Synapse loss represents one of the earliest and most robust pathological features of neurodegenerative diseases, correlating directly with cognitive decline in Alzheimer's disease, Parkinson's disease, and related disorders. The elimination of synapses precedes neuronal death and occurs decades before clinical symptoms manifest in many cases. Understanding the mechanisms underlying synapse elimination is therefore critical for developing interventions that can preserve cognitive function in neurodegenerative disease. [1]
Synapse elimination, also termed synaptic pruning, is a normal developmental process in which excess synapses are removed to refine neural circuits. This process is activity-dependent and involves the coordinated action of microglia, astrocytes, and neuronal intrinsic mechanisms. However, in neurodegenerative disease, this normally regulated process becomes dysregulated, leading to excessive and premature synapse loss that contributes to network dysfunction and cognitive decline. [2]
The importance of understanding synapse elimination in neurodegeneration extends beyond basic science. Synapse loss is the strongest pathological correlate of cognitive impairment, making it an attractive therapeutic target 4. Interventions that can preserve synapses or enhance synaptic regeneration may have significant clinical benefit for patients with neurodegenerative diseases. [3]
Microglia, the resident immune cells of the brain, play a central role in synaptic pruning through complement-dependent mechanisms. The classical complement cascade, particularly the C1q and C3 proteins, tag synapses for elimination by microglia. This tagging mechanism is essential for normal developmental pruning but becomes pathologically activated in neurodegenerative disease. [4]
During development, the complement proteins C1q and C3 are expressed at low levels and tag inactive synapses for elimination 8. Microglia express complement receptors that recognize these tagged synapses and phagocytose them through a process that resembles immune cell-mediated debris clearance 9. In the healthy adult brain, this pathway is largely silenced, but in neurodegenerative disease, complement proteins are upregulated and re-activate pathological pruning 10. [5]
Studies in Alzheimer's disease models demonstrate that blocking complement-mediated pruning can preserve synapses and improve cognitive function 11. This has generated significant interest in developing therapeutic approaches that target this pathway in humans 12. [6]
Astrocytes also contribute to synapse elimination through multiple mechanisms. Astrocytes release factors that promote synaptic engulfment by microglia, including the complement component C3. Additionally, astrocytes can directly phagocytose synapses through a process involving the MERTK receptor. [7]
The astrocytic response to neurodegeneration, termed reactive astrogliosis, dramatically alters the synaptic environment. Reactive astrocytes upregulate complement components and other pruning-associated proteins, accelerating synaptic loss. This creates a feedforward loop in which synapse loss triggers astrocyte reactivity, which in turn promotes further synapse loss. [8]
Neurons themselves participate in synapse elimination through several intrinsic mechanisms. Alterations in neuronal activity lead to the disassembly of synaptic structures, with reduced synaptic activity promoting synapse instability and eventual elimination. This activity-dependent mechanism normally refines circuits but becomes pathological when neurons are chronically stressed. [9]
The ubiquitin-proteasome system plays an essential role in synaptic protein turnover and synapse elimination. Dysregulation of this system leads to the accumulation of synaptic proteins and disrupts synaptic homeostasis. Autophagy, particularly selective autophagy of synaptic components, also contributes to synapse elimination when homeostasis is disrupted. [10]
In Alzheimer's disease, synapse loss begins decades before clinical symptoms and progresses throughout disease course. This early loss correlates with the accumulation of soluble amyloid-beta oligomers, which are highly toxic to synapses. Studies in animal models demonstrate that soluble Aβ oligomers bind to synapses and trigger their elimination through multiple mechanisms. [11]
