Alzheimer's disease and inclusion body myositis (IBM) are disorders frequently found in the elderly and characterized by the presence of amyloid-β peptide (Aβ) aggregates. We used Caenorhabditis elegans that express Aβ in muscle cells as a model of IBM, with the aim of analyzing Aβ-induced muscle pathology and evaluating the consequences of modulating Aβ aggregation.
First, we tested whether the altered motility we observed in the Aβ transgenic strain could be the result of a compromised neuromuscular synapse. Our pharmacological analyses show that synaptic transmission is defective in our model and suggest a specific defect on nicotine-sensitive acetylcholine receptors (AChRs). Through GFP-coupled protein visualization, we found that synaptic dysfunction correlates with mislocalization of ACR-16, the AChR subunit essential for nicotine-triggered currents.
Histological and biochemical analysis allowed us to determine that copper treatment increases the amyloid deposits and decreases Aβ oligomers in this model. Furthermore, copper treatment improves motility, ACR-16 localization, and synaptic function and delays Aβ-induced paralysis. Our results indicate that copper modulates Aβ-induced pathology and suggest that Aβ oligomers are triggering neuromuscular dysfunction. Our findings emphasize the importance of neuromuscular synaptic dysfunction and the relevance of modulating the amyloidogenic component as an alternative therapeutic approach for this debilitating disease.
Alzheimer's disease (AD) and inclusion body myositis (IBM) are disorders frequently found in the older population (Askanas and Engel, 2007; Selkoe, 2007). Whereas both diseases affect different tissues, nervous system and muscle, respectively, they have similar features, and most molecules known to be involved in IBM are also present in AD (Rebolledo et al., 2008). The amyloid-β (Aβ) peptide is the main constituent of both senile plaques (SPs) in AD (Masters et al., 1985) and intracellular amyloid aggregates in skeletal muscle cells in IBM (Askanas and Engel, 2006). Because of the inflammatory component of IBM, immunosupressor therapies have been tried in patients without satisfactory results (Dalakas et al., 2001; Barohn et al., 2006). Therefore, the amyloidogenic component acquires greater relevance, and Aβ modulation may provide an alternative therapeutic approach.
It is not yet clear whether the amyloid aggregates trigger the onset of amyloidogenic diseases or whether they are the end product of the cell protective machinery. Current evidence shows that Aβ oligomers, rather than SPs or Aβ fibrils, are the toxic species (Lesne et al., 2006; Lacor et al., 2007). On the other hand, although it is clear that SPs are structured as metal-enriched aggregates that accumulate Cu2+, Fe3+, and Zn2+ (Lovell et al., 1998) and that the aggregation state of Aβ peptide is modulated by metals and metal chelators (Bush et al., 1994b; Cherny et al., 2001), the role of copper as deleterious or beneficial in the context of amyloidogenic and neurodegenerative diseases is controversial (Strausak et al., 2001; Cerpa et al., 2005; Kessler et al., 2005).
Caenorhabditis elegans experimental models for AD and IBM are strains that overexpress the human Aβ3–42 peptide (McColl et al., 2009). When Aβ is expressed in muscle cells, pathological behaviors include paralysis in Aβ thermo-inducible strains (Link et al., 2003) and slow movement in liquid medium in constitutive Aβ strains (Minniti et al., 2009). Serotonin hypersensitivity, altered chemotaxis, and learning deficits have been reported when Aβ is expressed in neurons (Wu et al., 2006; Dosanjh et al., 2010). The toxic Aβ species responsible for these effects have not been thoroughly identified, but they are likely to be Aβ oligomers (Wu et al., 2006) rather than mature Aβ amyloid aggregates.
It is currently known that the neurotoxic effects of Aβ peptide in AD compromise synaptic function before the onset of cell death (Selkoe, 2002). For instance, there is a decrease of synaptic proteins before the onset of plaque formation in murine models of AD (Mucke et al., 2000) and AD patients (Masliah et al., 2001a,b). Given the similarities between AD and IBM, we hypothesize that in IBM the early pathological events might be associated with the neuromuscular synapse.
We found that a C. elegans model of IBM develops functional and structural defects at the neuromuscular level. We also show that copper can improve the Aβ-induced dysfunction and decrease soluble oligomeric Aβ species. Our findings emphasize the importance of neuromuscular synaptic dysfunction in IBM and the relevance of modulating the amyloidogenic component as an alternative therapeutic approach.
