Low-dose glyphosate exposure alters gut microbiota composition and modulates gut homeostasis

Abstract

The widespread use of glyphosate, a broad-spectrum herbicide, has resulted in significant human exposure, and recent studies have challenged the notion that glyphosate is safe for humans. Although the link between disease states and glyphosate exposure is increasingly appreciated, the mechanistic links between glyphosate and its toxic effects on human health are poorly understood. Recent studies have suggested that glyphosate may cause toxicity through modulation of the gut microbiome, but evidence for glyphosate-induced gut dysbiosis and its effect on host physiology at doses approximating the U.S. Acceptable Daily Intake (ADI = 1.75 mg/kg body weight) is limited. Here, utilizing shotgun metagenomic sequencing of fecal samples from C57BL/6J mice, we show that glyphosate exposure at doses approximating the U.S. ADI significantly impacts gut microbiota composition. These gut microbial alterations were associated with effects on gut homeostasis characterized by increased proinflammatory CD4⁺IL17A⁺ T cells and Lipocalin-2, a known marker of intestinal inflammation.

1. Introduction

Glyphosate-based herbicides have been widely used in agriculture for decades and are the primary herbicides used worldwide due to the wide availability of genetically modified, glyphosate-resistant food crops. Approximately 8.6 billion kg of glyphosate have been applied since 1974, with two-thirds of this amount used in the past 10 years (Hebert et al., 2019). Glyphosate is frequently detected in our water (Sanchis et al., 2012), soils (Silva et al., 2018), food crops (Kolakowski et al., 2020, Louie et al., 2021), and even urine which has raised concerns over its overall effect on human health (Eaton et al., 2022, Grau et al., 2022, Nomura et al., 2022). Currently, the U.S. EPA has set the acceptable daily intake (ADI) or chronic Reference Dose (cRfD) at 1.75 mg/kg/day (Mao et al., 2018).

Glyphosate kills weeds by inhibiting the 5-enolpyruvylshikimate-3 phosphate synthase (EPSPS) enzyme, which is a central enzymatic step of the shikimate pathway responsible for the biosynthesis of aromatic amino acids (phenylalanine, tyrosine, and tryptophan) (Schonbrunn et al., 2001). The shikimate pathway is primarily found in plants and some microorganisms, where many off-target effects of glyphosate have been reported (Zobiole et al., 2011, Newman et al., 2016). As mammals do not have the shikimate pathway, it was initially thought that glyphosate could not negatively impact human health. However, multiple studies have linked glyphosate exposure with various diseases, including cancers (Zhang et al., 2019). These findings have increased focus on understanding the potential mechanism through which glyphosate exposure can cause toxic effects in humans. In this regard, the gut microbiota has emerged as a possible link between glyphosate and adverse health effects reported in humans. Trillions of bacteria (gut microbiota) living in the human gut play a critical role in maintaining the healthy state of the human through the regulation of several host physiological processes, including the development and maintenance of the immune, endocrine, and nervous systems (Fan and Pedersen, 2021). As bacteria utilize the shikimate pathway, glyphosatecould alter gut microbiota composition by inhibiting gut bacteria harboring glyphosate sensitive EPSPS enzymes. Multiple in vitro studies have verified that many gut resident microbes are sensitive to glyphosate exposure. One study showed that the common beneficial bacteria Lactobacillus and Bifidobacterium were more susceptible to glyphosate exposure than the pathogenic bacteria Clostridium perfringens and Salmonella typhimurium (Shehata et al., 2013). However, the biological relevance of glyphosate-induced gut dysbiosis on observed effects on host physiology remains controversial. While some studies report significant changes to gut microbiome composition after glyphosate exposure, the effect of glyphosate at doses relevant to human exposure is lacking, with many studies administering doses much higher than the U.S. ADI (Nielsen et al., 2018, Aitbali et al., 2018, Motta et al., 2018, Tang et al., 2020, Barnett and Gibson, 2020, Mesnage et al., 2021, Van Bruggen et al., 2018). Additionally, the glyphosate formulation used in different studies significantly affects gut microbiome alterations and bacterial growth inhibition in vitro, making it difficult to separate the effect of commercial adjuvants from the active ingredient, glyphosate (Clair et al., 2012, Nielsen et al., 2018, Poppe et al., 2019). Thus, the impact of glyphosate-induced gut microbial alterations on host physiology is poorly understood.

