Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Amyloid precursor protein elevates fusion of promyelocytic leukemia nuclear bodies in human hippocampal areas with high plaque load

Amyloid precursor protein elevates fusion of promyelocytic leukemia nuclear bodies in human... The amyloid precursor protein (APP) is a type I transmembrane protein with unknown physiological function but potential impact in neurodegeneration. The current study demonstrates that APP signals to the nucleus causing the generation of aggregates consisting of its adapter protein FE65, the histone acetyltransferase TIP60 and the tumour suppressor proteins p53 and PML. APP C-terminal (APP-CT50) complexes co-localize and co-precipitate with p53 and PML. The PML nuclear body generation is induced and fusion occurs over time depending on APP signalling and STED imaging revealed active gene expression within the complex. We further show that the nuclear aggregates of APP-CT50 fragments together with PML and FE65 are present in the aged human brain but not in cerebral organoids differentiated from iPS cells. Notably, human Alzheimer’s disease brains reveal a highly significant reduction of these nuclear aggregates in areas with high plaque load compared to plaque-free areas of the same individual. Based on these results we conclude that APP-CT50 signalling to the nucleus takes place in the aged human brain and is involved in the pathophysiology of AD. Keywords: Alzheimer’s disease, APP-CT50, PML, IPSC-derived cerebral organoids, 3D culture, Nuclear complexes, Human brain, Amyloidogenic plaques, HSV, Viral defence Introduction the generation of three fragments: (i) the secreted Increased amyloidogenic processing of the amyloid extracellular domain (sAPPβ), (ii) the β-amyloid pep- precursor protein (APP) occurs in sporadic Alzhei- tide (Aβ), and (iii) the APP C-terminal fragment mer’s disease (AD) [1], in familial AD with mutations (APP-CT). The secreted fragment (i) was reported to in APP or in its processing enzymes [2], and in trisomy provoke neurotrophic effects [4], Aβ (ii) is the main 21 patients [3]. Amyloidogenic APP processing causes component of amyloidogenic plaques [5] and APP- CT (iii) was suggested to play an important role in a nuclear signal transduction pathway [6]. APP-CT, which has been reported to exist in different isoforms *Correspondence: thorsten.t.mueller@rub.de David Marks, Natalie Heinen, Lisa Bachmann: shared first authors with 50aa in length being the most stable one (APP- Department of Molecular Biochemistry, Cell Signalling, Faculty CT50), indeed is a remarkable protein fragment as it of Chemistry and Biochemistry, Ruhr University Bochum, Bochum, is intrinsically unstructured [7]. Though, this changes Germany Full list of author information is available at the end of the article upon interaction with FE65, causing APP-CT50 to fold © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Marks et al. acta neuropathol commun (2021) 9:66 Page 2 of 16 into a three-dimensional conformation that can be Geltrex (Gibco) according to the manufacturer’s analysed by x-ray crystallography. APP-CT50 is capa- protocol, and split before reaching 70% confluency. ble to enter the nucleus establishing a protein complex HEK293T cells were seeded and incubated in DMEM consisting of additional proteins like FE65, TIP60, and (Gibco) with 10% heat inactivated FBS (Gibco), 1% BLM [8]. The presence of the histone acetyl trans- Penicillin/Streptomycin and 1% L-glutamine (Gibco), ferase (TIP60) and the DNA helicase (BLM) in the to a confluency of 70%. For overexpression assays, ster - complex points to a functional role in essential bio- ile precision cover glasses (1.5 H Marienfeld Superior) logical mechanisms such as gene expression, DNA were placed into a 24-well cell culture plate (Sarstedt) replication/damage/repair or chromatin modification. and coated with 0.01% poly-l-ornithine solution (Sigma Indeed, a variety of target genes like GSK3β, IDE, and Aldrich). The respective plasmids were transfected APP have been proposed to be APP-CT50 dependently via the K4 Transfection Kit (Biontex) according to regulated [9–12]. manufacturer’s recommendations for 24-well culture Promyelocytic leukemia nuclear bodies (PML-NBs) plates. After 24 or 48  h, cells were briefly washed with are multiprotein complexes with PML as the main DPBS (Gibco) and fixed in 4% paraformaldehyde in building component [13]. A diverse set of nuclear pro- PBS. Cover glasses were mounted with the Shandon teins have been identified as permanent or transient Immu-Mount solution (Thermo Scientific) on glass PML-NB-binding partners [14]. PML-NBs are highly slides and dried overnight at RT. For immunofluores - dynamic structures with respect to mobility, composi- cence staining, HEK293T cells were seeded in 8-well- tion, architecture, and function [15]. While their pre- µ-slide-ibi Treat (ibidi , Martinsried, Germany) and cise biochemical functions have not been elucidated transfected using calcium phosphate transfection after yet, they have been linked to many aspects of chroma- 24 h. Cells were fixed with Roti -Histofix 4% (4% phos - tin biology, including transcription, histone modifica- phate buffered formaldehyde solution; Roth, Karlsruhe, tion, repair and recombination, degradation, hence DE) for 20 min at 37 °C, and permeabilized and blocked genome maintenance [16]. Transcription of PML is with 5% normal goat serum (NGS) in 0.3% (w/v) Triton strongly upregulated by interferons and p53 [17], caus- X-100/PBS for 30  min at RT. The cells were incubated ing a significant increase in the number and size of the with primary antibodies diluted in 1% BSA/0.3% Tri- bodies. Recent studies revealed an emerging role of ton X-100/DPBS (mouse anti-HA (BioLegend, 901501; PML-NBs as coregulatory structures of both type I and 1:1000), mouse anti-myc (NEB/Cell Signalling, 2276; type II interferon responses [18]. Within this work, we 1:1500) o/n at 4  °C. For the secondary antibody stain- demonstrate that PML nuclear bodies interact with ing and the cell staining, goat-anti-mouse AF568 (Inv- highly mobile APP-CT complexes and progressively itrogen, A11004, 1:1000) was used, together with HCS form immobile large nuclear structures with relevance CellMask Deep Red Stain (ThermoFisher Scientific, for AD pathophysiology. H32721, 1:5000), Hoechst33342 (10  mg/mL in H2O, Applichem, A0741, 1:1000) in DPBS (1% BSA, 0.3% Tri- Methods ton) and incubated for 1 h at RT. Vector constructs The plasmids encoding the fusion proteins (FE65- Immunoprecipitation EGFP, FE65-mCherry, APP-CT-GFP, TIP60-EGFP, HEK 293  T cells were seeded in 10  cm dishes and co- TIP60-HA, TIP60-BMP, EGFP-PML, PML-HA, transfection was performed 24  h after seeding. Whole PML-myc, p53-EGFP, Daxx-EGFP, H2A-mTurquoise, cell extracts were prepared 24  h after transfection by HIPK2-EGFP, HP1ß-EGFP, UBE2D2-mCherry, and scraping the cells from the dish with a cell scraper, WRN-EGFP) described in this paper, were gener- washing the cell pellet in ice-cold PBS, extracting with ated using the In-Fusion HD cloning kit (Takara Bio) 1  ml interaction buffer (50  mM Tris pH 8, 150  mM according to manufacturer’s instructions or were pur- NaCl, 5  mM EDTA, 0.5% NP40, 1  mM DTT, 1  mM chased (Addgene). Amplification and purification of PMSF, 1× complete protease inhibitor cocktail), fol- the plasmids were done according to standard proto- lowed by sonication (15  s at 95% amplitude) using a cols. An overview of all constructs used in this study is Sonopuls mini20 device (Bandelin, Berlin, Germany). given in Additional file 1: Fig. S7. The lysates were centrifuged (15,000g , 15 min, 4 °C) and the supernatant was transferred to a new reaction tube. Cell culture, transfection, and immunofluorescence Input samples of the lysates were stored separately. Stem cells (iPS CD34) were cultured in StemFlex Immunoprecipitation (IP) was carried out with the (Gibco) on 35  mm dishes, coated with Matrigel or µMACS isolation kits for tagged proteins from Miltenyi M arks et al. acta neuropathol commun (2021) 9:66 Page 3 of 16 Biotec (Bergisch-Gladbach, Germany). The eluates, as attachment plate (Corning) with 500  µL Neural Induc- well as input samples of the lysates were subjected to tion (NI) -medium. Every day, half