Strategies of positron emission tomography (PET) tracer development for imaging of tau and α-synuclein in neurodegenerative disorders

Abstract

Alzheimer's disease (AD) is a progressive neurodegenerative disorder, characterized by the presence of extracellular amyloid plaques consisting of β-amyloid peptides (Aβ) and intracellular neurofibrillary tangles (NFTs) composed of hyperphosphorylated tau (pTau) protein in the brain. Genetic and animal studies strongly indicate that Aβ, tau and neuroinflammation play important roles in the pathogenesis of AD. Several staging models showed that NFTs correlated well with the disease progression. Positron emission tomography (PET) imaging has become a widely used non-invasive technique to image NFTs for early diagnosis of AD. Despite the remarkable progress made over the past few years, tau PET imaging is still challenging due to the nature of tau pathology and the technical aspects of PET imaging. Tau pathology often coexists with other proteinopathies, such as Aβ plaques and α-synuclein aggregates. Distinguishing tau-specific signals from other overlapping pathologies is difficult, especially in the context of AD, where multiple protein aggregates are present, as well as the spectrum of different tau isoforms (3R and 4R) and conformations. Moreover, tracers should ideally have optimal pharmacokinetic properties to penetrate the blood–brain barrier (BBB) while maintaining specificity, low toxicity, low non-specific binding, rapid uptake and clearance from the brain, and formation of no radiolabeled metabolites in the brain. On the other hand, Parkinson's disease (PD) is a progressive neurodegenerative movement disorder characterized by the abnormal accumulations of α-synuclein in neurons. Heterogeneity and the unclear pathogenesis of PD hinder early and accurate diagnosis of the disease for therapeutic development in clinical use. In this review, while referring to existing reviews, we focus on the design strategies and current progress in tau (NFTs) targeting new PET tracers for AD; evolution of non-AD tau targeting PET tracers for applications including progressive supranuclear paralysis (PSP) and corticobasal degeneration (CBD); new PET tracer development for α-synuclein aggregate imaging in PD and giving an outlook for future PET tracer development.

Advances in the positron emission tomography (PET) tracer development for imaging of tau in Alzheimer's disease (AD) and non-AD, and for imaging of α-synuclein in Parkinson's disease (PD).

Protein aggregation and neurodegenerative diseases

Neurodegenerative disorders, such as Alzheimer's and Parkinson's, represent a group of conditions that are characterized by progressive degeneration of the structure and function of the central nervous system (CNS). Alzheimer's disease (AD), the most common form of dementia, is characterized by memory loss and decline in cognitive function. On the other hand, Parkinson's disease (PD) affects movement, often including tremors due to the degeneration of dopamine-producing neurons. Both disorders share a common thread of protein misfolding followed by neuronal damage leading to the gradual deterioration of brain function.

According to the World Health Organization (WHO), there are about 55 million people worldwide living with dementia.¹ AD is a progressive neurodegenerative disorder and may contribute to 60–70% of all dementia cases worldwide.¹ AD is characterized by the presence of extracellular amyloid plaques consisting of β-amyloid peptides (Aβ) and intracellular neurofibrillary tangles (NFTs) composed of aggregated hyperphosphorylated tau (pTau) protein in the brain.² Genetic and animal studies strongly indicate that Aβ, tau and neuroinflammation play important roles in the pathogenesis of AD.²

Tau, a microtubule-associated protein, encoded by the microtubule-associated protein tau (MAPT) gene, is highly expressed in the human CNS with high enrichment in the neuronal axons. Under normal physiological conditions, tau regulates the assembly and stabilization of microtubules for effective axonal transport and synaptic plasticity.^3,4 However, under abnormal physiological and certain pathological conditions, Aβ plaques along with neuroinflammation could cause aberrant post-translational modifications of tau, mainly phosphorylation, which leads to its disassembly from microtubules, followed by aggregation and redistribution to cell bodies and dendrites leading to neurodegeneration (Fig. 1).^5,6

Fig. 1. Schematic showing shrinkage of Alzheimer's brain and tau protein's association with microtubules and propagation to form NFTs. Yellow dots represent hyperphosphorylation on the tau protein. (This figure was drawn using BioRender).

In general, human tau is differentially expressed in six isoforms during cell development and differentiation as a result of the alternative splicing of pre-mRNA that was generated from the MAPT gene transcript, and is located on the chromosome 17q21.31, and consists of 16 exons.^7–9 The tau exons 2, 3, 4a, 6, 8 and 10 are alternatively spliced following the structured mechanisms of inclusion and exclusion phenomena.¹⁰ However, in the human CNS, six main MAPT transcript variants are generated by the alternative splicing of exons 2, 3, and 10, and then translated into different tau protein isoforms (Fig. 2). The six tau isoforms exist in the range of 352 (55 kDa) to 441 (62 kDa) amino acids depending on the presence or absence of exon 2, exon 3, and exon 10 of the MAPT gene.¹¹ For exons 2 and 3, each encodes 29 amino acids in the N-terminus (termed as ‘N’), while exon 10 encodes 31 amino acids in the C-terminus as it localized in the microtubule-binding domain (MBD) or termed as repeat (R). The inclusion or exclusion of exons 2 and 3 indicates the tau isoform with 2N (both exons 2 and 3 are inserted) or 1N (only exon 2 is inserted) or 0N (neither is inserted), while exon 10 causes the presence of either four repeats (4R, includes exon 10 as ‘R2’) or three repeats (3R, excludes exon 10). Therefore, among six tau isoforms in the human CNS, three isoforms can be represented as ‘four repeat tau’ that include exon 10 (4R, with 0N, 1N and 2N), while the other three isoforms as ‘three repeat tau’ that exclude exon 10 in the microtubule-binding domain (3R, with 0N, 1N and 2N).

Fig. 2. Human MAPT gene splicing pattern and tau isoforms. (This figure was drawn using BioRender).

In healthy humans, the ratio of 4R/3R is approximately 1 : 1. However, this ratio is perturbed in several tauopathies due to the varying extent of hyperphosphorylation on either 3R or 4R tau. Thus, the tauopathies can be categorized based on the presence or absence of 3R or 4R tau isoforms as shown in Table 1.¹² In addition, the binding ability of the hyperphosphorylated tau to the microtubule subsides, hence detaching to self-assemble into various aggregates that include straight filaments (SFs), randomly coiled filaments (RCFs), twisted filaments (TFs) and paired helical filaments (PHFs) which predominantly appear as NFTs (Fig. 1), a hallmark in the AD brains.^13,14

Classification of tauopathies based on the presence or absence of the 3R or 4R tau isoform (PiD: Pick's disease; FTLD-MAPT: frontotemporal lobar degeneration with mutations in the MAPT gene; FTDP-17: frontotemporal dementia with parkinsonism-17; PSP: progressive supranuclear paralysis; CBD: corticobasal degeneration; AGD: argyrophilic grain disease; GGT: globular glial tauopathy; AD: Alzheimer's disease; DS: Down syndrome; PART: primary age-related tauopathy; CTE: chronic traumatic encephalopathy; ARTAG: age-related tau astrogliopathy).

Predominant tau isoform	Tauopathy classification
3R	PiD, FTLD-MAPT, FTDP-17
4R	PSP, CBD, AGD, GGT, FTLD-MAPT, FTdp-A7, ARTAG
3R and 4R	AD, DS, FTLD-MAPT, FTDP-17, PART, CTE

Clinically, NFTs correlate well with the severity of the cognitive decline in AD patients.^15,16 The spread of NFTs in the brain tissue occurs 20 years or later after initial Aβ deposits appear in the brain, and the pattern follows Braak staging.^17–20 Therefore, non-invasive detection and quantification of NFTs using positron emission tomography (PET), a molecular imaging technique, have become a sustained practice for a number of PET research groups over the years for AD diagnosis. PET is a non-invasive functional imaging technique in which a small molecular probe labeled with a radiotracer (a positron-emitting isotope) is used for visualization and measurement of physiological processes. In general, upon administration, the molecular probe binds to a specific target in a living system with high affinity and specificity, and it can be traced and quantified. In this regard, early Aβ targeting PET probes have been developed and the history and evolution of Aβ targeting probes have been reviewed elsewhere.²¹

Some reports show that the Aβ accumulation in the brain does not correlate well with the progression of neurodegeneration/cognitive impairment in AD,²² and Aβ deposits have been shown to accumulate early in the brain at risk of developing AD.²³ In addition, Aβ plaques are found not only in AD patients but also in cognitively unimpaired older people and patients with other neurodegenerative disorders.^24,25 A positive Aβ PET scan alone may not accurately diagnose AD, but may help to increase clinical certainty of AD diagnosis. Therefore, Aβ PET probes were extensively used to diagnose preclinical stages of AD and in the clinical trials of Aβ directing antibody-based drugs aducanumab, lecanemab and donanemab which were approved by the United States Food and Drug Administration (FDA).^26,27 In those AD clinical trials, Aβ PET probes were used to monitor the dose-dependent reduction of brain Aβ plaques after treatment. Therefore, the use of tau PET imaging in conjunction with Aβ PET imaging has become a powerful tool for early diagnosis of AD and to distinguish AD patients from non-AD patients.

Recent advances in the in vivo tau PET staging in conjunction with Braak staging enabled the development of new models for AD biomarkers to stage the AD progression in living people. In contrast, distribution and neuropathological changes of Aβ plaques and NFTs in postmortem brains can be staged using gold standard methods the CERAD²⁸ (Consortium to Establish a Registry for Alzheimer's Disease) score and Braak staging,^17–19 respectively. An early AD biomarker staging model²⁹ by Jack et al. (in 2010) related the disease stage to the AD biomarker abnormality. In their study, they used temporally ordered five main AD biomarkers (Aβ, tau-mediated neuronal injury and dysfunction, brain structure, memory, and clinical function) and proposed that, first Aβ changes occur, then tau, followed by alterations in the brain structure, then memory loss, and finally clinical impairment. The authors depicted disease stages as parallel sigmoidal curves showing occurrence of rapid changes in biomarkers in the early stage and then slow changes as the disease progressed. In the revised model³⁰ reported in 2013 by Jack et al., individuals were indexed by time in the x-axis (not shown here), rather than by the disease stage, which states that the initial Aβ pathophysiology could accelerate the antecedent limbic and brainstem tauopathy.

In this series, recently, Therriault et al.²⁰ have come up with a new model of tau PET imaging using the [¹⁸F]MK6240 probe in conjunction with the Braak staging system showing the variability in magnitude and topography of pathological tau. In this study, instead of plotting biomarker abnormalities vs. time in the x-axis as Jack et al.³⁰ have done, Therriault et al. have plotted PET based Braak staging in the x-axis. They have devised the PET based version of Braak staging connecting with AD biomarkers (Fig. 3) for 324 living individuals in a TRIAD cohort. The biomarkers include Aβ, pTau (pTau₁₈₁, pTau₂₁₇, pTau₂₃₁ and pTau₂₃₅) in cerebrospinal fluid (CSF), pTau in plasma (pTau₁₈₁ and pTau₂₃₁), neurodegeneration and cognitive impairment. The cohort includes: 179 cognitively unimpaired people, 80 people with mild cognitive impairment and 65 people with advanced AD. Based on the tau PET [¹⁸F]MK6240 tracer distribution in each participant, they identified the nonlinear biomarker trajectories in relation to the topography of cerebral tau pathology (Fig. 3), and concluded that: 1) as per the spatial extent of tau PET, the progression of tau tangles follows Braak pathological staging; 2) early tau biomarkers arise in CSF at Braak stage II; 3) pTau progression beyond stage II or later requires Aβ accumulation as a prerequisite. This study establishes a model of Braak staging using topological information from tau PET imaging, and to monitor the severity of AD in living individuals. Therefore, further advancements in new PET probe development for imaging at picomolar concentrations with high affinity and specificity are still necessary for effective early diagnosis and clinical monitoring of AD.

Fig. 3. AD biomarker abnormalities in relation to the topography of cerebral tau pathology showing biomarker abnormalities vs. PET-based Braak staging in the x-axis (taken from the original article by Therriault et al., Nature Aging, 2022, 2, 526–535, under Creative Commons Attribution 4.0 International License).

In this review, we emphasize the history of tau PET tracer development (after 2015) focusing on the design and new synthetic approaches to PET tracers in continuation of existing reviews^12,31,32 on PET probe design. We will also discuss the evolution of 4R-tau targeting PET tracer development and its current status. At the end, we would also review the cryo-EM structural determination of tau filaments with PET ligands, and recent advances in α-synuclein PET probe design and development for imaging in Parkinson's disease (PD).

Early Aβ-directing PET probes for AD diagnosis and FDA approval

There are several reviews published discussing the early development of Aβ-directing PET probes and their FDA approval for non-invasive imaging of AD in the human brain.^12,31 Initial Aβ-directing PET probe Pittsburgh compound B (PiB) or [¹¹C]PiB is a radioactive analogue of a very well-known histological staining agent thioflavin-T. [¹¹C]PiB has been the most widely used PET tracer for Aβ imaging. However, due to its short half-life (¹¹C, t_1/2 = 20.3 min), the use of [¹¹C]PiB was limited to centers with cyclotron and radiochemistry facilities on-site, and it was not usable in the context of multi-center clinical trials. Therefore, tracers with longer half-life (for example: ¹⁸F, t_1/2 = 109.8 min) were discovered and successfully used to image Aβ plaques non-invasively. In this series, [¹⁸F]-florbetapir (Amyvid, aza-stilbene derivative), [¹⁸F]-flutemetamol (Vizamyl) and [¹⁸F]-florbetaben (Neuraceq, a stilbene derivative) were developed and approved by the US FDA in 2012, 2013 and 2014 respectively for intravenous use as diagnostic tools for Aβ imaging in AD.

