Prolongation of metallothionein induction combats Aß and α-synuclein toxicity in aged transgenic Caenorhabditis elegans

Neurodegenerative disorders (ND) like Alzheimer’s (AD), Parkinson’s (PD), Huntington’s or Prion diseases share similar pathological features. They are all age dependent and are often associated with disruptions in analogous metabolic processes such as protein aggregation and oxidative stress, both of which involve metal ions like copper, manganese and iron. Bush and Tanzi proposed 2008 in the ‘metal hypothesis of Alzheimer’s disease’ that a breakdown in metal homeostasis is the main cause of NDs, and drugs restoring metal homeostasis are promising novel therapeutic strategies. We report here that metallothionein (MT), an endogenous metal detoxifying protein, is increased in young amyloid ß (Aß) expressing Caenorhabditis elegans, whereas it is not in wild type strains. Further MT induction collapsed in 8 days old transgenic worms, indicating the age dependency of disease outbreak, and sharing intriguing parallels to diminished MT levels in human brains of AD. A medium throughput screening assay method was established to search for compounds increasing the MT level. Compounds known to induce MT release like progesterone, ZnSO4, quercetin, dexamethasone and apomorphine were active in models of AD and PD. Thioflavin T, clioquinol and emodin are promising leads in AD and PD research, whose mode of action has not been fully established yet. In this study, we could show that the reduction of Aß and α-synuclein toxicity in transgenic C. elegans models correlated with the prolongation of MT induction time and that knockdown of MT with RNA interference resulted in a loss of bioactivity.

www.nature.com/scientificreports/ MT release in ageing worms. ZnSO 4 , quercetin, apomorphin and dexamethasone induced MT, decreased proteotoxicity of Aß and α-synuclein in strains CL2659 and NL5901 and prolonged lifespan in the wild type strain N2 (Fig. 2). Compounds for novel neurotherapeutics against NDs, such as clioquinol (CQL) thioflavin T (Th T) and emodin decreased proteotoxicity by MT induction and all but clioquinol prolonged lifespan in the wild type strain N2 (Fig. 3). Sesamin was not able to induce MT ( Table 1). Knockdown of MT even resulted in a loss of function of the bioactivity of emodin in the AD assay with strain CL2659 (Fig. 3, Table 2).

Results and discussion
Breakdown of metallothionein induction in aged Aß expressing worms. In 2002 Mijazaki et al. 28 showed that the expression of MT and its mRNA was up-regulated in the healthy aged rat brain, whereas treatment with lipopolysaccharide (LPS)-induced expression of MT only in young but not in aged rat brain regions. Based on this study, we monitored MT expression in transgenic worms with (CL2120) and without (CL2122) Aß expression. We could show that MT in healthy organisms slightly increases with age whereas in Aß expressing worms an intense induction in the young adults was followed by a breakdown during ageing (Fig. 1a-c) accompanied by an accumulation of iron, copper and manganese 7 . Therefore, we hypothesized that prolonging the time span of MT release might be a promising therapeutic target in NDs.
Establishment of a novel medium throughput screening assay method to search for compounds that prolong time of Mt induction in aged transgenic C. elegans model of AD. To screen compounds for their ability to prolong MT induction we established a robust medium throughput screening assay based on C. elegans strain CL2120. CL2120 MT is GFP tagged and can be detected by fluorescence in a multiwell plate reader at em/ex 450-535 nm. By a serial dilution step in a multiwell plate, we determined the optimal worm density for a strong signal to be between 30 and 50 worms per well (Fig. 1c). The first measurement (day 0) was performed when larvae in the 4th stage were transferred from the petri dish to the 96 well plate containing media with compounds. The second time point indicated the highest MT level and was between d4 and d7. The last time point indicated the breakdown of MT induction between d7 and d9 (Fig. 1a, c). To guarantee optimal feeding of the worms throughout the whole experiment, we measured the optical density at 600 nm (OD 600 ) for the determination of the bacterial clearance. In this way, the change in optical density of bacteria over time was quantified and the necessary amount of E. coli OP50 was added based on the absorbance value 34 . The optimal concentration of the E. coli food source was determined to be 5 mg/ml with an OD600 of 0.9 (Fig. 1d) Compounds known to induce MT protected worms against Aß-and α-synuclein toxicity and prolonged lifespan. For the assay evaluation procedure, we hypothesized that the