β-amyloid induces a dying-back process and remote trans-synaptic alterations in a microfluidic-based reconstructed neuronal network
© Deleglise et al.; licensee BioMed Central Ltd. 2014
Received: 16 August 2014
Accepted: 14 September 2014
Published: 25 September 2014
Recent histopathological studies have shown that neurodegenerative processes in Alzheimer's and Parkinson's Disease develop along neuronal networks and that hallmarks could propagate trans-synaptically through neuronal pathways. The underlying molecular mechanisms are still unknown, and investigations have been impeded by the complexity of brain connectivity and the need for experimental models allowing a fine manipulation of the local microenvironment at the subcellular level.
In this study, we have grown primary cortical mouse neurons in microfluidic (μFD) devices to separate soma from axonal projections in fluidically isolated microenvironments, and applied β-amyloid (Aβ) peptides locally to the different cellular compartments. We observed that Aβ application to the somato-dendritic compartment triggers a “dying-back” process, involving caspase and NAD+ signalling pathways, whereas exposure of the axonal/distal compartment to Aβ deposits did not induce axonal degeneration. In contrast, co-treatment with somatic sub-toxic glutamate and axonal Aβ peptide triggered axonal degeneration. To study the consequences of such subcellular/local Aβ stress at the network level we developed new μFD multi-chamber devices containing funnel-shaped micro-channels which force unidirectional axon growth and used them to recreate in vitro an oriented cortico-hippocampal pathway. Aβ application to the cortical somato-dendritic chamber leads to a rapid cortical pre-synaptic loss. This happens concomitantly with a post-synaptic hippocampal tau-phosphorylation which could be prevented by the NMDA-receptor antagonist, MK-801, before any sign of axonal and somato-dendritic cortical alteration.
Thanks to μFD-based reconstructed neuronal networks we evaluated the distant effects of local Aβ stress on neuronal subcompartments and networks. Our data indicates that distant neurotransmission modifications actively take part in the early steps of the abnormal mechanisms leading to pathology progression independently of local Aβ production. This offers new tools to decipher mechanisms underlying Braak's staging. Our data suggests that local Aβ can play a role in remote tauopathy by distant disturbance of neurotransmission, providing a putative mechanism underlying the spatiotemporal appearance of pretangles.
KeywordsMicrofluidics Neuronal network Axonal degeneration Tau Synapse loss
The two main hallmarks of Alzheimer's disease (AD) are extracellular deposits composed of β-amyloid peptide (senile plaques) and intracellular filamentous aggregates composed of self-assembled hyperphosphorylated Tau proteins (neurofibrillary tangles, NFTs). Histopathological studies show that these hallmarks spread, each in their own stereotyped fashion, within specific regions of the brain during disease evolution ,. This progression follows neuro-anatomical pathways and could be the sign of nexopathy-related processes . A large body of evidence indicates that neurons affected in AD follow a dying-back pattern of degeneration, where abnormalities in synaptic function and axonal integrity long precede somatic cell death ,. Since neurons are highly polarized, this raises the question whether local Aβ and Tau protein abnormalities in the vicinity of different neuronal subparts (somatic, dendritic, axonal) lead to local degenerative processes or could initiate distant dysfunction within neurons or even within neuronal networks, through (trans-) synaptic alterations. However, though various mechanisms initiating local primary dysfunctions have been described, the molecular or structural reasons underlying distant breakdown of specific networks and the relationship between amyloid and Tau pathology are still unclear.
Recent advances in the development of microfluidic devices (μFD) facilitate microenvironment development adapted to neuronal structures and subdomains with easy access and control ,. We have previously designed and developed a μFD system to polarize neuronal networks using funnel-shaped micro-channels, also called axonal diodes, allowing the reconstruction of an orientated functional cortico-striatal neuronal network in vitro. In the present study, we used a similar μFD-based approach to study the molecular mechanism of neurodegenerative processes in compartmentalized neurons and neuronal networks. By compartmentalizing axon terminals from cell bodies of cortical neurons, we demonstrate that axons are partially resistant to axonal Aβ-peptide exposure, while somatic treatments trigger an anterograde degeneration signal in axons, inducing a dying-back pattern. We then reconstructed an oriented cortico-hippocampal network in μFD, and showed that somato-dendritic deposits of Aβ on cortical neurons trigger a rapid cortical presynaptic disconnection concomitant with a glutamate-dependent postsynaptic hippocampal tau-phosphorylation before any axonal and somatic cortical degeneration.