of the media was SDS-PAGE and immunoblotting. exchanged with 500  µL fresh NI-medium. On day 12, the embryoid bodies were embedded in droplets of Immunoblotting Matrigel (Corning) and incubated for 25  min at 37  °C Protein concentrations were determined using the for Matrigel polymerization. Afterwards, the droplets Bio-Rad protein assay system (Bio-Rad Laborato- were transferred to a 50  mm dish with differentiation ries, Richmond, CA). Equal amounts of protein were medium without vitamin A (DM-A) medium for fur- resolved by SDS-PAGE using a 10% acrylamide gel ther incubation at 37 °C in a 5% CO atmosphere. Four and subsequently transferred onto polyvinylidene dif- days later, the medium was changed to DM+A and luoride (PVDF) membranes (Amersham Hybond, GE the developing cerebral organoids (COs) were main- Healthcare) via the PerfectBlue tank electro blotter tained at 37  °C with 5% C O until experiments were (Peqlab, Erlangen, Germany) with 350 mA for 90 min. performed. To minimize unspecific binding, the membranes were blocked in 5% (w/v) non-fat dried milk powder Histology and immunohistochemistry in TBST for 30  min at RT. Membranes were probed The COs were removed from the media, washed with with primary antibodies against GFP (rabbit, poly- PBS, and fixed with 4% paraformaldehyde in PBS for clonal, 1:2000, Santa Cruz, sc-8334), HA tag (mouse, 90 min at 4 °C. After washing with PBS, organoids were monoclonal, 1:1000, BioLegend, 901501) and p53 incubated in 30% sucrose solution for cryoprotection at (mouse, monoclonal, 1:200, Novus Biologicals, NBP2- 4  °C overnight. The next day, the COs were embedded 29419) diluted in blocking solution overnight at 4  °C. in a 1:1 mixture of 30% sucrose and Tissue-Tek O.C.T. The membranes were washed three times with TBST, embedding medium (Science Services, SA62550-01), before they were incubated with HRP-conjugated sec- snap-frozen on dry-ice, and then stored at − 80 °C until ondary antibodies (1:10,000, NXA931 and NA934, cryosectioning. Frozen COs were sliced into 15 µm sec- GE Healthcare Europe, Freiburg, DE) also diluted in tions using a cryostat (Leica CM3050S), mounted on ™ ™ blocking solution for 1 h at RT. Visualization of bound SuperFrost slides (ThermoScientific ), and stored at antibodies occurred via enhanced chemiluminescence − 80 °C until further use. (ECL) with the ECLplus Western Blotting Substrate For immunohistochemistry, COs and brain tis- from Pierce (Rockford, IL, USA) according to the sue sections were thawed for 2  min in PBS. To apply manufacturer’s instructions. After incubation with the the biotin-avidin system used for the enhancement of substrate, the detection of the generated signal was fluorescence, the sections were first blocked with avi - carried out with the ChemiDoc MP Imaging System din for 10  min and, after washing twice with PBS for (Bio-Rad Laboratories GmbH, Feldkirchen, Germany). 4  min, blocked with biotin for 10  min. After washing twice with PBS for 4  min, sections were blocked and permeabilized in 0.1% Triton X-100, 5% goat serum in Cerebral organoids PBS for 1 h at RT, followed by incubation with primary Cerebral organoids were generated according to the antibodies in a humidified chamber overnight at 4  °C. protocol from Lancaster and Knoblich [19] with minor Primary antibodies were diluted in 0.1% Triton X-100 modifications, all media compositions remained in PBS as follows: APP-CT (mouse, Millipore MAB343, unchanged. Briefly, at day 0, iPS CD34 positive cells ™ 1:100), PML (rabbit, Novus Biologicals NB100-59787, were detached and harvested using TrypLE (Ther - 1:400), PML (mouse, Abcam ab96051, 1:200), TIP60 moFisher, Germany). Afterwards, DMEM/F12 was (mouse, Abcam Ab54277, 1:400), FE65 (mouse, Acris added to the detached cells and the cell number was AM32556SU-N, 1:400), FE65 (rabbit, Santa Cruz calculated using a Neubauer chamber. Next, 9000 cells/ sc-33155, 1:400), β-tubulin III (mouse, StemCell 01409, well were seeded into a 96-well ultra-low attachment 1:100), p53 (mouse, Novus Biologicals NB200-103, plate (Corning) with a total amount of 150  µL hESC- 1:100). Sections were incubated