2.1 Preparation of α-syn PFFs

Both human and mouse α-syn were expressed recombinantly in Escherichia coli strain BL21(DE3) (cat. no. C2527H, New England Biolabs) and purified as described previously (Hoyer et al., 2002). Briefly, the cells from 1 L minimal medium expression culture wereThe usage “the cells from 1 L minimal medium expression culture” is not clear. Please check. lysed by freeze–thaw cycles followed by sonication, boiled for 15 min and centrifuged at 48.000 g for 45 min. From the supernatant DNA was precipitated by adding streptomycin (10 mg/ml) to the ice-cold stirred solution. After another centrifugation step, α-syn was precipitated from the supernatant by adding ammonium sulfate to 0.36 g/ml. The pellet was resuspended in 25 mM Tris/HCl, pH 7.7 and further purified by anion-exchange chromatography on a 30-mm POROS HQ column (cat. no. 1232212, PerSeptive Biosystems). To prepare monomeric α-syn without any aggregates, the protein was dialyzed against phosphate-buffered saline (PBS; cat. no. A0964, Applichem) buffer, pH 7.4, centrifuged at 106,000 g for 1 hr at 4℃ and filtrated through 0.22-µm ULTRAFREE-MC centrifugal filter units (cat. no. UFC30GV0S, Merck Millipore). The final protein concentration was set to 5 mg/ml.

For fibrillization in PBS buffer, monomeric human and mouse α-syn were incubated at 37℃ with constant stirring at 200 rotations per minute (rpm). Progress of fibrillization was monitored with a thioflavin T fluorescence assay as described before (Hoyer et al., 2004). Briefly, a 1–2 μl aliquot of fibrils was mixed with 2 ml of a solution of 5 µM thioflavin T (cat. no. T3516, Sigma) in 50 mM glycine/NaOH, pH 8.2, at different incubation times and the fluorescence emission was measured between 460 and 600 nm on a Varian Cary Eclipse fluorescence spectrophotometer (Agilent Technologies). Fibrillization of human α-syn reached steady state after 7–9 days while fibrillization of mouse α-syn took up to 11 days. Monomeric protein and PFFs were shock frozen in liquid nitrogen and stored at −80℃ until usage. Immediately before usage, PFFs were diluted to the final concentration and sonicated (30 s, 1 cycle, 10% power; Bandelin SonoPlus). Electron microscopy images were acquired from PFFs before and after sonication (Figure S1). PFFs were applied to 400-mesh copper grids (cat. no. G2400C, Plano) and stained with 1% uranyl acetate (cat. no. 8473, Merck). Images were acquired using a Philips CM120 electron microscope (Philips) equipped with a TemCam 224A slow scan CCD camera (TVIPS) at a defocus of 2.3 μm, as previously described (Tatenhorst et al., 2016). For all cell culture experiments, mouse-derived α-syn PFFs were employed. For all animal experiments, human-derived monomeric α-syn and α-syn PFFs were employed.

2.2 Animal experiments

Wild-type C57Bl/6J mouse (RRID: IMSR_JAX:000664) couples purchased from Charles River were used for breeding in the Central Animal Care Unit of the University Medical Center Goettingen, Germany. The study received approval from the institutional ethics committee and the animals were treated according to the regulations of the local animal research council and legislation of the State of Lower Saxony, Germany (ethics approval number: 33.9-42502-04-15/1982) in an exploratory study. Sample size calculations were performed using GPower 3.1 software (Faul et al., 2009) according to the guidelines of the legislation of the State of Lower Saxony, Germany (ANOVA, a-priori analysis, effect size f = 0.4, power = 0.8). Mice were housed in individually ventilated cages (IVC, Tecniplast) with standard ad libitum food, water, and 12-hr dark/light cycle. No randomization was performed to allocate subjects in the study, no exclusion criteria were pre-determined. Altogether 233 animals (127 males, 106 females) were used in this study. The average body weight of all mice was 26.2 g, the number of animals used per group is indicated in the figure legends. This study was not pre-registered.

Female and male mouse pups were treated with carbonyl iron (cat. no. C3518, Sigma Aldrich) in three different dosages (60, 120, or 240 mg/kg body weight) dissolved in 5% sorbitol (cat. no. S1876, Sigma Aldrich) in H₂O via oral gavage between postnatal days 10 and 17, as previously described (Carboni et al., 2017). The control group was treated with vehicle only. To avoid injuries, special feeding tubes for mouse pups were used for the gavage (cat. no. FTP22-25, Instech Laboratories).

