In vivo induction of membrane damage by β-amyloid peptide oligomers
Exposure to the β-amyloid peptide (Aβ) is toxic to neurons and other cell types, but the mechanism(s) involved are still unresolved. Synthetic Aβ oligomers can induce ion-permeable pores in synthetic membranes, but whether this ability to damage membranes plays a role in the ability of Aβ oligomers to induce tau hyperphosphorylation, or other disease-relevant pathological changes, is unclear. To examine the cellular responses to Aβ exposure independent of possible receptor interactions, we have developed an in vivo C. elegans model that allows us to visualize these cellular responses in living animals. We find that feeding C. elegans E. coli expressing human Aβ induces a membrane repair response similar to that induced by exposure to the CRY5B, a known pore-forming toxin produced by B. thuringensis. This repair response does not occur when C. elegans is exposed to an Aβ Gly37Leu variant, which we have previously shown to be incapable of inducing tau phosphorylation in hippocampal neurons. The repair response is also blocked by loss of calpain function, and is altered by loss-of-function mutations in the C. elegans orthologs of BIN1 and PICALM, well-established risk genes for late onset Alzheimer’s disease. To investigate the role of membrane repair on tau phosphorylation directly, we exposed hippocampal neurons to streptolysin O (SLO), a pore-forming toxin that induces a well-characterized membrane repair response. We find that SLO induces tau hyperphosphorylation, which is blocked by calpain inhibition. Finally, we use a novel biarsenical dye-tagging approach to show that the Gly37Leu substitution interferes with Aβ multimerization and thus the formation of potentially pore-forming oligomers. We propose that Aβ-induced tau hyperphosphorylation may be a downstream consequence of induction of a membrane repair process.
KeywordsAlzheimer’s disease β-amyloid Tau Caenorhabditis elegans Pore-forming toxin
Alzheimer’s disease is characterized by the deposition in the brain of senile plaques, composed largely of the β-amyloid peptide (Aβ). The amyloid cascade hypothesis posits that in Alzheimer’s disease, accumulation of Aβ in the brain ultimately leads to changes in tau metabolism, which leads to the deposition of tau in neurofibrillary tangles (NFTs) that are the likely proximal cause of the neuronal loss observed in this disease . While there is still significant controversy as to whether the amyloid cascade hypothesis explains Alzheimer’s pathology [28, 39, 72], there is extensive evidence (in vitro and in vivo) that exposure of neurons to Aβ can lead to tau hyperphosphorylation, a potential driver of tau deposition [23, 51, 87]. Multiple tau kinases have been identified (reviewed in ), as have pathways by which these kinases could be activated to increase tau phosphorylation [15, 55, 89]. CDK5 and GSK3β have emerged as strong candidates for disease-relevant tau kinases [27, 50], and activation of both these kinases can occur downstream of calpain activation. Calpains are calcium-dependent proteases that can cleave P35 to produce P25, a strong activator of CDK5  or directly truncate GSK3β, leading to its activation . Many studies have also demonstrated that exposure of neurons to Aβ can lead to calcium influx [30, 77, 78] and thus there exists a well-supported molecular mechanism explaining how Aβ exposure can lead to tau hyperphosphorylation. Less clear is the biological rationale for Aβ-induced tau phosphorylation. Is this a selected function of Aβ, an indirect consequence of an evolved interaction of Aβ with cells, or an essentially random interaction with potentially deleterious outcomes?
Addressing “why” Aβ induces tau phosphorylation is complicated by uncertainty as to what, if any, selected function the Aβ peptide has. Similarly unresolved is how extracellular Aβ induces intracellular increases in calcium. Aβ toxicity in numerous models can be attenuated by blocking the NMDA-type glutamate receptor [5, 11, 24], suggesting that this calcium channel could be responsible for Aβ-dependent calcium influx. Similarly, the prion protein has been claimed to be a receptor for Aβ oligomers [7, 44, 74] and to moderate Aβ-dependent calcium influx . Other studies have implicated AMPA glutamate receptors [2, 73, 88] or L-type voltage gated calcium channels  in cytoplasmic calcium increases resulting from Aβ exposure. While multiple calcium channels could be involved in Aβ-dependent calcium influx, other studies suggest that Aβ oligomers could act independently of endogenous calcium channels by directly forming a calcium-permeable pore. It has long been known that synthetic Aβ can form ion-permeable channels in synthetic membranes , a finding that has often been replicated [29, 41, 71]. Although ring-like structures of Aβ oligomers have been visualized by atomic force microscopy in synthetic membranes , Aβ membrane pores in pathologically-relevant cells have not been directly visualized or assayed.
One approach to determining if Aβ pores are relevant to Aβ neurotoxicity is to identify substitutions in the Aβ sequence that block pore formation in synthetic membranes, and then determine if these substitutions alter Aβ toxicity. Kim et al.  noted a sequence motif (Gly-XXX-Gly-XXX-Gly) present in both the C terminus of Aβ and the transmembrane domains of bacterial channel proteins, and they proposed that this motif (the “glycine zipper”) could facilitate α-helical interactions driving Aβ oligomer assembly in membranes and subsequent pore formation. These researchers demonstrated that leucine substitutions at these glycine residues (particularly Gly37Leu) prevented synthetic Aβ 1–42 from forming ion channels in synthetic membranes and also significantly reduced Aβ toxicity in Neuro 2A cells. Fonte et al.  extended these results and demonstrated that the Gly37Leu substitution both reduced Aβ toxicity in an in vivo transgenic C. elegans model that expresses human Aβ and prevented synthetic Aβ oligomers from inducing tau hyperphosphorylation in hippocampal neurons. Peters et al.  subsequently demonstrated that a pentapeptide derived from the glycine zipper sequence (GLMVG) inhibited Aβ synaptotoxicity. These observations are consistent with pore formation underlying Aβ neurotoxicity, but they do not directly demonstrate pore formation, and they cannot exclude other interpretations, such as the possibility that the glycine zipper substitutions interfere with interactions with specific cell surface receptors.
