Understanding the endosomal network in neurodegeneration

Preceding any other neuropathology in Alzheimer’s Disease (AD) brain, neuronal early endosomes (EE) enlarge more than two-fold in response to abnormal activation of Rab 5 GTPase and upregulated transcription of Rab5 effectors. This Rab5 'phenotype' is selective for AD, both late onset and familial early onset subtypes linked to APP, including Down syndrome (DS). The EE phenotype on primary DS fibroblasts and AD/DS mouse models depends on elevated levels of APP-βCTF and not Aβ. APP-βCTF recruits the adaptor protein APPL1 to endosomes where it stabilises the activated state of rab5, thus linking βCTF directly to AD-related rab5 hyper-activation. APPL1 KD abrogates these events in neurons of DS-model mice, while its over-expression in wt mice enlarges Rab5-endosomes, as expected. Notably, gene polymorphisms conferring increased AD risk, including APOE4, SORL1, RIN3 etc. also hyper-activate Rab5 and enlarge EE, in some cases independently of APP-βCTF.

Extent of endosome swelling correlates with their slower rates and stalling during axonal/dendritic transport to the nucleus, including endosomes carrying NGF responsible for cholinergic neuron survival. Basal forebrain cholinergic neuron (BFCN) degeneration in APP-based AD models is reversed by lowering APP-βCTF or rab5 over-activation. Reproducing neuronal rab5 activation in mice phenocopies pathological consequences seen in DS model mice: EE enlargement and impaired axonal transport, tau hyperphosphorylation, LTP deficits, dendritic spine atrophy, BFCN neurodegeneration, and cognitive deficit. Neflamapimod (NFMD), a p38alpha kinase inhibitor that blocks rab5 hyper-activation at sub-micromolar concentrations, reverses the foregoing phenotype in DS model mice. A phase 2 clinical trial of NFMD in Lewy Body Dementia, NFMD improved cognitive and motor impairments attributable to prominent BFCN deficits in LBD. Support: NIA.

Professor Cullen's laboratory has sought to define the molecular machinery that serve to orchestrate the sorting of receptors, channels, transporters, polarity and adhesion molecules, and their associated proteins and lipids (collectively termed ‘cargoes’), through the endosomal network. Using structural and biochemical approaches coupled with in cellulo and in vivo analysis they have identified several evolutionary conserved multiprotein assemblies that conduct cargo sorting through the network, including human SNX3-Retromer and SNX27-Retromer complexes. This has provided new insight into cellular processes including nutrient sensing and signalling, cell adhesion and migration, and structural and functional synaptic plasticity, and has, in turn, begun to reveal how deregulation of endosomal sorting is associated with human disease. A current major focus is to apply an array of organelle-restricted proteomic approaches to unbiasedly quantify the global effect of retromer perturbation on the function of neurons and supporting cells within the context of brain development and age-related neurodegenerative disease. Read the preprint report.

The complexity, inaccessibility to experimentation of the human brain, lack of good animal models, and phenotypic heterogeneity and age dependence of most neurodegenerative diseases make finding causal pathways very difficult. Professor Petsko presents specific criteria to establish causality and validate a target and therapeutic approach. Given the complexity of the problem, such criteria should reflect an integrative biology approach: combining information and experiments from human genetics (are there causal mutations in the pathway?), molecular biology (are key molecules in the pathway dysregulated in the disease?), cellular biology (when disrupted, does the pathway cause the disease’s cardinal cellular pathologies, and can they be mediated by renormalising the pathway?), and anatomical biology (when disrupted, does the pathway cause the disease’s cardinal anatomical pathology?).