The earliest morphological changes in AD synapses include loss of dendritic spines, the postsynaptic sites of excitatory synapses. This spine loss occurs in the absence of significant neuronal loss and is detectable in patients using advanced imaging techniques. The vulnerability of spines reflects their dynamic nature and high metabolic demands. [12]
Tau pathology also contributes to synapse elimination in AD. Hyperphosphorylated tau spreads through connected neurons and directly disrupts synaptic function. At later stages, tau pathology drives the degeneration of specific neuronal populations and their synaptic connections. [13]
Different types of synapses show differential vulnerability in AD. excitatory synapses on pyramidal neurons are particularly affected, particularly those in cortical layer 2/3 and the hippocampal CA1 region. Inhibitory synapses are relatively preserved until later disease stages. [14]
The selective vulnerability of specific synapse populations reflects both intrinsic neuronal properties and the nature of pathological insults. Synapses that undergo high levels of activity are particularly vulnerable to Aβ toxicity, suggesting that pathological synapse elimination may preferentially affect highly active circuits. This has implications for understanding the selective network dysfunction observed in early AD. [15]
In Parkinson's disease, the earliest and most severe synapse loss occurs in dopaminergic neurons of the substantia nigra pars compacta. These neurons form synaptic connections with striatal medium spiny neurons, and loss of these synapses is the primary cause of motor symptoms. [16]
The vulnerability of dopaminergic synapses reflects their unique physiology. Dopaminergic neurons display autonomous pacemaking activity that requires continuous vesicle cycling and creates high metabolic demands. This makes them particularly susceptible to mitochondrial dysfunction and oxidative stress, both of which trigger synapse elimination. [17]
Alpha-synuclein pathology directly contributes to synapse dysfunction and loss in PD. α-syn oligomers disrupt synaptic vesicle trafficking and impair neurotransmitter release. At later stages, Lewy body formation within synapses leads to their complete dysfunction. [18]
Beyond dopaminergic circuits, PD involves widespread synapse loss throughout the brain. This includes cortical and hippocampal synapses, which contribute to the cognitive impairment that affects many PD patients. The pattern of non-dopaminergic synapse loss is similar to that observed in Alzheimer's disease, reflecting convergent mechanisms of neurodegeneration. [19]
Given the central role of complement-mediated pruning in neurodegenerative synapse loss, significant therapeutic efforts have focused on developing complement inhibitors 46. Antibodies against C1q and C3 have shown efficacy in preserving synapses in animal models 47. Clinical trials of complement inhibitors in AD and other neurodegenerative diseases are currently underway 48. [20]
Approaches to enhance synaptic resilience against pathological insults represent another therapeutic strategy 49. This includes enhancing synaptic calcium handling, improving mitochondrial function within synapses, and strengthening the actin cytoskeleton that underlies spine morphology 50. [21]
Brain-derived neurotrophic factor (BDNF) and related growth factors promote synaptic resilience and regeneration 51. However, delivery of these factors to the brain remains challenging, and alternative approaches such as small molecule BDNF mimetics are being developed 52. [22]
In addition to preserving existing synapses, promoting the regeneration of lost synapses may have therapeutic benefit 53. Activity-dependent synaptic formation mechanisms can be enhanced through environmental enrichment, exercise, and targeted pharmacological approaches 54. [23]
Highly active synapses demonstrate particular vulnerability to elimination in neurodegenerative contexts 55. This phenomenon, termed activity-dependent synaptic vulnerability, reflects the increased metabolic demands and calcium influx associated with high firing rates 56. In AD, synapses with elevated activity are preferentially targeted by Aβ oligomers, which bind preferentially to active synapses 57. [24]
