The inferior olive (IO) supports motor learning by supplying the cerebellum with critical sensory and motor input. In adult rats, that input includes externally generated limb stimulation. In contrast, the IO of postnatal day 8 (P8) rats does not exhibit responses to external stimuli. Instead, IO activity primarily reflects corollary discharges associated with the production of self-generated limb twitches during active (REM) sleep. Because corollary discharges are necessary for the computation of internal models, we tested the hypothesis that IO-related corollary discharge is necessary for the expression of cerebellar-dependent feed-forward activity during development. First, by recording from the IO of P12 and P20 rats of both sexes, we confirmed the presence of twitch-related corollary discharge at both ages; however, whereas the IO at P20 responded to limb stimulation, the IO at P12 did not. Next, using a protocol for selectively lesioning the climbing fibers that connect the IO to the cerebellum, including the interpositus nucleus (IP), we confirmed that lesioning at P12 prevents the IP's expression of corollary discharge at P13. Finally, we assessed the necessity of IO input to the cerebellum for the typical development of an internal model by lesioning climbing fibers at P12 or P19 and testing for the model's expression in the thalamus at P20. Only when the lesions occurred at P12 was the expression of the internal model severely disrupted. These findings provide the most direct evidence to date linking twitch-related corollary discharge to the developmental emergence of a cerebellar-dependent internal model.Significance statement In adult animals, internal models of movement generated by the cerebellum are essential for smooth, precise, and adaptive behavior. It is known that these internal models must develop in early infancy in an experience-dependent fashion; however, it is not known how sensorimotor experience acting on the nascent cerebellar circuit gives rise to adult functions. Here, in pre-weanling rats, we show that self-generated signals during sleep provide the early experiences that shape cerebellar internal models. Our findings lay a foundation for understanding how neural activity during sleep shapes typical and atypical brain development.
Expectations or prior beliefs about the world modulate sensory processing at the behavioral and neural levels. Bayesian models predict that such priors compensate for input uncertainty to optimize sensory judgments. Although Bayesian behavior is prevalent across sensorimotor systems, the relationship between priors and Bayesian inference is not obligatory. Priors may simply shift one's internal decision boundaries without interacting with sensory uncertainty at all. We recently showed that humans and monkeys use both Bayesian and non-Bayesian strategies when reporting judgments of visual stability across saccades, despite using priors in both cases. While they increased prior use to compensate for internal, movement-driven sensory uncertainty in a Bayesian manner, they decreased prior use when faced with external, visual image uncertainty. The latter, "anti-Bayesian" pattern was best explained by a model in which category boundaries were adjusted by the prior but susceptible to image noise. Here, in two male rhesus macaques we recorded neural activity in the frontal eye field (FEF), a prefrontal region important for visuosaccadic behavior. We toggled between subjects' prior use for Bayesian and anti-Bayesian behavior via trial-by-trial manipulation of the two uncertainty conditions. First, we found that FEF activity signaled the priors in both conditions. The prior-related modulation of activity, however, predicted only the anti-Bayesian, categorization behavior. The results suggest that neural activity in the FEF reflects the use of a flexible decision boundary for the perception of visual stability and, more generally, that neural mechanisms for Bayesian inference and visual categorization are dissociable and distributed in the primate brain.Significance Statement Appreciating a visual scene depends not only on retinal input, but also on priors about the world. Foreknowledge interacts with visual inputs to improve reactions and decisions. One way the brain combines priors and inputs is by using Bayes' rule to model optimal outcomes. A simpler way is by categorizing inputs with prior-adjusted boundaries. Here, we tested how neurons in primate frontal cortex use priors: for Bayes' rule, or for flexible categorization? A key feature of the study was to use a single perceptual task that was varied trial-by-trial to yield either Bayesian or categorization behaviors. We could then establish which behavior the neurons encoded. The implications extend beyond visuomotor behavior to broader neurocomputational mechanisms of prior use for cognition.
Callosal projections connect both cortical hemispheres via the corpus callosum, allowing bilateral integration of sensory information. Callosal axons originate mainly from layer (L)2/3 pyramidal neurons in primary sensory areas and project to homotopic contralateral regions. The projections display a stereotyped layer-specific pattern, targeting distinct dendritic domains of contralateral L2/3 neurons. In mouse somatosensory cortex, such precise innervation emerges in an activity-dependent manner during the second and third postnatal weeks but the molecular determinants are largely unknown. Using in utero electroporation of fluorescent reporters to label axonal and dendritic arbors of L2/3 neurons through postnatal development, we show that loss-of-function of Grin3a (gene encoding the non-conventional NMDA receptor GluN3A subunit) disrupts region and layer-specific contralateral targeting by callosal axons without affecting early axonal navigation or midline crossing. Rather than concentrating at the border between primary/secondary somatosensory cortex (S1/S2), callosal axons in male and female GluN3A knockout mice form a second column laterally in S2. Within the S1/S2 border, axonal arbors fail to innervate their normal destinations in L1 and outer 2/3 and shift towards inner L2/3 regions. Analysis of dendritic architecture revealed that GluN3A deletion drives proximal bifurcation and premature branching of apical dendrites of L2/3 neurons, with inward expansion of recipient dendritic trees matching callosal axon profiles in Grin3a knockouts. Together with conditional loss of function experiments, our results suggest that the dendritic patterning of postsynaptic L2/3 neurons directs the position of callosal axons within their target fields and implicate GluN3A in the postnatal timing and specificity of this process.Significance statement The formation of functional neural circuits requires a precise spatiotemporal overlap of incoming axons with specific dendritic domains of the target postsynaptic neurons, along with plasticity mechanisms that will stabilize some of the connections formed. Tight regulation of these processes during critical postnatal windows ensures input selectivity across brain regions, neuronal subtypes, and within distinct subcellular compartments, ultimately giving rise to highly stereotyped neural circuits that support accurate information transfer. Our study uncovers a previously unrecognized role for non-conventional N-methyl-D-aspartate (NMDA) receptor GluN3A subunits in coordinating the segregation of callosal axon projections onto defined dendritic domains of layer 2/3 neurons and specific areas of the mouse somatosensory cortex.
