Translocator Protein-18 kDa (TSPO) Positron Emission Tomography (PET) Imaging and Its Clinical Impact in Neurodegenerative Diseases

1Microglial cells are the resident macrophages of the central nervous system (CNS) that play different roles in both physiological and pathological conditions, by maintaining brain parenchyma integrity and involving in a wide range of neurodegenerative diseases. Microglia represent approximately 5–15% of all cells in the human brain. The brain homeostasis is achieved in part though the ability of microglia to regulate inflammation, such as cytotoxicity, repair and regeneration. In parallel to their well-known immune-modulatory functions, microglia are highly dynamic cells which contribute to the synaptic remodelling/plasticity as well as synaptogenesis, synaptic transmission and pruning.

2Several cellular actors including neurons, astrocytes, T-cells and the blood brain barrier (BBB) modulate microglial function and form a very dynamic network. In particular, microglia have been proposed to act as sensors, effectors and injury recipients of the brain homeostasis breakdown. In fact, upon injury, various afferents stimuli (e.g., soluble factors, cellular interactions) induced morphological (i.e., amoeboid, rod, multi-nucleated, epithelioid or dystrophic state) and functional changes of microglia, known as microglial “activation”.

3In front of homeostatic disturbances (e.g., vascular or tissue damage), microglial cells shift from sensing activity to a reactive state. Originally based on peripheral monocytes/macrophages, the reactive phenotype of microglial cells is dichotomized into “classical activation” or M1, pro-inflammatory and “alternative activation” or M2, anti-inflammatory reactions. According to this concept, microglia phenotypes are defined by triggering responses to cytokines and microbial agents. In fact, interferon-γ from T helper cell type 1 (Th1) causes M1 activation, a response usually associated with the struggle against intracellular pathogens. In M1 state, macrophages release pro-inflammatory cytokines such as IL-1β, IL-6, TNF-α or ROS/RNS. On the other hand, the secretion of IL-4 or IL-13 from Th2 promotes M2a polarization and supports tissue repair/regeneration. M2b cells are induced by the immune complex, secrete IL-10 and regulate the immune response. When macrophages release TGF-β or IL-10 (M2c reaction), anti-inflammatory effects are observed. Microglia polarization/action is a process highly context-dependent (nature of activating stimulus; e.g., PAMPS and/or DAMPS) which is integrated in a time-dependent fashion (acute versus chronic injury).

4First described as a peripheral benzodiazepine receptor (PBR), the translocator protein-18 kDa (TSPO) has been fortuitously identified by Braestrup et al., using rat kidneys as a control tissue during central benzodiazepine receptor (CBR) binding studies using [3H]-diazepam. This secondary binding site for diazepam, which is distinct in its structure from the CBR, was shown to be abundantly distributed in peripheral tissues. It has been renamed as 18-kDa translocator protein (TSPO) to better reflect the cellular functions and tissue distribution of this protein. TSPO is a highly hydrophobic five transmembrane domain protein mainly situated in the outer mitochondrial membrane. TSPO is widely distributed in most peripheral organs including kidneys, nasal epithelium, adrenal glands, lungs and heart, whilst the highest concentrations are in the steroid producing tissues; is also minimally expressed in resting microglial cells in the healthy brain. Although modestly expressed in healthy brain parenchyma, TSPO is substantially upregulated predominantly in microglia cells in a number of neurodegenerative and neuroinflammatory diseases during the microglia activation process. In addition to activated microglia, Lavisse et al. showed that reactive astrocytes could also overexpress TSPO using a model of selective astrocyte activation in the rat striatum. Likewise, after brain injury induced by focal cerebral ischemia, both microglia and astrocytes have been found to overexpress TSPO, with a dissimilar distribution within the infarcted lesion. Thus, astrocytes were mainly observed in the rim surrounding the lesion core, whereas microglia were abundant in the core of infarction. Furthermore, cells of the mononuclear phagocyte lineage, such as peripheral macrophages, also express the TSPO and, in the case of a disrupted BBB, also infiltrate the damaged CNS due to the vascular permeability increased by neuroinflammation.

5TSPO is thought to be involved in a wide array of vital cellular functions including porphyrin transport and heme synthesis, regulation of cell proliferation, steroid biosynthesis and programmed cell death. TSPO’s functions in metabolic homeostasis remain unclear, especially in the supra-physiologic stimulation of microglia. Most studies point out that high TSPO expression attenuates ROS generation. Gatliff and Campanella recently specify to what extent TSPO drives the mitochondrial quality control. Indeed, more than TSPO (over)expression, it seems to be the expression ratio of TSPO to VDAC1 (voltage-dependent anion channel 1) that regulates ROS production. Subsequently, this ROS generation leads to an accumulation of defective mitochondria due to an ineffective mitophagy PARK2-dependent mitochondriopathy, which is involved in several neurodegenerative diseases such as Parkinson’s disease (PD), Huntington’s disease (HD) and Alzheimer‘s disease (AD). Others authors suggest a role of TSPO in oxidative stress by directly facilitating the breakdown of porphyrin. The use of TSPO ligands in rodent models of neuroinflammation (e.g., kainic acid excitoxicity or intracerebroventricular infusion of lipopolysaccharide (LPS)) highlighted TSPO’s involvement in hippocampal gliosis and especially in microgliosis.

6The colocalization of TSPO upregulation and activated glia has been observed in human degenerative disorders including AD, PD, multiple sclerosis (MS), (HD) and amyotrophic lateral sclerosis (ALS). As a result, this increased density of TSPO in CNS diseases has been considered to be a relevant and an early biomarker to evaluate neuroinflammation over disease progression.

7Positron Emission Tomography (PET) imaging of microglia has widely grown over the last 25 years, through the development of radiopharmaceuticals targeting several molecular biomarkers of microglial activation. In this context, many biological targets of neuroinflammation for both diagnosis and therapeutic purposes have been identified. Among them, TSPO has become the reference molecular biomarker for microglia PET imaging. Thus, TSPO PET imaging has been used for both improving the knowledge regarding the role of neuroinflammation in CNS diseases and to assess the efficacy of novel anti-inflammatory therapeutic strategies. Yet, the upregulation of TSPO PET radioligands’ binding is not discriminant between activated microglia, astrocytes and infiltrated macrophages. Indeed, in AD, Venneti et al. observed that astrocytes are not a negligible factor in the increased TSPO PET radioligand binding. Moreover, TSPO overexpression in both astrocytes and microglia seems to be temporally distinct depending on the stage of disease and the disease itself. Thus, TSPO PET imaging is not selective of microglia activation, but depicts a broader multicellular inflammatory reaction.