Early tau PET tracer development and FDA approval of AV-1451 (Tauvid™)

2-(1-(6-((2-(Fluoro-¹⁸F)ethyl)(methyl)amino)naphthalen-2-yl)ethylidene)malononitrile ([¹⁸F]-FDDNP) (Fig. 5), a naphthalene-core based compound, was the first PET radiotracer developed (in 2002) to determine the localization of NFTs and Aβ plaques in the brains of living AD patients.³³ This compound was discontinued from further development due to poor selectivity between Aβ and NFTs, and low in vivo PET signals in AD patients. Therefore, to improve NFT selectivity, quinoline-core based PET tracers (S)-[¹⁸F]THK-5117 and (S)-[¹⁸F]THK-5351 were developed³⁴ (Fig. 5) from the initial BF series of [¹¹C]BF-158 (structures are not shown here).³⁵ In the THK series, enantiomeric (S)-[¹⁸F]THK-5117 was found to have better pharmacokinetics properties and signal to noise ratio compared to the (R)-enantiomer. Further refinement of the structure (S)-[¹⁸F]THK-5117 with the introduction of a pyridine moiety in place of the phenyl ring resulted in a new and more hydrophilic PET tracer (S)-[¹⁸F]THK-5351. This new PET tracer (S)-[¹⁸F]THK-5351 displayed high binding affinity for hippocampal regions of the brain homogenates from AD brains and fast white-matter tissue washout compared to (S)-[¹⁸F]THK-5117.³⁴ Autoradiography of brain sections showed that (S)-[¹⁸F]THK-5351 had higher selectivity to NFTs and a higher S/N ratio than (S)-[¹⁸F]THK-5117. In addition, (S)-[¹⁸F]THK-5351 exhibited favorable pharmacokinetics in mice without defluorination. (S)-[¹⁸F]THK-5351 also displayed faster kinetics, higher contrast, and faster white matter washout than (S)-[¹⁸F]THK-5117 in first-in-human PET studies in AD patients. However, it was later discontinued from further development as a tau PET tracer in patients with AD, which is due to its off-target binding and high affinity for monoamine oxidase-B (MAO-B). Since several reviews have been published on these PET tracers, in this review, we will focus on the design strategies, and new and improved synthetic approaches of PET tracers for clinical production.

Fig. 5. Structures of [¹⁸F]FDDNP, (S)-[¹⁸F]THK-5117, (S)-[¹⁸F]THK-5351, Tauvid™ “a.k.a. [¹⁸F]AV-1451, [¹⁸F]T807 or [¹⁸F]flortaucipir”, and [¹⁸F]AV-680 or [¹⁸F]T808.

Tauvid™ development, FDA approval and clinical production

[¹⁸F]AV-1451 or Tauvid™ or [¹⁸F]flortaucipir was initially discovered as a tau PET tracer at Molecular Imaging Biomarker Research, Siemens Medical Solutions, led by H. C. Kolb, and then licensed to Avid Radiopharmaceuticals, a subsidiary of Eli Lilly. The discovery campaign was initiated with the autoradiography screening of human AD brain sections with over 800 compounds resulting in two lead series containing benzimidazoles and carbazoles. Structure–activity relationship (SAR) studies and further optimization led to the discovery of two lead compounds [¹⁸F]AV-1451 (¹⁸F-T807)³⁶ and [¹⁸F]AV-680 (¹⁸F-T808).³⁷ Initial first-in-human studies of [¹⁸F]AV-1451 and [¹⁸F]AV-680 revealed high affinity for tau in the cortical region which correlated well with the Braak staging.³⁸ Both of these tau PET tracers displayed over 25-fold selectivity for NFTs (in the nanomolar range) over Aβ without off target binding to white matter, but [¹⁸F]AV-680 showed superior pharmacokinetics compared to [¹⁸F]AV-1451. However, ¹⁸F accumulation in the skull from the metabolic defluorination of [¹⁸F]AV-680 resulted in its discontinuation from further development. Therefore, the more metabolically stable PET tracer [¹⁸F]AV-1451 was advanced into further development and eventually approved by the FDA (in 2020) as a first-in-class imaging agent to image tau pathology in patients being evaluated for Alzheimer's disease. Several other reviews have discussed the development of [¹⁸F]AV-1451.^12,31,32 Since its approval by the FDA, [¹⁸F]AV-1451 has been widely used in AD therapeutic drug development trials as a tau PET imaging agent along with Aβ PET tracers to monitor patients' response to experimental AD drugs, such as recent aducanumab (a monoclonal antibody developed by Biogen-Eisai as a medication to treat AD). Therefore, its widespread use as a PET imaging agent and the development of robust and new methods for large scale synthesis and routine clinical production have been of interest for many PET research groups across the world. In 2013, Shoup et al. reported³⁹ a concise, one-step radiosynthesis of [¹⁸F]T807 a.k.a. [¹⁸F]AV-1451 using a GE TRACERlab™ FX_FN radiosynthesis module. The synthesis involves the use of a more soluble, N-Boc protected nitro precursor, which can directly be radiolabeled by the nucleophilic fluorination reaction with potassium cryptand [¹⁸F]fluoride (K[¹⁸F]/K₂₂₂) in DMSO at 130 °C for 10 min with concurrent N-Boc deprotection as this method offers a fast one-pot synthesis (Fig. 6). In this series, an updated or alternative synthesis of [¹⁸F]AV-1451 was published by Gao et al.⁴⁰ and Mossine et al.⁴¹ in 2015 and 2017, respectively (not discussed here).

Fig. 6. An updated radiosynthesis of [¹⁸F]AV-1451 for clinical production by Shoup et al.³⁹.

An alternative and high-yield automated production method of [¹⁸F]AV-1451 under GMP-compliance by using the commercially available Synthra RNplus Research module was reported by Jiang et al. in 2021.⁴² In their method, a trimethylammonium precursor (AV 1622) was used instead of an N-Boc protected nitro derivative (as shown in Fig. 6) as a leaving group to improve the HPLC purification process after radiolabeling, and full quality control tests were conducted under GMP compliance for clinical use. Automated radiosynthesis (in the RNplus Research module) involves nucleophilic radiofluorination of AV-1622, acid deprotection of the Boc group followed by purification of [¹⁸F]AV-1451 on a semipreparative HPLC column and the final formulation via solid phase extraction (SPE) yielding [¹⁸F]AV-1451 with high radiochemical purity for clinical use (Fig. 7).

Fig. 7. Synthetic scheme for automated production of [¹⁸F]AV-1451 by Jiang et al.⁴² (RCY = radiochemical yield; RCP = radiochemical purity).

[¹¹C]PBB3 development and its updated synthesis

In 2013, Maruyama et al. reported a new class of compounds bearing phenyl/pyridinyl-butadienyl-benzothiazoles/benzothiazoliums (PPBs) as a new class of PET radiotracers, for visualizing the tau aggregates in living brains of AD patients.⁴³ PBBs were developed by stretching the core structure of the well-known histological, prototypical fluorescent dye, thioflavin-T (a precursor to the well-known Aβ staining agent) (Fig. 4) by inserting two carbon–carbon double bonds between aniline and benzothiazole groups and radiolabeling with carbon-11. In this series, the most prominent PET tracer was [¹¹C]PBB3 (Fig. 8).⁴³ A clinical study of the PET tracer [¹¹C]PBB3 in the human brain showed a robust signal in the hippocampal regions where tau pathology is enriched, and had high binding selectivity for NFTs over Aβ plaques (>40–50 fold). Even though [¹¹C]PBB3 was demonstrated to interact with 3R- and 4R-tau aggregates in the human brain tissue with higher levels than [¹⁸F]AV1451,⁴⁴ due to its off-target binding in the regions of dural venous sinuses of all human subjects and the presence of a brain-penetrating metabolite, photoisomerization of the double bonds (leading to chemical and radiochemical purity) resulted in new analogues to pursue in this series. Several articles and reviews have discussed the clinical PET images of this tracer.^12,31,32

Fig. 4. Structures of early Aβ-directing PET probes for AD diagnosis.

Fig. 8. Structure of [¹¹C]PBB3.⁴³.

Since the synthesis of [¹¹C]PBB3 from both patented and published studies lacks details with low yields,^43,45 in 2015, Wang and co-workers⁴⁶ reported an updated synthesis of [¹¹C]PBB3 in order to have access to PET research for further development in this series. Their updated synthesis shows that [¹¹C]PBB3 was initially accessed via one pot-two-step synthesis from the key precursor TBS-protected N-desmethyl-PBB3 resulting in low radiochemical yields and long overall reaction times (structure not shown here). Therefore, to improve the radiochemical yield with a shorter overall reaction time, they have revised the radiosynthesis by adopting the method from the [¹¹C]PiB synthesis^47,48 (a similar structure to [¹¹C]PBB3) in which the radiolabeling step was achieved from the unprotected hydroxyl group precursor via fully automated, one-pot-one-step radiosynthesis from N-desmethyl-PBB3 (via N-[¹¹C]methylation using [¹¹CH₃OTf]) with a shortened synthesis time and with relatively good radiochemical yields (20–25% decay-corrected radiochemical yield) (Fig. 9).

Fig. 9. One-pot radiosynthesis of [¹¹C]PBB3 by Wang et al.⁴⁶ (RCY = radiochemical yield; SA = specific activity).

New tau PET tracers from 2016–present

[¹⁸F]PM-PBB3 a.k.a. [¹⁸F]APN-1607 development and its clinical studies

To overcome the technical issues (rapid conversion into metabolites, etc.) associated with [¹¹C]PBB3, Tagai et al. (in 2021)⁴⁹ modified the chemical structure of [¹¹C]PBB3 into a more metabolic stable tracer [¹⁸F]PM-PBB3, aiming at unambiguous detection of tau aggregates in individuals with AD and non-AD symptoms such as frontotemporal lobar degeneration (FTLD) in which 4R-tau is the predominant isoform. [¹⁸F]PM-PBB3 is a fluorinated-propanol modification of PBB3 with advantages over [¹¹C]PBB3, such as higher metabolic stability, higher PET scan throughput and broader availability of ¹⁸F reagents for radiolabeling. The initial radiosynthesis of [¹⁸F]PM-PBB3 was achieved in two steps by reacting the tosylate precursor of [¹⁸F]PM-PBB3 with [¹⁸F]KF followed by acid deprotection yielding [¹⁸F]PM-PBB3 (Fig. 10). For clinical applications and to obtain higher levels of chemical and radiochemical purity, an updated version of the radiosynthesis of [¹⁸F]PM-PBB3 was reported by Ohkubo et al. in 2021.⁵⁰ In this automated radiosynthesis approach, [¹⁸F]epifluorohydrin (prepared in one step from epoxypropyl tosylate) was used as an [¹⁸F]alkylating agent to obtain a high purity radiotracer (Fig. 11).

Fig. 10. Radiosynthesis of [¹⁸F]PM-PBB3 by Tagai et al.⁴⁹ (RCP = radiochemical purity; SA = specific activity).

Fig. 11. Updated and automated radiosynthesis of [¹⁸F]PM-PBB3 by Ohkubo et al.⁵⁰ (RCY = radiochemical yield; RCP = radiochemical purity; Am = molar activity; EOS = at the end of synthesis).

A clinical study⁴⁹ of [¹⁸F]PM-PBB3 in human subjects (Fig. 12) revealed increased binding/retention in the neocortical and limbic regions of the brain where both 3R- and 4R-tau isoforms are predominantly present in AD patients, while retention in the subcortical region indicates the presence of 4R-tau isoforms (deposits) in a PSP patient. In a sharp contrast, a low radio signal was obtained in the parenchyma of elderly healthy control (HC) brains after 30 min post injection of radioligand resulting in low retention and fast clearance. However, the clearance of [¹⁸F]PM-PBB3 was profoundly slow in AD and PSP patient brains (reflecting an increased SUVR) due to high tau-burden, which indicates the specific binding of radioligand [¹⁸F]PM-PBB3 to tau aggregates. In addition, the peak uptake of [¹⁸F]PM-PBB3 in the brains of the same individuals compared to [¹¹C]PBB3 was approximately 2-fold higher. No off target/non-specific radioactivity retention was observed in the HC from [¹⁸F]PM-PBB3 in the basal ganglia and venous sinuses and neocortical areas, while high retention and low retention were observed in these areas for [¹¹C]PBB3 implying the superiority of [¹⁸F]PM-PBB3 in this series. Moreover, [¹⁸F]PM-PBB3 and [¹¹C]PBB3 were found to be photostable (do not absorb light) at wavelengths longer than 500 nm UV-vis (<500 nm cutoff wavelength) as photo-isomerization was a known issue for [¹¹C]PBB3. In addition, the [¹⁸F]PM-PBB3 probe also captured FTDL-type tau pathology (data not shown here) allowing identification of AD and non-AD tauopathies. APRINOIA Therapeutics is leading the development of this series and this radiotracer is under clinical study with estimated completion by Nov 2025.⁵¹

Fig. 12. High-contrast PET images of AD and PSP tau pathologies in humans enabled by [¹⁸F]PM-PBB3 in comparison with [¹¹C]PBB3 and [¹¹C]PiB (this image was taken from the reference Tagai et al., @Elsevier)⁴⁹ (SUVR = standardized uptake value ratio; HC = healthy control).

