The following primary antibodies were used in this study: Myc monoclonal antibody (M4439, Sigma-Aldrich, Merk KGaA, Darmstadt, Germany, 1:1000), β-actin antibody (A5441, Sigma-Aldrich, Merk KGaA, 1:5000), TARDBP (190782-2-AP, Proteintech Europe Manchester, UK, 1:1000), histone H4 (SAB4500313, Merk KGaA, 1:2000), caspase-3 (9665, Cell Signaling Technology, Danvers, Massachusetts, USA 1:1000), LC-3B (2775, Cell Signaling Technology, 1:1000), p62 (GTX100685, GeneTex Inc, Irvine, California USA, 1:500), anti-rabbit peroxidase-conjugated secondary antibody (AP132P, Merk KGaA, 1:3000) and anti-mouse peroxidase-conjugated secondary antibody (AP124P, Merk KGaA, 1:3000); anti-rabbit, anti-mouse Alexa 488 (A-11001, Thermo Fisher Scientific, Waltham, Massachusetts, USA, 1:1000) or 647-conjugated secondary antibody (A-21244, Thermo Fisher Scientific, 1:1000). All antibodies were used at the dilution recommended by the manufacturer’s instructions.

L-BMAA, (+)-L-β-N-Methyl-αβ-diaminopropionic acid hydrochloride (B-107, Sigma-Aldrich, Merk KGaA), L-Alanine (05129, Sigma-Aldrich, Merk KGaA), L-Serine (S4500, Sigma-Aldrich, Merk KGaA), Puromycin (P8833, Sigma-Aldrich, Merk KGaA).

SH-SY5Y neuroblastoma cells (CRL-2266, ATCC, Rockville, MD) and SH-SY5Y-pCHOP cells were cultured in DMEN/F12 (Thermo Fisher Scientific) supplemented with 10% fetal calf serum (FCS, Thermo Fisher Scientific). The SH-SY5Y-pCHOP cell line was generated by stable transfection of the plasmid expressing zsGreen under the control of the ER stress-responsive promoter of DNA damage-inducible transcript 3, also known as the C/EBP homologous protein (CHOP) gene. The plasmid construct was generated by cloning the CHOP (Gene ID: 1649) promoter region from -954 to +91 into the SacI and HindIII sites of the pZsGreen1-1 plasmid (Takara Bio Inc, Kusatsu, Japan). Individual clones were selected using 400 µg/ml of G1418 (Gibco, Thermo Fisher Scientific) and analyzed for their response to ER stress.

Human Embryonic Kidney (HEK) 293T cell line (CRL 3216 HEK 293T ATCC, Rockville, MD) and primary cutaneous fibroblasts from healthy individuals, from patients with sporadic ALS (carrying a heterozygous 1144G-A transition in exon 6 of the TARDBP gene, resulting in an A382T substitution in TDP-43-encoding gene) and patients with familiar ALS without mutations in the most common ALS- related genes (48) were cultured in DMEM (Thermo Fisher Scientific) supplemented with 10% foetal calf serum. All cells were grown in an incubator at 37°C, with a humidified atmosphere containing 5% CO2. Trypsin (0.5 µg/ml, 68 mM EDTA, Thermo Fisher Scientific) was added to splitting cells and then diluted in fresh medium.

Expression and the transduction protocol were previously described (49). SH-SY5Y were plated at 40% of confluence and, the day after, were transduced by recombinant adenovirus for 1 h in a serum-free medium. After 1 h, the medium was replaced by a medium containing 1% serum. At the indicated time points, the cells were washed twice by cold PBS 1X and lysed in Laemmli buffer 1X.

Microcystis auroginosa isolated from artificial lake Coghinas (Sassari, Italy), were cultured at adequate temperature and light condition (21 ± 1°C with a photoperiod of 16:8 light:dark cycle and irradiance of 50–70 mol photons m-2 s-1). Cell abundance was determined in subsamples (5–10 mL) of fixed samples using an inverted microscope (Axiovert 25, Zeiss, Oberkochen, Germany). 1x 10⁶/ml cells were centrifugated and washed with cold PBS 1X than the obtained pellet was homogenized and sonicated in 1 ml of cold water according to (50–52) After 3 cycles of freeze and thaw the supernatant was collected and the protein quantification of crude extract was measured with Thermo Scientific NanoDrop 2000 Spectrophotometer. The optical density of the sample at 730 nm was used to determine cell destruction.

