Choroid Plexus calcification correlates with cortical microglial activation in humans: a multimodal PET, CT, MRI study; 2022; Butler et al

[SNT Gatchaman]

Published version linked in a later post
*****

Preprint
Choroid Plexus calcification correlates with cortical microglial activation in humans: a multimodal PET, CT, MRI study
Tracy Butler, X. Hugh Wang, Gloria C. Chiang, Yi Li, Liangdong Zhou, Ke Xi, NimmiWickramasuriya, Emily Tanzi, Edward Spector, Ilker Ozsahin, Xiangling Mao, Q. RayRazlighi, Edward K. Fung, Jonathan P. Dyke, Thomas R. Maloney, Ajay Gupta, Ashish Raj, Dikoma C. Shungu, P. David Mozley, Henry Rusinek, Lidia Glodzik

Background
Choroid plexus (CP) within brain ventricles is well known to produce CSF. Additional important CP functions are now recognized including critical modulation of inflammation. Recent MRI studies have demonstrated CP enlargement in human diseases including Multiple Sclerosis and Alzheimers Disease, and in association with neuroinflammation measured using translocator protein (TSPO) PET. The basis of MRI—visible CP enlargement is unknown.

Purpose
Based on tissue studies demonstrating CP calcification as a common pathology associated with aging and disease, we hypothesized that previously—unmeasured calcium within CP contributes to MRI—measured CP volume, and may be more specifically associated with neuroinflammation.

Materials and Methods
We performed a retrospective analysis of PET—CT studies performed between 2013—2019 on a single scanner using the TSPO radiotracer 11C—PK11195. Subjects included controls (n=43) and patients diagnosed with several non—inflammatory neuropsychiatric conditions (n=46.) Cortical inflammation / microglial activation was quantified as nondisplaceable Binding Potential (BPnd.) CP and ventricle volume were measured using Freesurfer. CP calcium was measured semi—manually via tracing of low—dose CT acquired with PET and automatically using a new CT/MRI method. The contribution of CP calcium, CP overall volume, ventricle volume, subject age, sex and diagnosis to BPnd was assessed using linear regression.

Results
89 subjects (mean age 54+/—7 years; 52 men) were included. Fully—automated CP calcium quantification was accurate (ICC with semi—manual tracing = .98.) The significant predictors of cortical neuroinflammation were subject age (p=.002) and CP calcium volume (p=.041), but not ventricle or CP volume.

Conclusion
CP calcium volume can be accurately measured using low—dose CT acquired routinely with PET—CT. CP calcification — but not CP overall volume — was associated with cortical inflammation. Unmeasured CP calcification may be relevant to recent reports of CP enlargement in human inflammatory and other diseases. CP calcification may be a specific and relatively easily—acquired biomarker for neuroinflammation and CP pathology.

MedRxiv

 

[SNT Gatchaman]

Key Results

• Choroid plexus (CP) calcification volume can be reliably quantified using semi-manual tracing on low-dose CT acquired with PET-CT, and fully automatically using our new, accurate (ICC with semi-manual tracing = .98) CT/MRI method.
  
  
• CP calcification and age –but not overall CP volume– significantly predicted 11C-PK11195 PET-measured cortical neuroinflammation in 89 subjects.
  
  
• CP calcification is a relatively easily-assessed, previously-overlooked potential biomarker for neuroinflammation and CP pathology.

This preprint has the potential to be a little jaw-dropping — in a "from the Dept. of It Was Staring Us in the Face All Along" kind of way.

We've discussed the difficulties of diagnosing neuroinflammation, by imaging or other biomarkers. Newer techniques of imaging microglial activation include MRI and PET-CT/MR with TSPO as a radiotracer.

Radiologists have learned to dismiss findings of eg increased perivascular spaces and choroid plexus calcification because they are a) quite commonly seen and b) of unknown pathological significance. CP calcification is very common in the elderly so its presence was usually put in the "accelerated ageing" basket.

Now this study is retrospectively looking at choroid plexus calcification in studies where they are proving (by non-invasive proxy) neuroinflammation and showing a correlation. This would be in the context of chronic neuroinflammation (although that may be the only type) and once established the calcification would presumably persist even if the neuroinflammation resolved, but the volume of calcification might reduce with resolution.

The authors are suggesting this might be a specific biomarker of neuroinflammation and one that is simple to obtain — a low dose non-contrast CT of the head. If this turns out to be valid many ME patients could even have their previous head CTs retrospectively evaluated.

I'll read through the paper itself and summarise some quotes.

