Papanikolaou A, Graykowski D, Lee BI, Yang M, Ellingford R, Zünkler J, Bond SA, Rowland JM, Rajani RM, Harris SS, Sharp DJ, Busche MA. Selectively vulnerable deep cortical layer 5/6 fast-spiking interneurons in Alzheimer's disease models in vivo. Neuron. 2025 Jul 23;113(14):2265-2279.e7. Epub 2025 May 8 PubMed.

Recommends

Please login to recommend the paper.

Comments

  1. This elegant study adds a crucial new dimension to our understanding of early circuit-level pathology in Alzheimer’s disease. By combining in vivo two-photon Ca²⁺ imaging with high-density single-unit recordings using Neuropixels probes in head-fixed mice, the authors identify a selective and early dysfunction of fast-spiking parvalbumin (PV) interneurons in cortical layer 5 (L5) of the V1 primary visual cortex, occurring prior to amyloid plaque deposition. Mechanistically, they demonstrate that PV hypoactivity is causally linked to a deep-layer-specific reduction of NPTX2 in excitatory neurons and GluA4 in PV interneurons, leading to a substantial loss of excitatory inputs onto these cells.

    While previous work linked PV interneuron hypoactivity to network hyperexcitability and memory deficits (Palop and Mucke, 2016; Verret et al., 2012), this study stands out for its laminar precision and in vivo single-cell resolution, both rarely achieved in AD models. The authors provide a compelling example of how early cortical inhibition failure can emerge in a layer- and cell-type-specific manner, revealing subtle but potentially pivotal circuit vulnerabilities that precede classical histopathological changes.

    As with any significant advance, the findings raise several thought-provoking questions:

    1. Is local PV interneuron hypoactivity sufficient to generate epileptiform discharges in V1, or is broader circuit instability required? LFP analysis of the existing Neuropixels data could help clarify this point.
    2. What molecular or activity-dependent signals drive the downregulation of NPTX2 in excitatory neurons?
    3. Does this deep-layer PV interneuron vulnerability extend to associative and limbic cortices more directly implicated in memory and cognition?

    Regarding the last question, several studies in the hippocampus (Kam et al., 2016; Slutsky, 2024; Soula et al., 2023; Zarhin et al., 2022) have shown that early network hyperexcitability in AD models is often state-dependent, emerging more prominently during sleep rather than wakefulness in early disease stages, and preceding cognitive symptoms. In contrast, Papanikolaou et al. report PV hypoactivity during wake states in V1 L5 of young AD model mice. Whether this discrepancy reflects fundamental differences in circuit function between the hippocampus and primary sensory cortices remains an open question.

    The results also resonate with earlier studies of laminar-specific electrophysiology in V1 of healthy mice. For example, L5/6 excitatory neurons exhibit higher firing rates and  lower burstiness in comparison to L2/3, and selective downregulation during NREM sleep, a feature not observed in superficial layers (Senzai et al., 2019). 

     It would be informative to know whether control mice in the Papanikolaou study display similar laminar dynamics, and how these properties are altered in the AD model, particularly during sleep, which was not examined here.

    Future Neuropixels recordings across the sleep–wake cycle in freely behaving mice will be needed to determine whether early L5 PV interneuron dysfunction is a generalizable hallmark of AD pathology or varies with brain regions and vigilance states. Overall, this study provides a technically rigorous platform for probing the earliest stages of single-neuron and circuit-level dysfunctions in Alzheimer’s disease.

    References:

    Kam K, Duffy ÁM, Moretto J, LaFrancois JJ, Scharfman HE. Interictal spikes during sleep are an early defect in the Tg2576 mouse model of β-amyloid neuropathology. Sci Rep. 2016 Jan 28;6:20119. PubMed.

    Palop JJ, Mucke L. Network abnormalities and interneuron dysfunction in Alzheimer disease. Nat Rev Neurosci. 2016 Dec;17(12):777-792. Epub 2016 Nov 10 PubMed.

