The role of the gut non-bacterial microbiome (virome, mycobiome, archaeome) and its impact on obesity

Abstract

The epidemic of obesity and metabolic syndrome is a major public health concern internationally. There is increasing knowledge and research in areas of appetite regulation and drivers of obesity but there is still a gap on how the interactomes are altered in a metabolically dysregulated human body. The human microbiome has been implicated in the pathogenesis of obesity. While the association of gut bacteriome dysbiosis is well described in obesity and metabolic syndrome, there is a lack of an integrative understanding about the roles of the non-bacterial microbiome (virome, mycobiome, and archaeome) in the pathogenesis and protection of obesity and metabolic syndrome. Accumulating studies have revealed that the non-bacterial microbes in the gut, including viruses/phages, fungi, and archaea, are profoundly altered in obesity, and impact host adiposity and physiology in nuanced manners. In this review, we aim to provide a comprehensive view on the role and the mechanisms of the gut virome, mycobiome, and archaeome in obesity. These insights will shed light on the translational value as well as the future research directions for harnessing the gut non-bacterial microbial entities in the therapeutics and prevention of metabolic diseases.

1. Introduction

Obesity and metabolic syndrome are a growing public health problem worldwide [1], with a prevalence of 16% of adults worldwide in 2022 [2]. The understanding of obesity as a disease entity has changed over the years, but the key tenets remained a net positive energy balance sustained over time [3]. This perturbation in energy homeostasis involved a complex interplay of intrinsic host biological and behavioural factors together with environmental factors [3]. Clinically, the management of obesity and strongly linked metabolic disorders like type-2 diabetes mellitus (T2DM) involves a combination of behavioural therapy [4], pharmacotherapy [5], and bariatric surgery [6]. These mechanisms induce neurohormonal and/or anatomical alterations, among others, to restrict caloric intake and achieve negative energy balance and weight loss.

The human microbiome comprises all the micro-organisms living symbiotically within and on an individual [7]. The gastrointestinal tract, which houses numerous microorganisms including bacteria, eukaryotic and prokaryotic viruses, fungi and archaea, contributes significantly to the diverse human metagenome [8]. The advent of 16S ribosomal RNA sequencing techniques has enabled researchers to characterise the human microbiome, namely the bacteriome, avoiding tedious culturing techniques [9]. Over the years, numerous studies have attempted to elucidate the alterations of the bacteriome with obesity [[10], [11], [12], [13], [14], [15]] and T2DM [16,17].

The pioneering studies that investigated the role of the gut bacteriome on obesity was through germ-free murine models. It was shown that fecal transplantation of conventional mice to germ-free mice resulted in an increased body fat content despite reduced energy intake [18]. The causality of the microbiome in obesity was then shown through transplantation of a human endotoxin-producing Enterobacter isolated from an obese human into germ-free mice [19]. Such initial findings have also been shown in human studies, where fecal microbiota transplantation (FMT) from lean donors improved insulin sensitivity in individuals with obesity [20]. However, despite the initial enthusiasm, significant weight loss has not been consistently observed, limiting the translation value on FMT as a weight loss strategy. This may in part be due to the complex nature of the disease biology, which is itself a chronic relapsing condition, and the gut microbiome is a complex ecosystem with redundancy of function, and a complex network of interactions amongst various microbes, and with the host. Thus, rather than a single microbe or group of microbes, it may be an entire ecosystem, or ecotype, that influences the host phenotype. The initial small human clinical trials of FMT and improvement in insulin sensitivity [20,21], implied that elements within the transplanted fecal matter, as an entire “lean” ecotype, whether predominantly due to bacteria, fungi, or virus-like particles, did influence host energy metabolism.

The advancement of sequencing techniques has allowed for the rapid identification of viruses in the human metagenome, known as the virome [22], which was previously challenging to isolate and culture due to its need for an obligate host. The quantity of viruses in the human gut likely outnumbers the number of bacteria by up to 10 times [23,24], with almost 90% being prokaryotic viruses or bacteriophages [25,26]. Numerically, this makes viruses the most successful biological entities within humans. This constellation of the gut virome of bacteriophages, the phageome, can directly alter the host bacteriome. Murine studies have demonstrated that fecal virome transplantation (FVT) can alter the host obese phenotype [27].

Shotgun sequencing techniques have also uncovered other inter-kingdom interactions, including the interplay of fungi and bacteria on gut dysbiosis and inflammation [28]. The gut mycobiome is also an emerging area of research, as it has similar potential for translational impact through fungal probiotics such as that of Saccharomyces spp. In addition, recent studies have also identified human-associated archaea in the gut as well as respiratory tract, spurring efforts to understand its role in heath and disease [29]. Thus, it is likely that there are inter-kingdom interactions that exist within the human microbiome, and its impact on obesity cannot be fully understood simply by studying the bacteriome. Thus, the aim of this review is to provide a scoping review of the non-bacterial human microbiome, namely the virome, mycobiome, and archaeome, and the inter-kingdom interactions and its associations with the obese phenotype.

2. The gut virome and obesity

In contrast to the bacteriome, the gut virome, which comprised the total population of virus-like particles (VLPs) in the human gastrointestinal tract, is relatively understudied [30]. The gut virome comprises of a large proportion of prokaryotic viruses (mostly bacteriophages), eukaryotic viruses, as well as plant viruses [24,30,32]. Most plant viruses are likely derived from dietary sources [31].