Cells were plated at 60% confluence. After 16 h growth medium was removed, cells were washed with PBS1x and the indicated treatment was added in MEM (Thermo Fisher Scientific) supplemented with 1% FBS and incubated for different exposure times.

Cell viability was determined by an MTS assay (CellTiter 96 Aqueous One Solution Reagent, Promega, Madison, Wisconsin, USA) according to the manufacturer’s instructions. Absorbance at 490 nm was measured in a multilabel counter (Victor X5, PerkinElmer, Waltham, Massachusetts, USA).

Measurement of produced intracellular reactive oxygen species was determinate with a commercial kit (DCFDA - Cellular ROS Assay Kit/Reactive Oxygen Species Assay Kit, ab113851, Abcam, Cambridge, UK) according to the manufacturer’s protocol. DCFDA is deacetylated by cellular esterases to a non-fluorescent compound, which is later oxidized by ROS into 2’,7’–dichlorofluorescein (DCF). DCF was detected by fluorescence spectroscopy with excitation/emission at 485 nm/535 nm in a multilabel counter (Victor X5, PerkinElmer).

Stable cell line SH-SY5Y-pCHOP, expressing zsGreen under the control of CHOP promoter were plated 1*10^⁵ in a 24-multiwell plate and transduced or not with adenoviral particles coding for TDP43^{WT/M337V/A382T} vectors. 24 hours after infection cells were treated with LCS-MaCe total extract or L- BMAA in base growth media Dulbecco MEM/F12 medium containing 10% FBS without phenol red. Following 24h of treatments, GFP fluorescence was detected with excitation/emission at 485 nm/535 nm of 1s in a multi-reader plate (Victor X5, PerkinElmer).

Protein content was determined using Bradford protein assay (27813, Sigma-Aldrich, Merk KGaA). Uniform amounts of protein extracts were loaded by standard SDS/PAGE. Samples were then electroblotted on Pierce Nitrocellulose Membrane, 0.45 µm (Thermo Fisher Scientific). Then, membranes were incubated in a blocking solution of 3% low-fat milk, diluted in PBS 1X-Tween 0.05% solution with the indicated primary antibody for 16 h at 4°C. Secondary antibody were used to reveal immunocomplexes using LiteUP WB Chemiluminescent Substrate (EMP002005, Euroclone, Hercules, California, USA). Chemiluminescent imaging was obtained using ChemiDoc XRS+ System (Bio-Rad Laboratories, Hercules, California, USA). The apparent molecular weight of proteins was determined by calibrating the blots with pre-stained molecular weight markers (Bio-Rad Laboratories).

Treated cells were washed twice with cold PBS 1X, lysed in cold RIPA buffer (50 mM Tris- HCl, pH 8, 150 mM NaCl, 1% NP-40, 0.1% SDS, 0.5 mM sodiumdeoxycholate) supplemented with 1x Protease inhibitor cocktail (Complete™, Mini, EDTA-free Protease Inhibitor Cocktail, Sigma-Aldrich, Merk KGaA), a mixture of phosphatase inhibitors, 1 mM NaF, 1 mM Na₃VO₄, (Sigma-Aldrich, Merk KGaA) and sonicated. Lysates were centrifuged (twice to prevent contamination caused by carrying over) for 30’ at 100,000*g at 4°C, the pellets were re-sonicated and re-centrifuged at 100,000*g for 30’ at 4°C and the supernatant was collected as RIPA buffer soluble fraction. RIPA buffer-insoluble pellets were dissolved in urea buffer (7 M urea, 2 M thiourea, 4% CHAPS, 30 mM Tris- HCl, pH 8.5) and sonicated. Soluble and insoluble fractions were subsequently analyzed by immunoblotting.

