Role of the microbiota-gut-brain axis in postacute COVID syndrome

Abstract

The COVID-19 pandemic has resulted in the infection of hundreds of millions of individuals over the past 3 years, coupled with millions of deaths. Along with these more acute impacts of infection, a large subset of patients has developed symptoms that collectively comprise “postacute sequelae of COVID-19” (PASC, also known as long COVID), which can persist for months and maybe even years. In this review, we outline the current knowledge on the role of impaired microbiota-gut-brain (MGB) axis signaling in the development of PASC and the potential mechanisms involved, which may lead to a better understanding of disease progression and treatment options in the future.

INTRODUCTION

The global severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic has infected hundreds of millions of people worldwide over the past 3 years, leading to millions of deaths following acute infection. In a meaningful subset of patients, symptoms comprising chronic postacute sequelae of COVID-19 (PASC) emerge following the resolution of acute infection, which has become a cause for concern. It is estimated that 1 in 13 (7.5%) of adults in the United States who have recovered from acute COVID-19 disease have PASC symptoms, defined as symptoms lasting 3 or more months following recovery, according to the Centers for Disease Control (CDC; 1). Although epidemiological studies leveraging electronic medical records found that sex, severity of initial infection, and certain comorbidities (including obesity, diabetes, and lung disease), coupled with socioeconomic status, all impacted the risk of development of PASC (2), the mechanisms that lead to prolonged disease remain to be fully elucidated. A few studies on the impact of the SARS-CoV-2 vaccine on the risk of PASC have suggested that vaccination may offer minor protection against PASC, possibly limited to fewer memory problems (3) and/or new onset of health conditions (4). Time between vaccination and infection and viral strain may determine the beneficial impact of vaccines on the development of PASC, with protection against PASC from vaccination greater for omicron compared with delta strains of SARS-CoV-2 (5). However, this remains a preliminary conclusion requiring additional studies.

Infection with the respiratory viral pathogen SARS-CoV-2 is associated with significant lung pathology; however, there is also a significant burden of pathology outside the respiratory tract. This burden includes gastrointestinal (GI) symptoms in a large subset of patients during acute COVID infection, including diarrhea, nausea, loss of taste, vomiting, and abdominal pain. Although these GI symptoms are not directly associated with patient mortality (6), they can be debilitating and are considered detrimental to the overall quality of life. GI symptoms were seen in 12% of patients following acute infection and up to 22% in patients suffering from PASC (7), representing a significant patient population. Fatigue, sleep disturbance, mood disorders, and cognitive dysfunction (including “brain fog” and memory impairment) are common psychological manifestations of postacute COVID infection (8, 9). These mental health outcomes have been seen even in those patients not hospitalized during the acute phase of their COVID infection (8, 10). The mechanisms leading to prolonged GI symptoms and increased mental health disorders in patients with PASC remain unknown. Observational evidence reveals gut microbiome compositional changes in patients with PASC, including a decreased abundance of beneficial butyrate producers (11), implicating the GI tract in disease (Fig. 1). Together, persistent GI and chronic neuropsychiatric symptoms in the presence of potentially long-lasting gut dysbiosis suggests that impaired postinfectious microbiota-gut-brain (MGB) axis signaling might account for long-term symptoms in patients with PASC. In this article, we discuss evidence for such a proposition and hypothesize potential underlying mechanisms.

Figure 1.

Microbiota-gut-brain axis and PASC. SARS-CoV-2 can enter the gastrointestinal (GI) tract and infect intestinal epithelial cells, causing local inflammation. Mucosal immune cells entering the circulation may subsequently gain access to the central nervous system, impacting the brain and causing cognitive deficits and brain fog. Alternatively, intestinal inflammation may trigger neuroimmune interactions, thereby activating sensory neurons and impacting cognitive function via a neural route. These potential signaling pathways suggest that microbiota-gut-brain communication may participate in PASC symptoms in at least a subset of patients. PASC, postacute sequelae of COVID-19; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

MICROBIOTA-GUT-BRAIN AXIS

The MGB axis is a bidirectional communication pathway between the gut and the brain that is heavily modulated by the gut microbiota (12). Changes in the fidelity of this communication pathway can result in disease. For example, depression, anxiety, and cognitive deficits are common behavioral impairments that are associated with altered gut microbial colonization in multiple preclinical animal models of disease and in patient populations (13). Overall, proper MGB axis signaling is crucial for the maintenance of overall health and well-being, with impairments in communication seen in patients with GI (irritable bowel syndrome, inflammatory bowel disease), metabolic (obesity, diabetes), and neurological (autism, Parkinson’s, Alzheimer’s) diseases (12).