In recent years, modulation of the gut microbiota has gained significant attention for its association with various disorders, including metabolic, gastrointestinal, and autoimmune diseases (Fan and Pedersen, 2021, Zhang et al., 2020, Qiu et al., 2022). The gut microbiome mediates these effects on human health in multiple ways, including the production of metabolites that interact with the colonic epithelium and lamina propria to alter immune and metabolic responses (Zheng et al., 2020). These microbial metabolites include short-chain fatty acids (SCFAs), secondary bile acids, neurotransmitters, and gaseous molecules (Koh and Backhed, 2020). For example, SCFAs are known to modulate gene expression in gut resident T cells through histone deacetylase (HDAC) inhibition leading to an increase in protective, anti-inflammatory regulatory T cells (Tregs) (Smith et al., 2013, Waldecker et al., 2008). However, in a dysbiotic state, characterized by an increase in harmful bacteria (pathobionts) and a decrease in commensal bacteria, metabolite synthesis shifts leading to a proinflammatory immune response.

Although, glyphosate has been shown to modulate gut bacteria and influence host physiology at high doses, the effect of glyphosate on gut microbiome composition and the health of the host at doses similar to the U.S. ADI is largely unknown. Therefore, we undertook this study to determine the effect of glyphosate on gut microbiota at the U.S. ADI as well as to decipher the mechanism by which glyphosate-modulated gut microbiota composition may induce adverse health outcomes. Specifically, we aimed to investigate gut microbial alterations after pure glyphosate exposure in mice at doses of 0 μg/ml (control), 1μg/ml, 10μg/ml (dose similar to the U.S. ADI at 1.75 mg/kg/day) (Mao et al., 2018), and 100μg/ml. We show that low-dose glyphosate exposure can modulate gut microbiota composition, especially by reducing the abundance of known beneficial commensal bacteria Bifidobacterium pseudolongum and Lactobacillus sp., which is accompanied with a decrease in microbial SCFA biosynthesis pathways. Additionally, glyphosate exposed mice showed increased fecal pH levels and increased proinflammatory markers. Thus, our study highlights the ability of glyphosate to cause gut dysbiosis and a potential mechanism through which exposure to glyphosate can negatively affect host physiology.

2. Methods

2.1. Animal housing and experimental design

Four weeks old C57BL/6J mice were purchased from Jackson Laboratories (Bar Harbor, ME). Animals had access to ad libitum Milli-Q water and were housed in a University of Iowa mouse facility under standard conditions. Mice were fed 7013 irradiated diet (Envigo, Madison, WI), and the bedding , Cellu-nest^™ (Sheppard Specialty Papers, Framingham, MA) was mixed between cages upon arrival to normalize the microbiome before glyphosate exposure. The circadian period was 12:12 light:dark cycles and mice were housed in cages of 4-5. Cages were allocated into three exposure groups: 1 μg/ml glyphosate, 10 μg/ml glyphosate (dose approximating U.S. ADI at 1.75 mg/kg/day), 100 μg/ml glyphosate, and one control group (n = 6, each group). For exposure groups, pure glyphosate (Research Products International, Mount Pleasant, IL) was administered through drinking water for 90 days. Fecal pellets were collected at 30 days, 60 days, and 90 days and immediately frozen at −80 °C. Blood was collected at 30 days, 60 days, and 90 days for serum isolation. After 90 days, mice were euthanized using CO₂ and colon tissue was harvested for isolation of lamina propria cells.