Professor Petsko applies these criteria, Koch’s Postulates for Neurodegenerative Diseases if you will, to the endosomal trafficking pathway regulated by the retromer multi-protein assembly. In conjunction with the retromer cargo receptor SORLA (product of the SORL1 gene), this pathway satisfies the integrative biology criteria: (1) rare alleles of SORL1 are causal for familial AD that phenocopies the common form; (2) retromer is deficient in the brains of both familial and sporadic AD patients; (3) disruption of SORL1-retromer dependent trafficking recapitulates the cardinal cellular pathologies of the disease in both patient-derived neurons and animals, and increasing retromer function rescues them; and (4) disruption of SORL1-retromer trafficking in animals causes synaptic dysfunction leading to neurodegeneration in the trans-entorhinal cortex, where AD pathology usually begins in the human disease. Professor Petsko therefore proposes the SORL1-retromer trafficking pathway as causal in many, if not most, cases of Alzheimer’s disease. Targeting this pathway may be of considerable therapeutic benefit.

Endolysosomal dysfunction as a converging mechanism in FTLD spectrum disorders

Christy Hung and Rickie Patani

The Francis Crick Institute, UK

Amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) were traditionally considered as two distinct neurodegenerative diseases which affect motor neurons and frontotemporal lobes, respectively. However, accumulating evidence suggests some convergence in molecular and genetic pathology, which may argue against their taxonomy as completely distinct entities. The valosin-containing protein (VCP/p97) gene is highly relevant, being associated with both FTD and ALS. Yet, the overlapping pathogenic pathways shared by both forms of the disease have remained unclear. This question is important because it provides clues to the most critical steps for therapeutic targeting. To address this, human induced pluripotent stem cells (iPSCs) carrying VCP mutations, which are causal for familial ALS and FTD, are directly differentiated into highly enriched and functionally validated cortical neurons and spinal cord motor neurons. Mutations in VCP lead to abnormal accumulation of enlarged endosomes, major defects in lysosomal function and defective axonal lysosome transport in both cortical and motor neurons. Taken together, these results suggest that the dysfunction of the endolysosomal-autophagic system represents a convergent mechanism shared by ALS and FTD.

Electron cryo-tomography of human iPSC-derived neurons to study structural changes in Alzheimer’s disease

Bronwen Foster^1,2, Michael Grange¹, Selina Wray² and Charles Arber²

¹Rosalind Franklin Institute, UK
²UCL Queen Square Institute of Neurology, UK

The endosomal lysosomal network (ELN) is disrupted in a variety of neurodegenerative diseases. In Alzheimer’s disease (AD), structural changes in the ELN have been described as the earliest documented AD-specific changes in mouse models and human neurons, specifically increased fusion and enlargement of Rab5 endosomes. Amyloid precursor protein (APP) processing occurs in endosomes, and APP-β C-terminal fragments directly over-activate Rab5, causing increased endocytosis and defects in endosome-mediated trafficking and signalling. This directly links aberrant APP processing and dysfunction of the ELN. Structural insight into these subcellular changes is required to appreciate the early molecular changes governing the AD mechanism. Electron cryo-tomography (cryo-ET) is an in-situ technique enabling visualisation of intracellular structures in their native context of the cell. This technique is being used to investigate cortical neurons derived from induced pluripotent stem cells (iPSC) harbouring mutations in APP and PSEN1 which are causative of familial AD. iPSC-derived cortical neurons can be grown on electron microscopy grids, allowing specific structures to be targeted using multi-modal approaches. AD-specific ELN defects visualised in situ in a variety of AD patient-derived iPSC lines will provide insight into the earliest fundamental molecular changes in the AD mechanism.

