The gut microbiota is a fairly complex ecosystem. Luckey (1972) found in 1972 that the number of human intestinal microbiota can reach 10¹⁴ species, and this data has been accepted and quoted by most scholars after publication. The amount of genetic material in the intestinal microbiota exceeds the human genome by more than 100 times (Bäckhed et al., 2005). The intestinal microbial community is diverse and dominated by five main bacterial phyla: Bacteroidetes, Firmicutes, Actinobacteria, Proteobacteria and Verrucomicrobia. Among these five phyla, the phylum Bacteroidetes and the phylum Firmicutes accounted for more than 90% of the total population. Most of the bacteria under the phylum Bacteroidetes belong to the genera Bacteroides and Prevotella. And bacteria under the phylum Firmicutes that dominate the gut microbiota include the genera Clostridium, Eubacterium and Ruminococcus (Eckburg et al., 2005). The proportion of various microbiota is different depending on the site of the GI tract. The upper GI tract was mainly enriched with Firmicutes, Proteobacteria and Lactobacillaceae, while the microbiota of lower GI tract was mainly composed of Bacteroidetes, Firmicutes, and A.muciniphila (Scheithauer et al., 2016; Meijnikman et al., 2018). In addition to taxonomic classification, the human microbiota can be categorized into three different enterotypes: Bacteroides, Prevotella or Ruminococcus. Bacteroides-dominated enterotype is characterized by saccharolytic and proteolytic activities involved in synthesis of riboflavin, biotin, ascorbate and pantothenate. The enterotype of Prevotella-dominated plays the role of mucin glycoprotein degraders involved in the synthesis of thiamine and folate. The last enterotype is dominated by Ruminococcus, characterized by membrane transportation of sugars and mucin degrading activities (Arumugam et al., 2011). A study from Taiwan categorized enterotypes into Bacteroides, Prevotella and Enterobacteriaceae and claimed that Enterobacteriaceae may be a new subtype in Asian populations (Liang et al., 2017). However, the concept of enterotypes has been debated due to the high degree of variability observed in the gut microbiota between individuals and the fact that many of the data show that there are not three completely discrete clusters (Knights et al., 2014). It is indisputable that regardless of enterotypes, some members of the microbial population act as a “core microbiota,” while others are more of a “flexible pool.” The “core microbiota” is composed of host-adapted microbes reproducibly included in the gut microbiota made up of different environmental combinations and determined by genetic factors. While the “flexible pool” is usually obtained from water, food and various components of the environment, which contributes to the adaptation of the host (Shapira, 2016).

Human cells coexist with bacteria for a long time, and there is a complex process of material and energy exchange between them. Intestinal microbiota plays a crucial role in many key metabolic processes of human body, such as SCFAs, amino acids, bile acids and vitamin synthesis (Montandon and Jornayvaz, 2017). There is no doubt that the gut microbiome is extremely relevant to human health and disease status. In many disease states, such as metabolic syndrome, inflammatory bowel disease (IBD), cardiovascular diseases, malignant tumors, etc., the structure and function of intestinal microbiota are obviously different from those of healthy bowel microbiota (Tilg and Moschen, 2014; Shin et al., 2017; Harsch and Konturek, 2018; Eibl and Rozengurt, 2019). Regulating the structure and function of intestinal microflora can improve the disease status of the body to a certain extent. So, targeting the gut microbiome may provide new therapeutic thoughts for some intractable diseases. In particular, the role of gut microbiota in T2DM and obesity has attracted the attention of many researchers (Boulangé et al., 2016; Pedersen et al., 2016; Cani, 2018; Chobot et al., 2018; Zhao et al., 2018). Gut bacterial therapies that improve insulin sensitivity by altering the composition of the gut microbiota have become a new therapeutic modality.