Transgenic C. elegans strains have been constructed that express human Aβ42 [16, 17, 18, 47, 82], and these strains have a variety of phenotypes, depending on where the transgene is expressed. A confound of these models is that the detectable Aβ is intracellular when assayed by immunohistochemistry, so the degree of outside-in Aβ toxicity (i.e., extracellular Aβ affecting neighboring cells) is unclear. To circumvent this limitation, we have developed a “feeding” model, where C. elegans is fed E. coli engineered to secrete human Aβ, and the cellular effects of this exogenous peptide are assayed in intestinal cells. One rationale for this model is that the C. elegans intestine does not express candidate Aβ receptors (e.g., prion protein, NMDA glutamate receptors, α7nAChR, etc.), and thus any physiological response to Aβ is unlikely to be receptor-mediated. The transparency of C. elegans and the existence of relevant transgenic reporter strains allows the effects of Aβ exposure to be followed in live intact animals. Using this model, we show that exposure to wild type Aβ42 (but not Aβ42 Gly37Leu) induces acid-sphingomyelinase-dependent endocytosis that parallels the response to a known pore-forming toxin, CRY5B. We find that this response to Aβ is calpain-dependent and is altered by loss-of-function mutations in the C. elegans orthologs of BIN1 and PICALM, two Alzheimer risk genes identified in genome-wide association studies [61, 84]. In hippocampal cultures, we show that the SLO pore-forming toxin induces calpain-dependent tau phosphorylation in primary neurons. Furthermore, exogenous sphingomyelinase itself can induce increased tau phosphorylation in these neurons. Finally, we use a novel tagging method to show that the Gly37Leu substitution does inhibit Aβ multimerization in a cellular context, thereby rationalizing why this Aβ variant is incapable of inducing membrane repair or tau phosphorylation. Taken together, these results support the view that tau hyperphosphorylation may be a downstream consequence of a membrane repair process, and that exogenous Aβ can induce membrane repair because of its ability to oligomerize into membrane pores.
Materials and methods
C. elegans strains and maintenance
C. elegans strains were generally maintained at 20 °C using standard methods . Strains containing the rynEx60 transgene were passaged at 25 °C to select for retention of this pha-1 rescuing extrachromosomal transgene.
C. elegans strains used in this study.
Construction of E. coli feeding strains
To engineer E. coli strains capable of secreting human Aβ, plasmids were constructed by inserting the wild type or Gly37Leu Aβ sequence between the NcoI and SacI sites of the arabinose-inducible pBAD gIII vector (Invitrogen), yielding pCL241 and pCL242, respectively. These plasmids were transformed into E. coli strain LMG194, allowing tight control of arabinose induction.
Feeding protocol and endosome scoring in C. elegans
Overnight cultures of E. coli strains were diluted 1:100 in Luria broth (+ 100 μg/mL ampicillin) and grown for 2 h in an orbital shaker at 37 °C. Twenty percent arabinose was added to final concentration of 0.2%, and the cultures were grown for another hour. These E. coli cultures were then spread on NGM plates (nematode growth media, ) containing 0.2% arabinose and 100 μg/mL ampicillin and used the next day. L4 worms were transferred onto the fresh plates and allowed to feed for 4 h at 25 °C. Worms were imaged with a Zeiss Axiophot microscope (40X objective) and endosomes were counted in the anterior 50 μm of the intestine.
Hippocampal neuronal culture
Different protocols were used for the preparation of hippocampal neurons in the Stein and Silverman labs. Results obtained in replicate experiments using the two different protocols were highly reproducible.
(Stein lab) Hippocampal neurons were isolated from rat E18 hippocampus (BrainBits, LLC) according to the supplier’s protocol. The neurons were plated at 10–20,000 cells/cm2 on Lab-Tek CC2 4-well chamber slides coated with poly-D-lysine (Sigma P0899) and grown at 37 o C. The medium, Neurobasal + 2% B27 and 0.5 mM GlutaMax (Invitrogen), was supplemented with 25 mM glutamate for the first 4 days; thereafter the neurons were fed twice a week with medium lacking glutamate and were used for experiments after 9–16 days in vitro (DIV).
(Silverman lab) Primary hippocampal neuronal cultures from E18 embryonic rats (Charles River, USA) of either sex were prepared as described by Kaech and Banker  and kept in PNGM primary neuron growth media (Lonza, Basel, Switzerland). The glial feeder layer was derived from murine neural stem cells as described by . All experiments with animals were approved by and followed the guidelines set out by the Simon Fraser University Animal Care Committee, Protocol 943-B05.
SLO activation and treatment of cells
SLO (Aalto Bio Reagents), a thiol-activated pore-forming toxin that loses activity upon storage at − 700 C, was reactivated before each experiment by incubating an aliquot of 3 mg/ml SLO (in Tris-buffered saline, pH 8.5) with an equal volume of 20 mM DTT for 10–15 min at room temperature. Activated SLO and vehicle were diluted with medium and added directly to the cells as 6X concentrates, or half the old medium on the cells was removed and replaced by a 2X concentration of SLO or its vehicle control. The toxicity of each aliquot of activated SLO was evaluated by exposing rat insulinoma cells (RIN5F in RPMI 1640 with 10% Hyclone FBS) to 0–2000 ng/ml SLO for 2 h and measuring viability by the reduction of MTT (Sigma M5655). The results indicated that SLO had little or no gross toxicity at < 200 ng/ml and killed half the cells at 600–800 ng/ml. We treated neurons with 50–100 ng/ml SLO for 1–2 h.