Materials and Methods
Transgenic strains CL2120 (dvIs14[unc-54/hAβ1–42(pCL12)+mtl-2::GFP (pCL26)]), CL2122 (dvIs15[pPD30.38(unc-54 vector)+mtl-2::GFP(pCL26)]) (Fay et al., 1998), CL2006 (dvIs2[pCL12(unc-54/hAβ1–42)+pRF4(rol-6(su1006)]) (Link, 1995), CL6176 (smg-1(cc546)I; dvIs19(Pgst-4::GFP)III;+/nT1(unc d;let-?IV;V);dvIs27 [pAF29(Pmyo-3::Aβ42)+pRF4]X), and CL6180 (smg-1(cc546)I; dvIs19(Pgst-4::GFP)III; skn-1(zu67) IV/nT1(unc d;let-?IV;V); dvIs27[pAF29 (Pmyo-3::Aβ42)+pRF4]) (Dostal et al., 2010) were a gift from Dr. C. Link, University of Colorado, Boulder, CO.
Strains carrying GFP fusion protein Pmyo3::ACR-16::GFP or Punc-29::UNC-29::GFP (Francis et al., 2005) were a gift from Dr. M. Francis, University of Massachusetts, Worcester, MA.
RB918 (acr-16(ok789)V), CB193 (unc-29(e193)I), CL4176 (smg-1(cc546ts)I; dvIs27[pAF29(myo-3/Aβ1–42/let UTR)+pRF4(rol-6(su1006)]X), and its standard control CL802 (smg-1(cc546ts)I; rol-6(su1006)II) (Link et al., 2003) were obtained from the Caenorhabditis Genetics Center. We constructed the following strains by standard genetic crosses: ANM30 (dvIs2[pCL12(unc-54/hAβ1–42)+pRF4(rol-6(su1006)]; Pmyo3::ACR-16::GFP); ANM31 (rol-6(su1006)II;Pmyo3::ACR-16::GFP); ANM40 (dvIs14[unc-54/Aβ1–42(pCL12)+mtl-2::GFP(pCL26)];acr-16(ok789)V); and ANM43 (dvIs2[pCL12(unc-54/hAβ1–42)+pRF4(rol-6(su1006); Punc-29::UNC-29::GFP).
Nematode propagation and treatments.
The worms were cultured on regular nematode growth medium (NGM) culture plates seeded with the bacterial strain OP50 (Brenner, 1974). Worm strains were maintained at 20°C, except for CL4176, CL802, CL6176, and CL6180, which were maintained at 16°C, and shifted to 23–25°C to induce Aβ expression. For copper treatment, NGM was supplemented with CuCl2 (Sigma) at the indicated concentrations, and the worms were treated from the embryo stage (CL2120, Cl2122, and CL2006 strains) or from the time of temperature upshift (CL4176, CL802, CL6176, and CL6180 strains) until the experiments were performed.
To obtain developmentally synchronized populations, gravid hermaphrodites were subjected to the hypochlorite method (Lewis and Fleming, 1995), and the resulting eggs were seeded onto the previously described agar plates. The animals were collected at indicated times by washing the plates with M9 buffer (Lewis and Fleming, 1995), precipitated by gentle centrifugation and removal of M9. The nematodes were then processed as described below. Hermaphrodite worms were used for all the experiments.
Thioflavine-S (Th-S) staining was performed as described previously (Fay et al., 1998). Briefly, the worms washed from the plates were fixed with 4% paraformaldehyde in PBS, pH 7.4, for 24 h at 4°C. The fixative solution was removed, replaced by permeabilization solution (125 mm Tris, pH 7.4, 1% Triton X-100, 5% β-mercaptoethanol), and incubated at 37°C for 24 h. The animals were washed three times in PBS-T (PBS plus 0.1% Triton X-100), stained in 0.125% Th-S (Sigma) in 50% ethanol for 2 min, and destained for another 2 min in 50% ethanol. The stained samples were resuspended in PBS and mounted in fluorescence mounting medium (Dako).
(Trans, trans)-1-bromo-2,5-bis-(3-hydroxycarbonyl-4-hydroxy) styrylbenzene (BSB) was purchased from AnaSpec and prepared according to the manufacturer's instructions. The assay was performed according to the protocol established by Toledo and Inestrosa (2009) with the following modifications. Briefly, synchronized worms were cultured and collected at different stages, fixed with 4% paraformaldehyde, and permeabilized as indicated above. Then, the worms were incubated with BSB (0.02% resuspended in 50% ethanol) for 4 h at 37°C. The worms were then washed with lithium carbonate (1 m), washed with water, and finally mounted for microscopy observation. BSB staining intensity was estimated using the WCIF ImageJ software.
