Mapping microglia and astrocyte activation in vivo using diffusion MRI, 2022, Garcia-Hernandez et al

Mapping microglia and astrocyte activation in vivo using diffusion MRI
Raquel Garcia-Hernandez, Antonio Cerdán Cerdá, Alejandro Trouve Carpena, Mark Drakesmith, Kristin Koller, Derek K Jones, Santiago Canals, Silvia De Santis

While glia are increasingly implicated in the pathophysiology of psychiatric and neurodegenerative disorders, available methods for imaging these cells in vivo involve either invasive procedures or positron emission tomography radiotracers, which afford low resolution and specificity.

Here, we present a noninvasive diffusion-weighted magnetic resonance imaging (MRI) method to image changes in glia morphology. Using rat models of neuroinflammation, degeneration, and demyelination, we demonstrate that diffusion-weighted MRI carries a fingerprint of microglia and astrocyte activation and that specific signatures from each population can be quantified noninvasively.

The method is sensitive to changes in glia morphology and proliferation, providing a quantitative account of neuroinflammation, regardless of the existence of a concomitant neuronal loss or demyelinating injury. We prove the translational value of the approach showing significant associations between MRI and histological microglia markers in humans. This framework holds the potential to transform basic and clinical research by clarifying the role of inflammation in health and disease.

Link | PDF | Supplementary Materials

 

Induced neuroinflammation in rats vs healthy controls. "Phase 1" style evaluation in healthy humans.

 

Yes. We're all crying out for a biomarker and the expectation has been that it would be found in the form of an immune cell/mediator signal or something similar, obtainable from blood. I have found it ironic that my own specialty has not helped me in the slightest (I only ever had a chest radiograph through this). However, I had always hoped that neuroimaging might advance to help us all. We could never biopsy the brain, but perhaps this technique will be a step away from "all tests normal". Also, sounds potentially useful for many conditions.

Tantalising to think this could end up being an equivalent to MS patients having their demyelinating plaques shown on MRI (or CT), to facilitate diagnosis and support disease surveillance.

I need to read the paper through in depth, and would like to see some pretty pictures in (affected) humans soon. This sort of technique would offer potential to see more directly into a hidden world that has only previously been inferred. Much work has been done to suggest that neuroinflammation, persistent glial hyperactivation etc are a big part of the symptomatology. And of course this would underscore the validity of the E in the original ME, which for many years has been so widely disputed by naysayers that this can not possibly be encephalitis/encephalomyelitis.

One of the aspects to note is the suggestion that this technique could relatively easily translate onto the fleet of in-service clinical MRIs, which are generally 3T strength. Previously, it might have been thought that research magnets at 7T or beyond would be needed for resolution (eg vanElzakker has a study on brainstem at high-field strength). They won't even approach mainstream for a decade or two.

Some in long Covid groups are commenting on the similarity to their past experiences post concussion. Perhaps that reflects commonalities neuroinflammation +/- BBB failure, although not post-traumatic in our situation.

 

Unfortunately it wouldn't be a question of changing some presets on our current machines and pushing the big blue start button. The manufacturers would need to translate, evaluate for accuracy and safety and then produce an add-on package. Even if it was ready and validated in humans from a research perspective it would be 2-3 years minimum, as medical device certification is molasses for mostly very good reasons. Also it would be expensive as an add-on, so won't be coming to our hospital magnets as a diagnostic soon, without a champion pushing for it. I can't see neurologists thinking of FND/ME patients first up in this scenario!

I guess if there is merit for diagnosing large numbers of LC patients - then who knows? - perhaps it could be warp-speeded into deployment. But for now I think the promise is with our researchers who might be able to power up their neuroinflammation studies. The targeted tracer PET imaging that we currently have suffers from very poor spatial resolution.

ETA: DWI is pretty low spatial res itself but can map onto structural sequences. All of that is without the radiation penalty of a PET tracer and the accompanying CT, so much easier to gain ethics for human subjects.

 

I've had a read through now and I need to reverse some of my earlier off-the-cuff comment. I've added a link to the supplementary materials in the first post. As far as I can tell from their methods, you probably could acquire the data on any clinical machine right now — so that aspect is better than I thought earlier. This is at high res, but I'll check with one of my techs whether it's feasible with typical setups.

Diffusion-weighted imaging is a basic MRI technique that looks at impairment to the free movement of water molecules (Brownian motion) to make inferences about anatomy and pathology. Eg in ischaemic necrosis (such as stroke), dying cells are swollen as their Na/K pumps fail and the increased intracellular water (which was moving freely in the extracellular space before) can't move around so much now. Once the cell walls' integrity is lost, that reverses and diffusion restriction becomes less, often seen 24-48 hours after stroke onset.

The gold though is in the model they have developed, which allows them to separate out the cell compartments and pathological processes. This looks like it's a custom MATLAB model they've developed, which is able to discriminate neuroinflammation in the presence or absence of demyelination and neuronal loss. The animal studies leading to this model are quite elegant (if you're not the rat).

The multi-compartment model they've developed then relates back to postmortem studies that define glial geography as it varies through the human brain. The result is a model of "spheres" and "sticks", such that small sphere size, stick fraction and stick dispersion correlate with microglial morphology. This is quite mind-blowing as an advance.

As they comment, traditionally we often report regions of non-specific "damage". We'd like to say more but often are only proposing possibilities. The authors appear to be offering the ability to define which cell compartments are involved in activation — microglia or astrocyte — regardless of the associated damage in terms of demyelination or neuronal loss. This would really help in defining pathology.

I expect the model would remain proprietary (pretty valuable IP I'd imagine). Not sure if the general concept could be re-engineered past the gaps in their description. I'll run it past a few people that understand these things better than me. It's been 25 years since I studied MRI physics and I only do routine brain MRI.

Give me a little time to get some feedback, but at this stage I'm not sure anyone else can progress with what we have available. So we're probably behind where we would need to be to do a trial in ME or MS, Parkinson's etc.

 

It certainly does sound like a major boost to 'seeing' what's going on in brains. Even with its limitations, some 'before, during and after PEM' scans might reveal something useful.