During social activities, people coordinate their movements by exchanging visuomotor information. Interpersonal coordination arises from two complementary processes, involving voluntary (planned) or spontaneous (emergent) synchronization, whose neurofunctional underpinnings are unknown. We investigated the brain correlates of these two synchronization processes during an fMRI finger-tapping task. Thirty participants (17 females, 13 males) were scanned while reproducing a target tempo either synchronizing with (Joint Action, JA) or resisting (Non-interactive, NI) the partner's tapping, whose hand was always visible to the scanned participant. Faster, slower, or equal tempi were used to induce within-dyad tempo discrepancies. Results revealed the emergence of tempo contagion: participants tapped faster in response to their partner's faster taps and vice versa for slower taps. The magnitude of this interpersonal contagion effect was similar across conditions but associated with the activity of different neural structures. Tempo contagion correlated positively with lateral occipitotemporal cortex (LOTc) activations in the JA condition but negatively with cerebellar activations in the NI condition. This suggests visuomotor information is exploited in opposite ways depending on the task instructions: the LOTc activity contributes to co-regulation to achieve synchronization in JA, whereas the cerebellar activity contributes to preserve individual motor stability in NI, preventing tempo contagion. A negative functional connectivity between the cerebellum and LOTc supports this latter result. These findings have implications for understanding the interplay between planned and emergent synchronization during motor interactions.Significance Statement During social activities, people can synchronize with each other intentionally or automatically. Here, the neural correlates of visuomotor social synchronization are investigated for the first time. We scanned brain activity while dyads voluntarily synchronized with the partner or resisted social synchronization. Surprisingly, both tasks activated the same brain areas, but visuomotor information was exploited in opposite ways by different neural structures. Visual brain regions contributed to coordination during social synchronization, while the cerebellum reduced temporal entrainment to the partner, thereby facilitating a focus on one's own tempo. We document that the brain flexibly uses the same visuomotor information differently depending on synchronization goals, balancing between matching others' timing and preserving one's own during social interactions.
Protein tyrosine phosphatase δ (PTPδ) is involved in Sema3A-induced dendritic elaboration of cortical pyramidal neurons through the activation of Fyn tyrosine kinase. However, the in vivo substrates of PTPδ remain largely unknown. Phospho-tyrosine proteome analysis of Ptpδ -/- mice brains from both male and female revealed that Signal Regulatory Protein α (SIRPα) was hyperphosphorylated at the carboxy-terminal Tyr501 residue in the knockouts. Immunohistochemistry with anti-phospho-Tyr501 SIRPα antibody showed the hyperphosphorylation of SIRPα in various regions including the olfactory nerve layer, cortex, striatum, thalamic reticular, and hypothalamic nuclei in Ptpδ -/- brain sections. These regions partially correlated with PTPδ expression. In the primary cultured wild-type mouse dorsal root ganglion neurons, Sirpα knockdown or overexpression of a cytoplasmic deleted SIRPα mutant partially blocked the Sema3A-induced growth cone collapse response. However, overexpression of non-phosphorylated mutants of SIRPα did not alter the response. This suggests that SIRPα is involved in the Sema3A-induced collapse response in an independent manner of phosphorylation and/or dephosphorylation. In the primary cultured cortical neurons, Sirpα knockdown or the overexpression of SIRPα-Tyr501Phe mutant attenuated Sema3A-induced dendritic growth. In cultured Sirpα -/- cortical neurons, re-expression of wild-type SIRPα, but not of SIRPα-Tyr501Phe, restored Sema3A-induced dendritic formation. In vivo analyses revealed ectopic expression of SIRPα-Tyr501Phe in cortical layer II/III pyramidal neurons with misoriented apical dendrites and attenuated basal dendrite arborization. Similar irregular cortical dendrites were observed in Sirpα -/- and Ptpδ -/- brains. Collectively, our results demonstrate that PTPδ regulates the dendritic elaboration of cortical pyramidal neurons through the dephosphorylation of SIRPα.Significance Statement The signaling mechanism of tyrosine phosphorylation and dephosphorylation in the dendrite morphogenesis is largely unknown. Sema3A facilitates the dendritic growth of cortical pyramidal neurons through the activation of protein tyrosine phosphatase δ (PTPδ) and Fyn tyrosine kinase. We here show that Signal Regulatory Protein α (SIRPα) is an endogenous substrate for Fyn and PTPδ. The Tyr501 residue of SIRPα is phosphorylated by Fyn and subsequently dephosphorylated by PTPδ. This phosphorylation/dephosphorylation turnover is essential for Sema3A-induced cortical dendritic growth. Sirpα -/- and Ptpδ -/- mice showed irregular dendritic phenotypes similar to those observed in Sema3a -/- and Fyn -/- mice. This is the first evidence showing that SIRPα and its phosphorylation/dephosphorylation participate in Sema3A-regulated dendritic elaboration of cortical pyramidal neurons.
