Transfer of microbiota from lean donors in combination with prebiotics prevents excessive weight gain and improves gut-brain vagal signaling in obese rats

In the United States, 42.4% of the adult population was obese in 2017–2018.¹ Obesity increases the risk of developing cardiometabolic diseases, such as diabetes and hypertension, as well as psychological disorders, and represents a financial burden with an estimated $260.6 billion spent on health care costs.² Most importantly, obesity and its associated comorbidities have a significant negative impact on the perceived quality of life.³

It is generally accepted that obesity results from an disparity in energy intake compared to energy expenditure.⁴ Energy metabolism in the human body is a highly complex, multifactorial process.⁴ The collection of microorganisms that inhabit the gastrointestinal tract (GI), known as the gut microbiota, plays a critical role in regulating host energy metabolism, thus contributing to metabolic health.⁵ Obesity is associated with deleterious changes in gut microbiota composition, or dysbiosis.⁶ Members of the Firmicutes and Bacteroidetes phyla make up ~80% of the gut microbiome, and studies have shown, in humans and rodents, that in obesity the relative abundance of Firmicutes increases while Bacteroidetes decreases.^7–11 Similar changes are observed in response to HF feeding.¹¹ Independent of length of dietary intervention, HF diet consumption leads to microbiota dysbiosis marked by a decrease in microbial diversity and increased pro-inflammatory potential of the gut microbiota associated with impaired GI permeability.^12,13 Colonization of germ-free (GF) animals with microbiota from obese or HF fed donors induces an obese phenotype in the conventionalized animals characterized by increased caloric intake and body fat accumulation.^14,15

Given the role of vagal signaling in regulation of intake and the distribution of vagal terminals along the gastrointestinal lamina propria, it is likely that changes in microbiota composition affect vagal signaling since vagal afferent neurons (VAN) express receptors for bacterial byproducts.^16,17 The vagus nerve provides bidirectional communication between peripheral organs and the brain. Within the GI tract, mechanosensitive vagal sensory terminals respond to distension, and chemosensitive vagal afferent terminals respond to nutrients to suppress meal size.¹⁸ Postprandial signals increase neuronal activity in the nucleus tractus solitarius (NTS), the central site of vagal sensory termination.¹⁹ Release of gut satiety peptides, such as cholecystokinin (CCK) and glucagon-like peptide 1 release (GLP-1) activate vagal afferents to signal meal termination and induce satiety.²⁰ Cocaine- and amphetamine-regulated transcript (CART) neuropeptide expression is induced in VAN, whose cell bodies are located in the nodose ganglion (NG), upon feeding and this response is mediated by CCK.^21,22 CART and CCK-mediated vagal signaling is further associated with c-Fos expression, a marker of neuronal activation, in the NTS and other central areas involved in the control of food intake.²³ HF feeding and/or obesity have been shown to reduce VAN sensitivity to tension, nutrients, and GI hormones as well as postprandial NTS activation. These changes have been linked to food overconsumption.^24–35 In addition to homeostatic mechanisms, gut-innervating VAN project to limbic brain regions, and this gut-reward circuit is also critical for meal termination.³⁶

Besides alteration in vagal function, our lab has previously shown that chronic HF feeding alters vagal structure, causing withdrawal of vagal c-fibers from the NTS. The timing of NTS vagal withdrawal coincides with onset of weight gain, hyperphagia, and insensitivity to GI CCK.^12,13,37,38 Nerve injury-induced or diet-induced vagal withdrawal can be followed by NTS reinnervation (sprouting). The progression time for vagal remodeling is dependent on the type of diet, with more energy dense diets causing more rapid progression, suggesting that HF diet consumption triggers a dynamic structural remodeling in the vagal system. It is unclear whether functionality improves along with density. In rats, reinnervation does not appear to fully restore function, as animals remain hyperphagic and insensitive to CCK.^13,39

Critically, transfer of microbiota from HF fed rats, a dysbiotic microbiota, into germ-free (GF) or microbiota-depleted rats significantly increased recruitment of immune cells in the NG and NTS as well as decreased the density of vagal afferents innervating the NTS. Microbiota-depleted rats colonized with a HF-type microbiota also display reduced sensitivity to the GI satiety peptide CCK³⁷ and dampened food -associated reward.¹⁵ Conversely, normalizing microbiota composition or inhibition of immune cells in HF rats prevented the loss of vagal innervation at the level of the NTS and preserved CCK sensitivity.^13,40 These data show that HF-driven changes in microbiota composition and associated inflammation are necessary and sufficient to alter vagal signaling.