2.1. The gut virome diversity and obesity

Similar to bacteria diversity, the gut virome appeared less diverse in obesity compared to lean subjects [33,34,36,37] (Figure 1). In a post hoc analysis of gut viral bioinformatics data from patients before and after bariatric interventions, it was shown that the gut viral alpha diversity tended to increase following weight loss interventions including exercise and dietary interventions, as well as bariatric surgery [33]. In a meta-analysis of available stool metagenomic sequences from 862 individuals from 5 countries, virome alpha diversity and richness were significantly reduced in obese subjects compared to controls [34]. Extrapolating from findings of gut bacteria, it is possible that the changes in alpha-diversity seen in the gut virome may also be a result of weight loss [35]. The beta diversity distances between obese and lean subjects were also significantly increased, demonstrating distinct gut virome profiles in obese subjects compared to healthy controls [34]. However, significant geographic factors may also influence the gut virome as well. In a study comparing the gut virome in largely ethnic Han Chinese patients with obesity and type 2 diabetes mellitus from 2 separate geographic cohorts, Yang et al. demonstrated that the viral alpha diversity was only reduced in obese subjects in Hong Kong compared to lean controls but there were no significant differences in alpha diversity in the Kunming cohort [36], though it can be argued that the difference may potentially be related to the Hong Kong obese cohort having a much higher mean BMI of 33.4 kg/m² compared to the Kunming obese cohort at 29.8 kg/m² [36]. Future studies can help to clarify confounders such as diet and water quality, stress and other potential jurisdictional-based factors.

Figure 1.

A summary of the associations of the non-bacterial microbiome with obesity (Created by biorender.com).

2.2. Prokaryotic viruses and obesity

Much of the human gut virome were prokaryotic viruses, or bacteriophages [22,25]. Bacteriophages infect bacteria, through its lytic and lysogenic life cycles [38]. The majority were from the order Caudovirales, which are double-stranded DNA bacteriophages with the families Myoviridae, Podoviridae and Siphoviridae, as well as the proposed family CrAssphages, and the single-stranded DNA Microviridae [24,38,39]. The association between bacteriophages on the host phenotype is less clear, as almost 50% of the viruses could not be identified to the family or lower taxonomic levels [40].

A cross-sectional study on murine fecal samples have demonstrated that the total viral content of leptin deficient mice was significantly elevated compared to control mice [37]. The fecal viral RNA content was also correlated with body weight, fat mass, glucose and the Firmicutes phyla, while negatively correlated with Bacteroidetes phyla, though there was a lack of speciation of which viruses were enriched, which limited the utility of the findings on viral diversity and composition [37]. Quantification of the viral genetic material alone also did not provide information with regards to the viral diversity or richness in the murine gut. However, it provided some evidence that inter-kingdom interactions may exist between the gut virome and bacteriome in influencing the host obesity phenotype. Another criticism was whether isolated viral DNA and RNA sequences represented transitioning viral sequences or were part of the gut ecosystem. This was an important consideration in gut virome studies, as there may be spatial differences in virome composition amongst mucosal and luminal viruses [41]. High fat diet resulted in increased temperate phages in the mucosa more so than the gut lumen [41]. A high fat diet was also shown to be associated with a reduction in lysogenic cycle phages, particularly Siphoviridae family of the order Caudovirales, with a concomitant increase in lytic cycle phages such as Microviridae [42]. This was associated with a concomitant increase in the gut Firmicutes compared to Bacteroidetes phyla of the corresponding bacteriome [42], a controversial occurrence that was commonly reported in the common obese phenotype, though these findings were still under investigation [41].

Another observation was seen with relation to crAssphages and its Bacteroidetes host. CrAssphages are the most abundant bacteriophage of the human virome, comprising of up to 90% of total viral genetic material in certain patients [25]. However, up until the last 10 years, the majority of the crAssphage genes did not have any matching database sequences and its prey was not known [25]. Clustered regularly interspaced short palindromic repeats (CRISPR) have been utilised by bacteria as an immunity mechanism to defend against foreign DNA sequences like bacteriophages [25,43]. CRISPR spaces are short nucleotide sequences representing foreign DNA (ie. from a previous bacteriophage infection) that are interspersed between the CRISPR array, serving as a form of immunologic memory of a past bacteriophage infection [25,43]. Analysing the corresponding CRISPR spacers, Dutilh et al. proposed the bacteria phylum Bacteroidetes as the in-situ prey for crAssphages [25]. This was subsequently confirmed, where ΦcrAss001 was demonstrated to infect the commensal bacteria Bacteroides intestinalis [44]. Despite not being a lysogenic phage, ΦcrAss001 infections typically do not result in an obliteration of Bacteroides intestinalis levels, but rather a perpetual steady state of host and phage concentration though mechanisms such of reversible acquisition of phage resistance, delayed lysis and pseudolysogeny [45] (Figure 1). This complex relationship has been shown where a decrease in crAssphage abundance was not associated with an expected proliferation of Bacteroides intestinalis as expected if it was a pure predator-prey relationship [46]. Rather, it was significantly associated with an overabundance of Bacilli class of the Firmicutes phylum, which was observed in children with obesity and metabolic syndrome [46].