Treatment with calpain inhibitor (PD 150606)
PD150606 (Sigma D5946) was dissolved in DMSO, aliquoted and stored at -20o C. It was added to cells at a final concentration of 30 μg.
Treatment with sphingomyelinase
Ten units of lyophilized B. cereus sphingomyelinase (SMase, Sigma S7651) were dissolved in 200 μl of cold 50% glycerol in sterile Ca2+/Mg2+ free PBS, yielding 50 mU/μl. Because SMase loses activity even when stored at − 700 C, it was best used fresh. Neurons were treated with vehicle control and SMase diluted to 2.5 mU/ml in Neurobasal/B27 for 2 h.
Antibodies for immunostaining neurons or C. elegans
Primary antibodies used to stain neurons were mouse monoclonal PHF1 to p-Ser396/p-Ser404 tau (kind gift from Dr. Peter Davies), mouse monoclonal AT8 to p-Ser202/p-Thr205 tau (Pierce, Thermo) and rabbit polyclonal K9JA to total tau (Dako A0024). Secondary antibodies (goat anti-rabbit IgG Alexa Fluor 488, goat anti-mouse Alexa Fluor 594 and donkey anti-mouse IgG Alexa Fluor 555) and Prolong Gold Antifade mounting medium were from Invitrogen. The biarsenical dye used was from the TC-FlAsH™ II In-Cell Tetracysteine Tag Detection Kit (Molecular Probes, Eugene, OR). The anti-RME-1 antibody was obtained from the Developmental Studies Hybridoma Bank, and anti-Aβ mouse monoclonal 6E10 was purchased from BioLegend.
Neurons grown on chamber slides (Stein lab) were fixed for 15 min in 4% p-formaldehyde/4% sucrose, permeabilized 7 min in 0.25% Triton X-100, and blocked 1–2 h with 5% goat or sheep serum (all solutions in PBS). Slides were stained with primary antibodies (K9JA at 1:500, and AT8 or PHF1 at 1:50), secondary antibodies at 1:500, and mounted with ProLong Gold AntiFade. Images were obtained by using a defined scanning pattern to view the total tau fluorescence in sequential fields in each well, capturing images of fields with an average number of cells (typically 5–20 cells per field) using a 10X or 20X objective on a Zeiss Axioskop epifluorescence microscope equipped with 3i Slidebook image analysis software. Identical conditions and exposure times were used to capture an average of 18.5 fields of cells for each treatment presented in our figures, and the ratio of p-tau to t-tau was determined for each image. The mean for each treatment was calculated and compared to its control, and Student’s t-test was used to determine if the difference of the means was significant. Similarly, neurons grown on coverslips (Silverman lab) were fixed in 4% paraformaldehyde for 15 min, blocked in 0.5% fish skin gelatin and 0.1%Triton X-100 in PBS, and immunostained with K9JA and AT8 or PHF1. In this case, to quantify tau phosphorylation, histograms were generated using ImageJ from the fluorescence intensity of each pixel across several images, and the average intensity was calculated . Appropriate thresholds were applied to eliminate background signal before histogram analysis, and 17 images per experimental condition were analyzed from at least three independent neuronal cultures.
Preparation of Aβ oligomers
Aβ oligomers were prepared from synthetic Aβ (BioSynthesis) using the “ADDL” preparation originally described by Lambert et al. , and subsequently used by us to characterize wild type and Gly37Leu oligomers . Briefly, peptides were solubilized in hexafluoroisopropanol (HFIP) and desiccated in microfuge tubes, then dissolved in fresh, anhydrous DMSO (Sigma Hybri-Max D-2650) to make a ~ 5 mM solution. This solution was then diluted to ~ 100 μM with cold F12 media without phenol red (Biosource) and aged 24 h at 4 °C. The samples were centrifuged at 14,000 g for 10 min at 4 °C to remove any insoluble material, and the supernatants stored at 4 °C.
Neurons were treated with vehicle, SMase, SLO, and PD 150606 as described above and lysed in RIPA buffer containing Complete Protease Inhibitor Cocktail (Roche) and Halt Phosphatase Inhibitor Cocktail (Thermo Fisher). Samples (10 μg) were resolved on 10% SDS–polyacrylamide gels and transferred to polyvinylidene fluoride (PVDF) membranes. Membranes were incubated with the following primary antibodies overnight at 4 °C: PHF-1 (1:1000), AT8 (1:250), K9JA (1:2000), and anti–tubulin (1:1000). The membranes were imaged using Fujifilm LAS4000 Luminescent Imager. Densitometric scanning and quantitative analysis were carried out using ImageJ.
For assaying Aβ expression in engineered E. coli strains, bacterial cultures were spun down in a tabletop centrifuge (3 min at 10,000 rpm), and pellets were frozen at − 80 °C until use. Pellets were solubilized in RIPA buffer supplemented with AEBSF (Sigma P2714) and quantitated using Pierce BCA Protein Assay (Thermo Fisher 23,225). Protein samples were boiled in sample buffer (4% BME in NuPAGE LDS sample buffer (Invitrogen NP0007) and run at 180 V on NuPAGE 4–12%Bis-Tris Gels (Invitrogen, NP0321) using MES SDS Running Buffer (Invitrogen NP0002). Gels were transferred to 0.45 μm supported nitrocellulose (GE Osmonics WP4HY00010) using 20% methanol, 39 mM glycine, 48 mM Tris base at 21 V for 108 min. Prestained Rainbow size markers (Amersham Biosciences RPN755) were used to size bands. Blots were visualized by Ponceau stain, boiled for 3 min in PBS, blocked in TBS-Tween + 5% milk (100 mM Tris7.5, 150 mM NaCl, 0.1% Tween-20) and probed with Aβ antibody 6E10 (BioLegend 803,002) at a 1:1000 dilution. Secondary HRP-conjugated antibody (Sigma A5906 mouse) was used, and the blot was developed in ECL Plus (Amersham RPN2132).