2.2. DNA isolation and metagenomic sequencing

Mouse fecal samples collected from glyphosate exposed and control C57BL/6J mice were used for microbial DNA extraction using Qiagen PowerSoil DNA Isolation Kit (Qiagen,Germantown, MD) following the manufacturer's protocol and described by us previously (Shahi et al., 2019). wereSamples were sequenced using MiSeq Illumina platform and analyzed at CosmosID Inc (CosmosID, Rockville, MD) to obtain a taxonomic abundance table, and Metacyc annotated metabolic pathways.

The downstream analysis of the taxonomic table was performed in R v4.1.0 (R Core Team, 2021) using phyloseq (McMurdie and Holmes, 2013), microbiome (al.), and vegan packages (Oksanen, 2020). Faith's phylogenetic diversity was calculated as alpha diversity measures. Those taxa with <100 reads in at least 20% of the samples were removed. Sequence reads for each sample was further normalized to their median sequencing depth. Weighted unifrac distances and Bray-Curtis dissimilarity metrics were used for beta diversity analysis. Bacterial species were analyzed for differential abundance using LEfSe (Linear discriminant analysis Effect Size) with default parameters in microbiomemarker package in R (Y, 2022). Significantly altered taxa were further tested using the Kruskal Wallis test to identify differentially abundant bacterial taxa between 0 μg/ml, 10 μg/ml, and 100 μg/ml groups after adjusting for multiple comparisons using the Benjamini-Hochberg correction (q<0.25). For functional analysis, Metacyc pathways were analyzed for differential abundance using LEfSe with default parameters in the microbiomemarker package in R.

2.3. Lipocalin-2 and pH measurements

Fecal pellets diluted 1:10 (v/v) in deionized water were homogenized and centrifuged at 10,000 g for 5 min at room temperature to collect fecal supernatant. Lipocalin-2 concentrations were measured using the Mouse Lipocalin-2/NGAL DuoSet ELISA kit (R&D Systems, Minneapolis, MN). The pH of fecal supernatant was measured using a pH microelectrode (Thermo Fisher Scientific, Waltham, MA).

2.4. Isolation of colonic lamina propria cells and flow cytometry

Colons were harvested from different groups (n = 4-5 mice per group), rinsed with PBS, and Peyer's patches were removed. Colons were cut into approximately 0.5 cm sections and incubated for 20 min two times in a 37μ C rotating incubator in HBSS media (BioWhittaker, Walkersville, MD) containing 2% FBS, 10 mM EDTA, and 2mM DTT to remove the epithelial layer. Tissue pieces were rinsed in 2% FBS HBSS and were incubated for 1 hour in a 37°C rotating incubator in 10% FBS RPMI, 1-1.5mg/mL Collagenase III (Worthington, NJ), and 0.4μg/ml Dispase (Thermo Fisher, Waltham, MA). Digested tissue was strained, and cells were collected by centrifugation at 900g for 10 min. Cells were resuspended in 4mL of 40% Percoll, overlayed on 80% Percoll and spun down at 500g for 20 min without a break followed by collection and washing. Lamina propria cells were then stimulated for 6 hours with cell stimulation cocktail at 1X concentration (Tonbo Biosciences, San Diego, CA). Cells were acquired on Cytek Aurora (Cytek Biosciences, San Diego, CA) and analyzed with FlowJo (Ashland, OR) software. To determine surface marker expression, single-cell suspensions were incubated with specific monoclonal antibodies (mAbs) at 4°C for 30 min and washed. For intracellular staining of cytokines/transcription factors, cells were fixed and permeabilized using FoxP3/Transcription Factor Staining Buffer Set (eBioscience, San Diego, CA) followed by incubation with mAbs at 4°C for 30 min. The mAbs used were CD45 30-F11( BD Biosciences, San Jose, CA), CD4 RM4-5( BD Biosciences), FoxP3 FJK-116s, (eBioscience, San Diego, CA), and IL17A TC11-18H10(eBioscience).