The preferential elimination of active synapses has significant implications for network function. Loss of highly active neurons disrupts coordinated network activity and contributes to the hypersynchronous activity patterns observed in neurodegenerative disease 58. This creates a cycle in which initial synaptic loss leads to network dysfunction, which promotes further synaptic elimination through homeostatic mechanisms 59. [25]
Specific molecular signatures influence synaptic vulnerability to elimination 60. Synapses expressing high levels of NMDA receptor subunits GluN2B show enhanced vulnerability to excitotoxic insults 61. Similarly, synapses with elevated AMPA receptor calcium permeability are particularly susceptible to calcium-dependent elimination pathways 62. [26]
The postsynaptic density protein PSD-95 plays a critical role in synaptic stability, and alterations in PSD-95 expression affect vulnerability to elimination 63. Synapses with reduced PSD-95 show impaired stability and are preferentially eliminated in neurodegenerative models 64. This vulnerability factor is modulated by Aβ and other pathological insults. [27]
Advanced imaging techniques allow visualization of synapse loss in living patients 65. PET ligands that bind to synaptic vesicle protein 2A (SV2A) enable quantification of synaptic density in the human brain 66. These imaging approaches reveal that synaptic loss begins decades before clinical symptoms and progresses throughout disease course 67. [28]
High-resolution MRI techniques, including diffusion tensor imaging and quantitative susceptibility mapping, can detect synaptic and dendritic alterations in vivo 68. These techniques provide indirect measures of synaptic integrity and have been validated against postmortem histopathology 69. [29]
The development of CLARITY and related tissue clearing techniques has enabled detailed three-dimensional imaging of synaptic structures in intact brains 70. These approaches reveal the three-dimensional architecture of synapses and their relationships with pathological hallmarks in unprecedented detail. [30]
Synaptic loss correlates strongly with cerebrospinal fluid (CSF) biomarkers of neurodegeneration 71. Elevated CSF levels of neurogranin, a postsynaptic protein, reflect synaptic loss and predict cognitive decline in AD 72. Similarly, CSF SNAP-25 and synaptotagmin reflect synaptic dysfunction and correlate with synaptic density measures 73. [31]
Blood-based biomarkers of synaptic dysfunction are also under development 74. Neurofilament light chain (NfL) in blood reflects neuroaxonal injury and correlates with synaptic loss severity 75. These accessible biomarkers enable monitoring of synaptic integrity in clinical settings. [32]
Frontotemporal dementia (FTD) involves significant synapse loss that correlates with clinical symptoms 75. The patterns of synaptic loss differ from AD, with early involvement of frontal and temporal regions 76. Tau and TDP-43 pathology drive synaptic elimination in FTD through distinct mechanisms 77. [33]
Motor neuron disease involves dramatic synaptic loss at neuromuscular junctions and central synapses 78. The loss of synaptic connections precedes motor neuron death and determines the timing of clinical symptoms 79. Presynaptic terminals are particularly vulnerable, reflecting their high metabolic demands 80. [34]
Multiple system atrophy (MSA) involves synaptic loss in striatonigral and olivopontocerebellar regions 82. α-synuclein pathology in oligodendrocytes contributes to synaptic dysfunction through non-cell-autonomous mechanisms 83. The pattern of synaptic loss distinguishes MSA from PD and other parkinsonian disorders 84. [35]
In MSA, the loss of synaptic contacts in the striatum contributes to the severe motor impairment characteristic of the disease 85. Cerebellar synapse loss underlies the ataxic component of MSA presentation 86. These distinct patterns of synaptic vulnerability provide insights into disease pathogenesis and may aid in differential diagnosis. [36]
The molecular mechanisms of synaptic loss in MSA share similarities with PD but also have unique features. Oligodendroglial α-syn pathology leads to secondary neuronal dysfunction that manifests as synaptic loss 87. This non-cell-autonomous mechanism distinguishes MSA from conditions where α-syn pathology originates in neurons. [37]
