Abstract Identifying the molecular cascade underlying neuronal degeneration in Alzheimer’s disease has been hampered by cellular heterogeneity and the limitations of donor-level classification. By integrating 851,682 cortical layer 2 and 3 excitatory neuron transcriptomes from 557 individuals across four independent snRNA-seq datasets of the human prefrontal cortex, we reconstruct neuronal degeneration as a continuous, stage-resolved transcriptional trajectory. Ordering neurons by collective pathological burden and clinical manifestation reveals that degeneration unfolds asynchronously within individual brains. Neurons in a single brain can simultaneously occupy early, intermediate, and late pathological states, a continuum entirely obscured by conventional approaches. The trajectory captures discrete transcriptional inflection points defining successive stages of vulnerability linked to development of neuropathology. Systematic analysis of the complete human kinome and phosphatome along this trajectory identifies a temporal hierarchy of phosphorylation dysregulation which collectively creates a permissive environment for tau pathology to escalate, defining stage-specific molecular nodes in the degenerative cascade.

Abstract Recently, the amyloid-beta (Aβ) targeting antibody lecanemab has demonstrated modest therapeutic efficacy in slowing cognitive decline in people with Alzheimer’s disease (AD). Lecanemab clears amyloid plaques from the brain; however, plaque load does not correlate strongly with cognitive function. The strongest neuropathological correlate of cognitive decline in AD is synapse loss, which is exacerbated in the halo surrounding neuritic amyloid plaques where Aβ accumulates in remaining synapses. Here, we hypothesised that, through clearing plaques and the associated halo of soluble Aβ that can directly damage synapses, lecanemab could temper plaque-associated synapse loss. High-resolution imaging of temporal cortex tissue from people who died with AD (N=20) and age-matched controls (N=19) reveals lecanemab staining within individual pre and post-synaptic excitatory terminals in addition to plaque staining. The percentage of pre-synapses containing lecanemab-positive Aβ was over 200% higher in AD and the percentage of post-synapses was over 150% higher in AD than control tissue, with highest levels of synaptic lecanemab staining observed near plaques. These data demonstrate that lecanemab antibody recognises Aβ within synapses, warranting future work to determine whether lecanemab treatment slows cognitive decline, at least in part, through both clearing plaques and facilitating clearance or neutralisation of synaptic Aβ.

The pathological cascade of Alzheimer's disease (AD) begins decades before clinical symptoms, yet modelling this slow and progressive process in experimental systems has remained challenging. Current models largely depend on Tau mutations associated with primary Tauopathies and fail to reproduce AD-relevant neurofibrillary tangle formation, even when human Tau isoforms and amyloid pathology are combined. Here, we xenotransplanted human neurons into Rag2 −/− and Rag2 −/− App NL−G−F mice and exposed them to AD-derived Tau seeds via parenchymal brain injections, providing both the pathological nucleation event and the amyloid-rich environment needed to compress decades of human disease into the lifespan of a mouse. Whilst AD-Tau seeding alone was sufficient to induce Tau misfolding in human neurons, amyloid exposure markedly accelerated Tau aggregation, driving neurofibrillary tangle formation and elevated insoluble pTau181 and pTau231 in aged xenografts. This effect was abolished in human Tau-knockout neurons, although LAMP1 positive neuritic dystrophy persisted. Late-stage Tau pathology was absent in non-grafted App NL−G−F mice, indicating a unique human neuronal susceptibility to Tau aggregation. Together, these findings demonstrate that the human neuronal environment is a critical determinant of neurofibrillary tangle formation, and that amyloid causally accelerates this process, thereby establishing the Tau seeded xenograft model as a platform to dissect the interaction between amyloid and Tau in a human-relevant system.

Microglia surrounding amyloid plaques in Alzheimer’s disease (AD) strongly upregulate HLA class II genes, but whether this transcriptional state reflects antigen presentation, and how it relates to genetic risk, remains unclear. Here we identify a broad repertoire of endogenous HLA-bound peptides in human brain extracts from individuals with and without AD. Using human microglia xenografts isolated after exposure to plaques, we show that amyloid pathology induces an HLA-expressing microglial state that presents a similarly diverse spectrum of peptide antigens. This response was accompanied by increased expression of costimulatory and coinhibitory molecules, supporting a transition towards an antigen presenting microglial state. Loss of PICALM, a major AD risk gene involved in endolysosomal trafficking, markedly amplified the microglial HLA response to amyloid pathology, and was associated with increased intracellular Aβ accumulation and modulation of immune regulatory proteins. These findings identify amyloid pathology as a driver of antigen presentation in human microglia and establish endolysosomal trafficking as a genetic regulator of this disease-associated immune state. Unexpectedly, the microglial immunopeptidome was not dominated by Aβ or tau, suggesting that amyloid plaques act less as a dominant antigenic source and more as a trigger of broader self-antigen presentation in the CNS by microglia.