The immune system plays a crucial role in MGB axis signaling; it is a primary responder to altered host-microbe interactions following dysbiosis of the gut microbiota, such as during enteric bacterial infection. Inflammation, in both the gut and the brain, is a prominent feature seen in many animal models of MGB axis disruption (14, 15), and in clinical diseases where altered MGB axis communication is observed (13). In addition, the extensive intrinsic neural network within the wall of the GI tract, which communicates with the extrinsic autonomic nervous system, may also provide an important means by which signals from the gut are transmitted to the central nervous system (CNS), including the brain. Although the exact pathways through which MGB axis communication is maintained remain to be completely elucidated, it is now clear that altered MGB axis signaling may cause significant and chronic symptoms in patients with multiple complex diseases.

The MGB axis has been extensively characterized in mouse models, allowing for increasingly in-depth mechanistic studies. Studies using germ-free (GF) mice have revealed a crucial role for the gut microbiome in regulating the brain and behavior, as exemplified by the presence of anxiolytic behaviors and cognitive deficits in these animals compared with colonized control mice (16, 17). Additional models of disrupted gut colonization, including those produced by antibiotic administration (18) or bacterial infection (14, 17), have revealed cognitive deficits similar to those seen in GF mice, further highlighting the importance of the gut microbiome in maintaining cognitive function. The precise pathway(s) by which microbes regulate cognition are not yet fully defined. However, the presence of altered neurogenesis (19), impaired myelination (18), and immature resident microglia cells (20) in dysbiotic states suggests potential mechanisms.

Given this important role of the gut microbiome in regulating cognitive function, it is tempting to speculate that many PASC-associated cognitive symptoms could result from impaired MGB axis signaling that persists after the resolution of acute infection. Whether these signaling deficits are due to an altered composition of the gut microbiota, GI pathophysiology, and/or downstream factors that lead to behavioral and cognitive impairments remain to be determined. However, mouse models of COVID infection should serve as excellent preclinical settings to study the MGB axis in PASC. For example, mice engineered to express the human receptor for the COVID virus, angiotensin-converting enzyme 2 (hACE2), have allowed the study of COVID and its sequelae following both acute (21) and chronic (22) infection. In fact, multiple preclinical animal models, including those in mice, ferrets, hamsters, and nonhuman primates, exist that can develop postinfectious PASC-like disease (23) and have the potential to identify MGB axis deficits induced by acute infection that may contribute to chronic symptoms in patients.

“BRAIN-FOG” AND NEUROLOGICAL IMPAIRMENTS

One of the top three complaints reported by patients suffering from PASC is the experience of “brain fog.” This colloquial term typically refers to symptoms including cognitive impairment in domains of attention and processing speed, short-term memory loss, and reduced mental acuity. These symptoms are also often seen in patients with other presumed postviral disease entities, such as chronic fatigue syndrome (CFS) or fibromyalgia (24). Neurological disorders that emerge following a viral infection can occur either directly via infection of the CNS, such as the olfactory bulb, or indirectly due to the host response to infection. In a recent COVID and cognition study in a community sample of patients, neurological and psychiatric symptoms during the initial acute phase of infection, or alternatively neurological, GI, and cardiopulmonary/fatigue symptoms during the persistent illness, predicted the subsequent experience of cognitive symptoms by patients (25). Cognitive deficits after COVID are primarily observed in the early stages of PASC, typically in the first month postinfection (24). PASC-associated neurological and psychiatric/mood disorders occur at a higher frequency in female versus male patients (26), similar to what is seen for the distribution of patients with CFS and fibromyalgia, suggesting the potential for a common pathway of pathogenesis. Altered hypothalamic-pituitary-adrenal (HPA) axis communication, significantly regulated by the gut microbiota and seen in all three disorders (27, 28), suggests that altered stress responses may underlie PASC and perhaps explain the sexually dimorphic association with PASC. However, the overall pathogenesis of symptomatology in patients with PASC is poorly understood and warrants additional studies to improve patient outcomes.