Studies have also attempted to demonstrate potential causality of the prokaryotic virome in influencing the obese phenotype through FVT studies [27]. Cecal viromes isolated from low fat diet mice were transplanted into high fat diet mice models and high fat diet mice models pre-treated with ampicillin to eliminate the gut bacteriome [27]. Transplantation of the lean mice virome to high fat diet mice resulted in a significant decrease in percentage weight gain at 4- and 6-weeks, when compared to non-transplanted high fat diet controls [27]. Oral glucose tolerance test of high fat diet mice treated with FVT from lean mice were also similar to that of lean diet controls and significantly lower than that of non-transplanted high fat diet controls, suggesting a normalisation in glucose homeostasis [27]. The converse was also seen, where transplanted viromes from high fat diet mice resulted in more weight gain than chow-derived viromes [47]. These observations suggest that viromes can influence obesity but remain a subject of investigation. These FVT findings may be mediated through an indirect phage-mediated response. It is possible that at any given timepoint, any gut bacterial enterotype [48] or entero-subtypes, have a corresponding viral enterotype in a steady state of transient passage, lysis and lysogeny. Perturbations in the viral enterotype results in an eventual modification of the gut bacterial composition, to resemble that of the host phenotype of the donor virome. This inter-kingdom modification of the gut microbiota has been shown through FVT experiments, where FVT alone from high fat diet mice was able to modify the ileal bacterial density of standard diet mice towards the high fat diet donors [49].

2.3. Eukaryotic viruses and obesity

The eukaryotic viruses make up a small proportion of the gut virome in health [32,50]. The most prevalent eukaryotic DNA virus is from the Anelloviridae family [39]. The role of Anelloviridae in human diseases is still unclear [39]. There were reports of coinfections seen with hepatitis B and human immunodeficiency virus infections as well as its DNA being detected in certain human cancers [51], but till date there were no associations reported with human obesity. Another ubiquitous eukaryotic virus is the Circoviridae family [39]. Circoviridae were more commonly sequenced in fecal samples of children with type 1 diabetes mellitus compared to healthy controls, which may suggest a potential immunomodulatory role in the disease pathogenesis [52]. The human adenovirus 36 (Ad-36), which was transmitted by the fecal-oral route [53], had been shown in twin studies to be associated with a significant, though modest, increase in BMI (24.5 vs 23.1 kg/m²) and total body fat mass (29.6 vs 27.5%) compared to the antibody-negative twin [54] (Figure 1). Individuals seropositive for Ad-36 were also more likely to be obese, even after adjusting for other environmental factors [54]. Proof-of-concept vaccines in murine models resulted in a 23% significant increase in epididymal fat pad weight in mice inoculated with live Ad-36 compared to inactivated Ad-36 controls 4-days after inoculation, with no change in mice body weight [54]. The proposed mechanisms of Ad-36 on infected individuals include an increased glucose uptake, reduced leptin production, increased food intake due to reduced noradrenaline [55] and increased chronic inflammation from increasing Macrophage Chemoattractant Protein-1 (MCP-1) [56]. However, as these studies were observational in nature, the causality of Ad-36 to the obesity epidemic remains controversial.

2.4. Future work on gut virome and obesity

Given the light of the above discussion, more in-depth analysis of the virome and in other cohorts across are needed, with more resolution into the different classes and families rather than just its diversity and composition. In addition, though virome diversity appeared to parallel bacteriome in obesity, it may be still premature to conclude the casuality of the correlation. Rather than isolated studies looking at the virome or bacteriome, an inter-kingdom approach should be taken. One exciting area involves the dynamic interactions between the phage and its bacteria prey, and its impact of the host obesity phenotype.

3. The gut mycobiome in health and obesity

Unlike the gut virome, the composition of fungi in the gastrointestinal tract is smaller, comprising 0.1% of the total gut microbiome, though they still play a pivotal role in intestinal homeostasis and disease [57]. The healthy adult mycobiome is dominated by the genus Candida, Saccharomyces and Cladosporium [57,58] namely of the Ascomycota phyla [8]. They are present in the gastrointestinal tract, even in the acidic gastric environment [8,59] and the pancreas, which was supposed to be sterile [60]. The gut mycobiome is significantly influenced by dietary and environmental factors [57]. In murine studies, fungal communities were significantly different between mice fed standard chow diet compared to high fat diet [61]. In human studies, individuals on a carbohydrate rich diet have an increase in Candida spp [62]. (Figure 1), while urban dwellers have a higher proportion of Saccharomyces cerevisiae [62]. This may in part be contributed by the high levels of foods containing yeast such as Saccharomyces often found in bread and beer [63]. Candida spp. also appeared to be predominant with older-age individuals [64].

3.1. The gut mycobiome diversity

In contrast to the gut virome and bacteriome, the alpha diversity of the gut mycobiome appeared to be similar in lean and obese states [61,63,65], in both murine [61,63] and human studies [65]. However, the composition of the mycobiome were significantly different [61]. Some of the reasons may relate to the relatively low abundance of the gut fungal communities in contrast to the bacteria or viral communities, which may account for the insignificant diversity changes seen.