Study has shown that changes in modern lifestyles have led to a decrease in the diversity of the gut microbiota in many developed populations (Clemente et al., 2015). This reduced diversity may promote the development of metabolic disorders. Individuals with low gut microbiota abundance have been found to be susceptible to obesity, insulin resistance and dyslipidemia (le Chatelier et al., 2013). Lactobacillus, Prevotella, Bacteroides, Desulfovibrio, and Oxalobacter spp. are decreasing in the gut microbiota of low-gene-count patients compared with high-gene-count patients. Functional changes in the microbiota of low-gene-count patients primarily consist of a decrease in butyrate-producing bacteria and an increase in the ratio of Akkermansia to Ruminococcus gnavus, which lead to decreased methane production potential, decreased hydrogen production potential, enhanced mucus degradation, and increased peroxidase activity (le Chatelier et al., 2013). The metabolic disturbances resulting from this imbalance of anti-inflammatory and pro-inflammatory bacterial species puts people at increased risk of suffering from T2DM (Ouchi et al., 2011). Approximately 86% of patients with T2DM are overweight or obese, and obesity is considered the greatest risk factor for T2DM (Daousi et al., 2006; Singer-Englar et al., 2019). A number of studies have reported significant changes in the gut microbiota of patients with obesity and T2DM compared to healthy adults (Qin et al., 2012; Karlsson et al., 2013; Tims et al., 2013; Forslund et al., 2015; Dao et al., 2016; le Roy et al., 2018), manifested in an overall increase in the abundance of Bacteroides and Prevotella copri as well as a decrease in the abundance of Akkermansia, Roseburia and Ruminococcus.

Metformin comes from Galega officinalis L., and later the function of lowering blood glucose has been found in animals, and more and more studies on its hypoglycemic effect have been conducted, and the effect has been furtherly confirmed. There has also been a great deal of research on its mechanisms. Stepensky et al. (2002) stated that metformin was better able to lower blood glucose by intestinal administration. Subsequently, researchers boldly hypothesized that metformin acted through the gut. Duca et al. (2015) found that the hypoglycemic effect of metformin was reduced by blocking the cAMP pathway in the gut, and speculated that the target of metformin was mainly in the gut. Buse et al. (2016) confirmed again that the gut is the main locus of hypoglycemic effects of metformin through comparable plasma levels of metformin. Meanwhile, due to its safety and few side effects, metformin has been widely studied, and it has been found that its effect is not only limited to reducing hyperglycemia, but also delaying the development of diabetes complications (Scheen and Paquot, 2013; Yaribeygi et al., 2019; American Diabetes Association, 2020). Therefore, people have found that it has many extra-hypoglycemic effects, such as anti-inflammatory, anti-tumor, anti-aging and so on. Among them, the anti-inflammatory effect has attracted the most attention, because in many diseased states, besides specific symptoms the body is generally in a state of chronic inflammation.

The mechanism of metformin has been studied endlessly, but the mechanism has not yet been fully clarified. In recent years, with advances in gene sequencing technology, more and more research has focused on the effects of metformin on the gut microbiota. It has been possible to confirm the association of metformin with certain intestinal microbiota. Most animal studies (Lee et al., 2018; Wang Z. et al., 2018; Zheng et al., 2018; Brandt et al., 2019; Ji et al., 2019; Ahmadi et al., 2020; Chung et al., 2020; Ryan et al., 2020) used C57BL/6 mice to construct a high-fat diet (HFD)-induced diabetes model followed by metformin treatment. Some studies have also used db/db mice (Chen et al., 2018; Zhang et al., 2019) or KKAy mice (Gao et al., 2018). Although the structure of the intestinal microbiota is very complex and the experimental design varies among studies, such as the different timing of HFD induction, distinctions in the duration of metformin treatment, and variations in mouse strains, but the impact of metformin on gut microbiota derived from these studies are nearly consistent. For example, the proportions of phyla Bacteroidetes (Gao et al., 2018; Ahmadi et al., 2020; Ryan et al., 2020) and Verrucomicrobia (Gao et al., 2018; Lee et al., 2018; Ryan et al., 2020) and genera Akkermansia (Chen et al., 2018; Gao et al., 2018; Lee et al., 2018; Zheng et al., 2018; Ji et al., 2019; Zhang et al., 2019; Ryan et al., 2020) and Bacteroides (Chen et al., 2018; Gao et al., 2018; Lee et al., 2018; Ryan et al., 2020) were significantly increased in the metformin treatment group. Similar results were also manifested in the rat model as in the mouse model (Zhang et al., 2015; Cui et al., 2019). Details are shown in Table 1.