The 5-methylcytosine (m⁵C) RNA modification regulates multiple aspects of RNA metabolism; however, its contribution to pathological pain remains poorly understood. Here, we investigated the role of the m⁵C reader Alyref in a mouse (of either sex) model of complete Freund's adjuvant (CFA)-induced chronic inflammatory pain. We observed a marked and sustained upregulation of Alyref in nociceptive neurons of the dorsal root ganglion (DRG) following CFA administration. Conditional deletion of Alyref in NaV1.8⁺ nociceptive neurons significantly exacerbated thermal and cold hypersensitivity, mechanical allodynia, and hyperalgesia. These behavioral abnormalities were accompanied by pronounced activation of microglia and astrocytes in the spinal dorsal horn, along with elevated expression of proinflammatory mediators, indicating enhanced neuroinflammation and central sensitization. Mechanistically, Alyref directly binds to m⁵C-modified Ccn3 mRNA and promotes its nuclear export, thereby maintaining cytoplasmic Ccn3 expression in DRG neurons. Loss of Alyref impaired Ccn3 mRNA export, reduced Ccn3 expression, and was associated with increased level of MMP-9, a key mediator of neuroinflammatory pain signaling. Collectively, these findings identify Alyref as a protective regulator in chronic inflammatory pain by restraining neuroinflammation and limiting the emergence of neuropathic pain-like features. Our study highlights a previously unrecognized role of m⁵C-dependent RNA regulation in nociceptive sensitization and suggests Alyref as a potential therapeutic target for pathological pain.Significance Statement Understanding the molecular and cellular mechanisms underlying initiation and maintenance of pathological pain is critically important in identifying novel therapeutic targets. Our observation of a marked and sustained upregulation of Alyref, the reader protein for 5-methylcytosine (m⁵C) RNA modification, in nociceptive neurons of the dorsal root ganglion (DRG) of mice with chronic inflammatory pain, suggests an important role of RNA modifications in regulating pain. Conditional knockout of Alyref in NaV1.8⁺ nociceptive neurons results in exacerbated thermal and cold hypersensitivity, mechanical allodynia, and hyperalgesia. We identify Ccn3 mRNA as a key target for Alyref. Our findings demonstrate that proper maintenance of cytoplasmic Ccn3 expression by Alyref is essential to suppress MMP-9-mediated neuroinflammatory signaling, thereby limiting the development of pathological pain.
The threshold for sensory detection varies with fluctuations in intrinsic cortical activity. When stimulus-encoding cortical populations are more excitable, stimuli elicit stronger neural responses that are more likely to be detected. However, the detection of a stimulus is also more likely when cortical populations are less excitable because there is less background "noise". Therefore, it is unclear how the variable states of cortical activity interact to impact sensory detection. We hypothesize the answer depends on the spatiotemporal structure of intrinsic activity states across stimulus encoding and non-encoding populations. To test this, we examined intrinsic and target-evoked population activity across cortical Area MT in common marmosets (Callithrix jacchus; one male and one female) while they performed a threshold visual detection task. We compared detection performance based on target-evoked responses and the state of intrinsic activity in the larger surrounding population. We find that the intrinsic activity in the surrounding, non-encoding population predicted trial-by-trial detection performance better than the population encoding the target-evoked response. Furthermore, we find that the detection performance of the monkey was best predicted by the divergence in activity between the encoding and surrounding non-encoding population. These findings suggest that, rather than a source of noise or irrelevant to sensory processing, the distributed spatiotemporal state of intrinsic activity directly influences how sensory signals are represented in cortical populations and can influence perceptual thresholds in visual detection.Significance Statement Prior research into how variability in neural activity impacts perception has often focused on neural populations that encode relevant sensory information. However, the role of variable intrinsic activity in nearby, non-encoding populations and their contribution to sensory representations is less well understood. We found that the state of intrinsic activity in non-encoding populations was a better predictor of performance on a threshold visual detection task than the evoked-response magnitude. These results suggest that the state activity in broader neural populations plays a larger role in sensory computations relevant to perceptual decisions than previously regarded.