    Senzai Y, Fernandez-Ruiz A, Buzsáki G. Layer-Specific Physiological Features and Interlaminar Interactions in the Primary Visual Cortex of the Mouse. Neuron. 2019 Feb 6;101(3):500-513.e5. Epub 2019 Jan 8 PubMed.

    Slutsky I. Linking activity dyshomeostasis and sleep disturbances in Alzheimer disease. Nat Rev Neurosci. 2024 Apr;25(4):272-284. Epub 2024 Feb 19 PubMed.

    Soula M, Maslarova A, Harvey RE, Valero M, Brandner S, Hamer H, Fernández-Ruiz A, Buzsáki G. Interictal epileptiform discharges affect memory in an Alzheimer's disease mouse model. Proc Natl Acad Sci U S A. 2023 Aug 22;120(34):e2302676120. Epub 2023 Aug 17 PubMed.

    Verret L, Mann EO, Hang GB, Barth AM, Cobos I, Ho K, Devidze N, Masliah E, Kreitzer AC, Mody I, Mucke L, Palop JJ. Inhibitory interneuron deficit links altered network activity and cognitive dysfunction in Alzheimer model. Cell. 2012 Apr 27;149(3):708-21. PubMed.

    Zarhin D, Atsmon R, Ruggiero A, Baeloha H, Shoob S, Scharf O, Heim LR, Buchbinder N, Shinikamin O, Shapira I, Styr B, Braun G, Harel M, Sheinin A, Geva N, Sela Y, Saito T, Saido T, Geiger T, Nir Y, Ziv Y, Slutsky I. Disrupted neural correlates of anesthesia and sleep reveal early circuit dysfunctions in Alzheimer models. Cell Rep. 2022 Jan 18;38(3):110268. PubMed.

    View all comments by Inna Slutsky

  2. This is an exciting and timely study that adds an important layer to our understanding of early circuit dysfunction in Alzheimer’s disease. The finding that deep-layer (L5/6) parvalbumin-expressing interneurons exhibit suppressed spiking before overt plaque pathology, and that this is driven by reduced excitatory input resulting from decreased NPTX2 and GluA4 expression, highlights a spatially and cell-type-specific vulnerability that has not been fully appreciated.

    What’s especially compelling is the demonstration that restoring NPTX2 expression in excitatory neurons not only rescues excitatory drive but also normalizes spiking of L5/6 PV interneurons. Interestingly, we showed previously that NPTX2 protects synapses by inhibiting complement activation and limiting microglia-mediated synapse pruning (Zhou et al., 2023). The convergence of molecular, physiological, and histopathological data across studies supports a model in which NPTX2 functions both as a coordinator of excitatory input to PV interneurons and as a safeguard against complement-mediated synaptic loss in AD.

    The present study provides mechanistic insight linking NPTX2 loss to early inhibitory dysfunction and altered cortical information processing. This may help explain why reduced CSF NPTX2 levels are strongly associated with cognitive decline and predict the transition from asymptomatic to symptomatic AD (Soldan et al., 2023). 

    Looking ahead, it will be important to determine whether similar spatial- and cell-type-specific deficits occur in human AD, and whether therapeutic elevation of NPTX2 can help restore cortical inhibitory balance and preserve cognitive function.

    References:

    Zhou J, Wade SD, Graykowski D, Xiao MF, Zhao B, Giannini LA, Hanson JE, van Swieten JC, Sheng M, Worley PF, Dejanovic B. The neuronal pentraxin Nptx2 regulates complement activity and restrains microglia-mediated synapse loss in neurodegeneration. Sci Transl Med. 2023 Mar 29;15(689):eadf0141. PubMed.