efficacy of compounds in prolonging the time of MT induction in strains CL2120 or CL2659 will correlate with a reduction in Aß and α-synuclein toxicity burden. MTs are mainly induced by heavy metals like Zn, Cu, Cd, Hg and others. Constantinidis and Burnet hypothesized that supplementation with zinc could prevent or delay the onset of dementia 35,36 . Many human trials were undertaken e.g. Constantinidis reported about improved memory, understanding, communication and social interaction of AD patients aged between 56 and 86 years when Zn was administered 11 . In our experiment with the transgenic strain CL2659 the supplementation of growth medium with ZnSO 4 was able to significantly increase  37 . Apomorphine is used in clinics for the therapy of PD. It has pleiotropic biological functions because it is antioxidative and upregulates NGF synthesis in cultured mouse astrocytes 38 . Another study reported that apomorphine stimulates degradation of intracellular Aß in a mouse model of AD 39 . The compound also exerts protective effects on neurons mainly via up-regulation of MT 16 . 10 µM apomorphin increased MT in our assay with strain CL2120 too and protected worms from proteotoxicity in both Aß (Fig. 2b: CL2659) and α-synuclein ( Fig. 2b: NL5901) assay. Furthermore, the same concentration even prolonged the lifespan in the wild type strain ( Fig. 2b: N2).  www.nature.com/scientificreports/ Glucocorticoids like progesterone are able to induce MT expression 40 . Pre-treatment with the synthetic glucocorticoid analogue dexamethasone (DXM) suppressed the formation of α-synuclein cytoplasmic aggregates in neuroblastoma cells after incubation with copper 27 . In our experiments, 30 µM and 3 µM DXM prolonged MT induction (Fig. 2c: CL2120), reduced proteotoxic burden in strains CL2659 (Fig. 2c: CL2659) and NL5901 (Fig. 2c: NL5901) and 300 µM and 30 µM prolonged the lifespan in the wild type strain N2 (Fig. 2c: N2).
The flavonoid quercetin (QC) was able to induce MT in hepatoma cells and protected them against oxidative stress at a concentration of 10 µM 41 . In C. elegans quercetin prolonged the mean lifespan by 15% by increasing stress resistance 42 . Further studies showed that quercetin ameliorated Alzheimer's disease pathology and protected cognitive and emotional function in aged triple transgenic Alzheimer's disease model mice. Extracellular β-amyloidosis, tauopathy, astrogliosis and microgliosis in the hippocampus and the amygdala have been decreased after treatment with quercetin 43 . QC at 33 µM prolonged timespan in which MT content was elevated (Fig. 2d: CL2120) and reduced proteotoxicity in our assays with strains CL2659 (Fig. 2d: CL2659) and NL5901 (Fig. 2d: NL5901). Further the same concentration significantly prolonged lifespan in the wild type strain N2 (Fig. 2d: N2). However, bioactivity of QC in the paralysis assay with CL2659 has not been lost after knockdown of MT with RNA interference (Fig. 3c: CL2659: MT knockdown). One explanation for this could be that QC mediated lifespan extension in C. elegans is modulated by age-1, daf-2, sek-1 and unc-43 44 . There has been shown that daf-2 and age-1 are not only responsible for longevity of C. elegans but further provide heavy metal resistance 45 . Therefore, we assume that MT induction is not the solely mode of action of QC in AD and PD transgenic nematodes.

Novel neuroactive lead compounds decreased proteotoxicity of Aß and α-synuclein by prolonging of Mt induction in C. elegans.
After assay evaluation, we investigated if the mode of action of thioflavin T (Th T), clioquinol (CQL), sesamin and emodin against Aß toxicity is based on prolonging the time of MT release. Alavez and colleagues showed that the amyloid binding compound Th T was able to maintain protein homeostasis during aging and extended lifespan and suppressed human Aß associated toxicity in C. elegans models depending on the protein homeostasis network regulator heat shock factor 1 (HSF-1), the stress resistance and longevity transcription factor SKN-1, molecular chaperones, autophagy and proteosomal functions 46 . Further treatment with Th T prevented Aß fibrillation in double transgenic AD mice 47 . In our study, Th T prolonged MT induction in both strains CL2120 (Fig. 3a: CL2120) and CL2659 (Fig. 3a: CL2659-MT) at concentrations between 1 µM or 10 µM and 100 µM. We could show for the first time that in the Parkinson assay 100 µM Th T was able to reduce α-synuclein ( Fig. 3a: NL5901). In the paralysis assay performed with strain CL2659 (Fig. 3a: CL2659), 50 and 10 µM Th T prolonged the time until the paralysis phenotype indicating a reduction in Aß toxicity. Interestingly 10 µM Th T prolonged whereas 100 µM shortened the lifespan in the wild type strain N2 (Fig. 3a: N2).