Prolonged postnatal maturation of the primate amygdala is thought to be driven, at least partially, by continued neural maturation within the paralaminar nucleus (PL). At birth, the PL is densely populated with post-mitotic glutamatergic neurons that gradually mature throughout postnatal life. This active process is likely supported by microglia, which promotes synaptic maturation. Our previous work showed that maternal separation alters microglia development across the infant to adolescent transition (n=19 females, 4 males). Here, in the same cohort, we examined whether morphologic microglial changes are associated with alterations in the numbers of pre-synaptic terminals (SYN1+ puncta), post-synaptic terminals (PSD95+ puncta), and putative excitatory contacts (SYN1-PSD95 colocalization), and whether these synaptic elements are engulfed by phagocytic microglia. In maternally reared macaques, SYN1+ puncta, PSD95+ puncta, and putative synaptic contacts decreased, while microglial (IBA1+) volume, CD68+ content, and engulfment of synaptic elements increased between infancy and adolescence. These findings suggest greater pruning of all synaptic elements by adolescence. Maternal separation altered this trajectory, resulting in increased phagocytic activity and engulfment of synaptic elements primarily during infancy. Maternal separation also resulted in a 50% reduction in mature PL neurons by adolescence, suggesting maturational failure, cell loss, or both by adolescence. These findings demonstrate that early life stress disrupts normative synaptic pruning and microglia-synapse interactions in the developing primate PL. Increased synaptic engulfment in infants with disrupted care is associated with premature, aberrant pruning and highlights a potential cellular mechanism through which early environmental insults could change PL neural development by adolescence.Significance Statement The paralaminar nucleus (PL) of the amygdala is an important substrate for the delayed post-natal development of the amygdala in human and nonhuman primates. Gradually maturing glutamatergic neurons in this region, and the microglia that support them, are exposed to life events that may shape their development. We recently found that maternal separation in infants produces aberrant hyper-ramified microglia in the PL beginning in infancy and persisting into adolescence. Examining neuron-microglial interactions in the same cohort, we now find increased phagocytic engulfment of synaptic elements by microglia after maternal separation in infancy only, with a reduction in PL mature neurons that is apparent by adolescence. Together, these data suggest a mechanism for altered PL maturation, instigated by disrupted maternal care.
Trauma to the spinal cord initiates an inflammatory response that causes secondary damage, which collectively can result in loss of function below the level of the injury. The unbalanced risk-benefit-ratio of methylprednisolone led to development of therapeutic nanoparticles (NP) that associate with circulating monocytes and neutrophils to reduce inflammation and secondary damage, and improve functional recovery in a female mouse model of cervical hemisection spinal cord injury. Herein, we investigate the mechanisms occurring during the acute phase of injury by which NPs directly and indirectly modulate the phenotype and trafficking of monocytes and neutrophils, and computationally catalog the communication network among cell types within the injury microenvironment. Using adoptive transfer to monitor trafficking, NP treatment reduced the extent of myeloid cell recruitment to the injury, yet did not impact the composition of adoptively transferred monocytes or neutrophils. The proportion of inflammatory monocytes was reduced with NP treatment, and single cell sequencing analysis indicated increased polarization towards pro-regenerative phenotypes. Sequencing analysis also demonstrated that outgoing signals from monocytes and neutrophils influenced the phenotype of numerous cell types, including endothelial cells, fibroblasts, oligodendrocyte progenitor cells, and Schwann cells. Signaling between cell compartments involves a combination of soluble and matrix signals, with NP treatment enhancing expression of genes associated with anti-inflammatory phenotypes, angiogenesis, neuroprotection, and promotion of axon outgrowth or decreasing expression of inhibitors to regeneration. Collectively, NP delivery leads to direct and indirect effects on monocytes and neutrophils, which subsequently influence gene expression and intercellular signaling networks that promote a pro-regenerative environment.Significance Statement Spinal cord injury is a serious trauma that results in significant loss of motor function and lacks successful clinical treatment options. We have previously demonstrated that polymeric nanoparticles can attenuate inflammation to reduce the severity of secondary injury, resulting in improved functional outcomes. Here, we computationally explore the mechanisms through which these nanoparticles alter the inflammatory response and improve the regenerative trajectory of the injury.
In this study, we investigated whether the architecture of brain interactions at rest maintains a representation of individual behavioral skills. Specifically, we aimed to identify a minimal set of topological features that capture an electrophysiological mechanism underlying the encoding of a motor skill. We tested whether brain network topology at rest could model individual performance in a motor task, such as manual dexterity, in 86 subjects of either sex from the Human Connectome Project. Using a machine learning procedure, we identified an optimal fingerprint that accurately modeled individual manual dexterity, encompassing four Participation Index-based connector hubs in the alpha frequency band, involving the parietal cortex. A vulnerability analysis, in which we simulated disconnections of the involved hubs, revealed that two of them were critical, resulting in a significant drop in predictive performance. We combined these features to propose a functional "refocusing" mechanism: hubs progressively prune connections external to their modules when dexterity increases, while maintaining an internal representation of dexterity performance. Such inhibition and maintenance are well aligned with the role of the alpha band reported in the literature. These findings suggest that the architecture of interactions at rest, by combining few topological features in the alpha band, encodes stable behavioral traits, such as motor skills.Significance statement Behavior can be recovered from the architecture of brain communication at rest, using high-dimensional models, typically based on whole-brain activity or dense connectomes. However, obtaining a low-dimensional and interpretable encoding mechanism remains challenging. Thus, here we aimed at identifying a compact, topological fingerprint that encodes individual motor performance. We revealed a novel electrophysiological mechanism of functional refocusing, where connector hubs inhibit connections external to their functional modules to maintain an internal representation of the motor skill. To our knowledge, this is the first time motor dexterity has been predicted solely from resting-state connectivity, i.e., without task-related modulation. The topography of these hubs and the involved frequency band are consistent with current literature on inhibition and maintenance of functional interactions.