Cells were plated at 2*10^5 in 24-multiwell plates on 12mm glass coverslips, fixed with 4% paraformaldehyde in PBS 1X and permeabilized with 0.2% Triton X-100 in PBS 1X for 15’. After a blocking step of 1 h in 5% BSA, diluted in PBS 1X–0.05% Tween-20, cells were incubated with the indicated primary antibody diluted in blocking solution, overnight at 4°C, and then incubated with the proper secondary antibody, diluted in blocking solution, for 1h at RT; DAPI (D9542, Sigma Aldrich, Merk KGaA), fluorescent staining of DNA content and nuclei, was added diluted 1∶1000 in blocking solution. Cells were then analyzed with a Leica TCS SP5 confocal microscopy, with LAS (Leica Application Suite) lite 170 image software (Advance Fluorescence 2.7.3.9723). Using the 3D viewer in LAS Lite 170, we visualize and analyze three-dimensional datasets acquired through confocal microscopy.

Cells treated as indicated were diluted in cold 1XPBS, centrifuged and resuspended in 1XBinding buffer (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl2, pH 7.4) at the concentration of 10^6cells/mL. Subsequently, Annexin V-FITC staining solution and Propidium iodine (PI) staining solution were added at the final concentration 20 μg/mL. Cells were then analyzed using a flow cytometer FACSCanto (Becton Dickinson, Franklin Lakes, NJ, USA). Flow cytometer acquired 3x10⁴ total events. Data analysis was carried out using BD FACSDivaTM software 6.1.3.

The results are presented as means ± SD of independent experiments as indicated. For bands analysis in Western blot experiments, after image acquisition, the protein bands were quantified by densitometry and normalized to the specific loading control using Quantity One software (Version 4.6.8, Bio-Rad Laboratories). Statistical evaluation was conducted by one-way ANOVA and Bonferroni’s multiple comparison post-test. Values significantly different from the relative control are indicated with *, **, or *** symbols when p < 0.05, p < 0.01, and p < 0.001, respectively.

The causal relationship between neuronal degeneration and exposure to cyanotoxins, such as L-BMAA, is still quite controversial. To establish a reliable neurotoxicity assessment model, we opted to employ the human-derived neuroblastoma cell line SH-SY5Y. These cells, derived from a human metastatic neuroblastoma, have dopaminergic, cholinergic, glutamatergic and adenosinergic properties and are widely used as a model to study NDs (53). As mentioned in the introduction toxicity of cyanobacteria can be influenced by the mode of exposure; To mimic natural exposure to neurotoxins, we utilized both pure synthetic L-BMAA and crude cellular extracts obtained from a strain of Microcystis aeruginosa isolated from Coghinas Lake (LCS-MaCe, in which L-BMAA was measured, maximum 17.84 μg L-1 (54),), mimicking natural exposure to neurotoxins. LCS-MaCe was obtained using the freeze-thaw method, that is considered a standard method to lyse cells and it is recommended by U.S. Environmental Protection Agency (EPA) Method 546 (50, 51).

In preliminary experiments testing different protocols, doses, and exposure times for L-BMAA (data not shown) we observed a reduction in cell viability when SH-SY5Y cells were exposed to increasing concentrations of L-BMAA or LCS-MaCe and tested by MTS assays ( Figure 1A ). Both treatments induce a significant reduction in cell viability, which can be rescued by adding to cell medium the aminoacids L-serine or alanine ( Figure 1B ).

To discern the cause of the reduction in the observed cell viability we evaluated oxidative or ER stress.

Exposure to the pure toxin L-BMAA or to LCS-MaCe caused a significant increase in oxidative stress in SH-SY5Y cells, as assessed by measuring both cellular ROS levels ( Figure 1C ) and transcriptional activation of the antioxidant responsive element (ARE), a key regulatory element of many cellular defense enzymes, that drives the luciferase reporter expression ( Figure 1D ).

