Akkermansia muciniphila is a promising probiotic

Summary Akkermansia muciniphila (A. muciniphila), an intestinal symbiont colonizing in the mucosal layer, is considered to be a promising candidate as probiotics. A. muciniphila is known to have an important value in improving the host metabolic functions and immune responses. Moreover, A. muciniphila may have a value in modifying cancer treatment. However, most of the current researches focus on the correlation between A. muciniphila and diseases, and little is known about the causal relationship between them. Few intervention studies on A. muciniphila are limited to animal experiments, and limited studies have explored its safety and efficacy in humans. Therefore, a critical analysis of the current knowledge in A. muciniphila will play an important foundation for it to be defined as a new beneficial microbe. This article will review the bacteriological characteristics and safety of A. muciniphila, as well as its causal relationship with metabolic disorders, immune diseases and cancer therapy.


Introduction
Several microbial species are getting increasing attention for their role in modulating the gut microbiota. At present, many diseases and conditions have been reported to be closely related to gut microbiota , so it is of great interest to improve the host health by modulating the intestinal bacteria. Akkermansia muciniphila (A. muciniphila) is a strict anaerobe recently isolated from human faeces and uses the mucin as the sole sources of carbon and nitrogen elements (Derrien et al., 2004). This mucin degrader is affected by the nutrients in the mucus layer located at a close distance to the intestinal epithelial (Belkaid and Hand, 2014). Due to this unique function and its high universality and richness in almost all life stages, A. muciniphila has opened new avenues for the application in next-generation therapeutic probiotics (Collado et al., 2007;Derrien et al., 2008;Belzer and de Vos, 2012;Cani and de Vos, 2017). A series of studies have revealed that A. muciniphila regulated metabolic and immune functions, thus protecting mice from high-fat diets Everard et al., 2013). Further analysis confirmed A. muciniphila can degrade mucin and exert competitive inhibition on other pathogenic bacteria that degrade the mucin (Belzer and de Vos, 2012). These findings provide a rationale for A. muciniphila to become a promising probiotic. However, products containing A. muciniphila are currently not available worldwide. The exact mechanism underlying A. muciniphila interacts with host remains unknown. Based on previous human and animal studies, extensive assessment for A. muciniphila is still needed. Here, we will summarize and provide the updated information on the bacteriological characteristics, safety, pathogenicity, antibiotic resistance of A. muciniphila and its effects on host health and diseases.

Characteristics of A. muciniphila
Akkermansia muciniphila is a bacterium of oval shape, strictly anaerobic, non-motile and gram-negative and forms no endospores (Fig. 1). It was historically discovered in 2004 at Wageningen University of the Netherlands when searching for a new mucin-degrading microbe in human faeces (Derrien et al., 2004). Akkermansia muciniphila is the first member and the only representative of the phylum Verrucomicrobia in the human gut (Miller and Hoskins, 1981;Derrien et al., 2010), which is relatively easy to detect (Rajilic-Stojanovic and de Vos, 2014). The genome of A. muciniphila strain MucT (=ATCC BAA-835T=CIP 107961T) involves one circular chromosome of 2.66 Mbp, which shared a limited number of genes (29%) with its closest relatives in the Verrucomicrobia phylum (van Passel et al., 2011). Recently, Guo et al. (2017) reported a high genetic diversity of A. muciniphila by whole-genome sequencing, with 5644 unique proteins assembling a flexible open pangenome. They further classified A. muciniphila into three species-level phylogroups, which demonstrated different function features.
It is widely distributed in the intestines of human and animals (Belzer and de Vos, 2012;Lagier et al., 2015). Akkermansia muciniphila was originally classified as a strictly anaerobic bacterium, but a recent study found that it can tolerate low levels of oxygen, with an oxygen reduction capacity to be 2.26 AE 0.99 mU mg À1 total protein (Ouwerkerk, et al., (2017b). This property is similar to some intestinal anaerobic colonizers such as Bacteroides fragilis and Bifidobacterium adolescentis, which could still survive after exposure to ambient air for 48 h. Akkermansia muciniphila is abundant in the host intestinal mucosal layer, with a largest number in the caecum. It is found to be ubiquitous in the guts of healthy adults and infants, and accounts for 1-4% of the total gut microbiota starting from early life (Derrien et al., 2008).
Akkermansia muciniphila is one of the normal gut symbionts throughout our life (Collado et al., 2007). This bacterium can stably colonize the human gut within 1 year after birth, and its abundance in the gut eventually reaches the same level as that in healthy adults (Collado et al., 2007;Derrien et al., 2008), but gradually decreases in the elderly (Collado et al., 2007). Previous phylogenetic and metagenomic studies based on hundreds of subjects have found that A. muciniphila is one of the top 20 most abundant species detectable in the human gut (Collado et al., 2007(Collado et al., , 2012Qin et al., 2010;Arumugam et al., 2011;Thomas et al., 2014;Drell et al., 2015). In addition, A. muciniphila is reported to be present in human milk . Human milk can act as a carrier for the transfer of A. muciniphila from mothers to infants, thereby explaining its presence in the gastrointestinal tract of newborn infants (Collado et al., 2007). At this life stage, A. muciniphila can successfully colonize the gastrointestinal tract with the active acid resistance system and the ability to degrade human milk oligosaccharides in newborn infants' stomach (Bosscher et al., 2001).

