Towards Improved Clinical Diagnosis of Argyrophilic Grain Disease Using Brain Imaging

Sakurai et al. [1] performed magnetic resonance imaging (MRI) morphometry analyses in autopsy-confirmed patients with argyrophilic grain disease (AGD) and depicted gray matter volume or cortical thickness loss in limited regions of the bilateral limbic lobes. AGD is a common cause of dementia in the elderly [2] and is classified as a 4-repeat tauopathy, which has a distinct pattern of tau filament fold [3]. Nevertheless, the clinical diagnosis, including the brain imaging of patients with AGD, remains inconclusive. In contrast, it is well established that the distribution of argyrophilic grains (AGs) exhibits a characteristic pattern, which begins in the ambient gyrus (stage 1), and spreads to the amygdala and medial temporal cortices (stage 2), and finally to the insular and anterior cingulate cortices (stage 3) [4]. Since the distribution pattern of grain pathology in AGD is different from that of neurofibrillary tangles in Alzheimer’s disease (AD), according to Braak’s staging, analyses of the distribution of atrophy using structural MRI could provide clues for differential diagnosis. In the report by Sakurai et al., the anterior medial temporal cortex, including the ambient gyrus, was clearly identified as the atrophic region in patients with AGD. The ambient gyrus is a small area located in an anterior and dorsomedial portion relative to the entorhinal cortex that has received less attention in MRI assessments than the entorhinal cortex affected in patients with AD [5]. Whether the detection of atrophy of the limbic lobe, including the ambient gyrus, is useful in practice for diagnosis of individual cases should be considered in a larger prospective cohort study.

AGD and senile dementia of the neurofibrillary tangle type (SD-NFT) are observed in the advanced age individuals and are often misdiagnosed as AD [6, 7]. Based on amyloid positron emission tomography (PET) studies, AGD, and SD-NFT are thought to be major causes of dementia in suspected non-Alzheimer’s disease pathophysiology (SNAP) [8]. In addition to MRI, cerebral perfusion single photon emission computed tomography (SPECT) or [¹⁸F] fluorodeoxyglucose PET may be useful in the differential diagnosis of AGD and AD by highlighting right-left asymmetry of regional cerebral blood flow (rCBF) or glucose metabolic rate in the cerebral cortex of patients with AGD [9]. We previously reported that patients with SNAP, including autopsy-proven AGD [16], showed a milder reduction in rCBF in the posterior cingulate cortex, precuneus, and parietal lobe, relative to the reduction observed in patients with AD [10]. Thus, combining MRI and SPECT could improve the diagnostic accuracy of AGD. It is also important to evaluate asymmetry and the distribution of abnormal findings.

The susceptibility of patients with AGD to comorbidities and other pathologies is a major issue that needs to be recognized. AGD is often comorbid with AD [11] and transactive response DNA-binding protein 43 kDa (TDP-43) pathology [12], which may accelerate atrophy of the medial temporal lobe [13], including the ambient gyrus and the adjacent brain structures. In the Sakurai et al. study, TDP-43 pathology was positive in three out of eight cases and was found in the amygdala and hippocampus, consistent with stage 2, according to the limbic predominant age-related TDP-43 proteinopathy (LATE) neuropathologic changes (NC) [14]. In this report, the impact of TDP-43 pathology on AGD could not be examined due to the small number of cases. In a previous study, we reported the brain imaging and clinicopathology of four elderly patients with prominent medial temporal lobe atrophy with AGD: one patient had stage 3 LATE-NC with stage 1 AGD [15] and A3B3C2 AD pathology; one had stage 3 AGD with stage 2 LATE-NC and A0B2C0 AD pathology [16]; and two patients with late-onset semantic dementia had TDP-43 type A (stage 3 LATE-NC) with stage 3 AGD pathology and A3B3C2 or A3B3C3 AD pathology [17]. These findings suggest that TDP-43 pathology could accentuate the medial temporal lobe atrophy associated with AGD. However, since a PET tracer for TDP-43 is lacking, and measurement accuracy of TDP-43 in blood or cerebrospinal fluid remains insufficient, the confirmation of comorbid TDP-43 pathology in vivo is difficult.

Although tau imaging revealed tau pathology in AD, conflicting results have been reported for tau imaging of non-AD 4-repeat tauopathy, including progressive supranuclear palsy (PSP) and corticobasal degeneration (CBD) [18]. There are few reports of successful visualization of AGs using tau PET [19]. Recently, high-contrast in vivo tau imaging using [¹⁸F]-PM-PBB3 in biopsy or autopsy-confirmed PSP and CBD patients has been reported [20]. This probe clearly stained the coiled body, argyrophilic threads, tufted astrocytes, and astrocytic plaques in PSP/CBD brains in vitro, but not AGs, which may be harder to detect. We examined an autopsy-confirmed patient with AGD who showed high accumulation of [¹⁸F] THK-5351 in the ambient gyrus, where AGs and astrogliosis were evident [16]. We also reported THK-5351 PET findings in early-stage semantic dementia and showed accumulations not only in the anterior medial temporal cortex but also in the anterior lateral temporal cortex [21]. Contrarily, a patient with AGD and comorbid TDP-43 pathology showed restricted THK-5351 accumulation in the anterior medial temporal cortex [16]. Neuropathologically, TDP-43 deposition in LATE is milder than SD and not necessarily restricted to the ambient gyrus [14]. Therefore, THK-5351 accumulation restricted to ambient gyrus may be a characteristic finding in pure AGD. [¹⁸F] THK-5351 was developed as a tau tracer but was found to reflect monoamine oxidase B (MAO-B) activity associated with astrogliosis. For AGD diagnosis, it may be clinically useful to evaluate astrogliosis PET together with tau PET.

When considering disease-modifying therapies for AD, clinicians should pay more attention to AGD. Since some patients with AGD pathology do not develop dementia but have psychiatric disorders [22], additional studies are required to determine whether these changes may represent the prodromal stage of AGD or merely age-related changes.