3.2. Yeasts in obesity

Yeasts are unicellular eukaryotes within the fungi community, with Candida spp. being the most prevalent commensal and pathogenic fungi of the human host [66]. The impact of Candida spp. on obesity remains unclear. In murine models, perturbations of fungal communities with low dose antifungals that depleted Candida and Saccharomyces populations resulted in more pronounced body fat increases under high fat diet conditions [67]. In a separate case–control study comparing relative abundances of fungal microbes in obese children with normal weight controls matched by gender, obese children had a significantly less diverse yeast population, with lower abundances of Candida and Saccharomyces spp. [68]. However, in a cross-sectional study comparing normal weight, overweight and obese adult through fungal plate culture, there was a higher prevalence of Candida albicans and Candida kefyr, which also inversely correlated with serum high-density lipoprotein levels [69]. In another separate study, there was a higher prevalence of Candida parapsilosis in obese subjects compared to healthy human controls [70]. Most likely, the change in Candida spp. in relation to obesity, may reflect an increased dietary consumption of simple carbohydrates, which was likely more prevalent in patients with obesity [71]. While the above studies are seemingly at odds with one another, the predominance of Candida spp. likely both promotes and reduces the impact of metabolic syndrome via different mechanisms, outcomes of which are dependent on the predominating pathway. For example, Candida spp. help in the breakdown of complex polysaccharides in the gut, which may then be fermented on by short-chain fatty acid (SCFA) producing bacteria, such as Ruminococcus spp. [68], which may help in reducing insulin resistance [72]. Formyl-methionine, another metabolite of Candida spp. metabolism, in contrast, contribute to the development of atherosclerosis [73]. Candida parapsilosis has been shown in murine models to produce extracellular lipases that can digest dietary lipids, which likely contributed to increased gut absorption of fatty acids, leading to increased energy uptake and obesity [70]. Overall, predominance of Candida spp. with increased carbohydrate intake in the gut, likely influences the host obese phenotype through indirect mechanisms like modulating the gut bacteriome [68], thereby affecting energy uptake [70] and promoting atherosclerosis [73,74] as well as chronic inflammation [75,76].

Unlike Candida spp, the role of Saccharomyces spp. on the obese phenotype is less evident [71,77]. Saccharomyces boulardii, which is used as probiotic, has been shown to reduce hepatic steatosis, inflammation [78] and fat mass [79]. It can also modulate the gut bacteriome through modulation of the 2 most dominant bacteria phyla (Firmicutes and Bacteroidetes) [[79], [80], [81]], as well as other benefits independent of host gut colonisation [82], including reduction of chronic inflammation [82] (Figure 1). Yet, Saccharomyces boulardii, is a distinct strain, first isolated from plants [83], that does not naturally colonize the human host [83]. It plays its role well as a probiotic as it can survive through gastric acid and alkaline bile salt environments, yet not colonize the gut in the long term, but only promoting these effects transiently [83]. This contrasts that of other Saccharomyces spp, which has been shown mixed results with regards to obesity and metabolic syndrome [61,71]. In murine models, alterations in Saccharomyces spp, in contrast to Saccharomyces boulardii, were associated with dysfunctional metabolism, including increasing fatty accumulation in the liver and triglyceride concentrations [84], though Saccharomyces cerevisiae were also reported to be more abundant in mice fed a chow diet compared to high fat diet [61]. In human studies, Saccharomyces spp. abundance was observed in obese children [68]. Higher abundance of fecal Candida and Saccharomyces spp. were also seen in children who eventually develop type-1 diabetes mellitus [85], though it could be argued that type-1 diabetes mellitus may be more an immunologic phenomenon and pathogenetically different from T2DM, where insulin resistance is the hallmark. The exact role of Saccharomyces spp. on the obese phenotype now is still unclear, and more research is required to better understand such correlations between Saccharomyces spp. and metabolic syndrome.

Schizosaccharomyces pombe was another yeast that may have a potential probiotic potential [86]. In human correlation studies, it was significantly correlated with lean controls compared to obese subjects [86]. When transplanted into mice fed a high-fat diet, it resulted in a slower weight gain, reduced hepatic fat accumulation, reduced fat mass and improved insulin sensitivity [86]. Other yeasts related to mycobiome dysbiosis that were correlated with obesity included Pichia spp., which was positively correlated with obesity [65,69,87] and Tilletopsis spp., which was negatively correlated with obesity respectively [61,87]. Future research into the potential role of certain yeasts in modulating the effects of metabolic syndrome and obesity, such as through probiotics, are much anticipated.

3.3. Moulds in obesity

In contrast to yeasts, the impact of moulds on obesity were less well described. Fungi can exist as unicellular yeasts, multicellular moulds, or as dimorphic fungi [88]. Moulds are multicellular with branching filamentous extensions known as hyphae [88]. In one of the first studies to describe the human mycobiome in obesity, Mar Rodríguez et al., reported that Mucor spp. had a negative correlation with obesity [65,89]. Mucor spp. were more abundant in non-obese controls, and its relative abundance was increased following weight loss in obese subjects [65] (Figure 1). In patients with newly diagnosed T2DM, opportunistic fungal pathogens such as Aspergillus [90] were more abundant compared to individuals with long-standing diabetes as well as individuals without diabetes [91], though unknown periods of relative hyperglycemia prior to the diagnosis of T2DM remains a confounder.