The prevailing view frames microglia and macrophages as guardians against amyloid beta (Aβ) accumulation in Alzheimer’s disease (AD). Here, we overturn this paradigm by demonstrating that human phagocytic cells, including differentiated THP-1 macrophages and hESC-derived microglia, are not merely passive responders but active producers of extracellular, seeding-competent Aβ42 fibrils, the amyloid species most strongly linked to parenchymal plaque formation and neurodegeneration. These cell-generated aggregates differ structurally and functionally from synthetic fibrils, displaying enhanced seeding and tau cross-seeding activity in biosensor models. Notably, Aβ42 fibril formation in this system requires active cellular processes and is exacerbated by loss of Triggering Receptor Expressed on Myeloid Cells 2 (TREM2), a major AD risk gene. Transcriptomic profiling reveals an early inflammatory response resembling microglial states observed in human AD models. Together, these findings support emerging evidence from in vivo studies that macrophages and microglia can influence amyloid seeding and introduce a human-relevant in vitro platform to explore how Aβ aggregation intersects with innate immune function and genetic risk. Our results reinforce the concept that microglia may play a dual role in AD, acting both as responders and inadvertent facilitators of amyloid assembly, with implications for early therapeutic intervention.

Alzheimer's disease (AD) is not an inevitable outcome of pathology but a dynamic process shaped by how brain cells respond to amyloid-β (Aβ) and tau. To disentangle these responses, we combined spatial transcriptomics and single-nucleus RNA sequencing of the superior frontal cortex from octogenarians living with or without dementia and from cognitively intact centenarians with comparable Aβ accumulation. We identified six distinct tissue domains representing a spatial pathological continuum of AD, with a key inflection point marked by a shift from Aβ-associated inflammatory changes to tau-associated cellular programs. This transition was accompanied by a change in microglial states, from early inflammatory to late antigen-presenting phenotypes, termed early and late plaque-induced gene (PIG) programs. Resilient individuals showed distinct pathological patterns: octogenarians without dementia lacked late PIGs, whereas centenarians showed late PIG activation that was uncoupled from tau accumulation. Together, these findings highlight divergent resilience-associated mechanisms in human aging and position microglial state transitions at the Aβ-tau interface as candidate points of resilience with potential therapeutic relevance.

Astrocytes and APOE are strongly implicated in Alzheimer’s disease (AD), yet the impact of astrocytes carrying different APOE variants on AD hallmarks remains incompletely understood. Here, we generate a chimeric model of AD by transplanting isogenic APOE3 or APOE4 human induced pluripotent stem cell-derived astrocyte progenitors into neonatal AD mice. Donor cells differentiate into human astrocytes that integrate into the cortex and display morphologies consistent with interlaminar-like astrocytes. APOE3 and APOE4 astrocytes differ in expression of APOE, which associates differentially with Aβ plaques. Notably, APOE3 astrocytes are associated with reduced Aβ burden, Tau pathology, and neuritic dystrophy, whereas APOE4 astrocytes exacerbate these processes. They also induce distinct microglial responses: APOE4 astrocytes enhance microglial clustering around Aβ plaques and promote a disease-associated microglia-like state, whereas APOE3 astrocytes reduce clustering and support a more homeostatic profile. These findings highlight a role for human astrocytes and APOE-dependent astrocyte functions in modulating AD-related pathology.

Tau pathology closely tracks neuronal loss in Alzheimer's disease, but how phosphorylated tau becomes lethal has remained unclear. Chen et al. identify a dual-hit mechanism: glucose hypometabolism removes a protective A20-mediated brake on necroptosis while phosphorylated tau scaffolds RIPK1 activation, driving tau-associated neuronal death.1.

A blood biomarker reveals the mechanistic shift from amyloid to tau pathology.