    Soldan A, Oh S, Ryu T, Pettigrew C, Zhu Y, Moghekar A, Xiao MF, Pontone GM, Albert M, Na CH, Worley P. NPTX2 in Cerebrospinal Fluid Predicts the Progression From Normal Cognition to Mild Cognitive Impairment. Ann Neurol. 2023 Oct;94(4):620-631. Epub 2023 Jul 25 PubMed.

    View all comments by Borislav Dejanovic

  3. This work by Papanikolaou et al. presents a series of neurophysiological investigations of neuronal activity across superficial and deep layers of primary visual cortex (V1) in AD mouse models.

    The key findings from V1 deep layers (L5/6) of AD mouse models—prior to plaque deposition—include: increases in burst firing from excitatory neurons, which was associated with higher calcium load and metabolic cost; reductions of NPTX2 in excitatory neurons; and a ~70 percent decrease in excitatory synapse count on parvalbumin interneurons. The authors present evidence that NPTX2 overexpression mitigates layer-specific burst firing in APP/PS1 mice.

    The findings suggest cellular mechanisms for underlying, and increasingly documented, posterior cortical abnormalities in AD—especially apparent in posterior cortical atrophy (“visual variant AD”). Of note:

    • V1 is traditionally considered spared until later stages of AD. Even the limited investigations of regional pathology in posterior cortical atrophy tend to report more burden of AD pathology within extrastriate and visual association than primary visual cortices (Hof et al., 1990; Tang-Wai et al, 2004; Abdi et al., 2025). 
    • However, the earliest pathological studies of PCA (“AD-Balint’s”; Hof et al., 1990) noted vulnerability of Meynert cells in V1 layers 5-6, analogous to excitatory neurons in L5/6 in the current study (Figure 5). In primates, these cells project to occipito-parietal regions—characteristically vulnerable in early stage PCA—and the superior colliculus. These regions are extensively connected with the frontal eye fields which appear vulnerable in later stage PCA—of likely relevance to the authors’ predictions regarding wider network dysfunction.

    While exercising caution regarding extrapolating neurophysiological findings from APP/PS1 mouse models to vulnerability of the human visual system in sporadic AD, the above prompts consideration of:

    1) laminar-level, excitatory and interneuronal vulnerability in AD, perhaps owing to structural and functional characteristics, such as long-range projections and metabolic demands, and:

    2) whether particular cellular vulnerability and/or recent evidence of a different microglial phenotype in posterior cortical atrophy (Abdi et al., 2025) relate to Nptx2 underexpression and microglia-mediated synapse loss.

    References:

    Hof PR, Bouras C, Constantinidis J, Morrison JH. Selective disconnection of specific visual association pathways in cases of Alzheimer's disease presenting with Balint's syndrome. J Neuropathol Exp Neurol. 1990 Mar;49(2):168-84. PubMed.

    Tang-Wai DF, Graff-Radford NR, Boeve BF, Dickson DW, Parisi JE, Crook R, Caselli RJ, Knopman DS, Petersen RC. Clinical, genetic, and neuropathologic characteristics of posterior cortical atrophy. Neurology. 2004 Oct 12;63(7):1168-74. PubMed.

    Abdi Z, Yong KX, Schott JM, Gatt A, Revesz T, Crutch SJ, Lashley T. Pathological Characterisation of Posterior Cortical Atrophy in Comparison With Amnestic Alzheimer's Disease. Neuropathol Appl Neurobiol. 2025 Apr;51(2):e70007. PubMed.

    View all comments by Keir Yong

  4. This study represents a very important milestone, clarifying the intrinsic mechanisms related to sleep disturbances, the hyperexcitability of the hippocampus, and the excitotoxicity that leads to synapse and neuronal loss in the onset of Alzheimer's disease. Based on these mechanisms, the ability to selectively enhance the functional activity of parvalbumin (PV)-positive GABAergic inhibitory interneurons throughout the brain, including deep brain regions, has become a key to early AD treatment. Since the intrinsic oscillatory rhythm of PV-positive GABAergic inhibitory interneurons is the gamma rhythm (Hadler et al., 2024), deep brain-reachable, weak-intensity transcranial magnetic stimulation interventions targeting the function of PV-positive GABAergic interneurons may solve the problem (Zhen et al., 2017) and hold promise as an effective preventive and therapeutic approach for early AD.