It is, however, unknown whether restoring normal microbiota after obesity has been established and/or following diet-driven alterations in vagal signaling can improve gut-brain signaling and promote weight loss. Weight loss in humans and rodents is associated with changes in microbiota composition;^41,42 but there is little evidence that restoring normal microbiota in obesity can improve metabolism. There is some circumstantial evidence in humans that probiotics use is associated with weight loss,⁴³ but no causal relationship has been established. This is partially because in most studies, microbiota manipulations strategies (antibiotics, probiotics) are generally introduced concomitantly to HF feeding.^44–49 While these models provide insights into the etiology of obesity, they do not address whether targeting the microbiota in an obese model can modulate body weight. Previous reports showed limited effects of prebiotic supplementation after prolonged dietary exposure. Vieira et al. supplemented inulin to rats that had been fed a HF or chow diet for five weeks and reported no effect of inulin on body weight, independent of diet, and no effect on caloric intake in HF fed rats.⁵⁰ Thus, the focus of this study was to evaluate the effect of gut microbiota transfer in obese, vagally-compromised, rats on gut-brain communication, food intake and body weight, 24-h feeding patterns, and willingness to work for a food reward. We found that transferring lean microbiota to HF-fed, obese animals prevented excessive weight gain, normalized feeding patterns, improved learning, and rescued vagally mediated signaling. To our knowledge, this is the first study that has demonstrated that improving the gut microbiota composition in an obese model, independent of diet, can ameliorate some of the detrimental consequences of excess adiposity.

The objective of this study was to evaluate the effect of gut microbiota transfer with inulin supplementation in obese, vagally compromised, rats on gut-brain communication, food intake and body weight, 24-h feeding patterns, and willingness to work for a food reward. The animals were divided into four groups (n = 10 each): LF – control fed regular chow, LF-LF – chow fed animals that received microbiota from chow fed donors, HF-LF – HF fed animals that received microbiota from chow fed donors and inulin supplementation, and HF-HF – HF fed animals that received microbiota from HF fed donors.

LF microbiota transfer in combination with prebiotics prevented excessive weight gain in hf-fed rats

Initial body weights were similar for all rats, and they were randomly assigned to either chow or HF diet. 24-hr caloric intake measured using the Biodaq system in a subset of animals from each group and body weights are shown in Figure 2. Prior to microbiota transfer, HF feeding led a significant increase in daily caloric intake compared to chow (t-test, p < 0.01) (Figure 2a). Consequently, HF fed rats gained significantly more body weight than LF rats (Two-way ANOVA, F³ 323) = 83.86; p < 0.0001) (Figure 2c).

Following microbiota transfer, HF-HF rats had a significantly higher daily intake than all other groups (One-way ANOVA, ps < 0.05). There was a significant effect of time (Two-way ANOVA, F¹⁰ 330) = 34.79; p < 0.0001) and diet (F³ 330) = 30.83; p < 0.0001) on weekly body weight gain (Figure 2d). Following recolonization, HF-LF animals gained significantly less body weight than HF-HF animals (p < 0.05) and the rate of body weight gain in the HF-LF group was comparable to that of LF controls. Since HF-LF animals were maintained on a HF diet, inulin was added to their water as a prebiotic to aid the effectiveness of the microbiota transfer.⁵⁶ Inulin intake did not affect water intake in the HF-LF group compared to the HF-HF group (Supplementary Figure S1).