Somato-dendritic exposure of cortical neurons to Aβ-peptide induces an axonal dying back pattern
Cortical somato-dendritic β-amyloid peptide exposure induces a rapid disconnection of cortico-hippocampal synapses, which precedes axonal degeneration
Selective cortical Aβ peptide exposure induces Tau Thr231 phosporylation in a distant fluidically isolated hippocampal neuron through synaptic NMDA receptor-dependent neurotransmission
A large body of evidence indicates that neurons affected in AD follow a dying-back pattern of degeneration, where such loss of axonal integrity precede somatic cell death and has a profound effect on neuronal network function. However, the molecular mechanism underlying dying-back of neurons and its consequences on the neuronal network in AD remain elusive and difficult to study in vivo. Using a new μFD system , we modeled for the first time the perforant pathway, known to be affected early in AD ,. Somato-dendritic Aβ cortical application within cortico-hippocampal network leads to a rapid presynaptic collapse before cortical axonal or somatic loss ,. Since these synapses were not exposed to Aβ, this suggests that local somato-dendritic Aβ deposits have fast remote toxicity on the unchallenged synapses. This could be due either to a self-propagation  and to rapid distribution of Aβ through axonal transport  or to the induction of a signal in the soma, which is transmitted through the neuron. We recently described similar distant synapto-toxicity following axotomy . Although no short-term morphological alteration of postsynaptic hippocampal neurons was observed, the Aβ-induced remote loss of cortical presynapses was concomitant to Tau Thr231 phosphorylation in the interconnected postsynaptic hippocampal neurons, and occurred well before axonal and somatic degeneration of cortical neurons. Thus local Aβ deposits generate fast propagation of degenerative signals across networks leading to early dysfunctions in remote areas. Our results show that local somato-dendritic (but not axonal) Aβ triggers distal-to-proximal axonal degeneration before any somato-dendritic abnormalities, a process reminiscent of dying-back pattern observed in various neurodegenerative syndromes . Thus Aβ toxicity depends not only on direct contact but also on the location of its subcellular deposition. Axons are relatively resistant to direct Aβ exposure, which is in line with our recent observation showing that axonal endings are resistant to direct pro-apoptotic insults . Our observation with JNK and caspase pharmacological inhibitors suggest that both enzyme families are implicated locally (within axon) in the process of axonal degeneration, as already observed with other neuronal death inducers . We also shows, for the first time, that axonal addition of NAD+ is protective against Aβ-peptide induced axonal degeneration (Figure 2). Since NAD+ is a well-established axo-protectant in the context of Wallerian degeneration, this raises the question whether there might be a NAD+ controlled molecular pathway implicated in both Aβ-induced axon degeneration and Wallerian degeneration. Therefore NAD+ related pathways might offer interesting targets to slow down dying back induced processes.
In some AD patients, increased Aβ is associated with complex disturbances of neuronal activity (e.g. epileptic activity)  and neurotransmission dysfunctions have been described in early phases of AD models ,. Such local circuit disturbance might potentially lead to a broader network disruption in remote areas through neuronal projections. Interestingly, we observed that a mild somatic glutamatergic stress exacerbates distant axonal toxicity of Aβ. This suggests that cumulative and multi-focal stresses might play an important role in disease progression, by switching from local (and probably recoverable) minor dysfunctions to extended neuronal alterations like permanent synaptic, axonal or even cell bodies loss. Several non-exclusive mechanisms may account for AD-related spatiotemporal progression in the brain. This may involve neurotrophic factors withdrawal, aberrant neurotransmission, synaptic loss, excitotoxicity and prion-like spread of pathogenic misfolded proteins like trans-synaptic spread of Aβ . Interestingly, in our paradigm, blocking NMDA receptors on hippocampal neurons during cortical somato-dendritic Aβ treatment prevented Tau Thr231 phosphorylation. These results are consistent with studies reporting that Aβ could potentiate potassium-evoked glutamate release from neurons ,.
While brain complexity, with its interconnected neuronal loops, complicates the in vivo analysis of pathophysiological initiation and spreading mechanisms, we were able for the first time to evaluate the distant effects of local cortical β-amyloid deposits on neuronal subcompartments and networks in μFD-based reconstructed cortico-hippocampal networks. We show that a strictly local somato-dendritic amyloid trigger is sufficient to recapitulate a dying-back process, and to initiate an oriented neuron-to-neuron progression of pathological events. Aβ peptide accumulation in the somato-dendritic compartment of cortical neurons leads to a fast anterograde propagation of degenerative signals toward endings, resulting in presynaptic collapse. This fast loss of cortical presynapses is associated with early trans-synaptic dysfunction such as NMDAR-dependent tau phosphorylation in postsynaptic hippocampal neurons. Hence, reconstructed cortico-hippocampal μFD networks offer a new tool to decipher mechanisms that could underlie dying back and Braak's staging.