Stereotactic injections of 12-week-old mice were conducted as described previously (Saal et al., 2015; Tatenhorst et al., 2014). The animals received an analgesic treatment with Metamizole (1.5 mg/ml) in drinking water, starting 2 days before surgery and ending 2 days after the injection of α-syn. Mice were anesthetized by intraperitoneal injection of ketamine (150 mg/kg body weight) and xylazine (10 mg/kg body weight) and fixed in a stereotactic frame. Eyes were protected by application of eye ointment. For all stereotaxic injection experiments, human-derived monomeric α-syn and α-syn PFFs were employed. Coordinates for intrastriatal injections into the right hemisphere were set relative to Bregma (anteroposterior axis: +0.4 mm, mediolateral axis: −1.8 mm, dorsoventral axis: −3.5 mm). After incision and trepanation, 2 µl of α-syn human monomer or PFFs solution (5 µg) was injected at an injection rate of 500 nl/min. The injection capillary was left in place for 4 min to prevent reflux, then slowly removed, and the incision was closed with tissue glue (cat. no. 129463, DermaBond, Ethicon). After surgery, mice were placed on a warming pad until wake up, and then returned to their home cage.

Motor behavior was analyzed using the rotarod test (cat. no. 47600, Ugo Basile) as previously described (Tatenhorst et al., 2016). Briefly, mice were placed on a rotating rod for 5 min with accelerating speed from 5 to 40 rpm. Up to five mice were tested simultaneously. Mice were pre-trained on two consecutive days before the first assessment to reach stability in rotarod performance. The test was conducted weekly, starting after stereotactic injections. All behavioral tests were performed in the afternoon. Each mouse performed three trials (5 min each) on the rotarod per test day with an intertrial interval of 30 min to reduce stress and fatigue. Per session, the time in seconds the mouse spent on the rod was recorded.

The novel object recognition (NOR) test to analyze short-term memory (STM) and long-term memory (LTM) was performed before α-syn PFFs injection and before killing (Bevins & Besheer, 2006; Tatenhorst et al., 2016). Each mouse was placed into an empty arena of 48 x 35 cm for 3 min to habituate. Next, two identical objects were placed inside the arena for 5 min to familiarize the mouse with the objects. After a 10-min recovery (STM testing condition), or after 24 hr (LTM condition), the respective mouse was placed again into the arena with one familiar and one novel object. The exploratory behavior of each mouse was videotaped for 5 min and afterwards analyzed with Ethovision XT 8.5 software (Noldus, Wageningen, The Netherlands). Based on the time spent with the novel object compared to the time spent with both objects, a discrimination ratio was calculated as previously described (Tatenhorst et al., 2016).

Animals were killed at 90 days post-α-syn injection (dpi). This time point was selected based on previous studies (Luk, Kehm, Carroll, et al., 2012; Luk, Kehm, Zhang, et al., 2012), demonstrating a well-advanced spreading of α-syn fibrils through the brain at this stage. For perfusions, mice received a lethal intraperitoneal injection of ketamine (300 mg/kg body weight (bw)) and xylazine (15 mg/kg bw) solution. Under deep anesthesia, mice were then transcardially perfused with 50-ml ice-cold PBS within 5 min followed by 50 ml of 4% paraformaldehyde (PFA, cat. no. A3813, Applichem) at pH 7.4 within 5 min. The brains were post-fixed with 4% PFA in PBS for 24 hr at 4℃ and afterwards transferred into 30% sucrose (cat. no. A2211, Applichem) in PBS for cryopreservation for 48 hr. Cryosections of 30-µm thickness were prepared on a cryostat (CM3060, Leica, Wetzlar, Germany) and either directly mounted on Superfrost slides (cat. no. J1800AMNZ, Menzel, Braunschweig, Germany) and frozen at −20℃, or stored free-floating in 24-well plates in PBS with 0.1% sodium azide (cat. no. A1430, Applichem) until usage for immunohistochemical staining.

