A Systematic Review on Dementia and Translocator Protein (TSPO): When Nuclear Medicine Highlights an Underlying Expression

Background: Translocator protein (TSPO) is a neuroinflammation hallmark. Different TSPO affinity compounds have been produced and over time, the techniques of radiolabeling have been refined. The aim of this systematic review is to summarize the development of new radiotracers for dementia and neuroinflammation imaging. Methods: An online search of the literature was conducted in the PubMed, Scopus, Medline, Cochrane Library, and Web of Science databases, selecting published studies from January 2004 to December 2022. The accepted studies considered the synthesis of TSPO tracers for nuclear medicine imaging in dementia and neuroinflammation. Results: A total of 50 articles was identified. Twelve papers were selected from the included studies’ bibliographies and 34 were excluded. Thus, 28 articles were ultimately selected for quality assessment. Conclusion: Huge efforts in developing specific and stable tracers for PET/SPECT imaging have been made. The long half-life of 18F makes this isotope a preferable choice to 11C. An emerging limitation to this however is that neuroinflammation involves all of the brain which inhibits the possibility of detecting a slight inflammation status change in patients. A partial solution to this is using the cerebellum as a reference region and developing higher TSPO affinity tracers. Moreover, it is necessary to consider the presence of distomers and racemic compounds interfering with pharmacological tracers’ effects and increasing the noise ratio in images.


Introduction
Translocator protein (TSPO), previously designated as peripheral benzodiazepine receptor (PBR), is a protein on the outer mitochondrial membrane with five transmembrane helical domains combined with a voltage-dependent anion channel and a nucleoside carrier. It has been well conserved throughout evolution and is expressed both by prokaryotes and eukaryotes. TSPO binds benzodiazepines, namely RO5-4864 and derivates of isoquinoline carboxamide including PK11195 [1]. In fact, this protein belongs to the family of benzodiazepine receptors which can be classically distinguished into two types of receptors: the central benzodiazepine receptor (CBR) localized in the central nervous system, through which benzodiazepines exert their pharmacological effects [2], and the peripheral benzodiazepine receptor (PBR), or peripheral benzodiazepine binding site (PBBS) (Schoemaker et al. 1981), the object of this systematic review. TSPO can be found in steroid cells as well as in the central nervous system. It is ubiquitous in cerebral phagocytic cells (microglia), which are the primary immune cells in the central nervous system, but it can also be found in the heart, lungs, kidneys, and liver as well [3][4][5][6]. The primary function of TSPO is its involvement in cholesterol transportation through the inner mitochondrial membrane for

Search Results
The research produced a total of 50 articles. A total of 36 articles was identified from PubMed and 14 from Scopus, while no studies were detected in the Cochrane Library, Medline, or Web of Science. Five duplicate records were removed. The references of the selected studies were examined to check for any additional relevant articles and 12 papers were identified. A total of 57 articles was screened by examining each abstract to identify studies with potential relevance. From the overall group of 57, 29 articles were excluded because they did not satisfy the inclusion criteria. Among them, 15 were reviews, two articles were related only to traumatic injury and 12 studies were not related. The remaining 28 articles were included and selected for quality assessment. The search strategy and selection criteria applied are represented in a flowchart (see Figure 1).

Study Characteristics
The main thematic areas of the selected studies can be summarized as follows: (1) studies of new radiopharmaceuticals with an animal model; (2) biodistribution studies conducted on animals and humans; and (3) studies performed only on humans. The majority of the selected articles were conducted by European researchers, followed by Asian, American, and Australian. Among the selected studies, 15 experiments on tracer biodistribution were performed on animals, six both on humans and animals, and seven on humans only.

Methodological Quality Assessment
The methodological quality of the papers included studies of a very high quality. Of the 28 selected studies, only two did not satisfy all the QUADAS-2 domains (see Table 1). From analyzing the results within each bias assessment domain independently (see Table 2), almost all studies obtained a low concern of bias and no more than one study showed high risk in one domain. Considering all four bias assessment domains, three studies reported unclear results, while one study had high-risk results. Regarding patient selection, only two studies reported unclear results, inflow, and timing: one study had unclear results, while in the references and standards used, only one study was high-risk, due to an insufficient number of details. A low concern of applicability of was found for all the studies (See Figure 2).

