Abstract
Background:
Flortaucipir (AV-1451) and pyridinyl-butadienyl-benzothiazole 3 (PBB3) are newly developed and commonly used positron emission tomography (PET) tracers to detect tau deposition in tauopathies, including frontotemporal dementia (FTD). [18F]PM-PBB3, as a second-generation compound, has not been described in FTD so far.
Objective:
We aim to explore the in vivo performance of [18F]PM-PBB3 tau PET in an FTD case caused by microtubule-associated protein tau (MAPT) mutation and compare the binding to different tau strains between AV-1451 and PBB3.
Methods:
We reported the clinical and FDG, [18F]AV45 amyloid and [18F]PM-PBB3 tau PET findings in a patient with FTD of P301L MAPT mutation. Based on our results and published data, we summarized and compared the different utilities of tau PET tracers of AV-1451 and PBB3 in FTD with MAPT mutation.
Results:
The patient demonstrated slightly diffuse [18F]PM-PBB3 tau deposition in cerebral lobes especially in the left frontal lobe overlapping with the hypometabolic region detected by FDG PET. From our analysis of 35 FTD patients with MAPT mutation who underwent tau PET, AV-1451 was positive in all (n = 11) patients with mutations known to cause three and four repeat (3R/4R) tau deposition and in 14.3% (n = 2/14) of 4R tauopathies, while positive PBB3 retention was found in all patients with both 3R/4R (n = 2) and 4R (n = 8) tau.
Conclusions:
[18F]PM-PBB3 tau PET assisted the diagnosis of FTD with P301L MAPT mutation, and might be useful in the in vivo detection of both 3R/4R and 4R tau domains in the brain of FTD with MAPT mutation.
INTRODUCTION
Frontotemporal dementia (FTD) is associated with the deposition of hyperphosphorylated tau protein known as the microtubule-associated protein tau (MAPT) in neurons and glia cells of the brain [1]. MAPT gene mutation is one crucial hereditary cause for FTD, of which the most common types are P301L, N279K, and intron 10+16 mutation [2]. While in east Asia, FTD with P301L mutation is not as frequently reported in western countries and the clinical spectrum remains to be elucidated [3].
Different MAPT mutation types result in neuropathological varieties of tau isoforms with 3 repeats (3R) and/or 4R. Mutations in exon 10 only affect 4R tau, while missense mutations in exons 1, 9, 11, 12, and 13 alter the ratio of 3R/4R tau [4]. Several positron emission tomography (PET) tracers including flortaucipir (AV-1451) and pyridinyl-butadienyl-benzothiazole 3 (PBB3) could detect in vivo tau deposition in the brain with various affinities to 3R or 4R tau isoforms [5]. AV-1451 demonstrated high binding affinity to paired helical filaments (PHF) composed of 3R/4R tau which was widely used in Alzheimer’s disease [6], while it was less effective in binding to straight filaments with only 4R tau in brain sections from progressive supranuclear palsy (PSP), corticobasal degeneration or FTD [7, 8]. The use of AV-1451 PET in FTD patients with the mutation known to cause 4R tauopathy was still controversial [7, 9]. To note, some cases suggested that PBB3 tracer had great advantages of detecting a broad range of tau inclusions in human brains [5, 10]. For brain samples from patients with FTD and parkinsonism linked to chromosome 17, PBB3 fluorescence sensitively detected ex vivo 3R/4R and 4R tau in brain sections of patients with G272V and N279K MAPT mutation respectively, while AV-1451 showed very weak signals [11]. A study also revealed a good binding potential of [18C]PM-PBB3 and significant relationship between its deposition in frontoparietal white matter and the severity of clinical symptoms in PSP patients [12]. Compared to 11C-labeled agent, [18F]PM-PBB3 was considered to be more rapid, sensitive, and accurate in screening tau burden and have less off-target signals [5]. Up to now, the imaging feature of [18F]PM-PBB3 tau PET has not been reported in familial FTD patients.
Here we described the detailed clinical and neuroimaging features of a behavioral variant FTD (bvFTD) case carrying P301L MAPT mutation, who underwent [18F]FDG, [18F]AV45 amyloid, and [18F]PM-PBB3 tau PET. The relationship between regional PET standardized uptake value ratio (SUVr) and cortical thickness across the same cerebral regions was analyzed. Then we summarized the AV-1451 and PBB3 tau PET results of FTD patients with MAPT mutations from a literature review.