2.3 Primary cell culture preparation and treatment

C57Bl/6J donor mice of E18.5 embryos received a lethal intraperitoneal injection of ketamine (300 mg/kg bw) and xylazine (15 mg/kg bw) solution. After explantation, the embryos were decapitated, the brains were dissected, and the primary cortex neurons were prepared in ice-cold calcium–magnesium-free medium (1x HBSS, cat. no. 14180-046, Gibco, with 0.12% bicarbonate solution, cat. no. 25080-60, Gibco). Cells were incubated in 1-ml Trypsin (≥ 9,000 U/mg; 1.87 mg/ml; cat. no. T9935, Sigma Aldrich) for 12 min at 37℃. After adding DNAse (~400 U/mg; cat. no. DN25-100, Sigma Aldrich), the suspension was centrifuged and Trypsin/DNAse was replaced by 1-ml fetal calf serum (FCS; cat. no. S0615, Biochrom) to stop digestion. FCS was then substituted with cell culture medium (Neurobasal medium, cat. no. 21103-049, Gibco, containing holo-transferrin, cat. no. A3124, Applichem, PSN, cat. no. 15640-055, Gibco, L-Glutamine 200 mM, cat. no. 25030-024, Gibco, and B27 supplement 1:50, cat. no. 17504-044, Gibco) and 120,000 neurons were seeded in the primary cell compartment of a microfluidic chamber (cat. no. 855.234.0044, Xona). The other compartment was filled with medium. One day later, half of the medium was exchanged with cytosine arabinoside-containing medium (AraC; cat. no. C6645, Sigma Aldrich) to reach a concentration of 2 µM AraC. On day in vitro 2 (div 2), medium was changed to medium containing 100-µM FeCl₂ (stock solution 5 mM FeCl₂, cat. no. 372870, Sigma Aldrich, diluted in 5% glucose, cat. no. G8769, Sigma Aldrich). The control condition was treated with an equal amount of 5% glucose. After 24 hr, cells were washed twice with medium. Subsequently, the medium was changed to α-syn containing medium (1 µg in 200 ml medium). For all cell culture experiments, mouse-derived α-syn PFFs were employed. After 24 hr, cells were washed. On div 7, 250,000 cells (preparation on E 18.5) were seeded into the secondary cell compartment. Medium change was performed every second day. On div 14, cells were fixed with 4% PFA in PBS for 10 min and stored in PBS at 4℃.

2.4 Quantification of ferric iron levels in the brain tissue of iron- and vehicle-treated mice

To analyze the levels of ferric iron in brain tissue of iron-treated and control mice, midbrain sections from animals that received a postnatal treatment with either vehicle (sorbitol) or 240 mg/kg Fe were mounted on object slides and immunostained with the Iron Stain Kit (Prussian Blue Stain Kit, cat. no. ab150674, Abcam). Stainings were performed following the manufacturer's instructions with small modifications. The incubation for the background staining (Nuclear FastRed, provided in the above mentioned kit) was reduced to 60 s after a 1:20 dilution. Quantitative analyses of iron-stained samples were imaged using the Mosaix module (AxioVision SE64 Rel. 4.8 software, Carl Zeiss) on a Zeiss Axioplan microscope using a 10x magnification. Iron-positive signals and background signals were selected using the iLastik software version 1.3.3 (Berg et al., 2019), followed by the generation of segmentation images for each section (contained only pixels for iron-positive regions). Segmentation micrographs were converted to 8-Bit, binarized and quantified using the ImageJ software (National Institute of Health, Bethesda, MD). Briefly using the ROI manager, the areas of substantia nigra (SN) pars compacta/pars reticulata, cortex, brainstem, and hippocampus were selected for each section based on the Paxinos mouse brain atlas (Paxinos & Franklin, 2003). The average pixel intensity was quantified for each selected region and normalized by the total selected area (in pixels). A schematic of the procedure is given in Figure S8a.

2.5 X-ray fluorescence imaging

For X-ray fluorescence (XRF) analysis, we used 6-month-old female and male mice, which were neonatally enriched with 60 or 120 mg/kg body weight iron. Control mice were treated with vehicle only. After killing the mice, brains were explanted, the hemispheres were separated, and cut into 5-mm-thick sections. These single sections were mounted in cryo matrix (cat. no. 14020108926, Leica) and cryofixed for 2 min in 2-methyl-butane (cat. no. M32631, Honeywell) placed in liquid nitrogen to avoid ice crystal formation. Samples were stored in liquid nitrogen until further analysis. Frozen brain samples were cut into 30-µm sections on a CM3060 Leica Cryostat, freeze-dried, and mounted between two layers of 4-µm ultralene foil (cat. no. 3525, SpexSamplePrep) as described before (Vogel-Mikuš et al., 2014). XRF imaging of freeze-dried brain sections was performed at the XRF beamline of Synchrotron Elettra, Trieste (Karydas et al., 2018). The samples were positioned at an angle of 45° and scanned by 75 × 75 µm beam at excitation energy of 10 keV. XRF was detected by a silicon drift detector (SDD) (XFlash 5030, Bruker, Nano GmbH), which was placed on the incident beam horizontal plane, perpendicular to the primary beam. The SDD has a 30 mm² nominal crystal area, a thickness of 450 µm and nominal resolution equal to 131 eV (at Mn-K) (Karydas et al., 2018).

XRF spectra were fitted by PyMCA (Solé et al., 2007) and quantified as previously described (Kump & Vogel-Mikuš, 2018). The XRF system was calibrated using pure metal foils (MicromatterTM Technologies Inc.) and validated using standard reference materials (cat. no. NIST1537a, Tomato Leaves, Sigma Aldrich). Sample thickness was determined using absorption that was detected by a photodiode placed behind the sample and scattering of X-rays on the sample (Kump & Vogel-Mikuš, 2018).