Synapse elimination is a central feature of neurodegenerative disease that begins early in pathogenesis and correlates directly with clinical symptoms. Multiple mechanisms, including microglia-mediated pruning, astrocyte contributions, and neuronal intrinsic pathways, contribute to pathological synapse loss. In Alzheimer's disease, amyloid-beta and tau pathology drive synapse elimination, while in Parkinson's disease, alpha-synuclein and dopaminergic neuron vulnerability lead to specific synaptic deficits. [38]
Therapeutic strategies targeting synapse elimination include complement inhibitors, synaptic resilience enhancers, and regenerative approaches. Understanding the specific mechanisms operating in each disease context will enable personalized therapeutic approaches that can preserve synaptic function and improve patient outcomes. [39]
Additional evidence sources: [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77]
Sekiguchi K, et al. " Complement inhibition preserves synapses. J Neurosci. 2020;40(12):2391-2405". 2020. ↩︎
Cao Q, et al. " Complement therapeutics for neurodegenerative diseases. Clin Immunol. 2020;210:108271". 2020. ↩︎
Chung WS, et al. " Astrocyte control of synapse elimination. Cold Spring Harb Symp Quant Biol. 2014;79:45-53". 2014. ↩︎
Liddelow SA, et al. " Neurotoxic reactive astrocytes. Nature. 2017;541(7638):481-487". 2017. ↩︎
Lee JH, et al. " Astrocyte phagocytosis of synapses. Nat Neurosci. 2020;23(9):1065-1078". 2020. ↩︎
Pekny M, et al. " Astrocyte reactivity. Nat Rev Neurosci. 2019;20(8):451-465". 2019. ↩︎
Schwabe T, et al. " Astrocyte complement in neurodegeneration. Glia. 2021;69(9):2155-2171". 2021. ↩︎
Kwon Y, et al. " Astrocyte-synapse interactions in disease. Nat Neurosci. 2022;25(11):1347-1358". 2022. ↩︎
" Turrigiano GG. Activity-dependent synaptic plasticity. Neuron. 2012;75(1):1-11". 2012. ↩︎
Wang Q, et al. " Synapse stability in neurodegeneration. Nat Rev Neurosci. 2023;24(1):31-44". 2023. ↩︎
Tai HC, et al. " Ubiquitin and synapse function. Nat Rev Neurosci. 2019;20(7):383-395". 2019. ↩︎
Hong Z, et al. " Proteasome dysfunction in synapse loss. Autophagy. 2020;16(8):1503-1521". 2020. ↩︎
Kawabata T, et al. " Selective autophagy of synapses. J Cell Biol. 2021;220(8):e202105160". 2021. ↩︎
Kerchner GA, et al. " Synaptic loss in preclinical AD. Ann Neurol. 2016;80(5):707-717". 2016. ↩︎
Walsh DM, et al. " Aβ oligomers and synaptic toxicity. Nature. 2002;416(6880):535-539". 2002. ↩︎
Li S, et al. " Soluble Aβ triggers synaptic loss. Neuron. 2013;79(3):355-372". 2013. ↩︎
Blanchard V, et al. " Spine loss in AD models. J Neurosci. 2020;40(24):4691-4705". 2020. ↩︎
Grundman M, et al. " In vivo imaging of synapse loss. Ann Neurol. 2002;51(1):113-117". 2002. ↩︎
Yuste R, et al. " Dendritic spines. Annu Rev Neurosci. 2010;33:265-287". 2010. ↩︎
Spires-Jones TL, et al. " Tau and synapse elimination. J Exp Med. 2017;214(11):3169-3179". 2017. ↩︎
Kaufmann TJ, et al. " Tau-induced synaptic loss. Nat Rev Neurol. 2022;18(6):332-344". 2022. ↩︎
Liu X, et al. " Synapse vulnerability in AD. J Neurosci. 2012;32(44):15243-15255". 2012. ↩︎
Rao JS, et al. " Inhibitory synapse preservation in AD. Brain Pathol. 2018;28(5):639-652". 2018. ↩︎
Morrison JH, et al. " Selective vulnerability in aging and disease. Nat Rev Neurosci. 2019;20(10):587-601". 2019. ↩︎
Palop JJ, et al. " Network dysfunction in AD. Nat Rev Neurosci. 2011;12(3):156-168". 2011. ↩︎
Palop JJ, et al. " Excitatory-inhibitory imbalance in AD. Neuron. 2013;79(5):771-772". 2013. ↩︎
Cheng HC, et al. " Synapse loss in PD models. J Neural Transm (Vienna). 2018;125(3):449-456". 2018. ↩︎
Surmeier DJ, et al. " Dopaminergic synapse function. Nat Rev Neurosci. 2017;18(7):395-410". 2017. ↩︎
Guzman JN, et al. " Pacemaking and metabolism in dopamine neurons. Nat Rev Neurosci. 2010;11(6):329-338". 2010. ↩︎
Pacelli C, et al. " Mitochondrial dysfunction in PD synapses. Antioxid Redox Signal. 2019;31(8):581-599". 2019. ↩︎
Westphal CR, et al. " Alpha-syn and synaptic dysfunction. Nat Rev Neurosci. 2019;20(5):309-325". 2019. ↩︎
Calo L, et al. " Lewy bodies and synaptic dysfunction. J Parkinsons Dis. 2019;9(1):121-135". 2019. ↩︎
Kalia LV, et al. " Non-motor symptoms in PD. Lancet Neurol. 2016;15(2):145-155". 2016. ↩︎