    References:

    Hadler MD, Tzilivaki A, Schmitz D, Alle H, Geiger JR. Gamma oscillation plasticity is mediated via parvalbumin interneurons. Sci Adv. 2024 Feb 2;10(5):eadj7427. Epub 2024 Jan 31 PubMed.

    Zhen J, Qian Y, Weng X, Su W, Zhang J, Cai L, Dong L, An H, Su R, Wang J, Zheng Y, Wang X. Gamma rhythm low field magnetic stimulation alleviates neuropathologic changes and rescues memory and cognitive impairments in a mouse model of Alzheimer's disease. Alzheimers Dement (N Y). 2017 Nov;3(4):487-497. Epub 2017 Sep 11 PubMed.

    View all comments by Yunfeng Zheng

  5. This paper investigates early circuit dysfunction in the visual cortex in a mouse model of Alzheimer's disease, focusing on the role of parvalbumin-positive (PV+) interneurons, the neuronal pentraxin 2 gene (NPTX2), and GluA4-containing AMPA receptors. The authors use a combination of large scale functional tools in vivo (two-photon calcium imaging and neuropixel single-cell electrophysiological probes), which enhanced single-cell resolution and provided structured laminar profiling of the cortical areas in an unprecedented manner compared to previous studies.

    The authors report that in early stages of AD there is hypoactivity of PV+ interneurons in the deep, but not superficial layers of the visual cortex, accompanied by increased burst firing of putative pyramidal neurons. They identify a reduction in NPTX2 expression, which is crucial for clustering GluA4-containing AMPA receptors at excitatory synapses on PV+ interneurons. This reduction leads to impaired excitatory drive onto PV+ interneurons, disrupting the excitation-inhibition balance and contributing to network hyperexcitability and altered visual cortex processing (orientation/direction selectivity). Reinstating network functionality via adenoviral vector delivery of NPTX2 was able to restore the neurophysiological deficits in the visual microcircuit.

    This paper strengthens the view that AD is a disorder of network dysfunction with synaptic integrity and input loss (rather than cell death) being the core change of early AD. It stresses that excitation inhibition balance, input coordination, and ultimately long-range subcortical connectivity (deep layer cortical output) are the downstream targets of such synaptic changes and emphasizes that not all synapses are equally vulnerable, pointing toward a synapse- and cell-type-specific degeneration model.

    The reported dynamics of NPTX2 and GluA4 loss also testify toward a layer- and region-specificity (deep layers of V1 only, not L2/3) and reinforces the idea of microcircuit-specific vulnerability, not a general synaptic failure. Finally, this work stresses the predictive footprint that sensory cortices may have in early AD dysfunction and reframes the idea that the quantitative changes taking place in sensory processing areas can act as a readable output for early AD dysfunction and potential diagnosis biomarker.

    This paper offers a beautifully detailed microcircuit model of early dysfunction, but whether this scales to the human brain remains unclear.  The authors do a great job sketching out a molecular mechanism that leads to network failure, sensory dysfunction, and the proof-of-concept rescue of function. The manuscript raises a number of questions that are central to the quest for early AD biomarkers and calls for further validation of the links that it has provided, namely the generalizability of these ideas and the feasibility of a roadmap to early intervention in humans.