2.6 Analysis of α-syn spreading pathology

To analyze α-syn spreading pathology within the mouse brain, sections mounted on object slides underwent antigen retrieval procedure in citrate buffer (cat. no. 5110.3, Roth, Karlsruhe, Germany; pH 6.0) at 80℃ for 30 min. After cooling to 21℃ and washing, sections were incubated for 20 min in 25-mM glycine (cat. no. A1377, Applichem) in PBS. Following the application of 0.1% Sudan black (cat. no. A1407, Applichem) for 15 min, sections were rinsed in water and washed in PBS. Sections were then incubated for 1.5 hr at 21℃ in a blocking solution (10% normal horse serum, cat. no. SHD3250YK, Linaris, Mannheim, Germany, 5% bovine serum albumin (BSA; IgG free, cat. no. 001000162, Jackson ImmunoResearch), 0.3% TritonX, cat. no. A4975, Applichem, 25 mM glycine in PBS). The sections were incubated with rabbit anti-pS129 α-syn antibody (1:1,000; EP1536Y, cat. no. ab51253, Abcam) in blocking solution at 4℃ overnight. After washing, sections were incubated with the secondary antibody (Cy3 goat anti-rabbit, 1:500, cat. no. 111-165-006, Dianova, Hamburg, Germany) for 1.5 hr at RT. After washing for 3 × 10 min, 4′,6-diamidino-2-phenylindole (DAPI, cat. no. A1001, Applichem) was applied for 2 min and after final washing, the sections were dried at 37℃ for 10 min and embedded with Mowiol (cat. no. 0713.2, Sigma Aldrich).

For a quantitative analysis of pS129-α-syn signal, immunostained samples were imaged using the MosaiX module of AxioVision SE64 Rel. 4.9 software (Carl Zeiss) on a Zeiss Axioplan microscope. Representative exemplary photomicrographs are provided in Figure S2. Images of the Cy3 channel were converted to 8-bit and the threshold function of ImageJ (National Institute of Health) was used to manually adjust the threshold for every image. A schematic overview of the analysis procedure is given in Figure S3. The adjusted threshold of the greyscale was ranging from 231 to 250. It was verified that the manually adjusted threshold values did not differ significantly between images from different animals (Figure S4). Next, the corresponding image of the DAPI channel was used to calculate the area (in pixel²) of the injected and the contralateral hemisphere separately as regions of interest (ROIs). pS129-α-syn-positive signal within ROIs was calculated by the ‘analyze particles’ tool of ImageJ. Results were given as percentages resulting from the signal area divided by the total area of the corresponding hemisphere.

To describe pS129-α-syn spreading to different brain regions, ImageJ was used to overlay the generated threshold image from quantification analysis with DAPI images showing the brain structures (Figure S3). Brain regions were adjusted according to the Paxinos mouse brain atlas (Paxinos & Franklin, 2003) and marked where the overlay images showed pS129-α-syn signal. In addition, the occurrence of pS129-α-syn signal was graded for each brain region using a five-stage scale (0–4) as follows: - no positive signal, + few (sporadic positive structures), ++ mild (several positive structures, but not widespread), +++ moderate (many positive structures covering a widespread part of the brain region), and ++++ severe (almost the whole brain region is covered with pS129-α-syn signal). Exemplary images per stage are illustrated in Figure S5. Six to eight animals were evaluated per group and a mean signal amount was calculated for each positively graded brain region. A color code was assigned to the stage-scaled graduation and the results were illustrated as heat maps. The Allen Mouse Brain Connectivity Atlas (©2020 Allen Institute for Brain Science. Allen Mouse Brain Connectivity Atlas, available from: https://connectivity.brain-map.org/) (Oh et al., 2014) was used to analyze pS129-α-syn spreading to anatomically connected regions. The connectivity platform contains data of different EGFP injection experiments and enables to follow the affected and connected brain regions starting from the injection site. Based on the selection of two injection experiments of the aforementioned platform (Exp. 158916311: https://connectivity.brain-map.org/projection/experiment/158916311 and Exp. 100142580: https://connectivity.brain-map.org/projection/experiment/100142580), of which the injection sites are located in the injection radius of our study, the anatomical connectivity of this study was reconstructed. Respective brain regions were illustrated in connectivity maps of the six selected brain sections for our analysis. In a final step, the six connectivity maps were compared with the pS129-α-syn spreading maps to visualize intersections.

