The PGE2 hypothesis generates a number of testable predictions. If persistent PGE2 upregulation and altered EP-receptor expression contribute to Long COVID, these changes should be detectable and should correspond to specific clinical and biological patterns. Importantly, these predictions can also produce findings that challenge or refute parts of the hypothesis.

The following approaches could help determine whether the proposed mechanism is present in Long COVID and whether EP-receptor profiles are associated with distinct patient subgroups.

→ Assessing and comparing PGE2 levels 
→ Comparing EP-receptor expression
→ Comparing mitochondrial EP3 densities
→ Relating EP-receptor profiles to symptom clusters

Testable predictions

• PGE2 / stable PGE2 metabolites
• EP1–EP4 expression
• EP receptor signatures in PBMCs
• Ratio of different th-cells or monocyte subsets
• relationships with cytokine clusters
• relationships with GPCR autoantibodies
• changes following PGE2-targeted interventions

Assessing and comparing PGE2 levels 

Establishing elevated, stable prostaglandin E2 Major urinary metabolite (PGE2-M/PGE2-MUM) levels could be an initial step. Groups of LC patients could be compared with age- and sex-matched control groups, as basal PGE2 levels differ between the sexes. PGE2 levels are lower in young females compared to young males, increase in females, and decline in males with aging ¹.

Comparing EP receptor expression

We have noted that high levels of PGE2 can downregulate EP3 expression and upregulate EP1, EP2, and EP4 expression (see section 3.1, Table 1). Within-patient samples could be compared, biopsies ¹ taken before before infection and after LC diagnosis to evaluate changes in EP receptor expression. Our hypothesis is that two categories will emerge: one, “EP3 up” (Category A), will have significantly higher EP3 expression than the other, “EP3 down” (Category B), before and after infection. To verify the presence of “EP3 up” or “EP3 down” profiles, and possibly other unbalanced EP expression profiles in LC, the receptor densities of EP1-EP4 could be assessed in age- and sex-matched patient and healthy control groups. Matching is important, since EP3 receptor density decreases with aging has been reported ¹, and our hypothesis presumes sex-based differences in EP receptor expression. Important to note: if the regulatory effects of chronic PGE2 exposure on EP receptor expression are applicable in this context, it is conceivable that most individuals, in Category A as well, would be expected to display reduced EP3 expression, with EP1 upregulation potentially present across both cohorts.

Comparing mitochondrial EP3 expression

Mitochondrial EP3 density in long COVID (LC) patients is also presumed to be lower compared to controls and pre-infection biopsies. We base this assumption on the following data: Upregulated PGE2 has been reported to impair mitochondrial function ^2,3. PGE2 signaling through EP2 and EP4 has been reported to play a role in disrupting mitochondrial function: Morotti et al. identified the PGE2 – EP2/EP4 signaling axis as a key mechanism underlying mitochondrial depolarization ⁴ and, as mentioned, high levels of PGE2 have been shown to downregulate EP3 expression on plasma membranes ⁴ Downregulation of EP3 could be a mechanism by which upregulated PGE2 causes mitochondrial dysfunction and, as a result, chronic fatigue – the signature symptom of long COVID.

Density of all four EP receptors should match specific symptom clusters

Our hypothesis suggests that the expression pattern of all four EP receptors (EP1, EP2, EP3, and EP4) in LC patients should match specific symptom clusters experienced by these patients. According to published data, these symptom clusters could be linked to specific distributions of these receptors. This could be investigated. For further details, see Appendix A, Multimedia component 3, section 1.

1              Liu, S. Z. et al. Prostaglandin E2/cyclooxygenase pathway in human skeletal muscle: influence of muscle fiber type and age. J Appl Physiol (1985) 120, 546-551 (2016). https://doi.org/10.1152/japplphysiol.00396.2015

2              Chen, J. et al. Age-induced prostaglandin E2 impairs mitochondrial fitness and increases mortality to influenza infection. Nature Communications 13, 6759 (2022). https://doi.org/10.1038/s41467-022-34593-y

3              Rashid, M. A. et al. The selective cyclooxygenase-2 inhibitor NS398 ameliorates cisplatin-induced impairments in mitochondrial and cognitive function. Frontiers in Molecular Neuroscience 16 (2023). https://doi.org/10.3389/fnmol.2023.1295991

4              Morotti, M. et al. PGE(2) inhibits TIL expansion by disrupting IL-2 signalling and mitochondrial function. Nature 629, 426-434 (2024). https://doi.org/10.1038/s41586-024-07352-w