Despite being essential mediators of pain processing, the molecular identity of N-methyl-D-aspartate receptor (NMDAR) subtypes in nociceptive dorsal horn circuits is poorly understood, especially between sexes and in humans. Given the importance of GluN2 subunits in shaping NMDAR function and plasticity, we investigated the expression and localization of specific GluN2 NMDAR variants in the dorsal horn of viable spinal cord tissue from male and female rodents and human organ donors. Analysis of single-cell/nuclei sequencing datasets and quantitative reverse transcriptase polymerase chain reactions (qRT-PCR) revealed that the GluN2A (GRIN2A) and GluN2B (GRIN2B) subunits are robustly expressed in dorsal horn neurons of mice, rats and humans, with moderate expression of GluN2D (GRIN2D). Immunohistochemistry (IHC) with antigen retrieval demonstrated that GluN2A, GluN2B, and GluN2D proteins are all preferentially localized to the superficial dorsal horn of both adult rats and humans, which is conserved between males and females. Surprisingly, we found that these GluN2 NMDAR subunits are enriched in the lateral superficial dorsal horn in rats but not in humans, while presynaptic and neuronal markers are symmetrically distributed across the rat mediolateral axis. A dramatic shift in localization of GluN2A to the lateral superficial dorsal horn was observed across later postnatal development (PD21-PD90) in both male and female rats, with a corresponding change in synaptic NMDAR currents. This discovery of changes in NMDAR subunit distribution during maturation and between species will shed light on the physiological roles of NMDARs and their potential as therapeutic targets for pain.Significance statement We used complementary single-cell/nuclei analysis, immunostaining, quantitative reverse transcriptase polymerase chain reactions, RNAscope in situ hybridization, and electrophysiological approaches to compare the relative expression of N-methyl-D-aspartate receptor (NMDAR) GluN2 subunits in dorsal horn spinal cord pain circuits of mouse, rat, and human spinal cord tissue. Through these comparisons, we find that the transcripts and proteins of the GluN2A, GluN2B, and GluN2D NMDAR subunits are robustly expressed in superficial dorsal horn neurons, with conserved expression across sex but important differences in expression and localization patterns across late development and between species. These discoveries shed light on the physiological roles of NMDARs and their utility as potential therapeutic targets for pain.
Blocking estrogen synthesis via aromatase inhibition helps prevent the recurrence of estrogen-receptor positive breast cancer but also results in cognitive, sleep and thermoregulatory disturbances. These side-effects diminish quality of life and contribute to treatment nonadherence in a large proportion of patients. DHED is a brain-selective prodrug that, in rodent models, converts to 17β-estradiol (E2) selectively in the brain without affecting the periphery. We investigated whether DHED could prevent side effects associated with the aromatase inhibitor letrozole in a primate model of aging. Chronic oral treatment with DHED in letrozole-treated male and female marmosets led to a robust increase in E2 levels across brain regions without affecting estrogen levels at the periphery. In addition, DHED treatment (1) improved memory at short delays and prevented letrozole-induced cognitive slowing in a hippocampal-dependent memory task; (2) normalized hippocampal neuronal membrane potential and excitability and (3) reduced sleep fragmentation. However, DHED treatment had opposite effects on thermoregulation in males and females, necessitating additional research in this area. Overall, the results suggest that DHED, which lacks estrogenic effects in peripheral tissues, could be a safe and effective novel hormonal therapy for improving quality of life in breast cancer patients treated with aromatase inhibitors.Significance statement Women with estrogen receptor positive (ER+) breast cancers take aromatase inhibitors to lower estrogens and reduce cancer recurrence. As a result, they often experience symptoms of estrogen deficiency that compromise treatment adherence. Here, we show that a brain-selective estrogen therapy administered orally via the prodrug DHED substantially increases estrogen levels in the marmoset brain without affecting the periphery and normalizes impairments in working memory, hippocampal neuronal excitability and sleep induced by aromatase inhibition. These findings in a translational primate model represent a significant advance for women's health by positioning DHED as a non-invasive, safe and efficient novel hormone therapy to improve quality of life of women with ER+ breast cancers.
The reversal of learning-induced synaptic potentiation through depotentiation may be important in certain types of forgetting. Here, we evaluated how synaptic plasticity induced by different stimuli in the hippocampus is affected by mechanistically distinct forms of depotentiation. In hippocampal slices obtained from male and female mice, we artificially induced long-term potentiation (LTP) using either a temporally spaced or compressed stimulation pattern. Using a combination of electrophysiology and protein quantification approaches, we found divergent molecular pathways recruited during depotentiation of spaced and compressed LTP. Depotentiation of both forms of LTP required glutamatergic activation of the NMDA receptor (NMDAR). However, depotentiation of compressed, but not spaced, LTP shared a requirement with long-term depression for intracellular non-ionotropic NMDAR signaling cascades mediated by the C-terminal domain of GluN1. Downstream of NMDAR signaling, AMPA receptor phosphorylation was also differentially modified during depotentiation of spaced and compressed LTP. Finally, we found that depotentiation of spaced but not compressed LTP required synaptic Arc. Altogether, we identify a role for non-ionotropic NMDAR signaling in synaptic depotentiation. Additionally, we reveal two mechanistically distinct forms of NMDAR-dependent depotentiation that can be selectively induced after different temporal patterns of LTP induction. Our findings have important implications for the regulation of physiological and pathological forgetting.  Significance statement Synaptic depotentiation, the reversal of learning-associated synaptic potentiation, is an important mechanism of forgetting. This study uncovers divergent signaling pathways mediating depotentiation in the hippocampus. We reveal that the type of long-term potentiation (LTP) induced can bias signaling pathways downstream of NMDA receptor activation during depotentiation. Specifically, we show divergent requirements for non-ionotropic GluN1 C-terminal domain interactions in depotentiation of LTP induced using temporally compressed versus spaced patterned activity. Further, depotentiation of spaced and compressed LTP are associated with unique biochemical signatures, including AMPA receptor phosphorylation states, consistent with divergent molecular mechanisms that act to dampen synaptic responses. Our results illuminate fundamental mechanisms that govern plasticity associated with forgetting, with implications for memory preservation in disease states.