For the in vitro experiment, fixed cells were incubated with 25-mM glycine in PBS for 20 min. After washing, cells were incubated with blocking solution for 1.5 hr at 21℃. Primary antibodies against α-syn fibrils (rabbit anti-α-syn pS129, 1:1,000, cat. no. EP1536Y; Abcam) and MAP2 (mouse anti-MAP2, 1:200, cat. no. MAB3418; Merck Millipore) were diluted in blocking solution and cells were incubated with the antibody solution overnight at 4℃. After washing and incubation with the secondary antibodies for 1.5 hr at 21℃ (Cy3 goat anti-rabbit, 1:500, cat. no. 111-165-006, Dianova, Hamburg, Germany; Cy5 goat anti-mouse, cat. no. 111-175-146, Jackson ImmunoResearch), cells were washed and counter-stained with DAPI for 2 min. The cells were mounted with Mowiol and finally imaged on a Zeiss Axioplan microscope (magnification 200x; AxioVision SE64 Rel. 4.9 software; MosaiX module). pS129- α-syn signal was assessed using the threshold function of ImageJ as described above. Per culture (n = 5) one threshold was defined, which was used to analyze the chambers of the different conditions. The secondary cell compartment was defined as region of interest (ROI). A percentage was generated from the signal area and ROI area and, afterwards, the average of one to three chambers per condition was calculated. Using the PBS condition as control, the relative signal (%) of the PBS condition was subtracted from the relative signal (%) of the monomer or PFFs conditions, respectively.

2.7 Microglia and T-cell quantification

To investigate the immune cell response after α-syn PFF injections, microglia and T cells were analyzed in the brain tissue of mice. Immunohistochemical staining for CD11b+ microglia was performed using 30-µm PFA-fixed coronal cryosections of the substantia nigra (SN) and striatum. Free-floating sections were pre-incubated for 3 hr at 21℃ on a shaker with 5% BSA/5% normal goat serum (NGS; cat. no. CL1200-100, Cedarlane Labs) in 0.01 M PBS before being incubated overnight at RT with the primary rat anti-mouse CD11b antibody (1:100, cat. no. MCA74G, Bio-Rad Laboratories), diluted in 1% BSA/1% PBS in 0.01 M PBS. To visualize the CD11b primary antibody, sections were then incubated for 2 hr with biotinylated secondary antibody against rat Igs (1:100, cat. no. BA-9401-0.5, Vector Laboratories) diluted in 1% PBS/1% NGS in 0.01 M PBS followed by avidin/biotin reagent (cat. no. 32020, Thermo Fisher Scientific) before incubation and staining with diaminobenzidine-HCl (DAB) and H₂O₂ (DAB substrate kit, cat. no. SK-4100, Vector Laboratories). Cell nuclei were counter-stained with Hematoxylin (cat. no. T865.2, Carl Roth).

Quantification of CD11b+ microglia was performed with a light microscope (Olympus BH2, Olympus, Hamburg, Germany) using an ocular grid covering a defined area (0.0625 mm²) at a final magnification of 400x. For each animal, the PFFs-injected and contralateral sides were investigated separately by quantifying on average two sections per region and two to three grids per section. The cell number of immune cells was then calculated per square millimeter.

Immunohistochemical stainings for CD3+ T cells were performed using 30-µm PFA-fixed coronal cryosections of the SN and striatum. Free-floating sections were pre-incubated for 3 hr at 21℃ on a shaker with 5% BSA/5% NGS in 0.01 M PBS before being incubated for 2 hr at RT with the primary rat anti-mouse CD3 antibody conjugated with Alexa Fluor® 488 (1:50, cat. no. 100210, Biolegend). DAPI staining was performed to label cell nuclei.

Quantification of T-cell numbers was performed using a Zeiss Axio Imager.M2 fluorescence microscope. Because of the low number of T cells, the entire regions of the SN (pars compacta and pars reticulata), the pre-commissural, commissural, and post-commissural parts of the striatum were analyzed, respectively. The cell number of immune cells was then calculated per square millimeter.

2.8 Statistical analysis

SPSS statistics 26.0 (IBM) and Graphpad Prism 5.04 (Graphpad Software) were used for statistical analyses. The Kolmogorov–Smirnov test was used to evaluate the normality of the data. The results indicated that when alpha =0.05, p > .05 confirmed that data were normally distributed. No test for outliers was conducted. Analysis of variance (ANOVA) with repeated measurements, one-way and two-way ANOVA, Kruskal–Wallis test, Fisher's exact test, two-tailed Student's t test, and Mann–Whitney U test were calculated as indicated. For multiple group comparisons, Tukey's post hoc tests or Dunn's post hoc test were used. Sample sizes of the experiments as well as used statistical tests are specified in the respective figure legends. All analyses were performed by a blinded investigator. Data are given as mean ± standard error of the mean (SEM). Differences were considered significant with p < .05 (*p < .05; **p < .01; ***p < .001).