Treating hippocampal neurons with Cys-tagged Aβ peptides
Rat hippocampal neurons (cultured in the Stein lab) were treated with wild-type and Cys-tagged Aβ peptides (synthesized by Bio Synthesis) and exposed to the FlAsH dye according to manufacturer’s instructions with modifications [TC-FlAsH™ II In-Cell Tetracysteine Tag Detection Kit (Molecular Probes)]. Briefly, rat hippocampal neurons were treated to 10 μM EDTA to suppress background fluorescence for 10 min, and were then exposed to 2.5 μM Cys-tagged Aβ peptides or wild-type Aβ peptides for 1 h at 37 °C. FlAsH dye was then applied to neurons at 1:800 for 30 min. BAL wash buffer was used to remove the excess of the FlAsH dye and was replaced by warm HBSS after 15 min. Neurons were then immediately fixed and prepared for immunohistochemistry.
A Zeiss Axiophot microscope equipped with digital deconvolution optics and a Nikon Structured Illumination Super-resolution (Light Microscopy Core Facility in Department of Molecular, Cellular and Developmental Biology, University of Colorado Boulder) were used for imaging C. elegans and rat hippocampal neurons. Images generated in the Silverman lab were acquired using a Leica DMI6000B inverted epifluorescence microscope using a 63 X 1.4 N.A. oil-immersion objective equipped with a cooled CCD camera controlled by MetaMorph (Molecular Devices).
A one-tailed paired two sample for mean t-test, was used to determine significance between pair wise comparisons of control and experimental conditions. Log rank statistics were used to analyze survival curves in the CRY5B exposure experiments (Fig. 5b). Significant differences between the treatments were analyzed by t-tests with equal or unequal variance at a 95% confidence interval. The experiments were carried out in three different cultures and at least 60–70 cells were analyzed per condition. For the blots, cell lysates were collected from at least 3 cultures.
Aβ induces a membrane repair process in C. elegans
pha-1(e2123) III; him-5(e1490) V; rnyEX60 (vha-6::mCherry, Pmyo-3:: GFP, pha-1 rescuing fragment)
unc-119(ed3) III; dkIs166(Popt-2::GFP::pgp-1; unc-119 rescuing fragment)
asm-1(tm5267) II; pha-1(e2123) III; him-5(e1490) V; rnyEX60
pha-1(e2123) III(?); amph-1(ok3443); him-5(e1490) V (?); rnyEX60
unc-11(e47) I; pha-1(e2123) III (?); him-5(e1490) V (?); rnyEX60
clp-4(ok2808) III; him-5(e1490) V (?); rnyEX60
unc-119(ed3) III; pwIS72[Pvha-6::GFP::rab-5, unc-119 rescuing fragment)
To examine the epistatic relationship between these genes, we repeated the Aβ feeding assay after knocking down amph-1 or unc-11 by RNAi in the reporter strain, with and without the clp-4 mutation. We found that the increased endocytosis induced by amph-1 RNAi was completely blocked by the clp-4 loss-of-function mutation, and the increased endocytosis induced by unc-11 RNAi was somewhat reduced (Fig. 4c). Thus, clp-4 is fully epistatic to amph-1 and may moderate unc-11 in the membrane repair process.
Induction of membrane repair increases tau phosphorylation in hippocampal neurons
Visualization of toxic Aβ oligomers using fluorescent biarsenical labeling
We synthesized both wild type and Gly37Leu Aβ 1–42 tagged with two cysteine residues at the C-terminus. The addition of the C-terminal cysteines did not prevent wild type Aβ from inducing tau phosphorylation in hippocampal neurons (Additional file 1: Figure S3 A). Furthermore, E. coli engineered to express the two-Cys version of wild type, but not Gly37Leu, Aβ, induced endocytosis in the C. elegans model (Additional file 1: Figure S3B). We exposed hippocampal neurons to these Cys-tagged peptides using the same protocols we used previously to assay Aβ induction of tau. After 1 h exposure to Aβ oligomer preps, the cultured neurons were stained with 1:800 FlAsH dye for 30 min, then fixed, permeabilized and immunostained for Aβ (mAb 6E10) (see Methods). When imaged by super-resolution microscopy, neurons exposed to (untagged) Aβ 1–42 had foci of Aβ immuno-reactivity associated with neuronal processes, but as expected these rarely overlapped with the FlAsH reagent signal (Fig. 8b). In contrast, a number of Aβ-positive foci co-stained with the FlAsH reagent when the neurons were exposed to Cys-tagged Aβ 1–42 (Fig. 8c). Importantly, these co-staining-foci were reduced to the level of the untagged Aβ control in neurons exposed to Cys-tagged Aβ Gly37Leu (Fig. 8d, quantified in Fig. 8e). We interpret this result to indicate that the Gly37Leu substitution interferes with the Aβ multimerization required to interact with the FlAsH reagent.