Deep brain stimulation (DBS) of the ventral capsule/ventral striatum (VCVS) can treat obsessive-compulsive disorder (OCD) and other psychiatric conditions. Yet, optimizing its clinical efficacy is a major challenge, often hindered by incomplete knowledge of how stimulation parameters and targets affect neural activity and behavior. VCVS DBS is thought to work in part by improving cognitive control, an important decision-making component that is impaired in OCD and other illnesses. The magnitude of this cognitive control enhancement was shown to be lateralized, with right-unilateral stimulation being the most effective. Prior work developed a preclinical model of VCVS DBS by leveraging the cognitive control construct, which can be measured in humans and rodents and is modulated by analogous brain circuits. However, this work did not address laterality effects observed in humans or examine left/right stimulation differences. These effects may be critical for maximizing therapeutic benefit while avoiding aversive outcomes. This study aimed to investigate lateralization in the rodent model, where bilateral stimulation of the mid-striatum was previously shown to improve cognitive control. Right and left-unilateral stimulation reduced response times without changing accuracy, replicating the cognitive control improvement from bilateral stimulation. With computational modeling, we show that bilateral and unilateral stimulation modifies the same decision-making variables to drive this behavior change. We also establish that females have the same cognitive control improvement from stimulation as males. These findings increase our understanding of cognitive control circuits and strengthen the validity of the rodent model as a translational platform to study VCVS DBS's therapeutic mechanisms.Significance Statement Here, we demonstrate that stimulating just one side of the brain (e.g. unilaterally) can be as effective as bilateral stimulation for improving cognitive control, the ability to adjust thoughts and decisions in response to environmental changes. These findings in rodents match results from prior human deep brain stimulation (DBS) studies, highlighting the validity of this preclinical model to study DBS's therapeutic mechanisms. We hypothesize that unilateral stimulation may be preferable to maximize cognitive benefits without causing off-target effects, while also reducing surgical invasiveness. Further, we demonstrate that females have the same cognitive control improvement from stimulation as males. Overall, this work answers important outstanding clinical questions regarding laterality and sex in DBS therapies for psychiatric illnesses.
In addition to supervised motor learning, the cerebellum also supports nonmotor forms of learning, including reinforcement learning (RL). Recent studies in animal models have identified core RL signals related to reward processing, reward prediction, and prediction errors in specific regions in cerebellar cortex. However, the constraints on these signals remain poorly understood, particularly in humans. Here, we investigated cerebellar RL signals in a computationally-driven fMRI study. Human participants performed an RL task without low-level sensorimotor contingencies (N = 32, N female = 24). We observed robust RL signals related to reward processing and reward prediction errors in cognitive regions of the cerebellum. These signals were not explained by oculomotor or physiological confounds. By manipulating the delay between choices and reward outcomes, we discovered that cerebellar RL signals are temporally sensitive: robust when feedback was delivered shortly following choices, but undetectable at supra-second feedback delays. Similar delay effects were not found in other areas implicated in reward processing, including the ventral striatum and hippocampus. Further, reward prediction error activity in the cerebellum was related to behavioral performance when feedback was delivered promptly, but not when it was delayed. Connectivity analyses revealed that during RL feedback, cognitive areas of the cerebellum coactivate with a network that includes the medial and lateral prefrontal cortex and caudate nucleus. Together, these results highlight a temporally constrained contribution of the human cerebellum to a cognitive learning task.Significance statement Reinforcement learning (RL) - the shaping of behavior through reward feedback - is an essential cognitive capacity. Previous work has focused almost exclusively on cerebral circuits that support RL, however recent work in animal models also implicates the cerebellum. Despite growing interest in the "cognitive cerebellum," its contributions to human RL remain unclear. We show that regions spanning lobules Crus I/II respond to rewards and encode reward prediction errors in a temporally sensitive manner. Our results build directly on research in model organisms, highlighting functional parallels across species and the importance of crosstalk between human and animal researchers. Including the cerebellum as a node in RL networks creates a potential new target for intervention in cases of reward processing dysfunction, like addiction.