Culturing A. muciniphila
Akkermansia muciniphila is divided into three specieslevel phylogenetic groups with distinct metabolic features, but current studies still focused on the strain MucT (=ATCC BAA-835T=CIP 107961T) (Guo et al., 2017). Akkermansia muciniphila is sensitive to oxygen, and its growth medium is animal-derived compounds. Therefore, the clinical application of A. muciniphila is very limited due to these limitations in culture conditions. Ottman et al. (2017a,b) established a genome-scale metabolic model to evaluate the substrate utilization abilities of A. muciniphila. It showed that A. muciniphila can utilize the mucin-derived monosaccharides fucose, galactose and N-acetylglucosamine. These additional mucinderived components might be needed for its optimal growth. Plovier et al. (2017) reported that A. muciniphila can be grown on a synthetic media, in which the mucin is replaced by a combination of glucose, N-acetylglucosamine, peptone and threonine. This synthetic medium is capable of culturing A. muciniphila at the same efficiency as the mucin medium, while avoiding all compounds that are incompatible with humans. At the same time, A. muciniphila grown on synthetic media was confirmed to be safe for human administration (Plovier et al., 2017). A recent study reported that the genome-scale metabolic model can be used to accurately predict growth of A. muciniphila on synthetic media (van der Ark et al., 2018). They found that glucosamine-6-phosphate (GlcN6P), which exists in the mucin and prompts the adaptation to the mucosal niche, is a necessity for A. muciniphila.
Moreover, Ouwerkerk et al. (2017a,b) proposed an efficient scalable workflow for the preparation and preservation of viable cells of A. muciniphila under strict anaerobic conditions for therapeutic interventions. An anaerobic plating system was used in this process to quantify the recovery and survival of viable cells of A. muciniphila. The preserved A. muciniphila cells showed very high stability with survival rate of 97.9 AE 4.5% for over 1 year at À80°C in glycerol-amended medium. These results might pave a way for future clinical studies using A. muciniphila as a therapeutic product.