8In this section, we summarize the progress being made in both the involvement of microglia in various neurologic conditions and the understanding of microglial activation in CNS disorders thanks to TSPO PET molecular imaging.

9A wide variety of endogenous compounds like diazepam, cholesterol or porphyrin interacts with TSPO at distinct binding sites. Because of the availability of high affinity and selective ligands, which can be labeled with various radioisotopes (11C, 18F), the distribution of TSPO can be visualized with PET technique. Historically, the benzodiazepine 11C-Ro5-4864 was the first ligand able to discriminate peripheral from central benzodiazepine receptors. Then, the most widely used TSPO PET radiopharmaceutical, namely 11C-(R)-PK11195, an isoquinoline carboxamide developed in the early 1980s, was the first non-benzodiazepine type compound identified to bind to TSPO with high affinity (human Kd = 2 nM). However, 11C-(R)-PK11195 clinical usefulness is narrowed by several major limitations, including the short half-life of carbon-11, a low brain bioavailability and a poor signal-to-noise ratio due to high nonspecific binding. To counteract these drawbacks, there has been a great effort for the development of second generation TSPO PET radiotracers, including phenoxyarylacetamides derivatives labeled with carbon-11 (11C-PBR28, 11C-DAA1106) or fluorine-18 (18F-FEPPA, 18F-PBR06), imidazopyridines derivatives (11C-CLINME) and pyrazolopyrimidines derivatives (18F-DPA-714). Indeed, a single nucleotide polymorphism in the TSPO gene (rs6971) leading to an amino-acid substitution (A147T) has shown to be the cause of binding affinity properties variation with a huge heterogeneity in PET images and their associated quantitative data. The fact that the second-generation radiopharmaceuticals have been found to be sensitive to this polymorphism leads searchers to develop rs6971-insensitive radioligands. These rs6971-insensitive radioligands, called third generation TSPO tracers, include 11C-ER176 and flutriciclamide (18F-GE180). Nonetheless, the clinical relevance of these compounds remains to be confirmed.

10TSPO PET tracers have been investigated in various preclinical and clinical studies and we have now reported major advances provided by the development of new tracers (Table 1).

11Parkinson’s disease (PD) is a common age-related neurodegenerative disorder that mainly affects the motor system. Symptoms include resting tremor, asymmetrical bradykinesia and muscular rigidity. PD results from the progressive degeneration of dopaminergic neurons, initially in the substantia nigra, which causes a clinical parkinsonism when the loss of the striatal dopamine reaches to 60%. PD is neuropathologically characterized by a neuronal rarefaction in the substantia nigra pars compacta and the presence of neuronal Lewy bodies (LB) inclusions, which are abnormal aggregates of α-synuclein. Since the discovery of activated microglia in a post mortem study of PD brains, objectified by a large amount of HLA-DR-positive microglial cells in the locus niger, “neuroinflammation” became one of the main culprits in the pathogenesis and in the progression of PD. Correlation analysis between TSPO tracer binding potential values and clinical parameters show that microglial activation PET imaging should predict the symptom severity, but only in early-stage PD patients (Table 2). Indeed, the nigral 11C-PK11195 accumulation was significantly positively correlated with the motor severity of the UPDRS assessment (unified Parkinson’s disease rating scale) in drug-naïve PD patients. In addition, PD patients presented significantly increased 11C-PK11195 uptake in the basal ganglia (striatum and pallidum), thalamus, brainstem and frontal, temporal and cortical regions. This TSPO overexpression in different cerebral regions supports the notion that microglial activation is not confined to nigrostriatal dopaminergic neurons within PD pathophysiology. Moreover, identified by positive CD68 immunoreactivity, amoeboid activated microglial cells were identified in both the striatum and the hippocampus of PD patients, suggesting that an inflammatory response occurs to the active α-synuclein pathology in the hippocampus area. Furthermore, TSPO PET imaging shows that the microglial shift, from a sensing to a reactive state, takes place at the early stage of PD and can occur even without dementia. This was corroborated by reported data from a rat model of PD, where the activation of microglia seems to precede dopaminergic neuronal cell loss. Given the fact that a chronological spread of LB inclusions occurred in α-synucleinopathies, and these α-synuclein can directly provoke microglial activation, we could expect to find, in a time-dependent fashion, a widespread TSPO signal in PET imaging. Interestingly, Gerhard and colleagues reported a lack of change in levels of cortical and subcortical TSPO tracer uptake for two years, despite the clinical deterioration of PD patients. Similarly, microglia TEP imaging followed-up with a second generation of TSPO radioligand (11C-DPA713) also failed (in ROI—Regions of interest—analysis) to show increased tracer binding at one year in earlier stage PD patients. Therefore, microglial activation assessed by TSPO PET imaging seems to be relatively static over time in PD, and only microglia phenotypes may change with the disease progression.

12Multiple system atrophy (MSA) is another form of atypical parkinsonism in which overlapping symptoms as well as tremor, akinesia and rigidity are associated with autonomic failure and cerebellar dysfunction. In contrast to PD, the accumulation of α-synuclein fibrils leads to the generation of glial cytoplasmic inclusions (GCI) which affects mostly the oligodendroglia rather than neurons.

13In MSA patients, significantly increased 11C-PK11195 uptake has been observed in the dorsolateral prefrontal cortex, putamen, pallidum, pons and substantia nigra, reflecting the known distribution of microglial and neuronal lesions in MSA. Additionally, TSPO is overexpressed in the white matter of MSA and Progressive Supranuclear Palsy (PSP) patients. In MSA, these results appear to echo the findings of neuropathological studies and could reflect the demyelination process which predominantly affects the MSA-C variant. Unlike the PD scenario progression, microglial activation in a mouse model of MSA-like neurodegeneration occurred parallel to the dopaminergic neuronal loss in the locus niger. Remarkably, the analysis of histopathological patterns of MSA brains revealed that the amount of GCIs and the level of microglial activity is reduced especially since the tissue injury becomes severe in white matter areas, unlike gray matter areas. If we generalize this concept of microglial (down/up)-regulation according to lesion severity in MSA, that could explain that 11C-PK11195 binding increased in the thalamus or decreased in the substantia nigra after 24 weeks of active disease, compared to baseline.

