Experimental design

Cells were administered various doses of METH to select the optimal dose to produce the most severe alpha-syn-related pathology, while avoiding to induce extensive cell death. At this purpose, a dose–response curve showing the occurrence of cell death was compared with a dose–response curve measuring alpha-syn-related cytopathology. The chance to analyze a highly reproducible in vitro system allows to count specific protein amount within alpha-syn positive areas. The ultrastructural analysis of these alpha-syn positive cell domains following METH was developed by combining light and electron microscopy (CLEM), similarly to that recently reported by Shahmoradian et al. (2019). A preliminary approach consisted of selecting specific areas visualized at light microscopy through semi-thin sections to further proceed with electron microscopy to carry out a quantitative assessment. Semi-thin slices stained with toluidine blue provided a further analytical step to confirm the placement of those pale cytosolic areas observed at Hematoxylin & Eosin (H&E) light microscopy, which characterize METH-induced cytopathology, and corresponding to strong alpha-syn staining as confirmed by pre-embedding immuno-peroxidase. Once these areas were identified through semi-thin sections, they were further dissected to carry on ultra-thin slices for electron microscopy, where the amounts of specific antigens could be detected by immuno-gold stoichiometry. These pale eosinophilic areas represented by H&E and toluidine blue histochemistry were evidenced as encircled red-dotted line. Within these areas, which correspond to those previously immuno-stained at light microscopy, a further validation of alpha-syn content was provided by ultra-thin sections (70–90 nm) from samples undergone pre-embedding immuno-peroxidase before proceeding with the counts of immuno-gold stoichiometry of specific protein content and plain ultrastructural morphometry of membranous organelles.

Thus, the first part of the study assessed specific (alpha-syn, p62, poly-ubiquitin, ubiquitin) protein amount at light microscopy (immuno-peroxidase and immuno-fluorescence) following METH administration. The second part of the study combined light and electron microscopy (CLEM) to count protein stoichiometry at TEM within corresponding cytosolic areas possessing strong alpha-syn immuno-staining at light microscopy. In these cytosolic areas featuring a cytopathology reminiscent of inclusions, the amount of alpha-syn was compared with other proteins, such as p62 and poly-ubiquitin in situ. In fact, these latter proteins were recently claimed as early markers of catecholamine cytopathology (Kurosawa et al. 2016; Noda et al. 2022; Sato et al. 2018, 2020, 2021). This ultrastructural analysis indicates that, within alpha-syn abundant areas, p62 prevails instead. Therefore, we added a reversed sampling focusing also on those areas which expressed the highest content of p62. In both cases, areas were measured to calculate the membranous organelles (lysosomes, autophagosomes, and mitochondria) compared with protein content.

Cell cultures

Pheochromocytoma PC12 cell cultures, purchased from IRCCS San Martino Institute (Genova, Italy), were kept in a wet atmosphere with 5% CO2 at 37 °C and were grown in RPMI 1640 medium (Sigma-Aldrich, St. Louis, MO, USA), supplemented with horse serum (HS, Sigma-Aldrich), fetal bovine serum (FBS, Sigma-Aldrich), and antibiotics (streptomycin and penicillin). Experiments were carried out when PC12 cells were in the log-phase of growth, corresponding to 70% confluence (Qiao et al. 2001; Song et al. 1998). Before experimental treatments, cells were seeded according to the different experimental procedures and incubated for 24 h at 37° C in 5% CO2. In detail, for trypan blue (TB) staining, cells were seeded at a density of 104 cells/well and placed within 24-well plates in 1 mL of culture medium; for light microscopy staining procedures, 5 × 104 PC12 cells were seeded on poly-lysine coverslips and placed in 24-well plates in a final volume of 1 mL/well. For TEM, 1 × 106 cells were seeded in culture dishes in a final volume of 5 mL.