Anatomical studies have pinpointed the trans-entorhinal cortex (TEC) as one of the subregions within the brain most vulnerable to Alzheimer’s disease (AD). Interestingly, endosomal trafficking defects have now become a well-accepted feature of AD, based on a growing body of evidence from genetic, molecular, and biology studies. However, whether and how endosomal trafficking is linked to anatomical vulnerability still remains unclear. Here, in addressing these questions, the group began by showing that neurons are enriched with a functional distinct retromer core organized around VPS26b, differentially dedicated to endosomal recycling. Using fMRI imaging, and in contrast to its paralog VPS26a, the group further demonstrated that depleting VPS26b in mice caused a specific and age-dependent dysfunction in the TEC region, a finding further validated by electrophysiology, immunocytochemistry, and behaviour. Surprisingly, repletion of VPS26b, both in vitro and in vivo, rescued the phenotypes seen upon VPS26b deficiency, further supporting the relevance of VPS26b for the proper functioning of the TEC region. Additionally, when the group turned to humans, VPS26b was found highly enriched in the TEC region of healthy control individuals, and reduced – together with retromer-related receptor SORL1 – in patients with AD. Finally, by regulating glutamate receptor and SORL1 recycling and its levels, VPS26b was shown to mediate regionally-selective synaptic dysfunction, as well as AD-associated amyloid-precursor protein misprocessing and tau secretion. Together with the trans-entorhinal’s unique network properties, hypothesized to impose a heavy demand on endosomal recycling, our results contribute to the overall hypothesis explaining AD’s regional vulnerability.

Disruption of retromer-dependent endosomal trafficking is considered pathogenic in late-onset Alzheimer's disease (AD). Here, to investigate this disruption in the intact brain, Dr Qureshi turns to a genetic mouse model where the retromer core protein VPS35 is depleted in hippocampal neurons, and then he repletes VPS35 using an optimised viral vector protocol. The VPS35 depletion-repletion studies strengthen the causal link between the neuronal retromer and AD-associated neuronal phenotypes, including the acceleration of amyloid precursor protein cleavage and the loss of synaptic glutamate receptors. Moreover, the studies show that the neuronal retromer can regulate a distinct, dystrophic, microglia morphology, phenotypic of hippocampal microglia in AD. Finally, the neuronal and, in part, the microglia responses to VPS35 depletion were found to occur independent of tau. Showing that the neuronal retromer can regulate AD-associated pathologies in two of AD's principal cell types strengthens the link, and clarifies the mechanism, between endosomal trafficking and late-onset sporadic AD.

Tauopathies, including Alzheimer’s disease and frontotemporal dementia, share the feature of apparent spatial spreading through neural networks, between diseased and healthy neurons. In this presentation, Professor Livesey will discuss the contribution of the endolysosome system to inter-neuronal transfer of pathogenic forms of microtubule-associated tau protein. Endolysosome dysfunction is emerging as an important pathogenic process in tauopathies, influencing both disease initiation and progression through the CNS. Autosomal dominant mutations in PSEN1 and APP causal for Alzheimer’s disease compromise human neuronal endolysosome function and autophagy. This results in extracellular release of cleaved forms of tau protein that contain the aggregation-prone microtubule-binding region (MTBR), which are efficiently taken up by neurons via LRP1-mediated endocytosis, which Professor Livesey proposes contributes to disease propagation. Whole genome CRISPR knockout screens in human neurons were used to identify the genes and pathways required for uptake of extracellular tau. In addition to identifying surface receptors for tau and routes for tau’s receptor-mediated endocytosis, the group found that the large multifunctional protein LRRK2 is a key regulator of neuronal tau endocytosis. Given LRRK2’s role in Parkinson’s disease pathogenesis, Professor Livesey will discuss the possible similarities in disease mechanisms involving the endolysosome system between tauopathies and other neurodegenerative diseases.

Dysfunction in the endo-lysosomal network (ELN) is an early driver of Alzheimer’s disease pathogenesis and therapeutics targeting this pathway could be novel and effective. However, understanding how the ELN functions in the diverse cell types of the brain and how ELN genes that confer increased AD risk modulate these functions are necessary. The group uses human induced pluripotent stem cell-derived neurons and glia to elucidate the role of the sorting receptor SORL1, an established AD risk gene, in ELN function. SORL1 is an adaptor that interacts with the multi-protein sorting assembly, retromer, to traffic cellular cargo. The most characterized cargo sorted by SORL1-retromer is the amyloid precursor protein APP.