14Apart from α-synucleinopathies, tauopathies such as PSP and corticobasal degeneration (CBD) belong to parkinsonian disorders. In addition to parkinsonism, the clinical features of PSP include early falls and abnormal eye movements (vertical supranuclear gaze palsy). The hallmark of PSP comprises the deposition of 4-repeat (4R) tau protein in neurofibrillary tangles and neuropil threads that are found in the basal ganglia and brainstem.

15In vivo microglia imaging showed increased 11C-PK11195 binding in basal ganglia, thalamus, midbrain, substantia nigra, pons, cerebellum and frontal lobe in PSP patients. A post mortem study by Fernández-Botrán et al. specified that the high degree of correlation between microglial activation, tau burden and the areas of degeneration (substantia nigra and subthalamic nucleus) previously described, involves a high expression level of IL-1β. Although the involvement of microglial activation as etiology or consequence remains unclear, these results suggest that there is an M1 reactive phenotype of microglial cells at a point of PSP progression. Similar to PD, while PSP patients clinically declined over 6 to 10 months, there were no significant changes in TSPO PET imaging.

16CBD is a rare neurodegenerative tauopathy in which the hyperphosphorylated 4R tau deposits concern neurons and glia, in both the cerebral cortex and the basal ganglia. In CBD, a recent report supports that astrogliopathy precedes neuronal loss, which leads to clinical symptoms as well as asymmetrical akinetic-rigidity, apraxia, disequilibrium and limb dystonia. Elevated 11C-PK11195 binding in the striatum, thalamus and substantia nigra is in consonance with neurological injury patterns in CBD, especially tau load. Besides the diagnostic interest of PET, molecular imaging has also been used as a tool to evaluate treatment efficacy in neurodegenerative disorders.

17Huntington’s disease (HD) is an autosomal dominant inherited trinucleotide (CAG) repeat neurodegenerative disorder in which the age at onset is inversely correlated with the size of the polyglutamine repeat expansion. This feature allows for the dichotomization of HD patients into premanifest HD, i.e., gene carriers are clinically asymptomatic, and manifest HD, that is to say, after motor onset of the disease. This mutation in the exon 1 of the gene encoding the Huntingtin protein (HTT) leads to a mutant HTT (mHTT) which is expressed by peripheral monocytes of premanifest HD patients, and induces both a loss of function (e.g., regulating intracellular transport, controlling apoptosis pathway) and a gain of toxic functions (e.g., abnormal protein-protein interaction, promoting pro-inflammatory cytokines). HD is neuropathologically characterized by the formation of intranuclear inclusions of mHTT resulting in striatal, cortical and hypothalamic atrophy which are associated with an accumulation of reactive microglia in these same regions. In line with these findings, in vivo microglia PET imaging showed an overexpression of TSPO in the striatum,cortical regions and hypothalamus (Table 3). According to lesion topology, which even depends on the stage of HD, symptoms include motor troubles (chorea, parkinsonism), cognitive decline (slowed mentation, impaired strategy) and psychiatric disorders (mood, irritability, apathy, depression). To such an extent, microglial activation seems to begin in cognitive regions, especially in associative striatum and correlated with cognitive dysfunction. Moreover, striatal 11C-PK11195 binding is strongly correlated with the five-year probability of disease onset in asymptomatic HD gene carriers. Furthermore, in manifest HD patients, the correlation between striatal TSPO tracer binding and the severity of motor symptoms (evaluated by the unified Huntington’s disease rating scale) could echo the ex vivo relationship found between neuronal loss and density of reactive microglia. Although HD is a monogenic disease, recent studies revealed that different genes, distinct from the HD locus itself, could change disease manifestation and progression. We hypothesize that these genetics and thus clinical features could explain the difference in microglial activation, objectified by ventral striatum TSPO tracer binding variability in premanifest HD patients.

18Alzheimer’s disease (AD) is the most prevalent neurodegenerative cause of dementia and is characterized by amnesic memory impairment, language deteriorate and behavioral changes due to cognitive decline. The neuropathological hallmarks of AD include extracellular peptide amyloid β (Aβ) forming neuritic plaques and neurofibrillary tangles (NFTs) of intraneuronal deposits of hyperphosphorylated/misfolded tau proteins. Each of these lesions has a characteristic distribution during the evolution of AD and has a more or less strong clinicopathological correlation. A transcellular propagation of abnormal tau proteins and Aβ occurred in AD. Regarding the neurofibrillary degeneration, it starts in the entorhinal cortex and then spreads to synaptically connected brain areas such as the hippocampus/limbic system, and finally affects the frontal and parietal lobes (i.e., associative isocortex). The amount and the topography of these NFTs match with the neuronal loss and correlate with the severity and the duration of dementia as well as lesions topology (e.g., neocortex, limbic system) is in consonance with the clinical features of AD (e.g., cognitive impairment). Moreover, tau hyperphosphorylation seems to be promoted by a pro-inflammatory environment (e.g., IL-6 release) and may support microglial activation, highlighting the vicious circle of tauopathy’s process in AD (for review, see Calsolaro et al., 2016). Microglia PET imaging supported these experimental results. Indeed, microglia PET imaging revealed greater radioligand uptake in AD patients than in healthy controls in the same regions which are sequentially affected by neurofibrillary degeneration: entorhinal cortex, hippocampus, frontal and parietal cortex (Table 4). Longitudinal studies exhibited an increase in cortical-activated microglia during the course of AD. Similarly, cross-sectional post mortem studies using the temporal neocortex of 40 AD subjects with a symptom duration ranging from 4 to 20 years showed a significant positive association between CDC68+ microglia and symptom duration. Comparison in TSPO density between mild cognitive impairment due to AD (MCI) and AD dementia yielded ambiguous results: significant difference was found between the two groups or was absent according to different studies. MCI represents a diagnostic entity for the transition between normal aging and AD dementia. Analogous to Aβ and tau that has already accumulated in the brain of MCI patients, PET imaging showed that microglial activation occurred before the onset of clinical symptoms of dementia. These results are consistent with the presence of neuroinflammatory biomarkers in the cerebrospinal fluid of pre-symptomatic AD subjects. Nevertheless, some studies provided contradictory results.