 

[SNT Gatchaman]

OK, this is a bit confusing. Their subjects included 3 groups: "CFS" (!!), "Gulf War Syndrome" and Parkinson Disease (non-dementia), which they are characterising as "non-inflammatory neuropsychiatric" vs HCs. They are evaluating cortical microglial activation, which they seem to be differentiating from "inflammation". I may be missing something fundamental in terminology or understanding with microglial activation and neuroinflammation.

A well-cited article in Nature Reviews: Neurology: Neuroinflammation and microglial activation in Alzheimer disease: where do we go from here? (2020) said:

Among the innate immune cells, microglia are the primary players in neuroinflammation.

But maybe microglial activation can occur up to some level before neuroinflammation proper develops??

It's getting a bit late so I'll re-read tomorrow and try and follow down some of their references too. They don't give raw data but it may be in their prior papers. If anyone else has capacity to look through I'd appreciate it. Regardless I think their hypothesis is really good, so replication with 2-4x participants would be good.

TBC

 

[Hutan]

Just in the meantime, some background on the choroid plexus, which forms the blood-CSF barrier:

From Development and functions of the choroid plexus–cerebrospinal fluid system, Nature 2015

The choroid plexus (ChP) is a secretory tissue found in each of the brain ventricles, the main function of which is to produce cerebrospinal fluid (CSF). Although the ChP–CSF system is essential for proper development of the nervous system owing to fluid pressure within the ventricles as well as myriad CSF-borne signalling factors, it is nevertheless one of the most understudied areas of neurobiology.

A highly organized tissue, the ChP consists of simple cuboidal epithelial cells surrounding a core of fenestrated capillaries and connective tissue. As the interface between peripheral circulation and the CNS, the ChP forms the blood–CSF barrier via tight junctions between adjacent epithelial cells to restrict free passage of solutes from blood into CSF, and vice versa.​

The blue is the cerebrospinal-spinal fluid and the red is the blood. There are the endothelial cells forming the capillary linings and choroid plexus epithelial cells lining the ventricular spaces where the CSF is. As I understand it, the choroid plexus are areas where there are a lot of villi, increasing the surface area of the connection between the capillaries and the CSF.


Here's another paper, this one is a 2020 one, and they are still noting how understudied the choroid plexus is:
Choroid plexus and the blood–cerebrospinal fluid barrier in disease, 2020

The central nervous system (CNS) is protected against harmful substances contained in the blood by the blood–brain barrier (BBB) and the blood–cerebrospinal (B-CSF) barrier [1,2,3]. Even though the B-CSF barrier is more accessible than the BBB for many foreign invaders, the BBB has received more attention regarding CNS pathology. Nevertheless, there is a growing body of evidence showing that the B-CSF barrier plays a crucial role in the spread of inflammatory reactions from the periphery to the CNS and contributes to the pathogenesis and progression of various neurological disorders.

If the B-CSF barrier allows blood-borne pathogenic components to enter the CSF, they would contribute to neurotoxicity and neuronal dysfunction. On the other hand, reduction in the efflux of waste products (and the consequent presence of neurotoxic residues inside the brain), could interfere with neurotransmission, and lead to various disorders. The mechanism(s) by which a dysfunctioning B-CSF barrier contributes to pathogenesis is still poorly understood.​

So, the choroid plexus is not the blood-brain barrier, it's the next line of defence. There's the possibility of the choroid plexus letting 'pathogenic components' into the brain, but also it not efficiently removing waste products from the brain.

It does sound like an interesting tissue to look at.

 

[Hutan]

Regarding the calcification, this image might help with understanding where the calcification occurs.

Clinical Imaging of Choroid Plexus in Health and in Brain Disorders: A Mini-Review, 2019



At the bottom of the image is a capillary with red blood. At the top is the blue CSF, and then the choroid plexus epithelial cells. The calcifications (the white blob) occur in the space between. Some calcification is normal in adult brains.

 

[Hutan]

PET frames were motion corrected using MCFLIRT30 within FSL 31. Nondisplaceable Binding Potential (BPnd) images reflecting the concentration of translocator protein (TSPO) expressed by activated microglia, irrespective of tracer delivery/blood flow, were generated from dynamic PET using a multilinear reference tissue model 32 implemented in the freely available software package FireVoxel (https://firevoxel.org.)

Differences in CP calcium volume, CP overall volume and cortical BPnd between the diagnostic groups and between women and men were assessed using ANCOVA, controlling for age.

The 2022 study quantifies the amount of calcification in the CP and the size of the CP and compares them to "cortical BPnd" which is a measure of the concentration of translator protein (TSPO). They say TSPO is expressed by activated microglia - and so I think activated microglia is assumed to be neuroinflammation.

Here are the study groups:

 

[Hutan]

When controlling for age, which differed significantly across diagnostic groups (GWI and CFS subjects were younger than controls and PD subjects) there were no group or sex differences in CP calcium volume, CP overall volume or cortical BPnd (p>.1 for all analyses.)