[¹⁸F]MK-6240 discovery and its preclinical evaluation

In 2016, Walji et al.⁵² at Merck reported the discovery of [¹⁸F]MK-6240 as a PET imaging agent for quantification of NFTs in AD patients (Fig. 13).

Fig. 13. Design strategy of MK-6240.

Their design strategy began with focusing on the molecular connectivity starting from the minimal pharmacophore identification, and then the selective incorporation of the ¹⁸F isotope over the ¹¹C isotope (due to a shorter half-life t_1/2 = 20.3 min) into the target molecule. They screened the Merck compound collection using a competitive binding assay on in vitro assembled tau filaments to identify a new reference (lead) radioligand, followed by the AD cortical homogenate assay as a secondary assay for lead optimization to identify the target radioligand. Initial lead identification resulted in compound 1, an aza-indole linked isoquinoline. Fluorination at the 6-position of the isoquinoline ring resulted in compound 2. The pharmacokinetic and intrinsic physicochemical properties of [¹⁸F]2 (radioligand of 2) in non-AD rhesus monkeys showed rapid distribution across the blood–brain-barrier (BBB) and rapid clearance but showed significant retention in the white matter. Several modifications were performed on 2 to reduce the polar surface area to decrease the overall lipophilicity, and amino group insertion in isoquinoline resulted in the lead compound MK-6240 with good physicochemical properties (calculated CNS-MPO (>5.5) and shake flask log D (2.9–3.3)). To enable in vivo PET evaluation of MK-6240 in rhesus monkeys, radiolabeled compound [¹⁸F]MK-6240 was prepared in one step from 5-diBoc-6-nitro precursor 3 in two synthetic steps, and the cGMP production of this tracer was reported by Collier et al.⁵³ (Fig. 14). Preclinical PET studies⁵⁴ in monkeys illustrated that the new tau PET tracer [¹⁸F]MK-6240 showed high specificity and selectivity for binding to NFTs and no off-target binding affinity to white matter or MAO-A/B (monoamine oxidase A or B). [¹⁸F]MK-6240 has been widely used in clinical trials in large cohorts as well as in a therapeutic trial of aducanumab.⁵⁵ [¹⁸F]MK-6240 is under clinical evaluation and no additional clinical data are available as of now.

Fig. 14. Scheme for the radiochemical synthesis of [¹⁸F]MK-6240.^52,53.

Discovery of [¹⁸F]JNJ-311 and [¹⁸F]JNJ-067 as tau PET tracers

Despite the high selectivity of [¹⁸F]AV-1451 for tau aggregates over Aβ plaques, it showed⁵⁶ high off-target affinity for monoamine-oxidase-A (MAO-A) and slow washout from the human brain. For a more sensitive and selective tau PET tracer, in 2017, Rombouts et al. at JNJ reported the discovery of a novel tau targeting 1,5-naphthyridine based PET tracer [¹⁸F]JNJ-311.⁵⁷ In their medicinal chemistry program, an initial mini-HTS of ∼4000 compounds was performed by applying two different selection strategies: 1) 2D fragment-based (ECFP6) similarity using internally developed software and 2) 3D shape/pharmacophore comparison (FastROCS), and found N-(6-methyl-2-pyridyl)quinolin-2-amine 4 as a HTS hit with good binding affinity for aggregated tau and moderate selectivity for aggregated Aβ (Fig. 15). Initial SAR exploration to identify structural elements and further medicinal chemistry optimization for identification of a suitable fluorination position led to the discovery of a radiofluorination candidate, 1,5-naphthrydine based tau binding ligand 5 (Fig. 15).⁵⁷ Radiofluorinated ligand [¹⁸F]JNJ-311 (Fig. 15) was synthesized in one step by a nucleophilic aromatic substitution reaction with ¹⁸F-fluoride on its trimethylammonium precursor.⁵⁸In vitro semiquantitative autoradiography studies on postmortem human brain sections of Alzheimer's disease patients showed high binding affinity to tau rich regions. In vivo PET scans in rhesus monkeys revealed moderate initial brain uptake (SUVR 1.9 at 1 min post injection of [¹⁸F]JNJ-311) and rapid washout from the brain. However, some bone uptake was detected.⁵⁸

Fig. 15. Structures of the hit compound 4, optimized ligand 5 and radiolabeled tau PET tracer [¹⁸F]JNJ-311.

Rombouts et al. (in 2019)⁵⁹ at JNJ reported a tau PET tracer [¹⁸F]JNJ-067, an optimized derivative of their previous radiotracer [¹⁸F]JNJ-311. During the [¹⁸F]JNJ-311 SAR exploration, they discovered that good potency and selectivity could be achieved when shifting the ‘nitrogen’ atom in the quinoline of 4 to the 6-position, a similar strategy seen in replacing the 2-aminopyridine with 4-aminopyridine to generate 5 (Fig. 15). Further medicinal chemistry exploration on the resulting intermediate (not shown here) led to the identification of fluorinated ligand 6 (JNJ-067) with an improved preclinical PK profile and affinity for aggregated tau. Radiolabeling of the nitro precursor by nucleophilic substitution with [¹⁸F]KF resulted in the discovery of [¹⁸F]JNJ-067 (Fig. 16).

Fig. 16. Structures of 4, intermediate ligand 6 and tau PET tracer [¹⁸F]JNJ-067.⁵⁹.

In vivo preclinical PET evaluation of [¹⁸F]JNJ-067 in a male rhesus monkey revealed a relatively high initial brain uptake and rapid washout, low off-target binding (Aβ, MAO enzymes), good brain penetration, and low risk of radiodefluorination and formation of radiometabolites.⁵⁹ Initial clinical studies⁶⁰ of [¹⁸F]JNJ-067 revealed tracer retention in the tau aggregate regions in both positive AD and MCI patients, and no retention was found in healthy controls. Some off target binding of [¹⁸F]JNJ-067 was found in the white matter and basal ganglia, and no additional clinical data are available as of now.

[¹⁸F]GTP1 (Genentech tau probe 1) an alternative PET ligand to [¹⁸F]AV-680 or [¹⁸F]T808

Initial clinical evaluation (Fig. 5) of one of the early PET tracers in the AV series of [¹⁸F]AV-680 or [¹⁸F]T808 showed a desirable brain kinetic profile and uptake in the specific regions of the brain associated with AD patients.⁶¹ However, significant metabolic instability resulted in apparent defluorination and accumulation of [¹⁸F]fluoride in the skull (bone), which hampered the further development of [¹⁸F]AV-680 or [¹⁸F]T808. To address the limitations associated with [¹⁸F]AV-680 or [¹⁸F]T808, in 2019, Sanabria Bohórquez et al. at Genentech reported⁶² a new PET tracer [¹⁸F]GTP1 (Genentech tau probe 1). The new probe design strategy was based on the hypothesis that the isotopic substitution of hydrogen by deuterium could improve the metabolic stability of the molecule, and this concept has been previously employed in PET tracer design to reduce the formation of free [¹⁸F]fluoride.^63,64 Therefore, they designed and synthesized the new deuterated probe [¹⁸F]GTP1, by the isotopic substitution of hydrogens on the carbon bearing an ¹⁸F in [¹⁸F]AV-680 or [¹⁸F]T808 by deuterium (Fig. 17). This new PET tracer exhibited high affinity and selectivity for tau over Aβ, with no binding to MAO-B in AD tissue. When tested in humans, [¹⁸F]GTP1 exhibited favorable dosimetry and brain kinetics without a sign of defluorination which is a major issue with [¹⁸F]AV-680 or [¹⁸F]T808, and specific binding to cortical regions of the brains of AD patients predicted to have tau pathology was observed. In addition, in a cross-sectional population, the PET tracer [¹⁸F]GTP1 exhibited specific and increased binding with AD severity and differentiated AD from cognitively normal controls (CN) (data not shown here).⁶²

Fig. 17. Structures of [¹⁸F]AV-680 or [¹⁸F]T808, and [¹⁸F]GTP1.

Since its discovery, it has widely been used in therapeutic trials to monitor the accumulation of tau aggregates in the brains of Alzheimer's disease patients. To further validate the metabolic stability of [¹⁸F]GTP1, in 2023, Marik et al. (at Genentech) assessed⁶⁵ its metabolic process in vitro in the influence of smoking, since smoking strongly stimulates the activity of the CYP1A2 enzyme. When the tracer was incubated with recombinant CYP1A2 (to evaluate the role of the enzyme in tracer metabolism), they found that the tracer [¹⁸F]GTP1 could form more than 11 high polar metabolites compared to the parent tracer, a small portion of defluorinated metabolite (∼2.6%). In addition, they found that the formed metabolites are not due to the activity of CYP1A2 suggesting no influence of smoke on the tracer metabolism, indicating the stability of [¹⁸F]GTP1 over [¹⁸F]AV-680 or [¹⁸F]T808.

The structure of [¹⁸F]GTP1 consists of a benzimidazopyrimidine core with a radiolabeled fluoride (¹⁸F) adjacent to two geminal deuterium atoms. The radiosynthesis of [¹⁸F]GTP1 can be achieved in one step from its deuterio-tosylate (7) precursor (Fig. 18). However, due to the short half-life of ¹⁸F (t_1/2 = 109.8 min), it is prepared as needed, and the synthesis of the deuterio-tosylate (7) precursor requires a robust synthetic route and it is not economical/practical to produce on a large scale for clinical use. Therefore, in 2020, White et al., at Genentech, reported⁶⁶ a scalable and improved method for the synthesis of the deuterio-tosylate (7) precursor.

Fig. 18. Schematic of one step radiosynthesis of [¹⁸F]GTP1 from its tosylate precursor.

Upon evaluation of the first-generation synthetic route of [¹⁸F]GTP1, White et al. found⁶⁶ some issues related to reproducibility, purification and low reaction yields in the key synthetic steps. Therefore, White et al. designed⁶⁶ a more robust, scalable, second-generation synthetic route in seven steps (Fig. 19) starting from commercially available N-Boc protected piperidine methyl ester 8. The key reaction step involved the annulation between α,β-unsaturated deuteron-amide 9 and phosphoramidate 10, followed by debenzylation to provide the alcohol precursor. The final step involves the tosylation of an alcohol precursor under acidic conditions to provide the tosylate precursor 7 in high yield, which can easily be converted into [¹⁸F]GTP1 in one step using K¹⁸F as shown in Fig. 18. Last update on the clinical trials of [¹⁸F]GTP1 was available as of November 2021.⁶⁷

Fig. 19. Scheme for the radiosynthesis and design strategy for [¹⁸F]GTP1.⁶⁶.

Development of [¹⁸F]RO-6958948

In 2017, Gobbi et al., at Roche, reported the identification of novel PET radiotracers for imaging of aggregated tau in AD.⁶⁸ Their internal discovery program aimed to find high affinity tau NFT ligands following the structural features of reported tau PET ligands [¹⁸F]AV-1451 and [¹⁸F]AV-680. The strategy was based on the hypothesis that these two structurally similar probes may share a common binding site on the amyloid β-sheet surface. Initial molecular design was focused on the simplification of the tricyclic aromatic core in [¹⁸F]AV-1451 and [¹⁸F]AV-680, followed by the heteroatom incorporation in the aromatic rings to reduce the lipophilicity of the molecule for improved specificity. They designed and synthesized 550 analogues using a medium to high throughput tau binding assay based on the displacement of [³H]AV-680 from brain tissue originating from AD patients, since [³H]AV-680 was shown to possess high affinity for tau aggregates in AD (K_D = 22 nm). Three ligands, RO-6924963, RO-6931643 and RO-6958948 (Fig. 20), were identified as high-specific displacers of the [³H]AV-680 binding site in the tau-rich cortical tissue sections obtained from donors of AD patients with Braak stages V and VI. In addition, these three ligands showed excellent selectivity to tau NFTs over Aβ plaques in vitro and exhibited no binding affinities in brain sections obtained from non-AD patients. The autoradiography displacement assay revealed no affinity to MAO-A or MAO-B.⁶⁹ Successful radiolabeling⁶⁸ with either [¹¹C] or [¹⁸F] resulted in tau PET ligands [¹¹C]RO-6924963, [¹¹C]RO-6931643 and [¹⁸F]RO-6958948 (Fig. 20). Interestingly, the tracer [¹⁸F]RO-6958948 emerged from the slight modification of displacement of carbon by nitrogen in the aromatic ring of [¹⁸F]AV-1451. Preclinical evaluation⁶⁹ of these tracers in tau-naïve baboons (intravenous administration) indicated good brain uptake, rapid washout, and a favorable metabolic pattern. When evaluated in Aβ-positive AD human subjects,⁷⁰ [¹⁸F]RO-6958948 exhibited the highest brain uptake (SUV = 3.5) and fast wash out compared to the other tracers [¹¹C]RO-6924963 (SUV = 3.0) and [¹¹C]RO-6931643 (SUV = 1.5). In addition [¹⁸F]RO-6958948 showed better contrast between areas of high vs. low tau aggregate accumulation. Therefore, superior characteristics enabled the evaluation of [¹⁸F]RO-6958948 as a PET tracer for quantitative assessment of tau accumulation in the human brain,⁷¹ and this tracer showed high discrimination between tau positive AD patients and the HC with a significant margin. Younger control (YC) and older control (OC) subjects showed a relatively homogeneous radioactivity distribution at 60–90 min post administration of [¹⁸F]RO-6958948, whereas AD subjects showed a heterogeneous pattern of radiotracer retention in the brain regions known to exhibit tau pathology (Fig. 21). [¹⁸F]RO-6958948 showed no off-target binding in the regions of basal ganglia, choroid plexus, thalamus and white matter; in contrast, off-target binding in the choroid plexus is the main issue with [¹⁸F]AV-1451. Since its first-in-human evaluation, [¹⁸F]RO-6958948 has been used in clinical development and no further data are available as of now.