3.1 Behavioral effects of iron treatment and injection of α-syn PFFs in mice

To investigate how iron treatment and α-syn PFFs injections influence motor and cognitive behavior, mice were tested on the rotarod and in the novel object recognition (NOR) test. Animals treated with different concentrations of iron (60, 120, or 240 mg/kg body weight) or sorbitol as vehicle and injected with α-syn monomer or PFFs performed the rotarod test weekly between the day of α-syn injection (doi) and killing (90 dpi). The NOR test was performed twice: directly before doi to investigate the effect of iron intoxication, and directly before killing to investigate combined effects of iron intoxication and α-syn PFFs injections (Figure 1a).

Analysis of the rotarod data revealed that iron and PFFs treatment did not affect the motor behavior (Figure 1b). Only for the variable “time before killingAccording to the style sheet, the word “sacrificed” should be avoided. Hence it has been changed to “killed” and “killing” according to the context. Please check.”, ANOVA with repeated measures indicated that the average of all mice shows a slight continuous decline of the time spent on the rotarod (**p < .01; overall mean 10 weeks: 241.0 ± 17.1 s versus overall mean at time of killing [0 weeks]: 223.3 ± 19.2 s).

The NOR test was used to test both short-term memory (STM) and long-term memory (LTM) by modifying the timeframe between habituation and recall phase (10 min versus 24 hr; Figure 1c). Before PFFs injection, high-dose iron-treated mice (0.49 ± 0.03) showed no significant difference in their STM as compared to controls (0.54 ± 0.04; Figure 1d). Regarding the LTM at this time point (Figure 1h), mice treated with 240 mg/kg iron (0.54 ± 0.02) performed significantly worse in LTM compared to mice of the 60 mg/kg iron group (0.64 ± 0.02; *p < .05), but not compared to controls (0.58 ± 0.03). At 90 dpi, neither an interaction effect of iron×α-syn on STM nor a main effect of the separate treatment conditions (iron or α-syn treatment) could be detected (Figure 1e–g). Focusing on LTM, the interaction effect of iron×α-syn was not significant and no main differences between the iron-treatment groups were found. However, two-way ANOVA revealed a significant main effect of α-syn injections showing moderately reduced discrimination ratios in LTM for PFFs-injected mice (0.63 ± 0.02 vs. 0.56 ± 0.02; Figure 1i–k; *p < .05).

As a control variable of general mobility, the distance moved within the NOR arena was analyzed for both time points (Figure S6). Before α-syn PFFs or monomer injection, no significant differences of mice treated with different iron concentrations were observed when compared to controls. Only mice of the 240 mg/kg iron group (1,286 ± 67 cm) showed a decreased distance moved compared to both the 120 mg/kg iron group (1622 ± 113 cm; *p < .05) and the 60 mg/kg iron group (1715 ± 79 cm; **p < .01) at 24 hr post-habituation phase. However, none of these groups was significantly different when compared to controls (1,472 ± 87 cm). At 90 dpi, no significant differences in movement were detected between the treatment groups.

3.2 Analysis of iron in tissue from iron-treated animals

Synchrotron micro-XRF analyses of freeze-dried SN sections as well as Prussian blue iron stainings of cryosections from 6-month-old mice were performed to verify increased brain iron in C57Bl/6J mice treated with carbonyl iron between postnatal days 10 and 17. In XRF analyses, quantitative iron distribution maps were generated after scanning SN sections with a 75-µm beam and an average iron value was extracted from the region of the SN of mice treated with 60 mg/kg Fe (25 and 31.2 µg/g), 120 mg/kg Fe (25.4 and 40.6 µg/g), or vehicle (7.1 and 4.4 µg/g) as indicated (Figure S7a, b). Mice treated with iron showed at least threefold increase of iron concentration within the SN compared to control mice (Figure S7c). A significant increase of iron was also confirmed by Prussian blue iron staining in the SN (vehicle: 0.006 ± 0.001 normalized average pixel intensity versus 240 mg/kg Fe: 0.010 ± 0.004 normalized average pixel intensity, Figure S8b) as well as in the cortex (vehicle: 0.031 ± 0.008 normalized average pixel intensity versus. 240 mg/kg Fe: 0.065 ± 0.033 normalized average pixel intensity, Figure S8c), and in the hippocampus (vehicle: 0.010 ± 0.002 normalized average pixel intensity versus. 240 mg/kg Fe: 0.017 ± 0.004 normalized average pixel intensity, Figure S8e). No significant difference could be detected in the brainstem (vehicle: 0.004 ± 0.002 normalized average pixel intensity versus. 240 mg/kg Fe: 0.006 ± 0.001 normalized average pixel intensity, Figure S8d).