Association of CP calcium, CP overall volume, ventricle volume, age, sex and diagnosis with cortical BPnd. As shown in Table 2, a multiple regression model (R2 = .257, F(6, 82) = 4.72, p < .001) showed that only age and CP calcium volume were significant predictors of cortical BPnd.

Age, CP calcium volume and cortical BPnd were correlated. Sex and the disease/control groups were not. So, none of the disease groups had more CP calcification or cortical BPnd than might have been expected for the ages of the groups.

They conclude that CP calcium volume is an easy way to assess neuroinflammation.

Limitations:

This study has several significant limitations. It uses TSPO PET to quantify cortical neuroinflammation indexed by activated microglia. However, TSPO PET is expressed not just by microglia but by several cell types in the brain 48 and microglial “activation” is a complex process which cannot be assessed using a probe for a single molecule 49. Furthermore, this study used a first generation TSPO radiotracer with inferior signal properties to newer tracers 28. However, TSPO PET is the only available method to assess microglial activation in vivo in humans, and we apply optimal image processing and analysis methods 34.

They note that TSPO may not be a perfect measure of microglial activation, and the TSPO radio tracer is an older version that isn't as good as more recent ones.

It can be considered a limitation that our subject group was heterogeneous and included subjects with a variety of disorders, including patients with PD in whom inflammation of substantia nigra in midbrain is pathophysiologically important 50. Results were not significant when the subject group was halved to include only normal control subjects. Addressing this limitation, there were no significant differences in CP calcification, CP volume or cortical BPnd between subject groups, we included diagnosis as a covariate in all analyses, and assessment of neuroinflammation was limited to the cortex, which is not considered to be involved in PD without dementia 50 or the other included disorders.

I guess this is the confusing thing that @SNT Gatchaman was talking about. They say that neuroinflammation is a feature of Parkinsons, but they did not find higher levels of the marker of activated microglial cells than would be expected given the older age of the Parkinsons group. They explain that away by saying that the assessment of neuroinflammation was limited to the cortex and that there isn't neuroinflammation in the cortex of Parkinsons(or GWI or CFS). But I think that means that the paper suggests that the CP calcification is only a potential marker of neuroinflammation when it is in the cortex?

So, it doesn't sound as though the CP calcification assessment will be of use to us. If we believe this study, there is neither neuroinflammation in the cortex of people with CFS and no abnormal calcification of the choroid plexus.

The paper is a bit light on detail. There were only 17 CFS subjects and we aren't told how they were diagnosed.

What do others think?

 

[SNT Gatchaman]

Thank you @Hutan for backgrounding and evaluating.

I guess this is the confusing thing that @SNT Gatchaman was talking about. They say that neuroinflammation is a feature of Parkinsons, but they did not find higher levels of the marker of activated microglial cells than would be expected given the older age of the Parkinsons group. They explain that away by saying that the assessment of neuroinflammation was limited to the cortex and that there isn't neuroinflammation in the cortex of Parkinsons(or GWI or CFS). But I think that means that the paper suggests that the CP calcification is only a marker of neuroinflammation when it is in the cortex?

Maybe it was because they were comparing the PET study, which only looked at the cortex (??): as opposed to the central grey matter. In Parkinson's they would expect neuroinflammation centrally (affecting movement), but not in the cortex in the absence of clinical evidence of dementia. It reads as if their entire participant cohort was categorised as "normal-in-terms-of-cortical-neuroinflammation", but they could closely correlate choroid plexus calcification volume with cortical microglial activity (which they considered within the range of normal?).

The key thing for us would be what do they mean by "CFS". If it's a loose diagnosis where patients might have fatigue-NOS and possibly no neuro symptoms this study will be unhelpful. But... there is potential for a similar evaluation in a well characterised group. (I wonder whether the BioBank studies that looked at pre- / post-Covid MRIs could do something with this).

I'll post some more quotes that caught my eye below.

 

[SNT Gatchaman]

The choroid plexus (CP), a highly vascularized filamentous structure within the brain ventricular system, constitutes the blood-CSF barrier and is well known to produce CSF.

Recently, additional, important CP functions have been recognized including clearance of toxins, sleep-wake regulation, secretion of hormones and growth factors, support of neurogenesis, and regulation of inflammation.

The role of CP in inflammation is especially critical: CP both initiates and modulates neuroinflammation, and CP dysfunction is considered highly relevant to the pathophysiology of multiple sclerosis.

Unlike virtually all brain parenchymal structures, which atrophy in association with aging and dysfunction, CP enlargement appears to be pathological. [...] The basis of CP enlargement remains uncertain, but based on autopsy and animal studies, has been posited to relate to CP basement membrane thickening and fibrous stroma expansion with protein, inflammatory cells and deposition of metal, in particular calcium.