We have developed a novel C. elegans model to investigate the cellular response to exogenous Aβ, and we observe an induction of endocytosis similar to that associated with membrane repair in response to exposure to the CRY5B pore-forming toxin [31, 38, 52, 85]. The observations that Aβ induction of endocytosis is blocked by a mutation in a gene encoding acid sphingomyelinase and the apparent membrane association of Aβ in induced endosomes also support the view that exogenous Aβ induces a membrane repair process as described in Fig. 1. To link membrane damage/repair with a classic marker of AD pathology, we also show that hippocampal neurons exposed to either a pore-forming toxin (streptolysin O) or exogenous sphingomyelinase display increased tau hyperphosphorylation comparable to that seen after exposure to Aβ. In addition, the increase in endosomes induced by either Aβ or CRY5B is inhibited by a loss-of-function mutation in clp-4 and enhanced by mutations in amph-1 or unc-11, which are orthologs of human LOAD genes BIN1 and PICALM, respectively. These results further tie the membrane damage/repair process to Alzheimer’s disease. Despite these similarities between the effects of Aβ-and CRY5B, feeding wild type worms Aβ-expressing E. coli does not result in the lethality observed in worms fed CRY5B. We attribute this difference to CRY5B being an evolved “professional” pore-forming toxin, assembled by a large protein that produces a pore size likely to be significantly larger  than a pore formed by Aβ oligomers. At present we cannot determine if the effects of Aβ in this model are due solely to the direct interaction of the peptide with the intestinal membrane, or whether proteins in the lumenal membrane are playing a role. We can conclude that suggested Aβ receptors (e.g., prion protein, NMDA glutamate receptors, α7nAChR) are highly unlikely to be involved, as there is no evidence these proteins are expressed in the C. elegans intestine.
The ability of the clp-4 deletion to prevent endocytosis induced by Aβ or CRY5B implicates calpain activity in the membrane repair process. Calpains are known to target a number of cytoskeletal proteins , and to moderate cytoskeletal rearrangement, for example in focal adhesion remodeling . Interestingly, in macrophages, calpain 2 cleavage of talin 1 is critical for endocytosis of the pore-forming “protective antigen” toxin produced by Bacillus anthracis . We envision calpain activation as a conserved process required to induce local cytoskeletal changes that promote endocytosis in response to membrane damage.
The increased accumulation of Aβ-induced intestinal endosomes in the amph-1 and unc-11 mutants could result from either increased endosome formation or reduced endosome disassembly (or both). Loss of amph-1 alters the distribution of the early endosome marker RAB-5 and leads to increases in the size of steady-state endosomes in the C. elegans intestine . We note that an increase in the number and size of RAB-5 endosomes in AD brains has long been recognized . BIN1, the ortholog of amph-1, has been implicated in the regulation of endocytosis in multiple contexts [65, 83, 86], even though its specific role in Alzheimer’s disease is unclear. The rs59335482 risk allele, a 3 bp insertion upstream of the BIN1 coding sequence, increases BIN1 transcription and is correlated with tau but not β-amyloid loads in AD brains . This study also reported that loss of Amph (the Drosophila ortholog of BIN1) suppressed tau pathology, but not Aβ pathology, in transgenic fly models based on ectopic expression of these AD-associated proteins in the fly eye. However, studies in mammalian neurons have indicated that loss of BIN1 promoted the propagation of tau pathology . Our studies cannot resolve the discrepancies between these reports, but they do establish that amph-1 in C. elegans (and potentially BIN1 in mammals) play a role in the membrane repair response to pore-forming toxins.
unc-11 encodes a C. elegans clathrin adaptor protein orthologous to AP180 and PICALM, two closely related mammalian proteins. The unc-11 gene is expressed at high levels in neurons and at lower levels in other tissues, including the intestine. Interestingly, null alleles of unc-11 do not prevent endocytosis of synaptic membrane, but do result in an increased size of synaptic endosomes . Similarly, RNAi knockdown of PICALM in mammalian neurons leads to synaptic vesicles with variably increased size , suggesting that these proteins primarily regulate the sorting and recycling of endosomes rather than endocytosis per se. Thus, the increase of endosome number (and size) in the intestines of unc-11 mutants exposed to CRY5B or Aβ likely results from altered endosome sorting or disassembly rather than upregulated endocytosis. Most studies support the view that PICALM expression is protective in AD, as full-length PICALM has been reported to be reduced in AD brain , and the protective allele of the primary AD-related SNP (rs3851179) is reported to increase PICALM mRNA levels in the brain .
The involvement of BIN1 and PICALM in endocytic processes has led multiple research groups to investigate their possible role in the production of Aβ, which largely occurs during the endocytic recycling of Amyloid Precursor Protein (APP) . Indeed, loss of BIN1 has been reported to increase the production of Aβ [3, 60], whereas knockdown of PICALM is reported to reduce Aβ production [37, 80]. If, as described above, the risk alleles of BIN1 and PICALM lead to increased and decreased expression of these two genes, respectively, then the reported effects of these genes on Aβ production are actually the opposite of what might be expected. Our studies explore an alternative (but not mutually exclusive) explanation for the association of these genes with AD risk: a role in the cellular response to extracellular Aβ. We find that loss of function mutations in amph-1 or unc-11 sensitize worms to the CRY5B pore-forming toxin (Fig. 5b), and lead to similar dysregulation of the endocytosis induced by CRY5B or Aβ. These results are consistent with BIN1 and PICALM risk alleles acting by modulating cellular responses to extracellular Aβ.