3. Results

3.1. Distinct microbiota between glyphosate exposed and control mice

To determine whether glyphosate at the U.S. ADI dose cause gut dysbiosis, we exposed mice to three doses of glyphosate in drinking water for 90 days and analyzed their fecal microbiota after 30-, 60-, and 90-days using shotgun metagenomics sequencing. These doses were selected to resemble a scale similar to the U.S. ADI of 1.75 mg/kg/day or 10-fold below and above the U.S. ADI dose, where a concentration of 10 μg/ml is approximately equal to the U.S. ADI assuming an average daily water consumption of mice of 4 ml. We observed no significant differences in phylogenetic diversity between doses (Figure 1A). However, phylogenetic diversity significantly decreased over time in mice treated with 10 μg/ml (R = −0.54, p = 0.0051) and 100 μg/ml glyphosate (R = −0.54, p = 0.0052), while no difference in phylogenetic diversity over time was observed in control or 1 μg/ml exposed groups (Figure 1B). These data indicate that long-term glyphosate exposure at the U.S. ADI or higher doses may exert selective pressure on gut bacteria favoring a more uniform community.

Figure 1. Distinct gut microbiota between glyphosate exposed and control mice.

A) Faith's Phylogenetic diversity of mouse gut microbiota after 30, 60, or 90 days exposure to glyphosate water (0 μg/ml , 1 μg/ml , 10 μg/ml , 100 μg/ml ) (n = 6). B) Change in Faith’s Phylogenetic diversity over time. C) Beta diversity of mouse gut microbiota after 30 days D) 60 days, and E) 90 days exposure to glyphosate measured using weighted unifrac distance metrics. Adonis2 test was performed for statistical differentiation between the groups for C, D, and E.

We analyzed beta-diversity at 30, 60, and 90 days to determine the compositional similarity between microbial communities. At 30 days, no significant difference was observed between the control and glyphosate-exposed groups (Figure 1C). Similarly, no significant difference between microbial communities was observed at 60 days (Figure 1D). However, at 90 days, there was a significant difference between microbial communities with 10 μg/ml and 100 μg/ml exposure groups showing distinct separation from the control group (p = 0.011) (Figure 1E). Due to the observed similarity between the 1 μg/ml exposure group and control, only control, 10 μg/ml, and 100 μg/ml groups were used for further downstream taxonomical analysis.

3.2. Glyphosate exposed and control groups show differentially abundant bacterial taxa

Next, we determined the microbiota compositional changes between control and glyphosate-exposed groups to understand which taxa may be sensitive or resistant to glyphosate exposure. In all groups, Bacteroidetes and Firmicutes were the dominant phyla. However, increasing glyphosate concentrations reduced Firmicutes and increased Bacteroidetes relative abundance at 90 days (Supplementary Figure 1A). This change was reflected in the Bacteroidetes to Firmicutes (B:F) ratio, where a significant increase in the B:F ratio was observed over time for the 100 μg/ml group (R = 0.6, p = 0.0083) (Supplementary Figure 1B). No significant change in the B:F ratio over time was observed for the control and 10 μg/ml groups (Supplementary Figure 1B). Furthermore, LEfSe was used to identify microbial taxonomic markers in the control and 10 μg/ml groups at 90 days. We found that Clostridia and Lachnospiraceae were important taxa distinguishing the 10 μg/ml group, while Actinobacteria and Bifidobacterium were enriched in the control group (Fig 2A, 2B). Similarly, comparison between the 100 μg/ml and control group at 90 days showed that Bacteroidetes was significantly enriched in the 100 μg/ml group, while the genus Lactobacillus was the major marker in control (Figure 2C, 2D). At an FDR of ≤25%, 13 taxa at the species level were differentially abundant. The top 7 species with an average of ≥4000 reads in at least one of the groups are shown in Figure 3. Multiple species belonging to the Lactobacillus genus showed lower abundance in glyphosate-exposed groups where Lactobacillus murinus and Lactobacillus sp. ASF360 were depleted in the 100 μg/ml group and Lactobacillus reuteri was depleted in the 10 μg/ml group. Additionally, Bifidobacterium pseudolongum showed lower abundance in the 10 μg/ml group compared to the control. Furthermore, Bacteroides acidifaciens showed higher abundance, and Enterorhabdus mucosicola showed lower abundance in the 10 μg/ml and 100 μg/ml glyphosate-exposed groups. A number of Bifidobacterium and Lactobacillus sp. are known probiotics with antitumor activity, antioxidant, antibacterial, and immunomodulatory activity (O'Callaghan and van Sinderen, 2016, Di Cerbo et al., 2016). Thus, our microbiome analysis showed that glyphosate exposure at U.S. ADI levels result in the loss of beneficial bacteria.