3.4. Future work on gut mycobiome

The gut mycobiome is often overlooked in contrast to the gut bacteriome. However, as discussed above, there is evidence that it may play a significant role in modulating, or even exert certain direct effects on host metabolism. Certain yeasts, in particular Saccharomyces boulardii and Schizosaccharomyces pombe may have the potential to be a probiotic that may be beneficial in modulating the host obese phenotype. Further bench research and future clinical studies with regards to the therapeutic role of fungal probiotics and its role in obesity and metabolic syndrome are keenly anticipated.

4. The gut archaeome and obesity

Archaea were previously classified as archaebacteria [92] until the proposal of the three domain system revolutionised by Carl Woese in 1990 [93]. They are single-celled prokaryotic organisms without a nucleus [94]. Yet, archaea have a different ribosomal ribonucleic acid (RNA) as well as different cell membrane lipids compared to bacteria [95]. Archaea are known to survive in extreme environments, though they are also found to be part of the human metagenome [96]. In contrast to the gut bacteriome, mycobiome or virome, the role of the gut archaeome in human health and disease is relatively unknown, and it comprises about 1% of the human gut microbiome [97].

Metanogenic archaea are the most prevalent gut archaea [97]. Methanogenesis results in the production of methane by gut archaea, commonly utilising end products of bacterial metabolism as substrates [97]. The most prevalent human gut archaea belong to the order Methanobacteriales, for which Methanobrevibacter smithii and Candidatus Methanobrevibacter intestini were the most dominant, accounting for more than 90% of the gut archaeome [97,98]. Previously, Candidatus Methanobrevibacter intestini was considered as the same species as Methanobreibacter smithii before it was re-classified as a separate species given its distinct genomic features compared to Methanobreibacter smithii [97]. To date, no archaea have been causally linked with any human disease, and are generally considered non-pathogenic [96,99].

Given its predominance in the human gut, most studies of the human archaeome focused on Methanobrevibacter smithii. However, existing literature show conflicting results on the relationship of M. smithii on host obesity. M. smithii has been implicated in gnotobiotic mice models to contribute to host adiposity through modification of the gut bacteriome via inter-kingdom interactions [100,101]. Gnotobiotic mice models demonstrated a symbiotic relationship through co-colonisation of M. smithii with Bacteroides thetaiotaomicron, a bacteria known to breakdown dietary polysaccharides [100]. In contrast to controls, B. thetaiotaomicron and M. smithii co-colonization increased the former’s ability to degrade polysaccharides [100]. This results in a metabolic shift towards acetate and formate production, which in turn helps to sustain B. thetaiotaomicron proliferation [100]. Formate is then utilised as a hydrogen-donor to produce methane, which in the process helps in M. smithii proliferation [100]. This in turn also impacts the host, as the metabolic shift increased the production of cecal acetate, driving increased intraluminal caloric extraction, increasing hepatic de novo lipogenesis and adiposity [100]. This was confirmed by measuring the epididymal fat pad of the mice models [100]. M. smithii co-colonized mice had a significantly higher fat pad weights, even with similar chow consumption [100]. The Prevotellaceae family of bacteria, which can produce hydrogen (H₂) through fermentation [102], was also postulated to exhibit a similar symbiotic relationship with M. smithii, through direct transfer of H₂ gas as a substrate for methanogenesis [102]. In a cross-sectional study comparing stool samples from 3 lean, 3 obese and 3 post Roux-en-y gastric bypass individuals, no methanogens were found in the stool samples of lean individuals [102]. In contrast, all the obese individuals had methanogens in their stool samples [102]. Of the 3 post Roux-en-y gastric bypass individuals, with a mean BMI in the overweight category (27.7 ± 4.1 kg/m²), methanogens were present in only 1 individual, and at a significantly lower abundance compared to the obese cohort [102]. The increased abundance of methanogens found in the study, were coupled with a concomitant enrichment of Prevotellaceae family of H₂-producers, again suggesting a possible link between dietary polysaccharide fermentation with cross-feeding of microbiota and M. smithii through acetate and H₂ production [102]. The resultant increase in energy harvest, through metabolites like acetate, may explain the correlation between methanogens and the obese phenotype [102]. This association was similarly demonstrated in a prospective pediatric cohort with childhood obesity as well, where M. smithii predominance and colonization were significantly correlated with BMI and higher weight scores respectively [103]. In the clinical setting, methanogens can be detected through volatile substances in the breath, thereby avoiding the inconvenience of stool sample collection [104]. This has been demonstrated in a cross-sectional study of obese patients, where methane-positivity in breath tests were significantly correlated with elevated BMI compared to methane-negative patients [104]. Similarly, in a large cohort study of 792 participants, presence of both methane and hydrogen on breath test was significantly associated with a higher BMI and body fat percentage [105].