Fluorescence images were acquired using the same exposure parameters with a 40× or 100× objective in a BX51 microscope (Olympus). The microscope was equipped with a digital camera, Micropublisher 3.3 RTV (JH Technologies). For ACR-16::GFP and UNC-29::GFP distribution analysis, 1- to 2-d-old worms were anesthetized with 20 mm NaN3 and photographed.
Individual animals of the same age were placed on an 80 μl drop of M9 buffer. After a 2 min recovery period, the worms were recorded for 1.5 min with a digital camera (Nikon Coolpix-4500), and the thrashes were counted while replaying the video in slow motion. A thrash is defined as a change in the direction of bending at the midbody (Miller et al., 1996).
Neuromuscular synaptic transmission assay.
Worms (25–30) of the same age were place on assay plates (NGM supplemented with 0.25 mm levamisole, 31 mm nicotine, or 20 mm nicotine). We then analyzed the drug-induced paralysis over time (Mahoney et al., 2006). As an internal control, we use the RB1131 [lev-10(ok1154)I] strain that is nicotine and levamisole resistant, because lev-10 is required for clustering of postsynaptic proteins (Gally et al., 2004).
Temperature-induced paralysis assay.
The assays were performed as described by Wu et al. (2006), with the following modification. At the L3 stage, worms were moved to newly seeded NGM plates with or without copper supplementation, and the temperature was upshifted to 23–25°C. The ratio of paralyzed to nonparalyzed individuals (90–120 worms per plate) over time was plotted.
These experiments were performed as described by Dostal et al. (2010) using the corresponding Escherichia coli from the Ahringer RNAi feeding library. Exposure to the skn-1 RNAi clone caused the treated animals to lay over 99% dead eggs, the expected skn-1 phenotype.
Worms were collected from the plates with M9 buffer, transferred to tubes, centrifuged, and washed two times to eliminate bacteria. The final worm pellet was resuspended in 100 μl of lysis buffer in the presence of protease inhibitors (50 mm HEPES, pH 7.5, 6 mm MgCl2, 1 mm EDTA, 75 mm sucrose, 25 mm benzamidine, 1 mm DTT, 1% Triton X-100) and frozen at 80°C. Samples were sonicated three times on ice for 15 s.
Western blot analyses.
For Aβ analysis, the crude protein extracts were centrifuged 5 min at 10,000 rpm to eliminate cuticles. The supernatant was transferred to a new tube, and total protein content was quantified using the BCA kit (Pierce). Equal protein quantities were suspended in loading buffer (50 mm Tris-HCl, pH 6.8, 100 mm DTT, 5% SDS, 10% glycerol, 0.02% bromophenol blue, 5% β-mercaptoethanol) and boiled for 3 min (adapted from Wu et al., 2006). The samples were subjected to electrophoresis on Tris-Tricine polyacrylamide gels at a constant voltage (100 V) at 4°C. The proteins were then transferred to PVDF membranes for 2 h at 4°C (300 mA). The membranes were boiled in PBS for 3 min before blocking with a solution of PBT (PBS and 0.1% Tween 20) plus 5% milk for 1 h at room temperature. We used the mouse anti-Aβ monoclonal antibody 6E10 (Millipore) at 1:1000 dilution or mouse anti-α-tubulin monoclonal antibody (Sigma) at 1:5000 dilution overnight at 4°C. We used goat anti-mouse HRP as a secondary antibody.
Equal protein quantity was spotted on 0.45 mm2 nitrocellulose (Millipore), blocked with PBS-T gelatin (PBS, 0.05% Tween 20, 0.3% gelatin), and incubated using the anti-oligomeric antibody A11 (Millipore) at 1:5000 dilution for 2 h at room temperature or overnight at 4°C. We used goat anti-rabbit HRP as a secondary antibody.
The data were analyzed using GraphPad Prism 5 software. Values are presented as mean ± SE, and t test analysis was used to compare treatments, except for mislocalization of GFP fusion proteins where a Z test was used to compare proportions (*p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001).
Transgenic C. elegans expressing Aβ show synaptic defects
Under normal culture conditions, we previously observed that adult worms expressing Aβ constitutively move significantly slower than the controls that do not express Aβ (Minniti et al., 2009). We hypothesized that the compromised motility of Aβ transgenic worms in liquid medium could be, at least in part, the result of Aβ-induced defects at the neuromuscular junction (NMJ), leading to synaptic transmission failure and decreased motor capacity. To test this hypothesis, we performed pharmacological assays with drugs known to affect NMJ function. Because Aβ is expressed in muscle cells, we evaluated postsynaptic NMJ function using levamisole and nicotine, agonists of the acetylcholine receptors (AChRs). C. elegans has two subtypes of AChRs at the NMJ that are differentially sensitive to these agonists (Richmond and Jorgensen, 1999). The AChR subunits responsible for levamisole-triggered currents are UNC-38, UNC-63, UNC-29, and LEV-1 (Fleming et al., 1997; Culetto and Sattelle, 2000). On the other hand, ACR-16 is the only reported AChR subunit responsible and essential for nicotine-triggered currents, and it is insensitive to levamisole (Francis et al., 2005; Touroutine et al., 2005). These drugs generate a sustained stimulation of AChRs and release of Ca2+, causing paralysis in the worms. The time course of paralysis is dependent on the efficacy of synaptic transmission (Gottschalk et al., 2005; Mahoney et al., 2006).