LF microbiota transfer in combination with prebiotics improved microbiota profile in hf-fed rats

Analysis of fecal microbiota composition is shown in Figure 3. Species richness post inoculation was evaluated using the Shannon index. HF animals had significantly lower species richness compared to LF animals (One-way ANOVA, F^3,19 = 9.741, p < 0.001) (Figure 3a), regardless of microbiota transplant. We did not observe differences in species evenness. A PERMANOVA analysis of (dis-)similarities between samples revealed a clear separation between the fecal microbiota of LF and HF rats (F = 5.0392; p < 0.001) (Figure 3b). There was no difference between LF and LF-LF groups (p = 0.113) as microbiota profiles from these groups clustered together. However, within the HF animals, HF-LF and HF-HF animals displayed significantly different microbiota profiles (p < 0.01). HF-LF animals clustered closer to the LF (LF and LF-LF) animals compared to the HF-HF group, yet their microbiota profile did not fully normalize (ps < 0.05). This can be further appreciated in a dendrogram presenting the Bray Curtis distance (Figure 3b). Although HF-LF animals are significantly different from LF (LF and LF-LF) (ps < 0.01), there is a small overlap, while HF-HF animals are significantly different from HF-LF and LF (LF and LF-LF) (ps < 0.01). Analysis of relative taxa abundance at the phylum level identified significant differences in the Bacteroidetes and Firmicutes phyla among the groups (Two-way ANOVA, F⁷ 146) = 412.7; p < 0.0001) (Figure 3c). HF-HF animals exhibited a significant decrease in relative abundance of members of the Bacteroidetes phylum compared to all other groups (ps < 0.001) as well as significant increase in the relative abundance of Firmicutes compared to all other groups (ps < 0.0001). HF-LF phyla profile was similar to the LF and LF-LF animals, with no differences in relative abundances between the groups

Belonging to the Bacteroidetes phylum, the species caccae and copri were found to be enriched in the HF-LF group compared to the other groups (One-way ANOVA, caccae, ps < 0.01; copri, ps < 0.05 vs. LF-LF and HF-HF). Conversely, the species uniformis was not transferred to HF-LF animals as we found in significantly higher abundance in LF compared to HF animals (One-way ANOVA, ps < 0.05) (Figure 3d). Within the Firmicutes phylum (Figure 3e), species belonging to the SMB53 genus were found to be enriched in LF (LF and LF-LF) compared to the HF-HF animals (One-way ANOVA, ps < 0.05) and this trait was successfully transferred to the HF-LF group (HF-LF vs. HF-HF, One-way ANOVA, p < 0.05). In addition, species belonging the genus Dorea, and the species producta and celatum were present in significantly lower abundance in the LF groups (LF and LF-LF) compared to HF-HF animals (One-way ANOVA, ps < 0.05) and again, this trait was successfully passed to the HF-LF group (HF-LF vs. HF-HF, p < 0.05). Other Firmicutes abundances differed based on diet regardless of microbiota transfer, species of the Ruminococcus genus were found to be significantly enriched in LF compared to HF animals (One-way ANOVA, ps < 0.001) while species of the genus Oscillospira were significantly depleted in LF compared to HF animals (One-way ANOVA, ps < 0.05). Overall, the HF-LF group displayed a unique microbiota profile with similarities with both the LF and HF groups, showing the microbiota transfer improved but did not fully normalize the animals’ microbiota profile.

Functional prediction analysis is shown in (Supplementary Figure S2). Supplementary Figure S2A shows the relative abundance of genes identified in our samples, per group, involved in metabolic pathways. In general, we observed a decrease in the abundance of genes pertaining to KEGG metabolism categories in HF-HF animals compared to all other cohorts. Consistent with a prior in mice fed a high fat diet,⁵⁷ HF-HF animals had significantly lower abundance of genes involved in amino acid metabolism (Supplementary Figure S2B; ps < 0.05), energy metabolism (Supplementary Figure S2C; ps < 0.05), glycan biosynthesis and metabolism (Supplementary Figure S2D; ps < 0.01) compared to LF and HF-LF animals. In addition, abundance of genes involved in biosynthesis of lipid metabolism (Supplementary Figure S2F; p<0.05), metabolism of cofactors and vitamins (Supplementary Figure S2H; p<0.05), secondary metabolites (Supplementary Figure S2G; p<0.05) and metabolism of other amino acids (Supplementary Figure S2E; p<0.05) was significantly lower in HF-HF compared to HF-LF animals. There were no differences in carbohydrate metabolism, metabolism of terpenoids and polyketides, nucleotide metabolism, and xenobiotics biodegradation and metabolism.