In a mouse model of SARS-CoV2 infection, using mice administered an adeno-associated virus expressing hACE2, impaired hippocampal neurogenesis, decreased oligodendrocytes, and myelin loss were observed associated with increased levels of cytokines/chemokines in the cerebrospinal fluid (most notably CCL11) compared with sham-infected controls (29). In addition, increased microglial activation was observed as indicated by increased Iba1 and CD68 costaining in subcortical white matter, in the absence of an overall increase in total number of microglia (29). The authors also observed related findings in humans, in patients with lasting cognitive symptoms after COVID, including brain fog, also exhibiting elevated serum CCL11 levels (29). Serum CCL11 levels have also been shown to increase with age and contribute to decreased adult neurogenesis and impaired learning and memory in mice (30). In postmortem tissue samples, moreover, humans with SARS-CoV-2 infection had increased CD68⁺ microglia/macrophage reactivity in the brain compared with individuals without SARS-CoV-2 infection (29). Together, these findings suggest that altered cognitive function in PASC may be mediated by increased neuroinflammation that persists after resolution of acute infection.

CHRONIC DYSBIOSIS OF THE GUT MICROBIOME

Colonization of the gut microbiome begins in early life and remains dynamic for the first 5 years of life (31). In adulthood, the microbiome is more stable, with fluctuations in composition and functional potential occurring as part of the normal dynamics of the microbiome and homeostatic interactions with the host and dietary components (32). In general, the adult microbiome remains stable yet responsive to physiological changes over decades (33). International travel, for example, can cause a temporary shift in the composition of the gut microbiome (34). The microbiome is typically then restored upon completion of travel; however, this can be abrogated following bacterial or viral infection, leading to diarrhea, in a subset of individuals (34). An altered composition of the gut microbiota is also associated with multiple diseases, although whether these alterations are a cause or consequence remains unclear in many instances.

Multiple studies over the course of the COVID-19 pandemic have revealed that the infection impacts the composition of the gut microbiome. A recent systematic review found that gut bacterial diversity was decreased in both patients with acute COVID-19 and those who have recovered from COVID-19, coupled with a decreased abundance of butyrate producers and increases in taxa with proinflammatory properties (35). This gut perturbation continued into the recovery phase after the resolution of an initial infection, suggesting that it may persist and could conceivably contribute, at least in part, to PASC in some individuals. Indeed, certain metagenomic ecological clusters were identified in severe COVID and predicted the development of PASC (36). Moreover, multiple commensal bacteria known to have beneficial immunomodulatory potential, including Faecalibacterium prausnitzii, Eubacterium rectale, and strains of Bifidobacteria, were underrepresented in infected patients at up to 30 days after the resolution of disease (37). This depletion of beneficial taxa was associated with increased proinflammatory cytokines in the plasma of patient, suggesting that these strains may normally be involved in maintaining appropriate immune responses to pathogens while preventing overt inflammatory responses (37). In addition, the oral microbiome was found to contain a significant increase in proinflammatory taxa, including Prevotella and Veillonella, in patients who developed prolonged PASC symptoms (38). These alterations in the oral microbiome of patients with PASC resembled those seen in patients suffering from CFS, providing a potential link between the two disorders (38). Together, these studies suggest that the gut microbiome may serve as a trigger through which PASC is initiated in patients with acute COVID, resulting in persistent symptoms well past the resolution of infection. Furthermore, given the known linkages between the gut microbiome and cognitive function via the MGB signaling axis, as discussed earlier, effects of SARS-CoV-2 infection on specific microbial communities could modulate beneficial bacterial metabolites, subsequently impacting neurogenesis or neuroinflammation in the brain, and in turn contributing to the cognitive symptoms of PASC specifically.