with the biotinylated medium (containing 4 ng/mL bFGF and 50 µM ROCK- secondary antibody (goat anti-mouse IgG Biotin, Life Inhibitor) per well. On day 3, half of the media was Technologies B-2763, 1:100) diluted in 0.1% Triton exchanged with 150 µL of hESC-medium without bFGF X-100 in PBS for 1  h at RT in a humidified chamber. and ROCK-Inhibitor. Subsequently (day 6), the embry- Following washing twice with PBS for 4  min, sections oid bodies were transferred to a 24-well ultra-low Marks et al. acta neuropathol commun (2021) 9:66 Page 4 of 16 were incubated with Avidin-TRITC (1:1000) and a non- nuclear bodies. The detection threshold was adjusted to biotinylated secondary antibody (donkey anti-rabbit measure objects with a positive generated score, com- FITC, Santa Cruz sc-2090, 1:100) diluted in 0.1% Triton puted by the software, to further discriminate against X-100 in PBS for 45 min in a humidified chamber pro - the background. The bodies were tracked within the cells tected from light at RT, and subsequently washed twice over a time span of 300  s and the speed was calculated with PBS for 4 min. using the integrated software. In case of thioflavin-S counterstaining of amyloid plaques, sections were incubated with 0.1% aqueous thioflavin-S solution, washed twice with PBS for 4 min, STED microscopy washed with 30% ethanol followed by 50% ethanol for For STED, GFP fusion proteins in fixed cells were 5  min each, and finally washed twice for 4  min with labelled with Alexa Fluor 647-coupled GFP nano- PBS. bodies (GFP-booster gb2AF647-50, Chromotek, For counterstaining of nuclei, DAPI solution Germany) at 1:100 dilution. Endogenous and overex- (0.001  mg/mL) was added to the sections for 15  min pressed PML was immunofluorescently labelled with while protected from light, then slides were washed anti-PML antibody (rabbit, ABD-030, Jena Bioscience, twice with PBS for 4 min and mounted. Germany, 1:500), followed by secondary antibody coupled with STAR 580 STED dye (goat-anti-rabbit, Imaging and tracking ST580-1002-500UG, Abberior, Göttingen, Germany, Cells were either imaged after fixation and mount- 1:100). Stained cells were embedded in ProlongGold ™ ™ ing (Shandon Immu Mount solution, Thermo Sci- with DAPI (Thermo Fisher Scientific, Germany) and entific) on glass slides or for life cell imaging directly covered with 12  mm round cover glasses (Thickness using the integrated incubation chamber of the Leica 0.17 ± 0.01  mm). Gated STED images were acquired (Mannheim, Germany) TCS SP8 microscope system on a Leica TCS SP8 STED microscope equipped with (37 °C and 5% CO ). Samples were imaged using a 63× a 100× oil objective (HC PL APO CS2 100×/1.40 Oil) water (1.2 NA) or 100× oil objective (1.4 NA). Fluo- according to protocols established for nuclear bod- rophores were excited with 405/488/514/561  nm laser ies by Okada and Nakagawa [20]. Pixel size in STED lines performing a sequential scan beginning with the acquisition was applied automatically in LAS-X soft- most red-shifted wavelength. Images were recorded ware (Leica, Mannheim, Germany) for the most red- into 1024 × 1024 images at a scan speed of 200  Hz shifted dye (AF 647), usually resulting in a pixel size with HyD detectors. Tile scans were imaged through of less than 20 × 20  nm. STED beam alignment was the selection of 800 × 800  µm areas (5 × 5 tiles) in x- performed before each imaging session between the and y-direction. Additionally, z-stacks (n = 5) of 2  µm pulsed white light laser and the 592 nm depletion laser. between each plane (8 µm in total) were recorded and DAPI, Alexa Fluor 488, Star 580 and Alexa Fluor 647 merged via the maximum projection tool in the LASX- were excited with laser lines 405  nm, 488  nm, 580  nm software (Leica, Mannheim, Germany). The fluores- and 635 nm of the white light laser, respectively. Emis- cence intensity curves were measured along the cell sion was captured through band pass settings 430– nucleus within a region of interest (ROI) and the chro- 470  nm, 505–550  nm, 590–620  nm and 648–720  nm, matogram was normalized using the quantitative tools respectively. Depletion of STAR 580 