We performed linear regression analysis between identified taxa and the phenotypic outcomes assessed in this study. Several correlations were found to be significant and are shown in Supplementary Figure S3. There is a negative correlation between abundance of members of the Bacteroidetes phylum, Copri (enriched in LF, LF-LF, and HF-LF animals) and Uniformis (enriched in LF animals), and body weight gain. There is a positive correlation abundance of members of the Firmicutes phylum, Dorea and Producta (enriched in HF-HF animals), and body weight gain. There is a negative correlation between abundance of SMB53 (enriched in LF, LF-LF, and HF-LF animals) and body weight gain. We also identified a negative correlation between Copri and Uniformis (Bacteroidetes) abundance and acquisition time for an operant task. The species Dorea was positively correlated with acquisition time. Abundance of the species Caccae (enriched in HF-LF rats) and Copri was positively correlated with the number of CART⁺ neurons, while abundance of SMB53 was negatively correlated with the number of cFos⁺ neurons.

LF microbiota transfer in combination with prebiotics normalized feeding patterns and improved acquisition time for an operant task in hf-fed rats

Feeding patterns were analyzed using a BioDaq food intake monitoring system (Figure 4). Pre-inoculation, HF feeding significantly increased meal size during the light phase (t-test, p < 0.05) (Figure 4c). There were no significant differences in meal size in the dark or meal number in the light or dark phase (Figure 4a, g e and e). Post-inoculation, colonization with a LF microbiota normalized light phase meal size (HF-LF vs. HF-HF, One-way ANOVA, p < 0.05). HF-LF animals also displayed a small reduction in meal number during the dark cycle, but this only reached significance when compared to the LF-LF groups (One-way ANOVA, p < 0.05, Figure 4f). The HF-HF group still displayed a significant increase in meal size during the light phase compared to the LF groups (LF and LF-LF, One-way ANOVA, ps < 0.05). There were no significant differences among groups in meal size in the dark or meal number in the light (Figure 4b, h)

A progressive ratio schedule was used to assess the animals’ drive for fat and sucrose pellets (Figure 5). There were no differences in responses for fat versus sucrose reward pellet; thus, data were combined. Pre inoculation, LF rats achieved FR3 and FR5 criteria significantly faster (t-test, ps < 0.05) than HF rats (Figures. 5a, c). Time to achieve FR criteria has been interpreted as a measure of motivation¹⁵ as well as a measure of cognitive function and ability to learn.^58,59 There was no difference in breakpoint between LF and HF animals (Supplementary Figure S4A). Post inoculation, colonization of HF fed rats with a LF microbiota (HF-LF) led to a significant improvement in task acquisition as HF-LF rats achieved FR3 and FR5 criteria as quickly as the LF animals. Although only a trend was observed during FR3 acquisition, LF (LF and LF-LF) and HF-LF animals reached FR5 criteria significantly faster than HF-HF animals (One-way ANOVA, ps < 0.05). Again, there was no difference in breakpoint amongst the cohorts (Supplementary Figure S4B).

LF microbiota transfer in combination with prebiotics rescued postprandial NG and NTS activation in hf-fed rats

Refeeding a fasted animal significantly increases CART immunoreactivity in the NG of lean¹¹ but not HF fed obese animals.⁶⁰ Post-prandial vagal activation is dependent on the amount of kcal consumed as well as macronutrient makeup, with fat consumption leading to greater activation, for the same amount of kcal, than carbohydrates and proteins.⁶¹ Following refeeding, CART⁺ neurons were present in the NG of all animals, however there were no increase in the number of CART⁺ neurons in the HF-HF rats compared to the LF groups, despite receiving a fat-rich meal. HF-LF rats had a significantly higher number of CART positive neurons compared to all other groups (One-way ANOVA, ps < 0.001) (Figure 6a-e). Similarly, there were no differences in the number of NTS c-Fos⁺ neurons in the NTS between the HF-HF rats and the LF animals while the HF-LF animals had a significantly higher number of c-Fos⁺ neurons than LF-LF and HF-HF animals (One-way ANOVA, ps < 0.05) (Figure 6f-j). This supports our hypothesis that improving the gut microbiota composition ameliorates vagal signaling independent of diet since HF-LF animals are exhibiting greater neuronal activation following a meal than HF-HF rats.

