Does TSPO PET Measure Microglia Activation or Density?

Researchers rely on TSPO PET imaging to evaluate neuroinflammation in living people. But what does this signal actually measure? In a May 11 preprint on bioRxiv, researchers led by David Owen at Imperial College London and Sandra Amor at Amsterdam UMC present evidence that human microglia, unlike their mouse counterparts, do not turn up TSPO expression in response to inflammatory stimuli. Why, then, does the TSPO PET signal rise in neurodegenerative disease? In postmortem brain sections from people with amyotrophic lateral sclerosis or multiple sclerosis, the researchers found more microglia than in control brain. The higher TSPO signal may indicate microglial density, not activation as it does in mice, they proposed. The data highlight that mouse models do not always reflect human biology. The paper is under review.

  • Unlike in mouse, microglia from people do not increase TSPO expression in response to inflammation.
  • In postmortem AD, ALS, and MS brains, TSPO expression per microglia remains the same as, or lower than, in control brain.
  • Instead, microglial numbers are greater, suggesting TSPO might serve as a marker of their density rather than their state.

"Researchers will have to stop referring to TSPO PET as a marker of inflammatory activity in human disease,” David Brooks at Imperial College London wrote to Alzforum. For his part, Tharick Pascoal at the University of Pittsburgh wrote, “The results presented in this preprint refine the interpretation of TSPO PET findings.” Pascoal believes increases in microglial density could still reflect inflammation.

Others agreed the new data may spur the field to rethink what TSPO imaging means. “The source of the increase in TSPO PET signal has never been truly elucidated,” noted Hervé Boutin at the University of Manchester, U.K. (full comments below).

TSPO PET imaging has been used for 20 years, but has remained strangely vague. Early tracers were crude, not allowing the signal to be quantified. Later, researchers discovered that about 10 percent of people with Caucasian or African heritage did not bind the tracer. As a graduate student, Owen identified the cause, a A147T variant in TSPO that is absent in mice (Owen et al., 2012). More recently, Owen found another mouse/human difference, reporting that TSPO expression did not rise in activated primary human microglia as it does in mouse microglia (Owen et al., 2017). 

To follow up on this, joint first authors Erik Nutma at Amsterdam UMC and Nurun Fancy and Maria Weinert at Imperial College London combed through the literature on gene expression in cultured mouse and human microglia and macrophages. In a meta-analysis of 10 mouse and 42 human datasets, in vitro mouse myeloid cells consistently boosted TSPO in response to inflammatory stimuli such as lipopolysaccharide or interferon-γ, while human myeloid cells either maintained normal TSPO expression or dampened it.

Does the same thing happen in vivo? Nutma and colleagues examined postmortem brain or spinal cord sections from people who had Alzheimer’s disease, ALS, or MS, comparing these to postmortem tissue from mouse models of the conditions. They assessed the amount of TSPO per microglial cell by immunohistochemistry as well as by mass cytometry, where labeled cells are sent through a mass spectrometer.

Overall, the data replicated the in vitro findings. In microglia from APP knock-in, P301S, SOD1, and experimental autoimmune encephalomyelitis (EAE) mice, TSPO expression was up about threefold compared to controls. By contrast, in the AD hippocampus, ALS spinal cord, and MS white-matter lesions, there was no difference in TSPO expression per microglia compared to healthy control tissue. Instead, microglial density shot up threefold in ALS sections and 14-fold in MS lesions. Curiously, the AD hippocampus showed no difference in microglia density. Owen noted that prior studies have reported mixed results for TSPO imaging in this region of AD brain, with some showing an increase in tracer binding and others not (Venneti et al., 2008; Gul et al., 2020). 

Why do mouse and human microglia respond differently to inflammatory insults? Looking for clues, the authors compared the TSPO promoter region in rodents and primates, as well as in other mammals such as sheep, pigs, and rabbits. In mice, rats, and Chinese hamsters, but not any other mammals, they found a binding site for the transcription factor activator protein-1 (AP1). AP1 switches on in response to infection, and activates pro-inflammatory genes. Thus, in mice and rats, but not people, TSPO is part of the inflammatory response. The authors bolstered this conclusion with an analysis of publicly available RNA-Seq data. In mice and rats, TSPO expression was linked to pro-inflammatory networks, but in people, it was linked to cellular bioenergetics. The results suggest TSPO fulfills a different role in primates than in rodents.

Other researchers called this a key finding. “This paper raises the important point that TSPO is actually a byproduct of microglia activation and/or proliferation, and reflects, in fact, an increase in metabolic activity and not a pro- or anti-inflammatory phenotype,” Boutin noted. Brooks agreed that TSPO has nuanced roles. “My view has always been that TSPO expression by microglia is nonspecific and could reflect activation of phagic or protective microglial phenotypes as well as toxic inflammatory activity,” Brooks wrote.

