Pathological correlations of [F-18]-AV-1451 imaging in non-alzheimer tauopathies

Abstract

Objective: Recent studies have shown that positron emission tomography (PET) tracer AV-1451 exhibits high binding affinity for paired helical filament (PHF)-tau pathology in Alzheimer's brains. However, the ability of this ligand to bind to tau lesions in other tauopathies remains controversial. Our goal was to examine the correlation of in vivo and postmortem AV-1451 binding patterns in three autopsy-confirmed non-Alzheimer tauopathy cases.

Methods: We quantified in vivo retention of [F-18]-AV-1451 and performed autoradiography, [H-3]-AV-1451 binding assays, and quantitative tau measurements in postmortem brain samples from two progressive supranuclear palsy (PSP) cases and a MAPT P301L mutation carrier. They all underwent [F-18]-AV-1451 PET imaging before death.

Results: The three subjects exhibited [F-18]-AV-1451 in vivo retention predominantly in basal ganglia and midbrain. Neuropathological examination confirmed the PSP diagnosis in the first two subjects; the MAPT P301L mutation carrier had an atypical tauopathy characterized by grain-like tau-containing neurites in gray and white matter with heaviest burden in basal ganglia. In all three cases, autoradiography failed to show detectable [F-18]-AV-1451 binding in multiple brain regions examined, with the exception of entorhinal cortex (reflecting incidental age-related neurofibrillary tangles) and neuromelanin-containing neurons in the substantia nigra (off-target binding). The lack of a consistent significant correlation between in vivo [F-18]-AV-1541 retention and postmortem in vitro binding and tau measures in these cases suggests that this ligand has low affinity for tau lesions primarily made of straight tau filaments.

Interpretation: AV-1451 may have limited utility for in vivo selective and reliable detection of tau aggregates in these non-Alzheimer tauopathies. ANN NEUROL 2017;81:117-128.

Figure 1

Coronal in vivo [F-18]-AV-1451 PET images from the three study subjects (A–C). PET images from a cognitively normal 85-yo male (D) and a 78-yo male clinically diagnosed with mild AD (E), using the same color scaling (SUVR values ranging from 1 to 2), are shown for comparison. The three study subjects exhibited increased [F-18]-AV-1451 in vivo retention predominantly in basal ganglia and midbrain (subject #1 PSP (A), subject #2 PSP (B) and subject #3 MAPT P301L mutation carrier (C)). Elevated uptake in basal ganglia and midbrain, and some lower retention in the medial and inferior temporal regions was also observed in the 85-yo cognitively normal control (D). PET imaging from a 78-yo male clinically diagnosed with AD showed the typical in vivo AV-1451 distribution pattern of AD described in Johnson et al., with high retention in the inferior/lateral temporal, parietal, occipital and precuneus/posterior cingulate regions (E). Abbreviations: AD = Alzheimer dementia; MAPT = microtubule-associated protein tau; PET = positron emission tomography; PSP = Progressive Supranuclear Palsy; yo = years old.

Figure 2

Representative images from phosphor screen (A, C, E) and nuclear emulsion autoradiography (B, D, F) of brain slices from the three study subjects. The three cases showed a similar autoradiographic pattern with non-detectable AV-1451 binding across multiple brain ROIs known to contain high burden of tau pathology, including the basal ganglia, a region that exhibited elevated in vivo retention. Only two regions in these brains exhibited a strong autoradiographic signal, the EC (asterisk) (reflecting age-related tangle pathology that served as internal positive control for our autoradiographic experiments), and midbrain sections containing the substantia nigra (asterisk) (reflecting tracer off-target binding) (A, C, E). In both regions [F-18]-AV-1451 binding was almost completely blocked after incubating the slides with 1 μM unlabeled AV-1451. Microscopic images from slices dipped in the nuclear emulsion and immunostained with mouse PHF-1 antibody (a kind gift of Dr. Peter Davies) showed negligible silver grain deposition colocalizing with PSP tau inclusions and tau grains in the MAPT P301L mutation carrier (B, D, F). As expected, neuromelanin-containing neurons of the substantia nigra were strongly labeled by silver grains (D, F). Scale bars = 20 μm (B right top and bottom images, D bottom images and F), 50 μm (B left top image and D top images), 1 cm (A, C, E). Abbreviations: EC = entorhinal cortex; HPC = hippocampus; MAPT = microtubule-associated protein tau; MRI = magnetic resonance imaging; NFT = neurofibrillary tangles; PET = positron emission tomography; PSP = progressive supranuclear palsy.