CQL is the prototype of the novel drug PBT2, which is effective in phase 2 clinical trials for AD and HD 48 . CQL has been shown to be neuroprotective, by decreasing brain aggregate load and restored reduced insulin levels in R6/2 HD mice 49 . A double-blind phase 2 clinical trial demonstrated the efficacy of clioquinol treatment in producing effects on plasma Aβ and zinc ion (Zn 2+ ) levels. The drug was well tolerated and inhibited cognitive decline in patients who, untreated, otherwise experienced deterioration 50 . PBT2 delayed the onset of paralysis in a C. elegans model of PolyQ overexpression 51 . In our study 10-100 µM CQL induced MT release in strains CL2120 (Fig. 3b: CL2120) and CL2659 (Fig. 3b: CL2659). At 100 µM, CQL reduced α-synuclein fluorescence in strain NL5901 (Fig. 3b: NL5901). At 100 µM and 50 µM CQL significantly prolonged the time until paralysis in strain CL2659 (Fig. 3b: CL2659). Neither 100 µM nor 10 µM CQL were able to prolong the lifespan in the wild type strain N2. Both concentrations even shortened it (Fig. 3b: N2). We studied the gene expression of metallothionein-1 (mt-1) and -2 (mt-2) after the CQL treatment in strain CL2120. The mt-1 and mt-2 expression remained unchanged in treated worms showing that the increased MT-2 protein level observed in CL2120 strain (Fig. 3b) must be a result of post-transcriptional regulation (Fig. 4a, b).
Emodin is a bioactive antraquinone present in some prescriptions of traditional Chinese medicine for cerebral protection activity. Cognitive deficits of hyperhomocysteinemia rats were improved by emodin. The animals had better behavioral performances, so that in the hippocampi the neuron loss decreased and synapse-related proteins increased. Further Aß overproduction and tau hyperphosphorylation were eliminated by emodin 52 . Emodin inhibited zinc-induced neurotoxicity in neuroblastoma SH-SY5Y cells 53 . Further Aloe-emodin has been shown to be neuroprotective to NMDA-treated retinal ganglion cells by Cu-Zn superoxide dismutase 54 . Anthraquinone-2-sulfonic acid prevents death of primary neurons by mechanisms like caspase inhibition and AKT activation. This compound may be a lead to develop a novel neurotherapeutic antraquinone-based drug 55 . In our study, emodin induced and prolonged MT release at 400 µM, 74 µM and 37 µM in CL2120 (Fig. 3c: CL2120) and CL2659 (Fig. 3c: CL2659-MT). α-synuclein expression was decreased with 37 µM emodin (Fig. 3c: NL5901) and when treated with 74 µM and 37 µM emodin the time until Aß expression-induced paralysis was prolonged in strain CL2659 (Fig. 3c: CL2659). At the highest dose (400 µM), emodin induced MT (Fig. 3c: CL2120), but was toxic in both CL2659 paralysis assay (Fig. 3c: CL2659) and NL5901 based Parkinson assay (Fig. 3c: NL5901). Bioactivity of 37 µM emodin in the paralysis assay has been lost when MT 2 has been knocked down with RNA interference and has been reduced when MT 1 has been knocked down (Fig. 3c: CL2659 with MT knockdown). Further 37 µM emodin prolonged lifespan in the wild type strain N2 (Fig. 3c: N2). We also studied the mt-1 and mt-2 expression in emodin treated CL2120 strain (Fig. 4a, b). We found a modest increase in mt-1 expression (Fig. 4a). The mt-1 expression showed an increasing trend in three and five day treated worms, while a reverse trend was observed for mt-2 (Fig. 4c). Our observations suggest that mt-1 and mt-2 may act differently to regulate the metallothionein levels in worms. www.nature.com/scientificreports/ The test compound sesamin has been shown previously to act protective against Aß toxicity and to extend the lifespan in C. elegans 55 . We wanted to clarify if these effects were due to an increase of MT. Only a high dosage (300 µM) was able to induce MT expression (Fig. 3d: CL2120), but resulted in a significant toxicity in the PD assay (Fig. 3d: NL5901) and in the paralysis assay ( Fig. 3d: CL2659). As reported before 55 56 µM and 28 µM sesamin were protective against Aß toxicity (Fig. 1d: CL2659) in our study, but α-synuclein fluorescence was unaltered (Fig. 3d: NL5901). Therefore, we assume that sesamin is protective through another mechanism than MT activation.