METHODS
Clinical data collection and patient consents
We personally inquired about the detailed medical history of the family and examined the proband and her younger brother. The study has been approved by the institutional review board of Huashan Hospital Fudan University and the patient and caregivers signed an informed consent form. DNA from peripheral blood samples was extracted and was sequenced by the second-generation sequencer Illumina Hiseq X-ten. Sanger sequencing was performed in the detected mutation sites to exclude false positive and to verify the presence of co-segregation in the proband’s family member.
MRI acquisition and cortical thickness calculation
The proband was scanned at the radiology department of Huashan Hospital on a Siemens MAGNETOM Verio 3.0T MRI. Parameters of 3-dimensional (3D) T1 magnetization prepared rapid acquisition gradient echo (MPRAGE) sequence were as follows: repetition time/echo time = 1780/2.93 ms, flip angle is 0, FOV 24 cm, in-plane matrix 352×512 and slice thickness 1.00 mm. Cortical thickness of 68 predefined regions of interests (ROIs) was calculated from T1 images using Freesurfer 6.0 (http://surfer.nmr.mgh.harvard.edu/fswiki/).
PET acquisition and processing
3D PET images were acquired by a Siemens Biograph 64 PET/CT scanner at PET center of Huashan Hospital. The cerebral glucose metabolism, amyloid and tau distribution were measured 60 min, 90 min, and 90 min after the intravenous injection of 185 MBq (0.1 mCi/kg) [18F]FDG, 370 MBq (10 mCi) [18F]AV 45 and 370 MBq (10 mCi) [18F]PM-PBB3 and lasted for 8, 20, and 20 min separately. There was at least one day interval for each PET scan and a delay of no more than 10 days between PET and MRI scans. Visual inspection for PET images of the proband was performed by an experienced nuclear medicine physician. We calculated the SUVr of 68 predefined ROIs using cerebellar grey matter as reference with SPM12 (http://www.fil.ion.ucl.ac.uk/spm) and cat12 (http://www.neuro.uni-jena.de/cat/) software implemented in MATLAB 2018b (Mathworks Inc., MA). Briefly, all PET images were coregistered with T1 images, spatially normalized in the Montreal Neurological Institute template and then smoothed.
Statistical method
The Spearman’s correlation analysis was used to assess the relationship between regional [18F]FDG, [18F]PM-PBB3 SUVr, and corresponding cortical thickness using SPSS v20.0 (SPSS Inc., Chicago, IL).
Search strategy and study selection for literature review
Our aim was to select all cases of clinically diagnosed FTD with MAPT mutation who underwent in vivo tau PET examination in case reports or cohort studies. Articles published in English were carefully searched from the PubMed and Cochrane Library databases from January 2013 to February 2020, since the tau PET tracer of AV-1451 or PBB3 was first reported in 2013 [13]. The following terms were used as keywords: “(tau PET OR AV-1451 OR flortaucipir OR PBB3 OR pyridinylbutadienyl-benzothiazole 3) AND (frontotemporal dementia OR frontotemporal degeneration OR MAPT mutation)”. We hand screened references of articles or reviews on this subject and identified additional citations. For repeated cases in different papers, the selection tendency was given to the one with more details. The flow chart of the review was shown in Fig. 1.

Flow chart of the article selection procedure.
Clinical information was extracted from each article as followings: family history, type of MAPT mutation, clinical phenotype, tau PET tracers, pathology, and major findings. We categorized the patients into 4R or 3R/4R tau pathology according to the MAPT mutation type or pathological findings if possible, and compared the major findings from these studies, trying to specify the different utility of AV-1451 or PBB3 tau PET in distinct kinds of tauopathies.
RESULTS
Clinical presentations
Proband (patient III.8)
The patient was a right-handed Chinese woman with 11 years of education. She worked as manager before retirement. She had personality changes like becoming more irritable and caring less about other family members since age 53. Later, she developed impulsive actions, inappropriate social behavior, anomic aphasia and memory impairment, and could not live independently.
On neurological examination, the patient showed agitation and diminished attention. She had dysfluent speech, impaired word-finding ability, and semantic paraphasia. Limb muscle strength was normal. Unsteady gait, bradykinesia, ataxia, ocular movement problems, limb rigidity, or cortical sensory loss were not found. She was unable to cooperate in the Mini-Mental State Examination test. The Frontal Behavioral Inventory score was 32 by inquiry to her husband. The proband’s brain MRI showed symmetrical frontal and temporal lobes atrophy. Thus, the patient met the international consensus criteria for probable bvFTD [14].