The C. elegans transgenic worms that constitutively express the Aβ peptide in muscle cells and control worms (72 h of age) were assayed on NGM plates supplemented with one of the drugs mentioned above, and the course of paralysis was plotted. We did not find significant differences between the Aβ transgenic C. elegans strain and its control when we assayed them on levamisole plates (Fig. 1A). However, we observed significant resistance to nicotine in the Aβ transgenic strain compared with its control (Fig. 1B). Resistance is also observed in assays using lower nicotine concentrations (20 and 25 mm) (Fig. 1F and data not shown). Older worms (120 h of age) also show the same response to levamisole and nicotine (Fig. 1C,D). These data suggest that not all postsynaptic AChRs are affected in this model; only those sensitive to nicotine but insensitive to levamisole appear to be compromised by the presence of Aβ.
To support the idea that Aβ is affecting the nicotine-sensitive receptors, we performed a genetic test. The acr-16 mutants do not show obvious locomotory defects on agar plates, whereas the levamisole receptor (unc-38, unc-63, unc-29 or lev-1) mutants show uncoordinated movement. The acr-16/unc-29 (or unc-38) double mutant has a synthetic phenotype where the uncoordinated movement is exacerbated (Francis et al., 2005). If Aβ was interfering with ACR-16 but not with the levamisole (UNC-29-containing) receptor, the uncoordinated phenotype of unc-29 mutants would be intensified with Aβ expression. Using thrashing assays, we observed that the expression of Aβ increases the severity of the uncoordinated phenotype of unc-29(e193) mutants (Fig. 1E). This result suggests that the Aβ peptide is affecting molecules different from the AChRs sensitive to levamisole. Using the same strategy, we found that the Aβ worms are not statistically different from the Aβ worms in the acr-16 background (Aβ, 107.8 ± 3.9 thrashes/min; acr-16/Aβ, 97.79 ± 2.8 thrashes/min; p = 0.0513), suggesting that ACR-16 may be one of the Aβ targets. We also performed nicotine sensitivity assays (20 mm) (Fig. 1F). The acr-16 worms are resistant to 20 mm nicotine. We observed that, as we describe above, the worms that express Aβ are resistant to paralysis compared with control (no Aβ) worms. If Aβ was affecting the ACR-16-containing receptors, worms expressing Aβ in the acr-16 background should have a phenotype very similar to that of the Aβ worms. Aβ and Aβ/acr-16 worms show paralysis curves that are not statistically different at this nicotine concentration (20 mm), except for the last point of the curve (90 min), where the Aβ/acr-16 worms show improved synaptic transmission. This behavioral improvement of Aβ worms in the absence of ACR-16 may indicate that ACR-16 is indeed one of the Aβ targets (direct or indirect) since its presence seems to be necessary for Aβ to exert full synaptic dysfunction.
The ACR-16 nicotinic acetylcholine receptor is mislocalized in Aβ transgenic C. elegans
Since we found that C. elegans expressing Aβ have decreased sensitivity to nicotine compared with the control, we hypothesized that Aβ could be downmodulating the activity of ACR-16-containing receptors, which could explain, in part, the nicotine resistance. One possible mechanism is that the presence of Aβ could be directly or indirectly altering the subcellular localization of ACR-16 receptors. To test this possibility, we crossed the Aβ-expressing strain with a strain carrying ACR-16::GFP or UNC-29::GFP fusion proteins in muscle cells, and we analyzed the distribution of the fluorescent marker on the ventral nerve cord. Control worms expressing these markers show a punctuate pattern over the ventral nerve cord (Fig. 2A,C). Aβ expression does not change nerve cord localization of the UNC-29 subunit of the levamisole-sensitive receptors (Fig. 2D). In contrast, ACR-16::GFP localization is conspicuously altered in worms that express the Aβ peptide. We observe a less structured punctuate pattern over the ventral nerve cord and accumulation of fluorescent puncta in the muscle arms (Fig. 2B). In fact, the muscle arms can now be clearly observed because of the presence of the fluorescent-tagged receptors. The correct localization of the receptors at the nerve cord makes it virtually impossible to clearly visualize the muscle arms.