Primary culture in microfluidic chips
Microfluidic chips were realized as described in . The design used for network reconstruction comprises two culture chambers each connected to two reservoirs and separated by a series of 500 μM-long asymmetrical micro-channels (3 μM high, tapering from 15 μM to 3 μM) . E16 embryos (Swiss mice, Janvier, France) were micro-dissected in GBSS (without CaCl2 and MgCl2) + 0.1% glucose (Merck), digested with papain (20U/mL) and mechanically dissociated in DMEM containing DNAse. 120.103 cortical cells and 45.103 hippocampal cells were seeded in the chamber on contact with the wide and narrow sides of micro-channels, respectively. Cells were cultured in DMEM glutamax (Invitrogen) supplemented with 10% FBS (PAA), N2, streptomycin/penicillin and B27 (Invitrogen). The culture medium was renewed every 3-5 days. Cortical axon entered the micro-channels and reached the hippocampi-containing chamber in 4-5 days. The cortico-hippocampal oriented network was maintained routinely up to 2-3 weeks in vitro.
Oligomeric and fibrillar Aβ peptide preparations
Oligomeric and fibrillar forms of Aβ1-42 (Tocris Bioscience, MN, USA), were produced according to  and controlled by electron microscopy (Additional file 2: Figure S2). Briefly, lyophilized peptides were solubilized at 1 mM in 1, 1, 1, 3, 3, 3,-hexafluoro-2-propanol (HFIP, Sigma Aldrich). After 30 min of incubation at RT, HFIP was evaporated overnight and peptides were dried (Speed Vac, 1 h 4°C). Then, Aβ peptide stock solution were obtained byresolubilizationet 5 mM in dimethylsulfoxide (DMSO, Sigma Aldrich). To obtain oligomers, Aβ stock solution was diluted at 100 μM in phenol-free DMEM-F12 medium (Life technologies), incubated 24 h at 4°C and centrifuged at 20 000 g (10 min; 4°C) before supernatant (soluble Aβ fraction) collection. To obtain fibrils, Aβ stock solution was diluted at 100 μM in HCl 10 mM and then incubated at 4°C for 24 h.
Aβ sample aliquots (10 μM) were allowed to adsorb onto carbon-coated 200-mesh copper grids (EMS, PA, USA) for 2 minutes before blotting off. Then, grids were incubated for 45 seconds in uranyl acetate 2.5% (w/v) to produce a negatively stained protein loaded grid. Images were recorded on a Zeiss 912 omega electron microscope (Carl Zeiss Group, Oberkochen, Germany).
The following compounds were used: Aβ1-42 peptide (1428, Tocris), Aβ25-35 peptide (H-1192, Bachem), control Aβ35-25 peptide (H2964, Bachem), glutamate (G8415, Sigma), MK-801 (0924, Tocris),5 mM NAD+ (Sigma), 50 μM z-VAD-fmk (R&D Systems, Minneapolis, MN, USA) and 50 μM SP 600125 (Sigma). Oligomeric and fibrillar forms of Aβ1-42 were produced according to  and controlled by electron microscopy. To ensure fluidic isolation between compartments, a hydrostatic pressure difference was generated by over-pressurizing the non-treated chamber as described in .
Cultures were fixed (4% PFA, 20 min RT) and immunostained as described in . Primary antibodies included anti-α-tubulin-FITC (F2168, sigma), anti-MAP2 (M4403, Sigma), anti-Synaptophysin (S5768, Sigma), anti-Vglut1, anti-α-synuclein (4179, cell signaling), anti-pTau Thr231 (44746G, Invitrogen). Species-specific secondary antibodies (coupled to Alexa 350, 488, 555) were used (Invitrogen). Images were acquired with an Axio-observer Z1 (Zeiss) fitted with a cooled CCD camera (CoolsnapHQ2, Ropert Scientific) and images were analyzed using ImageJ.
Quantification of axonal fragmentation
For axonal fragmentation analysis and quantification we used fluorescence microscopy (immuno-detection of tubulin), phase contrast, and picture analysis according to a previously described protocol . Briefly, we used a macro developed in NIH ImageJ software that utilizes the Otsu thresholding algorithm, and a particle analyzer algorithm of ImageJ. The total area of axonal regions with circularity greater than 0.9 was determined and normalized by the total axonal area, which was measured from the thresholded image. This ratio, termed fragmentation index, is an indicator of the average axonal fragmentation level and is used in statistical comparisons. Indices of 0.005, 0.083, and 0.157 correspond to <5%, 50%, and >95% fragmentation, respectively.