Family history
The proband’s maternal grandmother (patient I.2) had behavioral abnormalities and died from unknown causes. The proband’s mother (patient II.2) became irritated since 40 s and died at age 60 because of appendicitis. One of her older sisters (patient III.2) died of cancer in her 20 s. Another older sister (patient III.6) became disinhibited at 55 since when she easily quarreled with others and died at age 60. The proband’s younger brother (patient III.9) gradually exhibited slow reaction and a decline in naming objects and then memory loss at age 52. His Mini-Mental State Examination score was 16 with 12 years’ education. MRI showed brain atrophy predominantly in bilateral temporal lobes (Fig. 2).

Pedigree of the patient’s family.
Genetic tests
The P301L MAPT mutation was identified in the proband and her younger brother (patient III.9).
PET results
From visual assessment, for FDG PET, the radioactive uptake was significantly decreased in bilateral temporal lobes and moderately decreased in bilateral frontal lobes with more remarkable in the left side. Amyloid PET was negative in this patient. The [18F]PM-PBB3 PET showed slightly diffuse tracer binding in cerebral lobes, especially in the left frontal cortex overlapping with the hypometabolic area in FDG PET, and much higher tracer retention in the bilateral thalamus, midbrain, and pons (Fig. 3). However, we did not find the relevance between cortical thickness and cerebral [18F]FDG SUVr (rho = –0.070, p = 0.573) or [18F]PM-PBB3 SUVr (rho = 0.206, p = 0.092), or between cerebral PBB3 SUVr and FDG SUVr (rho = 0.004, p = 0.977) in quantitative analysis.

FDG, tau, and amyloid PET - MRI images in the proband. [18F]FDG PET - MRI, [18F]PM-PBB3 PET –MRI, and [18F]AV45 PET - MRI infusion images of the proband were shown.
The usefulness of tau PET in FTD with MAPT mutation
A total of 35 cases were collected for the analysis (see Table 1), including 25 cases with [18F]AV-1451, nine cases with [11C]PBB3, and our case with [18F]PM-PBB3 tau PET tracer. Tau PET tracer uptake in the frontal and temporal lobes with or without distribution of tracer retention corresponding to pathology, glucose hypometabolism, or brain atrophy was defined as positive for tau PET. For [18F]AV-1451, 52% of the FTD cases (13/25) were positive, of whom 84.6% (11/13) with the mutation of R406W, V337M, and Q351R known to cause 3R/4R tau aggregation in the brain. While among the negative cases, all carried the MAPT mutation associated pathologically with 4R tau deposition, comprising P301L, S305N, N279K, and IVS 10+16. For PBB3, there was [11C]PBB3 retention in nine cases including seven cases with 4R tau and two cases with 3R/4R tau (five cases without detailed information) and mildly elevated [18F]PBB3 deposition in our case.
Tau PET in FTD with MAPT mutations
MAPT, microtubule-associated protein tau; 3R, 3-repeat; 4R, 4-repeat; FTD, frontotemporal dementia; bvFTD, behavioral variant FTD; FTDP, FTD with parkinsonism; FTDP-17, FTD with parkinsonism linked chromosome 17; Park, parkinsonism; PPND, pallidopontonigral degeneration; SUVr, standardized uptake value ratio; NA, not available. *Pathology in the pathology-PET imaging correlation analysis was from the proband’s father without genetic tests. #Pathology in the pathology-PET imaging correlation analysis was from in proband’s affected mother and unrelated the same mutation carrier. aThis article was from a poster presentation at the 2016 Alzheimer’s Association International Conference.
Pathology-tau PET imaging correlation analyses have been performed in nine patients. Two cases further supported the usefulness of [18F]AV-1451 in 3R/4R tauopathies, although one case’s pathology was from the proband’s affected mother and unrelated the same mutation carrier [1, 15]. Two other cases showed conflicting results of [18F]AV-1451 in 4R tauopathies. Neither in vivo detectable tau signals nor correlation with ex vivo tau measures from the brain section was found in one case [7], while the other showed abnormal tau signal in cerebral lobes recapitulating tau pathology in proband’s father [9]. The remaining five cases of the total nine cases in favor of PBB3 binding to various tau inclusions comprising 4R and 3R/4R tau were from a poster presentation at the 2016 Alzheimer’s Association International Conference [16].
DISCUSSION
Here we reported the clinical and structural MRI and PET imaging findings of brain glucose metabolism, amyloid, and especially tau protein with a new generation tracer of [18F]PM-PBB3 in a case of bvFTD with family history and P301L MAPT mutation, and further explored the relationship between FDG, PBB3 retention, and cortical atrophy. Then the usefulness of AV1451 and PBB3 tau PET in FTD with MAPT mutation was summarized from a review.