Study Characteristics
The main thematic areas of the selected studies can be summarized as follows: (1) studies of new radiopharmaceuticals with an animal model; (2) biodistribution studies conducted on animals and humans; and (3) studies performed only on humans. The majority of the selected articles were conducted by European researchers, followed by Asian, American, and Australian. Among the selected studies, 15 experiments on tracer biodistribution were performed on animals, six both on humans and animals, and seven on humans only.

Methodological Quality Assessment
The methodological quality of the papers included studies of a very high quality. Of the 28 selected studies, only two did not satisfy all the QUADAS-2 domains (see Table 1). From analyzing the results within each bias assessment domain independently (see Table  2), almost all studies obtained a low concern of bias and no more than one study showed high risk in one domain. Considering all four bias assessment domains, three studies reported unclear results, while one study had high-risk results. Regarding patient selection, only two studies reported unclear results, inflow, and timing: one study had unclear results, while in the references and standards used, only one study was high-risk, due to an insufficient number of details. A low concern of applicability of was found for all the studies (See Figure 2).

Included
Additional records identified through the reference of selected studies (n = 12)
Chauveau et al. radiolabeled SSR180575 with 11 C ( Figure 3) and performed in vivo and in vitro imaging using a model of rodent acute neuroinflammation comparing with [ 11 C](R)-PK11195 [24]. They injected 0.5 µL of (R,S)-α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic (AMPA) into the right striatum of anaesthetized rats. Higher uptake of [ 11 C]SSR180575 in the AMPA-lesioned striatum was observed. In the right striatum, the uptake was higher for [ 11 C]SSR18057 than [ 11 C](R)-PK11195 while the uptake in the contralateral striatum was lower for [ 11   Another study by Chauveau et al. [26], compared 11 , two molecules that belong to the pyrazolopy rimidine class (see Figure 3). The study was conducted in vivo and in vitro with a murin model of neuroinflammation. In vitro, 11 C-DPA-713 and 18 F-DPA-714 had a similar capa bility in the detection of neuroinflammatory areas and a similar signal-to-noise ratio. In vivo, the fluorinated compound had the highest ratio of ipsilateral to contralateral uptak and a grated greater binding affinity compared to 11 C-DPA-713 and 11 C-PK11195.
Another intriguing study is the research of Vignal et al. [27]. They applied the opti mized [ 18 F]FEPPA (see Figure 4) in a mouse model of cerebral inflammation elicited by th intraperitoneal injection of Salmonella enterica serovar typhimurium lipopolysaccharide preceding a 24 h PET scan. The tracer was synthesized through nucleophilic substitution with a tosylated precursor. The tracer distribution was dependent on TSPO mouse expres sion as well as in the heart and kidneys, as observed during small animal PET acquisition Western blotting showed a 2.2-fold greater expression of TSPO in the brain of treated mic than in the control. Another study by Chauveau et al. [25], compared 11 , two molecules that belong to the pyrazolopyrimidine class (see Figure 3). The study was conducted in vivo and in vitro with a murine model of neuroinflammation. In vitro, 11 C-DPA-713 and 18 F-DPA-714 had a similar capability in the detection of neuroinflammatory areas and a similar signal-to-noise ratio. In vivo, the fluorinated compound had the highest ratio of ipsilateral to contralateral uptake and a grated greater binding affinity compared to 11 C-DPA-713 and 11 C-PK11195.
Another intriguing study is the research of Vignal et al. [26]. They applied the optimized [ 18 F]FEPPA (see Figure 4) in a mouse model of cerebral inflammation elicited by the intraperitoneal injection of Salmonella enterica serovar typhimurium lipopolysaccharides preceding a 24 h PET scan. The tracer was synthesized through nucleophilic substitution with a tosylated precursor. The tracer distribution was dependent on TSPO mouse expression as well as in the heart and kidneys, as observed during small animal PET acquisition.
Western blotting showed a 2.2-fold greater expression of TSPO in the brain of treated mice than in the control.
