Akkermansia beyond muciniphila - emergence of new species Akkermansia massiliensis sp. nov.

BACTERIAL SPECIES OF THE GENUS AKKERMANSIA

The genus Akkermansia belonging to the phylum Verrucomicrobiota of Gram-negative bacteria was first described by scientists at the University of Wageningen, Netherlands, after isolation and characterization of a mucin-degrading strain from a healthy adult, which they named Akkermansia muciniphila and is represented by the type strain MucT (= ATCC BAA-835)^[1]. Since then, two more validly published species under The International Code of Nomenclature of Prokaryotes (ICNP) of the genus Akkermansia have been identified: A. glycaniphila isolated from reticulated python feces (PytT)^[2] and A. biwaensis (type strain WON2089T)^[3].

In addition, several new Akkermansia species have been proposed, such as Candidatus A. intestinavium isolated from the gut of birds^[4], Candidatus A. intestinigallinarum isolated from the gut of hens^[4], and the human gut isolates A. massiliensis sp. nov. with strain Marseille-P6666T (= CSUR P6666 = CECT 30548) as the type strain^[5] and Candidatus A. timonensis with Akk0196 as the type strain^[5]. Figure 1 shows the genomic phylogenetic tree and corresponding proteome of the current and proposed Akkermansia species generated using the BV-BRC server^[6]. Interestingly, the recently named A. biwaensis and the proposed Candidatus A. timonensis type strains are 99.5% similar based on genome-wide average nucleotide identity (gANI); hence, they represent the same species. A. biwaensis is on the notification list for taxonomic validation^[7]. A recent phylogenomic analysis of 367 high-quality Akkermansia isolates and metagenome-assembled genomes suggested the existence of at least 25 Akkermansia species at the genomic level^[8].

Figure 1. Genomic phylogenetic tree and corresponding proteome^[6]. Type strains of validly published species are denoted in red.

Based on extensive metagenomic assembly, Akkermansia strains in the human gut were grouped into five candidate species including A. muciniphila based upon species-level genome bins (SGBs), with significant genomic diversity despite similar 16S rRNA gene sequences^[9]. Karcher et al. highlighted that this surprisingly high similarity based on 16S rRNA (at least 98%) is likely the reason for the wrong taxonomic assignment of several cultivated strains to the A. muciniphila species and leading to an underestimated diversity of the genus Akkermansia and less researched and characterized Akkermansia species beyond A. muciniphila^[9]. The publication also highlighted that the corrin ring biosynthesis operon genes are consistently present only in two candidate species, namely SGB9227 and SGB9228, but not A. muciniphila. A comparison of the public genome of cultured isolates from NCBI within SGB9228 to the proposed type strain for A. massiliensis sp. nov. Marseille-P6666T shows that SGB9228 corresponds to the proposed A. massiliensis species (> 95% gANI similarity), and SGB9224 similarly corresponds to A. biwaensis.

