The Microbiota-Gut-Brain Axis May Be the Missing Link Between Autism Spectrum Disorder and Anorexia Nervosa

Loss of GBB and BBB integrity and neuroinflammation are both associated with various neurological disorders, including depression, schizophrenia, ASD, and AN (Garcia-Gutierrez et al., 2020; Liang et al., 2018; Obrenovich, 2018; Rudzki and Szulc, 2017; Gibson and Mehler, 2018; Lobzhanidze et al., 2019; Seitz et al., 2019; Fields et al., 2018; van de Wouw et al., 2017; Karakula-Juchnowicz et al., 2017). In addition to causing BBB permeability and neuroinflammation, elevated circulating cytokines and bacterial metabolites have been shown to directly alter brain activity, further supporting the hypothesis that dysbiosis-induced leaky gut can cause brain dysfunction (Mangiola et al., 2016; Obrenovich, 2018; Liang et al., 2018; Liu and Zhu, 2018; Aresti Sanz and El Eidy, 2019).

MGB axis dysfunction as a mechanism of ASD pathogenesis

One of the fastest growing areas of MGB axis research is focused on its implications for the etiology of autism spectrum disorder (ASD). ASD is a neurodevelopmental disorder marked by stereotyped and repetitive behaviors, impaired social communication, and sensory processing disorders (Israelyan and Margolis, 2019; Galiana-Simal et al., 2017). While the prevalence of ASD in the United States is as high as 1 in 54, the etiology of the disorder is still unclear, in part due to its heterogeneity of presentation (CDC, 2020; Israelyan and Margolis, 2019; Vissoker et al, 2015; Sanctuary et al, 2018; Xu et al, 2019; Azhari et al, 2018).

A growing body of neurological, immunological, and gastrointestinal-based theories points to the potential role of the MGB axis and gut dysbiosis in ASD pathogenesis (Azhari et al., 2018; Srikantha and Mohajeri, 2019; Fattorusso et al., 2019; Tye et al., 2018). This is supported by the high frequency of gastrointestinal and immunological comorbidities in children with ASD. The literature reports a four times greater prevalence of GI comorbidities in children with ASD than without, and the common comorbidity of constipation has been found to correlate with symptoms like social impairment, aggression, and compulsivity (Sanctuary et al., 2018; Israelyan and Margolis, 2019; Fattorusso et al., 2019). Additionally, the correlation between ASD severity and immune dysregulation (food allergies, asthma, and systemic inflammation evidenced by upregulated proinflammatory cytokines, natural killer cells, interferons, and autoantibodies) further points to the potential role of MGB axis dysfunction (Tye et al., 2018; Vissoker et al., 2015; Azhari et al., 2018; Fattorusso et al., 2019). Thus, a complex phenotypic pattern of gastrointestinal, immunological, and neurological dysfunction has been identified, suggesting a connection between ASD and MGB axis dysfunction (although a causal relationship is yet to be established by human studies) (Azhari et al., 2018; Srikantha and Mohajeri, 2019; Fattorusso et al., 2019; Xu et al., 2019). The induction of ASD-associated traits (such as repetitive behaviors and related social and cognitive dysfunctions) in mice through microbiome manipulation demonstrates that gut dysbiosis indeed plays a role in ASD, and trials of Microbiota Transfer Therapy show promise for improving ASD symptoms and GI comorbidities; thus, it is worthwhile to continue to investigate the patterns, mechanisms, and causal relationships underlying this connection, especially within the context of human ASD patients (Kang et al., 2017; Wang et al., 2019; Liang et al., 2018; Srikantha and Mohajeri, 2019; Rudzki and Szulc, 2018; Mangiola et al., 2016; Sanctuary et al., 2018; Fattorusso et al., 2019).

While GI disorders do not present in all cases of ASD, and neither do autoimmune and anxiety disorders, the correlation between these three comorbidities in a subset of cases leads some to believe that there is a particular immune-mediated subtype of ASD distinguished by degree of systemic inflammation (Tye et al., 2018; Fattorusso et al., 2019; Azhari et al., 2018). Given prior research, it seems likely that this subtype might be found in patients with co-occurring ASD, GI disorders, and immunological disorders; this is explained by Azhari and colleagues’ gut-immune-brain paradigm (2018).