and AF 647 was of the LASX-software tool (Leica, Mannheim, Ger- performed with the 775 nm depletion laser. The power many). For the 3-dimensional imaging, several z-stacks of the depletion laser was optimized for each dye to (n = 10) of 1  µm step size were recorded and the obtain highest resolution while minimizing bleach- 3-dimensional image was generated using the LASX- ing. Imaging conditions were fine-tuned on several software tool. cells before application of the optimized settings for The track analyser of the Hyugens object tracker wiz - final images. Each dye was imaged in sequential scans ard was used to study the 3-dimensional motion of the to avoid spectral overlaps. While hybrid detector gain nuclear bodies of cells that were previously transfected was set to 100%, excitation laser intensity was set such with and without PML. Therefore, ROIs containing to prevent pixel saturation. Images were obtained nuclear bodies only or background only were selected for using a pixel dwell time of 100  ns. Photon time gating the tuning of the detection filters via linear discrimina - was employed by collecting lifetimes between 0.3 and tion analysis (LDA) and the subsequently tracking of the 6.0  ns. To compensate for inevitable signal intensity M arks et al. acta neuropathol commun (2021) 9:66 Page 5 of 16 loss during STED acquisition, the excitation laser All analysed cell types exhibited the same pheno- power was set three–fivefold higher than in conven- types—from cells with many tiny aggregates (Fig.  1B, tional confocal mode. When using STED channels, the arrow) to those with a few large speckle-like structures pinhole was set 1.0 Airy Units. In non-STED channels (Fig.  1B, arrowhead; an overview image in Additional the pinhole was set to 0.49 Airy Units to allow for sub- file  1: Fig. S1). Live cell imaging demonstrated a phe- Airy super-resolution confocal microscopy according notype of the complexes resembling a highly dynamic to the HyVolution II mode of the Leica SP8 microscope circular structure, suggesting being membrane coated system. All images were deconvolved with Huygens aggregates. However, electron microscopy analy- Professional Software (Scientific Volume Imaging B.V., sis (Fig.  1C; Additional file  1: Fig. S2) as well as Cell- Hilversum, The Netherlands) using the deconvolu- Mask membrane stain (Additional file  1: Fig. S3) tion pre-settings in Huygens software applying Classic argued against this hypothesis and rather revealed a Maximum Likelihood Estimation (CMLE) algorithms. donut-like shape of the intranuclear aggregates with an electron-dense border and an electron-poor cen- Cell profiler data analysis tre. According to present literature [9–11], a potential Brain tissue slides acquired from patients with Alz- function of the APP-dependent nuclear aggregates is heimer’s disease in different stages of severity, were the modulation of gene expression in dependence of stained with Thioflavin, DAPI and anti-PML antibod- yet unknown stimuli. In order to test this hypothesis, ies (as described before). The slides were imaged using transfected cells were fixed and stained with an anti- the Leica TCS SP8 confocal microscope system with a Histone3-K9ac antibody recognizing transcriptional 100× oil objective (HC PL APO CS2 100×/1.40 Oil) active loci. Indeed, high resolution STED indicated and tile scans (10 × 10 tiles) containing amyloidogenic active gene expression (positive staining in red) within plaques were recorded. Additionally, several plaque- large ring-like structures (Fig.  1D, arrows; Additional free regions were captured as control (n = 9 from 3 file 1: Fig. S4). individuals). PML bodies and cell nuclei were identi- fied and counted using the CellProfiler Software. In The nuclear APP‑CT complex associates to two tumour the tile scans, rectangular areas around the plaques suppressor proteins were defined as ‘plaque near’ (n = 31 from 8 indi- Next, we aimed to identify the protein composi- viduals) and the surrounding area as ‘plaque distant’ tion of these structures in more detail. Therefore, we (n = 28 from 7 individuals), not including cells close to extracted