Understanding long COVID—The role of serotonin in cognitive impairment

A recent study by Wong et al. was published in the journal “Cell” and illuminates a potential role for serotonin reduction in mediating cognitive impairment following postacute sequelae of COVID (PASC) or Long COVID.¹ This research highlight explores the mechanisms underlying viral infection-mediated serotonin reduction, and unveils therapeutic targets which could alter the way we approach Long COVID in clinical practice.

In the aftermath of the COVID-19 pandemic, patients are increasingly presenting with debilitating symptoms persisting for months after acute SARS-CoV-2 infection.² The constellation of symptoms, collectively termed Long COVID, are heterogeneous and involves multiple body systems. Despite the significant impact of Long COVID on healthcare resources and patients' quality of life, the mechanisms underlying these symptoms remain largely enigmatic. However, in a groundbreaking study, Wong et al. illuminate a potential link between viral-induced inflammation, serotonin reduction, and cognitive deficits in individuals suffering from Long COVID.¹

Wong et al. analysed a range of metabolites in the serum of patients exhibiting symptoms of both acute and post-acute COVID-19 infection, and found amino acid metabolites, such as serotonin were depleted throughout the acute and chronic phases of infection. Interestingly, patients with Long COVID had lower serotonin levels compared to those who fully recovered from their initial infection. Indeed, serotonin levels in the bloodstream were predictive of long-term symptom burden after initial COVID infection, and strongly suggests a potential role for serotonin in the pathophysiology of Long COVID. Similar reductions in serotonin levels were found in other infections such as varicella-zoster virus and lymphocytic choriomeningitis virus, suggesting that reduced serotonin may be a shared characteristic of systemic viral infections.

The study then turned to mouse models of viral infection to characterise the mechanisms underpinning serotonin reduction, and found increased type 1 interferon (IFN) signalling. Importantly, IFN signalling was persistently upregulated in long COVID, and inhibition of the IFN alpha receptor prevented viral infection-induced serotonin depletion. This effect was abolished in mice with impaired IFN signalling.

Serotonin is predominantly synthesized in the gastrointestinal tract, where it is produced from an essential dietary amino acid called tryptophan.³ This study also found that individuals with acute and persistent COVID-19 infection had reduced plasma tryptophan levels, hinting at a potential limitation in serotonin production during viral infections. This suggests individuals with congenital or acquired tryptophan deficiency may be more susceptible to developing long COVID. RNA sequencing of small intestinal epithelium found that genes involved in amino acid absorption were significantly downregulated following viral infection. Interestingly, promoting alternative pathways for serotonin production prevented reduction of serotonin following COVID infection.

The study also evaluated serotonin storage in platelets, and found that platelet counts were substantially decreased following viral infections. Indeed, chemically induced thrombocytopaenia in mouse models using polyinosinic:polycytidylic acid (poly(I:C)) injections diminished serotonin levels in plasma and in isolated platelets. Thrombocytopenia was found to be dependent on the implicated IFN signalling pathway, highlighting the role of this pathway in serotonin reduction.

Cognitive impairment is a prominent feature of Long COVID, often characterized by symptoms such as difficulty concentrating, problems with attention, and memory deficits.⁴ This study demonstrated memory deficit in mouse models following exposure to various viral infections. This deficit was dependent on IFN signalling and platelet depletion. These findings indicate a role for serotonin reduction in the cognitive impairment associated with Long COVID. Remarkably, treating mice with fluoxetine (a selective serotonin reuptake inhibitor [SSRI]) and tryptophan supplementation were effective in restoring normal cognitive performance. This highlights a potential therapeutic target in the treatment of Long COVID-related cognitive impairment.

Interestingly, despite peripheral reduction of serotonin, brain serotonin levels were unaffected following SARS-COV-2 infection. This raised the possibility that peripheral serotonin influences the brain through afferent sensory neurons, perhaps the vagus nerve. Indeed, activation of vagal neurons restored hippocampal neuron activation and memory formation in poly(I:C)-treated mice. In vitro experiments with vagal neurons cultured from nodose ganglia also showed a robust response to serotonin treatment. The authors also found selective expression of the ionotropic serotonin receptor 5-HT3 along the vagus nerve. Pharmacological activation of the 5-HT3 receptor normalized hippocampal neuron responses and cognitive performance during viral inflammation. These findings ultimately suggest that serotonin reduction impairs cognitive function by dampening vagal signalling.

This study outlines a mechanism on how serotonin levels might be altered during viral infections, and its impact on brain function (Figure 1). Viral-induced inflammation, may drive the depletion of serotonin through reduced synthesis due to downregulated tryptophan absorption, disrupted serotonin storage through a low platelet count, and also increased monoamine oxidase (MAO) expression. Serotonin depletion then leads to consequences such as cognitive impairment which may be mediated by decreased vagal and hippocampal activation.

The implications of these findings are far-reaching. First and foremost, the study sheds light on the profound effects of persistent viral reservoirs. There is growing evidence of persistent viral components and sustained high levels of IFN in the blood long after the acute phase of infection.⁵ This study indicates that these viral remnants and the resulting IFN response could be directly attributed to the symptomatic profile found in long COVID. Furthermore, the authors highlight a dysregulated amino acid uptake as a sequelae of viral infection, potentially resulting in nutrient deficiencies. Further studies are required to investigate levels of melatonin and vitamin B3, of which tryptophan is a precursor, in long COVID. Moreover, the study points to hypercoagulability and the formation of microthrombi as common features of both acute and post-acute SARS-CoV-2 infection, which has been suggested previously.⁶ The study also highlights the link between serotonin depletion and cognitive impairment, consistent with neurological symptoms reported by long COVID patients.⁴

Finally, the study outlines potential preventative and therapeutic interventions in the clinical management of long COVID. The mouse model used, poly(I:C) treatment, was an effective way to diminish peripheral serotonin levels where they showed impaired memory. Evidence with animal models in this study demonstrates that serotonin levels can be replenished, and cognitive function can be restored through precursor supplementation or the use of selective serotonin reuptake inhibitors (SSRIs). Furthermore, a recent randomised-controlled trial demonstrated that replenishing gut microbiota reduced the incidence of several long COVID symptoms, including fatigue and cognitive impairment.⁷ Gut microbes can induce vagal nerve signalling and produce several metabolites including tryptophan and serotonin. Whilst the role of probiotics in improving memory is previously established, this study suggests a role for targeted microbiome therapy for cognitive impairment in the context of long COVID. Ultimately, this study, along with recent findings linking mood disorders with memory impairment in long COVID, calls for a comprehensive assessment of serotonin signalling as a potential therapeutic target for neurocognitive manifestations of PASC.⁸

Shuvam Sarkar: Conceptualization (lead); investigation (lead); writing—original draft (lead); writing—review and editing (equal). Olivia Monteiro: Funding acquisition (lead); resources (lead); visualization (lead); writing—review and editing (equal). Both authors have read and approved the final manuscript.

The authors declare no conflict of interest.

The authors have nothing to report.