and cholesterol (Supplementary Fig. 2D) after 4 weeks of injections. Therefore, differences in food intake of body weights consequent upon Botox injections did not impact fecal lipid output.
Reduced Nutrient Absorption Is Associated With Faster GI Transit
Accelerated transit has been reported to be associated with reduced absorption of nutrients, including that of glucose and fat (24,25). We first measured total GI transit using the carmine dye method and first detected red carmine color in the feces of saline-treated mice at ∼6 h (Fig. 8A). A robust decrease in transit time (∼110 min) was observed in Botox-treated mice, indicating that overall GI transit is accelerated. We next determined the section(s) of the intestine targeted by the accelerated transit by gavaging the mice with FITC-dextran, collecting the intestines at 15, 30, 45, 60, 90, 120, 180, 240, and 360 min later, and sectioning the tissue into four 8-cm segments (sections (S) 1 to S4 from duodenum to cecum) for determination of fluorescence. As shown in Fig. 8B–F and Supplementary Fig. 3, Botox-induced accelerated transit was observed to be most robust in S1 and S2, which includes the duodenum and proximal jejunum. In section S1 of control mice, maximum fluorescence was observed at 30 min (Fig. 8B) and disappeared after 1 h (Supplementary Fig. 3A). In contrast, section S1 of Botox-treated mice reached maximum fluorescence by 15 min, markedly dropped by 30 min, and disappeared by 1 h (Fig. 8B). Sections S2–S4 showed similar trends, with peak fluorescence reaching earlier and disappearing earlier in Botox-treated mice compared with saline-treated controls (Fig. 8C–F and Supplementary Fig. 3B and C). Although fluorescence in S3 and S4 from Botox-treated intestines appeared faster, the peak was sustained for approximately the same amount of time as saline controls (∼30–45 min), which indicates that transit rates are comparable in the distal part of the small intestine. Taken together these data demonstrate that Botox accelerates transit in the proximal small intestine that includes the injection site at the duodenum and at least 12 cm of the proximal jejunum.
Increased PYY Release After Botox Treatment
Botox-treated mice had significantly lower food intake, which accounts, at least in part for the observed weight loss. To determine whether this is attributable to changes in hunger and satiety-regulating enteroendocrine hormones, we determined plasma levels of GIP, GLP-1, CCK, PYY, and leptin in response to a mixed-meal challenge. Hormones were measured using the Luminex Multiplex Assay. Insulin and C-peptide levels were also measured as part of the Multiplex Assay.
PYY levels were found to be significantly higher in Botox-treated mice (weeks 4–6 after treatment) compared with controls (Supplementary Fig. 4). No significant differences were observed in the AUCs for GIP (Supplementary Fig. 5A and B), GLP-1 (Supplementary Fig. 5C and D), CCK (Supplementary Fig. 5E and F), and leptin (Supplementary Fig. 7A and B). However, there was a trend toward increased GIP, GLP-1, CCK, and leptin levels in the Botox-treated mice. Levels of GLP-1 and PYY, secreted predominantly by L cells in the distal small intestine and colon, peaked later (60 min) than the proximal enteroendocrine hormones in the saline-treated controls. In contrast, early spikes (30 min) were observed for GLP-1 and PYY release in Botox-treated mice. Given the faster transit in the proximal small intestine after Botox treatment (Fig. 8) and the fact that nutrients reach the distal sections earlier, it is plausible that nutrient-stimulated enteroendocrine release by the distal enteroendocrine cells occurs earlier. However, unlike PYY, the increase in the AUC for GLP-1 did not reach statistical significance.