Fig. 20. Structures of Roche's tau PET ligands [¹¹C]RO-6924963, [¹¹C]RO-6931643 and [¹⁸F]RO-6958948.

Fig. 21. SUVR (60–90 min) images of [¹⁸F]RO-948 for representative YC, OC, and 3 levels of AD subjects (this figure was taken from the original reference).⁷¹.

Discovery of [¹⁸F]PI-2620

In 2019, Kroth et al. reported the discovery and preclinical evaluation of [¹⁸F]PI-2620, a next-generation PET ligand for the assessment of tau pathology in AD and other tauopathies.⁷² Their discovery goal was to identify a ligand with 1) high affinity and selectivity to pathological tau aggregates over other targets in the brain; 2) favorable brain uptake and fast wash out from the brain; 3) lack of defluorination and no formation of metabolites in the brain; and most importantly 4) high selectivity among AD and non-AD tauopathies such as PSP, Pick's disease, etc. Therefore, they initiated the discovery program by screening the Morphomer™ library, in a human AD brain homogenate assay, to identify small molecule ligands that bind to the selected target without binding to off-targets such as Aβ and MAO-A/B. Pyrrolo[2,3-b:4,5-c′]dipyridine (core in RO-948) and pyrido[4,3-b]indole core (as seen in AV-1451) structures found as hits show high affinity for tau deposits with low off-target binding to Aβ. A secondary screening was performed to assess the off-target binding ability of ligands using MAO-A reversible binding ligand [¹⁸F]FEH and reversible MAO-B binding ligand [¹⁸F]deprenyl resulting in 9H-pyrrolo[2,3-b:4,5-c′]dipyridine as an optimal core with improved selectivity. Synthesis of regioisomers (total of 10) around the fluoropyridine ring (from AV-1451) and SAR exploration revealed the ligand PI-2620 (Fig. 22, [¹⁸F] analogue) as a lead molecule when evaluated for target affinity to aggregated tau (pIC₅₀ > 8, in tau brain homogenates). For the in vitro and in vivo evaluation, the radioligand [¹⁸F]PI-2620 (Fig. 22) was synthesized from its N-Boc or N-trityl protected nitro precursor by direct nucleophilic substitution with ¹⁸F, and the N-Boc or N-trityl protecting groups showed improved solubility of the precursor.

Fig. 22. Scheme showing the evolution of [¹⁸F]PI-2620 (top) and its radiosynthesis (bottom).⁷².

Preclinical evaluation in human AD brain homogenates indicated that this ligand exhibited high affinity for 3R/4R tau aggregates and no off-target binding to Aβ. Autoradiography of [¹⁸F]PI-2620 on brain sections from AD patients revealed a good specificity to pathological tau aggregates and also showed selectivity for non-AD tauopathies such as PSP (only the 4R tau isoform) and Pick's disease (only the 3R isoform).

In 2020, Mueller et al., at Life Molecular Imaging GmbH (formerly Piramal Imaging), reported the first-in human tau PET imaging study of [¹⁸F]PI-2620 in patients with AD and healthy controls.⁷³ [¹⁸F]PI-2620 showed good brain uptake (in 5 min post injection) and fast washout from non-target regions of the brain. High localization was observed in AD subjects in the regions of the temporal and parietal lobes, precuneus, and posterior cingulate cortex. The uptake of [¹⁸F]PI-2620 in neocortical regions clearly correlated with the severity of cognitive impairment. [¹⁸F]PI-2620 PET images clearly discriminated between AD and HC subjects, no defluorination was observed and [¹⁸F]PI-2620 showed high selectivity for aggregated tau over Aβ when tau PET was obtained in the same AD subject in comparison with [¹⁸F]-florbetaben amyloid PET (Fig. 23).

Fig. 23. Comparison of PET images of Aβ selective PET [¹⁸F]-florbetaben and [¹⁸F]PI-2620 (this figure was taken from the original reference).⁷³.

In 2023, Malarte et al. reported⁷⁴ the discriminative binding pattern of recent second generation tau PET tracers PI-2620, MK-6240 and RO-948 in tau pathology of AD and primary tauopathies of corticobasal degeneration (CBD) and progressive supranuclear palsy (PSP). They investigated the binding studies and tau complexity using postmortem brain sections and autoradiography of large and small frozen brain sections from AD, CBD and PSP patients. In their study, among tritium labeled tracers, [³H]PI-2620 showed comparable binding affinity in all three tauopathy models, but varying binding site densities in the order AD > CBD > PSP. All three tritium labeled tracers [³H]PI-2620, [³H]MK-6240 and [³H]RO-948 behaved similarly in AD brain tissue, but their competitive binding studies indicated that [³H]PI-2620 detected multiple binding sites in AD, CBD and PSP indicative of its selectivity and tau heterogeneity. Moreover, when compared to [³H]MK-6240 and [³H]RO-948, the tracer [³H]PI-2620 exhibited high specificity in CBD and PSP brain tissues. These initial results indicate the different binding affinities of second-generation tau tracers and help researchers understand heterogeneity in AD vs. non-AD tauopathies and to develop new highly selective tau tracers to discriminate between 3R and 4R tau isoforms. In addition, in a recent report by Varlow and Vasdev,⁷⁵ the [³H]PI-2620 tracer was shown to have in vitro binding affinity to human postmortem brain tissues of pathologically diagnosed patients with chronic traumatic encephalopathy (CTE). When compared to [³H]AV-1451 (off target binding to MAO-A of CTE brain tissue), [³H]PI-2620 and [³H]MK-6240 were shown to bind to CTE tau in severe or mixed pathology cases, indicating the potential use of [¹⁸F]PI-2620 for evaluation of human subjects with suspected CTE. Since the discovery⁷² and first-in-human study report⁷³ of [¹⁸F]PI-2620, it has been widely used for imaging in both AD and non-AD human subjects.^76–82 According to a press report on August 28th, 2024, by Life Molecular Imaging (LMI), the FDA has granted fast track designation to [¹⁸F]PI-2620 for clinical development in AD, PSP and CBD.⁸³

Miscellaneous tracers under development

a) Benzimidazopyridine derivatives

Ono et al., at Koto University in Japan, have been actively pursuing the discovery of PET tracers for imaging aggregated tau in AD.^84–86 Their initial PET ligand discovery program began with the SAR studies of heterocyclic phenylethenyl and pyridinylethenyl benzimidazole derivatives and found phenylethenyl benzimidazole SBI-2 (Fig. 24) as a tau selective SPECT (single-photon emission computed tomography) tracer for tau imaging in AD. However, its low binding affinity to tau aggregates and photoisomerization of the ethenyl group, a similar issue observed with the [¹¹C]PBB3 series, confined their further study as a PET tracer.⁸⁴ However, cyclization and functionalization led to the discovery of photostable tau SPECT ligand [¹²⁵I]BIP-NHEt (Fig. 24), with good tau binding affinity over Aβ, with high brain uptake and fast clearance from the mouse brain.⁸⁵ Further structural optimization at the 3-position followed by SAR studies revealed [¹⁸F]BIPF1 (Fig. 24) as a tau target PET probe with high selectivity for tau over Aβ (tau/Aβ ratio = 34.8), high brain uptake (6.22% ID g⁻¹ at 2 min postinjection) and fast clearance (2.77% ID g⁻¹ at 30 min postinjection) in mice without forming reactive metabolites.⁸⁶ In their recent study⁸⁷ in 2021, they designed and synthesized five analogues of [¹⁸F]BIPF1 with various substitutions at the 7-position with various atoms or groups resulting in [¹⁸F]Me-BIPF (Fig. 24) as a novel tau PET radioligand with improved brain uptake properties among new five analogues (not shown here). However, 7-methyl analogue [¹⁸F]Me-BIPF and other 4 analogues exhibited comparable tau selectivity compared to previous [¹⁸F]BIPF1, but the highest brain uptake (6.79% at 2 min post injection) with a 2 min/60 min ratio (3.59) was observed for [¹⁸F]Me-BIPF. These new findings suggest that further optimization at the 7-position would be an effective strategy for further optimization of this series of PET ligands for the discovery of a highly tau selective radiotracer for imaging of aggregated tau in AD. No additional data of [¹⁸F]Me-BIPF are available as of now.

Fig. 24. Structural evolution of benzimidazopyridine-based PET ligands.⁸⁷.

b) LM229, an [¹¹C] labeled PET ligand and a derivative of the PBB3 PET tracer

Exploration of tau pathology in rat models (no earlier report on these animal models) and translation of those models into human subjects have been an understudied area in the tau PET ligand discovery process. The advantage of using a rat model is the bigger size of the rat brain (six times bigger than the mouse brain), which would enable more precise radioactivity quantification in the brain regions. Therefore, to develop new PET tracers to image tau pathology in transgenic mouse and rat models of neurodegeneration, in 2021, McMurray et al. reported⁸⁸ the synthesis of novel ligands based on the existing PBB3 structures⁴³ since PBB3 tracers have been used to PET imaging of tau aggregates in mouse models and humans.

The synthesis of unlabeled ligands was accomplished using a convergent synthetic route with the Horner–Wadsworth–Emmons (HWE) reaction as a key reaction step in the coupling of the aldehyde intermediate A with phosphonate B to generate a library of PBB3 analogues. The ligand LM229 (Fig. 25) was found to have high binding affinity to the recombinant tau fibrils with K_D = 3.6 nM. LM229 demonstrated high specificity to pathological 4R tau in living cultured neurons, mouse brain sections (human transgenic P301S tau mouse) and transgenic rat brain sections (truncated human 151–351 3R/4R tau aggregates). In addition, LM229 also exhibited high specificity to tau aggregates in brain sections from AD (3R/4R) and PSP (4R only) patients. Therefore, this has been the first protocol demonstrating the binding affinity of PET imaging ligands in rat models. Finally, LM229 was shown to cross the BBB, and a radiolabeling method was established to generate the radioligand [¹¹C]LM229 from its amine precursor 10 (Fig. 25) for in vivo translational studies. No additional in vivo evaluation of [¹¹C]LM229 are reported as of now.

Fig. 25. Scheme for the design strategy and [¹¹C]LM229 radioligand synthesis.⁸⁸.

c) [¹⁸F]N-Methyl lansoprazole or [¹⁸F]NML as a tau PET imaging agent

New PET ligand discovery based on the known proton pump inhibitor lansoprazole for tau aggregate imaging in various tauopathies has been of long-standing interest in the research groups of Scott at the University of Michigan, USA, and Riss at the University of Oslo, Norway.⁸⁹ A previous report⁹⁰ on the binding affinity (2.5 nM) of lansoprazole for heparin-induced tau aggregates in vitro inspired them to repurpose and develop tau PET imaging agents based on lansoprazole for clinical use. Therefore, they developed several analogues^91–93 and reported their preclinical evaluation. In 2020, they reported the first-in-human evaluation of [¹⁸F]N-methyl lansoprazole ([¹⁸F]NML) as a tau PET imaging agent, and an improved radiochemical synthesis for clinical use (Fig. 26).⁸⁹ The radiosynthesis of [¹⁸F]NML involves a novel method for radiofluorination of the gem-difluoro enol ether precursor with [¹⁸F]KF in DMSO at 90 °C (Fig. 26). This clinical production method yielded good noncorrected radiochemical purity, high molar activity and excellent radiochemical purity (>97%).⁸⁹ The first-in-human studies of [¹⁸F]NML were evaluated in 11 subjects that include healthy controls (n = 4), MCI/AD patients (n = 6) and a PSP patient (n = 1). [¹⁸F]NML exhibited good brain uptake, reasonable pharmacokinetics in the healthy controls, and no adverse effects in all 11 subjects. However, PET images of [¹⁸F]NML revealed no significant brain retention (very low) in the MCI/AD or PSP patients when compared to its in vitro tau affinity for 3R and 4R tau.⁹⁴ In addition, there was no evidence of binding specificity of [¹⁸F]NML to tau aggregates in vivo. Therefore, the authors claimed that the clinical translation of this tracer was unsuccessful and discontinued from further development.