The paralysis assays performed with strain CL2659 showed that knockdown of either MT-1 or -2, alone, did not worsen the paralysis phenotype (Fig. 3c). Furthermore, quercetin (33 µM) treatment rescued the paralysis phenotype after MT-1 and MT-2 knockdown in CL2659 strain. Emodin (37 µM) mediated rescue was partly affected after MT-1 knockdown and was completely abolished after MT-2 knockdown in CL2659 strain. These observations indicate that MT-1 and MT-2 may show functional redundancy, however further studies are required to evaluate this hypothesis. Also, pharmacological compounds may act differently to ameliorate the paralysis symptoms, and precise mechanism of actions will be addressed in future work.  . Relative expression of mt-1 (a) and mt-2 (b) following treatment with test compounds. 150 worms of CL2120 strain were treated with 40 µM emodin, 100 µg/ml clioquinol and 100 µM ZnSO 4 . Relative fold change was measured using comparative ΔΔCT method. Gene expression data was normalized to rps-18. (c) 150 worms of CL2120 strain were treated with 40 µM emodin for a period of 3 and 5 days and quantitative RT PCR was performed to measure relative mt-1 and mt-2 expression. One-way ANOVA followed by Tukey's multiple comparisons test was used to measure statistical significance (*p ≤ 0.01). All experiments were performed in triplicates. Data represents means ± SEM.

Methods
Maintenance of C. elegans. Caenorhabditis elegans were maintained according to the protocol of the CGC (Caenorhabditis Genetics Center), University of Minnesota, Minneapolis, MN 55455 USA. Cultivation and preparation of media and agar plates were done according the protocol of Stiernagle 56 . All worms are kept at 16 °C. C. elegans were growing on plates containing nematode growth medium (NGM) seeded with Escherichia coli strain OP50 (CGC) as described 57 . Large numbers of developmentally synchronized worms grew in solid culture and were harvested at L3 stage (CL2659) or L4 stage (CL2120, CL2122, NL5901). They were suspended at defined density in 96 well plates containing SOF medium (S-medium with OP50 and 5-fluorodeoxyuridine (FUdR)) and test compounds 58,59 . In vivo fluorescence was measured with a fluorescence microplate readerphenotypical changes like paralysis with a stereo light microscope. In order to maintain an age-synchronized population we used the egg prep method from the CGC. To prevent the population from producing progeny, 5-fluorodeoxyuridine was used. In our procedure, a synchronized population is exposed to 60 µM FUdR just as it reaches sexual maturity 60 .
Measuring the depletion of E. coli food source. OP50 were diluted 1:1 in liquid broth (LB) in a 96 well plate starting with a concentration of 6 mg/ml in triplicates and absorbance was measured at 600 nm (OD 600 ) 34 (Fig. 1d).

Determination of optimal worm density and MT expression pattern.
To follow protein expression in C. elegans, we used the transgenic strain CL2120, where MT is tagged to GFP and Aß is expressed. As control strain CL2122 was used. Real-time fluorescence intensity was measured with a fluorescence multiwell plate reader. For high throughput assays, the protocol of Leung et al. 59 was used, where defined densities of developmentally synchronized fluorescent worms were added to 96-well plates (Fig. 1c). To determine the optimal worm concentration we serially diluted CL2120 L4 1:1 in a 96 well plate starting with a concentration of 256 worms per well and measured the fluorescence at 450/535 nm daily.
MT assay with strain CL2120. 50 µl SOF medium were added to each well of a 96 well plate. 10 µl of compounds dissolved in 1% DMSO were added in triplicates in different concentrations. L4 larvae were harvested from the NGM agar plates and suspended in SOF medium. 40 µl of worm suspension were added to each well of the 96 well plates at the concentration of 30-50 worms/well. MT expression was followed by measuring GFP at d0, d4, d8 by the fluorescent multiwell plate reader at 450/535 nm. OD 600 was measured on d0 and d3 and d6 to observe depletion of food source. OP50 were added after 3 days at the time after the food source was usually depleted.
Assaying α-synuclein toxicity with strain NL5901. The handling for the assays of alpha-synuclein toxicity followed the same protocol as used in the MT assay. α-Synuclein expression was followed by measuring GFP at d0, d3 and d5 by the fluorescent multiwell plate reader at 450/535 nm. OD 600 was measured on d0 and d3 to observe depletion of food source. OP50 were added after 3 days because this was the time after which food source was usually depleted.