AKKERMANSIA MASSILIENSIS SP. NOV. - A RE-EXAMINATION OF ITS PRESENCE IN THE GUT MICROBIOME

In the survey published by Karcher et al.^[9], they reported the prevalence of A. muciniphila in gut metagenomes from adults (n = 7,995) as 34%, A. massiliensis sp. nov. (SGB9228) as 8%, and A. biwaensis (SGB9224) as 2%. While it is true that the high similarity of the 16S rRNA genes among these species and the lack of reference sequences in taxonomic databases contributed to their historical under-classification in microbiome datasets, these species can be distinguished using newer high-resolution clustering algorithms such as DADA2 that defines amplicon sequence variants (ASVs) based on single base pair differences^[10]. If utilizing the commonly sequenced V4 region of the 16S rRNA gene (515f/806r primer pair), there are four base pair differences between each A. muciniphila (GenBank Accession: AY271254) and A. massiliensis sp. nov. (GenBank Accession: ON014381) and likewise between A. muciniphila and A. biwaensis (GenBank Accession: LC711105), which provides sufficient resolution to distinguish ASVs from the species within Akkermansia. Thus, we examined the prevalence and abundance of these species in three studies where fecal 16S rRNA V4 amplicon data were collected: (1) The American Gut Project^[11] (n = 5,000), which involved a crowd-sourced sampling of adults; (2) IsoMic clinical study (ClinicalTrials.gov identifier NCT05150184), which was comprised of healthy adults [Body mass index (BMI) 18.5-25 kg/m²] from 5 ethnicities residing in the United States; and (3) Next Generation Probiotics for Metabolic Health (NGPs for MH) clinical study (ClinicalTrials.gov identifier NCT04229082), which included lean (BMI 18-25 kg/m²) and obese (BMI 27.5-35 kg/m²) women residing in Finland. Similar to Karcher et al.^[9], we found A. muciniphila to be both more prevalent and abundant in all of the cohorts compared to A. massiliensis sp. nov. [Table 1]. The prevalence of A. massiliensis sp. nov. was higher in the American Gut Project, IsoMic, and NGPs for MH lean cohorts (14.7%, 11.5%, and 26.3%, respectively), compared to the average abundance (8%) reported by Karcher et al.^[9]. The exception was the NGPs for MH obese cohort, where, intriguingly, the colonization of A. massiliensis sp. nov. was not detected in any of the individuals. Interestingly, only a small percentage of individuals were co-colonized with both A. municiphila and A. massiliensis sp. nov., suggesting that these two species may be mutually exclusive. This agrees with Karcher et al.^[9], who also showed a strong negative correlation between their co-occurrence in 4,171 human metagenomes. A. biwaensis was undoubtedly the most rarely detected human-associated Akkermansia species, but we found its prevalence was 8.9% in the American Gut Project cohort, with an average abundance of 0.2%. This prevalence is perhaps higher than expected for this almost completely unexplored species in human microbiomes. These findings support the exploration of potential metabolic health benefits associated with the colonization of A. massiliensis sp. nov. and warrant future mechanistic studies to better understand the interplay among A. muciniphila and the less abundant Akkermansia species and their impact on the host.

Two studies have looked in more detail at associations between colonization with different Akkermansia species and respective health outcomes. Karcher et al. reported that in a sample set of 3,311 genomes coming from different metagenomic datasets and isolates, only A. muciniphila was significantly negatively associated with BMI^[9]. In addition, a more recent analysis by Mueller et al., who re-analyzed metagenomic datasets of patients with obesity, inflammatory bowel disease, and response to cancer immunotherapy, reported that there were no significant correlations between abundances of A. muciniphila and A. massiliensis and obesity in children^[12]. Only A. muciniphila was, however, found to be significantly reduced in samples from patients with ulcerative colitis compared to healthy controls. In addition, patients with non-small-cell lung cancer undergoing immune checkpoint inhibitor treatment colonized with A. muciniphila clade AmIa had an increase in survival compared to patients with no detectable A. muciniphila. Mueller et al. reported the prevalence of the different Akkermansia species similar to those in Table 1, with A. muciniphila being the most prevalent species [32.5% of 1,088 samples, followed by A. massiliensis (11.9%) and A. biwaensis (5.2%)]^[12]. This lower prevalence of Akkermansia species other than A. muciniphila means that much larger datasets are required to thoroughly investigate correlations between human health outcomes and the relative abundance of these Akkermansia species.