Xenotransplantation enables the interrogation of human neuron-specific vulnerabilities to Alzheimer’s pathology within a physiologically relevant in vivo context. While amyloid-beta (Aβ) is known to disrupt synaptic integrity, it remains uncertain whether the synaptotoxicity observed in vitro accurately models the disease. Here, we establish a xenotransplantation paradigm in which human neurons integrate into the brains of amyloid precursor protein (APP) transgenic mice that develop amyloid plaques. Using a genetically encoded pre-synaptic reporter, we label human pre-synapses post engraftment to assess early-stage pathology. We demonstrate that extracellular Aβ plaques induce localized synaptic damage in human neurons, characterized by local pre-synaptic loss and the formation of dystrophic neurites. Notably, this pathology is restricted to the plaque microenvironment and does not result in widespread pre-synaptic degeneration. Our findings establish this human-mouse chimera model as a platform for dissecting Aβ-induced synaptic pathology and reveal that extracellular Aβ exerts compartmentalized yet impactful toxicity on human pre-synapses.

Alzheimer's disease (AD) is a complex neurodegenerative disorder characterized by progressive cognitive decline. Although amyloid-β and tau pathologies remain central to our understanding of AD, growing evidence suggests that disrupted lipid metabolism and impaired bioenergetics are closely linked to these hallmark features. Genetic, lipidomic and functional studies point to alterations in cholesterol, phospholipids and polyunsaturated fatty acids, which can influence mitochondrial function, organelle communication and glial responses. These processes are further modulated by apolipoprotein E (APOE) genotype, sex differences and systemic metabolic states such as obesity and diabetes, contributing to neuroinflammation and cognitive decline. Although findings are sometimes conflicting, an emerging theme is that lipid and energy metabolisms are central to how genetic and environmental risk factors shape AD pathogenesis. This integrated perspective highlights lipid and bioenergetic pathways as promising therapeutic targets, where metabolic modulators, lipid-directed interventions and lifestyle strategies may complement amyloid-based therapies and offer opportunities for precision approaches, particularly in women and APOE ε4 carriers.

How polygenic risk translates into cellular dysfunction remains largely unknown in Alzheimer’s disease and related disorders. Here, we selected 31 donors broadly spanning the Caucasian-based polygenic risk distribution for AD and show that human polygenic architecture modulates microglial immune responses to amyloid pathology in vivo. We developed a “microglia village” model by xenotransplanting pooled iPSC-derived microglia from these genetically diverse donors into amyloid-bearing (AppNLGF) and control (AppHu) mouse brains, allowing the effect of genetic background to be separated from shared environmental influences. Marked inter-donor transcriptomic differences were observed in hMG derived from homeostatic, non-amyloid brain environments, demonstrating divergent baseline states across individuals. Amyloid exposure induced highly varied expression of MHC class II genes across donor, which correlated with individual’s AD polygenic risk scores. These findings demonstrate that polygenic risk can be decoded into functional immune phenotypes in human microglia and establish a scalable in vivo platform to dissect the genetic regulation of cellular responses in complex brain disorders.

While social and medical debate about the efficacy and safety of anti-Aβ immunotherapy is ongoing, one thing that emerged is that we have little understanding of the working mechanisms of these antibodies and this lack of knowledge complicates the interpretation of the clinical results. Here, we aimed to establish if microglia are required for the efficacy of Lecanemab, one of the most promising FDA-approved disease-modifying therapy for AD (Van Dyck et al. N Engl J Med 2023). To do so, we crossed App NL-G-F mice with Csf1r ΔFIRE/ΔFIRE mice (Rojo et al. Nat Commun 2019) to generate mice that show key features of Aβ pathology but genetically lack mouse microglia. We then assessed the effect of Lecanemab treatment on Aβ load and neuritic dystrophy. We demonstrate that Lecanemab lacks efficacy in the absence of microglia. On the other hand, when we xenotransplant human microglia into the brain of these mice (as described in Mancuso et al. Nat Neurosci 2019), we show that Lecanemab treatment significantly ameliorates both Aβ load and neuritic dystrophy. Furthermore, by employing scRNAseq on sorted human microglia, we demonstrate that Lecanemab treatment affects the transcriptome of the microglia by inducing a number of genes related phagocytosis, interferon response and immune activation. Functionally, we also established that Lecanemab-treated human microglia ingest more amyloid-β in vivo . Overall, we provide the first evidence that microglia are crucial for the efficacy of anti-Aβ immunotherapy and provide real insight into the working mechanisms of this first disease-modifying therapy for AD.