Cell treatments

A stock solution of METH (kindly gifted by Forensic Medicine, University of Pisa) 10 mM was obtained by dissolving in 1 mL of culture medium 2.3 mg of METH. Aliquots of the stock solution were diluted in the culture medium to obtain the treatment solutions. In detail, PC12 cells were exposed to increasing doses of METH, ranging from 1 up to 1000 µM, for 72 h. Control cultures were kept in the same volume of culture medium for the same time interval. This time interval was selected based on previous studies (Fornai et al. 2004; Lazzeri et al. 2018, 2021). At the end of the treatments, PC12 cells were washed in PBS and processed according to the various experimental procedures. After the pilot dose–response study, the dose of METH which was selected to analyze various antigens and ultrastructural morphometry was 100 µM for 72 h.

Hematoxylin and eosin (H&E) histochemistry

Cells were fixed in a 4% paraformaldehyde phosphate-buffered solution (PBS) for 15 min, washed in PBS, and then immersed for 15 min in the hematoxylin solution (Sigma-Aldrich). The hematoxylin staining was stopped through repeated washing in running water. After, cells were immersed within the eosin solution (Sigma-Aldrich) for a few minutes and washed out again to remove the excess of dye. Finally, after dehydration in increasing alcohol solutions, cells were clarified in xylene, covered with DPX mounting medium (Sigma-Aldrich) and observed under a Nikon Eclipse 80i light microscope (Nikon, Tokyo, Japan) equipped with a digital camera connected to the NIS Elements software for image analysis (Nikon, Tokyo, Japan).

Cell count was carried out under light microscope, using a 20 × magnification; the number of H&E-stained cells in each experimental group was counted and expressed as the mean percentage ± SEM of the control group (which corresponds to 100%). Data refer to three independent experiments.

Adobe Photoshop CS4 Extended program (version 11.0, Adobe Systems Inc., San Jose, CA, USA) was used to create the artwork.

Fluoro-Jade B (FJB) histo-fluorescence

Cells were fixed with a solution of paraformaldehyde 4% for 5 min, washed in PBS, and immediately incubated with 0.06% potassium permanganate for 10 min at room temperature. After washing in distilled water, cells were incubated for 20 min in a FJB solution prepared by dissolving 0.01% FJB (Merck Millipore, Billerica, MA, USA) in acetic acid. Cells were incubated with 0.0004% of this FJB solution for 20 min and then cover-slipped with mounting medium. FJB-positive cells were analyzed using a Nikon Eclipse 80i light microscope (Nikon, Tokyo, Japan), equipped with a florescence lamp and a digital camera connected to the NIS Elements software for image analysis (Nikon, Tokyo, Japan).

Cell count was carried out under fluorescence microscope at 20 × magnification. The number of FJB-fluorescent cells was expressed as the mean number ± SEM for each experimental group. The intensity of the fluorescent signal was measured under florescence microscopy using the software Image J (NIH, Version 1.8.0_172, Bethesda, MD, USA) and values are expressed as the mean percentage ± SEM of optical density (assuming controls as 100%) from N = 90 cells/group. Data refer to three independent experiments.

Adobe Photoshop CS4 Extended program (version 11.0, Adobe Systems Inc., San Jose, CA, USA) was used to create the artwork.

Trypan blue (TB)

Cells were collected and centrifuged at 800×g for 5 min, and the cell pellet was suspended in 0.5 mL of culture medium. Twenty-five µL of cell suspension was incubated for 10 min in a solution containing 1% TB in PBS and 10 µL of this solution were injected into a Bürker chamber and analyzed under an Olympus CKX 41 inverted microscope (Olympus Corporation, Tokyo, Japan). Viable and nonviable cells were counted and values were expressed as the mean percentage ± SEM of TB-positive cells out of the total cells. Data refer to three independent experiments.

Immuno-histochemistry

After washing in PBS, PC12 cells were fixed with 4% paraformaldehyde in PBS for 15 min, and incubated with 0.1% TritonX-100 (Sigma-Aldrich) for 15 min in PBS.