However, the results in humans are not entirely consistent with those in animals. This may be due to the significant differences in gut microbial diversity between humans and animals (Peng et al., 2013). Moreover, clinical trials in patients with diabetes are affected by many complex factors, such as diet, race, comorbidities, and drug combinations. An analysis of a T2DM cohort based on metagenomic sequencing and validated in an independent cohort found that metformin-treated patients had an increase in Escherichia coli and a decrease in Intestinibacter, and that the abundance of several intestinal microbial genera was more similar to that of normal control levels, particularly Akkermansia (Forslund et al., 2015). Moreover, A double-blind study of patients with T2DM treated with metformin led to significant changes in the relative abundance of more than 80 bacterial species compared with placebo after 4 months, in which an increase in Escherichia and Akkermansia muciniphila and a decrease in Intestinibacter were also observed, whereas most of the changes with placebo occurred in phylum Firmicutes and phylum Proteobacteria (Wu et al., 2017). Similar results were found in several studies on the effects of metformin on the gut microbiota of patients with obesity or T2DM (de la Cuesta-Zuluaga et al., 2017; Hiel et al., 2020), and even a study on metformin-induced changes in the gut microbiota of healthy young men likewise observed an increase in Escherichia/Shigella and a reduction in Intestinibacter (Bryrup et al., 2019). Details are shown in Table 1.

Nevertheless, the alterations in the gut microbiota of diabetic individuals by metformin treatment show consistent results in certain bacteria, in both humans and rodents. This was almost always demonstrated by enrichment of mucin-degrading A. muciniphila and SCFA-producing bacteria such as Bifidobacterium bifidum and Butyrivibrio (de la Cuesta-Zuluaga et al., 2017). However, this discrepancy between the human and animal studies suggests that we should still focus our researches in clinical setting and interpreting the mechanisms of action of the human gut microbiota. However, the reported results on the effect of metformin on human gut microbiota diversity are inconsistent. In patients with newly diagnosed T2DM, Tong et al. (2018) reported an increase in gut microbiota diversity with metformin based on Chao1 enrichment estimates, whereas Sun et al. (2018) observed a slight decrease in alpha diversity. In studies of healthy individuals, Bryrup et al. (2019) found no significant changes in gut microbial richness and diversity after metformin treatment. However, Elbere et al. (2018) reported a significant decrease in gut microbiota diversity after 24 h of metformin administration using the Shannon index. Overall, the therapy of metformin was associated with changes in the abundance of specific bacterial genera in the human gut microbiota, but the effects on gut microbiota diversity were variable, which highlights the need for further research to understand the potential mechanisms and clinical significance.

Akkermansia muciniphila, a strictly anaerobic, endospore-free ovoid gut bacterium, is true of the phylum Verrucomicrobia (Derrien et al., 2004). It mainly colonizes the outer mucus layer of the GI tract, and uses the mucin of the GI tract as a carbon and nitrogen source to maintain its growth. The dynamic balance between the consumption of mucin and the production of mucin by goblet cells is obtained to maintain the stability of the mucus layer (Sicard et al., 2017). And the abundance of A. muciniphila is obviously reduced in patients with obesity, T2DM, and cardiovascular disease. At the same time, a large number of studies have shown that the content of mucin-degrading bacteria A. muciniphila in the intestinal microbiota is significantly increased after the administration of metformin (Hur and Lee, 2015; de la Cuesta-Zuluaga et al., 2017; Wu et al., 2017; Lee et al., 2018; Verdura et al., 2019), which can make it reach 20% of the total microbiome (de la Cuesta-Zuluaga et al., 2017; Cani, 2018). Moreover, some studies have verified that metformin directly promotes the growth of this bacterium in vitro (Wu et al., 2017).

