Glial activation among individuals with Neurological [PASC]: A Brain Fog Positron Emission Tomography Study of [18F]-FEPPA, 2025, Clouston+

[Nightsong]

Abstract
Background: This study examined the regional distribution of glial activation in essential workers with neurological post-acute sequelae of coronavirus disease 2019 (COVID-19) infections (N-PASC).

Methods: We injected ≤185 MBq of [18F]-FEPPA as an intravenous bolus and positron-emission tomography over two hours. To measure distribution volume (VT) we acquired an arterial input function in 24 essential workers (14 N-PASC, 10 Never-COVID-19 Controls, of whom eight successfully placed arterial lines). Individuals with low binding affinity were excluded from this study, and VT was adjusted for translocator protein genotype. Analyses that passed the false discovery rate are reported.

Results: Participants at midlife survived mild to moderate COVID-19 without hospitalization but reported onset of post-acute sequelae for, on average, 22 months before undergoing neuroimaging. Hippocampal VT was higher (VT=1.70, 95% C.I. = [1.30-2.21], p=0.001) in participants with persistent brain fog after COVID-19, reflecting an increase of 10.58 ml/cm3 in VT (area under the receiver-operating curve, AUC= 0.95 [0.85-1.00]). At a cutoff of 10.6, sensitivity/specificity/accuracy were 0.88/0.93/0.91.

Conclusion: The results from this study imply that neuroimmune response is a distinct and identifiable characteristic of brain fog after COVID-19. Results suggest that [18F]-FEPPA could be used to support PASC diagnosis.

Link | PDF

 

[forestglip]

As far as I can tell, this is only comparing a group of people who had COVID and continue to have brain fog many months later, with a group who got scans before the pandemic. Seems premature to connect the difference in scans to brain fog specifically, but maybe there's some justification in the paper.

 

[Hutan]

Changes in translocator protein (TSPO) expression result from glial or macrophagic hyperactivity [15] and that cerebral gliosis can therefore be measured using [18F]-FEPPA with positron emission tomography (PET). The purpose of this study was to use [18F]-FEPPA to examine the severity and distribution of widespread neuroinflammation in the cortex, white matter, or cerebellum, and to determine if cerebral gliosis was more evident in points of entry or vulnerability including in the hippocampus or in the orbitofrontal and medial frontal cortices in N-PASC.

The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines [16].

Could be worth checking out those STROBE guidelines some time.

Participants are essential workers who were recruited from September 2022 – June 2023 from a clinic-based monitoring program that has been collecting data for the past two decades on workers living in Long Island, NY [17]. This program was proactive in recording the onset of COVID-19 in NY [18], has linked onset and severity to polygenic and clinical risk factors [19], and has previously described the development of PASC symptom at midlife in non-hospitalized essential workers [5]. Due to the occupational inclusion criteria this cohort is predominantly male, and most participants are under the age of 70.

N-PASC cohort: Non-hospitalised Covid-19 infection; lasting neurological symptoms including brain fog for at least 6 months and still present at teh time of scanning.
Controls: people from the same monitoring program, prior to the pandemic, healthy at the time of scanning.

Just by the way, on the use of TSPO scanning:

Because [18F]-FEPPA binds poorly among individuals who have a genotype indicating low affinity for TSPO polymorphism, we excluded individuals whose genotyping indicated this affinity [20]. TSPO genotype was based on the rs6971 polymorphism and was measured using blood samples collected at screening. Following recommended procedures 21, 22, all subjects were genetically profiled and classified as high-affinity (HABs, Ala147/Ala147), and mixed-affinity (MABs, Ala147/Thr147). Low-affinity binders (LABs, Thr147/Thr147) were excluded from further study.

 

[Hutan]

The outcome of this study was the cerebral distribution volume for Fluorine F 18–labeled N-(2-(2-fluoroethoxy)benzyl)-N-(4-phenoxypyridin-3-yl)acetamide ([18F]-FEPPA) as measured using dynamic PET. [18F]-FEPPA is a radioactive ligand that was designed to bind to mitochondrial translocator protein 18-kDa (TSPO), which is most abundantly expressed in M2-activated macrophages and microglia [23], making [18F]-FEPPA a particularly valuable radioligand for the study of neuroinflammation in N-PASC.

So, a radioactive molecule that binds with TSPO, which is a marker of macrophage and microglia activation. The scanning then allowed 3D modelling of where in the brain the radioactive molecule showed up.

To enable kinetic modeling, an automated blood sampling system (Twilite; Swisstrace, Zurich, Switzerland) was used to continuously measure arterial blood radioactivity for the first six minutes of the scan, and manual arterial blood samples were collected at 2.5/7/15/30/60/90/120-minute time-points following injection. Tracer and metabolite levels were measured using high-performance liquid chromatography analysis for manual blood draws [25].