Another intriguing study is the research of Vignal et al. [27]. They applied t mized [ 18 F]FEPPA (see Figure 4) in a mouse model of cerebral inflammation elicite intraperitoneal injection of Salmonella enterica serovar typhimurium lipopolysacc preceding a 24 h PET scan. The tracer was synthesized through nucleophilic subs with a tosylated precursor. The tracer distribution was dependent on TSPO mouse sion as well as in the heart and kidneys, as observed during small animal PET acq Western blotting showed a 2.2-fold greater expression of TSPO in the brain of treat than in the control.   [28]. A dynamic PET scan of 100 min was executed. After cerebellum and basal ganglia uptake were evaluated with a complete washout in  [29]. After 30 h post-injection, a plateau in the adrenal glands was seen. Maximal activity was in the heart, lungs, liver, spleen, and kidneys at 5-15 min post tracer administratio blood concentration was very low during the exam. Higher activity was detecte olfactory bulbs compared to the rest of the brain. Thyroid uptake rose from 5 m post-injection, then diminished at 3 h to 4%, remaining steady until 6 h post-i Competition studies conducted of PK11195 and Ro 5-4854 confirmed the spec CLINME binding to TSPO linking sites.  Figure 6) in an AMPA-induced excitotoxic murine model through SPECT imaging [28]. After 30 min to 6 h post-injection, a plateau in the adrenal glands was seen. Maximal activity was reached in the heart, lungs, liver, spleen, and kidneys at 5-15 min post tracer administration, while blood concentration was very low during the exam. Higher activity was detected in the olfactory bulbs compared to the rest of the brain. Thyroid uptake rose from 5 min to 1 h post-injection, then diminished at 3 h to 4%, remaining steady until 6 h post-injection. Competition studies conducted of PK11195 and Ro 5-4854 confirmed the specificity of CLINME binding to TSPO linking sites. h post-injection, a plateau in the adrenal glands was seen. Maximal activity was reached in the heart, lungs, liver, spleen, and kidneys at 5-15 min post tracer administration, while blood concentration was very low during the exam. Higher activity was detected in the olfactory bulbs compared to the rest of the brain. Thyroid uptake rose from 5 min to 1 h post-injection, then diminished at 3 h to 4%, remaining steady until 6 h post-injection. Competition studies conducted of PK11195 and Ro 5-4854 confirmed the specificity of CLINME binding to TSPO linking sites. Tran et al. tested a variety of TSPO ligands (11a-c and 13a-d) with a 2-phenylpyrazolo[1,5-a]pyrimidin-3-yl acetamide structure by conducting an in vitro binding assay [30]. A major part of the compound had a better affinity than DPA-714, with particular affinity and lipophilicity in compound 11a. It was radiolabeled with 18 F and PET imaging was conducted in an LPS-induced neuroinflammatory murine model. [ 18 F]11a demonstrated a high affinity for inflammation sites, confirmed by an immunohistochemical exam conducted on the dissected brain, which showed how the PET uptake was in correspondence with areas of activated microglia.
Quiao et al. conducted a synthesis study on compound 6 (GE387) radiolabeled with 18 F [16]. The racemic [ 18 F]6 was studied in male Wistar rats: the tracer showed low Tran et al. tested a variety of TSPO ligands (11a-c and 13a-d) with a 2-phenylpyrazolo[1,5a]pyrimidin-3-yl acetamide structure by conducting an in vitro binding assay [29]. A major part of the compound had a better affinity than DPA-714, with particular affinity and lipophilicity in compound 11a. It was radiolabeled with 18 F and PET imaging was conducted in an LPS-induced neuroinflammatory murine model. [ 18 F]11a demonstrated a high affinity for inflammation sites, confirmed by an immunohistochemical exam conducted on the dissected brain, which showed how the PET uptake was in correspondence with areas of activated microglia.
Qiao et al. conducted a synthesis study on compound 6 (GE387) radiolabeled with 18 F [16]. The racemic [ 18 F]6 was studied in male Wistar rats: the tracer showed low sensitivity to the polymorphism rs6971 in human embryonic kidney cell lines and modest affinity in murine brains because the animals were healthy.