19Unlike NFTs, amyloid plaques deposition showed a lesser stereotypical spatiotemporal pattern of progression; they accumulate mainly in the isocortex, and the amyloid burden does not correlate with the severity or the duration of dementia. Supported by various biomarker-autopsy correlation studies, Jack et al. proposed that AD biomarker curves assume a sigmoidal shape as a function of time, involving a threshold effect and a late deceleration of tau and Aβ loads. With current data on microglial activation, it seems difficult to give a similar trend to the progression of neuroinflammation in AD dementia. However, microglial activation states obviously play a significant role in aging and age-related disease, especially in AD. Aβ (both, soluble and fibrillary) can bind to several Toll-like receptors (TLR) and co-receptors expressed on microglia, leading to its activation and promoting the formation of clusters of activated microglia surrounding amyloid plaques in AD brains. In line with these ex vivo findings, TSPO PET imaging showed higher microglial activation in various cortical regions known to be the seat of Aβ deposits (e.g., prefrontal, temporal, occipital, precuneus, cingulum) in AD dementia patients, compared to healthy subjects. This Aβ-induced inflammatory environment, combined with effect of immunosenescence on microglia, lead to a situation in which microglia that initially exhibited a neuroprotective M2 phenotype at the beginning of AD, become confined to a classical activated state (i.e., M1) with strongly reduced Aβ-clearing capabilities in the advanced stage of AD.

20Correlation analysis between TSPO tracer binding and AD severity provided heterogeneous findings. In most studies, in parallel to cognitive deterioration, objectified by a decreased MMSE score or increased CDR-SB, TSPO expression density in cortical regions was linearly raised. To the same extent, Kreisl et al. exhibited that early onset AD (EOAD) patients presented significantly higher 11C-PBR28 than later onset AD (LOAD) patients, both at baseline and follow-up. This overexpression of activated microglia in more active AD (i.e., early onset) may confer on it the status of a deleterious process. Other studies found no link. Conversely, one of the surprising results of the study of Hamelin et al. is that MMSE scores correlate positively with TSPO tracer binding. Moreover, the radioligand (18F-DPA-714) GCI was greater in slow decliners than in fast decliners, suggesting an early neuroprotective effect of microglial activation in AD.

21In light of such antipodic findings, we cannot exclude the fact that TSPO PET imaging has limited abilities to robustly reflect microglial activation in AD. In fact, Banati et al. suggested that the overexpression of TSPO in activated microglia may be related to structural changes in mitochondria due to morphological changes in microglial cells. On the basis of that postulate, we hypothesize that the profound morphological deterioration of microglial cells which occurred in AD could be a source of signal variability in TSPO PET imaging.

22Frontotemporal lobar degeneration (FTLD) is a clinically, pathologically and genetically heterogeneous syndrome characterized by progressive clinical features such as behavioral and personality changes, cognitive decline and parkinsonism. These symptoms result from progressive neurodegeneration and, ultimately, focal atrophy of the frontal and temporal cortex in which inclusions of hyperphosphorylated tau protein in neurons and glial cells (FTLD-tau) and inclusions of TDP-43 (FTLD-TDP) compose the two main histopathological hallmarks of this disease. FTLD has a strong genetic component; mutations causing FTLD include MAPT (microtubule-associated protein tau) and progranulin. Interestingly, proganulin is expressed by activated microglia and polarizes microglial cells into M2 anti-inflammatory phenotypes, suggesting that microglia polarization deficiency could play a role in FTLD. Indeed, Cagnin et al. demonstrated an enhanced uptake of tracer (11C-PK11195) in the typically affected frontotemporal regions. Two MAPT pre-symptomatic gene carriers exhibited microglial activation in either, posterior cingulate cortex, region which sends projection to the superior temporal sulcus, or in the medial frontal cortex. Elevated putaminal uptake of TSPO tracer in frontotemporal dementia patients is consistent with the neuropathological (involvement of the basal ganglia) and clinical features (parkinsonism) of FTLD.

23Amyotrophic lateral sclerosis (ALS) is a rare adult-onset neurodegenerative disorder in which upper and lower motor neurons degenerate leading to progressive muscle paralysis then respiratory failure with a fatal outcome, typically within 2–4 years from onset. This motoneuropathy that occurred from 5% to 20% in a positive family history of ALS, fronto-temporal demence (FTD) or both, results in fasciculation, cramps, muscle atrophy associated with hyperreflexia and spasticity. Central inflammation has been proposed to play a role in the motor neuron death process in ALS. TSPO PET imaging supported this assumption. Indeed, ALS patients presented significantly greater TSPO tracer binding in motor cortex than healthy volunteers. Moreover, “microgliopathy”, assessed by CD68- and Iba1-reactive microglial cells, correlated with the disease progression and is linked to the severity of upper motor neuron deficits in the corticospinal tract of ALS patient autopsies. These ex vivo findings were clinically sustained by PET studies, where TSPO expression density was correlated with motor symptoms severity, assessed by the UMND (Upper Motor Neuron Disorders) score (Table 5). Interestingly, the one study that described an elevated TSPO radioligand uptake (11C-PBR28) in the brainstem of bulbar-onset ALS patients was also the one in which a statistical correlation was found between SUVr values and ALSFRS-R (revised ALS functional rating scale). Given the fact that this functional scale is more accurate to evidence lower motor neuron involvement, this association is probably due to the fact that microglial activation in the brainstem could reflect inflammation around the lower motor neurons of the cranial nerves. The murine model of ALS, in which microglial cells overexpressed mutant-superoxide dismutase (mSOD1), the main cause of non-sporadic ALS, point out the pathogenic role of microglia in ALS; mutant protein may be more involved in disease progression than in its initiation. Inversely, activated microglia in the temporal region could be involved in the early stage of ALS, and then progressively decrease. Indeed, this cerebral area was not affected in ALS patients with a more advanced stage. Like AD dementia, a switch from the classically activated M1 phenotype to the alternatively activated M2 microglial state takes place during the progression of ALS disease.

24Bipolar disorder (BD) is a chronic mood trouble in which episodes of elevated mood and manic state are typically alternated with major depressive access, interspersed by euthymic periods. Although the pathophysiology of this severe mental illness remains to be elucidated, a large body of literature suggests that several pathophysiological processes may sequentially and/or concurrently occur in BD, as well as corticolimbic control circuits deterioration, monoaminergic neurotransmission imbalance or microglia homeostasis disruption (for review, see Sigitova et al., 2016). Among these, hippocampal neuroinflammation, assessed by enhanced 11C-PK11195 binding compared to healthy subjects, could echo the change in the cytokines pattern’s profile (TNF, IL-1, IL-2 and IL-6) of BD-I patients (Table 6).