Assaying Aß toxicity CL2659: paralysis assay. In this strain Aβ expression can be induced by temperature upshift in muscle cells 59 . A correlation between the increase of neurotransmission and progression of paralysis has been reported previously 62 . This strain has wild-type movement at the permissive temperature of 16 °C but becomes paralysed upon temperature upshift to 25 °C within approximately 48 h in liquid culture. Treatments that inhibit Aβ toxicity in this model (e.g. exposure to Ginkgo biloba extracts 65 alter the rate of paralysis in these worms. In accordance to the work of Dostal et al. 63 with some alterations, we used the screening protocol for measuring the rate of paralysis. The inducible Aβ expression does not lead to amyloid deposits and the paralysis phenotype appears independent of amyloid deposition 64 . Therefore, the acute toxicity of induced Aβ expression resulting from the accumulation of soluble oligomeric Aβ can be measured. Simultaneously, real time MT expression was followed by GFP fluorescence. For a high throughput screening method 50 µl SO medium were added to each well of a 96 well plate. 10 µl of compounds dissolved in 1% DMSO were added in triplicates in different concentrations. L3 larvae were harvested from NGM plates and suspended in SO medium. 40 µl of worm suspension were added to each well of the 96 well plates in the concentration of 10-20 worms/well. Aβ transgene expression in muscle cells was induced by temperature upshift from 16 to 25 °C and lasts until the end of the paralysis assay. Usually on d0 (before temperature upshift) and d2 (48 h after temperature upshift) the number of paralysed worms was scored under the dissecting microscope. The percentage of non-paralysed worms on d0 and d2 is shown in a bar graph using MS-Excel 2010. www.nature.com/scientificreports/ MT induction in CL2659-MT. Fluorescence of GFP-tagged MT expressing worms in the paralysis assay was measured on d0 and d2 by the fluorescent multiwell plate reader at 450/535 nm and MS-Excel 2010 was used for all calculations and plotting of data. The changes in fluorescence between wells with worms treated with the test compound and vehicle control on two different days were analysed using two tailed Student's t-test (n = 10-20/well).
Lifespan assay with N2. The lifespan assays were performed with some alterations according the protocol of Solis and Petraschek 65 . L4 worms were washed from the agar plates, pelleted and transferred to a 96 well plate containing S-medium with OP50, FUdR and compounds in triplicates. Vehicle control contained 1% DMSO. Immediately number of worms were counted per well. Usually we used 10-20 worms/well. Counting after transferring worms to the 96-well plate marked time point 0. Counting of living worms was repeated after 18 days. After 6 days of transferring bacterial food source OP50 were added. After 18 days percentage of compound treated living worms were compared to worms from the vehicle control.
MT 2 and MT 1 knockdown with RNA interference. Induction of RNA interference by feeding was performed according the protocol of Conte et al. 67 . Clones carried in E. coli HT115(DE3) were purchased from GE Dharmacon (www.horiz ondis cover y.com) . Clone Id for MT 1 is: K11G9.6 ORF and clone Id for MT 2 is T08G5.10 ORF. Synchronized L1 larvae of CL2659 were transferred to agar plates inoculated with the E. coli strain carrying the double stranded RNA for RNA interference. L3 larvae were harvested and paralysis assay was performed as usual.
Gene expression analysis. Total RNA was extracted using TRIzol reagent 68 (ThermoFisher scientific, Austria). Quantitative RT PCR was performed using the iTaq Universal SYBR Green Supermix (Bio-Rad Laboratories G.m.b.H., Austria) as per manufacturer's instructions. Complementary DNA (cDNA) was synthesized with the ProtoScript II First Strand cDNA Synthesis Kit (New England Biolabs GmbH, Germany) using primers described by Chiang et al. 31 . Gene expression was standardized against rps-18 and comparative ΔΔCT method was used to measure the relative gene expression changes. All the experiments were performed in triplicates.
Statistical evaluation of paralysis. Raw data of the paralysis assay were analysed in Excel to keep track of paralysed populations in each well. For each well the coordinates in the plate, strain, drug and the total number of animals paralysed on day 0 (d0) and day2 (d2) were recorded. For generation of bar graphs, the median fraction of non-paralysed worms was given as percentage at d0 and d2. Compound treated and vehicle treated worms were compared using two tailed Student's-test (n = 10-20/well).

Statistical evaluation of gene expression.
For gene expression analysis, Graphpad Prism 8 was used to perform One-way ANOVA followed by Tukey's multiple comparisons test.
Fluorescence analysis. The fluorescence intensity of each well was measured with a microplate reader with the appropriate emission and excitation wavelength (Filter for our assay: GFP 450/20ex 535/20ex). The difference of fluorescence increase or decrease between compound treated wells and vehicle control on two different days was analysed using two tailed Student's t-test (n = 30-50/well).

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.