a set of proteins from the literature revealing the edge of the tile scans. The acquired data from Cell- a nuclear phenotype similar to the APP-CT50-depend- Profiler were exported to Excel, sorted in those two ent aggregates, which resulted in a list of the follow- groups and compared regarding the amount of PML ing proteins: Daxx, H2A, HIPK2, HP1ß, p53, PML, bodies inside the cells and the percentage of cells con- UBE2D2, and WRN. For all of these proteins, vectors taining those aggregates. encoding the respective candidate DNA sequence fused to a fluorescent protein cassette were cloned and Results co-transfected with expression constructs for the APP- APP‑CT induces gene expression‑active aggregates CT50 complex (APP-CT50/FE65/TIP60). The vast with donut‑like shape in a variety of cells majority of our candidates showed no co-localization Our study was initiated by studying the presence of with two exceptions. One protein has been described the potential nuclear APP signalling pathway in vari- before to be part of nuclear bodies [21] and to bind to ous cell lines including primary neurons (Fig.  1). As BLM [8, 22], as well as to reveal an unstable interac- published earlier, co-expression of FE65/TIP60 (omit- tion to TIP60 [23]: the tumour suppressor protein p53. ting APP-CT50) is sufficient to establish the nuclear Indeed, co-expression of p53-EGFP, FE65-mCherry dot-like phenotype [6, 8]. In order to not overwhelm and TIP60-BFP demonstrated a strong co-localization the cell with unnecessary expression constructs, we in nuclear aggregates (Fig.  2A). This was validated passed on APP-CT50 expression for some of the sub- by profiling of the individual fluorescence intensi- sequent experiments. Indeed, many different cell lines ties (Fig.  2B), all three fluorescent signals revealed the including cancer cells, fibroblasts and primary neu- same intensity course along the dotted line with peaks rons revealed the typical dot-like phenotype upon at position I, II, and III (Fig.  2B). Omitting FE65 co- co-expression of the complex components (Fig.  1A). expression demonstrated p53 co-localization to TIP60 Marks et al. acta neuropathol commun (2021) 9:66 Page 6 of 16 Fig. 1 Dynamic nuclear aggregates are present in various cells, lack a membrane coating and are transcriptionally active. (A) Upon co-expression of FE65-EGFP (green) and TIP60-HA (w/o fluorophore, stained using anti-HA tag antibody (red)), nuclear aggregates in various sizes are generated in multiple cell lines including neurons (for neurons no co-staining was done as a mCherry vector was additionally co-transfected to identify neuronal cell structure). Transfected vectors with respective fluorophore (EGFP, mCherry) are indicated. (B) Every cell type used in A demonstrated cells with many tiny (arrow) or few large spheres (arrowhead) (or transition states) supporting the hypothesis of sphere fusion over time (blue, APP-CT50; red, FE65-mCherry, TIP60-HA un-stained; a different overview image is also given in Additional file 1: Fig. S1). (C) Transmission electron microscopy of FE65/TIP60-HA transfected cells revealed an electron-dense ring structure (additional image in Figure S2). However, there was no evidence for a membrane sheath. Additionally, these results were also confirmed by a CellMask staining (Additional file 1: Fig. S3). (D) High-resolution STED imaging revealed that the inner core of the aggregates is positive for anti K9 acetylation histone 3 antibody staining (red, RFP) supporting the hypothesis of active gene expression within the aggregates M arks et al. acta neuropathol commun (2021) 9:66 Page 7 of 16 speckles alone as well (Fig.  2C), which is in line with dynamics of the nuclear complexes was compelling earlier results [23]. To further confirm the interac- and pointed to a spatial-temporally highly organized tion of p53 with APP-CT50-depending complexes, co- mechanism. Direct co-localization of PML with the immunoprecipitation assays (co-IPs) were performed APP-CT50 fragment (w/o FE65/TIP60 co-expression) (Fig. 2D). Sample conditions were selected as indicated was also observed in a minor percentage of