Figure 2. LefSe analysis shows differentially abundant taxa between control and glyphosate-exposed microbiota.

A) LefSe analysis between control and 10 μg/ml glyphosate exposed mice at 90 days, B) Cladogram of LefSe analysis in A representing differentially abundant taxa from A. C) LefSe analysis between control and 100 μg/ml glyphosate exposed mice at 90 days. D) Cladogram representing differentially abundant taxa from A (n = 6). Lefse was analyzed using “microbiomemarker” package in R with an LDA cutoff of 2 and Wilcox cutoff of 0.05.

Figure 3. Differentially abundant microbial species after 90 days of glyphosate exposure.

Box plots represent normalized abundances of significantly different bacterial species identified using LefSe in Figure 2. Kruskal-Wallis test followed by pairwise Wilcox test with Benjamini-Hochberg adjustment was performed to determine significance. Species with p≤ 0.05 and q≤0.25 are plotted. Species with abundance <5000 reads are not shown.

3.3. Functional differences between control and glyphosate-exposed microbiota

Alterations in gut microbiota composition also led to differences in the functional potential of the gut microbiome, which are tightly connected to host physiology and play important roles in immunomodulatory processes (Heintz-Buschart and Wilmes, 2018). MetaCyc metabolic pathways analysis showed distinct separation between control and 10 μg/ml glyphosate-exposed mice (Fig 4A) implying that the bacterial community in glyphosate-exposed mice is functionally different from control mice. Numerous pathways were significantly differentially abundant between control and glyphosate-exposed mice, including pathways related to SCFA production (Fig 4B). These pathways included "pyruvate fermentation to acetate and lactate II," "acetylene degradation," and "super pathway of glucose and xylose degradation," which were enriched in the control group. Intestinal SCFA production by gut bacteria is known to exert anti-inflammatory effects through the induction of regulatory T cells and can ameliorate DSS-colitis in vivo (Parada Venegas et al., 2019). Thus, our data suggest that glyphosate exposure modulates bacterial metabolic pathways that likely affect inflammation.

Figure 4. Functional differences between control and glyphosate-exposed gut microbiota.

A) PCoA plots representing differences in the diversity of functional pathways between control and 10 μg/ml glyphosate exposed mice at 90 days (n = 6), measured using Bray-Curtis dissimilarity metrics. B) LEfSe analysis between control and 10 μg/ml glyphosate exposed mice at 90 days, all taxa with LDA scores >3 are shown.

3.4. Glyphosate exposure alters lamina propria resident CD4⁺ IL17A⁺ T cells and markers of inflammation

To determine whether glyphosate can induce a pro-inflammatory response at concentrations similar to the U.S. ADI, we isolated immune cell populations from colonic lamina propria of mice exposed to 10 μg/ml glyphosate for 90 days.We observed that the frequency of CD4⁺IL17A⁺ cells (Th17) was significantly increased in the colonic lamina propria of glyphosate-exposed mice compared to mice form the control group (Fig 5A, 5C). However, there were no observed differences in the frequency of CD4⁺FOXP3⁺ regulatory T cells (Fig 5B, 5D). An increase in proinflammatory Th17 without any change in the anti-inflammatory Treg population can still shift the balance between Treg/Th17 towards proinflammatory Th17 response.

Figure 5. Glyphosate exposure modulates gut immune homeostasis.