LF microbiota transfer in combination with prebiotics decreased immune cells activation but did not affect the density of vagal afferents in the NTS in HF fed rats

Obesity is a low-grade chronic inflammatory disease.⁶² HF-induced changes in gut microbiota composition result in an increased proinflammatory state in the organism.⁶³ Microbiota-derived inflammation has been shown to detrimentally alter communication along the gut-brain vagal axis and increased microglia activation in the NTS is one of the hallmark phenotypic observations.^{12,13,38,40,64,65} HF-HF animals had a significant increase in the number of Iba1⁺ cells in the NTS as well as overall positive staining compared to LF (LF and LF-LF) (One-way ANOVA, ps < 0.05) (Figure 7). Colonization of HF fed animals with a LF microbiota (HF-LF) led to a significant reduction in NTS Iba1⁺ cells and staining (HF-LF vs. HF-HF, One-way ANOVA, p < 0.05). This reduction the NTS Iba1⁺ cells and staining observed in HF-LF animals indicates lower inflammatory response and thus improved metabolic health. However, staining against isolectin B4 showed no difference in vagal afferent density in the intermediate NTS among the groups (Supplementary Figure S5).

As the prevalence of obesity and obesity-related comorbidities continues to rise worldwide, it is of primary importance to develop therapies that can aid in preventing overeating and excessive weight gain. In the current study, we sought to understand whether normalizing gut microbiota composition in HF fed obese animals would offset the detrimental physiological effects of consuming an energy dense, high in fat and sugar, diet. Prior reports,^40,66 have shown that preventing dysbiosis in HF fed animals improves metabolic outcomes. Most of these interventions were started at the same time as HF diet introduction. Here, we report that transferring the microbiota of a LF rat into an already obese HF-fed animal (HF-LF) along with inulin supplementation prevented the excessive body weight gain characteristically observed with HF feeding, normalized meal patterns, and improved vagally mediated gut-brain axis signaling and learning.

Chronic consumption of a HF diet significantly decreases bacterial diversity and alters the gut microbiota profile. Consistent with existing data, our results show that HF animals had significantly lower bacterial diversity than LF animals.¹² HF animals recolonized with a HF microbiota (HF-HF) also displayed the hallmarks of HF feeding,⁶⁷ including increased Firmicutes abundance, in particular Dorea⁴⁰ and reduced Bacteroidetes. Although a LF microbiota transfer into a HF animal did not rescue bacterial diversity, the microbiota profile of these animals resembles more closely that of LF animals than HF animals. LF microbiota transplant to HF animals resulted in normalized – reduced – abundance levels, comparable to LF animals, of members of the Lacnospiraceae (Dorea and Producta) and Clostridiaceae (SMB53 and Celatum) families of the Firmicutes phylum. Firmicutes have been hypothesized to be more efficient at extracting energy from food promoting weight gain^7,68,69 and we observed positive correlations between Dorea and Producta relative abundances and body weight gain. In addition, lower abundances of these taxa have been associated with increased fecal SCFA, reduced adiposity and gut permeability, and improved cardiometabolic health profile.⁶⁹ SCFA stimulate the production and differentiation of enterocytes, improving mucus production, epithelial health and barrier function.^70,71 Increased GI permeability allows for translocation into the circulation of pro-inflammatory bacterial factor, such as lipopolysaccharide,^72,73 which has been shown to alter vagal signaling.⁷⁴ In addition to their trophic role, acetate and propionate can activate G-protein coupled receptors (GPRs), such as GPR43 and promote production of GI peptides including GLP-1.⁷⁵ Diet-driven reduction in GI SCFA could therefore both directly and indirectly impact gut-brain communication. In contrast, the relative abundance of several members this Firmicutes phylum was not affected by the microbiota transfer. The taxa Oscillospira, associated with lower BMI⁷⁶ and reduction of body weight,⁷⁷ was found in similar abundance levels in HF-LF and HF-HF animals. This is likely due to increased bile secretion in response to the HF diet as it has been shown that HF diet consumption increases bile secretion⁷⁸ and there is a positive correlation between bile secretion and Oscillospira abundance level.⁷⁹