Quantification of synapse loss
Synaptic disconnection was assessed through fluorescence microscopy by counting α-synuclein clusters affixed to MAP2 positive hippocampal dendrites. Images were all obtained with the same acquisition parameters. The images were similarly processed with ImageJ software before being used for quantification: the brightness/contrast of all control images was optimized manually to eliminate the background and to maximize the signal. The means of the minimum and maximum intensities were then calculated in the control condition, after which these settings were applied to all images. The images of the three α-synuclein/β-tubulin/MAP2 stainings were merged and the resulting image was used to define the zone where hippocampal dendrites were sufficiently innervated by cortical fibres. α-synuclein/MAP2 merges were then used for quantification.
Quantification of Tau phosphorylation
Tau phosphorylation was assessed by counting the number of neurons presenting pTau levels above a fixed threshold. Images were all obtained using the same acquisition parameters. The images were similarly processed with ImageJ software before being used for quantification: the brightness/contrast of all control images was optimized manually to eliminate the background and to maximize the signal. The means of the minimum and maximum intensities were then calculated in the control condition, after which these settings were applied to all images. All pTau images were then processed with the “Lookup Tables, fire” plugin to visualize the intensity of pTau staining with pseudo-colours. The number of neurons above a fixed colour threshold was then counted and normalized by the total number of neurons to have the percentage of hyperphosphorylated tau neurons.
Differences were assessed by ANOVA, followed, when appropriate, by a post-hoc Bonferoni test. For all analysis * p-value < 0.05; ** p-value < 0.01; *** p-value < 0.001.
This work was supported by “Foundation plan Alzheimer 2010-13”, ANR "PrionSensiTNF: Blan-1312-02" 2011-14 and ANR "Neuroscreen: 2011-RPIB-008-01" 2012-16. B.D. was funded by an MNRT fellowship.
- Braak H, Alafuzoff I, Arzberger T, Kretzschmar H, Del Tredici K: Staging of Alzheimer disease-associated neurofibrillary pathology using paraffin sections and immunocytochemistry. Acta Neuropathol 2006, 112(4):389–404. 10.1007/s00401-006-0127-zView ArticlePubMedPubMed CentralGoogle Scholar
- Thal DR, Rüb U, Orantes M, Braak H: Phases of A beta-deposition in the human brain and its relevance for the development of AD. Neurology 2002, 58(12):1791–1800. 10.1212/WNL.58.12.1791View ArticlePubMedGoogle Scholar
- Warren JD, Rohrer JD, Schott JM, Fox NC, Hardy J, Rossor MN: Molecular nexopathies: a new paradigm of neurodegenerative disease. Trends Neurosci 2013, 36(10):561–569. 10.1016/j.tins.2013.06.007View ArticlePubMedPubMed CentralGoogle Scholar
- Terry RD: Cell death or synaptic loss in Alzheimer disease. J Neuropathol Exp Neurol 2000, 59(12):1118–1119.View ArticlePubMedGoogle Scholar
- Scheff SW, Price DA, Schmitt FA, DeKosky ST, Mufson EJ: Synaptic alterations in CA1 in mild Alzheimer disease and mild cognitive impairment. Neurology 2007, 68(18):1501–1508. 10.1212/01.wnl.0000260698.46517.8fView ArticlePubMedGoogle Scholar
- Magnifico S, Saias L, Deleglise B, Duplus E, Kilinc D, Miquel MC, Viovy JL, Brugg B, Peyrin JM: NAD + acts on mitochondrial SirT3 to prevent axonal caspase activation and axonal degeneration. FASEB J 2013, 27(12):4712–4722. 10.1096/fj.13-229781View ArticlePubMedGoogle Scholar
- Millet LJ, Gillette MU: New perspectives on neuronal development via microfluidic environments. Trends Neurosci 2013, 35(12):752–761. 10.1016/j.tins.2012.09.001View ArticleGoogle Scholar