Copper treatment improves synaptic function in Aβ transgenic worms
We previously reported that copper treatment increases the number and size of amyloid deposits in this model (Minniti et al., 2009). At the time, we also showed some data that hinted at the possibility of copper being able to ameliorate the motility impairments in Aβ transgenic C. elegans. Here, we clearly show that copper treatment improves motility in Aβ transgenic worms at all ages tested (Fig. 3A). Therefore, we evaluated whether synaptic dysfunction and therefore nicotine resistance could also be prevented when the animals were cultured on 150 μm copper-supplemented plates for 72 h. The treatment was able to recover the normal response to nicotine in Aβ transgenic strains, resulting in a curve very similar to that of the control strain under regular culture conditions (Fig. 3B). Using a lower concentration of copper (10 μm), the normal response is only partially recovered in 72-h-old worms (Fig. 3C). Once again, a similar behavior was observed in older worms (120 h) treated with 150 μm copper (Fig. 3D). However, treatment with 10 μm copper did not improve behavior at this age (Fig. 3E). The response of the control strain (no Aβ) to nicotine was not affected by copper treatment (Fig. 3B–E), indicating that copper does not normally affect NMJ function at the concentrations we used.
Mislocalization of the nicotinic acetylcholine receptor ACR-16 is prevented by copper treatment in Aβ transgenic worms
Since in our functional experiments we observed improved synaptic function after copper treatment, we decided to test whether this enhanced function correlates with a more correct localization of the ACR-16 receptor. We analyzed the distribution of ACR-16::GFP on the ventral nerve cord in controls, Aβ transgenics, and Aβ transgenics treated with copper for 72 h (1 d old). We found that the altered localization of ACR-16::GFP observed in the Aβ transgenic worms (that extends to the muscle arms and away from NMJ contacts at the nerve cord) (Figs. 2B, 4C) is reverted after copper treatment (Fig. 4D,E).
We quantified the percentage of worms with one or more muscle arms containing mislocalized ACR-16::GFP in five groups (control worms, control worms treated with copper, Aβ-expressing worms, and Aβ transgenics treated with two concentrations of copper). The quantification shows that the presence of Aβ significantly increases ACR-16 mislocalization and that this mislocalization is prevented by copper treatment (Fig. 4F). Control worms that express ACR-16::GFP show no altered receptor distribution in the presence of copper (Fig. 4B), whereas the expression of the GFP-tagged protein, evaluated by Western blot, seems to be unaltered by copper treatment (data not shown).
Copper treatment delays Aβ-induced paralysis in a temperature-sensitive strain
To further evaluate the effect of copper treatment on Aβ-induced pathology, we used a temperature-sensitive C. elegans strain that expresses Aβ in muscle cells. The temperature upshift from 16 to 23°C during some larval stages induces abundant Aβ peptide expression, leading to progressive paralysis (Link et al., 2003). Representative pictures of the culture plates are shown in Figure 5. The control strain (no Aβ peptide expression) moves normally at 16°C, and temperature upshift to 23°C does not trigger any phenotypes (Fig. 5A). The Aβ transgenic worms maintain their mobility when cultured at 16°C (Fig. 5B); if there is a temperature shift to 23°C during the L3 stage, the worms become paralyzed (Fig. 5C). For this experiment, we started copper treatments at the same time of the temperature upshift to 23°C, when Aβ expression was induced. We could not do it during the embryo stage because the worms (both controls and Aβ worms) did not grow properly or died before they were adults (data no shown). It is possible that the different genetic background of the worms we used in this assay (smg-1) makes the worms more sensitive to copper toxicity during this developmental stage. Under these conditions, the progression of paralysis was delayed compared with that of the untreated worms (Fig. 5C,D,E). We used different copper concentrations, and we observed that this effect is concentration dependent (Fig. 5E).
The SKN-1 transcription factor is not required for copper protection
It was recently reported that the induction of the SKN-1 pathway in response to treatment with coffee extract leads to protection in a C. elegans model of Aβ toxicity (Dostal et al., 2010). C. elegans SKN-1 is the functional orthologue of the mammalian protein Nrf2 (nuclear factor-E2-related factor 2), controlling the induction of multiple genes involved in detoxification and antioxidant stress response (Kahn et al., 2008). Nrf2 activation has been suggested to be protective in several neurodegenerative conditions, including neurons exposed to Aβ and a mouse model of AD (Kanninen et al., 2008, 2009; Wruck et al., 2008). It was also suggested that Nrf2 induction can be modulated by copper (Calay et al., 2010; Wang et al., 2010; Simmons et al., 2011).