Safety and pathogenicity of A. muciniphila
Currently, a large number of researches on A. muciniphila mainly focused on explaining its relationship with diseases, but have not addressed the causality of the bacterium on the diseases (Tables 1 and 2). Several studies focusing on the direct interventions with A. muciniphila mostly used animal models (Everard et al., 2013;Hanninen et al., 2017;Chelakkot et al., 2018) (Table 3). Currently, there are no published open clinical trials of A. muciniphila for humans and therefore resulting in a lack of strong evidence on the safety of A. muciniphila in humans. This could explain why A. muciniphila has not been involved in food production or drug use. However, some preliminary studies have indicated this bacterium should be safe for interventions in human. Dubourg et al. (2013) reported that even when the abundance of A. muciniphila reached a high level of 60% in human following broad-spectrum antibiotic treatment, no adverse events occurred. Moreover, in an ongoing clinical study, Plovier et al. (2017) have first evaluated the safety and tolerability of A. muciniphila in overweight subjects. Both live and pasteurized A. muciniphila were observed to be tolerated and safe in individuals with excess body weight after 2-week oral administration of A. muciniphila.
As for the pathogenicity of A. muciniphila, it has not yet been clearly associated with any disease or sign of illness (Derrien et al., 2010). The potential pathogenicity of A. muciniphila was mainly due to its process from adhesion to degradation of the intestinal mucus layer, which may involve some initial pathogenic behaviours (Donohue and Salminen, 1996;Tuomola et al., 2001;Derrien et al., 2010). Unlike pathogens, A. muciniphila as a mucin-degrading agent mainly stays in the outer mucosal layer and does not reach the inner mucosal layer, but bacteria reaching the inner layer have been shown to be required for pathogenicity (Gomez-Gallego et al., 2016). Although degrading mucin itself is a pathogen-like behaviour (Donohue and Salminen, 1996), it is considered a normal process in the intestinal selfrenewal balance (Gomez-Gallego et al., 2016). Moreover, it is reported that A. muciniphila may maintain host intestinal microbial balance by converting mucin into beneficial by-products (Derrien et al., 2008). To date, there is no evidence that A. muciniphila alone causes pathogenicity; nevertheless, it is not known whether it may cause diseases in synergy with other bacteria.
Akkermansia muciniphila, as a gram-negative bacterium, contains lipopolysaccharide, but it is not associated with endotoxemia. This bacterium even reduced the endotoxin level associated with high-fat diets in mice (Everard et al., 2013). Mucin degradants are known to regulate host immune system through signals such as tumour necrosis factor alpha (TNF-a), interferon gamma (INF-c), interleukin-10 (IL-10) and IL-4 Collado et al., 2012;Andersson et al., 2013). There was evidence that a decreased level of the antiinflammatory cytokines IL-10 and IL-4 and an elevated level of pro-inflammatory cytokines TNF-a and IFN-c were associated with an increased level of A. muciniphila (Collado et al., 2012). From a genetic point of view, colonization of A. muciniphila in sterile mice did not cause side-effects or the upregulated expression of pro-inflammatory cytokines . Intestinal anti-inflammatory and protective effects were thought to be closely related to A. muciniphila (Png et al., 2010;Candela et al., 2012). Hence, we suggest that treatment with A. muciniphila should be safe with a rationale.