The ability of asm-1 mutations to suppress Aβ- or CRY5B-induced endocytosis supports the view that Aβ can instigate a membrane repair process analogous to that occurring in mammalian cells challenged by a pore-forming toxin such as SLO . We reasoned that if Aβ acts as a pore-forming toxin in the AD brain, known effects of Aβ exposure such as tau hyperphosphorylation might be replicated using pore-forming toxins. We find that exposure of rat hippocampal neurons to SLO results in calpain-dependent tau hyperphosphorylation at two epitopes classically associated with AD pathology. Although we cannot exclude the possibility that SLO mimics Aβ-induced tau phosphorylation by an unrelated pathway, the ability of exogenous sphingomeylinase to induce tau hyperphosphorylation supports the view that tau hyperphosphorylation can be a downstream consequence of membrane damage/repair. Attempts to induce endocytosis in C. elegans intestines by feeding worms B. cereus sphingomeylinase were unsuccessful, possibly due to reduced activity of the enzyme at the lower temperature required for worm maintenance (20 °C for C. elegans vs. 37 °C for cultured neurons) and/or degradation of the enzyme in the intestinal lumen. Our experiments examining the effects of exogenous sphingomeylinase on tau hyperphophorylation also cannot determine if sphingomeylin-based secondary messengers are playing a role. Similarly, determining whether tau phosphorylation is a functional component of membrane repair or simply an incidental consequence will require additional investigation.
A major observation supporting the possible relevance of toxic Aβ pores is the dramatically reduced toxicity of the Aβ Gly37Leu variant in both transgenic C. elegans models and mammalian neurons . This variant, which unlike wild type Aβ cannot induce ion-permeable channels in synthetic membranes, was investigated by the Bowie lab based on modeling studies that suggested it could not assemble pore-forming oligomers due to interference with a “glycine zipper” motif . However, it had not been demonstrated previously that the Gly substitution in this critical variant actually alters Aβ multimerization, and in fact the Gly37Leu substitution does not reduce the stable oligomer species assayed by SDS-PAGE (see Additional file 1: Figure S1C). We therefore sought an approach to assay Aβ multimerization in vivo that could capture potentially less stable, membrane-associated oligomers. Using hippocampal neurons exposed to dicysteine-tagged synthetic Aβ and detection of closely associated pairs of dicysteine tags by means of a membrane-permeant biarsenical dye, we provide evidence that the Gly37Leu substitution does indeed inhibit Aβ multimerization in a cellular context. We suggest that the Aβ Gly37Leu variant can be used as a control peptide for investigating Aβ oligomer effects and could be superior to typical Aβ sequence-scrambled peptides used for this purpose.
More than 20 years ago, the discovery by Arispe and colleagues that Aβ could form ion channels in synthetic membranes led them to propose that Aβ membrane pores could contribute to Alzheimer pathology . While multiple studies since have been supportive , this hypothesis has neither been directly supported (e.g., by the visualization of Aβ pores in AD pathological tissue), nor convincingly discounted. The failure to visualize Aβ pores in AD brains may not be surprising given the rapidity with which membrane pores are removed in mammalian cells (< 1 min in HEK293 cells; ), whereas the consequences of the repair process, such as hyperphosphorylation of tau, might accumulate. Our studies cannot directly prove the disease relevance of the pore-forming capacity of Aβ, but they do suggest a possible re-interpretation of existing human data. In particular, the demonstration of tau hyperphosphorylation as a downstream consequence of membrane damage/repair suggests mechanisms that may also be in play in non-AD tauopathies such as Chronic Traumatic Encephalopathy (CTE) and Niemann Pick type C disease . Multiple explanations have been advanced to explain the dramatically increased AD risk for ApoE4 allele carriers (reviewed in ); our results support the idea that this risk could be due to altered membrane repair resulting from the reduced cholesterol and phospholipid secretion observed in APOE4 glia and neurons .
This work was supported by US National Institutes of Health award R21 AG049693 to CDL and Natural Science and Engineering Research Council of Canada (327100) and the Canadian Institutes of Health Research (90396) awards to MAS. We would like to thank Andrew Gaines and Josh Morgenstern for assistance with pilot experiments. Some nematode strains were provided by the Caenorhabditis Genetics Center, funded by the NIH National Center for Research Resources. We would also like to thank Dr. Raffi Aroian for providing the CRY5B expression vector.