Yet, as discussed above, M. smithii has also been shown to be enriched in lean, non-obese individuals. The family Christensenellaceae of the phylum Firmicutes, were generally found to be highly enriched in lean and metabolically healthy patients [106]. In germ-free mice models, co-colonization of M. smithii with Christensenellaceae resulted in an opposite effect to that seen with B. thetaiotaomicron as discussed above. These mice had the lowest weight increase compared to models with low Christensenellaceae and M. smithii populations [107]. Unlike elevations in acetate and formate seen with co-colonization of M. smithii with B. thetaiotaomicron, cecal levels of butyrate and propionate were significantly elevated in Christensenellaceae and M. smithii co-colonized mice, compared to methanogen-negative mice [107]. This suggested that different gut bacteriome populations, may interact with methanogenic archaea differently to exert mixed effects on the obese phenotype, probably via a mutualistic relationship, through differential effects on bacteria organic or SCFA production [108] (Figure 1). Armougom et al. reported non-significant differences in the prevalence of M. smithii populations between lean and obese subjects, though there was still a higher copy number of M. smithii in obese subjects compared to lean controls [109]. Interestingly, M. smithii copy number was reported to be highest in recently hospitalised patients for anorexia nervosa, though it was debatable whether nosocomial contact or medications may have altered the gut microbiome [109].

4.1. Future work on the gut archaeome

In contrast to other aspects of the gut microbiome, to date, the gut archaeome is still relatively understudied. Some of the challenges include its relatively low abundance compared to the gut bacteriome [97,110], biases with DNA extraction [111] and difficulty culturing [112]. However, literature suggest that the gut archaeome, especially the methanogens like M. smithii, may exert their effects via inter-kingdom interactions with saccharolytic bacteria in the gut, to influence host metabolism and adiposity. In addition, inter-kingdom work should also focus on archaeal viruses, like the temperate archaeal virus (MSTV1), that infects M. smithii, in a similar lysis and lysogenic cycle, and its impact of the corresponding gut bacteriome, and the host phenotype [113].

A key motivator of behavioural change often stems from the desire to alleviate pain or discomfort. Unfortunately, the absence of any clinically significant differences in reported intestinal symptoms between patients with obesity and patients without obesity [105] indicate that these biochemical changes in methane and hydrogen alone may not serve as a compelling incentive for weight loss. Thus, non-invasive methane breath testing may hold promise as a potential volatile surrogate for the identification of gut methanogens, more in prognosticating the efficacy of obesity treatment.

5. Microbial inter-kingdom interactions in obesity

As illustrated above, the gut microbiome is dynamic, and many of the metabolic effects were mediated through inter-kingdom interactions between the non-bacterial microbiome, the gut bacteriome, and the host. For the gut fungi and archaea, bacteriome-mycobiome and bacteriome–archaeome interactions commonly represent either a syntrophic relationship, through cross-feeding mutual metabolic byproducts [114] or a competitive/antagonistic one. This was seen with methanogenesis involving the M. smithii and gut polysaccharide fermentation [100,101]. Possible mixed relationships were also described between Candida and Prevotella [61,63,115]. Prevotella abundance was shown in a randomized controlled trial to be associated with 1.8 kg more weight loss in individuals on a high fibre diet [116]. Fermentation of complex fibres by Prevotella results in the production of SCFA [117] that inhibits Candida proliferation [115]. Conversely, when an individual is on a carbohydrate-rich diet, Candida spp. may predominate, and outcompete with Prevotella spp. over simple polysaccharides, resulting in a more obesogenic phenotype and reduced SCFA production. Yet, syntrophic relationships may also co-exist, as starches broken down by Prevotella spp. may serve as substrates for Candida spp. metabolism [63].

In contrast to a mutualistic relationship exhibited by the fungi or archaea, viruses are dependent on inter-kingdom interactions with its obligate host for its replication. In inflammatory bowel disease (IBD), there was a significant increase in bacteriophage richness, namely the expansion of Caudovirales, with a concomitant decrease in bacterial diversity and richness, compared to household controls [118]. When viromes from patients with healthy and inflamed IBD tissue were transplanted into murine models of ulcerative colitis, the mice transplanted with healthy viromes showed a longer colonic length, reduced stool lipocalin and colonic interleukin-6 levels, demonstrating a reduction of colonic inflammation [119]. This was not observed in mice transplanted with IBD viromes, which showed persistence of gut inflammation, reduction in the intestinal barrier integrity and a promotion of macrophage activation [119]. We postulate that a similar phenomenon may also be in play in obesity, which is also a chronic inflammatory condition, though the fluctuations in virome and bacteriome diversities may not be as straightforward.

In FVT experiments by Rasmussen et al., the reduction in weight gain and glucose resistance seen following transplantation of lean mice virome to high fat diet mice were no longer seen when high fat diet mice were pre-treated with ampicillin [27]. This suggested that the initial antibiotic-mediated disruption of the gut bacteriome impacted the metabolic effects mediated by the virome transplantation [27]. When comparing the alpha diversity of the bacteriome and virome, ampicillin-treated high fat diet mice had a less diverse bacteriome, but contrastingly more diverse virome [27]. This was in stark contrast to lean mice and high fat diet mice with lean FVT, which showed a more diverse bacteriome, and less diverse virome [27]. Interestingly, high fat diet controls had a less diverse bacteriome and virome in both instances [27].