Colonization of A. muciniphila and its interaction with the host
The ability of A. muciniphila to adhere to the mucus layer was considered to be a beneficial probiotic characteristic (Derrien et al., 2010;Everard et al., 2013;Chelakkot et al., 2018;Hanninen et al., 2018). The intestinal mucosal layer mainly protects epithelial cells from microbial attacks and provides growth energy for microorganisms that use it as a nutrient. A low level of A. muciniphila in the intestine may result in the thinning of the mucosa, thus leading to a weakening of the intestinal barrier function, and making it easier for the toxins to invade the host. The relationship between A. muciniphila and the host is not only reflected in the intake, utilization and consumption of energy associated with glucose, protein and lipid metabolism, but also in the integrity of mucosal layer and related mucosal immune response. Akkermansia muciniphila not only participates in the host immune regulation, but also enhances the integrity of the intestinal epithelial cells and the thickness of the mucus layer, thereby promoting intestinal health (Everard et al., 2013;Reunanen et al., 2015).
Microorganisms on the surface of the intestinal mucosa are known to contribute more to host immunity, and A. muciniphila is a typical representative (Nieuwdorp  , 2014). The host's nutrient environment could affect the growth of A. muciniphila in the intestine. For example, the property of A. muciniphila degrading mucin can be defined as a competitive advantage when the host is in nutritional deficiencies such as during fasting and in malnutrition. This was confirmed by the experiment on hamsters that the abundance of A. muciniphila was significantly increased after fasting (Sonoyama et al., 2009). The level of mucin in the intestine of rats fed with arabinose or inulin was significantly increased, and this change also contributed to the abundance of A. muciniphila.
In turn, the host will also benefit from the colonization of A. muciniphila. A. muciniphila was colonized in the sparse mucus layer, and it therefore was closer to the intestinal epithelial cells than other microorganisms colonized in the intestinal lumen. Its metabolites, such as propionic acid, were also present in the mucus layer close to the intestinal epithelial cells and were easily accessible to the host. Propionic acid can act on the host through Gpr43 (G protein-coupled receptor 43), while other short-chain fatty acids through Gpr41, thus causing a series of downstream pathway changes to achieve immunomodulatory effects (Le Poul et al., 2003;Maslowski et al., 2009).
In vivo, A. muciniphila was colonized in sterile mice and the effective colonization was highest in the caecum . This may be explained by the reason that most of the mucin was produced in the caecum. The whole transcriptome analysis of intestinal tissue samples indicated that A. muciniphila regulated the expression of approximately 750 genes, with the changes mainly focused on genes associated with immune responses. In vitro, propionic acid and butyric acid are the main metabolites of A. muciniphila. A. muciniphila regulated the expression of 1005 genes in intestinal tissue, of which 503 genes were upregulated and 502 genes were down-regulated. While Faecalibacterium prausnitzii only affected the expression of 190 genes, of which 86 were upregulated, and 104 genes were downregulated (Lukovac et al., 2014). Consequently, A. muciniphila can regulate the host's metabolism and immune function. However, the causal relationship between the microbes and host genomes is very complicated and needs to be further evaluated (Wang et al., 2018a,b).

Akkermansia muciniphila regulated the balance between health and disease
Akkermansia muciniphila has recently been considered as a significant factor in human physiology, including homeostatic and pathological conditions. A large number of human and animal studies have addressed the associations between the abundance of A. muciniphila and various disorders and diseases (Tables 1 and 2). The decreased level of A. muciniphila is considered to be related to the development of some diseases. Amongst which, the majority were metabolic disorders and inflammatory diseases, including obesity, type 2 diabetes, inflammatory bowel disease (IBD), autism and atopy. However, Weir et al. (2013) found that the level of A. muciniphila was obviously elevated in patients with colorectal cancer compared with that in healthy individuals. This negative correlation might be associated with some confounders such as diet and medication. For example, food intake was greatly reduced in patients with colorectal cancer, while fasting is reported to be involved in increasing the level of A. muciniphila (Remely et al., 2015a,b). A small sample size of patients might be another influencing factor. Moreover, some studies showed that no relation with A. muciniphila-like bacteria was observed by metagenomic analysis (Zeller et al., 2014;Yu et al., 2017).
Recently, the research models of microbiome are facing a shift from focusing on association with a causality in recent years. For example, the beneficial therapeutic effects can be observed when the bacteria were administered in a viable form (Table 3). Consequently, A. muciniphila may become a biomarker of host health status, indicating the state of disease progression (Png et al., 2010;Swidsinski et al., 2011;Berry and Reinisch, 2013).
Unexpectedly, a recent study showed that pasteurized A. muciniphila can also prevent obesity and related complications, with the effectiveness be even better than live bacteria (Plovier et al., 2017). Even more exciting, the research team purified the outer membrane protein of A. muciniphila, Amuc_1100, which may exert this beneficial effect. Amuc_1100 was stable during pasteurization and interacted with Toll-like receptor 2 to improve intestinal barrier function and to perform part of the probiotic function alone. Consistent with this finding, Ottman et al. (2017a,b) also found that Amuc_1100 could activate TLR2 and TLR4 to increase IL-10 production and thus regulating immune response and intestinal barrier function This finding is significant and provides an important theoretical basis for the application of A. muciniphila in clinical treatments. However, the proved activity of A. muciniphila in pasteurized form has caused another controversial problem. The use of the term probiotic, which was specifically defined as live microorganisms by the Expert Panel from the Food and Agriculture Organization of the United Nations in 2001, may be misleading. A recent review stated that probiotic applications can be either live or dead forms (Hai, 2015). Regarding this modified definition, the Expert Panel previously declared that a dead probiotic is not approved. They Male 6-8 week C57BL/6 mice Interventional 1 ND: n = 5-7 2 HFD: n = 5-7 3 ND with AmEVs: n = 5-7 4 HFD with AmEVs: n = 5-7