Therefore, we investigated whether the SKN-1 detoxification pathway could be involved in the copper protection we observe in this model. We found that copper, at the concentrations used in our assays, is not able to induce the Pgst-4::GFP reporter gene in C. elegans (Kell et al., 2007) (Fig. 6A), where gst-4 (glutathione S-transferase) is a downstream effector of the SKN-1 phase II detoxification pathway (Kahn et al., 2008). To determine whether loss of SKN-1 can decrease or block the protection given by copper exposure, transgenic C. elegans expressing Aβ were fed with skn-1 RNAi (Dostal et al., 2010). The Aβ-induced paralysis in the temperature-sensitive strain (CL4176) is still delayed by copper treatment when skn-1 is knocked down (Fig. 6B). We also used the inducible Aβ strain that carries the skn-1 loss-of-function allele zu67 (CL6180) and its control (CL6176) that carries the wild-type copy of the gene, because skn-1 knockdown worms may still have residual skn-1 gene product. Again, our data show that the paralysis delay induced by copper treatment is maintained in the absence of SKN-1 (Fig. 6C).
Since we are particularly interested in neuromuscular junction dysfunction caused by Aβ accumulation in muscle cells, we performed nicotine assays to evaluate whether copper treatment could still improve synaptic function in the absence of SKN-1. The results show that knocking down skn-1 using the RNAi feeding protocol does not block the effect of copper in recovering the normal response to nicotine in Aβ transgenic worms (Fig. 6D). Note that the graph in Figure 6D is equivalent to that shown in Figure 3B, where skn-1 was not manipulated. Therefore, under the conditions of our experiments, the SKN-1 pathway (Nrf2 in mammals) does not appear to be involved in copper protection in this model of Aβ toxicity.
Copper treatment decreases the relative amounts of soluble Aβ oligomers
In the C. elegans strain that constitutively expresses Aβ, there is a correlation between copper treatment, increment of amyloid deposits (Minniti et al., 2009), improvement of motility, and recovery of a normal response to nicotine. In the temperature-inducible strain, copper treatment delays Aβ-induced paralysis. To investigate whether there was also an increment of amyloid deposits in this temperature-sensitive strain, we tested the status of Aβ aggregation in the temperature-sensitive strain by Th-S staining 30 h after temperature upshift. We were not able to observe quantifiable Th-S-positive deposits in either control or copper treatment conditions (Fig. 7A). It is likely that the time allowed for Aβ induction is not enough for the formation of Th-S-positive amyloid aggregates. Longer induction times are not possible because the animals become paralyzed.
Because we were not able to observe amyloid deposits with Th-S, we performed Western blot analyses using the 6E10 antibody, which recognizes residues 1–17 of Aβ, to determine the Aβ species present in the C. elegans temperature-inducible strain. We were able to visualize four bands: the monomer (∼4 kDa) and three oligomeric forms of 20, 28, and 32 kDa. This assay shows that the worms treated with 150 μm copper have lower quantities of these oligomeric forms than the untreated worms, with no changes in the relative quantities of the monomer (Fig. 7B). Our results suggest that copper may induce a shift from Aβ oligomeric species to less toxic fibrilar species that do not reach amyloid structure before the animal becomes paralyzed. Since immunoreactive bands from SDS-PAGE may not necessarily indicate the presence of oligomeric species in vivo (Bitan et al., 2005), we also used two other approaches to analyze the presence of Aβ oligomers.
We evaluated whether the increment in amyloid deposits observed in the strain that expresses Aβ constitutively correlates with a decrease in oligomeric forms of Aβ. We performed dot-blot assays in strain CL2120, using the A11 antibody that recognizes oligomeric species (Kayed et al., 2003). We observed that copper treatment (150 μm) significantly decreased the amount of oligomers compared with the same untreated strains (Fig. 7C). We obtained similar results using strain CL2006 (data not shown). In addition, we stained the worms with the dye BSB, a Congo Red derivative that has higher affinity for oligomers than for fibrils (Schmidt et al., 2001; Maezawa et al., 2008; Toledo and Inestrosa, 2009). BSB-positive areas are decreased after copper treatment (Fig. 7D,E), providing additional evidence for oligomeric species reduction.