In addition to the gut bacterial microbiome, increasing evidence suggests that SARS-CoV-2 infection impacts the gut virome. The virome exhibits significant interindividual heterogeneity and age-dependent characteristics, similar to the microbiome, and is predominantly composed of bacteriophages in the healthy adult gut (39). Infection with SARS-CoV-2 impacted both the enteric RNA and DNA virome, with changes persisting after the resolution of infection (40). The fecal virome in patients with active COVID-19 had increased stress-, inflammation-, and virulence-associated gene-encoding capacities, including for those important for bacteriophage integration, DNA repair, metabolism, and virulence associated with their bacterial host (39). Consequently, COVID-induced virome shifts are closely linked with microbiome-induced changes and their impact on bacteriophages (41). Patients with severe COVID demonstrated an increased abundance of opportunistic pathogens coupled with decreased butyrate-producing bacterial species. In addition, fecal viral abundance was inversely correlated with disease severity and immune activation (40). Finally, limited studies to date have also identified detrimental impacts of SARS-CoV-2 infection on the gut mycobiome. A pilot study performed in the early months of the pandemic found that infection was associated with an enrichment of fecal fungal pathogens, including Candida and Aspergillus (42). These viral-fungal coinfections contributed to unstable mycobiomes and were thought to have led to hyperactive immune responses and cytokine storm in response to fungal pathogen-associated molecular patterns (43). Taken together, these findings suggest that the broader impact of SARS-CoV-2 infection on the gut microbiome involves not only the infection-induced dysbiosis but also a subsequent impact on bacteriophages and fungi within the gut. Whether these alterations contribute to development of PASC or whether they can be exploited for therapeutic benefit remains to be determined.

GI SYMPTOMS IN COVID

Although the GI manifestations of acute SARS-CoV-2 infection are now well recognized (44), the extent to which these effects contribute to the subsequent occurrence of PASC remains unproven. SARS-CoV-2 directly infects and replicates within intestinal epithelial cells (45), given that they express high levels of the ACE2 receptor and the transmembrane protease serine subtype 2 (TMPRSS2), a cellular protease important for viral entry (46). Intestinal epithelial cell infection leads to local inflammatory responses and subsequent impacts on the gut microbiome, which may all contribute to GI symptoms (47). Studies using human gastric and intestinal organoids were crucial early in the pandemic to study viral entry and novel antiviral therapy against SARS-CoV-2-induced disease (48, 49). Large interindividual variability in susceptibility to infection was highly correlated to ACE2 expression levels (50). The ability of the SARS-CoV-2 virus to infect intestinal epithelial cells may represent a crucial component that dictates the potential for the development of MGB axis deficits in patients with PASC.

A recent meta-analysis found that GI manifestations of PASC, which include loss of taste, loss of appetite, abdominal pain, nausea, vomiting, and diarrhea or constipation, are not related to the initial severity of acute disease following COVID infection (7). In addition, a significant number of patients develop dyspepsia and irritable bowel syndrome following acute COVID infection. Indirect markers of gut barrier dysfunction, including serum peptidoglycan (PGN) and lipopolysaccharide (LPS) levels and fatty acid binding protein 2 (FABP2), were elevated in hospitalized patients with acute SARS-CoV-2 infection compared with healthy individuals, and barrier dysfunction plausibly could account for these findings (51). Increased serum levels of bacterial products may be an indicator of sepsis, and a subsequent “cytokine storm,” which represents a leading cause of COVID-related deaths in hospitalized patients (52). Fecal RNA shedding of SARS-CoV-2 was also prolonged in patients with ongoing GI symptoms, including abdominal pain, nausea, and vomiting (53). This prolonged fecal viral shedding suggests the presence of ongoing gut infection, after clearance of lung infection, and likely ongoing inflammation that could contribute to the presence of persistent PASC symptoms. In line with this, GI PASC symptoms seem unique to correlate with the presence of newly expanded cytotoxic CD8⁺ and CD4⁺ T-cell populations, specific to SARS-CoV-2 and emerging in the 2–3 mo after initial symptom onset (54).