Pike et al. synthesized a library of new 2-phenylindol-3-ylglyoxylamide derivatives of the general formula II (compounds 19-31), attributing a methyl group on the indole nitrogen. Then, they tested ligand 31 (Figure 7) which, thanks to its methyl group, was feasible to label with carbon-11 for positron emission tomography (PET) imaging in monkeys [30]. Peak radioactivity was observed in a region rich in TSPO 40 min after tracer injection, with a maximal accumulation in putamen between 12 and 32 min. After the pre-blocked administration of [ 11 C]PK 1119519 ([ 11 C]1, the rapid uptake in all the examined regions and rapid washout demonstrated the specific binding of [ 11 C]31. PET images at 4-100 min post-injection demonstrated a high uptake in the putamen and cerebellum and intermediate levels of accumulation in the cortical regions. Ligand 31 exhibited high-affinity binding in HABs and MABs, which was lower for low-affinity binders even if the authors suggested that further studies in the larger group were needed to establish if differences between binding toward MABs and HABs could be revealed. DAA1106 (Figure 8), a molecule with high binding selectivity for cerebral rat and monkey mitochondria with a weak affinity for CBR, GABA, and Kappa1 receptors, was tested by Kumata et al. [31]. They labeled DAA1106 with fluorine-18 in a previous study in 2007 through the fluorination of diphenyliodonium salt used as a precursor, but, due to the instability of the diphenyl iodonium precursor, the clinical use of the compound was not feasible. This was until the possibility of adding nucleophilic 18 F into an electron-rich benzene ring, Cu-mediated radiofluorination of pinacol arylboronic ester, or the use of arylstannane precursors with [ 30 and 60 min after tracer injection showed high uptake in the heart, lungs, spleen, and kidneys and, interestingly, a low uptake in bone, which indicated that no defluorination had occurred thanks to the direct introduction of fluorine-18 in the benzene ring. The polar metabolite generated in rats did not pass the blood-brain barrier or, if penetrated, did not remain in brain tissue due to its hydrophilicity. For the PET study, an ischemic rat brain model was used and demonstrated that [ 18 F]DAA1106 had a high binding specificity for TSPO-expressing ischemic areas. affinity in murine brains because the animals were healthy.
Pike et al. synthesized a library of new 2-phenylindol-3-ylglyoxylamide derivatives of the general formula II (compounds [19][20][21][22][23][24][25][26][27][28][29][30][31], attributing a methyl group on the indole nitrogen. Then, they tested ligand 31 (Figure 7) which, thanks to its methyl group, was feasible to label with carbon-11 for positron emission tomography (PET) imaging in monkeys [31]. Peak radioactivity was observed in a region rich in TSPO 40 min after tracer injection, with a maximal accumulation in putamen between 12 and 32 min. After the preblocked administration of [ 11 C]PK 1119519 ([ 11 C]1, the rapid uptake in all the examined regions and rapid washout demonstrated the specific binding of [ 11 C]31. PET images at 4-100 min post-injection demonstrated a high uptake in the putamen and cerebellum and intermediate levels of accumulation in the cortical regions. Ligand 31 exhibited high-affinity binding in HABs and MABs, which was lower for low-affinity binders even if the authors suggested that further studies in the larger group were needed to establish if differences between binding toward MABs and HABs could be revealed. DAA1106 (Figure 8), a molecule with high binding selectivity for cerebral rat and monkey mitochondria with a weak affinity for CBR, GABA, and Kappa1 receptors, was tested by Kumata et al. [32]. They labeled DAA1106 with fluorine-18 in a previous study in 2007 through the fluorination of diphenyliodonium salt used as a precursor, but, due to the instability of the diphenyl iodonium precursor, the clinical use of the compound was not feasible. This was until the possibility of adding nucleophilic 18 F into an electronrich benzene ring, Cu-mediated radiofluorination of pinacol arylboronic ester, or the use of arylstannane precursors with [ 18 F]F− and high radiochemical yields were introduced [33][34][35][36]. In their later paper, Kappal et al. ventured into the synthesis of [ 18 F]DAA1106 via the [ 18 F]fluorination of a spirocyclic iodonium ylide (SCIDY), compound (1), used as a precursor with [ 18 F]F−. The SCIDY precursor was demonstrated to be stable at room temperature for 60 days. DAA1106 was stable after radiolabeling with [ 18 F] for the duration of at least one PET scan. Biodistribution