A) Representative flow cytometric plots of CD4⁺ IL17A⁺ T cells or B) CD4⁺ FOXP3⁺ T cells isolated from the lamina propria after 90 days on 10 μg/ml glyphosate or facility water. Cells were stimulated with PMA and ionomycin in the presence of BFA. C) Frequency of CD4⁺ IL17A⁺ or D) CD4⁺FOXP3⁺ T cells. E) Lipocalin-2 levels in fecal after 90 days on 10 μg/ml glyphosate or facility water was determined using ELISA. F) Fecal pH of glyphosate after 90 days post glyphosate exposure (10μg/ml ) measured using pH microelectrode. P-values were determined by Mann Whitney test. *p < 0.05, (n = 4-5)

Additionally, fecal supernatant was collected to determine the effect of glyphosate-exposure on fecal pH and Lipocalin-2 levels, a known marker of gut inflammation. Both fecal Lipocalin-2 and fecal pH were significantly increased in glyphosate-exposed mice after 90 days (Figure 5E and 5F). Thus, our data indicate that exposure to ADI levels of glyphosate can promote gut inflammation, highlighted by an increase in CD4⁺IL17A⁺ T cell frequency, increased sera levels of Lipocalin-2, and higher fecal pH.

4. Discussion

The association between glyphosate and human toxicity remains controversial due to the lack of mechanistic studies directly connecting glyphosate exposure with human disease. Recently, however, glyphosate-induced gut dysbiosis has emerged as a potential mechanism through which glyphosate can affect human health due to the presence of glyphosate sensitive shikimate pathway in bacteria. In this study, utilizing shotgun metagenomic sequencing, we show that glyphosate exposure at levels approximating the U.S. ADI resulted in gut dysbiosis characterized by the depletion of beneficial gut bacteria. Specifically, we observed a reduced abundance of known beneficial bacteria especially Lactobacillus and Bifidobacterium in glyphosate-exposed mice compared to the control group. In the same glyphosate-exposed groups, we also observed changes in gut microbiota gene abundance, especially in pathways associated with SCFA production. Additionally, glyphosate-exposed groups showed increased markers of inflammation, including increased Lipocalin-2, CD4⁺IL17A⁺ gut resident immune cells, and increased fecal pH. Taken together, these data provide strong evidence that glyphosate exposure at the U.S. ADI can cause biologically relevant effects on host physiology through modulation of gut microbiota.

Our microbiome analysis showed that glyphosate exposure similar to the U.S. ADI was sufficient to shift gut microbiota composition. Specifically, longitudinal analysis of phylogenetic diversity showed a significant decrease in diversity over time for 10 μg/ml and 100 μg/ml groups. These data indicate that glyphosate may exert selective pressure on the gut microbiome where bacteria with sensitive EPSPS enzymes may be preferentially depleted, resulting in a competitive advantage for resistant bacterial species over time. Based on amino acid markers in the active site of EPSPS, Puigbo et al. conservatively estimate that 12-26% of bacterial species in the human gut are sensitive to glyphosate (Leino et al., 2021). Additionally, inhibition of EPSPS in gut bacteria is evidenced by the accumulation of shikimate pathway intermediates in rat ceca (Mesnage et al., 2021). This selective pressure on glyphosate-exposed microbiota in 10 μg/ml and 100 μg/ml groups was also apparent from our β-diversity analysis. After 30 days and 60 days of glyphosate exposure, no significant differences were observed between microbial communities. However, after 90 days of exposure, the β-diversity analysis showed distinct separation between the 10 μg/ml and 100 μg/ml groups compared to the control. Therefore, long-term glyphosate exposure is necessary for significant microbial changes to occur. Our findings align with previous studies that highlight exposure time as a limiting factor. Nielsen et al. reported limited effects on murine gut microbiota after 2 weeks of glyphosate exposure (Nielsen et al., 2018). However, exposure to Roundup, a commercial glyphosate-based herbicide preparation at doses below the U.S. ADI significantly impacted gut microbiota composition after 2 years in rats (Lozano et al., 2018).