For immuno-fluorescence experiments, cells were immersed for 1 h in a blocking solution containing 10% normal goat serum (NGS) in PBS at room temperature and then were incubated overnight at 4 °C in a solution containing the primary antibodies in PBS and 1% normal goat serum. In detail, the following primary antibodies (AbI) were used: (1) anti alpha-syn AbI (Abcam, Cambridge, UK), diluted 1:100; (2) anti-p62 AbI (Abcam), diluted 1:100; (3) anti-poly-ubiquitin AbI (Abcam), diluted 1:100; (4) anti-ubiquitin AbI (Sigma-Aldrich), diluted 1:100.

After rinsing in PBS, cells were incubated for 1 h with the appropriate fluorophore-conjugated secondary antibodies (i.e., Alexa 488, Life Technologies Carlsabad, CA, USA, or Alexa 594, Life Technologies) diluted 1:200. All these reactions were carried out within the well plate. After washing in PBS, cells were transferred on coverslip, mounted with the mounting medium Fluoroshield (Sigma-Aldrich), and finally observed under the Nikon Eclipse 80i light microscope (Nikon) equipped with a fluorescent lamp and a digital camera connected to the NIS Elements Software for image analysis (Nikon). Negative control cells were incubated with secondary antibodies only. For double fluorescence pictures, single fluorescent images were acquired independently, and then, they were merged using the NIS Elements Software (Nikon).

For immuno-peroxidase experiments, cells were incubated in 3% hydrogen peroxide (H2O2) for 20 min at room temperature to block endogenous peroxidase activity, and then were plunged in a blocking solution containing 10% NGS in PBS for 1 h at room temperature. Cells were incubated overnight at 4 °C with the primary antibody solution containing 2% NGS in PBS and the following AbI: the anti-alpha-syn AbI (Abcam) (1:2000), the anti-p62 AbI (Abcam, 1:2000), the anti-poly-ubiquitin AbI (Abcam, 1:2000), and the anti-ubiquitin AbI (Sigma-Aldrich, 1:2000).

The antigen–antibody reaction was revealed using the appropriate biotin-conjugated secondary antibodies (Vector Laboratories, Burlingame, CA, USA) diluted 1:200 for 1 h at room temperature, followed by avidin–biotin complex (Vector) for 1 h and the peroxidase substrate diaminobenzidine (DAB, Vector) for a few minutes. Finally, cells were dehydrated using increasing alcohol solutions. All these reactions were carried out within the well plate. After washing in PBS, cells were clarified in xylene and transferred on coverslips where DPX mounting medium (Sigma-Aldrich) was added before their observations at light microscopy (Nikon) equipped with a digital camera connected to the NIS Elements software for image analysis (Nikon, Tokyo, Japan). Negative control cells were incubated with secondary antibodies only.

The optical density of each single fluorescent or peroxidase-stained picture was measured using Image J software (NIH, Version 1.8.0_172, Bethesda, MD, USA). Values are given as the mean percentage ± SEM from N = 90 cells/group.

In double fluorescent pictures, merging areas were measured in µm2 using Image J software (NIH, Version 1.8.0_172, Bethesda, MD, USA) and values are given as the mean merging areas ± SEM per cell from N = 90 cells/group.

All data refer to three independent experiments.

Adobe Photoshop CS4 Extended program (version 11.0, Adobe Systems Inc., San Jose, CA, USA) was used to create the artwork.

Sample preparation for semi-thin slices and combined light and electron microscopy (CLEM)

At the end of the treatment, cells were centrifuged at 1000×g for 5 min, rinsed in PBS, and fixed in a solution of 2.0% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M PBS (pH 7.4) for 90 min at 4 °C. Cells were then washed out in PBS (0.1 M), post-fixed in 1% osmium tetroxide (OsO4) for 1 h, at 4 °C, and dehydrated in increasing ethanol solutions. Finally, they were embedded in Epoxy resin.