Discussion
Our study demonstrates that Botox injections into the proximal intestinal wall of DIO mice lead to weight loss associated with reduced food intake and significant improvements in hyperglycemia, hyperlipidemia, and glucose and fat tolerance. These metabolic improvements disappear by the end of 10 weeks, consistent with reversibility of Botox and its reported half-life in rodents. The mechanisms underlying these Botox-induced metabolic improvements appear to be multifactorial. They include, but are not limited to, reduction in food intake and reduced glucose and fat absorption, but do not appear to be dependent on changes in EE. Although reduced glucose and fat absorption associated with faster transit in the proximal small intestine is likely to contribute significantly to weight loss and observed metabolic benefits in this model, it cannot be concluded that this constitutes the principal underlying mechanism. It is entirely possible that in addition to local effects of Botox impacting the proximal intestine, Botox targets vagal neurotransmission, especially given that the proximal intestine is densely innervated by the vagus nerve. For example, the reduced food intake observed in this model could be a direct or indirect consequence of altered communication and feedback regulation loops operative along the gut-brain axis, investigation of which is beyond the scope of this study. Nevertheless, alteration of those pathways offers an attractive therapeutic avenue that warrants further investigation. The improvements are robust and, if translatable in humans, could potentially provide a groundbreaking approach for amelioration of obesity and T2D. In humans, Botox has been delivered endoscopically into the stomach wall and the pancreatic sphincter of Oddi for treatment of GI disorders (26–29). Thus, endoscopic delivery into the proximal intestinal wall is feasible and can be easily accomplished. If the observed benefits in mice are mimicked in humans, this approach could potentially lay the groundwork for a less invasive and relatively less expensive therapy that does not involve permanent and extensive realignment of the GI tract.
Botox injection into the proximal intestinal wall resulted in significant reduction in food intake, which could be a direct or indirect consequence of altered communication and feedback regulation loops operative along the gut-brain axis. The dense parasympathetic innervation of the proximal intestine, primarily by the vagus nerve, with 80% of this innervation being afferent, serves a primarily role in relaying information to the brain, potentially impacting the hunger and satiety-regulating agouti-related protein–positive neurons in the hypothalamus area. Additionally, the increase in PYY levels observed in the cohort injected with Botox may help promote increased and/or prolonged satiety. It is important to note, however, that the relative contribution of caloric restriction to the metabolic improvements appears to be limited, as demonstrated by the pair-feeding experiments. Plasma glucose, lipid profiles, and glucose tolerance improvements were more pronounced in the Botox-treated than in the saline-injected partners. Interestingly, the observed metabolic improvements were not associated with any significant changes in EE throughout the study period. Calorie restriction has been reported to lower EE in obese and nonobese humans (30–33), in the face of lower body weight, which is postulated to eventually reach a new “set point” such that the decline in EE is maintained at levels that equal energy intake (34–36). Although this metabolic adaptation is thought to safeguard against excessive weight loss during calorie restriction, it likely poses challenges to weight loss sustenance over long-term and predisposes to weight regain (34,37). Unaltered EE observed in our model alleviates this challenge and underscores the contribution of local and other systemic mechanisms beyond EE to the observed weight loss.
Botox treated mice lost ∼10–15% of body weight and showed significantly lower SNAP-25 expression, suggesting that downregulation or blockade of cholinergic neurotransmission may play a mechanistic role. However, SNAP-25, along with other components of the soluble N-ethylmaleimide–sensitive factor attachment protein receptor (SNARE) complex is also involved in calcium-dependent neurotransmitter release (38–40) and regulation of calcium-independent exocytotic events (41). SNAP-25 also interacts with voltage-gated calcium channels (N-, P/Q-, L-, and T-type), where they modulate calcium currents (42–46), thus indirectly impinging neurotransmission. Given this broad spectrum of effects mediated by SNAP-25, contribution of plausible specific cholinergic blockade to the observed benefits needs to be determined in future investigations.