CJ designed and performed both C. elegans and hippocampal neuron experiments, and contributed to the writing of this manuscript. CT performed C. elegans experiments, including the CRY5B survival studies. CMR constructed and characterized the E. coli strains expressing human Aβ. AA performed hippocampal neuron experiments, including tau immunoblots. GHS designed and performed hippocampal neuron experiments and contributed to the writing of this manuscript. MAS and CDL contributed to experimental design and the writing of the manuscript. All authors have read and approved of this manuscript.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 2.Alberdi E, Sanchez-Gomez MV, Cavaliere F, Perez-Samartin A, Zugaza JL, Trullas R, Domercq M, Matute C (2010) Amyloid beta oligomers induce Ca2+ dysregulation and neuronal death through activation of ionotropic glutamate receptors. Cell Calcium 47:264–272. https://doi.org/10.1016/j.ceca.2009.12.010 CrossRefPubMedGoogle Scholar
- 3.Ando K, Brion JP, Stygelbout V, Suain V, Authelet M, Dedecker R, Chanut A, Lacor P, Lavaur J, Sazdovitch V et al (2013) Clathrin adaptor CALM/PICALM is associated with neurofibrillary tangles and is cleaved in Alzheimer’s brains. Acta Neuropathol 125:861–878. https://doi.org/10.1007/s00401-013-1111-z CrossRefPubMedGoogle Scholar
- 5.Arbel-Ornath M, Hudry E, Boivin JR, Hashimoto T, Takeda S, Kuchibhotla KV, Hou S, Lattarulo CR, Belcher AM, Shakerdge N et al (2017) Soluble oligomeric amyloid-beta induces calcium dyshomeostasis that precedes synapse loss in the living mouse brain. Mol Neurodegener 12:27. https://doi.org/10.1186/s13024-017-0169-9 CrossRefPubMedPubMedCentralGoogle Scholar
- 10.Carrasquillo MM, Belbin O, Hunter TA, Ma L, Bisceglio GD, Zou F, Crook JE, Pankratz VS, Sando SB, Aasly JO et al (2011) Replication of BIN1 association with Alzheimer’s disease and evaluation of genetic interactions. J Alzheimers Dis 24:751–758. https://doi.org/10.3233/JAD-2011-101932 CrossRefPubMedPubMedCentralGoogle Scholar
- 13.Chapuis J, Hansmannel F, Gistelinck M, Mounier A, Van Cauwenberghe C, Kolen KV, Geller F, Sottejeau Y, Harold D, Dourlen P et al (2013) Increased expression of BIN1 mediates Alzheimer genetic risk by modulating tau pathology. Mol Psychiatry 18:1225–1234. https://doi.org/10.1038/mp.2013.1 CrossRefPubMedPubMedCentralGoogle Scholar
- 14.De Felice FG, Wu D, Lambert MP, Fernandez SJ, Velasco PT, Lacor PN, Bigio EH, Jerecic J, Acton PJ, Shughrue PJ et al (2008) Alzheimer’s disease-type neuronal tau hyperphosphorylation induced by A beta oligomers. Neurobiol Aging 29:1334–1347. https://doi.org/10.1016/j.neurobiolaging.2007.02.029 CrossRefPubMedGoogle Scholar
- 15.Deng J, Habib A, Obregon DF, Barger SW, Giunta B, Wang YJ, Hou H, Sawmiller D, Tan J (2015) Soluble amyloid precursor protein alpha inhibits tau phosphorylation through modulation of GSK3beta signaling pathway. J Neurochem 135:630–637. https://doi.org/10.1111/jnc.13351 CrossRefPubMedPubMedCentralGoogle Scholar
- 16.Diomede L, Di Fede G, Romeo M, Bagnati R, Ghidoni R, Fiordaliso F, Salio M, Rossi A, Catania M, Paterlini A et al (2014) Expression of A2V-mutated Abeta in Caenorhabditis elegans results in oligomer formation and toxicity. Neurobiol Dis 62:521–532. https://doi.org/10.1016/j.nbd.2013.10.024 CrossRefPubMedPubMedCentralGoogle Scholar
- 18.Duan Z, Sesti F (2013) A Caenorhabditis elegans model system for amylopathy study. J Vis Exp:e50435. https://doi.org/10.3791/50435
- 20.Fonte V, Dostal V, Roberts CM, Gonzales P, Lacor PN, Velasco PT, Magrane J, Dingwell N, Fan EY, Silverman MA et al (2011) A glycine zipper motif mediates the formation of toxic beta-amyloid oligomers in vitro and in vivo. Mol Neurodegener 6:61. https://doi.org/10.1186/1750-1326-6-61 CrossRefPubMedPubMedCentralGoogle Scholar
- 24.Guivernau B, Bonet J, Valls-Comamala V, Bosch-Morato M, Godoy JA, Inestrosa NC, Peralvarez-Marin A, Fernandez-Busquets X, Andreu D, Oliva B et al (2016) Amyloid-beta Peptide Nitrotyrosination Stabilizes Oligomers and Enhances NMDAR-Mediated Toxicity. J Neurosci 36:11693–11703. https://doi.org/10.1523/JNEUROSCI.1081-16.2016 CrossRefPubMedGoogle Scholar
- 26.Hassan WM, Dostal V, Huemann BN, Yerg JE, Link CD (2015) Identifying Abeta-specific pathogenic mechanisms using a nematode model of Alzheimer's disease. Neurobiol Aging 36:857–866. https://doi.org/10.1016/j.neurobiolaging.2014.10.016 CrossRefPubMedGoogle Scholar
- 36.Kagan BL (2012) Membrane pores in the pathogenesis of neurodegenerative disease. Prog Mol Biol Transl Sci 107:295–325. https://doi.org/10.1016/B978-0-12-385883-2.00001-1 CrossRefPubMedGoogle Scholar
- 42.Kurbatskaya K, Phillips EC, Croft CL, Dentoni G, Hughes MM, Wade MA, Al-Sarraj S, Troakes C, O’Neill MJ, Perez-Nievas BG et al (2016) Upregulation of calpain activity precedes tau phosphorylation and loss of synaptic proteins in Alzheimer’s disease brain. Acta Neuropathol Commun 4:34. https://doi.org/10.1186/s40478-016-0299-2 CrossRefPubMedPubMedCentralGoogle Scholar