Semi-thin slices (about 1 µm thick) were cut by ultra-microtome (Leica Microsystems, Leica Microsystems, Wetzlar, Germany), and they were stained with toluidine blue and observed at light microscopy (Nikon). These slices were ordered with ultra-thin slices (70–90 nm thick) to combine light and electron microscopy (CLEM) analysis of the corresponding areas. Ultra-thin slices (70–90 nm thick) were counterstained with uranyl acetate and lead citrate to be examined using a JEOL JEM SX100 transmission electron microscope (JEOL, Tokyo, Japan).

To provide internal validation, each semi-thin slice was followed by an ultra-thin slice (each including five series), which were processed for light and electron microscopy, respectively. This allowed to compare similar cytosolic areas according to a roughly 0.1 µm thickness interval between light and electron microscopy sampling. As a reference point, in these alternate slices, we used as a spatial reference the shape of vacuolated cytosolic domains, which were poorly stained by toluidine blue (light beam) and provided a poor contrast to the electron beam. The choice of selecting highly vacuolated cytosolic domains highly stained with alpha-syn was based on light microscopy data. In fact, the occurrence of clusters of alpha-syn was abundant in these cell regions. Immuno-peroxidase carried out post hoc at pre-embedding validated the occurrence of these antigens within ultra-thin slices.

Pre-embedding for immuno-peroxidase

At TEM, alpha-syn and p62 were also labeled by immuno-peroxidase. Cell pellets were fixed with 2.5% paraformaldehyde and 0.1 glutaraldehyde dissolved in PBS for 90 min. After washing in PBS, cell pellets were incubated in 0.002% hydrogen peroxide in 0.05 M Tris–HCl buffer, pH 7.6 for 2.5 min. Then, after washing in PBS, cell pellets were permeabilized in ethanol (10% for 5 min, 25% for 5 min, and 10% for 5 min) and pre-blocked with a solution containing 10% NGS and 0.2% saponin in PBS for 30 min. Samples were then incubated with anti-alpha-syn AbI (1:100, Abcam) or anti-p62 AbI (1:100, Abcam) diluted in 10% NGS and 0.2% saponin in PBS for 24 h. Then, samples were incubated with a solution containing the biotin-conjugated secondary antibodies (Vector) diluted 1:100 for 1 h at room temperature. After washing in PBS, samples were incubated in avidin–biotin peroxidase complex (Vector) for 1 h. After washing in PBS samples were incubated in 0.075% DAB (Vector) for a few minutes.

Samples were washed in PBS and osmicated, dehydrated, and embedded in Epoxin resin. Ultra-thin sections were cut by ultra-microtome (Leica Microsystems) and were observed under TEM (JEOL JEM SX100, JEOL, Tokyo, Japan).

Adobe Photoshop CS4 Extended program (version 11.0, Adobe Systems Inc., San Jose, CA, USA) was used to create the artwork.

Post-embedding for immuno-gold TEM

Ultra-thin slices from cell cultures were collected on nickel grids, and they were de-osmicated in aqueous solution saturated by sodium metaperiodate (NaIO4), for 15 min. Sections were washed three times for 10 min in ice-cold filtered PBS (pH 7.4) and the grids were treated with ice-cold PBS containing 10% NGS and 0.2% saponin to block non-specific antigens for 20 min at room temperature.

Primary antibodies were incubated in ice-cold PBS containing 1% NGS and 0.2% saponin in a humidified chamber overnight, at 4 °C. The following primary antibodies were used: anti-alpha-syn AbI (1:100, Abcam); anti-p62 AbI (1:100, Abcam); anti-poly-ubiquitin AbI (1:100, Abcam); anti-ubiquitin AbI (1:100, Sigma-Aldrich). Double immuno-gold staining was used to compare the amount and co-localization of the following: alpha-syn and p62, alpha-syn and poly-ubiquitin, and p62 and poly-ubiquitin.