The “gut-immune-brain” paradigm of ASD

One hypothesis grounded in MGB axis dysfunction is the “gut-immune-brain” paradigm, proposed by Azhari and colleagues (2018). Featuring parts of the gut-microbiota, old friends, and leaky gut theories, the paradigm states that gut dysbiosis leads to the autistic phenotype and its comorbidities through four interacting mechanisms: an increase in intestinal permeability (“leaky gut”), toxin production and exposure, aberrant immune response, and metabolic abnormalities (Azhari et al., 2018). This theory is supported by repeated findings of dysbiosis, abnormal levels of circulating bacterial metabolites, immune dysregulation, and impaired metabolic processes in association with ASD (Fattorusso et al., 2019; Israelyan and Margolis, 2019; Sanctuary et al., 2018; Srikantha and Mohajeri, 2019; Obrenovich, 2018; Rudzki and Szulc, 2018; Mangiola et al., 2016; Dempsey et al., 2019; Shultz et al., 2014; Garcia-Gutierrez et al., 2020; Nankova et al., 2014; MacFabe, 2012; Xu et al., 2019; Fields et al., 2018; Lobzhanidze et al., 2019). Given differential immune responses between males and females (for example, males exhibiting a more pro-allergic response), an immune-driven paradigm of ASD could also explain the overrepresentation of males compared to females in the autistic population (Fields et al., 2018; Werling and Geschwind, 2014). Furthermore, the explanatory power of the paradigm lies in its application to many of the leading proposed mechanisms of ASD and its ability to account for their interactions such as through positive feedback loops.

It is important to note that stress alone has been demonstrated to alter the gut microbiome as well as gut permeability, and so one might argue that a genetic and/or environmental predisposition to barrier permeability and/or immuno-inflammatory responses could cause enough physiological stress to induce dysbiosis (Liu and Zhu, 2018; Rudzki and Szulc, 2018; Liang et al., 2018). In this case, it is possible that dysbiosis is merely correlated with the immune response but does not cause it in the first place. However, dietary intervention, probiotic, and fecal transplantation trials demonstrating amelioration of stereotyped behaviors and comorbidities in ASD suggest that a causal relationship might indeed exist between dysbiosis and ASD pathogenesis (Rudzki and Szulc, 2018; Mangiola et al., 2016; Sanctuary et al., 2018; Srikantha and Mohajeri, 2019; Fattorusso et al., 2019).

Patterns of dysbiosis in ASD

While studies on ASD patient microbiomes have yielded heterogenous and at times conflicting results, a number of significant patterns have emerged. Analyses of the fecal microbiota of children with ASD compared with typical development peers and/or siblings reveals that Bacteroidetes, Clostridium, Desulfovibrio, Faecalibacterium, Proteobacteria (including Sutterella), Lactobacillus, and Actinobacteria are overrepresented in ASD; Bifidobacterium, Prevotella, Coprococcus, and Veillonellaceae are lacking, and overall species diversity is diminished as well (Fattorusso et al., 2019; Tye et al., 2018; Nankova et al., 2014, Azhari et al., 2018; Garcia-Gutierrez et al., 2020; Sanctuary et al., 2018; Fields et al., 2018; Oleskin et al., 2016; Xu et al., 2019; Wang et al., 2019; Dempsey et al., 2019). Additionally, higher Firmicutes: Bacteroidetes ratios have been noted in ASD (Srikantha and Mohajeri, 2019; Sanctuary et al., 2018). While no definite ASD microbiome profile has emerged, the Human Microbiome Project has demonstrated that the microbiomes of two healthy humans can vary greatly in comparison to one another, highlighting that perhaps it is not the presence of an ideal profile, but the balance between competing and symbiotic species that demarcates a healthy gut from dysbiosis (Fattorusso et al., 2019; Israelyan and Margolis, 2019; Fields et al., 2018). As van de Wouw and colleagues note, a less diverse gut microbiome might allow one species to dominate, influencing the host in ways otherwise impossible; additionally, a variety of dysbiotic profiles could disrupt the microbial ecosystem by hampering cross-feeding (the supplying of nutrients for one species by the metabolites of another). This is particularly relevant to SCFA production, a mechanism of interest for gut-related neurological disorder pathogenesis (2017).                       