Metabolic disorders and A. muciniphila
Akkermansia muciniphila is abundant in the gut microbiota of healthy individuals and exerts the effect of preventing and treating obesity, type 2 diabetes and other metabolic dysfunctions (Png et al., 2010, Santacruz et al., 2010Karlsson et al., 2012;Everard et al., 2013;Zhang et al., 2013). Previous studies found that its abundance was inversely proportional to the body weight of mice and humans (Derrien et al., 2010;Santacruz et al., 2010;Karlsson et al., 2012;Everard et al., 2013;Teixeira et al., 2013). Akkermansia muciniphila can significantly increase glucose tolerance and attenuate adipose inflammation in obese mice by inducing Foxp3 regulatory T cells (Shin et al., 2014). With the application of probiotics to overweight subjects after fasting, an obviously increased level of A. muciniphila was observed (Remely et al., 2015a,b). Moreover, an interventional study with Akkermansia showed that the level of blood lipopolysaccharide, which functioned as an indicator of gut permeability, was significantly decreased in obese mice after the administration of Akkermansia (Everard et al., 2013). Similarly, another study established that Akkermansiaderived extracellular vesicles could regulate the intestinal permeability and barrier integrity and thus affect the metabolic functions in mice with a high-fat diet (Chelakkot et al., 2018). Dao et al. (2016) reported that the baseline level of A. muciniphila in obese patients was negatively related to the fasting blood glucose, waist-tohip ratio and subcutaneous fat cell diameter. And after limiting energy intake for 6 weeks, patients with a high abundance of A. muciniphila at baseline had significantly improved insulin sensitivity and other obesity-related clinical indicators. Akkermansia muciniphila can be therefore used as a metabolic marker to indicate the reduction in the risk of obesity (Brahe et al., 2015), and it might be directly used to improve the glucose and lipid metabolism to treat obesity. Recently, Chelakkot et al. (2018) reported that compared to patients with type 2 diabetes, healthy human contained more A. muciniphila extracellular vesicles (AmEVs) in faeces. Another study found that the abundance of A. muciniphila was reduced in subjects with pre-diabetes and type 2 diabetes compared to subjects with normal glucose tolerance (Zhang et al., 2013). The relationship between A. muciniphila and type 2 diabetes was also reflected in cases using metformin (Lee and Ko, 2014). High levels of A. muciniphila in patients seemed to contribute to enhancing the efficacy of metformin (Shin et al., 2014). This was confirmed by the correlation between an increased A. muciniphila level and the effectiveness of metformin in a recent study (Forslund et al., 2015). Although the mechanisms involved are not fully understood (van Passel et al., 2011;Swidsinski et al., 2011;Everard et al., 2013;Cani and Everard, 2014;Shin et al., 2014), these animal experiments and related human studies have provided strong support for A. muciniphila in regulating energy homeostasis and glucose metabolism.
Several animal experiments and one human study have used A. muciniphila for direct intervention to evaluate its effectiveness in treating metabolic diseases. Initially in 2013 (Everard et al., 2013), Everard et al. reported that the abundance of A. muciniphila was decreased in mice with diabetes and obesity caused by high-fat diet, and the metabolic function of mice could be improved by intragastric administration of viable A. muciniphila. In 2017, Hanninen et al. (2017) established that transplanting the gut microbiota of mice with low incidence of diabetes, into the mice with high incidence of diabetes, did not reduce the morbidity of diabetes, but transplanting the single strain A. muciniphila into the mice with high incidence of diabetes can reduce the morbidity of diabetes. Chelakkot et al. (2018) reported that the intervention of oral administration with AmEVs may improve metabolic function by altering intestinal permeability and barrier integrity in high-fat diet mice. Thus, based on these direct interventional studies, A. muciniphila could be a very promising beneficial microbe for treating metabolic disorders. Most importantly, Plovier et al. (2017) have implemented a clinical study to evaluate the efficacy of A. muciniphila on metabolic syndrome. Currently, complete results have not been published, but the preliminary human data at least suggested that oral administration of this bacterium is safe. Altogether, these results demonstrate that A. muciniphila promises to be a potential therapy to treat metabolic diseases.