It is currently accepted that the Aβ peptide is the main player in AD and its effects are not limited to neuronal death in areas of Aβ accumulation. Several early effects in AD development, leading to loss of synaptic function, have also been attributed to Aβ (Selkoe, 2002). For instance, AD murine models overexpressing Aβ show an important decrease in immunoreactivity for synaptophysin (Mucke et al., 2000), which is also seen in AD patients (Masliah et al., 2001a,b). In addition, cellular cultures from IBM patient biopsies show evidence of abnormal innervations: NMJ proteins such as AChRs and acetylcholinesterase are diffusely distributed (McFerrin et al., 1998), suggesting a neuromuscular synaptic transmission disorder (Askanas et al., 1998). In our work, through pharmacological manipulation of synaptic transmission, we identified changes at the C. elegans NMJ attributable to Aβ expression, and we show that copper treatment can reverse these defects. We did not find copper upregulation of the SKN-1 phase II detoxification pathway, known to confer protection against Aβ toxicity. On the other hand, the histological and biochemical assays suggest that protection is given by decreasing Aβ oligomeric species. We found that C. elegans expressing Aβ have decreased sensitivity to nicotine compared with the control. This is a new phenotype for Aβ transgenic strains used as models of AD and IBM. Since there is no difference in the response to levamisole, we hypothesize that there might be a selective effect of Aβ on the AChRs, affecting only those that are sensitive to nicotine but insensitive to levamisole. This is supported by the results showing mislocalization of ACR-16-containing AChRs, whereas UNC-29-containing receptors are not affected.
The sequence of the C. elegans ACR-16 protein, the only essential component of the AChR-mediated nicotinic current (Francis et al., 2005; Touroutine et al., 2005), shows high identity with the α7 subunit of the vertebrate nicotinic AChR (Ballivet et al., 1996). The vertebrate nicotinic AChRs (nAChRs) containing the α7 subunit are localized in several brain areas that are involved in cognitive function (Dominguez del Toro et al., 1994; Drago et al., 2003), and their abnormal functionality has been related to several neuropathologies, including AD (Quirion et al., 1986; Perry et al., 1987, 1995). For instance, α7nAChRs colocalize with AD plaques (Wang et al., 2000a) and correlate with neurons that accumulate Aβ (Wevers et al., 1999) and with neurons most vulnerable to Aβ toxicity (D'Andrea and Nagele, 2006). Furthermore, there are data showing a likely Aβ/α7nAChR interaction (Wang et al., 2000a,b). However, the consequences of these interactions are a controversial issue: For instance, it was shown that the activity of α7nAChR is blocked by Aβ, supporting the hypothesis that Aβ could inactivate these receptors (Liu et al., 2001). On the other hand, other researchers showed that extracellular Aβ can selectively activate α7nAChRs (Dineley et al., 2002). Nevertheless, the prolonged exposure to Aβ leads to receptor inhibition, probably through desensitization. Additionally, the association of Aβ with α7nAChR can trigger Aβ internalization, generating intracellular accumulation and, eventually, toxicity (Nagele et al., 2002); however, it is not clear whether this internalization is caused by an indirect influence or to a direct binding of Aβ to AChRs. Buckingham et al. (2009), summarized evidence showing that the signaling pathways affected by intracellular Aβ accumulation seem to be the same as those intervened by extracellular Aβ treatment. The authors speculate that the activation of nAChR-dependent pathways by exogenous Aβ is actually the result of Aβ internalization and discuss that the partial block of α7nAChR responses by Aβ may be the result of Aβ-induced internalization of the receptors (Buckingham et al., 2009).
Given this background and our results showing that the expression of intracellular Aβ peptide leads to mislocalization of ACR-16-containing AChRs (Fig. 2B), we speculate that an Aβ/ACR-16 interaction could be altering the activity of ACR-16-containing receptors, which could explain the partial nicotine resistance. However, we do not know when and where Aβ could be interacting directly or indirectly with ACR-16. Studies from vertebrate α7nAChRs have been performed incubating the cells with Aβ solutions (Wang et al., 2000a,b), suggesting an interaction when the receptor is at the cell surface. However, the transgenic C. elegans strains express the Aβ peptide intracellularly in muscle cells (Link, 1995), as is the case of Aβ accumulation in IBM (Askanas and Engel, 2007). The Aβ−expressing strains used in this work were generated from a construct that contains a secretion signal sequence (Link, 1995) with the aim of obtaining a worm that could secrete high levels of Aβ peptide. However, the strains obtained had detectable Aβ levels only intracellularly, maybe because the short translation product of the minigene unc-54/Aβ could interfere with an efficient traffic through the secretory pathway. Alternatively, amyloid deposition inside the muscle cells could prevent the secretion of immunologically detectable Aβ (Link, 1995). In this scene, the interaction between ACR-16 and Aβ peptide may occur either in the secretory pathway as well as on the cell surface with some undetectable secreted Aβ. In our experiments, the ACR-16::GFP protein is found in the muscle arms, but we cannot distinguish whether this is a consequence of ACR-16 internalization or altered traffic to the postsynaptic site. Because of the reasons explained before, both alternatives are possible. Furthermore, although ACR-16 receptor mislocalization could be a mechanism for partially explaining the synaptic dysfunction observed in transgenic C. elegans, possible alterations in ionic channel properties of ACR-16, and also some other proteins involved in intracellular receptor trafficking, could occur because of the presence of Aβ. For instance, Aβ could be interfering with the Ror-receptor tyrosine kinase CAM-1 that is required for ACR-16 localization to the postsynaptic region at the NMJ (CAM-1 is not involved in the localization of other receptors such us UNC-29 and UNC-49 (GABA receptors) (Francis et al., 2005).