Fig. 26. Structure of lansoprazole and the radiosynthetic scheme of [¹⁸F]NML for clinical production.⁸⁹.

d) 2-Phenylquinoxaline and styrylquinoxaline derivatives as tau PET imaging agents

In search of novel PET tracers to image Aβ plaques in the brains of AD patients, Cui⁹⁵ and others^96,97 reported a series of papers on the discovery and in vitro and/or in vivo evaluation of 2-phenylquinoxaline scaffold based PET tracers. During the discovery process, Cui et al. observed that some of the ¹⁸F labeled 2-phenylquinoxaline derivatives were also found to recognize NFTs in vitro in the brain sections of AD patients. Those observations initiated them to develop novel PET tracers to image NFTs based on the optimization of 2-phenylquinoxaline scaffolds. Since the 2-phenylquinoxaline and THK derivatives share a common core (quinoline), they followed the THK optimization process to develop novel NFT targeting PET tracers by optimizing the 2-phenylquinoxaline scaffolds.^98–100 To achieve high binding affinity to NFTs, low off-target binding and high selectivity over Aβ plaques, the 2-phenylquinoxaline core was derivatized with a fluoropropanol side chain as in the THK series (Fig. 27).¹⁰¹ The fluoropropanol side chain not only improves the hydrophilicity of the tracer, but also results in better in vivo pharmacokinetics and higher selectivity for tau over Aβ plaques as evidenced by the development of the [¹¹C]PBB3 tau tracer.⁴³ Therefore, they designed and synthesized 6 (3 enantiomeric pairs) chiral fluoropropanol appended 2-phenylquinoxaline analogues with varying substituents of N,N-dimethylamino or N-methylamino groups at the 6- or 7-position on the quinoxaline moiety (Fig. 27).¹⁰¹

Fig. 27. New PET ligand design strategy for the 2-phenylquinoxaline scaffold following the optimization strategy of THK based PET tracers.¹⁰¹.

Radiolabeling of each analogue was achieved by the nucleophilic substitution of the tosylate precursor with [¹⁸F]KF. All synthesized probes displayed excellent fluorescence properties, and radiotracers (S)-[¹⁸F]16 and (R)-[¹⁸F]5 (Table 2) exhibited high selectivity for tau tangles evaluated on the brain sections from transgenic mice and AD patients. Further evaluation of (S)-[¹⁸F]16 and (R)-[¹⁸F]5 in a quantitative binding assay with AD homogenates revealed high affinity and selectivity for NFTs over Aβ plaques. In vitro autoradiography on AD brain tissue further confirmed the high affinity and selectivity of these two probes for tau over Aβ (Table 2). In vivo evaluation in normal ICR mice revealed good BBB penetration and suitable brain kinetics. (S)-[¹⁸F]16 displayed low off-target binding to MAO-B compared to (R)-[¹⁸F]5. In addition, (S)-[¹⁸F]16 exhibited high binding affinity, high selectivity for tau over Aβ, good brain uptake and low off-target binding to MAO-B compared to the existing [¹⁸F]THK-5351 tau tracer. Therefore, the preclinical evaluation data of (S)-[¹⁸F]16 clearly show its advantage over THK based tracers. Therefore, (S)-[¹⁸F]16 was selected for clinical validation studies in human subjects (AD patients), and no additional clinical data are available as of now.

Structures of radiotracers (R)-[¹⁸F]5 and (S)-[¹⁸F]16.

Structure of the PET tracer
Tau affinity (nM)a	4.1	10.3
Selectivity for tau over Aβa	30.5 fold	34.6 fold
BBB penetrationb	7.06% ID g−1	10.95% ID g−1
Brain kinetics:b brain2min/brain60min	10.1	6.5
IC50 (μM) for MAO-Bc	0.33	>10

^a In vitro.

^b In vivo.

^cSpectro-photometric assay (enzyme inhibition assay).

In this series, Wu and Cui et al., in 2023, reported¹⁰² a new tau selective PET tracer [¹⁸F]15 based on the 2-styrylquinoxaline scaffold, a derivative of their previous tracer (S)-[¹⁸F]16 (Fig. 28). The design strategy involves the introduction of a double bond between the quinoxaline moiety and benzene ring, and two ethoxy units (FPEG) onto the phenyl ring for skeletal extension and to improve pharmacokinetic properties, respectively, while the N,N-dimethylamine group on quinoxaline was kept intact for tau selectivity over Aβ. In vitro binding affinity of [¹⁸F]15 on recombinant K18-tau aggregates revealed high affinity and high selectivity for tau vs. Aβ plaques (Fig. 28). In addition, autoradiography studies and fluorescence staining profiles in brain sections of AD patients and tau-transgenic mice revealed additional evidence of its high selectivity for tau aggregates. Initial in vivo evaluation of [¹⁸F]15 in normal ICR mice revealed good brain uptake, a moderate washout ratio and the ability to penetrate across the BBB (Fig. 28). However, there is no evidence reported on the photo-stability of the double bond and radiodefluorination of the tracer [¹⁸F]15. Nevertheless, these initial results of the styrylquinoline scaffold [¹⁸F]15 indicate it as a promising PET tracer for further in vivo evaluation.

Fig. 28. Design strategy for the new tau selective PET tracer [¹⁸F]15.¹⁰².

e) [¹⁸F]FPND (2-(4-[18F]fluorophenyl)imidazo[1,2-h][1,7]naphthyridine)

In continuation of their discovery efforts for novel tau PET tracers, Liu and Cui et al. screened several aza-6,6,5-tricyclic cores and developed a ¹²⁵I labeled imidazonaphthyridine-based radiotracer [¹²⁵I]PND-1 (Fig. 29) for in vitro tau aggregate imaging in AD.¹⁰³ This aza-fused tricyclic derivative exhibited moderate in vitro PET properties. Therefore, for further expansion of this tracer's applications in vivo and preclinical evaluation as a PET imaging agent, they performed extensive SAR studies on [¹²⁵I]PND-1. Medicinal chemistry efforts were aimed at its substituent effect on imidazo[1,2-h][1,7]naphthyridine as well as varying ¹⁸F-labeling sites in [¹²⁵I]PND-1. Extensive SAR studies resulted in an ¹⁸F-labeling analogue [¹⁸F]FPND-4 as a new tau PET tracer in this series.¹⁰⁴In vitro autoradiography quantitative analysis on human brain tissue revealed its high affinity for tau aggregates (IC₅₀ = 2.80 nM), and some off-target binding to Aβ and MAO-A/B enzymes. In vivo PET imaging and evaluation in rodents and non-human primates exhibited desirable brain uptake (SUV = 1.75 at 2 min), rapid clearance (brain_2min/60min = 5.9) and minimal defluorination in rhesus monkeys. Therefore, these preliminary results warranted further evaluation in human AD subjects. No additional data are reported on this tracer as of now.

Fig. 29. Structures of [¹²⁵I]PND-1 (ref. 103) and [¹⁸F]FPND-4.¹⁰⁴.

f) Discovery and preclinical optimization of [¹⁸F]SNFT-1

To address the off-target (MAO-B) binding affinity of their previously developed tau PET tracer [¹⁸F]THK-5351 (Fig. 5), in 2023, Harada and Okamura et al. reported¹⁰⁵ the discovery and preclinical evaluation of a new PET tracer [¹⁸F]SNFT-1 (Fig. 30). Compound optimization was performed on [¹⁸F]THK-5351, followed by SAR studies and further lead optimization resulted in [¹⁸F]SNFT-1. An in vitro binding assay revealed its high binding affinity (K_D = 0.6 nM) and selectivity against tau aggregates. In addition, in a head-to-head autoradiography comparative assay with other second-generation tracers, [¹⁸F]SNFT-1 exhibited a high signal-to-background noise ratio as seen in [¹⁸F]MK-6240 indicating its usefulness as a novel PET tracer for imaging of tau aggregates. Furthermore, [¹⁸F]SNFT-1 showed high selectivity and affinity for 3R/4R tau aggregates but low binding affinity (IC₅₀ > 1000 nM) for both MAO-A/B enzymes and Aβ plaques. [¹⁸F]SNFT-1 showed a high initial brain uptake and rapid washout from the normal mouse brain and no sign of radio-metabolites was detected. Additional evaluation in in vitro and in vivo binding affinity and pharmacokinetic studies to address metabolism are needed before it is translated into human subjects to quantitatively monitor tau aggregates in AD patients.

Fig. 30. Structures of [¹⁸F]THK-5351 and optimized tracer [¹⁸F]SNFT-1.¹⁰⁵.

Evolution of 4R tau directing PET tracers

1. Discovery of [¹⁸F]CBD-2115 as a first-in-class radiotracer for imaging 4R-tauopathies

3R tau is mainly present in Pick's disease while 4R is in both PSP and CBD (see Table 1, for classification of tauopathies and terminology). AD and CTE were both characterized by the presence of 3R/4R tau isoforms with equal ratios in AD while unequal ratios in CTE. First and second generation tau targeting PET tracers have emerged and some of them are in clinical studies (except the FDA approved PET tracer [¹⁸F]AV-1451). However, 4R tau selective PET tracers are now under development even though some initial 4R selectivity seen by the existing PET tracers [¹⁸F]PI-2620 and [¹⁸F]PM-BB3. Therefore, 4R tau selective PET tracer development is still a challenging area of radiochemistry research. In 2019, a patent by CBD Solutions¹⁰⁶ disclosed 148 compounds with an objective of discovering new PET radiotracers to target the 4R tau isoform. Those compounds were designed following the established principles of CNS PET tracer development¹⁰⁷ and to target the 4R tau isoform (see Fig. 2) with high selectivity and affinity.

In 2021, Svensson (at CBD Solutions, Sweden), Mathis (Univ. of Pittsburgh) and Vasdev (Univ. of Toronto) et al. reported¹⁰⁸ the discovery of a highly 4R-seelctive tau imaging agent as a first-in-class radiotracer. Their discovery approach began with the selection of a lead ligand CBD-2115 from patented¹⁰⁶ 148 compounds based on the pyridinyl-indole scaffold (structure not shown here). ³H-labeled CBD-2115 was used for initial in vitro evaluation in P301S transgenic mouse tissue, and in human brain tissues obtained from AD, PSP, and CBD patients. In the 4R expressing mouse brain tissue, the K_D of [³H]CBD-2115 was found to be 6.9 nM with a B_max of 500 nM. Meanwhile in the human brain tissue, when compared with the first and second generation ³H-labeled tau tracers [³H]AV-1451 and [³H]MK-6240, [³H]CBD-2115 showed higher affinity for PSP (4.9 ± 0.2 nM) and CBD (27 ± 1 nM) than the other two reference tracers [³H]AV-1451 (PSP: 45 ± 4; CBD: 33 ± 2 nM) and [³H]MK-6240 (PSP: >50; CBD: >50 nM). In addition, computational studies revealed that CBD-2115 had favorable binding affinities for CTE tau protofibrils.¹⁰⁹

Radiolabeled [¹⁸F]CBD-2115 was synthesized from its N-Boc protected nitro precursor in two steps (Fig. 31). The physicochemical parameters for [¹⁸F]CBD-2115 were calculated and found to be within the acceptable range of the CNS PET radioligand. The dynamic PET imaging evaluation in rodents and nonhuman primates showed an initial brain uptake (SUV = 0.5–0.65) with fast washout. However, this initial brain uptake value is somewhat less than the normal acceptable value range (SUV = 2–5 at 2 min). Therefore, additional chemical modifications along with the tracer insertion site need to be optimized to improve the brain uptake and selectivity for 4R tau aggregates.

Fig. 31. Scheme for the radiosynthesis of [¹⁸F]CBD-2115.¹⁰⁸.

2. Discovery of [¹⁸F]Z3777013540 and [¹⁸F]Z4169252340

Due to the low abundance of 4R tau aggregates in non-AD tauopathies, in contrast to the high abundance of 3R/4R tau aggregates in AD, a more selective PET ligand is desired for improved in vitro evaluation for 4R tau aggregates. Therefore, in their continuing efforts on discovery of new 4R tau selective PET ligands, in 2023, Graham and others (Vasdev and Mathis) reported an in silico method for new ligand discovery¹¹⁰ based on their previously reported 4R tau selective ligand CBD-2115.¹⁰⁸ Identification of the binding structure using cryo-EM has become a center stage in structural biology.¹¹¹ A recent report on the binding site identification¹¹² of PM-PBB3 in AD tau aggregates initiated them to use an in silico method for new ligand discovery. The in silico search was initiated by screening of the 3.5B compound library from the Enamine REAL collection (screening details are not discussed here), followed by the visual inspection of the top 750 compounds resulting in a new ligand Z3777013540 (1-(5-(6-fluoro-1H-indol-2yl)pyrimidine-2-yl)peperidine-4-ol), with structural similarity to the previously reported CBD-2115. It was further validated using the calculated CNS MPO (3.8), CNS PET MPO (2.8) and BBB (3.55) scores of Z3777013540 over CBD-2115 (3.7, 1.9 and 3.18). In addition to the structural similarity, superior physicochemical properties enabled the selection of Z3777013540 as a potential lead. A subsequent molecular similarity search using Z3777013540 as a reference compound resulted in the identification of a new ligand Z4169252340 (4-(5-(6-fluoro-1H-indol-2-yl)pyrimidin-2-yl)morpholine (Fig. 32).

Fig. 32. Structures of Z3777013540 and Z4169252340 and the corresponding radioligands [¹⁸F]Z3777013540 and [¹⁸F]Z4169252340.