HF animals that received LF microbiota (HF-LF) also displayed increased relative abundances of bacteria belonging to the Bacteroidetes phylum. It is possible that members of the Bacteroidetes phylum are playing a role in the reduced weight gain observed in these animals as it has been previously shown that body weight loss is positively associated with increased Bacteroidetes abundance in obese human subjects.⁸⁰ We did find negative correlations between abundance of members of the Bacteroidetes phylum Copri (enriched HF-LF animals) and Uniformis (enriched in LF animals), and body weight gain. Despite not reaching statistical significance, HF-LF had a greater abundance of A. muciniphila compared to LF (LF and LF-LF) and HF-HF animals. Previous studies have shown that A. muciniphila is involved in regulation of intestinal lining integrity, adiposity⁸¹ and inflammation.^82,83 High fat diet significantly reduces the abundance of A. muciniphila,^84,85 however, prebiotic (oligofructose) supplementation has been shown to restore its abundance to normal levels.⁸⁴ In the present study, it is likely that inulin supplementation aided the reestablishment of A. muciniphila abundance in the HF-LF animals. Similar results have been reported in high fat fed mice that also received a probiotic cocktail (Lactobacillus acidophilus, Bifidobacterium longum, and Enterococcus faecalis − 1:1:1).⁸⁵

Similarly to previous reports,⁵⁷ HF feeding led to a reduction in several bacterial metabolism pathways, including amino acids, energy and glycan metabolism. These changes were not evident in HF-LF rats. Obese individuals have been shown to display significant reductions in bacterial glycan metabolism,⁸⁶ carbohydrate and amino acids metabolism.^87,88 This reduction in bacterial gene richness is associated with metabolic alterations, in particular low-grade inflammation.^89,90 Interestingly, in mice, HF diet- driven reduction in specific metabolism pathways can be reversed by antibiotic usage,⁹¹ thus changes observed may not be related to a general loss in bacterial diversity but to diet-driven changes in bacterial genes. By transferring lean microbiota into HF-fed rats, we may have counteracted the establishment of deleterious bacteria as well as alterations in specific bacterial gene expression and prevented the excessive weight gain typically induced by HF feeding.

In rats, chronic HF feeding triggers an increase in caloric intake and leads to body weight gain over time. Consistent with these reports, our results show that HF animals had a significantly higher caloric intake and gained significantly more weight than LF animals pre-inoculation. Increase in intake was driven by an increase in meal size, especially during the light cycle.⁹² While some effect of HF feeding on meal size may be driven by the novelty and increased palatability of the diet, HF rats that received HF microbiota still displayed increased meal size after being on the diet over a prolonged period of time. Thus, meal pattern disruptions elicited by a HF diet are long-lasting and have been linked to increased body fat accumulation in animal models⁹³ and human subjects.⁹⁴ However, improving the gut microbiota composition can improve feeding patterns since HF rats that received LF microbiota did not exhibit increased meal size despite being on a HF diet. It is noteworthy to mention that these rats did exhibit increased meal size prior to the microbiota transfer. Overall, HF animals that received LF microbiota transfer (HF-LF) consumed significantly less calories and gained significantly less body weight than HF animals that received HF microbiota (HF-HF). Male Wistar rats pair-fed a HF or HS (high sucrose) diet gained significantly more body weight than their LF counterparts.^95–97 Similarly, male Sprague-Dawley rats maintained on a HF (45% fat) diet for over six months gain significantly more body weight and fat mass than LF animals.³⁸ Here, we show that improving the gut microbiota composition to a lean profile counteracts the obesogenic effect of a HF diet independent of dietary changes. To our knowledge, this is the first time these findings have been reported in conventional animals fed a HF diet.