Immune diseases and A. muciniphila
A decreased abundance of A. muciniphila in children with atopic diseases indicated that it plays an important role in IgE-related atopic diseases (Drell et al., 2015). The correlation between a low level of A. muciniphila and immune response in atopic children suggested that A. muciniphila could interact with intestinal epithelial cells to produce IL-8 for immunomodulatory effects (Drell et al., 2015;Reunanen et al., 2015). In addition, the reduction in the number of A. muciniphila was closely related to the occurrence of IBD (Png et al., 2010;Rajilic-Stojanovic et al., 2013). The abundance of A. muciniphila was significantly decreased in the intestinal mucosa of IBD patients compared to that in healthy people (Png et al., 2010). Kang et al. (2013) recently found that AmEVs could regulate intestinal immunity and homeostasis and exert protective effects in the development of dextran sulfate sodium-induced colitis in mice. However, there is still a lack of human experiments that directly interfere with A. muciniphila to illustrate the causal relationship between this microbe and host immune diseases.

Cancer therapy and A. muciniphila
Recently, three consecutive articles published in 2018 have shown the importance of gut microbiota combined with anti-PD-1 antibody in cancer therapy (Gopalakrishnan et al., 2018;Matson et al., 2018;Routy et al., 2018). Routy et al. (2018) analysed the relationship between the therapeutic efficacy of immune checkpoint inhibitors and the gut microbiota in patients with different cancers. They found that the intestinal level of A. muciniphila was significantly increased in patients with a positive response to the immune checkpoint inhibitor PD-1 antibody. Furthermore, when the faecal microbiota from patients who responded positively to the immunotherapy were transplanted to sterile mouse, the corresponding positive response to the anti-PD-1 antibody was achieved. But when the faecal bacteria from patients who did not respond to the immunotherapy were transplanted to sterile mice, the native response was observed. Excitingly, the mice could recover their response to the anti-PD-1 antibody after oral administration of A. muciniphila. In addition, Matson et al. (2018) reported A. muciniphila abundance was observed in four metastatic melanoma patients with clinical response to anti-PD-1-based immunotherapy. After gavaged with faecal material from responding patient donors, improved tumour control and better efficacy of immunotherapy was observed in a mouse melanoma model. Gopalakrishnan et al. (2018) also found a higher level of good intestinal bacteria in the melanoma patients who responded to the treatment of PD-1 blockade. Combining three studies, Gharaibeh et al. (Gharaibeh and Jobin, 2018) concluded that there was a signal for more A. muciniphila in responders. The above results indicate that cancer immunotherapy combined with A. muciniphila as one of important probiotics in selective microbiota transplantation (Wu et al., 2019) is expected to achieve better clinical results for patients in the near future.
Consistently, Wang et al. (2018a,b) reported one patient with high-grade metastatic urothelial carcinoma showed immune checkpoint inhibitors (ICI)-associated colitis after a trial of combined CTLA-4 and PD-1 blockade. After ICI-associated colitis in the patient was successfully treated with FMT, donor-derived bacteria were observed to be effectively colonized in the patient's intestinal tract, with an obviously higher level of A. muciniphila. Consequently, A. muciniphila has shown its potential role in the treatment of cancer, and this role needs to be further confirmed by researchers.

Conclusions
Akkermansia muciniphila, as a potential probiotic that can make good use of gastrointestinal mucin, is inextricably linked to host metabolism and immune response. It promises to be a therapeutic target in the microbiotarelated diseases, such as colitis, metabolic syndrome, immune diseases and cancer. Preliminary human data suggest oral administration of A. muciniphila is safe, but its effect needs to be further verified in more human clinical trials in the near future.