Tritiated [³H]Z3777013540 was used for in vitro binding assay evaluation in the post-mortem brain sections of AD, PSP and CBD patients and revealed K_D (nM) values of 4.0, 5.1 and 4.5, respectively. For [³H]Z4169252340, the K_D (nM) values are 1.2 for AD, 1.6 for PSP and 1.7 for CBD. In addition, [³H]Z4169252340 showed lower binding affinity for aggregated α-synuclein and Aβ plaques. In vivo PET imaging of radiofluorinated ligand [¹⁸F]Z3777013540 in rats revealed good brain penetration and rapid washout from normal brain tissue. However, [¹⁸F]Z4169252340 demonstrated higher brain penetration and rapid clearance compared to [¹⁸F]Z3777013540 from normal brain tissue. Collectively, both PET tracers exhibited affinity to 4R tau in vitro and favorable brain penetration and clearance in rats, showing improved CNS MPO and BBB scores compared to CBD-2115. Therefore, these results indicated that further discovery of additional PET radiotracers for selective binding of 4R tau aggregates to discriminate between AD and non-AD tauopathies is warranted.

3. Discovery of [¹⁸F]OXD-2314

As discussed earlier in this review, CBD-2115, a first-in-class high selective 4R tau targeting ligand was reported recently.¹⁰⁸ Even though with high in vitro binding affinity to 4R tau aggregates, in vivo evaluation of [¹⁸F]CBD-2115 exhibited low brain uptake in rodents and non-human primates. Therefore, to improve the in vivo brain uptake and high selectivity of [¹⁸F]CBD-2115 for non-AD tau aggregates, in 2024, Lindberg and others (Svensson, Mathis and Vasdev et al.) reported a ligand-based design of [¹⁸F]OXD-2314 (ref. 113) (Fig. 33). By using a ligand-based design approach, over 150 pyridinyl-indole analogues of CBD-2115 (a.k.a. OXD-2115) were designed and synthesized for initial screening for 4R tau affinity. All leads were evaluated against [³H]OXD-2115 in in vitro human postmortem AD and PSP tissues, and computational prediction tools were used to assess the BBB penetration properties of each potential lead and found (2-(2-fluoro-6-(3-methoxypiperidin-1-yl)pyridin-3-yl)-1H-indol-5-ol), OXD-2314, as the most potent analogue. Tritium labeled OXD-2314 was used in in vitro evaluation in the brain tissues of AD, CBD, PSP, Pick's disease, and PD patients and healthy controls in comparison with first- and second-generation tritium labeled tau PET ligands AV-1451, PI-2620 and OXD-2115.

Fig. 33. Structures of OXD-2115, OXD-2314 and radiofluorinated tracer [¹⁸F]OXD-2314.¹¹³.

In a competitive binding assay, [³H]OXD-2314 was found to exhibit 2-fold potency in PSP tissue (K_D = 2.4 nM), 10-fold potency in CBD tissue (K_D = 2.1 nM), equipotency in AD tissue and good affinity in PiD tissue (K_D = 1.1 nM).¹¹³ In the competitive binding assay, OXD-2314 was displaced by either AV-1451 or PI-2620 indicative of high binding affinity for tau aggregates. In silico BBB prediction models also indicated high BBB penetration for OXD-2314 compared to the reference OXD-2115.

Radiolabeling was performed on the N-Boc protected nitro precursor using cyclotron generated [¹⁸F]F⁻ to produce [¹⁸F]OXD-2314 (Fig. 33). Dynamic PET imaging of [¹⁸F]OXD-2314 in rats and non-human primates showed good brain uptake (SUV = >2) and rapid clearance from the brain (Fig. 34). Radiometabolite studies in rats and non-human primates indicate the presence of only polar radiometabolites in blood plasma. In addition, ex vivo analysis of the brain homogenate from rats indicated low levels of radiometabolites in the brain samples not passing through the BBB. No in vivo model was reported in the study for PET imaging of tau aggregates using [¹⁸F]OXD-2314. Nevertheless, in vitro and in vivo PK results indicated that [¹⁸F]OXD-2314 is a first-in-class non-AD tau PET radiotracer evaluated in vivo and is under process for first-in-human studies upon regulatory clearance.

Fig. 34. Representative [¹⁸F]OXD-2314 PET images in the rat brain (left) and the macaque brain (right) (figures were reproduced from the original reference by Lindberg et al.,¹¹³ under Creative Commons Attribution 4.0 International License).

Binding structure determination of PET ligands to AD tau filaments using cryo-EM

In the brain of an individual with AD, tau filaments are composed of paired helical filaments (PHFs) and straight filaments (SFs).¹¹⁴ The common structural unit of PHFs and SFs is a double helical stack of C-shaped subunits, revealed by negative-stain electron microscopy.¹¹⁵ In contrast, the amino- and carboxy-terminal regions of tau likely adopt random conformations, forming a fuzzy coat. In AD, tau filaments are formed from full-length tau within cells but lose most of their fuzzy coat in extracellular ghost tangles.¹¹⁶ These filaments exhibit a cross-β structure.¹¹⁷ PHFs and SFs are differentiated by their unique inter-protofilament packing arrangements. The ordered core, characteristic of amyloid structures with rich β-sheets, spans amino acids 306–378 in the 441-amino-acid tau isoform.¹¹⁸ Structural insights into how tau PET ligands bind to disease-related tau filaments are crucial for guiding ligand optimization. Understanding the binding sites and affinities can help in the early detection and accurate diagnosis of AD, as well as better monitoring of disease progression over time.

Many tau PET ligands, such as flortaucipir or AV-1451, contain heterocyclic aromatic moieties. Although the first-generation tracer flortaucipir could not be visualized to bind to AD PHFs, it was rather intriguingly observed to bind to tau filaments with a CTE fold.¹¹⁹ Flortaucipir interacted with these CTE tau fibrils specifically within the more open curve of their C-shaped protofilaments (Fig. 35). While second-generation PET tracers have been developed to reduce off-target binding and optimize pharmacokinetic properties, their direct binding to disease-relevant tau filament-folds remains to be characterized. Additionally, a cryo-EM structure of the PET tracer APN-1607 or PM-PBB3 at relatively low resolution has been modeled to show that PET tracers bind end-to-end with the plane of the aromatic rings parallel to the fibril axis.¹²⁰ From the cryo-EM binding model, two major sites were identified in the β-helix of PHFs and SFs, as well as a third major site in the C-shaped cavity of SFs (atomic model not shown here).

Fig. 35. Flortaucipir binding to CTE type I filaments. The atomic model for tau (orange) and flortaucipir (purple) for the CTE type I filaments (image was taken from the original reference by Shi et al.,¹²⁰ under a Creative Commons Attribution 4.0 International License).

In 2023, Merz et al. reported the cryo-EM structure of tau fibrils from AD patients bound to the PET ligand GTP-1, a second-generation tau PET tracer that is being evaluated in clinical trials.¹²¹ The structure, resolved at 2.7 Å, resembles PHFs obtained from AD brain tissue, where two C-shaped protofilaments interact laterally to form fibrils, providing insights into GTP-1 binding (Fig. 36). The additional, well-resolved density identified the ligand bound to a solvent-exposed C-shaped cleft in each protofilament, comprising residues within the amyloid core (pseudo-repeats R3 and R4) and including strands β6 and β7. These strands are separated by a kink at Gly355, forming a cleft that matches the convex shape of GTP-1. Using molecular dynamics simulations and density functional theory, the authors revealed that GTP-1 interacts with the tau PHF through π–π stacking, precisely matching its phenyl ring with the side chains of residues in the binding site, with a 1 : 1 stoichiometry. Importantly, a specific orientation of the piperazine ring perpendicular to the symmetry of the fibrils is observed. This precise matching might underlie the high affinity and specificity of GTP-1. However, it remains to be determined whether this binding arrangement is preserved at the lower tracer concentrations used in PET imaging.

Fig. 36. The atomic model of tau PHFs (purple) and GTP-1 (green). Each GTP-1 molecule was oriented at an angle relative to the tau backbone (image was taken from the original reference by Merz et al.,¹²¹ under a Creative Commons Attribution 4.0 International License).

The binding structure of MK-6240, a second-generation tau PET tracer, has been revealed using cryo-EM.¹²² Incubation of extracted AD tau filaments with a high concentration of MK-6240 showed the characteristic PHFs of AD, consisting of stacks of back-to-back C-shaped protofilaments. Cryo-EM analysis showed that each C-shaped protofilament contained additional density within a small cleft bordered by residues Q351 and I360 (Fig. 37). At a 1 : 1 stoichiometry with tau monomers, MK-6240 molecules were observed to be stacked 4.8 Å apart in a slanted, staggered arrangement along the fibril axis. A significant observation was the alignment of tracer molecules within the fibril, where they were stacked diagonally and closely positioned along the fibril axis, forming what appears to be a molecular adhesive, a similar binding characteristic that has been reported for Genentech's GTP-1 tau tracer.

Fig. 37. The atomic model of the MK-6240 binding site on PHFs. The cryo-EM density (white) is compared with the atomic model of the tau fold (blue) and MK-6240 (orange) (image was taken from the original reference by Kunach et al.,^122,123 under a Creative Commons Attribution 4.0 International License).

These findings suggest that a specific orientation of the ligand, where the aromatic rings are angled relative to the fibril symmetry, could be crucial for achieving high binding affinity. This orientation may represent a fundamental motif that future small molecule targeting filaments would need to adopt to effectively interact with tau or similar fibrillary structures. Understanding this ligand orientation could guide the design of new molecules with optimized binding properties.

α-Synuclein PET radiotracers for imaging in Parkinson's disease (PD)

Neurodegenerative diseases such as Parkinson's disease (PD), dementia with Lewy bodies (DLB), and multiple system atrophy (MSA) are characterized by abnormal accumulations of α-synuclein (α-syn) in neurons, nerve fibers, or glial cells.^124,125 These diseases are collectively known as α-synucleinopathies. PD is the most common progressive neurodegenerative movement disorder, affecting around 7–10 million people worldwide; while the exact biochemical mechanisms underlying PD remain unclear,¹²⁶ several pathogenic mechanisms have been proposed and reviewed elsewhere.¹²⁷ Early and accurate diagnosis is crucial for timely intervention to slow down the neurodegenerative process, and it is particularly challenging, hindering the discovery of new therapeutic targets and the evaluation of treatment effectiveness.¹²⁸ Developing a tracer that could specifically detect α-syn deposition would revolutionize clinical trials by ensuring that the appropriate patient groups are selected and by aligning trial outcomes with the disease's underlying pathology.¹²⁹

α-Syn is a small (14 kDa) acidic protein widely distributed throughout the CNS, making up about 1% of cytosolic proteins in the brain, as well as being present in blood cells and other tissues.^130,131 Although the exact normal function of α-syn is not fully understood, it is suggested to play a role in synaptic plasticity.¹³² α-Syn is primarily located in the neocortex, hippocampus, substantia nigra, thalamus, and cerebellum.¹³³ While it can be found in the nucleus of mammalian neurons, it is mostly present in presynaptic terminals in either a free or membrane-bound state.¹³⁴ PD patients exhibit abnormal intraneuronal inclusions known as Lewy bodies (LBs) and Lewy neurites (LNs), primarily found in the substantia nigra.¹³⁵ These filaments contain α-syn as a major component. Similarly, patients with DLB have LB and LN filaments, but these are mainly distributed in the cerebral cortex, unlike in PD.¹³⁶ In MSA patients, these filamentous components are present not only in neurons but also in oligodendrocytes and the cytoplasm of nerve cells.¹³⁷

α-Syn is encoded by the synuclein alpha (SNCA) gene and can be subdivided into three main regions, each responsible for different properties: the N-terminal amphipathic region (1–60 residues), the non-amyloid-β component (NAC) hydrophobic region (61–95 residues), and the C-terminal acidic region (96–140 residues).¹³⁸ The N-terminus is characterized by amphipathic repetitions that tend to form an α-helix structure, while the NAC region is identified as the most aggregation-prone region. The C-terminus is negatively charged and is involved in Ca²⁺ binding and chaperone-like activity.¹³⁹ Through alternative splicing, α-syn exists in several isoforms.¹⁴⁰ Extensive in vivo post-translational modifications (PTMs) including phosphorylation, truncation, acetylation, or nitration have also been reported on the synuclein protein.¹⁴¹

The native α-syn is present in synaptic terminals, the nucleus of neuronal cells, mitochondria, endoplasmic reticulum (ER), Golgi apparatus (GA), and the endolysosomal system.¹⁴² Under pathological conditions, α-syn aggregation¹⁴³ begins with the accumulation of a soluble α-syn monomer into oligomers and eventually fibrils then culminates in the formation of LBs, a pathological hallmark of PD (Fig. 38). It remains unclear if a specific conformation or other processes trigger this aggregation converting soluble native α-syn into tightly stacked, β-sheet rich fibrils.¹⁴⁴ The complexity of the aggregation process, influenced by numerous parameters, makes it challenging to understand. Environmental factors, such as exposure to pesticides, heavy metals, and oxidative stress, are associated with an increased risk of PD.¹⁴⁵ Several single amino-acid mutations of α-syn, including A53T, E46K, H50Q, G51D, A30P, and A53E, have been identified in familial PD patients and are causative for early-onset pathology with different clinical symptoms.^146,147 Among these, A53T is the first hereditary mutation of α-syn discovered in Italian and Greek families with autosomal dominant and early-onset PD.¹⁴⁸ The NAC domain of α-syn, particularly the amino acid sequences 71VTGVTAVAQKTV82 or 66VGGAVVTGV74, is crucial for fibrillation; removing these sequences prevents fibril formation entirely.^146,149 PTMs also impact aggregation; for instance, less than 5% of the soluble, monomeric α-syn is phosphorylated under physiological conditions. Meanwhile, 90% of α-syn is phosphorylated in PD patients, indicating a close relationship between phosphorylation on residue S129 and α-syn aggregation.¹⁵⁰

Fig. 38. α-Synuclein in the pathogenesis of PD. Putative mechanisms leading to α-syn intracellular accumulation and the formation of Lewy bodies in motor neuronal cells. (This figure was drawn using BioRender).