ER stress was assessed evaluating gene activation of C/EBP homologous protein (CHOP), one of the components of the ER-stress-mediated apoptosis pathway. Under physiological conditions, CHOP is expressed at low levels and localizes in the cytoplasm; however, under stress conditions, an up-regulation of CHOP mRNA levels and its protein accumulation are observed (55, 56). To monitor CHOP activation, we took advantage of a stable cell line SH-SY5Y-pCHOP expressing zsGreen (Zoanthus sp. human codon-optimized green fluorescent protein 1) under the control of the human CHOP promoter, previously generated in our laboratory.

GFP autofluorescence corresponding to CHOP transcriptional activation, was measured after L-BMAA or LCS-MaCe exposure by both fluorescence measurement by multi plate reader ( Figure 1E ) and autofluorescence analysis by confocal microscope ( Figure 1F ). Notably, the observed increase in GFP autofluorescence is comparable to the intensity induced by Sodium Arsenite, a known inducer of acute stress (57).

Since both oxidative and ER stress can induce cell death, a flow cytometry analysis for necrosis, apoptosis was performed labeling the cells by propidium iodide (PI) and Annexin V. The results indicated that neither apoptosis nor necrosis appeared to be the primary determinant of the effects of L-BMAA or LCS-MaCe ( Figures 2A, B ). These data were confirmed by Western blot ( Figure 2C ) where no Caspase-3 cleavage was detected by the different treatment compared to staurosporin (STS) treatment used as positive control. On the contrary, when we evaluated two autophagic markers, LC3B and p62, we were able to highlight a specific induction by L-BMAA or LCS-MaCe treatments ( Figures 2D–H ), indicating an impairment of the autophagic flux (58).

Since autophagy is a catabolic process for unnecessary or dysfunctional cytoplasmic contents mediated by lysosomal degradation pathways and autophagy mediates both TDP43 (59) and α-synuclein turnover (60), two of the major components of protein aggregates in degenerating neurons respectively in ALS and PD, we evaluated the effects of L-BMAA or LCS-MaCe on the protein levels of TDP-43 and α-synuclein. Exposure to cyanotoxins increases the endogenous TDP-43 protein level, and at high exposure times we were able to also demonstrate an accumulation of high molecular weight TDP-43 ( Figures 3A, B ). Furthermore, confocal analysis in SH-SY5Y cells revealed a moderate cytoplasmic TDP-43 relocalization ( Figure 2E , left column), another typical sign of TDP43 proteinopathy.

To evaluate the effects of L-BMAA or LCS-MaCe on α-synuclein, since this protein is barely detectable in SH-SY5Y cells, we overexpressed it by adenoviral transduction (61). Western blot analysis shows a significant accumulation of α-synuclein ( Figures 2C, D ) upon L-BMAA or LCS-MaCe exposure. To assess whether the observed α-synuclein accumulation was due to a defect in protein turnover, we performed a treatment with puromycin, a drug that inhibits translational elongation, in presence or absence of L-BMAA. Cells were treated or not with L-BMAA for 24h and then with puromycin for the indicated time. Notably, cells treated with L-BMAA and puromycin exhibited a significant reduction in α-synuclein turnover compared to the cells untreated by L-BMAA ( Figures 2F, G ). At 6 hours in L-BMAA untreated cells more than 50% of α-synuclein has been degraded while no differences are detectable in L-BMAA treated cells ( Figure 2G ). Conversely, the expression of β-actin remained unaffected by the treatments, serving as a control for protein expression in the cells. Finally, accumulation of α-synuclein accumulation was confirmed by confocal analysis ( Figure 2E -right column).

In summary, our findings indicate that exposure to L-BMAA or LCS-MaCe impacts autophagic pathways and replicates various molecular features typical of neurodegeneration, including the accumulation of TDP43 and α-synuclein.

Considering that ALS is generally thought to progress as a consequence of genetic susceptibility and environmental influences, and that TDP-43 has been observed as the major component of ubiquitinated inclusions in post-mortem tissues of ALS patients and patients with frontotemporal dementia (41), we decided to extend our analysis by evaluating the effects of L-BMAA or LSC-MaCe exposure in different ALS cellular models (49).