For decades, CP calcification has been recognized as a normal phenomenon, with gradually increasing calcification noted with increasing age.

To our knowledge, decades-old recognition of increasing CP calcification with aging and certain diseases has not been linked to more recent understanding that CP has important roles beyond CSF production, in particular, critical regulation of neuroinflammation.

These results shed light on the basis of pathological CP enlargement increasingly recognized in association with human inflammatory and neurodegenerative diseases, and suggest the potential for CP calcification to serve as a sensitive, specific and relatively easily-acquired biomarker for both neuroinflammation and CP dysfunction.

The novel result of this study is that CP calcification is independently associated with cortical inflammation when accounting for age, ventricle volume and CP volume, suggesting a possible unique contribution of CP calcification to inflammation.

This possibility is supported by recent findings in an animal model of familial brain calcinosis (commonly referred to as Fahr’s disease in humans) directly linking brain calcification to activated microglia. In this study, activated microglia were beneficial in regulating calcification and preventing neurodegeneration. Current findings showing an association between activated microglia and calcification in subjects free from inflammatory disease are broadly consistent with such a homeostatic interplay between microglia and calcification.

Calcium in CP has recently been shown to be uniquely dynamic

Calcium in CP is found mainly in structures called psammoma bodies, consisting also of collagen, phosphorus and iron. Psammoma bodies are present only in CP and in certain tumors, and are considered a hallmark of CP aging and dysfunction.

 

[SNT Gatchaman]

I'm not sure about their comments re Fahr's disease. As you can see in the linked page there is substantial brain parenchymal calcification, so not sure how this relates to the important CP neuroinflammatory modulation functions they initially described.

We need to see their raw data. It's possible that they are showing pathological microglial activation, but which they think is in normal range, occurring in the absence of neuroinflammation (eg controlling it) when actually their framing of CFS and GWI +/- PD as non-inflammatory is incorrect to start with.

Perhaps the choroid plexus is another mechanism whereby body inflammation can be mirrored as neuroinflammation (cf VanElzakker's vagus inflammation sensing). In this case across the blood-CSF barrier rather than blood-brain barrrer.

I hope the paper's reviewers follow these points up.

 

[SNT Gatchaman]

Now published in AJNR as —

Choroid Plexus Calcification Correlates with Cortical Microglial Activation in Humans: A Multimodal PET, CT, MRI Study
T. Butler; X.H. Wang; G.C. Chiang; Y. Li; L. Zhou; K. Xi; N. Wickramasuriya; E. Tanzi; E. Spector; I. Ozsahin; X. Mao; Q.R. Razlighi; E.K. Fung; J.P. Dyke; T. Maloney; A. Gupta; A. Raj; D.C. Shungu; P.D. Mozley; H. Rusinek; L. Glodzik

BACKGROUND AND PURPOSE:
The choroid plexus (CP) within the brain ventricles is well-known to produce cerebrospinal fluid (CSF). Recently, the CP has been recognized as critical in modulating inflammation. MRI–measured CP enlargement has been reported in neuroinflammatory disorders like MS as well as with aging and neurodegeneration. The basis of MRI–measured CP enlargement is unknown. On the basis of tissue studies demonstrating CP calcification as a common pathology associated with aging and disease, we hypothesized that previously unmeasured CP calcification contributes to MRI–measured CP volume and may be more specifically associated with neuroinflammation.

MATERIALS AND METHODS:
We analyzed 60 subjects (43 healthy controls and 17 subjects with Parkinson’s disease) who underwent PET/CT using 11C-PK11195, a radiotracer sensitive to the translocator protein expressed by activated microglia. Cortical inflammation was quantified as nondisplaceable binding potential. Choroid plexus calcium was measured via manual tracing on low-dose CT acquired with PET and automatically using a new CT/MRI method. Linear regression assessed the contribution of choroid plexus calcium, age, diagnosis, sex, overall volume of the choroid plexus, and ventricle volume to cortical inflammation.

RESULTS:
Fully automated choroid plexus calcium quantification was accurate (intraclass correlation coefficient with manual tracing = .98). Subject age and choroid plexus calcium were the only significant predictors of neuroinflammation.

CONCLUSIONS:
Choroid plexus calcification can be accurately and automatically quantified using low-dose CT and MRI. Choroid plexus calcification—but not choroid plexus volume—predicted cortical inflammation. Previously unmeasured choroid plexus calcium may explain recent reports of choroid plexus enlargement in human inflammatory and other diseases. Choroid plexus calcification may be a specific and relatively easily acquired biomarker for neuroinflammation and choroid plexus pathology in humans.


Link | PDF (American Journal of Neuroradiology)