Post-prandial gut-brain communication plays a key role in regulating satiation and meal termination.^67,98–100 Previous studies from our lab have demonstrated that HF feeding affects the gut-brain vagal system structure and function, somewhat independently of one another. HF diet consumption affects the density of vagal terminals in the NTS and that these effects are dependent both upon length of time on the diet and diet composition.^12,13 One week of a 60% HF diet exposure led to a transient decrease in vagal fibers in the NTS followed by sprouting, noticeable three weeks on the diet.¹³ In contrast, we have observed significant reduction in vagal fiber density in the NTS after four weeks and 8 weeks of 45% HF diet exposure.^12,37,40 Taken together, these results suggest that the morphological reorganization of the gut-brain vagal axis observed in response to a HF diet fluctuates over time

Gut-brain mediated satiation signaling is the major determinant of meal size GI satiety peptides, such as CCK and GLP-1 activate vagal afferents to trigger meal termination.^101,102 CART neuropeptide expression is induced in the NG upon feeding, and this response is mediated by CCK.^21,22 CART and CCK-mediated vagal signaling is further associated with c-Fos expression, a marker of neuronal activation, in the NTS and other central areas involved in the control of food intake.^23,60 These satiation and satiety signals are known to be compromised, i.e., blunted, in HF feeding and obesity as a result of decreased sensitivity to CCK.¹⁰ Post-prandial vagal activation is dependent on the amount of kcal consumed as well as macronutrient makeup, with fat consumption leading to greater activation, for the same amount of kcal, than carbohydrates and proteins.⁶¹ There were no differences in the number of CART⁺ neurons in the NG and c-Fos⁺ neurons in the NTS following refeeding between HF-HF rats and LF animals, despite the HF-HF rats consuming a fat rich meal. Colonization of HF animals with a LF microbiota resulted in a significantly greater number of CART⁺ neurons in the NG and c-Fos⁺ neurons in the NTS after refeeding compared to all other groups. We also identified positive correlations between abundance of the species caccae and copri and the number of CART⁺ neurons in the NG. This supports our hypothesis that improving the gut microbiota composition ameliorates vagal signaling independent of diet since HF-LF animals are exhibiting greater neuronal activation following a meal than HF-HF rats.

Diet and microbiota-driven alteration in vagal signaling are accompanied by an increase presence of Iba1⁺ positive immune cells along the gut-brain axis.^{12,13,38,40,64,65} Recruitment of these cells appears to be necessary for diet-driven alterations of vagal signaling as their pharmacological inhibition prevents diet-associated vagal remodeling and hyperphagia.¹³ Similarly to what had previously been reported,^13,37,38,40 we observed an increased in the number of Iba1⁺ cells (and overall staining) in the NTS of HF-HF rats. Conversely, HF animals that received LF microbiota do not exhibit an increase in inflammatory markers (Iba1⁺ cells) in the intermediate NTS, showing that changes in microbiota composition are sufficient to decrease immune cells recruitment along the gut-brain axis, independently of diet. This reduction in neuroinflammation may mediate the improved post-prandial neuronal response we observed in HF-LF rats. Sprouting following initial withdrawal of vagal terminals from the NTS does not appear to restore function, as rats fed a HF diet remained insensitive to gut satiety peptides.^13,39 Consistent with these findings, our results show no significant difference in the density of vagal afferents in the intermediate NTS among the groups. It is likely that since these animals have been on HF diet for over six months, there was an initial decrease in the density of VAN when the diet was first introduced followed by sprouting and restoration of structure in the HF-HF animals did not translate in reestablishment of function.