α-Syn fibrils derived from patients with different α-synucleinopathies also show structural differences.^{139,151–153} PD and MSA fibrils are characterized by flat and twisted polymorphs, while DLB fibrils are cylindrical without twists.^154,155 The exact role of α-syn in PD, DLB, and MSA remains unclear. However, early α-syn oligomers and aggregates (protofibrils) formed during the aggregation process are thought to contribute to neurotoxicity by disrupting cellular homeostasis, impacting various intracellular targets, leading to mitochondrial toxicity, enhancing inflammatory responses, and causing synaptic and endoplasmic dysfunction.^156,157 Despite these challenges, some structural aspects of α-syn fibrils have been uncovered. For example, in 2018, Li et al. distinguished between two structurally different forms of fibril species, termed protofilaments.¹⁵⁵ A molecular blind docking study identified three possible binding regions of small molecules toward α-syn fibrils: binding sites 2 (Y39-S42-T44; BS2), 9 (G86-F94-K96; BS9), and 3/13 (K45-V48-H50 and K43-K45-V48-H50; BS3/13).¹⁵⁸ In 2023, Sun et al. reported a cryo-EM structure of α-synuclein fibrils containing the hereditary A53E mutation and highlighted the structural differences, not only at the interface between proto-filaments but also between residues packed within the same proto-filament.¹⁵² However, it remains to be evaluated if these structures can mimic the aggregate fibrils in vivo.

This knowledge helps to explain why developing an α-syn PET tracer is difficult. First, Aβ or tau fibrils share similar β-pleated sheets to α-syn, while the concentration of α-syn aggregates within the brain is 10- to 50-fold lower than that of Aβ or tau.^159,160 Additionally, the structural variation of α-syn between patient brain-derived in vivo aggregates and in vitro α-syn fibrils poses a significant challenge for α-syn tracer development.¹⁶¹ Structural heterogeneity between in vivo brain-derived materials from PD and MSA patients has also been reported.¹⁵³ There is considerable uncertainty regarding the predictive and clinical relevance of in vitro assays or in vivo models reliant on fibrils. Therefore, the screening of new α-syn tracers should rely solely on patient-derived materials and access to well-characterized human tissue with established α-syn pathology is restricted. Moreover, high-throughput screening (HTS) methods for such materials are currently unavailable, and this bottleneck significantly hinders the progress of PET tracer development in this field.

Development of α-syn PET tracers

A variety of chemically diverse molecules have been developed to target α-syn over the past years. These compounds have emerged from various strategies, including rational drug design and HTS approaches. Initially, fluorescent dyes thioflavin-T (ThT) and thioflavin-S (ThS) were the first used as α-syn binders, although they lacked selectivity.^162,163 Since then, numerous efforts have been made to develop α-syn binders with improved binding affinity and selectivity.¹⁶⁴ However, in August 2024, the U.S. FDA issued a ‘Letter of Support’ for the use of the α-synuclein seed amplification assay (αSyn-SAA)¹⁶⁵ biomarker in clinical trials of PD and related diseases.¹⁶⁶ Here, we review the recent discovery of some α-syn PET ligands for imaging in PD.

2-(2-[2-Dimethylaminothiazol-5-yl]ethenyl)-6-(2-[fluoro]ethoxy)benzoxazole, also known as BF227 (Fig. 39A), was initially designed as a PET ligand to image Aβ-plaques in AD patients.^167–169 Interestingly, it was then found to show a higher affinity to α-syn fibrils than binding to Aβ1–42 fibrils.¹⁷⁰ The ability of BF227 to detect α-syn in human brain tissue has been reported in PD patients. However, it was later found that BF227 could not detect any α-syn inclusions in MSA patients without developing Aβ co-pathology, suggesting that BF227 is not capable of imaging α-syn fibrils in vivo.¹⁶⁸ These studies highlight the different conformational forms and densities of α-syn in MSA and PD tissue and the influence on the outcome of PET ligand studies leading to false conclusions. PBB3, a tau PET ligand, has also been assessed by in vitro fluorescence and autoradiographic labeling of brain sections from α-synucleinopathy patients, and the present data imply that α-synuclein pathology in LB disease is undetectable by PBB3.¹⁷¹ Hence, BF227 derivatives were subsequently developed to enhance affinity and selectivity for α-syn.¹⁶⁹ In 2023, Xiang et al.¹⁷² identified a brain-permeable PET tracer F0502B (Fig. 39A) with high binding affinity for α-syn. Structure wise, their optimized compound F0502B (Fig. 39A) shares similarities with BF227 in the benzoxazole (or benzothiazole)-ethenyl segments of their molecules. In this study, they began by screening 23 commercial compounds with backbones similar to BF227. These compounds were tested in vitro with preformed fibrils (PFFs) of α-syn, tau, and Aβ. From this initial screening, the compound showing the strongest binding activity toward α-syn PFFs was selected. Following this, 21 derivatives were designed and synthesized based on the initial compound's structure. These derivatives were then tested in primary cultured neurons containing aggregates of α-syn, tau, and Aβ to evaluate their binding selectivity. Compounds showing positive results were further tested with brain sections containing α-syn deposits. Through multiple rounds of screening with in vitro fibrils, intraneuronal aggregates, mouse models, and human brain sections, the compound F0502B emerged as a promising candidate.

Fig. 39. Development of α-syn PET tracers. (A) Chemical structure of BF227 and F0502B. (B) Structure of [¹⁸F]ACI-12589; (a) transversal images at the level of the middle cerebellar peduncles in a control participant and patients with DLB, MSA-C and PD; (b) transversal images at the level of the basal ganglia in a control participant and patients with DLB, MSA-P and PD. SUVR images have been created using the occipital cortex as a reference region (this figure was reproduced from the original reference by Smith et al.,¹⁷³ under Creative Commons Attribution 4.0 International License).

F0502B is a promising lead compound for imaging aggregated α-syn in synucleinopathies, with a K_D value of 10.97 nM to α-syn fibrils. Compared to that, the affinity of F0502B to Aβ1–42 or tau fibrils was much lower (Aβ fibrils K_D: 109.2 nM, tau fibrils K_D: 120.5 nM). F0502B also showed only specific binding with PD patients but not AD or control tissues, as the K_D and B_max values for aggregated α-syn in PD patients were 3.68 nM and 14.95 fmol nmol⁻¹, separately. Autoradiography also proved that F0502B selectively and specifically associated with LBs in brain sections from PD but not AD patients. Even more strikingly, the 2.8 Å resolution structure of α-syn fibrils in complex with F0502B was also determined and it was found that F0502B inserts into a deep cavity on the fibril surface constructed from residues K80-E83 and Y39-E46 with its phenol head residing inward and the fluoro tail pointing outward. By further evaluation of the diagnostic ability of the PET tracer [¹⁸F]F0502B by using monkey models of PD, higher signals in the PFF- and virus-injected groups than the control group were observed (data not shown here). These encouraging findings support the continued research and development of [¹⁸F]F0502B as a PET tracer. It meets key criteria for an ideal α-synuclein radioligand, such as high BBB permeability, high affinity and selectivity, rapid clearance, and no binding in healthy control rhesus macaques, and holds potential for detection of α-synuclein and to monitor the effects of disease-modifying treatments.

[¹⁸F]ACI-12589 (Fig. 39B) is also a novel α-syn PET tracer for PD diagnosis, identified from the screening of AC Immune's Morphomer® platform based on its α-synuclein-binding properties, and was brought into clinical development in 2020 as a non-invasive diagnostic tool.¹⁷⁴ During an oral presentation at the Alzheimer's Association International Conference, AC Immune reported preclinical results for [¹⁸F]ACI-12589 and it displayed fast brain uptake, very low non-specific retention, rapid metabolism, and a high free fraction. In their recent publication in 2023, Hansson et al.¹⁷³ further evaluated the affinity and specificity of [¹⁸F]ACI-12589 for pathological α-synuclein. For a preclinical study, [¹⁸F]ACI-12589 showed binding affinity to both a familial PD case and a MSA case, with dissociation constants (K_D) of 17 nM and 28 nM, respectively. The authors believed that binding affinities were similar across different synucleinopathy cases. Autoradiography also showed only weak binding to β-amyloid using an AD post-mortem brain. By performing a direct saturation binding experiment with a homogenate containing β-amyloid, TDP-43 and tau aggregates, ACI-12589 was proved to have in vitro selectivity for α-syn (data not shown here). Subsequently, 42 participants were recruited to evaluate the performance of [¹⁸F]ACI-12589 in vivo (Fig. 39B). The radiotracer demonstrated good stability throughout the scan, maintaining 66.7 ± 8.4% of the MSA parent fraction at 90 minutes. In the cerebellar white matter and middle cerebellar peduncles of participants, a high PET signal for [¹⁸F]ACI-12589 was observed. Compared to those with parkinsonism-dominant MSA (MSA-P), retention in cerebellar structures was greater in cerebellar ataxia-dominant MSA (MSA-C) participants. Basal ganglia involvement was possible as tracer uptake in the lentiform nuclei of MSA-P. In general, [¹⁸F]ACI-12589 is a useful tool for the diagnostic work-up of MSA.

Benzothiazoles are established pharmacophores for targeting α-syn. A study on 2-styrylbenzoxazole by Verdurand et al. demonstrated¹⁷⁵ high affinity for α-syn (K_D = 3.3 ± 2.8 nM) with moderate selectivity over Aβ but it failed to bind in human brain sections. To develop a more effective α-syn PET tracer, a library of 2-styrylbenzothiazoles was designed and screened. Di Nanni et al. identified¹⁷⁶ two promising compounds: PFSB (K_i = 25.4 ± 2.3 nM) and its less lipophilic analogue, MFSB, with even higher α-syn affinity (K_i = 10.3 ± 4.7 nM), both selective over Aβ (Fig. 40). In vitro autoradiography on human brain sections showed up to four times higher specific binding in MSA cases, confirming selective α-syn binding. In vivo PET imaging demonstrated that [¹⁸F]MFSB crosses the BBB (data not shown here).¹⁷⁶

Fig. 40. Structures of [¹⁸F]PFSB and [¹⁸F]MFSB.¹⁷⁶.

3,5-Diphenylpyrazole is also a promising scaffold to develop small molecules that detect aggregated α-syn. Diphenylpyrazole based drug candidates, such as anle138b and anle253b,^177,178 have been proved to inhibit pathological aggregation of α-syn in vivo by binding to α-syn fibrils. MODAG-001 (clog P = 3.85) is a new derivative within this compound class, which is generated by modifying anle253b by exchanging the bromophenyl moiety with bromopyridine to reduce lipophilicity. MODAG-001 showed strong binding to α-syn fibrils (K_D = 0.6 ± 0.1 nM) and a moderate affinity to tau (K_D = 19 ± 6.4 nM) and Aβ fibrils (K_D = 20 ± 10 nM).¹⁷⁹ [¹¹C]MODAG-001 displayed BBB permeability in mice but was not able to detect aggregated α-syn in brain sections of DLB patients. [¹¹C]MODAG-005 has recently been published with good brain penetration, rapid clearance from brain tissue and low metabolite formation in rodents and non-human primates, but it has only been evaluated on one first-in-human patient with clinically established MSA (data not shown here).¹⁸⁰

Low molecular weight arylpyrazolethiazole derivatives were designed based on structure–activity relationship analysis of known ligands like anle138b to achieve high selectivity for α-synuclein (α-syn) fibrils while optimizing physicochemical properties suitable for PET tracers. Using in vitro competition binding assays with [³H]MODAG-001 against recombinant α-syn and Aβ1–42 fibrils, Bonanno et al. identified¹⁸¹ APT-13, which exhibited an inhibition constant of 27.8 ± 9.7 nM and a selectivity greater than 3.3-fold for α-syn over Aβ. Radiolabeled [¹¹C]APT-13 demonstrated strong brain penetration and rapid clearance in healthy mice, positioning it as a promising lead compound for PET tracers targeting α-syn aggregates in vivo (data not shown here). This compound class holds potential for further SAR studies and will require additional evaluation in clinical settings.¹⁸¹

The pyridothiophene scaffold, initially patented by AC Immune as part of a diverse group of bicyclic compounds tested for binding to α-syn and Aβ plaques, demonstrated strong affinity for α-syn with minimal Aβ staining. A pyridothiophene-based compound shows promise for investigating PET radiotracers targeting α-synuclein. In their 2024 publication, Pees et al. synthesized¹⁸² and evaluated 47 derivatives of the lead compound asyn-44 (Fig. 41) (K_D = 1.85 nM), using autoradiography with [³H]asyn-44. Several of these derivatives exhibited low nanomolar K_i values ranging from 12 to 15 nM. Autoradiography studies confirmed that [³H]asyn-44 effectively binds to brain sections from donors with MSA and PD. Preliminary PET imaging with [¹⁸F]asyn-44 in healthy rats demonstrated high initial brain uptake (>1.5 SUV) and moderate washout (∼0.4 SUV at 60 minutes). However, high levels of radiometabolites hindered further PET studies. Developing additional pyridothiophene derivatives with enhanced affinity and metabolic stability will be crucial for advancing the use of this tracer in understanding these diseases.¹⁸²

Fig. 41. Structure of asyn-44.¹⁸².