SCFAs (such as acetic acid, propionic acid and butyric acid) are usually metabolites produced by various SCFAs-producing bacteria in the gut that metabolizes carbohydrates such as dietary fiber (Rau et al., 2018), and they are essential for the health of the gut, the body and even the brain. SCFAs can affect brain function by directly or indirectly regulating the gut-brain axis through immune, vagal, endocrine and other humoral pathways (Dalile et al., 2019). They also suppress appetite as well as regulate energy homeostasis by stimulating the secretion of hormones such as GLP-1 from intestinal L cells (De Silva and Bloom, 2012). There is evidence that metformin can increase the abundance of microbiota which can all produce SCFAs such as Butyrivibrio, Bifidobacterium, Megasphaera, Prevotella, and so on (de la Cuesta-Zuluaga et al., 2017).

Lactobacillus is a class of anaerobic or partly anaerobic, non-bacteriophage bacteria belonging to the phylum Firmicutes. The bacteria of this genus have a strong ability to break down sugar, and the main end product is lactic acid, which can protect the intestinal barrier and reduce the inflammatory response. Metformin has been found to increase the abundance of Lactobacillus in rodents with HFD (Zhou et al., 2016; Bauer et al., 2018). In recent studies, this trend has also been observed in certain clinical studies in patients with T2DM (Xu et al., 2022).

One study Forbes et al. (2018) indicated that the gut microbiota of patients with immune-mediated inflammatory diseases such as IBD, ankylosing spondylitis, rheumatoid arthritis and systemic lupus erythematosus differed sharply from normal controls. For example, the abundance of Intestinibacter is sensibly higher in the gut of patients with Crohn’s disease (CD). In addition, Bacteroides fragilis is the only bacterium that has been proved to beget abscess formation, has a potent virulence factor, and is the most common anaerobic pathogen (Wexler, 2007). Other bacteria associated with the disease are Clostridioides difficile [here we use its new nomenclature according to Zheng et al. (2020)], an anaerobic enteric pathogen that can cause severe diarrhea and lead to death (Sandhu and Mcbride, 2018). Metformin declines the number of disease-associated bacteria such as Intestinibacter spp., Bacteroides fragilis, and Clostridioides difficile (de la Cuesta-Zuluaga et al., 2017; Elbere et al., 2018; Bryrup et al., 2019).