Such findings may potentially be explained by the altered dynamics of the predator-prey relationship between the bacteriophages and bacteria in the gut, as it transitions to a “piggyback the winner” model of growth, depending on the abundance of its host and resources [120]. By using the lambda phage model of lysis and lysogeny, the switch to lysogeny on the lambda phage host Escherichia coli, may be influenced by the nutrient milieu [[121], [122], [123]], suggesting a potential link with dietary and nutrient intake on the gut virome composition. As seen in ΦcrAss001 phage and Bacteroides intestinalis, a steady state level of predator and prey persists, even though ΦcrAss001 phage is a lytic phage [45]. As the “piggyback the winner” model postulates, in settings where bacterial populations are rapidly expanding, certain strains develop a competitive growth advantage [120]. By delaying host cell lysis or through lysogeny (or pseudolysogeny for lytic phages), the phages can “piggyback” on the abundant bacterial host and maximize its chances of infecting the highest number of bacterial strains and hence maximising its own ability to replicate [120]. This creates a steady state of phage and host abundance, without proliferating at the expense of the host [113]. In instances of indiscriminate bacteriome depletion with broad spectrum antibiotic treatment, there is generalised depletion across most phyla. Majority of bacterial hosts have sustained cellular damage that may limit its proliferative ability and fitness. Lysogeny will only disadvantage the phage in such situations, as the host cell gene expression is not geared towards replication, but cellular repair, which does not contribute to replication of the prophage genetic material. Thus, lysis of all infected host cells while they were still viable, thus “abandoning the sinking ship”, will help maximise the phage progeny, thus accounting for the less diverse bacteriome, both as a result of antibiotic depletion and phage lysis, but more diverse virome due to generalised release of new daughter phages from the infected hosts, across multiple bacterial phyla. Resistant bacterial strains to antimicrobial treatment are likely a minority, thus its low abundance may not impact the overall bacteriome diversity. However, in cases of caloric restriction, such as after bariatric surgery, active dietary or pharmacotherapeutic interventions, reduced gut intraluminal nutrient transit reduces the resources available to gut bacteria. However, this does not result in an indiscriminate obliteration of most gut microbes. Metabolic activity may be decreased but the bacterial strains remain viable. In addition, certain microbes thrive in low caloric conditions, such as Akkermansia muciniphilia and Alistipes sp. [124]. The overall persistence of the overall bacteriome community, with the selective advantage of specific strains in low caloric conditions, likely contributed to the increased overall evenness in the gut microbial habitat [125], hence increasing bacteriome diversity. This is supported by the reported increase in Shannon index rather than the observed counts in alpha diversity, which considered both abundance and evenness [27]. Erez et al. have demonstrated that phages switch to lysogeny if concentrations of a communication peptide produced by the infected host into the surrounding is sufficiently elevated [121]. Given host fitness is not impaired, and replication still persist, albeit at reduced frequencies, the overall adoption of lysogeny by most phage populations will continue to be a more low-risk survival strategy. This likely contributed to the resultant decrease in viral release and hence reduction in virome diversity. The “kill the winner” approach may be transiently activated, in situations where a temporary growth advantage of a bacterial strain may proliferate at the expense of other susceptible bacterial hosts, thus maintaining overall community compositions. In obese individuals on the contrary, increased gut nutrient availability allows for the proliferation of more bacterial strains, though there may be a competitive growth advantage of the more abundant phyla at the expense of other less rare groups. This results in an overall reduction in evenness and bacteriome diversity. Phages specific to the bacterial strains with a proliferative advantage will “piggback the winner”, entering lysogeny (or delayed lysis and pseudolysogeny for lytic phages), maximising their own genetic material proliferation. Phages specific to less dominant strains may also continue to adopt a lysogenic strategy if the communication peptide threshold is crossed [121] or convert to a lytic strategy if the strains are at a proliferative disadvantage. However, given that these are usually rare bacterial taxa, the overall virome diversity hence is also correspondingly decreased. Such a finding will need to be confirmed via future studies, such as through the investigation of the role of temperate: lytic phages, such as Caudovirales: Microviridae ratio in obesity, or through studying free phage genetic material as a surrogate for phage lysis, to better help in our understanding of their role in inter-kingdom interactions and its contribution to obesity.

In addition to the virome, mycobiome and archaeome, other members of domain eukarya have also been implicated in human obesity. A cross-sectional study of 104 obese individuals in Spain found the presence of unicellular protozoa in 51% of stool samples, with Blastocystis hominis being the most common [126]. Majority of these patients have asymptomatic parasitosis [126]. Interestingly, patients with parasitosis were found to be significantly less insulin resistant compared to patients that were not colonized, though this was not significant in their subgroup analysis of individual protozoa, with no difference in nutritional deficiencies between groups [126]. A nation-wide observational study in Mexico of childhood obesity reported that children and adolescents with prior Ascaris lumbricoides or protozoa infections were more likely to have a higher BMI adjusted for age up to 12 years after parasitosis [127]. It is possible that exposure of the gut microbiome to these parasites, may result in persistent gut microbiota remodelling. Possibly, the competition for nutrients with parasitic infections, not only impacts the host, but also exerts a selection pressure on gut microbiota. This may select for an enterotype that is more efficient at luminal energy harvest, which eventually influences the host phenotype through modification of nutrient absorption [124,125], gut permeability and inflammation, as well as other neurohormonal effects.