Overall, the admittedly limited data available thus far suggest that persistent active GI infection with SARS-CoV-2 may be contributing to GI symptoms and gut dysbiosis. The corollary of this is that restoring the gut microbiome or its function(s) following infection, such as by administration of beneficial probiotics or replacement of short-chain fatty acids normally supplied by bacteria, could resolve GI pathology, inflammation, and persistent symptoms.

GUT-LUNG AXIS

The gut-lung axis is part of the shared mucosal immune system, which has the potential to coordinate immune responses across distant sites (55). Chronic lung and chronic GI diseases commonly occur together, with asthma or chronic obstructive pulmonary disease often seen in patients with inflammatory bowel disease or irritable bowel syndrome, for example (55). Similarly, given the important role of eosinophils in modulation of both GI and lung inflammation, allergic airway disease, and eosinophilic esophagitis often occur concurrently, however, the exact mechanisms that underlie this bidirectional relationship remain poorly understood (56). With respect to COVID-19, although patients with asthma are at an increased risk of developing PASC, patients with eosinophilic, or allergic, asthma may actually be protected (57), suggesting a protective role for eosinophils in long-term outcomes following infection.

The role of the gut microbiome in mediating host responses to respiratory viral infections is poorly understood. COVID-19-induced dysbiosis of the gut microbiome is thought to impact mucosal immune responses, increasing intestinal permeability and leading to secondary bloodstream infections, with the potential to cause severe complications (58). A recent study found that mortality in critically ill patients with COVID-19 was associated with increased representation of Proteobacteria in the fecal microbiota and decreased concentrations of fecal secondary bile acids (59). These microbiome and metabolite changes appeared to predict the trajectory of respiratory failure and subsequent death in this cohort of patients, with the authors ascribing a critical role to the gut-lung axis in disease prognosis (59). Therefore, beneficial modulation of the gut microbiota could serve as a novel means to treat lung disease. Similarly, with respect to patients with COVID-19, beneficial manipulation of the gut microbiome in the setting of viral infection may serve to limit downstream gut-lung mucosal immune responses and subsequent impairments in MGB axis communication, therefore reducing the incidence of PASC in susceptible individuals. Furthermore, viral infections, including SARS-CoV-2 and flu, were recently found to increase susceptibility to secondary fungal infections, including aspergillosis, via impaired humoral immunity (60). Given that the gut mycobiome can influence the lung mycobiome, as well as lung immunity (61), SARS-CoV-2-induced disruptions may detrimentally impact lung function, although the mechanisms remain to be fully determined.

CONCLUSIONS

We are continuing to learn how infection with SARS-CoV-2 and development of COVID leads to overt acute inflammatory responses that can persist, resulting in PASC in a subset of patients. Additional research into the role of the MGB axis in causing PASC will therefore help better identify those patients most at risk of long-lasting impacts of infection, including behavioral impairments and GI pathophysiology. Using a combination of preclinical models of SARS-CoV-2 infection, coupled with samples collected from infected patients, impairments in MGB axis signaling associated with PASC can be addressed systematically. Identifying pathways by which PASC may be prevented, such as by beneficially modulating the gut microbiome during recovery from acute disease (such as via dietary supplementation with beneficial butyrate-producing microbes, or their products), could have a significant downstream clinical impact in patients at high risk of developing PASC. Additional studies to test this hypothesis are urgently needed.

GRANTS

Studies from the authors’ group related to the subject matter of this article have been supported by National Institutes of Health Award R01AT009365 (to M.G.G.) and by an award to K.E.B. and M.G.G. from the Long COVID Seed Grant program of the UC Davis School of Medicine.

DISCLOSURES

Melanie Gareau is an editor of American Journal of Physiology-Gastrointestinal and Liver Physiology and was not involved and did not have access to information regarding the peer-review process or final disposition of this article. An alternate editor oversaw the peer-review and decision-making process for this article. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

M.G.G. and K.E.B. conceived and designed research; M.G.G. prepared figures; M.G.G. and K.E.B. drafted manuscript; M.G.G. and K.E.B. edited and revised manuscript; M.G.G. and K.E.B. approved final version of manuscript.