SH-SY5Y cells were transduced with adenoviral particles encoding WT or pathological TDP-43 mutant (A382T or M337V) and then exposed to varying doses of L-BMAA neurotoxin or LSC-MaCe. Confirming previous findings (49), a reduction in cell viability was observed following adenoviral delivery of myc-tagged mutant TDP-43 in SH-SY5Y neuronal cells ( Figure 4A ). Simultaneous exposure to increasing doses of L-BMAA or LSC-MaCe exacerbated the toxicity induced by both pathological variant A382T and M337V ( Figure 3A ). Analysis of transduced and basal TDP-43 expression through Western blot revealed characteristic bands, with higher molecular weights indicative of the 5X-Myc tag and lower bands suggesting degradation ( Figure 4B ). Interestingly, L-BMAA treatment determines the formation of TDP 43 degradation products for both transduced (asterisks in Figure 4B ) and endogenous TDP43 ( Figure 4B high exposure). Further characterization of the L-BMAA effect on TDP43 proteinopathy involved biochemical fractionation, highlighting a significant increase in insoluble TDP-43, a well-known marker of ALS ( Figures 4C, D ). Since TDP-43 is known to shuttle between the nucleus and cytoplasm and to form cytoplasmic protein aggregates under pathological conditions, we performed immunostaining using anti-myc tag antibody that only labels exogenous TDP-43. As shown in Figure 4E and quantified in Figure 4F , exposure to both pure L-BMAA or LCS-MaCe induces a stronger delocalization of mutant TDP-43 to the cytoplasm. Finally, we observe an exacerbation of both oxidative ( Figure 4G ) and ER stress ( Figures 4H, I ) in the presence of both mutant TDP43 and cyanotoxins.

Collectively our experimental data strongly suggest that in cellular ALS experimental models, exposure to L-BMAA or crude extracts from cyanobacteria can induce cellular stress comparable to the expression of pathological TDP-43 variants, ultimately exacerbating their phenotype in cellular ALS models.

To delve deeper into the influence of cyanotoxins on cellular functions, we used fibroblasts from both ALS patients and healthy controls. These fibroblasts from ALS patients provide a valuable platform for studying environmental triggers in a genetic background more closely resembling sporadic ALS (sALS). Additionally, fibroblasts exhibit characteristics relevant to TDP-43 protein metabolism and abnormal accumulation (62, 63). In light of these considerations, primary cutaneous fibroblasts from healthy donors, from patients with sporadic ALS (carrying a heterozygous 1144G-A transition in exon 6 of the TARDBP gene, resulting in an A382T substitution in the TDP-43 encoding gene) or from patients with familial ALS without mutations in the most common ALS-related genes were studied to elucidate the mechanisms of L-BMAA toxicity (64). Initial assessments of cell viability, both at low and high concentrations (1mM or 3mM), of L-BMAA or LCS-MaCe treatments after 24 hours, did not yield statistically significant decreases, although a noticeable trend was observed (data not shown). Subsequently, a time-course experiment revealed a statistically significant reduction in viability, particularly in fibroblasts from familial ALS patients carrying the mentioned TARDBP gene transition after 72h of treatment. ( Figure 5A ).

Furthermore, an escalation in oxidative stress was measured upon L-BMAA or LCS-MaCe exposure ( Figure 5B ), in all fALS and sALS fibroblasts, respect to healthy controls.

Subsequent evaluation of TDP-43 nuclear/cytoplasmic localization through immunofluorescence and confocal analysis unveiled a significant change in TDP-43 localization ( Figure 5C ). Abnormal cytoplasmic relocalization of TDP-43, predominantly affecting fibroblasts from ALS patients, was observed (highlighted with white arrows in Figure 5C , quantification in Figure 5E ). Treatment with both L-BMAA and LCS-MaCe induced features similar to those observed in TDP-43-overexpressing models: nuclear depletion and cytoplasmic delocalization. While cyanotoxin exposure impacted healthy control fibroblasts, TDP-43 delocalization and abnormal accumulation in the cytoplasm were significantly amplified in cells from ALS patients.