Stoichiometry staining was obtained through a solution containing gold-conjugated secondary antibodies (gold particle diameter, 10 nm or 20 nm, BB International, Cardiff UK) diluted 1:100, in PBS containing 1% goat serum and 0.2% saponin for 1 h, at room temperature. The size of immuno-gold particles was switched to validate the non-relevance of steric encumbrance in double immuno-staining. After rinsing in PBS, grids were incubated in 1% glutaraldehyde for 3 min, and they were washed in distilled water and further stained with uranyl acetate and lead citrate. Ultra-thin sections were finally observed at TEM (JEOL JEM SX100). Control sections were incubated with secondary antibodies only.

The number of immuno-gold particles related to alpha-syn, p62 and poly-ubiquitin proteins was expressed as the mean ± SEM within 2 µm2 selected areas from n = 30 cells per group.

Area of the tubulo-vesicular membranes and the cytosol within 2 µm2 selected areas was measured using Image J software (NIH, Version 1.8.0_172, Bethesda, MD, USA) at 6000 ×. Values are given as the mean percentage ± SEM from n = 30 cells/group.

Electron density of 2 µm2 selected areas was measured using Image J software (NIH, Version 1.8.0_172, Bethesda, MD, USA) at 6000 ×. Values are given as the mean percentage ± SEM of electron density measured in METH-treated cells compared with electron density measured in control cells (assumed as 100%) from n = 30 cells/group.

Adobe Photoshop CS4 Extended program (version 11.0, Adobe Systems Inc., San Jose, CA, USA) was used to create the artwork.

Extended statistics

In this paragraph, we report classic descriptive and inferential statistics, which were implemented by an extended explanation of arbitrary criteria used here to provide the optimal sampling for each procedure. A brief comment about validation of the procedure is provided as well. The effects of various doses of METH ranging from 1 to 1000 µM on cell damage (H&E) were expressed as the mean percentage of healthy cells ± SEM compared with control. The number of degenerating cells (FJB) was expressed as the mean ± SEM of stained cells. Degenerating cells visualized at TB was expressed as the percentage of stained cells within the whole population. The difference between groups were assessed by ANOVA with Sheffe’s post hoc analysis; H0 was rejected for p < 0.05.

The issue of combining different procedures and different techniques to count cell damage at light microscopy may produce discrepancies based on the specific procedure. For instance, the amount of severe cell damage assessed by H&E is based on the actual lack of cell structures (unique among light microscopy procedures used here) and the presence of remarkable alterations of cell shape and size and faint cytosol visible as pale eosinophilic cytosolic areas. In keeping with H&E staining, the amount of cell damage included those cells, where these alterations were severe. In the case of FJB, the count is based on the fluorescent area conventionally assuming that FJB immuno-fluorescence is synonymous of dying cell (which still may not be constantly true considering our lack of an in-depth knowledge of which and how many markers are responsible for FJB-induced fluorescence). In this case, additional bias may lead to non-damage dependent occurrence of some molecules, which target FJB, thereby providing a bias. This may explain why the amount of cell death assessed using FJB was the highest compared with all other methods (still considering that this may also depend on the highest sensitivity of this procedure). Data about cell death counted at TEM were inferred by counting the decrease of viable cells compared with control. Despite slight differences, the consistency across different techniques and between light and electron microscopy was remarkable concerning the amount of METH-induced cell damage, and values were quite steady, which internally validate each single procedure applied here.