In addition to its classic role in homeostatic regulation of feeding, gut-brain signaling has recently been shown to modulate food-driven reward. A recent study unveiled a direct, multi-synaptic pathway between the gut and the dopaminergic reward system. Han et al. showed that right NG vagal sensory fibers that innervate the gut project to the ventromedial NTS. These cells in turn project to the medial Parabrachial Nucleus, which then targets the Substantia nigra. Activation of right vagus-parabrachial-nigrostriatal pathway is required for nutrient sensing, sustained self-stimulation, and stimulus preference.³⁶ This gut to reward pathway is blunted in obesity and diet-induced changes in gut microbiota are sufficient to induce deficits in vagally-mediated reward.¹⁵ Progressive ratio (PR) responding has been widely used to study willingness to work for a natural reward or addictive substance in animal subjects. In recent years, studies have further shown that this task can also be used to evaluate learning, memory, and cognitive function.^58,59 In this task, animals are trained to emit a response to obtain a reward. Thus, acquisition time for the task can be reflective of learning capabilities. Breakpoint, which is the maximum number of responses the animal emits, relates to the amount of “work” a subject is willing to perform to obtain a reward. Consistent with a reduced food-driven reward, HF feeding has been shown to reduce motivation to work for a palatable reward.^103,104 In a recent study from our lab, we showed that HF rats exhibit significantly longer acquisition times for a fixed ratio (FR) 3 schedule compared to LF animals, but no difference in breakpoint. We further showed that colonization of GF rats with HF microbiota significantly lowered breakpoint and increased task acquisition time. These behaviors were directly linked to downregulation of the dopaminergic reward system as a result of vagal signaling alterations in response to changes in the microbiota composition.¹⁵ Consistent with these findings, HF feeding significantly increased the time to achieve FR3 and FR5 but had no effect on breakpoint. LF microbiota transfer to HF rats significantly decreased FR3 and FR5 acquisition time. Vazquez et. al. used acquisition time for a fixed ratio schedule task to assess long-term potentiation and memory, and determined that intact gut-brain communication through the vagus nerve improves learning and memory in rats.⁵⁹ This data further supports our hypothesis that improving the gut microbiota in HF animals restores gut-brain vagal signaling.

In conclusion, this study provides a proof of concept that improving the gut microbiota composition in diet-induced obesity can ameliorate many of the detrimental consequences of chronic high fat-high sugar diet consumption We found that microbiota transfer combined with inulin supplementation in obese rats resulted in a beneficial shift in the microbiota composition even when the animals remained on a HF diet. It is worth nothing that while the present study design does not allow us to distinguish between the effects of microbiota transfer and inulin or the synergistic actions of both, it has been previously shown that inulin supplementation alone after exposure to a HF diet for five weeks has no effect on body weight or caloric intake in rat.⁵⁰ Additionally, based on previous reports⁵⁴ the dose of inulin we used is not expected to independently affect food intake and body weight. Therefore, it is likely that changes we observed in metabolic outcomes were a result of a synergistic action of inulin and microbiota transfer. HF animals that receive LF microbiota gained less body weight, exhibited normalized feeding patterns, improved task acquisition, and had restored function of the gut-brain vagal signaling pathway. These data emphasize the importance of the gut microbiota population in the regulation of feeding behaviors not only in the development of but also in the maintenance of obesity, thus representing a potential therapeutic target to improve obesity and its comorbidities.

The authors would like to thank Shreya Pandya for her help with the meal patterning data

This work was supported by National Institutes of Health grant RNIHX0001216701.

No potential conflict of interest was reported by the author(s).

DM and CdLS analyzed the data, prepared the figures and wrote the main manuscript text.

GdL and CdLS designed the experiments.

JSK and RK conducted the animal work.

JSK, JA, SC and NM participated in data collection and analysis.

All authors reviewed the manuscript.

All data supporting the findings of this study are available within the paper and its Supplementary Information.

All raw data, including microbiota sequencing and IHC pictures can be accessed at:

https://www.dropbox.com/scl/fo/2mlhpk8z7w8r15xvjmg2g/ADWIt9BgTqk06eETZ6xkDcc?rlkey=bpztauv1rnv2kcvh90sfcj9jv&st=4s0u46aq&dl=0↗.

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19490976.2024.2421581↗

All animal procedures were approved by the University of Georgia Institutional Animal Care and Use Committee (#A2019 05–002) and conformed to the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals.