Intracellular amyloid formation in muscle cells of Aβ-transgenic Caenorhabditis elegans: determinants and physiological role in copper detoxification

In order to investigate the key intrinsic determinants of the human Aβ sequence for Aβ intracellular amyloid aggregation, we constructed transgenic C. elegans strains that express two different variants, Arctic (E22G) and NIC (V18A) (Fig. 1A, B, C and 1D), in muscle cells as described in the Methods section. Interestingly, despite having generated several transgenic lines that expressed the Arctic variant, all of them showed levels of expression far lower than those expressing the wild type or the NIC variant (Fig. 1C and 1D). One possibility is that the overexpression of the Arctic variant in C. elegans muscle cells is too harmful and therefore the only surviving transgenic lines were those with low Aβ Arctic expression. For our experiments we used the strain NIC2, with Aβ load similar to that of the Aβ wt expressing strain, and Arctic 6.

We then analyzed the degree of amyloidosis at different times during the animal's life to establish the pattern of intracellular aggregation. Fig. 2(A, B) shows the time course of amyloid formation as visualized by ThS staining. The animals were processed and stained at 48, 72, 96 and 120 h after being synchronized at the embryo stage. All transgenic strains show an increment in the total amyloid deposit area as they age (Fig. 2B). However, the animals expressing the Aβ variants show fewer amyloid aggregates than those that express the Aβwt at all ages. The worms expressing the NIC variant start showing amyloid deposits by 48 h but to a lesser extent than those expressing the Aβwt. Over the next time points (up to 120 h) the formation of amyloid deposits increased in these animals, but to a much lesser extent than in those expressing the Aβwt (Fig. 2A, B) as anticipated from previous in vitro experiments [14]. We were not able to detect any ThS positive spots until 72 h in the animals that express the Arctic variant, which suggests a delay in the formation of mature amyloid deposits as expected. However, in the next time points, the degree of amyloid formation in these transgenic animals was very poor, even lower than those expressing the NIC variant (Fig. 2A, B). We expect this low level of amyloid formation of the Arctic variant is due to the poor expression of the transgene compared with the strains carrying the wild type and NIC variants (Fig. 1C, D). Fig. 3 shows the expression of the three Aβ variants in the transgenic C. elegans lines by immunofluorescence (120 h). They all show expression of the Aβ peptide (green) in the muscle cells identified by phalloidin staining of actin filaments (red). As shown in the orthogonal analyses (Fig. 3B, D, F) the Aβ concentrated in aggregates is mainly located between the actin filaments, which are well observed in C. elegans overexpressing the Aβwt variant (Fig. 3A, B). Even though the Arctic variant formed very few and small amyloid deposits (probably as a consequence of low expression of the transgene (Fig. 1D)), the peptide accumulates in what appears to be fairly big amorphous aggregates, which were Aβ-immunoreactive (Fig. 3C, D). We do not observe clear muscle degeneration in any of the transgenic strains since actin fibers look fairly intact (Fig. 3).

To evaluate if the aggregation state of Aβ expressed in the muscle cells of the worms could be altered by environmental factors, we decided to expose the worms to copper, mainly because there is a large body of evidence indicating that the aggregation of Aβ in the brain can be modulated by transition metals [25, 38, 39]. We performed experiments with copper using the strain expressing the human wild type Aβ peptide, since it is the most commonly found in amyloidogenic diseases, the best characterized, and it behaves as expected in our previous studies. We first analyzed the formation of amyloid aggregates by staining synchronized populations of animals cultured on Cu²⁺, histidine and clioquinol supplemented medium every 24 h. Time zero was established as the time the eggs were harvested and seeded onto agar plates supplemented with the CuCl₂, histidine or clioquinol. The CuCl₂ concentrations used in the experiments were determined by preliminary studies of viability, fertility and Cu²⁺ incorporation in the wild type (N2) strain and analysis of stress induction in the hsp-16:GFP transgenic strain CL2070 [31, 40]. Based on these studies we decided to use a concentration of 150 μM CuCl₂, because it did not affect the survival of the worms. Cu-chelators were also used at a concentration of 150 μM. As previously reported, ThS reactive deposits were not observed in very young animals although their presence has been established even in embryos by staining with the more sensitive Congo-red derived dye X-34 [41]. We started seeing ThS positive deposits 48 h after the eggs where seeded on the plates, when most of the animals were L4 larvae (data not shown). Before this time, it was not possible to perform a quantitative analysis between the different treatments. The number of ThS positive deposits was quantified anterior to the pharyngeal bulb in the cephalic region of the worms (Fig. 4A), as described in the Methods section. Up to 72 h of treatment, the differences in the number of amyloid deposits between the animals exposed to CuCl₂, histidine or clioquinol did not appear significantly different from controls (Fig 4C). After 96 h of treatment, the formation of ThS-positive deposit was more abundant in the worms grown in CuCl₂ than in the worms cultured in the control medium (Fig 4B a–b, C). At this time, the worms cultured in the agar supplemented with Cu-chelators developed scarce amyloid deposits, significantly less than the number of deposits formed in the worms grown in the control medium (Fig. 4B a–d; C). The same pattern was observed at 144 h of treatment (Fig. 4C). We confirmed this quantification by measuring the total aggregate area in the worm's cephalic region at 96, 120 and 144 h (Fig. 4D). As expected, the CuCl₂ treatment increased the total deposit area, while histidine and clioquinol reduced it significantly. Moreover, at this time, the few aggregates formed in the presence of histidine and clioquinol seemed to have a more diffuse morphology than the ThS-amyloid aggregates formed in the presence of CuCl₂ and controls (Fig. 4Bc and 4d). In addition, the treatment with clioquinol often produced a virtual absence of discernible aggregates (Fig 4Bd).

Aβ has a strong affinity for Cu²⁺ [42], therefore, we set up to test whether there was a differential tolerance to the toxic effects of CuCl₂ between the Aβ expressing worms (CL2120) and the controls (CL2122). We reasoned that if Cu²⁺ becomes bound to Aβ, there would be less free Cu²⁺ to cause cell damage. To test this hypothesis we performed toxicity assays in which we exposed 96 h old Aβ-expressing worms and their controls to a range of copper concentrations ranging from zero (no copper added to the medium) to 450 μM in liquid medium during 24 h. In this way we were able to define survival curves for both populations and found a significant shift to the right for the Aβ transgenic animals (Fig. 5A). The lethal concentration at which half of the population dies (LC₅₀) was significantly higher for the animals expressing the Aβ peptide (366 ± 2 μM) than the controls (268 ± 8 μM). These results show that the Aβ transgenic worms are more resistant than controls to high Cu²⁺ concentrations suggesting that there might be less metal available to produce cell damage in this strain. We confirmed that the animals exposed to high concentrations of CuCl₂ (150 μM) during 96 and 120 h incorporated more Cu²⁺ (≈10 μg/g) than the animals grown on regular medium (≈ 0.1 μg/g) (Table 1). Moreover, Cu²⁺ supplementation did not induce changes in Zn²⁺, Fe³⁺, Mn²⁺ or Al³⁺ levels (data not shown). We did not observe any significant difference in constitutive Cu (or Zn²⁺, Fe³⁺, Mn²⁺ and Al³⁺) levels between the Aβ-transgenic C. elegans and their controls, indicating that the presence of Aβ does not affect Cu²⁺ absorption from the agar medium. Therefore, the aggregation of intracellular Aβ was mediated by direct reaction with elevated Cu²⁺. In fact, the Aβ-transgenic animals and their control counterparts exposed to CuCl₂, incorporated about 100 times more Cu²⁺ than the animals grown under normal culture conditions. However, we did not observe any statistically significant difference in the amount of incorporated Cu²⁺ between the Aβ-transgenic C. elegans and their controls under either condition (no CuCl₂ added and CuCl₂ supplementation) (Table 1), indicating that the presence of Aβ is not affecting Cu²⁺ transport, and suggesting that the effect observed on the experiments could be explained principally by the interaction between Cu²⁺ and the Aβ present in the muscle cells of the nematode. These results show that the Aβ transgenic worms are more resistant than controls to high Cu²⁺ concentrations.

In order to confirm that the protective effect of Aβ expression is limited to copper toxicity we performed CdCl₂ and ZnCl₂ toxicity assays. Both metals are known to be toxic and to elicit a stress response in C. elegans [43‐45]. Furthermore, zinc is known to be involved in the metallobiology of Alzheimer's disease [46]. We found that the survival curves (Fig. 5C, D) are almost identical between Aβ expressing worms and controls in both cases. The differences in LC₅₀ for cadmium (CL2120: 1.55 ± 0.02 mM; CL2122: 1.59 ± 0.03 mM) and zinc (CL2120: 4.9 ± 0.8 mM; CL2122: 4.8 ± 0.4 mM) are not statistically significant between the strains. These results exclude the possibility that Aβ overexpression may be providing a more general protective effect, at least to metal toxicity.

Finally, to correlate amyloid aggregate load with resistance to copper toxicity, we performed lifespan experiments to determine if Aβ expressing worms show longer lifespan than controls under 150 μM copper. We found that Aβ worms show a small but significant increase in median and total lifespan (Fig. 5B) demonstrating the protective effect of Aβ when the worms were chronically exposed to copper. When cultured on regular agar, the Aβ expressing worms and the control animals show a similar lifespan (Additional file 1).

To further corroborate the protective effect of Aβ expression to copper toxicity we performed thrashing assays [47]. These assays measure motor deficits in liquid medium and have been used extensively in studies of mutants that affect the function and/or structure of the C. elegans neuromuscular junction [47‐49]. We first evaluated the motility of Aβ expressing worms and controls that had been cultured in regular agar plates for 72 and 120 h (from the embryo stage). We found that the transgenic Aβ animals move 40% slower than controls (Fig. 6), indicating that Aβ is indeed affecting the worms' behaviour, even though this is not evident when they crawl on agar plates. Then, we evaluated the motility of Aβ expressing animals and their controls after being cultured on agar medium supplemented with 150 μM CuCl₂ during the same periods of time (72 and 168 h). We found that while the control animals show decreased movement at both time points after being cultured in copper rich agar plates, the Aβ expressing worms are not affected under the same conditions and they even show significant improvement (Fig. 6 and Additional files 2 and 3).

Distilled H₂O (dH₂O) used in buffer preparations was submitted to resin treatment in order to eliminate organic compounds, cations and anions. After treatment, the resultant dH₂O had a resistance value of 18 MΩ/cm.

Transgenic strains CL2120 and CL2122 were a gift of Dr. Christopher Link and were described earlier [28]. Briefly, CL2120 expresses the Aβ1–42 peptide under the control of the unc-54 (myosin heavy chain) promoter that drives expression to the body wall muscle cells. These animals form intracellular amyloid deposits constitutively in their muscle cells [69]. CL2122 has the same vector as CL2120, minus Aβ1–42, incorporated into the genome and therefore it does not form amyloid aggregates. Both strains have mtl-2/GFP as marker gene and therefore they express GFP in the intestinal cells. The worms were cultured on regular 35, 60 or 100 mm culture plates with NGM agar seeded with the bacterial strain OP50 [70]. The agar and the bacterial cultures were supplemented with histidine (Sigma), clioquinol (Sigma) and CuCl₂ (Sigma) at a concentration of 150 μM. All strains were maintained at 20°C. Worm cultures were synchronized by the hypochlorite method [71] and the resulting eggs were seeded onto the previously described agar plates. The animals were collected every 24 h by washing the plates with M9 buffer [71]. The worms were precipitated in a clinical centrifuge and the M9 removed. The nematodes were then stained as described below.