studies executed on rats al 30 and 60 min after tracer injection showed high uptake in the heart, lungs, spleen, and kidneys and, interestingly, a low uptake in bone, which indicated that no defluorination had occurred thanks to the direct introduction of fluorine-18 in the benzene ring. The polar metabolite generated in rats did not pass the blood-brain barrier or, if penetrated, did not remain in brain tissue due to its hydrophilicity. For the PET study, an ischemic rat brain model was used and demonstrated that [ 18 F]DAA1106 had a high binding specificity for TSPO-expressing ischemic areas. In 2012, in another study by Wadsworth et al., fluorine-18 analogues of DAA1106 were developed [37]. DAA1106 had a high affinity and less lipophilicity with better brain biodistribution. They valued the structure-activity relationships of the analogues by varying their aromatic rings. The reduction of their nitro groups was performed, followed by the reduction of their alanine. Acetylation or fluoroacetylation allowed them to obtain the desired molecule which was tested for affinity screening. Less lipophilicity was obtained by the substitution of phenyl groups with aliphatic groups. The aromatic rings, instead, gave rise to compounds with a higher affinity. The biodistribution of the compounds was studied in naïve Wistar rats. They had a higher whole brain uptake when compared to PK11195 but had a good capacity in discriminating high TSPO-expressing areas and lower ones. In vivo metabolic stability studies stated that compounds 27 and 30 reached the target. In a platelet binding assay however, the question of the affinity of the two binding sites (high affinity, low affinity) of TSPO ligands was not answered.

Studies Conducted on Animals and Humans
Some of the selected studies were performed on animals but also on humans for in In 2012, in another study by Wadsworth et al., fluorine-18 analogues of DAA1106 were developed [35]. DAA1106 had a high affinity and less lipophilicity with better brain biodistribution. They valued the structure-activity relationships of the analogues by varying their aromatic rings. The reduction of their nitro groups was performed, followed by the reduction of their alanine. Acetylation or fluoroacetylation allowed them to obtain the desired molecule which was tested for affinity screening. Less lipophilicity was obtained by the substitution of phenyl groups with aliphatic groups. The aromatic rings, instead, gave rise to compounds with a higher affinity. The biodistribution of the compounds was studied in naïve Wistar rats. They had a higher whole brain uptake when compared to PK11195 but had a good capacity in discriminating high TSPO-expressing areas and lower ones. In vivo metabolic stability studies stated that compounds 27 and 30 reached the target. In a platelet binding assay however, the question of the affinity of the two binding sites (high affinity, low affinity) of TSPO ligands was not answered.

Studies Conducted on Animals and Humans
Some of the selected studies were performed on animals but also on humans for in vitro examination, in vivo biodistribution, and metabolism studies or post-mortem binding examination. 11 C-PBR28 was demonstrated to be a valid radiopharmaceutical for the detection of glial response in a study by Donat et al. [9]. A transgenic mouse model of AD was used. The tracer was injected in the tail vein through a cannula and 60 min of dynamic PET scanning was executed. The activity of 370 MBq was also administered to AD patients and then 90 min PET images were acquired. CT scans were obtained for attenuation, scatter correction, and to add morphostructural information to the PET images. To improve the spatial resolution of the PET results, autoradiography with 3H-PBR28 and immunochemistry were executed to provide a tracer density of PBR28 binding sites in mouse brains. During the in vivo characterization of TSPO density through autoradiography, higher microglial representation was associated with higher 3H-PBR28 binding, with major representation in the cortical and hippocampal regions in the AD mouse model. Arlicot [36]. Arterial and venous samples were taken while two volunteers underwent whole-body acquisition 1 h after the tracer's administration. PET imaging post-processing showed a maximum cerebral uptake in the pons and cerebral radioactive peak within 5 min with a rapid clearance between 5 and 30 min. In whole-body scans, high uptake was seen in the vertebral bodies, gallbladder, heart wall, spleen, intestinal wall, kidneys, and adrenal gland. They also conducted biodistribution studies on mice highlighting similar results.