Bacteroidetes, Clostridium, and Desulfovibrio-produced propionate as a mechanism of ASD

There are several hypotheses pertaining to certain species that are found in elevated or diminished abundance which could explain the correlation between dysbiotic patterns in ASD and the etiology of the disorder. For example, the lack of Bacteroidetes, Clostridium, and Desulfovibrio in ASD patients lends support to a propionate-related theory of ASD pathogenesis (Garcia-Gutierrez et al., 2020; Nankova et al., 2014; MacFace, 2012). It is known that the aforementioned bacteria produce propionate (PPA), one of the three primary SCFAs, and the overgrowth of these species in ASD correlates with elevated levels of PPA in stool (Dempsey et al., 2019; Shultz et al., 2014; Lobzhanidze et al., 2019; Nankova et al., 2014; Srikantha and Mohajeri, 2019; Garcia-Gutierrez et al., 2020; Fields et al., 2018). It is known that PPA can cross the BBB, triggering an immune response; chronic PPA administration in mice is correlated with neuroinflammatory biomarkers (Garcia-Gutierrez et al., 2020; Lobzhanidze et al., 2019). There is also evidence that excess PPA exposure impacts amygdala development and function; this is pertinent to ASD as social communication and emotional processing is often impaired in the disorder (particularly gaze, attachment behavior, and social behavior) (Lobzhanidze et al., 2019). PPA exposure in mice decreases social motivation, reduces neuron number in the amygdala, and activates glial cells, causing further inflammation (Lobzhanidze et al., 2019). Although the data is currently inconclusive, there is evidence that structural changes of the amygdala correlate with degree of social impairment in ASD; the impacts of PPA on the amygdala could explain this association (Lobzhanidze et al., 2019). Furthermore, PPA is a rather convincing mechanism of ASD as murine modeling has shown an association between PPA and an array of ASD-related traits and comorbidities such as reduced food intake, hyperactivity, repetitive behaviors, obsessive/compulsive behaviors, mood disorders, impaired social behavior, altered fatty acid metabolism, altered BBB integrity, reduced gastric motility, developmental delay, GI dysfunction, immune dysfunction, and abnormal gastrointestinal serotonin secretion (Dempsey et al., 2019; Garcia-Gutierrez et al., 2020; Nankova et al., 2014; MacFabe, 2012; Lobzhanidze et al., 2019).

Faecalibacterium overgrowth as a cause of ASD-related immune dysfunction

Faecalibacterium is often overrepresented in the gut microbiomes of ASD patients (Dempsey et al., 2019; Garcia-Gutierrez et al., 2020; Xu et al., 2019; Azhari et al., 2018). This trait has been associated with systemic immune dysfunction (a common comorbidity of ASD), which is the result of altered butyrate (BA) production, where BA is a known metabolite of Faecalibacterium (Xu et al., 2019; Valles-Colomer et al., 2019). Some studies have measured high BA as correlated with high Faecalibacterium in ASD (Dempsey et al., 2019; Garcia-Gutierrez et al., 2020). This could also explain other comorbidities of ASD associated with high BA such as satiety/reduced intake, indigestion, constipation, altered social behavior, and immune dysfunction (Garcia-Gutierrez et al., 2020; Lin et al., 2012). However, the data and ensuing conclusions on the role of BA in ASD are far from conclusive: other studies found reduced BA in ASD and explained impaired GBB integrity as the intestinal mucosa is constructed from SCFAs including BA (Lobzhanidze et al., 2019; Srikantha and Mohajeri, 2019; Valles-Colomer et al., 2019). While different studies have found either increased or decreased levels of BA in ASD patients, it seems likely that either way, a BA imbalance contributes to ASD etiology. As a key SCFA, BA is implicated in other SCFA-related processes relevant to ASD and its comorbidities, such as serotonin production and secretion, inflammatory immune response, intestinal mucous production, vagal signaling, neurodevelopment, appetite regulation (via secretion of anorexigenic hormones GLP-1 and PYY), metabolism, gene expression (via histone acetylation and methylation), BBB maintenance, and lipid metabolism (Mangiola et al., 2016; Dempsey et al., 2019; Garcia-Gutierrez et al., 2020; Yang et al., 2019; Nankova et al., 2014, MacFabe, 2012; Oleskin et al., 2016; van de Wouw et al., 2017; Fields et al., 2018; Nankova et al., 2014).