Furthermore, studies performed in neuronal cell lines and primary neuronal cultures demonstrate that Botox blocks release not only of acetylcholine but also glutamate (47–50), aspartate (51), γ-aminobutyric acid (51,52), dopamines (52), serotonin (54), substance P (55,56), and calcitonin gene-related peptide (57–59) in brain synaptosomes, sensory neurons of the dorsal root ganglia, urinary bladder, and hind paws (47–59). However, in the enteric nervous system, botulinum neurotoxin A and B were found to recognize mainly the cholinergic neurons and their extensions in the intestinal submucosa (47,60,61), although a small proportion of serotonergic neurons and negligible fraction of vasoactive intestinal peptide, vesicular-glutamate transporter, and γ-aminobutyric acid–positive neurons were found to be targeted by both isoforms. Nevertheless, impact of possible blockade of serotonergic neurotransmission (and others as stated above) by Botox and their potential effects on weight loss and associated metabolic benefits cannot be precluded.
Additionally, while the proximal small intestine possesses intrinsic neural plexi in the smooth muscle layer as part of the enteric nervous system that allow formidable autonomous control over GI functions, the impact of central nervous system-mediated extrinsic neural inputs cannot be ignored. The sympathetic nervous system exerts predominantly inhibitory effects upon GI smooth muscle and mucosal secretion while also regulating blood flow via vasoconstriction. In contrast, parasympathetic innervation exerts both excitatory and inhibitory control over GI motility and tone (62,63) and is provided by the vagus nerve. The central terminals of vagal afferent nerve fibers innervate the nucleus tractus solitarius and use glutamate as a neurotransmitter. Botox has been reported to inhibit glutamate exocytosis in rodent cerebral cortical synaptosomes, cerebellar granule neurons, hind paws, and release in human skin (47–59). Although it remains to be determined whether Botox injected into the proximal intestinal wall impacts glutaminergic vagal terminals in the nucleus tractus solitarius and thereby affects glutamine secretion and/or release, plausible inhibition of glutaminergic neurotransmission elicits serious consideration as one of the potential modes of action.
In addition to lower food intake, a significant reduction in glucose and fat absorption was observed. Faster transit, as evidenced in our study by reduced recovery of [3H]3-OMG and [3H]OA in plasma after acute glucose and lipid loads, respectively, has been associated with limited nutrient absorption (64,65), owing at least in part to reduced exposure time of nutrients to luminal surfaces of enterocytes. The drastic reductions in lipid absorption were accompanied by increased fecal lipid output. Importantly, no diarrhea or steatorrhea was observed. Water and fiber (digestible and indigestible) composition of saline- and Botox-treated mice were comparable (data not shown). Absence of diarrhea or steatorrhea in our mouse model is encouraging for clinical testing of Botox for amelioration of obesity and diabetes.
A limitation of this approach is the reversible nature of the benefits, which calls for multiple injections, the efficacy of which needs to be examined. In human cosmetic applications, effects of Botox are potentiated by multiple dosages. If the observed effects are translatable in humans, this approach could lay the foundation for additional promising therapeutic approaches geared toward robust and sustained weight loss, mimicking some of the benefits of bariatric operations without its cost and complications. Clinical trials are warranted to determine whether Botox, when delivered endoscopically into the proximal small intestinal wall, could mimic the benefits of bariatric surgery without the cost and complications and avoiding permanent realignment of the GI tract.
This article contains supplementary material online at https://doi.org/10.2337/figshare.19609992.
Article Information
Acknowledgments. The authors thank the Vanderbilt Mouse Metabolic Phenotyping Core for assistance with energy expenditure experiments, and the Vanderbilt University Medical Center Molecular Cell Biology Resource Core for all technical assistance.
Funding. Internal funding was received from Department of Surgery, Vanderbilt University Medical Center. The EE ANCOVA analysis done for this work was provided by the National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases Mouse Metabolic Phenotyping Centers (www.mmpc.org) using their Energy Expenditure Analysis page (https://www.mmpc.org/shared/regression.aspx) and supported by grants DK076169 and DK115255.
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. S.S. performed experiments, analyzed data, and prepared the draft of the manuscript. J.An., B.B., J.Ad., C.J., B.C., K.D., and J.F. assisted in data acquisition. C.A.S. contributed to the discussion. N.A. envisioned the concept, interpreted the data, and edited and approved drafts. N.A. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.