Starting with the tetracyclic indole-based pharmacophore described in the previously reported by Okubo [37]. It demonstrated in rats a high specific binding, a relevant brain uptake with significant uptake and retention in the olfactory bulb, a region rich in TSPO expression, and a good clearance from the region with low TSPO density, that is, the striatum.
Chau et al. tested the higher affinity of the S-enantiomeric tricyclic indole compound, [ 18 F]GE-180, a third-generation tracer for the detection of TSPO [19]. In the Wistar rat heart, the S-enantiomer was demonstrated as having a higher affinity with a Ki of 0.87 nM, while the R-GE180 enantiomer had a Ki equal to 3.87 nM. Analogous results were demonstrated in humans (human colonic cell membranes). S-enantiomer radiolabeled with 18 F had a rapid clearance from blood followed by racemate and R-[ 18 F]GE-180. Brain retention in rats was higher in S-radiolabeled enantiomers than R-enantiomers at 10 and 30 min post-injection. S-[ 18 F]GE-180 also had a greater lung uptake than the R-enantiomer and racemate and was demonstrated to be stable in vivo without conversion to a distomer. Nag [38]. During the autoradiography assay, homogeneous binding between the cortical and subcortical regions was seen in whole hemisphere human brain slices of healthy subjects. Two cynomolgus monkeys underwent a PET scan and a rapid tracer accumulation was observed in the first 4 min with a decrease in around 20 min. [ 18 F]fluorovinpocetine had a good brain penetration and similar brain distribution pattern to [ 11 C]vinpocetine (high values in the thalamus and striatum, low in the cerebellum), but smaller regional differences.
Damont et al. synthesized a group of pyrazolo [1,5-a]pyrimidines, related to 2, DPA-714 (compound 2 in Figure 3 is represented labeled with 18 F) to test the in vitro binding affinity of TSPO [39]. Fluroalakyl-and fluoroalakynyl-analogues were created via Sonogashira coupling reactions. In competition experiments against [3H]1 ([3H]-PK11195) in a membrane rat heart, all compounds had a subnanomolar affinity compared to compound 2, but compound 12 of the fluoroalkyl series and compound 23 from the alkynyl series had the lowest Ki. The specificity of linking was exquisite in all compounds except for 29 and 30, which showed a modest affinity for central benzodiazepine receptor (CBR). New compounds (12, 21-24, and 28-30) have undergone oxidative metabolism investigation in humans, rats, and mouse hepatic microsome assays. In murine microsomes, 90% of biotransformation occurred in 20 min, while in human microsomes, the process was variable in different molecules (31% in 2, 91% in 29). In general, alkynyl derivatives were less metabolized than alkyl compounds. Oxidation did not generate fluoroacetate and so non-brain penetrant species could degrade PET image quality. On the basis of these results, compounds 12 and 23 were chosen for fluorine-18 radiolabeling. Wistar rats underwent PET scans with [ 18 F]12 and [ 18 F]23 7 days after AMPA-induced brain inflammation in the right striatum. Selective tracer uptake in the right striatum was seen 2 min post tracer injection and maintained for 60 min even though it became slightly lower. A better contrast was obtained with [ 18 F]23 due to the rapid washout in the non-lesioned striatum.

Studies Conducted on Humans
In this paragraph, studies conducted on humans are reported, including an in vivo biodistribution evaluation and post-mortem examination. Okello et al. conducted different studies on microglial activation and PET imaging. In their study in 2009, they used the fibrillar amyloid tracer 11 C-PIB (Pittsburgh compound B) and the peripheral benzodiazepine binding site ligand 11 C-(R)-PK11195 to explore the possible correlation between microglial activation and Aβ plaque in amnestic mild cognitive impairment (MCI) [40]. A total of 50% of the included 14 subjects with amnestic MCI had increased amyloid deposition but no correlation between the regional C-11-labeled PK-11195 binding and PIB uptake was seen in this group. The authors interpreted this result as due to the multifactorial activation of microglia as a response to amyloid deposition. 11 C-(R)-PK11195 was used also by Cagnin et al. for Alzheimer's type dementia (AD), including mild and early forms [41]. The in vivo detection demonstrated high levels of tracer binding in these patients, hinting that the microglial response is an event that could happen in the precocious phase of pathogenesis. In 2004, the same group of researchers demonstrated the presence of higher levels of 11 C-(R)-PK11195 in people with frontotemporal lobar degeneration, suggesting microglial activation that went hand-inhand with neuronal loss. Both in AD and frontotemporal dementia, the tracer uptake was independent of the augmented formation of amyloid plaque [41].