At present, there is little known about the possible health effects of Akkermansia species other than A. muciniphila. However, Kumar et al. have first described the identification and characterization of a novel species of the genus Akkermansia (Akkermansia strain DSM 33459) with beneficial effects in relation to metabolic health effects demonstrated in a diet-induced obesity (DIO) mouse model^[13]. The genome of this Akkermansia strain DSM 33459 showed 99.85% gANI similarity to the genome of the type strain A. massiliensis sp. nov. Marseille-P6666, indicating that strain DSM 33459 also belongs to the proposed species A. massiliensis sp. nov. and not to A. muciniphila species based on only 87.5% gANI identity. In the DIO model, Akkermansia DSM 33459 administration showed significant improvements in body weight, total fat weight, insulin and resistin levels after administration for 12 weeks^[13]. Mice consuming the high-fat diet also harbored a significantly greater abundance of native A. muciniphila in their intestinal microbiota compared to mice on normal chow. Akkermansia DSM 33459 was able to efficiently engraft and replace the native A. muciniphila in this model. This further highlights the possible competition of the two Akkermansia species, which may limit their co-existence in the gut and questions what mechanisms may be at play, such as potential differences in their growth rates, their ability to adhere to intestinal epithelial cells, and metabolic pathways related to sulfur reduction^[14]. Interestingly, in this mouse DIO study, pasteurized A. massiliensis sp. nov. did not show the same efficacy as compared to the live bacterium. This observation is different compared to what is known for A. muciniphila. Further studies are needed to explore the impact of pasteurization methods on the bioactivity of pasteurized A. massiliensis sp. nov.

In addition, van der Lelie et al. described a consortium of 11 strains (GUT-108), which was designed to address gut microbiome dysbiosis in inflammatory bowel disease, and showed that this consortium reversed experimental colitis in a mouse model^[15]. One of the 11 strains was named Akkermansia sp. GGCC_0220 and was suggested to belong to a new Akkermansia species based on gANI. Our analysis of the public whole genome deposit record (GenBank Accession: JABFCI000000000.1) for this bacterium indicates that Akkermansia GGCC_0220 is 98.7% similar to both A. massiliensis sp. nov. Marseille-P6666 and A. massiliensis sp. nov. DSM 33459, i.e., it belongs to the same species.

AKKERMANSIA MASSILIENSIS SP. NOV. - CHARACTERISTICS WITH RELEVANCE FOR UNIQUE MECHANISM OF ACTION

The human gut resident A. muciniphila has gained attention in recent years due to its potential role in promoting health. While research on Akkermansia is still progressing, several studies have suggested beneficial effects on health, including positive impacts on cardiometabolic health, inflammation, gut barrier integrity, outcomes to checkpoint blockade response in cancer immunotherapy, homeostatic immunity, and maintenance of microbial balance in the gut^[16,17]. Research on Akkermansia is moving beyond association studies, with research increasingly focusing on the elucidation of the mechanisms underlying efficacy. Several possible mechanisms have been published to demonstrate the dynamics of this commensal in the gut. Plovier et al. have shown very elegantly the role of an outer membrane protein Amuc_1100 in metabolic and gut health^[18]. Yoon et al. have identified another factor, P9 protein encoded by Amuc_1631, which interacts with ICAM2 and stimulates GLP1 secretion^[19]. Recently, it has been shown that mediated production of a tripeptide [Arg-Lys-His (RKH)] can negate inflammation via TLR4 in a sepsis model^[20]. Bae et al. have identified an active lipid molecule from the A. muciniphila membrane and the immunogenic activity of diacyl phosphatidylethanolamine with two branched chains (a15:0-i15:0 PE) has been shown to be via an unexpected toll-like receptor TLR2-TLR1 heterodimer^[21]. Interestingly, Kumar et al. have shown variation in a15:0-i15:0 PE between A. muciniphila and A. massiliensis sp. nov., where A. massiliensis sp. nov. shows a higher abundance of a15:0-i15:0 PE as compared to A. muciniphila^[13]. It will be worth investigating the impact of differential abundance of a15:0-i15:0 PE on the immunogenic properties of different Akkermansia strains.

Bacterial-derived extracellular vesicles have been investigated for their impact on physiological functions. Chelakkot et al. investigated the role of A. muciniphila-derived extracellular vesicles (AmEV) and showed that AmEV administration enhanced tight junction function, reduced body weight gain, and improved glucose tolerance in high-fat diet-induced diabetic mice^[22]. What is unknown at present is whether AmEV serve as carriers for any of the possible effector molecules such as a15:0-i15:0 PE, Amuc_1100, and P9 or they act via another mechanism to provide the described benefits.