    In particular, the authors do not explain why deep-layer interneurons in visual cortex are affected so early in this model. Is this a generalizable feature across cortical areas or specific to visual processing? Furthermore, they do not evaluate the extent of cross-species relevance of their findings nor whether the findings are related to this specific type of mouse model used. Our work finds that similar synaptic dysfunctionalities and hyperexcitable phenotypes occur more pronouncedly in the superficial rather than the deep layers of the human cortex (Gazestani et al., 2023; Dougalis et al., 2025). This raises the possibility of layer and cortical region-specific differences between mouse and human systems. If such laminar-specific discrepancies persist across models, it would suggest that current mouse paradigms may capture only aspects of the human pathology, warranting model diversification or adaptation.

    The prodromal parieto-occipital atrophy subtype of human AD (Ten Kate et al., 2018) is associated with distinct biomarker changes. Unfortunately, Papanikolaou and colleagues do not provide biomarker analysis that could link the mouse model visual deficits (and the molecular mechanisms reported) to human pathology, so the translational relevance and the link to Aβ plaques or to later tau neurofibrils remain unknown. Additionally, the transient nature of the reported temporal dynamics of the NPTX2-GluA4 deficit (present only at around 3 months in the APP/PS1 mice but not at 1.5 or 6 months) implies that a critical window of vulnerability or plasticity failure exists. The proposed AAV-NPTX2 rescue, while conceptually compelling, faces translational barriers, not only due to the current limitations of viral delivery in humans (particularly for widespread cortical targeting), but also because of the narrow (and currently undefined in humans) time window of opportunity, which may be difficult to detect in otherwise asymptomatic clinical populations. 

    To facilitate interpretation of the findings by Papanikolaou and colleagues, and to better reveal the diagnostic and clinical value of their study, further cross-species and model validation of the proposed mechanisms are needed. This should include neuroimaging work on early dysfunction of sensory systems, combined with validation of the correlative and predictive value of biomarkers such as p-tau, Aβ, microglial markers, and NPTX2. Longitudinal blood and CSF sampling will also be critical in characterizing the temporal dynamics of the molecular changes described. Finally, evaluating whether CSF or plasma NPTX2 levels correlate with early functional, behavioral and structural abnormalities in sensory cortices (using tools such as EEG, MEG, or fMRI), particularly in the parieto-occipital regions implicated in prodromal AD, would strengthen the translational value of the study. There is a long road ahead, and this study has illuminated a promising path forward.

    References:

    Gazestani V, Kamath T, Nadaf NM, Dougalis A, Burris SJ, Rooney B, Junkkari A, Vanderburg C, Pelkonen A, Gomez-Budia M, Välimäki NN, Rauramaa T, Therrien M, Koivisto AM, Tegtmeyer M, Herukka SK, Abdulraouf A, Marsh SE, Hiltunen M, Nehme R, Malm T, Stevens B, Leinonen V, Macosko EZ. Early Alzheimer's disease pathology in human cortex involves transient cell states. Cell. 2023 Sep 28;186(20):4438-4453.e23. PubMed.

    Dougalis A, Abushik P, Pelkonen A, Giudice L, Gomez-Budia M, Novosolova N, Valimaki N-N, Rezaie M, Nurkhametova D, Giniatullina R, Shakirzyanova A, Mali A, Rauramaa T, Stevens B, Hiltunen M, Leinonen V, Malm T. Early amyloid spine response and impaired synaptic transmission of pyramidal neurons in human biopsies with Alzheimer's Disease-related pathology. 2025 Jan 07 10.1101/2025.01.07.630516 (version 1) bioRxiv.

    Ten Kate M, Dicks E, Visser PJ, van der Flier WM, Teunissen CE, Barkhof F, Scheltens P, Tijms BM, Alzheimer’s Disease Neuroimaging Initiative. Atrophy subtypes in prodromal Alzheimer's disease are associated with cognitive decline. Brain. 2018 Dec 1;141(12):3443-3456. PubMed.

    View all comments by Antonios Dougalis

Make a Comment

To make a comment you must login or register.

This paper appears in the following:

News

  1. Does Loss of Neuronal Pentraxin 2 Start Network Dysfunction in Alzheimer’s?