The GI mucosal barrier with important physiological roles is a three-dimensional protective structure consisting of mechanical, chemical, immune and biological barriers. Many of the mechanisms of metformin action revolve around stabilization of the GI mucosal barrier. An emphatic increase in the expression of two markers of mucin levels, MUC2 and MUC5 genes, was observed after metformin treatment (Kyriachenko et al., 2019). This suggests that the increased abundance of A. muciniphila after metformin administration may be due to the fact that metformin increases the number of goblet cells, providing more substrate for the growth of A. muciniphila. Moreover, A. muciniphila was discovered to improve HFD-induced intestinal hyperpermeability and also affect intestinal barrier function by the way that upregulates the expression of tight junction and closure proteins (Li et al., 2016). In this way, it decreases the entry of pro-inflammatory lipopolysaccharides (LPS) into the circulation, thus reducing inflammation in the organism. Additionally, Everard et al. (2013). measured the concentration of endocannabinoids (ECs) in the gut of the mice treated with A. muciniphila and found that A. muciniphila also raised the release of ECs. Increased endocannabinoids inhibit monoacylglycerol lipase (Alhouayek et al., 2011), thereby reducing systemic inflammation, altering intestinal peptide secretion and increasing intestinal mucosal barrier thickness. It is well known that T2DM is a metabolic and inflammatory disease characterized by deteriorating islet function and elevated levels of inflammatory cytokines. Therefore, therapies targeting inflammation can also recover glycemic control in T2DM patients. In a study, the treatment with three different strains of A. muciniphila was found to reverse the low-grade chronic inflammatory state in HFD mice to a certain extent, including a significant increase in the gene expression of inflammatory and immunosuppressive factor interleukin (IL)-10, as well as a decrease in the mRNA levels of tumor necrosis factor-α (TNF-α), Monocyte chemoattractant protein-1, and Toll-like receptor 2 (TLR2) (Yang et al., 2020). In terms of alleviating inflammation, A. muciniphila also modulates Forkhead box protein 3 (Foxp3) in mouse adipose tissue (Choi et al., 2021), resulting in a diminution of regulatory T cells (Tregs), and anti-inflammatory while indirectly improving glucose homeostasis. In addition, a specific membrane protein, Amuc_1100, exists on the outer membrane of A. muciniphila, and many studies have reported that it induces the production of specific cytokines by binding to TLR2 and activating the downstream pathway of TLR2 (Figure 1), which in turn improves the host’s immune homeostasis and intestinal mucosal barrier function (Plovier et al., 2017; Song et al., 2023; Wang et al., 2024), and then improves the blood glucose level. The protein retains its active conformation and functions even after pasteurization (Plovier et al., 2017). Amuc_1100 protein also has the function of lowering blood lipids, which can reduce the content of low-density lipoprotein (LDL) and cholesterol (Lee et al., 2018). Plus, vesicles secreted by A. muciniphila reduce the expression of Toll-like receptor 4 (TLR4) (Figure 1), which impacts the nuclear factor-κB (NF-κB) pathway and thus lessens the secretion of the pro-inflammatory factors IL-6, IL-8 (Ottman et al., 2017; Ashrafian et al., 2019).

What’s more, metformin can increase the abundance of Bifidobacterium which can promote the growth of gastric mucin, thereby stabilizing the GI mucosa. Cani et al. (2009) identified that after increasing the quantity of Bifidobacterium in the intestine, the content of proglucagon mRNA was higher, which promoted the secretion of glucagon-derived peptides. During the course, glucagon-like peptide-2 (GLP-2) enhanced intestinal epithelial cells proliferation and reduced gut permeability, thereby stabilizing the intestinal mucosal barrier.

For the gut, SCFAs have important roles in regulating intestinal microbiota balance, maintaining electrolyte balance and improving intestinal immune function. Administration of metformin increases the abundance of SCFAs-producing microbiota (described previously, see Chapter 4.3), thereby increasing the production of SCFAs. The produced SCFAs are partially absorbed rapidly by colonocytes via monocarboxylate transporter protein (MCT), then undergo a series of oxidative reactions, and finally provide energy to the cells in the form of ATP (Dalile et al., 2019). The SCFAs not metabolized in colonocytes enter the portal circulation of the liver through the basolateral membrane and provide substrates for energy metabolism in hepatocytes. SCFAs are also implicated in the biosynthesis of glucose, fatty acids and cholesterol in hepatocytes. After the above metabolism, only a small fraction of SCFAs remains to reach the whole body through the blood circulation (Dalile et al., 2019).