25Major depressive disorder (MDD) is a highly prevalent condition, affecting nearly 4% of the adult population, which could complicate neurodegenerative disorders (e.g., MS, PD) or be isolated, without comorbidities. Although this depressive disorder mainly concerns mood disturbance, clinical features include a large range of symptoms as cognitive deficits, vegetative impairment (e.g., loss of appetite, sleeping difficulties) or behavioral deterioration (e.g., anhedonia, suicidal ideation). Several data, such as the association between central inflammation and depression-like symptoms in rodents, abnormalities in peripheral cytokines of depressed patients, or association between interferon alpha in the context of hepatitis C and depressive symptoms, have led some investigators to propose a microglial dysregulation theory for MDD etiology. Nevertheless, post mortem studies have questioned this assumption. Indeed, no evidence of microglial activation was found in the brain of MDD patients. In line with these results, microglia PET imaging did not reveal any statistical difference in 11C-PBR28 binding between healthy volunteers and patients with mild to moderate depression. However, Setiawan and colleagues recently found increased microglial activity (18F-FEPPA, VT, total distribution volume) in the anterior cingulate cortex of MDD patients, the region known to be implicated in the modulation of emotional behavior. Moreover, this PET imaging result was correlated with the MDE (major depressive episode) severity, objectified by a 17-item Hamilton depression rating scale (HDRS) score.

26Apart from epidemiological and genetic evidences of overlap between BD and schizophrenia (SCZ), molecular imaging of microglia of these psychiatric disorders also share some similarities, especially hippocampal TSPO overexpression compared to normal controls. SCZ is a chronic early adult-onset psychiatric disorder in which symptoms are typically divided into positive (e.g., hallucinations), negative (e.g., affective flattening, alogia) and cognitive (e.g., concentration and memory impairment). Multiple lines of evidence from immunohistochemical cell counting studies to clinical trials of a non-selective microglial inhibitor support the fact that microglia is involved in the pathophysiology of SCZ. Nevertheless, as with post mortem studies, microglia PET studies provided conflicting results. Indeed, some studies revealed significantly higher TSPO radioligand binding in SCZ patients in total gray matter, especially in the frontal and temporal lobes and the hippocampus than in controls subjects. However, other studies found no clear difference between normal controls and SCZ patients in various brain regions.

27Neuroinflammation imaging in psychiatry is in an early stage, and there is much more to investigate (for review, see Slifstein and Abi-Dargham, 2016), especially regarding TSPO imaging. Indeed, there is evidence that the dysregulation of steroidogenesis by TSPO is associated with psychiatric disorders. For instance, an association has been found between rs6971 polymorphism in the TSPO gene and the diagnosis of BD.

28Multiple sclerosis (MS) is a chronic autoimmune disease of the CNS in which migration of myelin-reactive T-cells into the CNS is followed by astrocyte and microglial activation, recruitment of peripheral macrophages and oligodendrocyte destruction, resulting in focal and diffuse demyelination. The clinical presentation of MS as well as symptoms produced by the episodes of neurologic dysfunction (i.e., “attacks” or “relapses”) are heterogeneous. Regarding the symptoms of MS, tingling, weakness, vision loss, incoordination or bladder dysfunction may occur and depend upon the site of neurologic involvement. As regards the clinical picture, the most typical form at the onset of the disease is relapsing-remitting MS (RRMS), which is characterized by attacks interspersing with neurologically stable periods. Over several years to decades, RRMS shifts into secondary progressive MS; the recovery between relapses becomes partial and gradually provokes persistent symptoms which lead to irreversible disability. Only a minority of patients (~10–20%) experience insidious worsening without acute attacks, a clinical course known as primary progressive MS (PPMS). The major substrates of the disease progression include demyelinated cortical lesions, diffuse cortical and deep gray matter degeneration, which are associated with the chronic activation of microglia. Microglia play a central role in maintaining the central chronic inflammation in MS, to the extent that its presence and distribution in the plaques allow for the neuropathological classification of the nature of the white matter lesions (i.e., active versus inactive) in secondary progressive multiple sclerosis (SPMS).

29Although the clinical and/or MRI discovery of dissemination of the lesions in time and space remains the cornerstone of MS diagnostics, TSPO PET imaging could be a promising technique to diagnose MS or to detect the conversion of RRMS to SPMS. In light of the limits of clinical assessment (e.g., measuring EDSS, expanded disability status scale) and MRI parameters to predict the progression of MS, TSPO PET studies have shown hopeful findings (Table 7). Indeed, activated microglia in SPMS relative to RRMS are characterized by both specific and more diffuse distribution in cortical regions as well as greater TSPO radiopharmaceutical uptake, especially in the thalamus and deep gray matter. Provided that these results will be replicated in larger cohorts, it paves the way to early intervention in MS, thanks to TSPO PET imaging, in order to prevent or efficiently treat SPMS. At the upstream level in the disease history, Giannetti and colleagues demonstrated in clinically isolated syndrome (CIS), the prodromal state of MS, that patients who developed MS at two years had higher 11C-PK11195 binding at baseline, especially in normal-appearing white matter (NAWM). Analogous to the relationship between MCI and AD, we hypothesize that NAWM TSPO tracer uptake of CIS subjects could allow for the establishment of an x-year probability of disease (i.e., MS) onset. This significant increase in NAWM TSPO density, compared to healthy volunteers, has also been found in RRMS and SPMS patients. In addition to detecting activated microglia in areas of the brain which appear normal in MRI, the NAWM, TSPO PET imaging could predict the development of some gadolinium-enhancing lesions. In fact, increased focal 11C-PBR28 binding in an area of NAWM at the time of the PET scan corresponded to a contrast-enhancing lesion at a follow-up MRI exam one month later, as was found in two MS patients, suggesting a role for early microglial activation in MS lesion formation. Apart from the fact that microglial activation may precede MS lesions, and that there is a greater TSPO radioligand uptake in RRMS than in SPMS, a body of evidence gives a deleterious status of activating microglia in MS. Among these, RRMS patients who had the greatest number of relapses (i.e., more clinically active) also had the highest 11C-PK11195 binding potential. Furthermore, association analysis revealed that TSPO tracer binding levels in various brain regions correlated positively with both MS score severity (e.g., EDSS) and disease duration. This set of results is in consonance with post mortem studies in which severe cortical microglia activation and demyelination are associated with a pejorative outcome. Nevertheless, some microglia PET studies provide conflicting results in terms of the association analysis in either EDSS or disease duration.