Nonetheless, more work is needed to confirm that TSPO does not rise with inflammation in human cells. Commenters suggested quantifying the amount of TSPO per cell more precisely by using techniques such as single-cell RNA-Seq, single-cell proteomics, or spatial transcriptomics, as well as autoradiography using the TSPO tracer.

Owen is doing some of that now. He noted that if the findings hold up and TSPO PET in fact measures microglial proliferation rather than activation, it would have practical implications for use of the tracer. For example, it would mean the technique likely could not be used to measure short-term responses to an anti-inflammatory drug.

“This paper highlights the urgent need to develop new imaging biomarkers that truly reflect the pro- or anti-inflammatory status of microglia,” Boutin noted.

A May 27 Science Advances paper suggests one alternative. Researchers led by Silvia De Santis and Santiago Canals at San Juan de Alicante, Spain, described a way to fine-tune diffusion-weighted MRI imaging to capture microglial and astrocyte activation. The technique relies on changes in the morphology and number of these cells. In a rat model of inflammation, it distinguished between the two glial cell types, and it disentangled inflammation from degeneration or demyelination, because each condition induced distinct morphological changes.

The researchers tested their method on six healthy people, and found that the resultant map of microglial density and morphology matched that seen in postmortem brain, suggesting the technique could be used in humans.—Madolyn Bowman Rogers

Comments

  1. This report is very interesting and questions the current assumption by researchers imaging TSPO with PET that they are detecting activated inflammatory microglia in neurodegenerations and multiple sclerosis. The authors find that TSPO expression can be seen at postmortem by microglia that do not have an inflammatory phenotype and, conversely, inflammatory microglia do not necessarily express TSPO. The authors suggest that TSPO PET is a marker of activated microglial density rather than inflammatory activity.

    My view has always been that TSPO expression by microglia is nonspecific and could reflect activation of a phagic or protective microglial phenotype as well as toxic inflammatory activity. Paul Edison and I have suggested that, in Alzheimer’s disease, prodromal cases first show raised cortical TSPO signals due to a rise in protective microglia trying to clear amyloid aggregates, which fails.

    TSPO signal then falls but, as tau tangles form, it rises again, reflecting microglial activation that now is inflammatory and toxic. By the time of end-stage disease seen at postmortem, microglial signal becomes very low as extensive neuronal death has occurred.

    What is clear is that researchers will have to stop referring to TSPO PET as a marker of inflammatory activity in human disease. Nutma et al.'s data makes it clear that raised TSPO signal primarily reflects microglial density and that these cells could have non-inflammatory phenotypes.

    View all comments by David Brooks

  2. This study combines animal models of AD proteinopathies and human tissue, analyzed with various techniques, to support results suggesting that while in rodents TSPO levels represent activated microglia, in humans TSPO better portrays microglia density than activation. This paper is one of the most comprehensive I have seen in this arena. It has a high methodological standard and presents robust results. The study is relevant because the uptake of PET tracers that quantify TSPO availability is currently interpreted as a proxy of microglial activation and, more broadly, neuroinflammation.

    The results presented in this preprint refine the interpretation of TSPO PET findings, suggesting that high uptake depicts a high density of inflammatory cells rather than a combination of their density and activation. High TSPO PET uptake showing microglia cell density, regardless of activation state, could still represent sustained inflammation. In addition, that TSPO availability represents the regional number of microglia would further support results with TSPO PET suggesting that microglia and tau physically propagate together across Braak stages (Sep 2021 news)

    The results raise important questions. If TSPO availability represents microglia density, what is the cause of the abnormally elevated microglial density in AD? And what is the role of neurons and astrocytes? To begin to answer these questions, it would be necessary to phenotypically characterize microglia in regions showing high density. This does not significantly differ from the current challenges ahead of TSPO PET studies, because it is known that the so-far-believed activation state observed with TSPO PET can be heterogeneous, requiring phenotypical characterization.

    These results need to be evaluated in the context of previous literature suggesting increased activation rather than an increase in the total number of microglia in AD (Hopperton et al., 2018). Understanding the reason for these conflicting results is crucial. Specificity/selectivity of TSPO antibodies for humans? Difference in the control groups (controls in the present study were up to Braak IV)?

    In summary, this is an important study that raises important questions about the complex link between AD and glial cells that need further elucidation.

    References:

    Hopperton KE, Mohammad D, Trépanier MO, Giuliano V, Bazinet RP. Markers of microglia in post-mortem brain samples from patients with Alzheimer's disease: a systematic review. Mol Psychiatry. 2018 Feb;23(2):177-198. Epub 2017 Dec 12 PubMed.