6. Small intestinal microbiome in obesity

Lastly, it is worth noting that most studies that investigated the gut microbiome had focused on stool samples as a surrogate for the composition of the entire gut microbiome. Studying stool have their inherent advantages, including its non-invasive nature for collection, allowing for repeated sampling with minimal risks to participants. However, many studies have shown that the stool microbiome may not be representative of the diverse gut microbiome [128] and can be affected by stool consistency [129,130]. It is also not representative of the mucosal microbiome [131]. Given that obesity is a metabolic disease related to nutrient uptake, and the small intestine is pivotal in nutrient absorption, understanding the role of the small intestinal microbiome is paramount in unravelling the complex interplay of microbiome–host interactions on the obese phenotype. In contrast to stool or even colonic samples, the small intestinal mucus layer is compositionally and functionally different from colonic mucus [132]. It is thinner, more likely to facilitate nutrient uptake [132], and contains antimicrobial peptides [133]. This results in a distinct small intestinal microbiome composition, which has been reported to have a better correlation with obesity compared to stool microbiota samples [134]. The different microbiome profiles exert distinct downstream metabolic effects in contrast to the colonic microbiome [135]. The small intestinal microbiome may more directly exert neurohormonal effects on the host, impacting gut hormones such as glucagon-like peptide 1 [136], while the colonic microbiome may better reflect energy harvest [10]. However, almost all small intestinal microbiome studies related to obesity thus far, have focused on the bacteriome [135,137]. In addition, unlike stool sampling, profiling the small intestinal microbiome is technically challenging, with most studies employing invasive endoscopic procedures [135,138,139], or conveniently sampling them intraoperatively during bariatric surgery [[140], [141], [142]]. Recently, non-invasive ingestible capsule-based sampling methods have been employed with promising results [143,144]. Non-invasive longitudinal profiling of the gut intestinal mucosa in the future, will be important to help us understand the complex inter-kingdom interactions amongst various microbial domains, and its impact on the host.

7. Challenges and translational impact

Knowledge of how the inter-kingdom interactions of the gut microbiome may impact the host can have a direct translational impact. This information can have a direct therapeutic impact, or it can serve as predictive or prognostic biomarkers, which may translate to personalised pharmacotherapies or surgeries in combating obesity (Figure 2). However, further research is required to move beyond correlations to more mechanistic studies, with subsequent translation into clinical studies. Significantly, unlike the bacteriome, the reference databases for the non-bacterial microbiome especially for the virome is incomplete, which have largely limited our ability to accurately identify and classify these viruses [145]. In addition, complex environmental interactions influence the host microbiome bidirectionally [146], with alterations at different time points, and at different sites along the entire gastrointestinal tract. These challenges have thus far, limited the ability to translate various clinical findings into any widely adopted clinical solution in the management of obesity.

Figure 2.

Summary of the possible translational avenues of the non-bacterial microbiome in obesity (Created by biorender.com).

It is however, still hoped that future phage-based therapies, or “provirotics” that help with the lytic-lysogenic switch in phages, may eventually have a benefit in modulating host responses, particularly for individuals with poor weight loss or metabolic outcomes with conventional pharmacotherapy or bariatric surgery. In addition, the concept of “infectobesity” [147,148] still needs to be further explored. The role of combination biomarkers, involving the gut virome, bacteriome, mycobiome, archaeome, with the host phenome, utilised in unison via a meta-omic approach, may serve as a much more powerful prognosticator of long term metabolic or cardiovascular risk. This may allow for early primary initiation of primary preventive strategies, including lipid-lowering agents or cyclooxygenase inhibitors, to reduce atherosclerosis, chronic inflammation and long-term mortality. Meta-omic biomarkers may also serve as predictors for weight loss outcomes to bariatric surgery or pharmacotherapy. It may also directly impact bariatric surgical interventions, such as allowing the surgeon to tailor the intestinal bypass limb length, or the sleeve diameter to the predicted weight response of the patient, thus minimising the probability of suboptimal clinical response or poor metabolic remission after bariatric surgery, as well as reducing the complications of malnutrition or reflux esophagitis, which may occur as a result of an excessive long intestinal bypass or excessively tight gastric sleeve respectively.

8. Conclusion

While host–bacteriome interactions are important in the understanding of host adiposity, the association and impact of the non-bacteria microbiome are increasingly being recognized to be equally important in understanding the complex interplay of host and microbe in metabolic disorders and obesity. Phage-bacterial, bacterial-fungal and bacterial–archaea interactions, as discussed above, can all impact the host obese phenotype. Better understanding of these inter-kingdom cross-talks between the host, bacteriome and non-bacterial microbiome can help us fill knowledge gaps in these areas, which have potential translatable outcomes. This may include prognostication of outcomes, through the discovery of convenient, non-invasive or even volatile biomarkers that may predict for successful or suboptimal weight loss outcomes, as well as therapeutic outcomes, such as the use of phage-based therapies or fungal based probiotics in combating metabolic syndrome and obesity.

CRediT authorship contribution statement

Koy Min Chue: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Sunny Hei Wong: Writing – review & editing, Supervision, Project administration, Conceptualization. Tao Zuo: Writing – review & editing, Visualization, Supervision, Project administration, Conceptualization. Yusuf Ali: Writing – review & editing, Visualization, Supervision, Project administration, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.