There were significant differences in taxonomy between the glyphosate-exposed and control microbiota in mice. In the 10 μg/ml group (dose approximating the U.S. ADI), the genus Bifidobacterium was a distinguishing feature of the control microbiome, while Lachnospiraceae and Clostridia were distinguishing features of the glyphosate-exposed microbiome. At the species level, B. pseudolongum showed lower abundance in the 10 μg/ml group compared to control. Bifidobacterium sp., including B. pseudolongum, are known probiotics and provide many beneficial functions to the host, such as enhancing intestinal barrier function, reducing proinflammatory cytokine expression, plus regulating the production and accumulation of reactive oxygen species (Yao et al., 2021). Additionally, in vitro studies have shown that Bifidobacterium are more sensitive to glyphosate in vitro than known pathobionts (Shehata et al., 2013, Kruger et al., 2013), and Bifidobacterium sp. are depleted in the gut microbiome of glyphosate exposed honeybees (Motta et al., 2018). Thus, our data is in agreement with previous studies highlighting the glyphosate sensitivity of Bifidobacterium.

L. reuteri was also depleted in the 10 μg/ml group, which is known to possess immunomodulatory capabilities and aid in defense against pathogen colonization through the secretion of antimicrobial compounds (Mu et al., 2018). Additionally, multiple in vivo studies have shown a reduced abundance of Lactobacillus after glyphosate exposure (Mao et al., 2018, Tang et al., 2020, Lozano et al., 2018). However, unlike Bifidobacterium sp., Lactobacillus sp. lack sensitive EPSPS enzymes and are often deficient in aromatic amino acid biosynthesis pathways (Leino et al., 2021, Zelante et al., 2013, Christiansen et al., 2008, Dempsey and Corr, 2022). Thus, Lactobacillus may be inhibited by glyphosate through an alternative mechanism other than direct inhibition of the shikimate pathway. Interestingly, Lactobacillus are considered nutritionally fastidious microbes and rely on the availability of many amino acids from their environment, including tryptophan. Additionally, tryptophan-rich conditions greatly promote the growth of L. reuteri in vivo (Christiansen et al., 2008, Dempsey and Corr, 2022, Zelante et al., 2013). Therefore, L. reuteri may rely on aromatic amino acid production from other gut microbes inhibited by glyphosate. Alternatively, Lactobacillus sp. may be inhibited by manganese (Mn) chelation. Glyphosate is known to chelate metal ions (Archibald and Duong, 1984), which are important for intestinal health, and Lactobacillus sp. utilize Mn to protect against oxidative damage (Archibald and Duong, 1984, Busbee et al., 2020). However, further research is needed to understand the glyphosate sensitivity of lactic acid bacteria in the murine gut. Additionally, glyphosate exposure at 10 μg/ml and 100 μg/ml doses led to a significantly increased abundance of B. acidifaciens, a species previously associated with murine colitis (Busbee et al., 2020, Berry et al., 2015). B. acidifaciens may be resistant to glyphosate exposure or thrive in the absence of more glyphosate sensitive gut bacteria.

Glyphosate-induced alterations to gut microbiota also led to changes in functional pathways, including the reduced abundance of pathways associated with SCFA production. These pathways included "pyruvate fermentation to acetate and lactate II," "acetylene degradation," and "super pathway of glucose and xylose degradation." SCFA acid production induces regulatory T cells in the colonic lamina propria and prevents gut inflammation (Smith et al., 2013, Waldecker et al., 2008). Nielsen et al. showed a negative correlation between acetic acid and glyphosate concentration in the cecum of rats treated with glyphosate (Waldecker et al., 2008). However, two studies showed no difference in SCFA levels after glyphosate exposure in rats and cows (Billenkamp et al., 2021, Mesnage et al., 2021). Thus, effect of glyphosate exposure on SCFA biosynthesis remains inconclusive and warrants further study. Multiple pathways associated with branched-chain amino acid (BCAA) biosynthesis were also significantly enriched in the control group compared to the 10 μg/ml group. BCAA (valine, isoleucine, and leucine) are hydrophobic and essential amino acids that account for 20-25% of dietary protein. Their importance in host physiology can be highlighted by the fact that they are used as nutritional supplements to improve mental and physical health. BCAA possess anti-inflammatory capabilities and suppress LPS-induced NO production in RAW 264.7 macrophages (Lee et al., 2017). Interestingly, Mesnage et al. also reported a significant decrease in metabolites associated with BCAA metabolism at different concentrations of glyphosate and the herbicide preparation, Roundup (Mesnage et al., 2015). However, it is unclear how changes in BCAA production by gut microbiota might modulate host physiology.