The researchers also took samples of arterial blood to measure the uptake of the radioactive molecule. I'm not yet sure why they did that. I assume that if there was microglial activation, the radioactive molecule would have been pulled out of the blood quicker and bound to the TSPO. Maybe there was some adjustment based on the levels of the radioactive molecule in the blood. Anyway.

 

[Hutan]

Although the abstract says 14, there were only 10 PASC participants.

To measure distribution volume (VT) we acquired an arterial input function in 24 essential workers (14 N-PASC, 10 Never-COVID-19 Controls, of whom eight successfully placed arterial lines).

In total, ten essential workers with N-PASC went through the [18F]-FEPPA imaging study.

Edit: I think there is an error in the abstract. Table 1 suggests that there were 10 PASC participants and 14 controls.

In this table, we found that the pre-pandemic group had lower BMI and appeared younger than the N-PASC group. The N-PASC group had much higher post-COVID-19 depressive symptoms. However, there were no differences in terms of TSPO genotype, premorbid cognitive ability, or sex. We recorded acute symptoms to help describe the sample, but since no symptoms were associated with any measure of neuroinflammation we did not report these results. However, arterial line placement was unsuccessful for two participants, resulting in VTbeing available for eight participants.

They then note that they weren't able to place an arterial line in 2 of the 10 people, so they only got scanning results for 8 people, but reported the demographic characteristics for 10 people.

A study of 8 PASC people, for whom we don't know the demographic characteristics compared to the healthy controls is a bit thin. The 10 PASC people were already older on average compared to the healthy controls (61 years compared to 56).

Each of the figures (Fig. 1A-E) show large, statistically significant, differences between N-PASC and never-COVID-19 controls.

There do seem to be real differences between the two groups in the volume of brain found to express the radioactive ligand though, significantly higher in the PASC group.

 

[Hutan]

Here's Figure 2. Top two rows are the PASC group; bottom two are the controls. Red indicates more radioactive ligand binding to the TSPO. As I said, there look to be real differences that differentiate the two groups but I've only skimmed the paper quickly and haven't understood all of it. I think it's worth a closer look, and following the research group's next move.




It would be good to understand better what the symptoms of the PASC group are, to assess if they match with ME/CFS. They do some assessment of correlations between marker binding and cognitive symptoms within groups, but I don't think they found much. In fact they seemed to find that the people with more putative neuroinflammation were thinking better.

 

[Hutan]

From the discussion:

They suggest others have found evidence of what they describe as neuroinflammation (glial activation)

Results from this study add to a growing literature describing neuroinflammation as a central aspect of Neurological PASC (N-PASC). Together with other studies that also support the view that neuroimaging and TSPO imaging can discriminate N-PASC effectively, this study suggests that N-PASC may be verified using neuroimaging-focused biomarkers including [18F]-FEPPA distribution volume.

These analyses revealed that, when compared to healthy controls, hippocampal [18F]-FEPPA distribution volume was able to identify individuals with N-PASC with sensitivity/specificity/F1 at the estimated cutoff of 0.88/0.93/0.88. If replicated, this degree of accuracy could support the use of [18F]-FEPPA to positively identify a diagnosis of N-PASC.

They tried hard to think of explanations for the finding that more TSPO was related to better cognitive performance in the PASC group, suggesting that the glial activation is adaptive. But I think, when they are working with data from 8 people, hypothesising is a bit premature.

The second reason for our findings could be fact that the TSPO protein, even if expressed by the microglia, seems to have different functions. In fact, meanwhile TSPO is clearly associated with pro-inflammation activated microglia cells in mice, recent studies suggest that while TSPO is an accurate indicator of microglial cell density that it may not be a clear indicator of pro-inflammatory activation [48]. This strengthens the possible view that the microglial cells, whose presence we found to be increased, might protect verbal fluency in N-PASC. Future work should determine whether monocytic response mechanisms and endothelial cell functioning, which uphold macrophagic reactivity, are also active and adaptive in N-PASC.

They seem to be questioning if the TSPO binding really does indicate inflammation....

 

[EndME]

From what I recall there was also a study by Hellemonds that tagged TSPO (the molecule used was a different one). They published a preprint that made it all over the news with 2 LC patients, but from what I've heard the full study actually had very mixed results (and hasn't been published yet even though it's been a few years). There have also been several other TSPO tagging studies in LC.

Edit: This is the study I meant (Hellemonds wasn't actually part of it as it was other dutch researchers).