trans- in the input blot (Fig.  2D, left). Co-IP (Fig.  2D, right) fected cells (not shown), however, the main phenotype using anti-HA tag antibody (TIP60-HA) revealed pre- revealed a uniform APP-CT50 signal without accu- cipitation of FE65 (as expected, Fig.  2D, white arrow) mulation in nuclear aggregates (Fig.  3A). Neverthe- and of p53 (red arrow), which validated its participa- less, cytosolic PML might also be bound to APP-CT50. tion in the complex. The TIP60/p53 interaction (inde- Additional expression of FE65 did not change the main pendent of FE65, in agreement to Fig.  2C) could be phenotype of uniformly localized APP-CT50 (Fig.  3B). confirmed by precipitating TIP60 via a GFP tag anti- Omitting TIP60 co-expression caused the generation body (white arrowhead). In addition to the p53-EGFP of a PML/APP-CT50/p53 complex (Fig.  3C). In con- signal, the presence of co-precipitated endogenous trast, the DNA helicase BLM is not co-localized to p53 was observed (3rd land, red arrowhead). Endog- PML (Fig.  3D). A more detailed analysis, utilizing 3D enous p53 is also detectable in conditions with TIP60 confocal imaging, revealed that APP-CT50 and PML co-expression (e.g. lane 7, blue arrow) suggesting an are associated with each other in all nuclear aggre- interaction of the (endogenous) tumour suppressor gates (Fig.  3E). Finally, the PML/APP-CT50 interac- protein p53 with the histone acetyltransferase TIP60 tion was confirmed by a co-immunoprecipitation assay in nuclear aggregates. (Fig. 3F). Co-expression of APP-CT with either PML1- The second candidate was a tumour suppressor pro- HA or PML1-myc demonstrated precipitation of PML tein as well, the promyelocytic leukemia protein (PML) (detection via anti-myc or anti-HA antibody, Fig.  3F, [13]. Co-expression of PML-EGFP, BFP-APP-CT50, white arrow) upon anti-GFP IP (independent experi- FE65-mCherry and TIP60 (w/o fluorophore) revealed ment given in Additional file  1: Fig. S5). This was true a close association of PML bodies to one or two APP- for two different APP-CT isoforms with 50 and 57  aa CT50 aggregates (Fig.  2E, zoom 1), which was also in length. High-resolution STED imaging further spec- confirmed by the fluorescence intensity line scan given ified the co-localization of PML and APP-CT50 within in Fig.  2E on the right. Indeed, the green PML peak is the PML bodies. APP-CT50 was either uniformly accompanied by two peaks of APP-CT50 (blue) and distributed or concentrated at the inner wall of the FE65 (red). Other cells of the same condition (zoom nuclear body (Fig. 3G). 2nd line) demonstrated association of three APP- CT50 complexes, whereas another phenotype (zoom 3rd line) demonstrated large APP-CT50 complexes with incorporated PML bodies. Collectively, the high (See figure on next page.) Fig. 2 The nuclear APP-CT50 complex associates to p53 and PML. (A) FE65 (red, mCherry) and TIP60 (blue, BFP), which are known to co-localize and form a dot-like structure, were transfected into HEK293T cells. With the addition of the tumour suppressor protein p53 (green, EGFP), the same pattern could be revealed (for example see white arrows), proving further co-localization and interactions of the proteins. (B) Co-localization of the protein complex is proven by tracking of the fluorescence intensity along the indicated arrow, which revealed peak intensities of each component (green: p53-EGFP, red: FE65-mcherry, blue: TIP60-BFP) supporting the association of all components within one complex. (C) Omitting FE65 and transfecting only p53 and TIP60 also revealed a co-localization (for example see white arrows), proving that the complex is independent of FE65. (D) P53 interaction with the APP-CT50 complex (FE65-EGFP/ TIP60-HA co-transfected to p53-EGFP versus control (EGFP-NLS, nuclear localization sequence)) was validated by co-immunoprecipitation (left side: input blot, right side: elution blot). IP using anti-HA tag antibody (against TIP60-HA) revealed precipitation of FE65 (white arrow) as well as of p53 (red arrow). Respective controls did not show a co-precipitation. TIP60-HA precipitation also occurred using anti-GFP as bait in well agreement