Depletion of beneficial gut bacteria and decrease in anti-inflammatory functional pathways after glyphosate exposure prompted us to determine whether these taxonomic and metabolic alterations were associated with a pro-inflammatory immune response. After 90 days of exposure to 10 μg/ml glyphosate, mice had a significantly increased frequency of gut resident CD4⁺IL17A⁺ T cells (Th17) and increased levels of lipocalin-2, a known marker of gut inflammation (Abella et al., 2015). Seminal studies utilizing Germ-free mice have shown that gut bacteria play a critical role in the induction of Th17 cells, which are linked with proinflammatory diseases such as colitis, multiple sclerosis, and obesity (Ivanov et al., 2009, Tesmer et al., 2008, Carding et al., 2015). Additionally, gut dysbiosis has been shown to promote inflammation at intestinal mucosal surfaces, including increased expression of lipocalin-2, which can favor inflammation via the recruitment of inflammatory cells and induction of proinflammatory cytokines (Carding et al., 2015, Schirmer et al., 2016). These data suggest that exposure to glyphosate may promote a proinflammatory environment through a reduction in L. reuteri, B. pseudolongum, SCFA production, or an increase in Bacteroides sp. However, it remains possible that glyphosate directly impacts gut inflammation, and studies, utilizing high doses of glyphosate, have shown that glyphosate modulates host inflammatory cytokine production (Pandey et al., 2019, Winstone et al., 2022, Qiu et al., 2022). Fecal pH was also significantly increased in 10 μg/ml glyphosate-exposed mice. SCFA-producing bacteria, including Lactobacillus and Bifidobacterium sp., thrive in acidic environments, and the production of SCFA is known to decrease colonic pH and promote the growth of commensal bacteria (Chai et al., 2021). Thus, an increased fecal pH associated with a reduction in known SCFA producers may be important in mediating glyphosate-induced gut dysbiosis.

5. Conclusion

In the present study, we found that glyphosate exposure, at doses similar to the U.S. ADI, can alter gut microbiota composition and modulate the neuro-immune-endocrine system resulting in a pro-inflammatory environment. Microbial alterations were characterized by the loss of beneficial bacteria (Lactobacillus and Bifidobacterium) and a reduction in SCFA producing microbial gene pathways. Additionally, these microbial alterations were accompanied by increased markers of inflammation, including increased Lipocalin-2, CD4⁺IL17A⁺ gut resident immune cells, and increased fecal pH. Collectively, our results suggest that low-dose glyphosate exposure approximating the U.S. ADI is sufficient to modulate gut homeostasis. This study also provides new insights into the mechanisms through which glyphosate effects host physiology. However, further research, including epidemiological studies, is needed to understand if chronic population exposure to glyphosate at current levels can have physiologically relevant effects on human health through modulation of the gut microbiome.

Highlights.

• Low-dose glyphosate exposure modulates gut microbiota composition

• Depletion of Bifidobacterium and Lactobacillus after glyphosate-exposure

• Gut microbial alterations are associated with a pro-inflammatory response

Data availability

The shotgun metagenomic sequences are deposited in NCBI under BioProject PRJNA880821. All other data needed to evaluate the conclusions in the manuscript are present in the manuscript and/or the Supplementary Materials.