Construction of the transgenic strains expressing the Aβ variants, Arctic (E22G) and NIC (V18A) (Fig. 1A), were obtained by site directed mutagenesis on the pCL12 plasmid (kindly provided by Dr. Christopher Link) that carries the human wt Aβ1–42 peptide (with a signal peptide in the N-terminus) under the control of the unc-54 promoter that drives expression to the body wall muscle cells [28]. The NIC mutation (V18A) induces a significant increment in the α-helical content of Aβ and dramatically diminishes fibrillogenesis in vitro [14]. The primers used to generate the Arctic (5'TGGTGTTCTTTGCAGGAGATGTGGGTTCAAA3' and its complementary anti-sense primer) and NIC (5'ATCATCAAAAATTGGCGTTCTTTGCAGAAGA3' and its complementary anti-sense primer) mutations were obtained commercially (Invitrogen). We obtained plasmids pPG110 and pPG120 carrying the Arctic and the NIC variant respectively as established by sequencing the DNA inserts. The transgenic lines were created using microparticle bombardment of the plasmid of interest and the selectable co-transformation marker unc-119 (pJN254) into unc-119 (e2498) mutants, as described earlier except that we used tungsten particles instead of gold particles [72]. The presence of each transgene was determined by PCR (primer sense 5'TGACCGTGGAAGATTA 3' located in the unc-54 promoter and primer antisense 5'GTCCAATGATTGCACCT 3' between bases 1208 and 1224 of the pCL12 plasmid and its derivatives pPG110 and pPG120 (Fig. 1B). The expression of the transgenes was established by western blot analysis (Fig. 1C and 1D) and immunofluorescence (6E10 antibody, Sigma) (Fig. 3). We obtained 4 transgenic lines that express the NIC variant and 10 expressing the Arctic variant. For Western analysis we choose lines in which the transgene was incorporated into the genome and therefore transferred to 100% of the progeny. For the analysis of intracellular aggregation (ThS and immunofluorescence) we choose lines with equivalent Aβ expression (NIC2). This was not possible for the Arctic variant since it we could not obtain strains expressing this variant at high levels, but we nevertheless analysed strain Arctic 6.

Worms were collected form 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 lysis buffer (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. The samples were sonicated 2 times on ice for 15 seconds and centrifuged 5 minutes at 10,000 rpm to eliminate the cuticles. The supernatant was transferred to a new tube and total protein content was quantified. Then, a volume corresponding to 50 μg protein was precipitated with acetone. The pellet was resuspended in loading buffer (50 mM Tris-HCl pH6.8; 100 mM DTT; 5% SDS; 10 % glycerol; bromophenol blue; 5% β-mercaptoethanol) and boiled for 3 minutes (adapted from Wu et al. 2006). The samples were subjected to electrophoresis on Tris-Tricine polyacrilamide gels at constant voltage (100 V) at 4°C. The proteins were then transferred to PVDF membranes for 2 hours at 4°C (300 mA). The membranes were boiled in PBS for 3 minutes before blocking with a solution of PBT plus 5% milk for one hour at room temperature. We used the anti-Aβ monoclonal antibody 6E10 (Chemicon) at 1:1,000 dilution or anti-α-tubulin monoclonal antibody (Sigma) at 1:5,000 dilution overnight at 4°C. We used goat anti-mouse HRP as secondary antibody. The densitometry analysis was performed with Matrix software.

We used the monoclonal anti-Aβ antibody 6E10 (Sigma) and as secondary antibody an anti-mouse Alexa Fluor 488 (Molecular Probes) pre-adsorbed with fixed worms. The fixation, permeabilization, and staining of the specimens were performed as described before [73]. Some samples were also stained with rhodamine-phalloidin (1:4,000) for the visualization of actin filaments in muscle cells.