The data indicates that in neurons, loss of SORL1 function converges on the early endosome, leading to endosome enlargement and mislocalization of several important neuronal cargo, including neurotrophin receptors and subunits of post-synaptic complexes. These 'traffic jams' have an impact on localization of synaptic machinery and neuronal function. In microglia, they demonstrate that loss of SORL1 impacts the lysosome, inducing lysosomal enlargement, impairing lysosomal enzymes and degradation of phagocytosed substrates. Finally, they show that enhancement of retromer function via a small molecules can rescue ELN dysfunction in neurons, however the degree of phenotypic rescue depends on whether there is a functional copy of SORL1 in the cell. Using a disease-relevant preclinical model, our work illuminates how the SORL1-retromer pathway can be therapeutically harnessed.

TMCC2, an apoE and APP-interacting endolysosomal protein, associates with Alzheimer’s disease pathology in the human brain.

Paul CR Hopkins, Claire Troakes, Andrew King and Guy Tear 
King’s College London, UK

Transmembrane and Coiled-Coil 2 (TMCC2) is an ER and endolysosome-associated protein that forms complexes with both apolipoprotein E (apoE) and the amyloid protein precursor (APP), two proteins central to the aetiology of Alzheimer’s disease (AD). Mutation of Drosophila TMCC2 (Dementin) can cause neurodegeneration with features of AD, ie accumulation of fragments of the APP-Like protein, synaptogenic defects, mis-localisation of cytoskeletal proteins and early death.

Here the investigators report the first investigation of TMCC2 in AD cases associated with apoE, mutations in APP Val717, and Down syndrome. TMCC2 and APP immunoreactivities are closely associated in neurons of both control and late onset AD cases. In all AD cases TMCC2 immunoreactivity was associated with dense core plaques and adjacent dystrophic neurites, but not with amyloid surrounding the core, nor with diffuse amyloid plaques or neurofibrillary tangles. In Down syndrome neuritic plaques, TMCC2 immunoreactivity additionally co-localized with amyloid having a spicular or threadlike pattern not seen in the non-Down syndrome cases examined.

Western blot revealed that TMCC2 exists as at least three protein isoforms, the relative abundance of which varied between the temporal gyrus and cerebellum, and was influenced by APOE genotype and/or dementia status.

Mutations in two Parkinsonism-linked endocytic genes SYNJ1 and DNAJC6 have synergistic effects on dopaminergic axonal pathology  

Xin Yi Ng¹, Yumei Wu^2*, Youneng Lin^1*, Sidra Mohamed Yaqoob^1*, Lois E Greene³, Pietro De Camilli² and Mian Cao^1,4

¹Programme in Neuroscience and Behavioural Disorders, Duke-NUS Medical School, Singapore
²Departments of Neuroscience and Cell Biology, Howard Hughes Medical Institute, Program in Cellular Neuroscience, Neurodegeneration and Repair, Kavli Institute for Neuroscience, Yale University School of Medicine, USA. Aligning Science Across Parkinson’s (ASAP) Collaborative Research Network, Chevy Chase, MD.
³Laboratory of Cell Biology, NHLBI, National Institutes of Health, MD, USA
⁴Department of Physiology, National University of Singapore, Singapore

^*These authors contributed equally

Parkinson’s disease (PD) is a neurodegenerative disorder characterised by defective dopaminergic (DAergic) input to the striatum. Mutations in two genes encoding synaptically-enriched clathrin-uncoating factors, synaptojanin 1 (SJ1) and auxilin, have been implicated in atypical Parkinsonism. SJ1 knock-in (SJ1-KI^RQ) mice carrying a disease-linked mutation display neurological manifestations reminiscent of Parkinsonism. Here Dr Cao’s group reports that auxilin knockout (Aux-KO) mice display dystrophic changes of a subset of nigrostriatal DAergic terminals similar to those of SJ1-KI^RQ mice. Furthermore, Aux-KO/SJ1-KI^RQ double mutant mice have shorter lifespan and more severe synaptic defects than single mutant mice. These include increase in dystrophic striatal nerve terminals positive for DAergic markers and for the PD risk protein SV2C, as well as adaptive changes in striatal interneurons. The synergistic effect of the two mutations demonstrates a special lability of DAergic neurons to defects in clathrin uncoating, with implications for PD pathogenesis in at least some forms of this condition.