Figure 3

Images from coronal brain sections of MRI (left) and their corresponding [F-18]-AV-1451 PET scans (middle) and autopsy tissue blocks (right) (A, C, E), and correlation analysis between in vivo uptake and in vitro binding (B, D, F) in the three study cases. Numbers displayed on PET images and in the graphs correspond to matching ROIs. No significant correlation was detected between in vivo uptake and in vitro binding of AV-1451 in any of the three cases (16-22 ROIs were analyzed in each case) (subject #1: R²=0.025, p=0.485; subject #2: R²=0.06, p=0.299; subject #3: R²=0.0002, p=0.963). Numbers correspond to the following anatomical regions: for subject #1: 1= middle frontal gyrus, 2 = superior precentral gyrus, 3 = inferior precentral gyrus, 4 = postcentral gyrus, 5 = superior temporal sulcus, 6 = superior frontal gyrus, 7 = superior frontal sulcus, 8 and 9 = frontal white matter, 10 = insula, 11 = planum polare, 12 = medium temporal gyrus, 13 = paracentral lobule, 14 = caudate, 15 = superior putamen, 16 = inferior putamen, 17 = inferior temporal, 18 = cingulate gyrus, 19 = thalamus, 20 = amygadala, 21 = head of the hippocampus, 22 = perirhinal cortex; for subject #2: 1 = caudate, 2 = putamen, 3 = external globus pallidus, 4 = internal globus pallidus, 6 = anterior nucleus of the thalamus, 7 = lateral geniculate nucleus of the thalamus, 8 = cingulate gyrus, 9 = superior frontal gyrus, 10 = insula, 11 = superior temporal gyrus, 12 = cerebellar dentate nucleus, 13 = cerebellar cortex, 14 = cerebellar peduncle, 15 = substantia nigra, 16 = red nucleus, 17 = periaqueductal grey, 18 = colliculus, 19 = basis pontis, 20 =pontine tegmentum; for subject 3: 1 = cingulate gyrus, inferior, 2 and 3 = frontal white matter, 4 = precentral grus, 5 = cingulate gyrus, superior, 6 = superior frontal sulcus, 7 = medial frontal gyrus, 8 = superior frontal gyrus, medial, 9 = superior frontal gyrus, lateral, 10 = putamen, 11 = insula, 12 = superior temporal gyrus, 13 = entorhinal cortex, 14 = amygdala, 15 = inferior temporal gyrus, 16 = medial temporal gyrus. Abbreviations: MAPT = microtubule associated protein tau; MRI = magnetic resonance imaging, PET = positron emission tomography; PSP = progressive supranuclear palsy; SUVR = standardized uptake value ratio.

Figure 4

(A) Microphotographs from brain tissue slices stained with PHF-1 antibody (kind gift of Dr. Peter Davis, left) and corresponding images displaying the optical threshold used to calculate tau burden using ImageJ software (NIH, right) (representative images from subject #1, PSP: precentral cortex; subject #2, PSP: caudate; subject #3, MAPT P301L mutation carrier: putamen). (B) Correlation analysis between tau burden and in vivo [F-18]-AV-1451 PET uptake in matching ROIs. In subject #1, the highest tau burden was detected in basal ganglia (3.83%), frontal medial (2.92%) and precentral (2.46%) areas. In subject #2, the highest tau burden corresponded to basal ganglia (1.85%), dentate nucleus of the cerebellum (1.63%) and EC (1.77%). In subject #3, the highest load of tau grains was detected in superior frontal (4.22%), medial frontal (1.85%) and medial temporal areas (1.84%). No significant correlation was detected between postmortem burden of tau aggregates and in vivo binding of [F-18]-AV-1451 in matching brain regions (subject #1: R²=0.01, p=0.70; subject #2: R²=0.09, p=0.48; subject #3: R²=0.11, p=0.31) (B). Numbers on the graphs correspond to the multiple ROIs analyzed: Subject #1: 1 = middle frontal, 2 = precentral, 3 = cingulate, 4 = superior temporal, 5 = insula, 6 = putamen, 7 = inferior temporal, 8 = HPC/EC, 9 = cerebellar cortex, 10 = superior colliculus, 11 = red nucleus, 12 = dentate nucleus; Subject #2: 1 = frontal, 2 = cingulate, 3 = basal ganglia, 4 = HPC/EC, 5 = midbrain, 6 = pons, 7 = dentate, 8 = cerebellum; Subject #3: 1 = HPC/EC, 2 = thalamus, 3 = insula, 4 = superior temporal, 5 = medial temporal, 6 = cingulate, 7 = prefrontal, 8 = superior frontal, 9 = medial frontal, 10 = inferior frontal, 11 = midbrain, 12 = cerebellum. Scale bar = 50 μm. Abbreviations: EC = entorhinal cortex, HPC = hippocampus, MAPT = microtubule associated protein tau; PET = positron emission tomography; PSP = progressive supranuclear palsy, ROI = region of interest.

Figure 5

Representative images of SDD-AGE membranes stained with total tau and PHF-1 antibodies (A, C, E) and correlation analysis between in vivo SUVR values and LMW and HMW tau measurements in matching ROIs (B, D, F). No significant correlation was detected between in vivo PET retention and postmortem measurements of total tau and PHF-tau assemblies, as reported by SDD-AGE, in subjects #1 and #3 (B and F). In subject #2, in vivo AV-1451 retention was significantly correlated with total tau HMW (R²=0.369, p=0.04), PHF-tau HMW (R²=0.383, p=0.032) and PHF-tau LMW (R²=0588, p=0.004). Abbreviations: HMW = high molecular weight; LMW = low molecular weight; PET = positron emission tomography; PHF = paired helical filaments; SDD-AGE = semi-denaturing detergent agarose gel electrophoresis; SUVR = standardized uptake value ratio.