Despite epidemiological and genetic data linking semantic dementia to inflammation, the topography of neuroinflammation in semantic dementia, also known as the semantic variant of primary progressive aphasia, remains unclear. The pathology starts at the tip of the left temporal lobe where, in addition to cortical atrophy, a strong signal appears with the tau PET tracer 18F-flortaucipir, even though the disease is not typically associated with tau but with TDP-43 protein aggregates. Here, we characterized the topography of inflammation in semantic variant primary progressive aphasia using high-resolution PET and the tracer 11C-PBR28 as a marker of microglial activation. We also tested the hypothesis that inflammation, by providing non-specific binding targets, could explain the 18F-flortaucipir signal in semantic variant primary progressive aphasia. Eight amyloid-PET-negative patients with semantic variant primary progressive aphasia underwent 11C-PBR28 and 18F-flortaucipir PET. Healthy controls underwent 11C-PBR28 PET (n = 12) or 18F-flortaucipir PET (n = 12). Inflammation in PET with 11C-PBR28 was analysed using Logan graphical analysis with a metabolite-corrected arterial input function. 18F-flortaucipir standardized uptake value ratios were calculated using the cerebellum as the reference region. Since monoamine oxidase B receptors are expressed by astrocytes in affected tissue, selegiline was administered to one patient with semantic variant primary progressive aphasia before repeating 18F-flortaucipir scanning to test whether monoamine oxidase B inhibition blocked flortaucipir binding, which it did not. While 11C-PBR28 uptake was mostly cortical, 18F-flortaucipir uptake was greatest in the white matter. The uptake of both tracers was increased in the left temporal lobe and in the right temporal pole, as well as in regions adjoining the left temporal pole such as insula and orbitofrontal cortex. However, peak uptake of 18F-flortaucipir localized to the left temporal pole, the epicentre of pathology, while the peak of inflammation 11C-PBR28 uptake localized to a more posterior, mid-temporal region and left insula and orbitofrontal cortex, in the periphery of the damage core. Neuroinflammation, greatest in the areas of progression of the pathologic process in semantic variant primary progressive aphasia, should be further studied as a possible therapeutic target to slow disease progression.

No abstract

BackgroundNeuroinflammation, along with beta amyloid and tau brain deposition, is a potential diagnostic biomarker and therapeutic target in patients with Alzheimer’s disease (AD). The role of inflammation at various stages of the AD process needs to be clarified. TSPO PET is useful to image brain inflammation, but some tracers have low affinity for TSPO and “second generation” tracers, with higher affinity, cannot image subjects with a low‐binder TSPO rs6971 genotype. We overcame this problem by using 11C‐ER176, a high‐affinity tracer allowing for imaging of all participants and with a more favorable metabolite profile than other high affinity tracers (Fujita, 2017). Then, we correlated the localization of neuroinflammation tau and amyloid in a subset of the participants.MethodWe imaged 15 patients with amnestic MCI (mean age 63±5, 7 women) and 15 healthy controls (HC) (8/15 women, mean age 65 ±7) using 11C‐ER176 PET. TSPO affinity was similar (MCI/HC = 2/3 low, 8/7 mixed, 5/5 high). A full factorial analysis was performed on V T values between groups at the regional level (Hammer’s atlas). Ten MCI patients had also amyloid PET (18F‐florbetaben or 18F‐florbetapir) and tau PET (18F‐flortaucipir). Standard uptake value ratios (SUVRs) were calculated for both tracers using the cerebellum as reference. Regional correlations among the three tracers were determined for each patient and an average calculated. All images were corrected for partial volume effect.ResultNeuroinflammation in MCI was bilaterally increased in precuneus and lateral temporo‐ parietal cortex, and right amygdala. There was a correlation between the localization of amyloid and tau in the brain (r=0.61). However, the correlation between neuroinflammation and tau was even higher (r=0.73); neuroinflammation also co‐localized with amyloid, but not as strongly (r=0.59).Conclusion 11C‐ER176 PET allowed for the identification in MCI of neuroinflammation in regions known to be involved in the AD process. Importantly, all subjects with any TSPO genotype could be studied. As expected, the localization of amyloid and tau in the brain was correlated, but the co‐localization of neuroinflammation with tau was even higher than between amyloid and tau. This finding highlights the importance of neuroinflammation as a biomarker of neurodegeneration and as a therapeutic target.

Background: Neuroinflammation, a hallmark of frontotemporal dementia, has been sparsely studied in vivo. However, neuroinflammation is a potentially important therapeutic target and biomarkers of inflammation would be welcome. To determine the anatomical specificity of a putative biomarker for inflammation, translocator protein 18 kDa (TSPO) imaging, we used 11 C-PBR28 PET to measure TSPO levels in patients with the non-fluent/agrammatic variant of primary progressive aphasia (nfvPPA) and/or apraxia of speech (AOS).Methods: Sixteen patients with nfvPPP/AOS underwent PET using 11 C-PBR28 and 18 Fflortaucipir. All patients were PET amyloid-negative except two, with a non-AD flortaucipir PET. Thirty healthy controls underwent PET imaging using 11 C-PBR28 (13 controls) or 18 F-flortaucipir (17 controls). Patients (9/16 women, mean age 65±5.9 years) did not differ significantly in age from the controls (17/30 women, mean age 66±9.3 years). V T values for 11 C-PBR28 were calculated with the Logan plot and a metabolitecorrected arterial input function. For 18 F-flortaucipir, the SUV ratio over the cerebellar gray matter was calculated for t = 80-100 min. All images were corrected for partial volume effect. A full factorial analysis was performed on V T and SUVR values between patients and controls at the voxel-wise level.Results: Compared to controls, patients showed increased V T and SUVR values (p < 0.001 uncorrected) in left inferior, middle and superior frontal gyri, as well as in the left putamen and pallidum. However, peak uptake of 18 F-flortaucipir was localized in Broca's area, epicenter of the pathology, while the distribution of 11 C-PBR28 V T was more extensive in the frontal lobe, including precentral gyrus, and other regions outside of the frontal lobe, as the supramarginal gyrus and superior temporal sulcus, as well as a small area in the posterior extent of the right inferior frontal gyrus.Conclusions: 11 C-PBR28 and 18 F-flortaucipir uptake patterns corresponded to anatomical areas known to be related to the specific language functions that are impaired in nfvPPA and apraxia of speech. Thus, 11 C-PBR28 seems to be a suitable radioligand for detecting neuroinflammation in nfvPPA. Furthermore, the more exten-