SORL1, the gene encoding the large multi-domained SORLA protein, has emerged as only the fourth gene that when mutated can by itself cause Alzheimer’s disease (AD), and as a gene reliably linked to both the early- and late-onset forms of the disease. SORLA is known to interact with the endosomal trafficking regulatory complex called retromer in regulating the recycling of endosomal cargo, including the amyloid precursor protein (APP) and the glutamate receptor GluA1. Nevertheless, SORLA’s precise structural-functional relationship in endosomal recycling tubules remains unknown. Here Dr Andersen addresses these outstanding questions by relying on crystallographic and artificial-intelligence evidence to generate a structural model for how SORLA folds and fits into retromer-positive endosomal tubules, where it is found to dimerise via both SORLA’s fibronectin-type-III (3Fn)- and VPS10p-domains. Moreover, he identifies a SORLA fragment comprising the 3Fn-, transmembrane, and cytoplasmic domains that has the capacity to form a dimer, and to enhance retromer-dependent recycling of APP by decreasing its amyloidogenic processing. Collectively, these observations generate a model for how SORLA dimer (and possibly polymer) formation can function in stabilising and enhancing retromer function at endosome tubules. These findings can inform investigation of the many AD-associated SORL1 variants for evidence of pathogenicity and can guide towards discovery of novel drugs for the disease.

Microglial activation is a neuropathological hallmark of the lysosomal storage disease Niemann-Pick type C (NPC). Whole-body knockout of NPC1 results in enhanced phagocytic uptake and impaired myelin turnover in microglia that precede neuronal death. Npc1-deficient microglia are characterized by accumulation of multivesicular bodies and impaired intracellular trafficking of lipids to lysosomes while lysosomal degradation remains preserved.

This work reveals that a specific loss of NPC1 in myeloid cells is sufficient to trigger pathology in other brain cells, suggesting that a pathological cascade initiated by microglial dysfunction may propagate to astrocytes and neurons.

Background: An estimated ~2-3% of early-onset Alzheimer’s disease (AD) patients carry a predicted rare damaging variant in the SORL1 gene, which encodes the SORLA receptor protein. Thus far, >500 unique variants have been identified, but to what extent each variant affects SORLA protein function is unclear. SORLA is cleaved at the neuronal membrane into soluble SORLA (sSORLA) and shed into the interstitial fluid, which flows into the cerebrospinal fluid (CSF). Here, the group tested whether the level of sSORLA in the CSF of AD patients reflects the pathogenicity of the SORL1 variant they carry.

Methods: sSORLA levels were determined in CSF from patients from the Amsterdam Dementia Cohort (ADC) using: (1) Western Blotting (WB) - 37 AD patients who carried a SORL1 variant with variable predicted pathogenicity; (2) CSF-proteomics in 606 individuals with AD, minor cognitive impairment (MCI) and normal cognition, including carriers of SORL1 variants with varying predicted pathogenicity. Additionally, the group tested whether changes in sSORLA levels in culture-medium of cell lines transfected with constructs encoding SORL1 variants were representative of changes observed in patient-derived CSF.

Results: Using WB, sSORLA levels in carriers of predicted likely pathogenic (n=8) and pathogenic (n=5) SORL1 variants were reduced to 48.6%±18.3% (p=6.8x10^-3) and 65.6%±32.0%, (p=2.0x10^-1) relative to AD patients who carry benign SORL1 variants (n=11). The in vitro cell-based functional assay demonstrated a similar decrease. With CSF-proteomics the group validated that sSORLA levels from carriers of predicted pathogenic variants (n=3) were significantly reduced relative to carriers of the WT allele (n=135); p=3.8x10^-7.

Discussion: Decreased sSORLA levels in patient-derived CSF or in in vitro cell-based functional assay may serve as a biomarker for damaging SORL1 genetic variants. These results encourage the investment in precise quantification methods of sSORLA levels.