Other PET radiotracer candidates for imaging α-synuclein include 4-methoxy-N-(4-(3-(pyridin-2-yl)-3,8-diazabicyclo[3.2.1]octan-8-yl)phenyl)benzamide (Fig. 42), which was reported by Kim et al. in 2023.¹⁸³ This compound was inspired by BF2846, as its congeners demonstrated a high binding affinity for α-syn in radioligand binding assays and ex vivo autoradiography studies.¹⁸⁴In vitro assays indicated that this probe binds strongly to recombinant α-synuclein fibrils (K_i = 6.1 nM) while exhibiting low affinity for Aβ fibrils. Additionally, it showed significant specific binding to tissues affected by tauopathies, as well as to those from PD and MSA, suggesting an affinity for tau aggregates. Notably, the specific binding to pathological α-syn aggregates in post-mortem brain tissues from MSA patients was significantly higher than that observed in PD tissues, as demonstrated by nuclear emulsion and autoradiography in tissue sections. Nonhuman primate (NHP) PET studies (Fig. 42) further confirmed its good brain uptake and rapid washout, which proves the potential of [¹¹C]4-methoxy-N-(4-(3-(pyridin-2-yl)-3,8-diazabicyclo[3.2.1]octan-8-yl)phenyl)benzamide as a PET tracer for imaging α-syn in MSA patients.¹⁸³

Fig. 42. Structures of the PET ligand 4-methoxy-N-(4-(3-(pyridin-2-yl)-3,8-diazabicyclo[3.2.1]octan-8-yl)phenyl)benzamide (top left) and its [¹¹C] labelled ligand (bottom left). The dynamic PET study in NHPs to investigate the pharmacokinetics of [¹¹C]4-methoxy-N-(4-(3-(pyridin-2-yl)-3,8-diazabicyclo[3.2.1]octan-8-yl)phenyl)benzamide as an in vivo brain PET tracer (right) (the above figure was reprinted with permission from the original article by Kim et al.¹⁸³ Copyright (2023) American Chemical Society).

As previously mentioned, PBB3, a PET ligand designed to detect pathological tau aggregates, was found to bind to α-synuclein lesions in patients with α-synucleinopathies. Endo et al.¹⁸⁵ then screened a series of PBB3 derivatives featuring an (E)-hex-2-en-4-yne linker, known as C05 series compounds, which demonstrated high reactivity and selectivity for α-synuclein pathologies compared to PBB3 and BF-227 (Fig. 39A). Among these, the small molecular compound C05-05 (Fig. 43) exhibited more favorable properties as an in vivo imaging agent. C05-05 successfully detected α-synuclein inclusions in living murine and NHP models through optical and PET imaging, showing slightly slower brain clearance than PBB3. Its high binding affinity for α-synuclein inclusions in brain tissues from PD, DLB, and MSA supports its clinical PET potential. Exploratory clinical PET assays confirmed that [¹⁸F]C05-05 (Fig. 43) can detect α-synuclein deposits in the midbrains of PD and DLB patients when compared with healthy controls. Their findings provide essential information on the PET probe's ability to capture Lewy-type α-syn pathologies in humans (Fig. 43).¹⁸⁵ As of now, no additional data are available on this probe.

Fig. 43. Structures of the PET ligand C05-05 and its radiolabeled version [¹⁸F]C05-05 (left). PET imaging of α-syn pathologies in MSA patients with [¹⁸F]C05-05 (right)¹⁸⁵ (this figure was reproduced from the original article under a Creative Commons license).

Summary and perspectives

Neurodegenerative disorders AD and PD are chronic conditions caused by the accumulation of protein aggregates, such as Aβ, NFTs and a-synuclein. Discovery of highly selective PET radiotracers for early, non-invasive detection of these diseases has been of interest for many research groups across the world. Early diagnosis of these diseases helps not only to assess the disease progression but also to aid in the discovery and development of disease modifying therapeutics for clinical use. Therefore, initial PET radiotracer discovery for imaging Aβ plaques in AD patients began with the identification of Pittsburgh compound-B (PiB) or [¹¹C]PiB, derived from well-known histological staining agent thioflavin-T. On the other hand, PET imaging of NFTs in AD has become the most powerful diagnostic tool to diagnose early symptoms of AD, and the research in this area has spurred from the discovery of [¹⁸F]FDDNP as a potential PET tracer for imaging Aβ plaques and NFTs. Several first generation tau PET tracers have been developed ([¹¹C]THK-5351, [¹⁸F]THK-5317, [¹¹C]PBB3) and evaluated in clinical settings including the FDA approved tracer [¹⁸F]AV-1451 (Tauvid™). [¹⁸F]AV-1451 displayed >27-fold selectivity toward NFTs over Aβ plaques but showed off-target binding to MAO-A. To address downside issues associated with the first-generation PET tracers (t_1/2, photoisomerization, off-target binding to white matter or MAO-A/B, etc.), second generation PET tracers have been discovered including [¹⁸F]RO-948, [¹⁸F]MK-6240, [¹⁸F]PI-2620, [¹⁸F]PM-PBB3, [¹⁸F]JNJ-311 and [¹⁸F]GTP1, and these are being evaluated in human subjects/clinical settings. Moreover, these second-generation PET tracers showed less off-target binding to the basal ganglia. [¹⁸F]MK-6240 showed very low off-target binding against white matter and MAO-A/B compared to [¹⁸F]AV-1451 which had high affinity for the MAO-A enzyme. Further chemical modifications and regio-isomer exploration on [¹⁸F]AV-1451 resulted in [¹⁸F]RO-948 and [¹⁸F]PI-2620. These two PET tracers showed reduced MAO-A binding, favorable tau selectivity and improved pharmacokinetics. Moreover, [¹⁸F]PI-2620 has been shown to bind to tau aggregates in AD and non-AD brain homogenates such as PSP and CBD, where the 4R isoform is predominant. Meanwhile, [¹⁸F]PI-2620 has also been shown to detect the 3R tau isoform in Pick's disease brain homogenates and [¹⁸F]PM-PBB3 has been shown to detect tau aggregates in AD, PSP and FTLD. However, PET imaging of tau aggregates in non-AD tauopathies has been limited and existing tracers [¹⁸F]PI-2620 and [¹⁸F]PM-PBB3 lack selectivity for either the 3R or 4R tau isoform. In this regard, [¹⁸F]OXD-2314 has been developed to image aggregated tau in non-AD tauopathy of PSP with high selectivity and specificity and with favorable brain kinetics and radio-metabolite profiles in rats and non-human primates. There is still much to be accomplished in the context of developing highly selective tau PET tracers to detect non-AD tauopathies where the 3R or 4R isoform is predominant.

Therefore, novel approaches are needed for PET ligand discovery, and it can be achieved by a) leveraging alpha-fold,^186,187 a new machine learning based approach, to rationalize the tau protein folding pattern to discriminate 3R vs. 4R tau isoforms; b) additional investigations pertaining to structural and binding site elucidation of tau protein isoforms using cryo-EM;^{122,188–191} c) utilizing new synthetic methodologies to incorporate radiolabels¹⁹² onto the newly discovered tau specific PET ligands with high radiochemical yields and purity, such as photoredox and SuFEx chemistry (SuFEx chemistry is one of the emerging biorthogonal ‘click’ reactions to attach fluorine atoms onto a chemical compound); d) additional investigations on the use of chiral functional groups on the PET ligand that may be advantageous for discriminate binding of 3R vs. 4R isoforms as evidenced by the OXD-2314 tracer.

The ongoing development of α-syn PET tracers marks significant progress in our ability to visualize and potentially diagnose α-synucleinopathies. These advancements have provided valuable insights into the pathophysiology of diseases such as PD and MSA. Several critical challenges remain in clinical settings. One of the primary challenges is ensuring the predictive and clinical relevance of α-syn PET tracers. Validating their ability to accurately detect and quantify α-synuclein aggregates in vivo is essential and requires robust comparative studies with gold standard methods and correlation with clinical outcomes. The application of high-throughput screening methods to evaluate potential tracers against patient-derived material is currently constrained, slowing down the development process. Addressing these challenges requires collaborative efforts across disciplines including neurology, radiology, and biochemistry to refine tracer design, enhance imaging protocols, and establish standardized evaluation criteria.

Notes and abbreviations

% ID g⁻¹

The percent of injected dose per gram of tissue

CBD

Corticobasal degeneration

AD

Alzheimer's disease

Aβ

β-Amyloid peptides

BBB

Blood–brain barrier

B _max

Total density (concentration) of receptors in each sample of tissue

CERAD

Consortium to Establish a Registry for Alzheimer's Disease

CNS

Central nervous system

CNS-MPO

Central nervous system multiparameter optimization

CSF

Cerebrospinal fluid

CYP1A2

Cytochrome P450 1A2

DLB

Dementia with Lewy bodies

ER

Endoplasmic reticulum

FTLD

Frontotemporal lobar degeneration

GA

Golgi apparatus

HC

Healthy control

HTS

High-throughput screening

HWE

Horner–Wadsworth–Emmons

K _D

Dissociation constant

LBs

Lewy bodies

LNs

Lewy neurites

MAO

Monoamine oxidase

MAPT

Microtubule-associated protein tau

MBD

Microtubule-binding domain

MCI

Mild cognitive impairment

MMSE

Mini mental state exam

MSA

Multiple system atrophy

NAC

Non-amyloid-β component

NFTs

Neurofibrillary tangles

OC

Older control

PD

Parkinson's disease

PET

Positron emission tomography

PHF

Paired helical filament

PiB

Pittsburgh compound B

PK

Pharmacokinetics

PSP

Progressive supranuclear paralysis

pTau

Hyperphosphorylated tau

PTMs

Post-translational modifications

RCF

Randomly coiled filament

SAR

Structure–activity relationship

SFs

Straight filaments

SNCA gene

Synuclein alpha gene

SPECT

Single-photon emission computed tomography

SuFEx

Sulfur(vi) fluoride exchange

SUV

Standardized uptake value

SUVR

Standardized uptake value ratio

TFs

Twisted filaments

ThS

Thioflavin-S

ThT

Thioflavin-T

YC

Younger control

α-syn

α-Synuclein protein

Data availability

No primary research results, software or code have been included and no new data were generated or analyzed as part of this review article. All images in this review article are reproduced from original articles under the Creative Commons Attribution 4.0 International License.

Conflicts of interest

The authors claim that there is no conflict of interest related to the manuscript.

Biographies

Biography

Shekar Mekala.

Shekar Mekala obtained his Ph.D in Organic Chemistry (2012) from the University of Oklahoma, Norman. He then moved to upstate New York to pursue his postdoctoral research at Syracuse University, and in the Center for Biotechnology & Interdisciplinary Studies at Rensselaer Polytechnic Institute (RPI), Troy, New York. In 2018, he joined the laboratory of Dr. Yue-Ming Li as a Research Fellow in the Chemical Biology program at Memorial Sloan-Kettering Cancer Center (MSK), New York. Currently, he is a Senior Research Scientist at MSK and his research has been focused on the a) early stage drug discovery, hit to lead generation and lead optimization of small molecule autophagy modulators for neurodegenerative disorders focusing on Alzheimer's disease (AD); b) design of chemical probes for novel protein target identification and validation; c) discovery of novel PET tracers for imaging of ‘tau’ protein in AD and non-AD tauopathies. He is originally from India and obtained master's degree in chemistry from Osmania University.

Biography

You Wu.

You Wu is a PhD candidate at Memorial Sloan Kettering Cancer Center, affiliated with the Tri-Institutional PhD Program in Chemical Biology. She earned her bachelor's degree in chemical biology from Nankai University. In 2022, she joined Dr. Yue-Ming Li's laboratory to continue her passion for Chemical Biology. Her research focuses on developing chemical probes to selectively target and label pathological tau protein associated with Alzheimer's disease. Additionally, she's also investigating the relationship between amyloid-beta accumulation and tau hyperphosphorylation.

Biography

Yue-Ming Li.

Yue-Ming Li is a Member and Professor of Chemical biology at Memorial Sloan-Kettering Cancer Center, and Professor of Pharmacology and Neuroscience at Weill Graduate School of Medical Sciences of Cornell University. He received his Ph.D. degree from the University of California, Berkeley and postdoctoral training at Harvard Medical School. Prior to coming to Memorial Sloan-Kettering Cancer Center in 2002, he was a scientist at Merck Research Laboratories. His lab studies on aging-related human disorders, such as Alzheimer's disease and cancer.