3.3 Quantification of α-syn spreading pathology in vivo

To assess the influence of iron on the spreading of α-syn pathology induced by the seeding of PFFs, we performed a threshold-based quantification of phosphorylated (pS129) α-syn. This post-translationally-modified subtype of α-syn has been reported to enhance its toxicity both in vivo and in vitro, as well as to increase the formation of α-syn aggregates (Chen et al., 2009; Fujiwara et al., 2002; Smith, 2005). Six coronal mouse brain sections named after their coordinates relative to Bregma were analyzed and the pS129-α-syn signal was normalized to the area of the region of interest and compared between the treatment groups. As expected, virtually no pS129-α-syn signal could be detected in the α-syn monomer-injected animals, independent of iron dosage (Figure 2).

At 90 days following α-syn PFFs injection, pS129-α-syn signal was detected in all six analyzed brain sections (Figure 2a–f). Interestingly, in the injected hemisphere, the five most rostral sections (Bregma +1.54, +0.38, +0.02, −1.58, and −3.08 mm, Figure 2a–e) did not show any significant difference in pS129-α-syn signal after iron treatment, although the control-treated group always showed numerically more pS129-α-syn signal as compared to the iron-treated groups. However, the most caudal section that also contained the SN (Bregma −3.28, Figure 2f) displayed a significantly decreased pS129-α-syn signal after high-dose iron treatment (240 mg/kg iron: 0.11 ± 0.03% vs. ctrl: 0.41 ± 0.11%; *p < .05).

Furthermore, we calculated the pS129-α-syn signal average of all available brain sections per mouse separately for signal in the contralateral (Figure 2g) and ipsilateral hemisphere (Figure 2h). At the contralateral hemisphere, the overall pS129-α-syn signal of PFFs-injected animals was markedly lower as compared to the injected hemisphere. No significant difference between groups was observed here. In the PFFs-injected hemisphere, mice showed significantly higher pS129-α-syn signal in the 240 mg/kg iron-treated group (0.10 ± 0.02%; *p < .05) as compared to controls (0.27 ± 0.05%).

3.4 Iron-dependent changes in α-syn spreading pathology in different brain areas

In order to analyze whether iron treatment changed the α-syn spreading pathology in PFF-injected mouse brains, brain sections were stained for pS129-α-syn. Spreading maps were generated for selected brain sections, similar to previously published analyses (Luk, Kehm, Carroll, et al., 2012; Masuda-Suzukake et al., 2013, 2014; Paumier et al., 2015). The spreading of α-syn pathology to different brain regions was quantified in six coronal brain sections and illustrated in heat maps to compare the pS129-α-syn spreading pattern between PFFs-injected mice treated with vehicle and high dosage iron (240 mg/kg).

As shown in Figure 3a, the overall α-syn spreading pattern between iron- and vehicle-treated mice was largely similar. In Tables S1–S6, all brain regions exhibiting α-syn pathology and the according semi-quantified pathology ratings are summarized corresponding to the six evaluated coronal brain sections. Between iron-treated and control animals, distinct differences in the quantity of pS129-α-syn aggregates within specific brain regions were found, which are visualized in specific subtraction heat maps (Figure 3b).

Here, in the injected hemisphere, the most prominent difference (at least two levels, e.g., + versus. +++) was observed within the cingulate cortex area 1, the primary motor cortex, perirhinal cortex, lateral amygdaloid nucleus, stria terminalis, and the amygdalohippocampal area. In all of these brain regions, the control group showed an increased pS129-α-syn signal compared to the iron group. Iron-treated mice showed more pS129-α-syn signal (two-level difference) at the contralateral hemisphere within the agranular insular area. Mostly, group differences showed only one rating level of distance such as – versus +or + versus ++. The quantification of regions showing a higher α-syn amount in either iron-treated or vehicle-treated mice is summarized in a fourfold table in Figure 3c. For the injected hemisphere, a higher number of regions with more α-syn signal was found in the PFFs-injected group with vehicle treatment as compared to iron-treated animals (*p < .05, Fisher's exact test).

3.5 Connectivity analysis

Based on the Allen Mouse Brain Connectivity Atlas (©2020 Allen Institute for Brain Science. Allen Mouse Brain Connectivity Atlas, available from https://connectivity.brain-map.org/) (Oh et al., 2014), pS129-α-syn spreading to anatomically connected brain regions and brain regions beyond the connectome was analyzed based on the data presented in Figure 3.