Background: Previously, we demonstrated that brain-penetrant complement C7 blocking antibody (mAb) mitigates brain inflammation, rescues synapse loss and improves cognition in Alzheimer’s disease (AD) mouse model (AppNL-G-F; PMID: 39215579), implicating membrane attack complex (MAC) in driving neuroinflammation in the model. Here, we extend this study, investigating the impact of anti-C7-mediated MAC inhibition on neuronal and glial health and function in the model. Method AppNL-G-F mice (6-9 months old, n = 12) were treated systemically with the recombinant brain penetrant mAb 73D1 anti-C7 mAb fused with nanobody (Nb62) targeting the transferrin receptor (TfR), to enable brain delivery (Nb62-r-mAb). The control AppNL-G-F group (n = 12) received the same C7-blocking antibody (73D1) without the brain-penetrant shuttle (control-r-mAb). Harvested brains were examined using immunofluorescence to detect neuronal degeneration (Fluoro-jade C), plaque pathology (Amylo-Glo) and microgliosis (CD11b and P2RY12). To measure impact on serum markers of neurodegeneration, the Alamar NULISAseq mouse panel (120 neurodegeneration-associated markers) was applied to serum samples from treated mice. Results: No differences were observed between Nb62-r-mAb treated and control mice in Aβ plaque number despite reduced Aβ in brain homogenates (TBH; p = 0.047) and suprislingly, plaque size was greater in Nb62-r-mAb treated mice (p = 0.024). Immunohistochemical analysis of hippocampus revealed fewer homeostatic (P2RY12-positive) and reactive (CD11b-positive) periplaque microglia in Nb62-r-mAb-treated mice (p = 0.023, p = 0.0449 respectively), indicating reduced microgliosis. Fluoro-Jade C IHC revealed fewer degenerating neurons in Nb62-r-mAb–treated mice (p = 0.005), demonstrating a neuroprotective effect. In support of these findings the Alamar NULISAseq mouse panel identified multiple serum markers of neurodegeneration as being altered between Nb62-r-mAb treated and control mice. Conclusion: Inhibiting C7/MAC in the brain of APPNL-G-F mice had minimal effect on amyloid plaques but reduced microgliosis and neuronal degeneration, providing further support for the therapeutic potential of brain-penetrant complement-inhibiting drugs for neuroprotection downstream of amyloid accumulation.

Microglia play a key role in the response to amyloid beta in Alzheimer's disease (AD). In this context, the major transcriptional response of microglia is the upregulation of APOE, the strongest late-onset AD risk gene. Of its three isoforms, APOE2 is thought to be protective, while APOE4 increases AD risk. We hypothesised that the isoforms change gene regulatory patterns that link back to biological function by shaping microglial transcriptomic and chromatin landscapes. We use RNA- and ATAC-sequencing to profile gene expression and chromatin accessibility of human microglia xenotransplantated into the brains of male APPNL-G-F mice. We identify widespread transcriptomic and epigenomic differences which are dependent on APOE genotype and are corroborated across the profiling assays. Our results indicate that impaired microglial proliferation, migration and immune responses may contribute to the increased risk for late-onset AD in APOE4 carriers, while increased phagocytic capabilities and DNA-binding of the vitamin D receptor in APOE2 microglia may contribute to the isoform's protective role.

Controversies over anti-amyloid immunotherapies underscore the need to elucidate their mechanisms of action. Here we demonstrate that Lecanemab, a leading anti-Aβ antibody, mediates amyloid clearance by triggering effector functions in the microglia. Using a human microglia xenograft model, we show that Lecanemab significantly reduces Aβ pathology and associated neuritic damage, while neither Fc-inactivated Lecanemab nor microglia deficiency elicit this effect despite intact plaque binding. Single-cell RNA sequencing and spatial transcriptomic analyses reveal that Lecanemab induces a focused transcriptional program that enhances phagocytosis, lysosomal degradation, metabolic reprogramming, interferon gamma genes, and antigen presentation. Finally, we identify SPP1/osteopontin as a major factor induced by Lecanemab treatment and demonstrate its role in promoting Aβ clearance. These findings highlight that effective amyloid removal depends on the engagement of microglia through Fc fragment, providing critical insights for optimizing anti-amyloid therapies in AD. ### Competing Interest Statement B.D.S. has been a consultant for Eli Lilly, Biogen, Janssen Pharmaceutica, Eisai, AbbVie and other companies and is now consultant to Muna Therapeutics. B.D.S is a scientific founder of Augustine Therapeutics and a scientific founder and stockholder of Muna Therapeutics. European Research Council, https://ror.org/0472cxd90, ERC-834682 CELLPHASE_AD VIB vzw KU Leuven and the Flemish Government, METH/21/05 Fonds voor Wetenschappelijk Onderzoek, G087523N KU Leuven, https://ror.org/05f950310 Queen Elisabeth Medical Foundation, https://ror.org/030x5z418 Stichting Alzheimer Onderzoek, SAO-FRA 20240017 Alzheimer's Association, AARF-22-968623