The pathogenesis of perioperative neurocognitive disorders (PND) involves a complex interplay of genetic vulnerability and environmental insults, with epigenetic regulation acting as a dynamic mediator. However, the cell-specific epitranscriptomic responses to perioperative stressors like sevoflurane anesthesia, and their functional consequences for cognitive decline, are not well defined. Here, we report that the m6A demethylase FTO is significantly upregulated in the medial prefrontal cortex (mPFC) of male mice exposed to sevoflurane anesthesia. Astrocytic FTO, but not neuronal or endothelial FTO, is highly sensitive to sevoflurane exposure. Conditional knockout of FTO in astrocytes attenuated sevoflurane-induced cognitive deficits, while astrocyte-specific FTO overexpression exacerbated sevoflurane-induced cognitive deficits. Mechanistically, astrocytic FTO mediated m6A demethylation of glutamate transporter-1 (GLT-1) mRNA, leading to enhanced GLT-1 protein expression and aberrant glutamatergic transmission. Sevoflurane exposure disrupted synaptic transmission, neuronal morphology, and calcium activity in the mPFC, which were rescued by astrocytic FTO deletion. Supplementation with the methyl donor S-adenosylmethionine (SAMe) normalized m6A levels and improved cognitive performance. This study demonstrates that astrocytic FTO is a critical epitranscriptomic modulator of sevoflurane-induced PND and a potential therapeutic target for PND.Significance Statement PND are a major clinical concern for which effective mechanism-based interventions are lacking. This study identifies astrocytic FTO as a cell-type-selective epitranscriptomic driver of sevoflurane-induced PND and establishes that its m6A-demethylase activity disrupts glutamate homeostasis by post-transcriptionally regulating the astrocytic glutamate transporter GLT-1. Astrocyte-restricted deletion of FTO preserves synaptic transmission, neuronal structure, and calcium dynamics, thereby preventing cognitive decline, while astrocytic FTO overexpression exacerbates deficits. Therapeutic restoration of m6A methylation with the methyl donor SAMe normalizes the epitranscriptomic landscape and rescues cognitive function. These findings reveal astrocytic m6A regulation as a previously unrecognized pathogenic mechanism and a druggable target for PND.
SLC6A8 encodes the creatine transporter (CRT), which mediates creatine transport across the plasma membrane in the brain, including the blood-brain barrier and neurons. Creatine transporter deficiency (CTD), caused by pathogenic variants in SLC6A8, leads to cerebral creatine depletion and cognitive impairment. Here, we investigated the developmental molecular mechanisms underlying CTD using the pathogenic c.1681G>C (G561R) variant of Slc6a8, which corresponds to a variant identified in SLC6A8 in a patient with CTD. In vitro analyses using HEK293 cells expressing mutant mouse CRT carrying the G561R variant demonstrated impaired N-glycan maturation and plasma membrane localization of the transporter, resulting in markedly reduced creatine uptake, consistent with previous reports on the corresponding human CRT variant. To investigate the in vivo effects of this pathogenic variant, we generated CRT-G561R knock-in mice by introducing the c.1681G>C point mutation into the mouse Slc6a8 gene using the CRISPR/Cas9 system. These male mice exhibited severe reductions in brain creatine levels, postnatal growth retardation, and impaired spatial memory, despite preserved gross brain morphology. Quantitative proteomic analyses of the hippocampus and cerebral cortex during postnatal development revealed region-dependent protein alterations in CTD. The hippocampus showed pronounced early postnatal remodeling involving proteins related to actin cytoskeleton organization and vesicle-mediated membrane trafficking, whereas the cerebral cortex exhibited a more gradual response involving creatine biosynthesis-related enzymes and later-emerging mitochondrial pathways, including the mitochondrial translation machinery. These findings demonstrate stage- and region-dependent proteomic remodeling during postnatal brain development in CTD.Significance Statement Creatine transporter deficiency (CTD) causes cerebral creatine depletion and intellectual disability; however, the developmental mechanisms linking creatine loss to brain dysfunction remain unclear. We performed developmental proteomic profiling of the hippocampus and cerebral cortex using a mouse model carrying a pathogenic Slc6a8 variant identified in patients with CTD. Creatine transporter dysfunction induces distinct region- and stage-dependent molecular responses during postnatal brain maturation. The hippocampus shows early alterations in cytoskeleton-dependent membrane trafficking pathways, consistent with impaired synaptic and circuit maturation, whereas the cerebral cortex exhibits progressive metabolic and mitochondrial adaptations. These findings suggest that impaired creatine-dependent energy buffering disrupts distinct developmental programs across brain regions, potentially contributing to cognitive dysfunction by hindering early hippocampal circuit maturation.
Threatening situations require animals to rapidly select appropriate defensive strategies, either disengaging behavior to avoid harm or engaging actions that allow escape or avoidance. The lateral habenula (LHb) is a key hub in aversive processing, and its projection to the rostromedial tegmental nucleus (RMTg) suppresses dopaminergic activity and promotes behavioral disengagement. However, although LHb neurons also project directly to the ventral tegmental area (VTA) and encode aversive signals, how this pathway contributes to learning and behavior remains poorly understood. In this study, we tested the hypothesis that VTA-projecting LHb neurons encode aversive signals that facilitate associative learning and promote escape behavior. Using a retrograde viral strategy, we targeted VTA-projecting LHb neurons and monitored calcium activity during active avoidance training in male and female mice. These neurons were activated by aversive stimuli and predictive cues as mice acquired avoidance responses and showed increased activity at movement onset during the tail suspension test (TST). Silencing LHb→VTA transmission impaired avoidance learning, prolonged escape latency, and reduced persistence and vigor of active responses in the TST, without affecting baseline locomotion. Anatomical and ex vivo electrophysiology revealed that LHb terminals innervate both dopaminergic (TH⁺) and non-dopaminergic (TH⁻) VTA neurons, exhibiting session-specific synaptic adaptations during avoidance learning. Together, these findings identify the LHb→VTA pathway as a source of aversive predicting signals required for the acquisition of avoidance behavior and the persistence of active responses in aversive contexts, supporting the idea that distinct LHb outputs may differentially regulate behavioral disengagement and active defensive responses.Significance Statement Threats demand fast decisions: freeze or act. We show that a direct pathway from the lateral habenula (LHb) to the ventral tegmental area (VTA) is essential for learning to avoid danger and for engaging active defensive responses. LHb→VTA neurons respond to aversive events, cues that predict them, and movement onset, linking prediction to action. Constitutive silencing of this pathway impairs avoidance learning and reduces adaptive escape behavior in threatening contexts. Anatomically, LHb inputs contact both dopamine and non-dopamine VTA neurons, and their synaptic responses change with learning, indicating plasticity in this circuit. These results identify a key pathway coupling threat prediction to action, informing models of stress-sensitive behavior.