We have previously shown that copper treatment increases both the number and area of amyloid deposits in the muscle cells of C. elegans expressing Aβ (Minniti et al., 2009). This observation is consistent with in vitro data showing copper induced acceleration of Aβ aggregation (Bush et al., 1994a,b). The intracellular amyloid deposits have been widely used as a pathological marker in C. elegans, and the reduction of these deposits has been considered beneficial in this model (Wu et al., 2006). However, our data suggest that the increased aggregation triggered by copper treatment actually improves behavioral deficits. Here, we show improved motility in liquid medium when transgenic Aβ worms are treated with CuCl2. The control strain, which does not express Aβ, shows abnormal development, compromised motility, and a general unhealthy appearance when cultured at the same concentration of CuCl2, supporting the idea that Aβ expression and aggregation can confer resistance to copper toxicity (Minniti et al., 2009). Furthermore, Aβ-induced paralysis in the temperature-sensitive strain is delayed by copper treatment in a concentration-dependent manner. However, it is likely that treatments with higher concentrations of copper exceed the protective role of Aβ and become harmful. Our results show that the protective effect of copper does not seem to be attributable to the activation of the SKN-1/Nrf2 phase II detoxification pathway, opposite to that found for coffee extract in this model (Dostal et al., 2010).
The improved motility in liquid medium and the delay in Aβ-induced paralysis triggered by copper exposure associated with an increment in Aβ aggregation suggest that increased amyloidosis could be decreasing the quantity of oligomeric species considered to be the toxic species in AD (Cleary et al., 2005; Lesne et al., 2006; De Felice et al., 2007; Lacor et al., 2007; Cerpa et al., 2008). Furthermore, Aβ oligomers are also present intracellularly in muscle fibers of IBM patients, where they could contribute to the pathogenic cascade (Nogalska et al., 2010). Our biochemical and histochemical analyses show strong evidence of Aβ soluble oligomer reduction after copper treatment (Fig. 7).
In summary, we have shown a new phenotype in C. elegans that expresses Aβ in muscle cells: NMJ defects manifested as resistance to nicotine that correlate with the mislocalization of ACR-16::GFP-containing receptors. We have also shown that copper can modulate Aβ-induced synaptic dysfunction and altered morphology and delay the development of paralysis in this invertebrate IBM model. The increment in amyloid deposits triggered by copper and the associated decrease of some oligomeric Aβ species is related with the improvement of Aβ-induced pathological traits. Our research suggests that neuromuscular synaptic dysfunction could be an important early feature of human IBM pathology and that the modulation of the amyloidogenic component in this disease could be a relevant therapeutic alternative.
This work was supported by Fondo de Financiamiento de Centros de Excelencia en Investigación Biomedicine Grant 13980001, the Millennium Institute of Fundamental and Applied Biology, the Center for Aging and Regeneration, Programa de Financiamiento Basal 12/2007, the Base Financing Program for Scientific and Technological Centers of Excellence, and Comisión Nacional de Investigación Cientifica y Tecnológica. D.L.R. has a predoctoral fellowship from CONICYT. We thank Dr. Michael M. Francis (University of Massachusetts, Worcester, MA) for his kind gift of the ACR-16::GFP and UNC-29::GFP strains, Dr. Enrique Toledo (Pontificia Universidad Católica de Chile) for help with BSB staining and A11 dot-blot assay, Dr. Christopher Link (University of Colorado, Boulder, CO) for his generous gift of C. elegans strains that express Aβ proteins in body wall muscle cells, and Dr. Olivia Casanueva for the skn-1 RNAi clone.
- Correspondence should be addressed to Nibaldo C. Inestrosa and Alicia N. Minniti, Center for Aging and Regeneration, Facultad de Ciencias Biológicas, P. Universidad Católica de Chile, Alameda #340, P.O. Box 114-D, 8331010 Santiago, Chile. or