SUMMARYDespite strong evidence supporting that both astrocytes and apolipoprotein E (APOE) play crucial roles in the pathogenesis and progression of Alzheimer’s disease (AD), the impact of astrocytes carrying differentAPOEvariants on key AD pathological hallmarks remains largely unknown. To explore such effects in a human relevant context, we generated a chimeric model of AD. We transplanted isogenicAPOE3orAPOE4human induced pluripotent stem cell (hiPSC)-derived astrocyte progenitors into neonatal brains of AD model mice. We show that at five to six months after transplantation, transplanted cells have differentiated into mature astrocytes (h-astrocytes) that often integrate in upper layers of one cortical hemisphere.APOE3andAPOE4h-astrocytes differentially express and secrete the APOE protein, which binds to Aβ plaques with an isoform-dependent affinity. Remarkably,APOE3h-astrocytes ameliorate Aβ pathology, Tau pathology and neuritic dystrophy. In contrast,APOE4h-astrocytes aggravate these AD processes. Moreover,APOE3andAPOE4h-astrocytes modulate microglia responses to Aβ pathology in opposite ways.APOE4h-astrocytes enhance microglia clustering around Aβ plaques and exacerbate DAM state whereasAPOE3h-astrocytes reduce microglia clustering and induce a more homeostatic state on plaque-associated microglia. These findings highlight a critical contribution of h-astrocytes not only to Aβ pathology but also to other key AD hallmarks in chimeric mice. In addition, our findings reveal that h-astrocytes with differentAPOEvariants and the different forms of APOE they secrete have a crucial role in AD progression.

Astrocytes are a numerous and diverse glial subtype specialised to carry out distinct roles involving maintaining homeostasis and effective functioning of the nervous system. To do so effectively, they respond to and secrete various proteins. In addition, astrocytes have been linked to Alzheimer’s disease (AD), where they are believed to become reactive and contribute to neuroinflammation. A key feature of this reactive gliosis is the secretion of inflammatory mediators. Although in some instances this can be protective, in others, secretion of inflammatory mediators can be harmful, thus possibly contributing to AD pathology. At present, there remains a limited understanding of global astrocytic membrane and extracellular protein profiles and potential AD-associated changes. Here, we aimed to address this using a proximity labelling-based approach. Specifically, we used a viral construct containing TurboID targetted to the ER under a GFAP promoter in order to study astrocyte-specific proteins trafficked through the classical secretory pathway in an AD mouse model. After characterising the construct, we investigated changes in proteins between AD and controls over time in response to amyloid pathology. We have identified protein changes which are now being validated using mammalian biofluids such as CSF and plasma. This work enables a better understanding of astrocyte-specific membrane and extracellular protein changes as the disease progresses in a mammalian model of AD. An enhanced understanding of this will not only provide insight into astrocyte biology more generally, but may ultimately be vital for identification of novel biomarkers and therapeutic targets for detecting and treating AD.

In the brain as in other organs, complement contributes to immune defence and housekeeping to maintain homeostasis. Sources of complement may include local production by brain cells and influx from the periphery, the latter severely restricted by the blood brain barrier (BBB) in healthy brain. Dysregulation of complement leads to excessive inflammation, direct damage to self-cells and propagation of injury. This is likely of particular relevance in the brain where inflammation is poorly tolerated and brain cells are vulnerable to direct damage by complement. We have developed novel anti-C7 antibodies (mAb) that efficiently inhibit formation of the pro-inflammatory membrane attack complex (MAC) in vitro and in vivo . Here we describe recombinant fusion proteins (FP) that replicate the MAC-blocking action of the mAb, and are designed to access the brain utilising “Trojan horse” shuttles. The Alzheimer model APP NL-G-F mice were treated systemically with native mAb to swamp peripheral C7 followed by the FP. Immunohistochemistry and ELISA were used to demonstrate FP entry into brain and show impact on the disease pathology. The recombinant FP showed complement inhibitory activity in vitro equivalent to their parent mAb and were able to cross an artificial BBB in transwells. The presence of the FP in brain homogenates of peripherally dosed animals was confirmed by ELISA. Treatment with the FP caused reduced levels of complement activation products C3b and terminal complement complex (TCC) in brain. Diolistics analysis showed significant increased neuronal spine density in treated mice compared to controls, demonstrating a protective effect of the FP on synaptic function. Mice treated with the drug showed significant improvements in cognition. The FP described are able to cross BBB and are potent inhibitors of complement in brain; impact on brain pathology was detected after just one week of treatment. The findings highlight the potential for complement inhibition as a therapy in Alzheimer’s disease.