ThS staining was performed as described before [28] with the modification that it was done in whole C. elegans populations instead of on individual worms. Briefly, the worms washed from the plates were submitted to a brief centrifugation at 3,000 rpm in microfuge tubes and fixed by adding 1 ml 4% paraformaldehyde in PBS pH 7.4 for 24 h at 4°C. The next day, the fixative was removed, replaced by 0.5 ml permeabilization solution (1% Triton X-100, 5% β-mercaptoethanol, 125 mM Tris, pH 7.4), and incubated at 37°C for 24 h. After removing the permeabilization solution the animals were washed in PBS-T (Triton X-100 0.1%) pH 7.4, stained in 0.125% ThS in 50% ethanol for 2 min and distained for another 2 min in 50% ethanol. The 50% ethanol was removed; the worm pellet was resuspended in a drop of PBS and a small aliquot observed under a Zeiss Axioplan microscope. If the background of the staining was too high a second wash with 50% ethanol was performed. The animals were finally mounted on a drop of mounting medium (DAKO), covered with a cover slip and sealed. The samples were coded, and the counting and analysis was performed double blind. The samples were then observed and photographed with a Zeiss Axioplan microscope equipped with epifluorescence or analysed in a Zeiss LSM 2000 MT confocal microscope. To quantify the amyloid aggregates, the ThS stained deposits present anterior to the pharyngeal bulb were counted in a minimum of 40 worms in three independent experiments.

Digital quantification of areas positive for ThS fluorescence were measured with the Sigma Scan Pro 5 or ImageJ. The analysis revealed the total average area of the deposits. The total area of deposits was divided by the total head area to standardize it to the worm size. Data from the image analysis microscopy were exported to a Sigma Plot file or GraphPad Prism 4 for statistical analysis. Results were expressed as mean ± standard error. Statistical significance was determined by one-way analysis of variance (ANOVA); p < 0.05 was regarded as statistically significant.

C. elegans worms were grown during 96 or 120 h on regular agar plates or agar plates supplemented with 150 μM CuCl₂. The animals were washed off the plates with dH₂O, placed on microfuge tubes and subjected to three washes with dH₂O, they were then heated 2 times at 100°C for 15 min. All samples were stored at 4°C. To avoid possible metal contaminations, the tubes used to collect the samples were previously treated with 0.5% HNO₃ for 24 h, washed with dH₂O 3 times and let dry at 20°C. The samples were digested in 200 μl of 12N HNO₃ (BDH, ARISTAR) overnight, the next day 200 μl of 1% HNO₃ were added. Samples were further diluted (1:10) in 1% HNO₃ for analysis by inductive coupled plasma mass spectrometry (ICP-MS). ICP-MS was performed using an UltraMass 700 (Varian Inc., Australia). Tissue samples were prepared in triplicate for all treatments and Cu²⁺ values for each sample was determined in triplicate. Cu²⁺ values are mean μg/g weight of each C. elegans sample.

Age-synchronous C. elegans populations were prepared by transferring eggs to a freshly seeded NGM plate. After 96 h, the worms were transferred to K medium (53 mM NaCl, 32 mM KCl) supplemented with different concentrations of CuCl₂ (up to 450 μM) on 24 well microtiter plates (15–20 worms/well) for 24 h at 20°C [44]. The number of dead worms was determined by the absence of touch-provoked movement when prodded with a platinum wire. Each independent experiment of a total of three was performed in triplicate.

Lifespan analysis was conducted at 20°C as described previously [74]. For the lifespan assays, 5–10 adult animals were allowed to lay eggs overnight on regular NGM plates seeded with OP50 bacteria. Parents were removed the next day and eggs were allowed to develop at 20°C. L4 larvae were transferred to new regular NGM plates or to 150 μM CuCl₂ supplemented plates. This time point was used as day 0 to account for differences in developmental time between strains. The worms were transferred to new plates every day during the first 4–5 days to avoid the mixing of generations. The number of dead individuals was verified every two days. A worm was considered dead when it did not respond when prodded two times with a platinum wire pick. Worms that died with larvae inside (eggs hatched before being laid) or those in which the intestine extruded from the vulva were removed from the sample. Lifespan curves were generated and analyzed with GraphPad PRISM^® version 5.0. Log-rank (Mantel-Cox) test was used for statistical analysis of survival curves.

Individual animals of the same age were placed on an 80 μl drop of M9 buffer. After a 2-minute recovery period the worms were recorded for 1.5 minutes and the thrashes counted. A thrash is defined as a change in the direction of bending at the mid body [47]. For the copper treatment experiments, the worms to be assayed were exposed to copper from the embryo (egg) stage in agar plates until the thrashing experiments were performed at 72 h (1 day-old adults), and 168 h later (5 day-old adults).