Advances in sequencing and bioinformatic tools have resulted in the identification of new Akkermansia phylotypes or species and highlighted the genomic and phenotypic diversity in the genus Akkermansia. The actual isolation of such novel Akkermansia strains enables further exploration in the field with the objective to understand if this variation is associated with unique properties and health benefits of the different Akkermansia species and phylotypes. This diversity also provides the opportunity to identify new mechanisms of action to understand differences in interaction between Akkermansia species and the host, raising interesting questions about the relative expression of effector molecules between different species and how this may impact efficacy. Recently, Kumar et al. have shown the health benefits of A. massiliensis sp. nov. in a DIO mouse model and have highlighted several new possible mechanisms of action^[13]. They have shown that A. massiliensis sp. nov. can degrade extracellular adenosine triphosphate (eATP). It has been shown that eATP is a key inducer of inflammation in the gastrointestinal tract and is strongly associated with inflammatory bowel disease^[23]. Extracellular ATP limits T follicular helper cells in the Peyer’s patches and thereby limits SIgA generation; thus, the degradation of eATP mediated by Akkermansia sp. in the lumen may play a vital role in modulating the immune structure in the gut^[24]. Kumar et al. also described microbial agmatine production by the A. massiliensis sp. nov. strain, and agmatine has been shown to have metabolic and neuroprotective effects^[13,25]. The in vitro production of agmatine was higher with A. massiliensis sp. nov. compared to A. muciniphila based on relative area concentrations. Pryor et al. have demonstrated that microbial agmatine is a major contributor to the agmatine pool in the gastrointestinal tract^[25]. If it can be shown that Akkermansia contributes to this pool, then exploring the role of Akkermansia in relation to the gut-brain axis presents an interesting avenue to explore. Further in vivo studies are needed to validate the physiological significance of eATP degradation and agmatine production by Akkermansia sp.

Karcher et al. showed that the absence of the corrin ring biosynthesis operon genes in A. muciniphila resulted in an inability to produce propionate in the absence of vitamin B₁₂^[9]. As depicted in Figure 2, we similarly measured the Vitamin B₁₂-dependent production of propionate in strains Akkermansia massiliensis sp. nov. DSM33459, Type Akkermansia massiliensis CECT 30548 and Type Akkermansia muciniphila ATCC BAA-835. Cells were grown in the defined media as published by Karcher et al.^[9]. Supernatants were collected at 48 h of growth and propionate was measured using HPLC. Strains A. massiliensis sp. nov. DSM33459 and Type A. massiliensis sp. nov. CECT 30548 showed similar production of propionate irrespective of the addition of B₁₂ in the media. Type A. muciniphila ATCC BAA-835 strain showed propionate production only in the presence of externally added B₁₂, suggesting its dependency on the external B₁₂ for propionate production as shown by Kirmiz et al. and Karcher et al.^[9,26]. Propionate production is dependent on the B₁₂-dependent methyl-malonyl CoA synthase reaction^[27]. The presence of the corrin ring biosynthesis operon and, hence, the ability to synthesize vitamin B₁₂ in A. massiliensis sp. nov. broadens its potential for human health. Recently, it has been shown that gut resident vitamin B₁₂-producing bacteria can modulate excitatory cholinergic signaling and behavior in the host Caenorhabditis elegans^[28]. Additionally, it is currently unknown, but possible that Akkermansia sp. may play a role in shaping the microbiome in the human gastrointestinal tract by providing the enzyme co-factor vitamin B₁₂ to other bacterial species.

Figure 2. Functional characterization of propionate synthesis pathway of Akkermansia species using representative isolates: Akkermansia massiliensis sp. nov strain DSM 33459, A. massiliensis strain Marseille-P6666T (CECT 30548), and Akkermansia muciniphila strain BAA-835T (ATCC BAA-835).

While these newly proposed mechanisms require further investigation and confirmation in in vivo studies, they do support further investigation of A. massiliensis sp. nov. in first-in-human intervention studies to study possible health benefits. The human gut commensal A. muciniphila remains the most studied beneficial bacterial species within the genus Akkermansia; however, the emergence of new species such as A. massiliensis sp. nov. presents new opportunities for the development of effective and targeted Akkermansia species-based interventions to provide health benefits.