- 52.Los FC, Kao CY, Smitham J, McDonald KL, Ha C, Peixoto CA, Aroian RV (2011) RAB-5- and RAB-11-dependent vesicle-trafficking pathways are required for plasma membrane repair after attack by bacterial pore-forming toxin. Cell Host Microbe 9:147–157. https://doi.org/10.1016/j.chom.2011.01.005 CrossRefPubMedPubMedCentralGoogle Scholar
- 56.Manucat-Tan NB, Saadipour K, Wang YJ, Bobrovskaya L, Zhou XF (2018) Cellular trafficking of amyloid precursor protein in Amyloidogenesis physiological and pathological significance. Mol Neurobiol. https://doi.org/10.1007/s12035-018-1106-9
- 58.Meng L, Zhang A, Jin Y, Yan D (2016) Regulation of neuronal axon specification by glia-neuron gap junctions in C. elegans. Elife 5. https://doi.org/10.7554/eLife.19510
- 59.Miranda CJ, Braun L, Jiang Y, Hester ME, Zhang L, Riolo M, Wang H, Rao M, Altura RA, Kaspar BK (2012) Aging brain microenvironment decreases hippocampal neurogenesis through Wnt-mediated survivin signaling. Aging Cell 11:542–552. https://doi.org/10.1111/j.1474-9726.2012.00816.x CrossRefPubMedPubMedCentralGoogle Scholar
- 61.Naj AC, Jun G, Reitz C, Kunkle BW, Perry W, Park YS, Beecham GW, Rajbhandary RA, Hamilton-Nelson KL, Wang LS et al (2014) Effects of multiple genetic loci on age at onset in late-onset Alzheimer disease: a genome-wide association study. JAMA Neurol 71:1394–1404. https://doi.org/10.1001/jamaneurol.2014.1491 CrossRefPubMedPubMedCentralGoogle Scholar
- 62.Nikkel AL, Martino B, Markosyan S, Brederson JD, Medeiros R, Moeller A, Bitner RS (2012) The novel calpain inhibitor A-705253 prevents stress-induced tau hyperphosphorylation in vitro and in vivo. Neuropharmacology 63:606–612. https://doi.org/10.1016/j.neuropharm.2012.05.011 CrossRefPubMedGoogle Scholar
- 63.Nonet ML, Holgado AM, Brewer F, Serpe CJ, Norbeck BA, Holleran J, Wei L, Hartwieg E, Jorgensen EM, Alfonso A (1999) UNC-11, a Caenorhabditis elegans AP180 homologue, regulates the size and protein composition of synaptic vesicles. Mol Biol Cell 10:2343–2360. https://doi.org/10.1091/mbc.10.7.2343 CrossRefPubMedPubMedCentralGoogle Scholar
- 67.Peters C, Espinoza MP, Gallegos S, Opazo C, Aguayo LG (2015) Alzheimer’s Abeta interacts with cellular prion protein inducing neuronal membrane damage and synaptotoxicity. Neurobiol Aging 36:1369–1377. https://doi.org/10.1016/j.neurobiolaging.2014.11.019 CrossRefPubMedGoogle Scholar
- 68.Peters C, Fernandez-Perez EJ, Burgos CF, Espinoza MP, Castillo C, Urrutia JC, Streltsov VA, Opazo C, Aguayo LG (2013) Inhibition of amyloid beta-induced synaptotoxicity by a pentapeptide derived from the glycine zipper region of the neurotoxic peptide. Neurobiol Aging 34:2805–2814. https://doi.org/10.1016/j.neurobiolaging.2013.06.001 CrossRefPubMedGoogle Scholar
- 75.Seshadri S, Fitzpatrick AL, Ikram MA, DeStefano AL, Gudnason V, Boada M, Bis JC, Smith AV, Carassquillo MM, Lambert JC et al (2010) Genome-wide analysis of genetic loci associated with Alzheimer disease. JAMA 303:1832–1840. https://doi.org/10.1001/jama.2010.574 CrossRefPubMedPubMedCentralGoogle Scholar
- 76.Seward ME, Swanson E, Norambuena A, Reimann A, Cochran JN, Li R, Roberson ED, Bloom GS (2013) Amyloid-beta signals through tau to drive ectopic neuronal cell cycle re-entry in Alzheimer’s disease. J Cell Sci 126:1278–1286. https://doi.org/10.1242/jcs.1125880 CrossRefPubMedPubMedCentralGoogle Scholar
- 79.Tam C, Idone V, Devlin C, Fernandes MC, Flannery A, He X, Schuchman E, Tabas I, Andrews NW (2010) Exocytosis of acid sphingomyelinase by wounded cells promotes endocytosis and plasma membrane repair. J Cell Biol 189:1027–1038. https://doi.org/10.1083/jcb.201003053 CrossRefPubMedPubMedCentralGoogle Scholar
- 80.Thomas RS, Henson A, Gerrish A, Jones L, Williams J, Kidd EJ (2016) Decreasing the expression of PICALM reduces endocytosis and the activity of beta-secretase: implications for Alzheimer’s disease. BMC Neurosci 17:50. https://doi.org/10.1186/s12868-016-0288-1 CrossRefPubMedPubMedCentralGoogle Scholar
- 82.Treusch S, Hamamichi S, Goodman JL, Matlack KE, Chung CY, Baru V, Shulman JM, Parrado A, Bevis BJ, Valastyan JS et al (2011) Functional links between Abeta toxicity, endocytic trafficking, and Alzheimer’s disease risk factors in yeast. Science 334:1241–1245. https://doi.org/10.1126/science.1213210 CrossRefPubMedPubMedCentralGoogle Scholar
- 87.Zempel H, Thies E, Mandelkow E, Mandelkow EM (2010) Abeta oligomers cause localized Ca(2+) elevation, missorting of endogenous Tau into dendrites, Tau phosphorylation, and destruction of microtubules and spines. J Neurosci 30:11938–11950. https://doi.org/10.1523/JNEUROSCI.2357-10.2010 CrossRefPubMedGoogle Scholar
- 88.Zhao WQ, Santini F, Breese R, Ross D, Zhang XD, Stone DJ, Ferrer M, Townsend M, Wolfe AL, Seager MA et al (2010) Inhibition of calcineurin-mediated endocytosis and alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors prevents amyloid beta oligomer-induced synaptic disruption. J Biol Chem 285:7619–7632. https://doi.org/10.1074/jbc.M109.057182 CrossRefPubMedGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. 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.