In iron-treated, but also in vehicle-treated mice with PFFs injection, the pS129-α-syn spreading signal was observed in brain regions of the connectome but also in other brain regions (Figure 4a–c). By adding up all pS129- α-syn-positive brain regions out of the six evaluated brain sections and dividing these regions into both connectome-based/connectome-independent brain regions, a fourfold table was designed to indicate that, in vehicle-treated mice, more connectome-associated brain regions were affected than in iron-treated mice. On the other hand, iron-treated mice show more connectome-independent brain areas affected (Figure 4d). In the following analysis, we aimed to assess which of the mouse groups (iron- or vehicle-treated) showed more pS129-α-syn staining, in connectome-specific brain regions or connectome-independent brain regions (Figure 4e–f), in order to investigate the influence of iron on connectivity-based pS129-α-syn-spreading. Therefore, pS129-α-syn rating averages of all affected brain regions (six evaluated brain sections) were considered and grouped for iron- and vehicle-treated animals. The graphs show the data for each of these groups individually. The average of these ratings of different brain regions was compared between both groups. Whereas vehicle-treated mice showed significantly higher α-syn pathology levels in affected connectome-specific brain regions (ctrl: 0.99 ± 0.09 vs. Fe: 0.70.± 0.07; *p < .05, Mann–Whitney U test, Figure 4e), in connectome-independent brain regions no significant group differences were found (ctrl: 0.64 ± 0.05 versus. Fe: 0.47 ± 0.03; Figure 4f).

3.6 Microglia and T-cell distribution in the striatum of iron- and α-syn PFFs-treated mice

To analyze if microglial and/or T-cell activation contributes to the iron-mediated changes in pS129-α-syn distribution following PFFs injection, microglia were quantified in the striatum (Figure S9a–c). Figure 5c shows an exemplary image of CD11b+ microglia, which were quantified in this analysis.

In the injected hemisphere, the average of microglia cell numbers of three striatal sections resulted in a significantly increased microglia accumulation in vehicle/PFFs mice (333.5 ± 13.1 microglia/mm²) compared to the monomer-injected mice (vehicle/monomer: 258.2 ± 13.6 microglia/mm² vs. 240 mg/kg Fe/monomer: 267.8 ± 15.5 microglia/mm²). Whereas no significant difference was observed between the vehicle/PFFs group and the 60 mg/kg Fe/PFFs group (277.5 ± 19.8 microglia/mm²), the vehicle/PFFs showed significantly increased microglia numbers compared to the 120 (254.1 ± 17.1 microglia/mm²) and 240 mg/kg Fe/PFFs groups (238.9 ± 11.1 microglia/mm²; Figure 5a). Furthermore, the injected hemisphere of one SN section was analyzed but no significant microglia differences were found between the groups (Figure S9d). In the contralateral hemisphere (Figure 5b), the vehicle/PFFs group (288.6 ± 10.5 microglia/mm²) showed a significantly increased microglia accumulation compared to the 60 (230.0 ± 12.3 microglia/mm²) and 240 mg/kg Fe/PFFs groups (226.2 ± 11.6 microglia/mm²). PFFs-injected mice treated with 120 mg/kg Fe showed unaltered microglia counts (242.7 ± 14.2 microglia/mm²). In addition, the vehicle/PFFs group exhibited an increased microglia accumulation compared to the vehicle/monomer group (237.4 ± 11.5 microglia/mm²). In the SN, no significant differences in microglia numbers were found between the groups (Figure S9e).

To assess a potential contribution of T cells to α-syn spreading, brain sections were additionally immunolabeled for CD3. However, our data showed no significant differences in T-cell counts between groups, neither in striatal nor in SN sections (Figure S10).

3.7 Contribution of iron to spreading of α-syn PFFs in vitro

After showing that iron treatment attenuated α-syn spreading throughout the brain of C57Bl/6J mice, we wanted to assess if this effect could also be observed in an in vitro PFFs spreading assay in microfluidic chambers. Primary murine cortex neurons were used to investigate if iron influences PFFs spreading in glia-free conditions. First, the cell toxicity of different iron dosages was assessed (Figure S11). Based on previously optimized FeCl₂ concentrations in cell culture models (Dai et al., 2013; Oexle et al., 1999; Solntseva et al., 2015), iron treatment was performed at a concentration of 100-µM FeCl₂, where no toxic effects were observed. Neurons were treated with either FeCl₂ or vehicle at div2 for 24 hr and afterwards for 24 hr with α-syn PFFs or monomers (Figure 6a).

Figure 6b displays an exemplary photomicrograph of cells immunostained with an antibody recognizing pS129-α-syn showing that α-syn pathology was transmitted to cells of the secondary cell compartment. pS129-α-syn signal within the secondary cell compartment was quantified in all different conditions (PBS, monomer or PFFs; with or without FeCl₂) and normalized to the respective PBS-only treated control condition (Figure 6c). In the secondary cell compartment, significantly more pS129-α-syn signal was detected in the PFFs-treated cells as compared to monomer-treated cells, but no influence of FeCl₂ treatment was observed in this model (FeCl₂/PFFs 0.015 ± 0.0048 vs. vehicle/PFFs 0.013 ± 0.0042). Figure S12 demonstrates that pS129-α-syn signal was present only intracellularly but not extracellularly.