RIKEN Center for Brain Science

Post-translational modifications regulate tau aggregation and propagation, yet how amyloid pathology shapes the tau modification landscape remains unclear. Using liquid chromatography-tandem mass spectrometry, we compared tau modifications in MAPT knock-in (MAPT KI) mice and MAPT/App double knock-in mice with AppNL-G-F amyloid pathology. Only subtle differences were detected, with a tendency toward increased acetylation within the repeat domain. Because K321 and K331 lie in the fibril-forming core, their roles were further examined. Acetylation at these sites was absent in cynomolgus monkey brains. To test functional relevance, we generated acetylation-mimicking MAPTK331Q knock-in mice on the MAPT KI background. Despite tau expression, these mice showed reduced tau phosphorylation at 24 months, with unchanged insoluble tau and seeding activity. Thus, K331 acetylation does not promote tau pathology.

Puncta adherentia junctions (PAJs) have been observed at glutamatergic excitatory synapses on glutamatergic excitatory neurons (E→E synapses) in the mouse hippocampus and are considered to be the mechanical adhesion sites between axon terminals and dendritic shafts. However, it remains unclear whether they are present at other types of synapses, such as glutamatergic synapses on GABAergic inhibitory neurons (E→I synapses) and GABAergic synapses on excitatory neurons (I→E synapses) and inhibitory neurons (I→I synapses). We showed here that the PAJ components, such as nectin-1, nectin-3, l-afadin, N-cadherin, β-catenin, and αN-catenin, were observed at all four types of synapses in cultured mouse hippocampal neurons. In the adult mouse hippocampus, these PAJ components were observed at a subset of I→I synapses on parvalbumin-positive GABAergic inhibitory neurons, similarly to their presence at E→E synapses in the CA1 and CA3 regions. In contrast, several PAJ components, including nectin-3, l-afadin, N-cadherin, β-catenin, and αN-catenin, were observed at E→I and I→E synapses, whereas nectin-1 was absent. These results indicate that the PAJs are formed at a subset of I→I synapses in addition to E→E synapses, where all examined PAJ components are present, whereas E→I and I→E synapses exhibit only partial localization of these components, reflecting molecular diversity among different synapses in the adult mouse hippocampus.

Identifying appropriate structures for generative or world models is essential for both biological organisms and machines. This work shows that synaptic pruning facilitates efficient statistical structure learning. We extend previously established canonical neural networks to derive a synaptic pruning scheme that is formally equivalent to an online Bayesian model selection. The proposed scheme, termed Bayesian synaptic model pruning (BSyMP), utilizes connectivity parameters to switch between the presence (ON) and absence (OFF) of synaptic connections. Mathematical analyses reveal that these parameters converge to zero for uninformative connections, thus providing reliable and efficient model reduction. This enables the identification of a plausible structure for the environmental model, particularly when the environment is characterized by sparse likelihood and transition matrices. Through causal inference and rule learning simulations, we demonstrate that BSyMP achieves model reduction more efficiently than the conventional Bayesian model reduction scheme. These findings indicate that synaptic pruning could be a neuronal substrate underlying structure learning and generalizability in the brain.

The Self-Determination Theory posits that extrinsic reward withdrawal often undermines intrinsic motivation. This undermining effect varies according to the reward category (e.g., monetary vs. praise). However, the mechanism underlying this variability remains unclear, partly because previous studies have conceptualized intrinsic motivation as a single construct without distinguishing liking (intrinsic value) from wanting (motivational drive). Failure to make this distinction may obscure the effects of rewards on each component. We hypothesized that extrinsic reward withdrawal would differentially affect intrinsic value and motivational drive. To test this hypothesis, we conducted functional magnetic resonance imaging (fMRI) on 54 participants who were randomly assigned to one of three conditions: control (no rewards), monetary reward, or social reward. Participants in the monetary and social reward groups played a stopwatch game twice, first with and then without extrinsic rewards. The control group received no extrinsic rewards during the task. We assessed intrinsic value through fMRI-measured activation of the reward system and motivational drive through voluntary engagement during the task intermission. Intrinsic value, measured by activation in predefined reward-related regions, declined in the control group but was preserved in both extrinsic reward groups after withdrawal. Thus, both monetary and social rewards were associated with preserved intrinsic value. However, intrinsic behavioral motivation declined in the monetary group but was maintained in the social group. These findings suggest that extrinsic reward withdrawal differentially influences intrinsic value and motivational drive across reward categories, thereby clarifying the variability of the undermining effect.