After SCFAs are absorbed in the colon, it works further via action on the G protein-coupled receptors GPR43 and GPR41 (Figure 1). SCFAs stimulate the proliferation of colonic epithelial L cells and are bound to GPR43 above them, increasing intracytoplasmic calcium ion and cyclic adenosine cAMP concentrations and increasing GLP-1 secretion (Chambers et al., 2018; Rodriguez et al., 2018) which inhibits glucose uptake by intestinal wall cells. Activation of GPR41 increases the secretion of intestinal peptide YY (PYY) and decreases the secretion of glucose-dependent insulin-releasing polypeptide (GIP) and growth hormone-releasing peptide, thereby reducing glucose uptake (Hu et al., 2018). GPR43 is distributed in adipose tissue, bone marrow, spleen, pancreas, peripheral blood mononuclear cells, small intestine, and mammary gland, while GPR41 is distributed in various tissues, especially in immune cells. Hence, SCFAs also have beneficial effects on peripheral tissues, controlling substrate metabolism and optimizing systemic insulin sensitivity. For example, in immune cells, SCFAs can inhibit inflammation by binding to GPR41 and GPR43 (Allin et al., 2015; Thorburn et al., 2015). SCFAs also induce browning of white adipose tissue, boost glucose tolerance, control blood glucose levels, and exert anti-obesity and anti-diabetic effects (He et al., 2020).

Moreover, SCFAs are inhibitors of histone deacetylase (HDAC). SCFAs inhibit HDAC of Treg, in turn affecting Treg production (Arpaia et al., 2013; Furusawa et al., 2013). Also, the inhibition of HDAC by SCFAs can further modulate the function of intestinal macrophages (Chang et al., 2014) and dendritic cells (Trompette et al., 2014; Meijnikman et al., 2018) and downregulate LPS-induced pro-inflammatory mediators, such as NO, IL-6, and IL-12, resulting in anti-inflammatory effects. There is also evidence that SCFAs decline the release of TNF-α from neutrophils (Lee et al., 2019), which is also one of the typical pro-inflammatory cytokines that trigger subclinical inflammation.

Notably, SCFAs also play a crucial role in maintaining the Intestinal mucosal barrier integrity. SCFAs can increase transcription of mucin genes (described previously, see Chapter 4.1), then upgrading intestinal barrier function (Willemsen et al., 2003; Gaudier et al., 2004). At the same time, the increased A. muciniphila would continue to degrade mucin and produce more SCFAs. A prospective study by Dao et al. (2016) verified the relationship. This study spotted that individuals with higher abundance of A. muciniphila had a better metabolic profile, and the concentration of SCFAs in such populations was positively correlated with A. muciniphila abundance.

Bile acids are a class of amphiphilic metabolites produced by hepatocytes and secreted by the bile ducts into the intestine, which have an important emulsifying effect on fat ingested by the body during digestion. The substantial increase in plasma bile acid concentrations (primary, secondary, total and unconjugated bile acid concentrations) with metformin may be explained by the fact that metformin increased the abundance of probiotic bacteria, especially Lactobacillus, and thus their bile salt hydrolase (BSH) gene expression. The BSH gene, a gene encoding the production of BSH by the intestinal microbiota, is associated with obesity and T2DM was significantly correlated (Jia et al., 2020). An essential part of cholesterol regulation by gut microbes is achieved by the hydrolysis of cross-linked bile salts by BSH, thus reducing serum cholesterol levels. The de-cross-linked bile acids have low water solubility and are readily secreted into the feces. Due to the loss of fecal bile acids and the decrease in bile salts, the liver increases the conversion of cholesterol to bile salts to maintain bile salt metabolic balance. It has been demonstrated that oral administration of BSH-containing Lactobacillus can reduce serum cholesterol levels in animals and humans, especially Limosilactobacillus fermentum [here we use its new nomenclature according to Naghmouchi et al. (2020), Zheng et al. (2020), Çiftci and Tuna (2021), Ng et al. (2022)]. However, no studies have revealed a correlation between the presence of specific bacteria and bile acid concentrations after metformin treatment (Wu et al., 2017).