Proteobacteria overgrowth as a cause of ASD-related immune dysfunction through LPS and TNF-ɑ

Proteobacteria, specifically Sutterella, have been found at elevated numbers in ASD patients and might contribute to ASD pathogenesis in line with the gut-immune-brain paradigm (Azhari et al., 2018; Fattorusso et al., 2019; Tye et al., 2018; Fields et al., 2018). An increase in Proteobacteria is associated with metabolic endotoxemia (low-grade systemic inflammation) through the overproduction and circulation of lipopolysaccharides (LPS) (van de Wouw et al., 2017; Fields et al., 2018). LPS interacts with toll-like receptor 4 (TLR4) to increase gut permeability, leading to a cascade of effects on CNS function, including altered SCFA circulation, vagal signaling, stress response and inflammatory immune response (Fields et al., 2018; Garcia-Gutierrez et al., 2020; Xu et al., 2019; Karakula-Juchnowicz et al., 2017). In mice, increased levels of circulating LPS are associated with repetitive behaviors, underscoring the implications of LPS for not only ASD comorbidities but also for hallmark traits (Fields et al., 2018).

Furthermore, LPS stimulates the secretion of TNF-ɑ from liver cells, leading to peripheral inflammation, inhibition of INF-β (which is essential for the intestinal mucosa lining), and impairment of tight junction formation in the GBB and BBB, thereby further perpetuating intestinal and brain permeability and systemic inflammation (Azhari et al., 2018; Karakula-Juchnowicz et al., 2017). Notably, TNF-ɑ correlates with the severity of ASD symptoms in general and GI dysfunction specifically (Garcia-Gutierrez et al., 2020; Srikantha and Mohajeri, 2019). The molecule is thought to influence CNS function by crossing the BBB, inducing neuroinflammation (Garcia-Gutierrez et al., 2020; Srikantha and Mohajeri, 2019). TNF-ɑ is also thought to act directly on the CNS to cause the release of corticotropin-releasing factor (CRF), an anorexigenic neuropeptide that also affects gastric emptying, immune response, and mood (Holden and Pakula, 1996; Furman, 2007). Moreover, elevated TNF-ɑ results in hypothalamic dysregulation manifesting in reduced insulin and increased cortisol, which further decreases IL-1, IL-2, and 5-HT secretion, and thereby impacts serotonergic signaling (Holden and Pakula, 1996).

Underrepresentation of Bifidobacterium as a mechanism of ASD through decreased SCFAs

Bifidobacterium has been repeatedly measured in decreased quantity in the microbiomes of ASD patients (Xu et al., 2019; Oleskin et al., 2016; Fattorusso et al., 2019; Azhari et al., 2018; Garcia-Gutierrez et al., 2020). Given that Bifidobacterium is a producer of SCFAs, low Bifidobacterium in ASD patients might explain decreased SCFAs and dysregulation of the gut mucosal barrier, anti-inflammatory pathways, vagal signaling, appetite, serotonin production, neurodevelopment, and gene expression (Xu et al., 2019; Mangiola et al., 2016; Dempsey et al., 2019; Garcia-Gutierrez et al., 2020; Yang et al., 2019; Nankova et al., 2014; Oleskin et al., 2016; van de Wouw et al., 2017; Fields et al., 2018; Srikantha and Mohajeri, 2019). It is thought that adequate amounts of metabolites from Bifidobacterium in particular are essential for the regulation of the hypothalamus-pituitary-adrenal (HPA) axis; thus, decreased Bifidobacterium levels may contribute to impaired stress response and explain elevated cortisol levels in ASD fecal samples (Oleskin et al., 2016; Rudzki and Szulc, 2018; Obrenovich, 2018). Furthermore, an altered stress response would perpetuate immune dysregulation, contributing to neuroinflammation, GBB and BBB permeability, and autoimmune comorbidities.

Underrepresentation of Prevotella as a mechanism of ASD

Finally, it is worth briefly noting decreased levels of Prevotella in ASD, given the bacterium is negatively associated with gut permeability and positively associated with GI dysfunction (Tye et al., 2018; Fattorusso et al., 2019; Fields et al., 2018; Garcia-Gutierrez et al., 2020). Low Prevotella is also associated with increased leptin and decreased ghrelin, possibly accounting for feeding-related comorbidities in ASD (Seitz et al., 2019).