Data about immuno-fluorescence for alpha-syn, p62, poly-ubiquitin, and ubiquitin stained alone or in combination (merging) were expressed as semi-quantitative fluorescent densitometry considering the fluorescent area of merging comparing controls with METH 100 µm. The comparison was carried out using ANOVA with Sheffe’s post hoc analysis; H0 was rejected for p < 0.05. This rough calculation due to intrinsic limits of non-linear relationship between protein amount and immuno-staining serves as a guide to better address molecular quantification of these proteins at TEM. Alpha-syn positive areas were selected by counting immuno-gold particles within selected circular areas of 2 µm2. This size was selected based on sampling the distribution of immuno-gold clusters observed in alpha-syn hot spots at CLEM in METH-treated cells. These consistently overlap with pale eosinophilic areas. The amount of immuno-gold for alpha-syn within these areas was different within various cytosolic domain ranging from 0 up to 12 immuno-gold. An arbitrary cut-off was set at 7 immuno-gold alpha-syn particles which was arbitrarily selected to define at sub-cellular level a cluster of alpha-syn. Within these areas of high alpha-syn content, the amount of p62 was counted as well. In this case, we used different immuno-gold particles owing different diameter (either 10 nM or 20 nM, respectively) to distinguish both antigens in the same ultra-thin section. In parallel experiments, we assessed that the diameter of the immuno-gold particles did not affect the count of the antigen, likely due to a lack of allosteric interference between the immuno-gold particles and primary antibodies on antigen epitopes. The comparison between p62 and alpha-syn immuno-gold was carried out using ANOVA with Sheffè’s post hoc analysis. Since p62 prevails at large compared with alpha-syn even in those areas being selected as the richest in alpha-syn content, a second step was necessary to re-assess the relative amount of these proteins by starting to select areas where p62 was most abundant (cut-off 100 immuno-gold particles per 2 µm2), independently by the amount of alpha-syn. Despite some overlapping between these antigens occurs, we found that a different placement was evident at electron microscopy, which did not emerge when the localization of both antigens was detected at light microscopy. In fact, the richest p62 immuno-gold areas analyzed at TEM may not contain high alpha-syn. Again, at TEM compared with light microscopy, the occurrence of p62 was strikingly more abundant than alpha-syn. Such a difference was magnified when alpha-syn and p62 richest areas were calculated following METH administration. In these experimental conditions, these 2 µm2-wide areas were analyzed concerning their structure. In fact, within these areas, the amount of altered lipid membranous organelles was counted as well. In detail, the area taken by membrane limited organelles (autophagosomes, lysosomes and mitochondria) was calculated independently within alpha-syn and p62 rich areas. Despite some areas feature a high content of both antigens, we found that vesicles limited organelles were more abundant within areas identified by p62 immuno-gold staining compared with alpha-syn immuno-gold staining. Within these areas, the mean electron-density was calculated. In calculating the specific kind of membranous organelles, we purposefully did not discriminate between the amounts of mitochondria compared with lysosomes or autophagosomes to harvest all membranous structures as previously described (Shahmoradian et al. 2019). The nuclear area was never considered in keeping with the cytosolic nature of the pathological process under analysis. In selecting the placement, the counted area was placed around the protein cluster under primary analysis. This led to discrepant, only partially overlapping regions depending on which protein was primarily counted. Again, a discrepancy exists considering non-protein vesicle crowding. This partial overlapping indicates that the region shape and size varies depending on which structure is considered as the hallmark. The choice of the protein hallmark was done at first based on classic literature (alpha-syn), and then, it was slightly modified based on actual quantitative findings (the excess of p62). Concerning the pattern of p62, we noticed a remarkable packing where p62 and poly-ubiquitin were expressed densely. The size of these small circular areas providing a sort of p62/poly-ubiquitin domain was roughly tenfold lower compared with the circles selected for alpha-syn clusters. Therefore, a further count was used within these 200 nm2 areas to better express the density of p62 and to make a comparison with the density of alpha-syn.

From a statistical perspective and a methodological approach, we found strong discrepancies between immuno-fluorescence/immuno-peroxidase at light microscopy and immuno-gold at TEM. In detail, at light microscopy, most cells appear to stain for alpha-syn and p62 according to an all (METH) or none (control) pattern. This contrasts with TEM showing a twofold difference between controls and METH. This is true also concerning various cell domains. This suggests that the statistical power of TEM to paint a scenario which is adherent to actual cytopathology is enormous compared with light microscopy. This latter procedure appears to provide negligible staining up to a level where a sort of dramatic staining enhancement takes place.

The striking discrepancy between immuno-stained area observed at light microscopy (almost as a clear-cut region compared with surrounding cytosol) with the faint, undefined border under TEM is due to the striking difference in magnification, which either neglects or highlights the continuum of cytopathology leading to undefined borders for these multi-faceted biological structures observed at high magnification.