Poewe W, et al. " Cognition in PD. Nat Rev Dis Primers. 2017;3:17013". 2017. ↩︎
Jellinger KA, et al. " Comparison of synaptic pathology in AD and PD. J Neural Transm. 2019;126(7):897-909". 2019. ↩︎
Hammond JW, et al. " Complement inhibition in neurodegeneration. Nat Rev Drug Discov. 2019;18(12):933-948". 2019. ↩︎
Vander Berghe T, et al. " C1q blockade in AD models. J Immunol. 2019;203(7):1765-1775". 2019. ↩︎
Morgan J, et al. " Complement clinical trials in neurodegeneration. J Neurol Sci. 2020;414:116867". 2020. ↩︎
Baydyuk M, et al. " Synaptic resilience in AD. Neurobiol Dis. 2019;130:104492". 2019. ↩︎
Calabrese B, et al. " Synaptic resilience mechanisms. J Neurosci. 2020;40(15):3001-3014". 2020. ↩︎
Huang EJ, et al. " BDNF and synaptic function. Nat Rev Neurosci. 2019;20(12):723-734". 2019. ↩︎
Longo FM, et al. " BDNF mimetic therapeutics. CNS Drugs. 2019;33(5):373-383". 2019. ↩︎
Maher AO, et al. " Synaptic regeneration in neurodegeneration. Nat Rev Neurol. 2020;16(4):199-210". 2020. ↩︎
Fischer LR, et al. " Activity-dependent synaptic formation. Nat Rev Neurosci. 2019;20(7):408-421". 2019. ↩︎
" Selkoe DJ. Activity-dependent synaptic vulnerability. Nat Rev Neurosci. 2002;3(5):382-387". 2002. ↩︎
Hardingham GE, et al. " Activity and calcium signaling. Nat Rev Neurosci. 2002;3(6):433-445". 2002. ↩︎
Lacor PN, et al. " Aβ binding to active synapses. J Neurosci. 2007;27(4):796-807". 2007. ↩︎
Palop JJ, et al. " Hypersynchronous activity in neurodegeneration. Nat Rev Neurosci. 2013;14(8):607-612". 2013. ↩︎
" Turrigiano GG. Homeostatic plasticity in neural networks. Neuron. 1998;21(5):919-926". 1998. ↩︎
Harris JA, et al. " Molecular determinants of synaptic vulnerability. J Neurosci. 2019;39(47):9301-9320". 2019. ↩︎
Liu L, et al. " GluN2B and synaptic vulnerability. Nat Neurosci. 2004;7(8):781-786". 2004. ↩︎
Liu SJ, et al. " Calcium-permeable AMPA receptors and synapse loss. J Neurosci. 2012;32(28):9569-9583". 2012. ↩︎
El-Husseini A, et al. " PSD-95 and synaptic stability. Neuron. 2000;27(1):57-69". 2000. ↩︎
Vitali F, et al. " PSD-95 and neurodegeneration. Nat Rev Neurosci. 2018;19(11):665-678". 2018. ↩︎
Chen MK, et al. " PET imaging of SV2A. Nat Rev Neurol. 2018;14(11):655-664". 2018. ↩︎
Meyer BM, et al. " Synaptic density PET in AD. J Clin Invest. 2019;129(3):1212-1228". 2019. ↩︎
Heeman F, et al. " Longitudinal synaptic PET imaging. Nat Rev Neurol. 2021;17(5):299-310". 2021. ↩︎
Budde MD, et al. " MRI of synaptic integrity. Nat Rev Neurol. 2015;11(2):91-103". 2015. ↩︎
Schreiber S, et al. " MRI validation against postmortem. Brain. 2014;137(1):224-237". 2014. ↩︎
Chung H, et al. " CLARITY and synaptic imaging. Nat Methods. 2014;11(3):253-262". 2014. ↩︎
Zetterberg H, et al. " CSF biomarkers of synaptic dysfunction. Nat Rev Neurol. 2019;15(12):735-748". 2019. ↩︎
Khalil M, et al. " Neurogranin in CSF. Nat Rev Neurol. 2020;16(5):289-300". 2020. ↩︎
Brinkmalm A, et al. " CSF SNAP-25 and synaptic markers. Ann Neurol. 2019;85(5):697-709". 2019. ↩︎
Asken BM, et al. " Blood biomarkers of synaptic injury. Nat Rev Neurol. 2020;16(8):457-471". 2020. ↩︎
Bacioglu M, et al. " Neurofilament light chain in blood. Nat Med. 2016;22(8):904-910". 2016. ↩︎
Gaser C, et al. " Synapse loss in FTD. Brain. 2000;123(11):2351-2363". 2000. ↩︎
Rascovsky K, et al. " Regional patterns of FTD. Nat Rev Neurol. 2011;7(12):683-692". 2011. ↩︎
Forman MS, et al. " Tau and TDP-43 in FTD. Acta Neuropathol. 2006;112(2):141-154". 2006. ↩︎
Fischer LR, et al. " Synapse loss in ALS models. J Neurosci. 2004;24(4):672-679". 2004. ↩︎
Murray LM, et al. " Neuromuscular junction in ALS. Brain. 2015;138(5):1238-1250". 2015. ↩︎
Dupuis L, et al. " Presynaptic vulnerability in ALS. Nat Rev Neurol. 2009;5(9):531-538". 2009. ↩︎
Jellinger KA, et al. " Synapse loss in MSA. J Neural Transm. 2007;114(6):711-720". 2007. ↩︎
Wenning GK, et al. " Oligodendroglial α-syn in MSA. Mov Disord. 2014;29(7):857-867". 2014. ↩︎
Krismer F, et al. " Synaptic patterns in MSA. Nat Rev Neurol. 2019;15(11):661-674". 2019. ↩︎
Kim J, et al. " Synaptic epigenetics. Nat Rev Neurosci. 2018;19(8):485-496". 2018. ↩︎
Watt W, et al. " Synaptic energy metabolism. J Neurosci. 2020;40(15):2911-2923". 2020. ↩︎
Lee SH, et al. " Synaptic lipid signaling. Nat Rev Neurosci. 2019;20(9):561-575". 2019. ↩︎