MGB axis dysfunction as a mechanism of AN pathogenesis

Anorexia nervosa (AN) is another neurological disorder which has recently received attention as a condition possibly linked to MGB axis dysfunction. With MGB axis-related traits and comorbidities such as gut dysbiosis, mood disorders, leaky gut, GI dysfunction, autoimmune disorders, sensory processing disorders, alexithymia, abnormal amygdala activity, impaired social functioning, serotonergic dysregulation, OCD, and ASD, AN is plausibly perpetuated by the immunological, nervous, and endocrine effects of gut dysbiosis (Hommer and Swedo, 2021; Gibson and Mehler, 2019; Bentz et al., 2016; Kerbeshian and Burd, 2009; van de Wouw et al., 2017; Galiana-Simal et al., 2017; Joos et al., 2011; Holden and Pakula, 1996; Karakula-Juchnowicz et al., 2017). Moreover, as a psychiatric disease dealing expressly with gut-related activities (such as eating), it makes sense to consider AN in relation to gut-brain communication. It is thought that interactions between the ENS and microbiota are involved in regulating AN-related processes such as food choice-related emotions, food intake, dopaminergic signaling, satiety, and hunger (Oleskin et al., 2016).

Once again, it is important to consider that many of the traits listed above (including dysbiosis and leaky gut) may be the result (and not the cause) of AN behavior, psychological or physiological stress, and/or other possible etiological pathways, and that a causal relationship between gut dysbiosis and AN has yet to be established (Breton et al., 2021; Gibson and Mehler, 2019; Glenny et al., 2017; Karakula-Juchnowicz et al., 2017; Liang et al., 2018). Repeated observations of dysbiosis in AN patients and existing knowledge linking MGB axis-related factors to AN symptomatology suggest that further investigation into a causal relationship between gut dysbiosis and AN is warranted (MacFabe, 2012; Breton et al., 2021; Gibson and Mehler, 2019; Seitz et al., 2019; van de Wouw et al., 2017; Glenny et al., 2017; Holden and Pakula, 1996; Karakula-Juchnowicz et al., 2017; Liang et al., 2018). Additionally, experiments modifying the microbiomes of mice have demonstrated a causal link between microbiome composition and anxious and depressive behaviors (Roubalovà et al., 2020). Finally, clinically noted overlaps between AN and ASD, paired with the earlier discussed evidence for a role of the MGB axis in the etiology of some forms of ASD, further support the consideration of AN as a MGB axis-related disorder (Kerbeshian and Burd, 2009; Galiana-Simal et al., 2017; Dattaro 2020).

Applying the gut-immune-brain paradigm to AN

Similar to Azhari and colleagues’ gut-immune-brain paradigm of ASD, an etiological pattern of AN can be derived from various MGB axis-related theories of neurological disorder in the literature (2017). Several sources propose gut dysbiosis as a causal factor of AN behaviors (van de Wouw et al., 2017; Karakula-Juchnowicz et al., 2017; Breton et al., 2021; Seitz et al., 2019). Both Karakula-Juchnowicz and colleagues and Seitz and colleagues hypothesize gut dysbiosis as a likely mechanism of AN pathogenesis through the induction of intestinal permeability and resulting immune response which ultimately interferes with mood and appetite (2017; 2019). This is supported by reports of a higher chance of autoimmune disorders in patients with AN (Seitz et al., 2019; Hommer and Swedo, 2017). While dysbiosis could occur for many reasons, it is known that malnutrition (for instance, as a result of restrictive feeding behaviors) alters the microbiome—for example, by promoting bacteria suited to a low-energy environment (Karakula-Juchnowicz et al., 2017; Glenny et al., 2017; Breton et al., 2021). In line with the prevailing thought that stress anticipates restrictive episodes, it is also known that stress alters the microbiome (Liu and Zhu, 2018; Rudzki and Szulc, 2018; Liang et al., 2018). It has been otherwise suggested that AN-related dysbiosis and GI permeability can be the result of infection and/or antibiotic use, which have been noted to commonly precede the onset of restrictive behaviors (Liang et al., 2018; Rudzki and Szulc, 2018; Holden and Pakula, 1996; Fetissov and Hökfelt, 2019). For example, it appears that Salmonella exposure may prompt overgrowth of the Enterobacteriaceae which are known to produce the ClpB mimetic of ɑ-melanocyte-stimulating hormone (ɑ-MSH), inducing satiety and behavior common to AN (Fetissov and Hökfelt, 2019).