Interesting estimations of ([ 18 F]5 were conducted in human brains by Feeney et al. [42] and previously, by Zanotti-Fregonara et al. [43], which stated 18 F-GE180 as an unfavorable tracer for TSPO brain imaging compared with 11 C-PBR28 because of 18 F-GE180's lower brain penetration.
Gulyás et al. [44] performed an in vitro autoradiography on human post-mortem brains of patients who suffered from Alzheimer's disease by evaluating the binding on TSPO of N-(5- . Their research demonstrated effective binding for both tracers with higher uptake in the hippocampus, temporal and parietal cortex, basal ganglia, and thalamus. Through a comparison with healthy age-matched controls, the tracer had high specificity in microglia-activated areas in the Alzheimer's brain, confirmed by immunohistochemical examination.
[ 18 F]GE180 was also studied in a pharmacokinetics study by Fan et al. in older healthy adults, which stated the two-tissue compartment model was the better model to describe brain kinetics and that 90 min was the ideal scan length for a good assessment [45].

Conclusions
TSPO represents a hallmark of neuroinflammation. The data from our systematic review demonstrate the huge effort in developing more and more specific tracers. The long half-life of 18 F (110 min) makes this isotope a preferable choice compared to the shorter life of 11 C (20 min). Moreover, new chemical synthesis technologies have permitted researchers to improve radiotracer production and allowed research to increase the use of radiolabeling procedures. However, an observed limitation to this research was the huge widespread neuroinflammation involvement that hampers the possibility of detecting a slight inflammation status change in patients. This issue finds a partial solution using the cerebellum as a reference region, as observed in the mentioned studies on 11 C-PBR28-TSPO. Improvement in the radiosynthesis and detection of molecules with higher TSPO affinity has permitted research to overcome this issue. To cite some results, [ 11 C]SSR18057 has a better capacity in discriminating healthy tissue compared to [ 11 C](R)-PK11195, while the choice of fluorinated compounds such as 18 F-DPA-714 has made it possible to obtain a greater binding affinity compared to 11 C-DPA-713 and 11 C-PK11195, as demonstrated by Chauveau et al.
Different synthesis strategies have been also exploited to ensure a better stability of the compound: the fluorination of diphenyliodonium salt used as a precursor was demonstrated to not be a feasible approach while the synthesis of [18F]DAA1106 via [18F]fluorination of the SCIDY precursor compound 1 was demonstrated to guarantee a more stable compound for a considerable time. Moreover, [18F]DAA1106 could be used to detect ischemic areas due to its high binding specificity for TSPO. DAA1106 also has a better brain distribution. In particular, the substitution of its phenyl groups with aliphatic groups permits obtaining compounds with less lipophilicity while its aromatic ring, instead, is responsible for higher affinity. Similar consideration could be directed at 18 F-GE180, which is an unfavorable tracer for TSPO brain imaging for lower brain penetration compared with 11 C-PBR28.
The introduction of [123I]-CLINME has been interesting in opening the possibility of conducting TSPO-SPECT imaging. The competition studies conducted with PK11195 and Ro 5-4854 have confirmed the specificity of TSPO binding even if thyroid uptake represents a relevant aspect to consider for dosimetry in humans.
Finally, the necessity to consider the presence of distomers or racemic compounds in the synthesis process should be emphasized, which could interfere with the pharmacodynamic and pharmacokinetic effects of the tracer, generating metabolites incapable of overcoming the blood-brain barrier and/or that could increase the noise ratio in PET or SPECT images. The complexity of these subjects was not evaluated in this review and requires further efforts.