Deep brain stimulation (DBS) of the subthalamic nucleus (STN) is an effective treatment for Parkinson's disease (PD). The outcome of DBS varies from patient to patient. We investigated the preoperative changes in the midbrain to predict DBS outcomes. Patients with PD who underwent bilateral STN-DBS at Juntendo University Hospital between June 2019 and October 2020 were included. We compared the Movement Disorders Society Unified Parkinson's Disease Rating Scale Part III score in the off state 3 months before and 1 year after surgery, and the difference in the score was defined as the motor outcome after DBS. Preoperative magnetic resonance images of the midbrain were evaluated using FreeSurfer and the One-Line Method. A total of 34 patients were enrolled. Multiple regression analysis with least squares indicated that the volume of the midbrain measured using FreeSurfer had no significance as an independent variable (p = 0.16), whereas the midbrain length evaluated using the One-Line Method had significance (p < 0.05) when the outcome of motor symptoms was set as the dependent variable. We showed that midbrain size evaluated using the One-Line Method was a good predictor of the motor outcome after STN-DBS, but volume measured using FreeSurfer was not. The One-Line Method was considered to detect changes in size specifically in the midbrain tegmentum, which is associated with PD motor symptoms. Thus, we present a simple and useful method for predicting poor outcomes after STN-DBS treatment in patients with PD with disease progression.

Abstract Understanding how molecular diversity across long-range neural circuits governs brain function remains a central challenge in neuroscience. Here, we developed a projection-specific spatial proteomics approach and revealed robust presynaptic molecular divergence across six projection-defined pathways of the medial prefrontal cortex (mPFC), including efferent projections to the basolateral amygdala (BLA), nucleus accumbens (NAc), thalamus (Thal), hypothalamus (HT) and cortex (CTX), as well as the afferent projection from the BLA to the mPFC. Among these pathways, we identify BLTP2 (KIAA0100), a previously uncharacterized transmembrane protein, as highly enriched in the projection from the mPFC to the BLA. BLTP2 localizes to excitatory presynaptic terminals and is enriched at synapses that are activated during memory formation. Loss of BLTP2 impairs synaptic structure and transmission, and reduces activity-dependent remodelling, resulting in selective deficits in contextual fear memory, anxiety-related behavior, and social behavior. Mechanistically, BLTP2 promotes presynaptic assembly by recruiting Neurexin 1. These findings reveal projection-specific presynaptic molecular diversity and provide mechanistic insights into circuit-level vulnerabilities in neuropsychiatric disorders.

Bovine leukemia virus (BLV) causes enzootic bovine leukosis (EBL), yet its pathogenic mechanisms remain largely unclear. In particular, the role of BLV genomic integration sites (IS(s)) in BLV-induced leukemogenesis has not been fully elucidated. Here, we identified a total of 235 ISs from 99 BLV-infected cattle with lymphoma, of which 4.3% and 46.8% were located within exon and intron, respectively, while no preferential integration into CpG islands or repetitive regions was observed. All identified ISs were distinct, and no identical sites were detected among the samples. We identified 246 genes related with these ISs and protein–protein interaction analysis of these genes demonstrated one “IS-Clustered genes” composed of 85 among 246 genes. This “IS-Clustered genes” contains 12 cancer genes (14.1%) with high significantly proportion. Notably, with 55 among 99 cattle tested (55.6%) harboring ISs within this cluster, suggesting its crucial involvement in BLV-induced pathogenesis. Furthermore, integrated analysis of known lymphoma- and IS-related genes and the 85 “IS-Clustered genes” showed that key genes formed a shared cluster, indicating a potential “common EBL-associated cluster.” These findings provide important insights into the role of BLV integration in EBL development and may contribute to elucidating its molecular mechanisms underlying onset of EBL. In addition, these findings may also aid in the development of therapeutic strategies and facilitate early diagnosis.