- Peyrin JM, Deleglise B, Saias L, Vignes M, Gougis P, Magnifico S, Betuing S, Pietri M, Caboche J, Vanhoutte P, Viovy JL, Brugg B: Axon diodes for the reconstruction of oriented neuronal networks in microfluidic chambers. Lab Chi 2011, 11(21):3663–3673. 10.1039/c1lc20014cView ArticleGoogle Scholar
- Mairet-Coello G, Courchet J, Pieraut S, Courchet V, Maximov A, Polleux F: The CAMKK2-AMPK kinase pathway mediates the synaptotoxic effects of Abeta oligomers through Tau phosphorylation. Neuron 2013, 78(1):94–108. 10.1016/j.neuron.2013.02.003View ArticlePubMedPubMed CentralGoogle Scholar
- Luna-Munoz J, Chávez-Macías L, García-Sierra F, Mena R: Earliest stages of tau conformational changes are related to the appearance of a sequence of specific phospho-dependent tau epitopes in Alzheimer's disease. J Alzheimers Dis 2007, 12(4):365–375.PubMedGoogle Scholar
- Jucker M, Walker LC: Self-propagation of pathogenic protein aggregates in neurodegenerative diseases. Nature 2013, 501(7465):45–51. 10.1038/nature12481View ArticlePubMedPubMed CentralGoogle Scholar
- Song HL, Shim S, Kim DH, Won SH, Joo S, Kim S, Jeon NL, Yoon SY: beta-amyloid is transmitted via neuronal connections along axonal membranes. Ann Neurol 2014, 75(1):88–97. 10.1002/ana.24029View ArticlePubMedGoogle Scholar
- Deleglise B, Lassus B, Soubeyre V, Alleaume-Butaux A, Hjorth JJ, Vignes M, Schneider B, Brugg B, Viovy JL, Peyrin JM: Synapto-protective drugs evaluation in reconstructed neuronal network. PLoS One 2013, 8(8):e71103. 10.1371/journal.pone.0071103View ArticlePubMedPubMed CentralGoogle Scholar
- Adalbert R, Coleman MP: Axon pathology in age-related neurodegenerative disorders. Neuropathol Appl Neurobiol 2013, 39(2):90–108. 10.1111/j.1365-2990.2012.01308.xView ArticlePubMedGoogle Scholar
- Palop JJ, Mucke L: Epilepsy and cognitive impairments in Alzheimer disease. Arch Neurol 2009, 66(4):435–440. 10.1001/archneurol.2009.15View ArticlePubMedPubMed CentralGoogle Scholar
- Palop JJ, Chin J, Roberson ED, Wang J, Thwin MT, Bien-Ly N, Yoo J, Ho KO, Yu GQ, Kreitzer A, Finkbeiner S, Noebels JL, Mucke L: Aberrant excitatory neuronal activity and compensatory remodeling of inhibitory hippocampal circuits in mouse models of Alzheimer's disease. Neuron 2007, 55(5):697–711. 10.1016/j.neuron.2007.07.025View ArticlePubMedGoogle Scholar
- Palop JJ, Mucke L: Amyloid-beta-induced neuronal dysfunction in Alzheimer's disease: from synapses toward neural networks. Nat Neurosci 2010, 13(7):812–818. 10.1038/nn.2583View ArticlePubMedPubMed CentralGoogle Scholar
- Kabogo D, Rauw G, Amritraj A, Baker G, Kar S: ss-amyloid-related peptides potentiate K + -evoked glutamate release from adult rat hippocampal slices. Neurobiol Aging 2008, 31(7):1164–1172. 10.1016/j.neurobiolaging.2008.08.009View ArticlePubMedGoogle Scholar
- Liang Z, Liu F, Iqbal K, Grundke-Iqbal I, Gong CX: Dysregulation of tau phosphorylation in mouse brain during excitotoxic damage. J Alzheimers Dis 2009, 17(3):531–539.PubMedPubMed CentralGoogle Scholar
- Stine WB, Dahlgren KN, Krafft GA, LaDu MJ: In vitro characterization of conditions for amyloid-beta peptide oligomerization and fibrillogenesis. J Biol Chem 2003, 278(13):11612–11622. 10.1074/jbc.M210207200View ArticlePubMedGoogle Scholar
- Taylor AM, Blurton-Jones M, Rhee SW, Cribbs DH, Cotman CW, Jeon NL: A microfluidic culture platform for CNS axonal injury, regeneration and transport. Nat Methods 2005, 2(8):599–605. 10.1038/nmeth777View ArticlePubMedPubMed CentralGoogle Scholar
- Kilinc D, Peyrin JM, Soubeyre V, Magnifico S, Saias L, Viovy JL, Brugg B: Wallerian-like degeneration of central neurons after synchronized and geometrically registered mass axotomy in a three-compartmental microfluidic chi. Neurotox Res 2011, 19(1):149–161. 10.1007/s12640-010-9152-8View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.