Subsequently, secondary bile acids, key signaling molecules produced by bile acids and gut microbiota metabolism, are bound to membrane receptors (Takeda G protein-coupled receptor 5, TGR5/G protein-coupled bile acid receptor 1, GPBAR1) and nuclear receptors (Farnesoid X Receptor, FXR), which stimulates colonic secretion of GLP-1 and PYY. It plays a role in promoting proliferation and inhibiting apoptosis in pancreatic 𝛽-cells, and directly scales insulin secretion up. Another study (Sun et al., 2020) demonstrated that serum fibroblast growth factor 19 (FGF19) and bile acid levels were simultaneously increased in patients with T2DM early in life. This may be due to the fact that bile acids entering the intestine stimulate intestinal secretion and expression of the gut-derived hormone FGF19 via bile acid-FXR signaling (Mulla et al., 2019), which stimulates hepatic glycogen synthesis in an insulin non-dependent manner, restrains hepatic gluconeogenesis, and exerts hypoglycemic effects. FGF19 also dramatically promotes glucose tolerance (Zhang et al., 2019).

There are other pathways for FXR signaling activation. Sun et al. (2018) showed that metformin treatment altered the metabolism of folate and methionine in mice, which in turn inhibited the growth of B. fragilis. The reduction of B. fragilis increased the level of Glycoursodeoxycholic acid in the gut, then led to the activation of intestinal FXR signaling. Several studies have implied that FXR activation relieves intestinal inflammatory state in IBD patients, as well as that colonic inflammation in patients with CD and in rodent models of colitis is linked to diminished expression of FXR mRNA (Vavassori et al., 2009). During inflammation in the liver, there is a sustained inhibition of bile salt export pump expression and reduced FXR expression. Hepatic transporter protein function is then decreased, which leads to both an increase in hepatic bile acid chelation and persistent inflammation. Activation of FXR stabilizes the nuclear co-repressor (NcoR) at the NF-κB response element on the IL-1β promoter, while decreasing the expression of TNF-α, IL-1β, IL-2, IL-6 and IFN-γ mRNA, thereby reducing the extent of inflammation (Vavassori et al., 2009).

In addition to regulating the gene expression of BSH of probiotic bacteria (described previously, see Chapter 4.3), metformin also modulates the protein-coding genes of another microbiota. Wu et al. (2017) discovered that the abundance of E. coli, Bifidobacterium adolescentic, and A. muciniphila was plainly increased in the intestine after metformin administration. They further proceeded transcriptomic analysis of fecal samples and found that some of the protein-coding genes of A. muciniphila and B. wadsworthia were dramatically regulated by metformin, and most of them encoded metalloproteins or metal transporter proteins. Metal ions (e.g., Ca2+, Mn2+, Zn2+, and Mg2+) play an important role in maintaining the structure and homeostasis of various proteins in microorganisms. And these proteins affect the metabolism of microorganisms (Murphy et al., 1981). Several studies have shown that metallothionein has a negative regulatory role in all kinds of organ tissues, various types (including LPS-related, allergic and oxidative) of inflammation (Inoue et al., 2009; Inoue and Takano, 2013). It has also been indicated that metformin is able to bind to metal ions, thus causing changes in chemical structure (Logie et al., 2012). However, further metabolomic and proteomic studies are still needed to investigate the mechanisms of relevant microbial metabolism-host interactions.

Metformin was found to reduce the number of C. difficile in the intestine in several studies (Elbere et al., 2018; Bryrup et al., 2019). C. difficile can release toxin A and toxin B (Popoff, 2018), which induce the production of pro-inflammatory cytokines and inflammatory mediators in a variety of cell lines (Zhang and Feng, 2016). And toxin A is able to trigger IL-8 secretion by human intestinal cells (He et al., 2002). It has also been reported that prostaglandin E2 (PGE2) secretion and cyclooxygenase-2 (COX-2) expression were significantly increased in colon cells exposed to toxin A either in vivo or in vitro (Kim et al., 2007). Therefore, metformin is able to decrease the release of pro-inflammatory factors by a reduction in the number of opportunistic pathogenic bacteria.