30Ratchford and colleagues provided proof of the concept that microglia PET imaging with 11C-PK11195 could also be a tool to assess disease-modifying drugs for RRMS efficacity. Indeed, radiopharmaceutical binding potential per unit volume was statistically decreased in the whole brain after one year of glatiramer acetate. This result supported the in vitro evidence of its mechanism of action in which an inhibition of transformation to an activated microglia form could be responsible for therapeutic effects.

31TSPO molecular imaging provides valuable information such as presence, intensity and localization of the microglia activation, but this uniparametric approach might be considered as one component of a multiplexed strategy. Indeed, TSPO imaging combined with further PET targets, imaging techniques (e.g., MRIf), or even biology, could allow for a deeper understanding of the role of activated microglia in neurodegenerative disorders. We organize this section into; first, unimodal approaches, that is to say, TSPO PET imaging coupled with other PET radiopharmaceuticals; and second, multimodal approaches associating microglia PET imaging with other modalities.

32Neuroinflammation is a complex physiological response to CNS injuries involving a biochemical molecular cascade involved in both tissue degradation and remodeling. Activation of microglia cells that leads to cell death is often related to oxidative stress through cytokines and chemokines release, glutamate-induced excitotoxicity, mitochondrial dysfunction or by protein aggregation. As described above, PET studies exploring neuroinflammation in CNS disorders have mainly been focused on the evaluation of the sole TSPO. Therefore, identifying and imaging other molecular biomarkers involved in inflammatory cascade would be a promising approach combined with TSPO imaging.

33PET imaging associating TSPO radioligands with tracers for other molecular targets involved in microglial activation has already allowed for the assessment of temporal progression of different biochemical mechanisms involved in neuroinflammation. Indeed, in the rodent model of cerebral ischemia, Martín et al. have assessed in parallel the in vivo expression of both α4β2 (nicotinic acetylcholine receptors) and TSPO in reactive glial cells using 2-18F-fluoro-A85380 and 11C-PK11195, respectively. Results have shown a progressive binding increase during the first week after cerebral ischemia, followed by a progressive PET-binding decrease in the ischemic territory. Likewise, differential time-dependent activation of TSPO and matrix metalloproteinases (MMPs), protease involved in tissue degradation and remodeling has been observed by PET co-imaging with the 18F-DPA714 and the 18F-BR-351, suggesting that their expression after stroke are subsequent or independent phenomena. While TSPO is known to be an early biomarker in the neuroinflammation cascade, another longitudinal investigation, using 18F-FSPG and 18F-DPA-714, in the cerebral ischemia model displayed an earlier expression of the cystine/glutamate antiporter (system xc-), involved in glutamatergic excitotoxicity, compared to TSPO imaging, evidencing its novel role on inflammation after cerebral ischemia. 18F-FSPG, a system xc- PET tracer, combined with 11C-PK11195 has also been used during the course of experimental autoimmune encephalomyelitis (EAE) induction in rats. In this MS model, 18F-FSPG showed on the one hand a significant increase of system xc-function in the lumbar section of the spinal cord at 14 days but, on the other hand, showed a significant increase of microglial/macrophage activation in the spinal cord and cerebellum two weeks after EAE induction. These dissimilar binding profiles between both PET tracers might be due to the existence of different microglial/macrophage populations throughout the CNS following MS. In the same way, to assess COX implication during the neurodegenerative process in AD, COX-1 and TSPO PET co-imaging provided details of the time course in the inflammatory process in a rodent model of neuroinflammation (induced by intrastriatal injection of lipopolysaccharide or quinolinic acid). High accumulation of COX-1 tracer (11C-ketoprofen methyl ester) was observed in the early phase, matching with the time course of microglial activation and macrophages, whereas the high accumulation of 11C-PK11195 started slowly and lasted until at least 14 days afterward, corresponding to changes in both the activation of microglia/macrophages and astrocytes.

34Thus, upregulated biomarkers in reactive glial cells complementary to TSPO are expected in order to fulfill the current clinician’s expectancies regarding the microglial role in neurodegenerative diseases and associated potential novel therapeutic approaches.

35Dopaminergic neurotransmission is a key player in various brain functions, including behavior and motor functions, and is involved in the pathogenesis of psychiatric (e.g., SCZ, depression) and neurological disorders (e.g., PD, parkinsonian syndromes). Movement disorders are strongly linked with dopaminergic impairment of the striato-pallido-thalamic-cortical axis in which the striatum receives afferent projections from cortical glutamatergic and nigral dopaminergic neurons. Molecular imaging of dopaminergic presynaptic integrity (e.g., DAT with 11C-CFT or 18F-FE-PE2I), dopamine synthesis and turnover (e.g., 18F-DOPA) and postsynaptic dopamine receptors (18F-Fallypride) are used to evaluate this dopaminergic pathway in various CNS conditions. Experimental models of PD suggest that microglial activation may lead to dopaminergic neuronal loss though iNOS-dependent oxidative stress. Three clinical studies have been performed to assess this association between microglial activation and dopamine terminal loss in PD. Concerning PD patients under antiparkinsonian medication, association analysis revealed that microglial activation did not correlate significantly with either putaminal 18F-DOPA signal or nigrostriatal DAT levels, objectified by 18F-FE-PE2I binding. Nevertheless, the mechanism of action of antiparkinsonian drugs is based on the enhancement of dopaminergic transmission in the brain, therefore in this situation PET imaging of the dopaminergic system is de facto biased, as previously demonstrated. The one study conducted on ten drug-naive patients with PD (rated at stage 1 to 2 on the Hoehn and Yahr scale) brought out the fact that 11C-PK11195 binding in the midbrain significantly and negatively correlated with those for 11C-CFT in the dorsal putamen, a region known to engage in motor execution. Moreover, this microglial activation correlated positively with the motor severity (UPDRS assessment), suggesting a causal deleterious relationship between microglial activation and nigrostriatal pathway impairment, both clinically (UPDRS) and biochemically (central hypodopaminergism). However, this association does not elucidate the sequence of events. Indeed, dopaminergic neuronal loss is accompanied by the release of endogenous constituents known as damage-associated molecular patterns (DAMPs), which are able to activate microglial cells.