For [18F]PM-PBB3 tracer, we found the increased uptake in the left frontal lobe consistent with the lowered metabolic area, and in the thalamus, midbrain, and pons of the proband. Frontal and/or anterior temporal atrophy on MRI and/or hypometabolism on PET has been added to the prerequisite of probable bvFTD in the diagnostic criteria [14]. However, from the quantitative analysis we did not observe their relationships with the tau deposition in cerebral lobes in our case. Recently, a study using [11C]PBB3 tau tracer showed disease progression greatly influenced the distribution and severity of tau deposition in the brain of FTD patients with N279K mutation [17]. There was regional PBB3 binding restricted to midbrain and medial temporal areas in patients with slow deterioration and more widespread retention in rapidly worse conditions [17]. This might explain the relatively limited tau distribution in cerebral lobes of our patient with three years’ interval between tau PET examination and disease onset. Follow-up [18F]PM-PBB3 PET could help further confirm this finding.
There was unusually increased PBB3 retention in the bilateral thalamus, midbrain, and pons of our patient. We first considered that it might be related to the off-target binding. As reported, common off-target sites of the second-generation tau agents [18F]PM-PBB3 included midbrain, thalamus, hippocampus, choroid plexus, and venous sinus, possibly due to a tendency to combine to monoamine oxidase, neuromelanin-rich, or mineralized tissue [5, 18]. In our routine examinations for patients suspected of Alzheimer’s disease or cerebral amyloid angiopathy, we also observed PBB3 retention in the thalamus and brain stem (data not shown). On the other hand, some subtypes of FTDs featured by tau isoforms with 4R domains showed tau deposition in the deep brain, such as PSP in the globus pallidus, putamen, subthalamic nucleus, and midbrain [19]. But our patient demonstrated no clinical evidence of unsteady gait, impaired ocular movement, or atrophy of the midbrain or superior cerebellar peduncles on structural MRI [20], and P301L MAPT mutation was well reported to be mainly related to FTD in East Asia [3] and western countries [21]. A pathology-tau PET correlation study is in great need to further testify this finding.
From our review, positive [18F]AV1451 retention was predominantly found in patients with 3R/4R tau of PHF. These higher signals were more likely observed in Alzheimer’s disease, well correlating with postmortem neurofibrillary tangles Braak stating in the brain [22], clinical severity of dementia [23], and regional gray matter atrophy [24]. However, it failed to show detectable signals in the brain with a greater proportion of 4R tau accumulation, such as cases with S305N, N279K, IVS 10+16 mutation, and three of four cases with P301L mutation.
Ten cases including ours confirmed in vivo application of PBB3 in FTD patients with both 4R and 3R/4R tauopathies, five of which with ex vivo autoradiography. PBB3 tracer retention was not only well associated with tau aggregation in brain sections, cognitive decline and gray matter atrophy in patients with Alzheimer’s disease [25], but observed in patients with PSP [19] and corticobasal degeneration [10]. Some researchers pointed out that the rapid metabolism of [11C]PBB3 might limit its in vivo binding [12]. [18F]PM-PBB3 was developed as the second generation of tau PET tracers, indicating a greater signal-to-background ratio and less off-target binding than [11C]PBB3 [5]. Thus, future cohort studies for tau PET may need to distinguish 4R and 3R/4R tauopathies. There are also interesting directions for designing head-to-head comparisons of PBB3 and AV-1451 PET and tau PET-neuropathology studies to confirm the different affinities of tau PET tracers.
Our study has several limitations. First, other family members of this proband refused further imaging examination; therefore, a whole [18F]PM-PBB3 tau PET imaging picture of this pedigree was not available. Second, since [18F]PM-PBB3 was documented to show higher affinity to 3R and 4R tau deposition and might have less off-target binding than [18F]AV-1451, we were unable to carry out a head-to-head comparison between [18F]PM-PBB3 and [18F]AV-1451. Third, [18F]PM-PBB3 PET data in healthy people were not available as controls.
Conclusions
We described the [18F]PM-PBB3 tau PET features in a bvFTD patient with P301L mutation and further explored the superiority of PBB3 in both 3R and 4R tauopathies from literature review. Further in vivo serial tau PET and neuropathology correlation studies are in great need to confirm the potentials of [18F]PM-PBB3 in clinical diagnosis, disease staging, and severity assessment.
Footnotes
ACKNOWLEDGMENTS
We thank APRINOIA Therapeutics for the precursor used in the patient of the case. This study is funded by National Key R&D Program of China (2016YFC1300503, 2017YFC1308201), National Natural Science Foundation of China (81971123), Shanghai Municipal Science and Technology Major Project (No. 2018SHZDZX01), and ZJLab.