Mosaic mutations in the X-linked cell adhesion molecule Protocadherin 19 (PCDH19) lead to epilepsy with cognitive impairment, whereas complete absence of functional protein, although possibly linked to autistic features, does not elicit any seizures. It is believed that mosaic expression of PCDH19 leads to defective neuronal communication, but whether further roles beyond cell adhesion are critical for PCDH19 function in the cortex is currently unknown. We confirm that the proteolytic processing of PCDH19, previously described in hippocampal neurons, also takes place in mouse cortical neurons in vivo and show that nuclear transport of its intracellular domain is mediated by importins. RNAseq analysis further indicates that the intracellular domain of PCDH19 leads to broad transcriptomic changes. Finally, we use in utero electroporation to provide the first in vivo data about the role of this cleaved intracellular domain in upper layer cortical neurons of male and female mice, where it reduces spine density through an increase in Xlr gene expression without affecting overall dendritic morphology. Our results suggest that PCDH19 could act as an activity sensor in a synapse to nucleus signalling pathway involved in synaptic homeostasis.Significance statement We investigate non-adhesive functions of the epilepsy-linked cell adhesion protein PCDH19 and uncover a signalling role for its intracellular domain in cortical neurons. Beyond its established function in cell adhesion, we show that proteolytic cleavage of PCDH19 and nuclear import of its intracellular fragment leads to transcriptomic changes that impact dendritic spine density through the upregulation of Xlr genes. These findings suggest that PCDH19 may act as a synaptic activity sensor, linking membrane dynamics to nuclear responses in the regulation of synaptic homeostasis.
HIV-1 infection often results in sensory neuropathy, with more than 60% of affected individuals developing chronic pain. Although viral proteins such as glycoprotein 120 (gp120) contribute to neuronal injury and pain hypersensitivity, their specific effects on nociceptive signaling remain unclear. Hyperactivity of N-methyl-D-aspartate receptor (NMDAR) in the spinal dorsal horn is a hallmark of neuropathic pain. Here, we determined how gp120 affects synaptic NMDAR activity in spinal excitatory and inhibitory neurons in male and female mice. Intrathecal gp120 enhanced expression of α2δ-1 and GluN1 in the dorsal root ganglion and spinal cord. Gp120 also increased α2δ-1-GluN1 interaction and their synaptic trafficking in the spinal cord. Functionally, gp120 induced hyperactivity of presynaptic NMDARs on primary afferent terminals and postsynaptic NMDARs in vesicular glutamate transporter 2 (VGluT2)-expressing excitatory, but not vesicular GABA/glycine transporter (VGAT)-expressing inhibitory, dorsal horn neurons. Importantly, gp120-induced hyperactivity of both presynaptic and postsynaptic NMDARs was eliminated by the α2δ-1 inhibitory ligand gabapentin or by an α2δ-1 C-terminal peptide that disrupts α2δ-1-NMDAR interactions. Correspondingly, treatment with the NMDAR antagonist, gabapentin, or α2δ-1 C-terminal peptide consistently reversed gp120-induced persistent nociceptive hypersensitivity. Furthermore, genetic deletion of Cacna2d1 or selective ablation of GluN1 in dorsal root ganglion neurons significantly attenuated gp120-induced nociceptive hypersensitivity. Together, these findings indicate that gp120 drives nociceptive hypersensitivity by augmenting presynaptic and postsynaptic activity of α2δ-1-bound NMDARs, thereby amplifying nociceptive transmission from primary afferents to spinal excitatory neurons. Targeting α2δ-1-associated NMDARs may therefore represent a promising therapeutic approach for HIV-associated chronic neuropathic pain.Significance Statement HIV-associated chronic pain affects a large proportion of people living with HIV and remains challenging to treat. This study reveals how the HIV viral protein gp120 augments transmission from peripheral nerves to spinal cord neurons to produce persistent pain. We show that gp120 strengthens the interaction between the α2δ-1 protein and NMDA receptors in the spinal cord. Remarkably, gp120 selectively potentiates synaptic NMDA receptor activity in spinal excitatory, but not inhibitory, neurons, leading to heightened nociceptive sensitivity. Importantly, blocking α2δ-1 or disrupting its coupling with NMDA receptors reverses these effects and alleviates pain-like behaviors in animal models. These findings uncover a molecular mechanism underlying gp120-induced central sensitization and identify a potential therapeutic target for chronic pain in people with HIV.