36Amyloid deposition, tangle formation, activated microglia and neuronal dysfunction are pathological processes involved in AD. However, the relative role of these processes in driving disease progression is still unclear. As stated above, microglia PET imaging in AD provides heterogeneous findings, hence the association analysis between fibrillar β amyloid, objectified by 11C-PIB binding, and microglial activation could (i) investigate longitudinal changes of microglial activation and amyloid; and (ii) assess the temporospatial relationship between these processes in AD. Indeed, a positive correlation was found in most of the cortical regions between amyloid burden and microglial activation. Longitudinal study revealed that this correlation was less pronounced at follow-up, and the location of the correlated regions presented a dynamic change, suggesting that microglial activation is not static and its progression during AD may fit with the model of Calsolaro et al.. Furthermore, the positive correlation between amyloid load and microglial activation is stronger in AD patients than in MCI patients, suggesting that microglial activation could be driven by the presence of amyloid deposition while other studies challenged these findings in MCI patients. Likewise, a positive correlation was seen in the parietal lobe and the anterior cingulate cortex in PDD patients. This combined approach of amyloid and microglial imaging is nevertheless submitted to debate by some studies pointing out no correlation or negative association with amyloid deposits in the posterior cingulate cortex in AD patients.

37On the same model as the amyloid/microglia imaging, with the emergence of new NFTs PET tracers, the combined use of tau and microglial imaging should also contribute to the pattern of distribution of the increased tau deposition and widespread microglial activation in AD subjects. Not only because post mortem and initial tau imaging studies indicate that the amount of tau deposition and their topographical distribution in the brain might be more relevant, and more tightly associated with neurodegeneration and cognitive decline than β amyloid deposition, but also because in studies of tau model mice using 18F-FEDAA1106 or 11C-Ac-5216, microglial TSPO expression was linked to neurotoxic tau pathology. Nevertheless, to our knowledge, no combined use of tau and microglial imaging study has been performed to date.

38Furthermore, recent studies suggest that neuroinflammation and α-synuclein disorders are interconnected processes which promote each other and perpetuate neurotoxicity. This interaction leads to distinct microglial polarization states and drives the chronic progression of neurodegeneration. Thus, it seems appropriate to combine neuroinflammation/α-synuclein imaging, for which efficient radiopharmaceutical are currently strongly expected.

39Microglia PET imaging (11C-PK11195) has been used in combination with 11C-methionine for the differential diagnostics between tumefactive MS lesion and low-grade glioma in a case report. Beside this diagnostic potential of TSPO PET imaging in MS as previously described, the hopeful results of myelin PET tracers in animal studies such as 11C-CIC or 11C-MeDAS (for review, see Faria et al.) lead us to believe that a multiplexed approach with these radiotracers could allow us to understand the relationship between the demyelination/remyelination process and microglial activation in MS.

40Cerebral glucose metabolic rate (rCMRGlc) measured by 18F-fluorodeoxyglucose (18F-FDG) PET is currently the most accurate in vivo method for the investigation of regional human brain metabolism, and is suggested to be to be an indirect measure of synaptic function. Thus, changes in neuronal activity induced by neurodegenerative diseases are reflected in an alteration in glucose metabolism.

41A significant negative correlation between microglial activation and glucose metabolism was present in the hippocampus, medial temporal cortex in AD patients, in the temporoparietal and occipital cortex and hippocampus in PDD patients, and in the hippocampus in MCI patients. This inverse correlation implies the direct damaging effect by the activated microglia on neuronal loss in AD and PDD, objectified by the glucose metabolism decrease. In addition, Fan and colleagues have used three combined molecular imaging radiopharmaceuticals (11C-PIB, 11C-PK11195 and 18F-FDG) to show a correlation between amyloid load, levels of microglial activation, and reduced glucose metabolism in AD and, to a lesser extent, in MCI and PDD subjects; furthermore, the correlation between amyloid load and glucose metabolism was more extensive at follow-up and was consistently associated with cognitive decline in AD patients, suggesting that, once synaptic dysfunction occurs, the disease is active and progressive.

42In order to investigate CNS disorders by a broader approach combining concomitantly complementary pathophysiological aspects, the use of TSPO PET in a multimodal neuroimaging strategy has increased significantly in recent years. In this framework, MRI and its different subtypes (i.e., structural, diffusion and functional MRI) are among the most democratized approaches, in combination with TSPO PET imaging.

43Brain atrophy assessed in structural MRI has been found in various CNS disorders. This is especially true in AD, where hippocampal atrophy correlates with clinical decline and assumes the status of both the diagnostic and imaging biomarker of AD progression. This approach was used to assess the relationship between microglial activation and neurodegenerative process in AD, PDD, SCZ and MS. Indeed, the amount of activated microglia in the whole cortex, medial temporal lobe and hippocampus was inversely correlated with hippocampal volume and 18F-FDG hippocampal uptake in AD and PDD patients. In front of this spatial agreement between microglial activation in the cortical projections from the hippocampus and hippocampal atrophy, the authors assumed that microglial cells in a reactive state might be one of the effectors of neurodegeneration. The same association between TSPO density increase and cerebral volume decrease reached significance in various cortical regions such as the inferior parietal lobule, precuneus, superior temporal cortex and especially in the prefrontal cortex, in which interconnection with the thalamus and hippocampus are strongly linked to episodic memory impairment in AD patients. On the contrary in BD-I patients, association analysis between 11C-PK11195 binding potential and hippocampal volume failed to exhibit a significant relationship. Quantitative magnetization transfer imaging (qMTI), an advanced structural MRI technique providing information about in vivo tissue integrity, is able to detect inflammation-related changes in brain microstructures. Furthermore, qMTI in MS patients has shed light on the pattern of tissue structural abnormalities in NAWM being greatest near the ventricles. In a way to characterize the temporal/sequential/spatial relationship between microglial activation and anatomically objective brain lesions, a coupling TSPO PET/qMTI could be a promising approach, especially in MS.

44Diffusion-weighted imaging, reflecting (ab)normal tissue diffusivity (measured by apparent diffusion coefficient (ADC)), has been used in combination with 11C-PK11195 in PD, MSA and PSP patients. Although the small sample size limits the scope of study’s conclusions about the impact of microglial activation on tissue damage, a positive correlation between microglial activation and ADC was found in the pons of atypical parkinsonism patients. Given the fact that ADC indirectly reflects cellularity, this association suggests a causal relationship between microglial activation/TSPO overexpression and gliosis.