    View all comments by Tharick Pascoal

  3. Although this is a preprint version and we should wait for the peer-review process to run its course, this paper addresses the question of the source of the increase in TSPO PET signal (i.e., increase in expression and/or increase in number of microglial cells) when imaging neuroinflammation in brain diseases, which has never been truly elucidated so far. This work is in agreement with numerous other reports showing that TSPO expression in microglia is not directly involved in neuroinflammatory processes but is more a by-product of microglia activation, and is actually more likely reflecting an increase in microglia metabolic activity and/or proliferation rather than a pro- or anti-inflammatory phenotype.

    However, one must note some limitations of the work in its current form, such as the low n number in many experiments, which underpowers the statistical analysis. Nevertheless, this work clearly underscores the urgent need to develop new imaging biomarkers and PET tracers that truly reflect the pro- or anti-inflammatory status of microglia in brain diseases such as Alzheimer’s disease.

    View all comments by Herve Boutin

Make a Comment

To make a comment you must login or register.

References

Research Models Citations

  1. APP NL-G-F Knock-in
  2. hTau.P301S
  3. SOD1-G93A (hybrid) (G1H)

Paper Citations

  1. Owen DR, Yeo AJ, Gunn RN, Song K, Wadsworth G, Lewis A, Rhodes C, Pulford DJ, Bennacef I, Parker CA, StJean PL, Cardon LR, Mooser VE, Matthews PM, Rabiner EA, Rubio JP. An 18-kDa translocator protein (TSPO) polymorphism explains differences in binding affinity of the PET radioligand PBR28. J Cereb Blood Flow Metab. 2012 Jan;32(1):1-5. PubMed.
  2. Owen DR, Narayan N, Wells L, Healy L, Smyth E, Rabiner EA, Galloway D, Williams JB, Lehr J, Mandhair H, Peferoen LA, Taylor PC, Amor S, Antel JP, Matthews PM, Moore CS. Pro-inflammatory activation of primary microglia and macrophages increases 18 kDa translocator protein expression in rodents but not humans. J Cereb Blood Flow Metab. 2017 Aug;37(8):2679-2690. Epub 2017 May 22 PubMed.
  3. Venneti S, Wang G, Nguyen J, Wiley CA. The positron emission tomography ligand DAA1106 binds with high affinity to activated microglia in human neurological disorders. J Neuropathol Exp Neurol. 2008 Oct;67(10):1001-10. PubMed.
  4. Gui Y, Marks JD, Das S, Hyman BT, Serrano-Pozo A. Characterization of the 18 kDa translocator protein (TSPO) expression in post-mortem normal and Alzheimer's disease brains. Brain Pathol. 2020 Jan;30(1):151-164. Epub 2019 Jul 25 PubMed.

Further Reading

Primary Papers

  1. Nutma E, Fancy N, Weinert M, Marzin MC, Tsartsalis S, Muirhead RCJ, Falk I, de Bruin J, Hollaus D, Pieterman R, Anink J, Story D, Chandran S, Tang J, Trolese MC, Saito T, Saido TC, Wiltshire K, Beltran-Lobo P, Philips A, Antel J, Healy L, Moore CS, Bendotti C, Aronica E, Radulescu CI, Barnes SJ, Hampton DW, van der Valk P, Jacobson S, Matthews PM, Amor S, Owen DR. Translocator protein is a marker of activated microglia in rodent models but not human neurodegenerative diseases. bioRXiv. 2022 May 11 bioRxiv
  2. Garcia-Hernandez R, Cerdán Cerdá A, Trouve Carpena Al, Drakesmith M, Koller K, Jones DK, Canals S, De Santis S. Mapping microglia and astrocyte activation in vivo using diffusion MRI. bioRxiv. June 11, 2021 bioRxiv

Follow-On Reading

Papers

  1. Nutma E, Fancy N, Weinert M, Tsartsalis S, Marzin MC, Muirhead RC, Falk I, Breur M, de Bruin J, Hollaus D, Pieterman R, Anink J, Story D, Chandran S, Tang J, Trolese MC, Saito T, Saido TC, Wiltshire KH, Beltran-Lobo P, Phillips A, Antel J, Healy L, Dorion MF, Galloway DA, Benoit RY, Amossé Q, Ceyzériat K, Badina AM, Kövari E, Bendotti C, Aronica E, Radulescu CI, Wong JH, Barron AM, Smith AM, Barnes SJ, Hampton DW, van der Valk P, Jacobson S, Howell OW, Baker D, Kipp M, Kaddatz H, Tournier BB, Millet P, Matthews PM, Moore CS, Amor S, Owen DR. Translocator protein is a marker of activated microglia in rodent models but not human neurodegenerative diseases. Nat Commun. 2023 Aug 28;14(1):5247. PubMed.

Annotate

To make an annotation you must Login or Register.