45fMRI is a technique that uses blood-oxygen-level-dependent (BOLD) contrast to detect changes in brain hemodynamics. Hippocampal functional connectivities to the subguenal and prefrontal and parietal regions correlated with the beck depression inventory scores and 18F-PBR111 DVR in MS patients. This correlation analysis between hippocampal 18F-PBR111 binding and functional connectivity, assessed by fRMI, lead authors to hypothesize that microglial activation is involved in MDE pathophysiology in MS patients. Besides “conventional fMRI”, innovative techniques like GluCEST (Glutamate chemical exchange saturation transfer) appear to be hopeful tools to assess and understand neuroinflammation, and may be even more useful in a multiplexed approach with microglia PET imaging.

46Beyond the interest of coupling MRI techniques with TSPO PET modality, recent advances in imaging of activated microglia have been shown promising results. Indeed, 2D optical imaging spectroscopy or optical projection tomography applied to a 3D spatial assessment of the development of neuroinflammation provided positive outcomes in rodent models.

47The understanding of the microglial activation status in neurodegenerative disorders could be reached by coupling microglia PET imaging to several biological techniques including metabolomics, transcriptomics, magnetic resonance spectroscopy (MRS) and measurement of central and/or peripheral pro-inflammatory components. Among these, 1H MRS in BD-I patients provided surprising results. Whereas no relationship was found between phosphocholine + glycerophosphocholine concentrations, cellular turnover biomarkers and microglial activation; N-acetylaspartate + N-acetyl-aspartyl-glutamate concentrations, representing loss of neuronal integrity when levels decreased, were positively associated with 11C-PK11195 binding in the left hippocampus. These findings challenged the neuroinflammation theory of psychiatric disorders in which activating microglia have a negative effect on neuronal survival.

48In MS patients with a higher inflammatory load, the concentration of myoinositol, a metabolite which has been proposed as a glial and especially as an astrocyte marker, was correlated with 11C-PBR28 DVR in WM lesions, suggesting the joint activation of astrocytes and microglia in MS.

49The association between inflammatory biomarkers (cytokine levels, or proteins carbonylation and TSPO PET imaging in neurodegenerative disorders was also the subject of several studies. Early stage dementia with Lewy bodies (DLB) and PD patients exhibited significantly greater levels of total carbonylated proteins in cerebrospinal fluid (CSF) than healthy controls, suggesting that relative aspecific oxidative modifications occurred in both neurodegenerative disorders. The lack of correlation between central inflammation measured by 18F-FEPPA VT and levels of several pro-inflammatory cytokines such as IL-1β, IL-6 and TNF-α supports the idea that microglial activation can occur in MDE even when peripheral inflammation is absent. In patients with recent onset of SCZ, the absence of this association has also been found between 11C-DPA713 VT and both plasma and cerebrospinal fluid IL-6 levels.

50As discussed above, TSPO PET imaging has been used in a large number of studies concerning neurodegenerative diseases over the past 20 years. However, such a major investment has not yet resulted in an expected impact for clinical practice. In AD or SCZ, for instance, our present analysis came to conflicting results, which are moreover based on small numbers (<100) of patients. The overall trend that could be drawn from these studies would suggest that early microglial activation could represent a beneficial response, whereas chronic exposure would induce detrimental effects by promoting neuronal death, thus contributing to the progression of the disease.

51A comparison between PET radioligands’ development, targeting TSPO and β-amyloid (Aβ) plaques, respectively, illustrates this assessment: while the first clinical uses of these radiopharmaceuticals were approximately contemporaneous, those of 11C-PK1195 being even much earlier than those of 11C-PiB and 18F-FDDNP (two of the first developed Aβ tracers), several Aβ-radiopharmaceuticals are nowadays routinely used to assess amyloid-β burden in a clinical setting. Aβ-PET represents a major advance in the evaluation of patients with cognitive impairment when AD is a possible cause, and affects the diagnosis and the resulting treatment recommendations. TSPO radiopharmaceuticals (18F-DPA714, 18F-FEPPA, 18F-PBR06, or 18F-GE180 being among the most promising ones), remain at the same time dedicated to specialized PET centers and restricted to research applications.

52The current state of knowledge, which clearly identifies the neuroinflammation process as a constitutive part of the pathophysiology of neurodegenerative diseases, might give a new impetus for TSPO PET imaging, especially for the noninvasive and longitudinal evaluation of anti-inflammatory therapies for these diseases. For instance, Ratchford et al. recently reported a longitudinal TSPO PET study evaluating microglial activation in MS. One year treatment with glatiramer acetate reduced TSPO binding significantly in both cortical gray matter and cerebral white matter, when compared with cerebellum. This promising approach gives excellent comparative data for longitudinal TSPO PET studies evaluating therapeutics in progressive diseases. Nevertheless, several drawbacks limit the ability of TSPO as a PET imaging molecular target in routine clinical practice, such as the identified polymorphism affecting the binding affinity properties of most of TSPO ligands and the TSPO multicellular expression, among others. Thus, we mentioned above the interest of TSPO imaging combined with further PET targets, but the dissimilar binding profile observed and its correlation with neuroinflammation cellular dynamics remains a major issue for understanding the pathophysiological and clinical relevance of what we are imaging.

53TSPO PET tracers were the most widely used and allowed us to better characterize microglia-related neuroinflammation in terms of intensity, localization and static/dynamic feature of microglial activation in various neurodegenerative troubles. Recent studies challenged the paradigm that activated microglia are detrimentally involved in neurodegenerative process. Indeed, microglia seem to play a role in CNS disorders with a relative dichromatic status that occurs in a time-dependent fashion. With the benefit of hindsight over 25 years, TSPO PET imaging showed several limitations in relation to the multicellular expression of the target or its similar expression, whatever the microglia polarization state. Thus, new PET biomarkers reflecting some specific functions of activated microglia (e.g., polarization, microglial oxidative stress-related) conjugated with a multiplex strategy combining several modalities (imaging, biology) may solve clinicians’ concerns about characterizing the role of activated microglia over the progression of CNS disorders.

54Finally, the usefulness of TSPO PET imaging as a major research tool has appeared inescapable over years. Indeed, it has allowed for better understanding of the neuroinflammation process in CNS disorders, and the multiplex approach, either unimodal or multimodal, would certainly broaden the knowledge of this issue. Nowadays, however, the main challenge remains to identify the scope of TSPO PET imaging for a transfer in routine clinical use whether as a diagnostic, prognostic or therapeutic effects assessment tool.