to results obtained in part C (white arrowhead). Notably, high levels of endogenous p53 co-eluted in the same condition (red arrowhead), whereas a moderate endogenous p53 signal was observable in control conditions (red arrow). (E) A second tumour suppressor protein was identified to associate with the APP-CT50 complex: the promyelocytic leukemia protein PML. Different phenotypes of association were observed, e.g. one or two APP-CT50 (blue, BFP)/FE65 (red, mCherry) dots ( TIP60-HA was co-transfected w/o fluorophore) associated with a single PML (green, EGFP) aggregate (first and second zoom-in row, compare fluorescence intensities). Alternatively, large APP-CT50/FE65 complexes with enclosed PML-dots were found (third row) Marks et al. acta neuropathol commun (2021) 9:66 Page 8 of 16 e M arks et al. acta neuropathol commun (2021) 9:66 Page 9 of 16 APP‑CT50 depending complexes drive PML complex 96  h (middle row), whereas expression of PML/FE65/ generation that are also present in the aged human brain APP-CT50/TIP60 demonstrated nuclear body forma- To further investigate this interaction, live cell imag- tion already after 24  h (below row). After 48  h, some ing experiments were performed in HEK293 cells cells revealed formation of super-aggregates within with ectopic expression of the aggregate components. the nucleus (white arrow). We subsequently aimed to Expression of FE65-EGFP, TIP60 (w/o fluorophore) investigate whether these complexes are present in the revealed highly dynamic ring-like structures moving human brain as well (Fig. 4E). Human hippocampal fro- throughout the nucleoplasm (Fig.  4A, arrowhead, con- zen brain samples from 15 AD patients were used to focal image, structures coloured according to z-level; study the co-localization of APP-CT50 and PML. We full video given in the Additional file  2). To investigate observed a strong co-localization of PML and APP- the influence of PML in these dynamics, the complexes CT50 in the human brain (Additional file  1: Fig. S6, an (based on the FE65-mCherry signal) were tracked in antibody raised against APP 643–695 was used in order APP-CT50/FE65/TIP60-transfected cells with versus to identify APP-CT50 in the nucleus). Tracking of the without PML co-expression (Fig.  4B). The velocity of fluorescence intensity along the white arrow (includ - the nuclear structures in the presence of PML (Fig. 4B, ing three aggregates) validated the strong co-localiza- diagram  1, red graph) was significantly reduced com - tion. We also investigated the localization of FE65 in pared to the condition without PML (black graph). the same manner (Fig. 4F) and demonstrated that FE65 Moreover, the average distance from the track ori- co-localized with PML in the human brain as well. gin was significantly less in PML co-expressing cells Both co-localizations were evident in the brains of AD (Fig.  4B, diagram 2). The mean speed was 0.38 in PML patients with different Braak stages. As all samples were versus 0.76  µm/s in non-PML co-expressing cells obtained from individuals older than 65 years, we next (Fig.  4B, diagram  3). These results suggest a mutual aimed to study whether this phenotype is age-depend- trapping function of APP-CT50 aggregates and PML, ent. To study this question in a human 3D model, we for which Fig.  4C shows the typical phenotype of up differentiated iPS cells to cerebral organoids (Fig.  4G). to three APP-CT50 aggregates (red) bound to a single In contrast to the aged human brain, immunofluores - PML body (green). In order to understand the condi- cence analysis demonstrated no co-localization of APP- tions for aggregation in more detail, we monitored CT50 with PML in cerebral organoid sections. transfected cells over time (Fig. 4D). Pure PML expres- sion (upper row) revealed a homogenous distribution Reduction of PML bodies occurs in human hippocampal of the protein within the whole cell (cytoplasm and brain areas with high plaque load nucleus). Co-transfection of APP-CT50 (blue) and In order to study a potential pathophysiological rel- FE65 (red) caused late aggregate formation mostly after evance, we examined human brain tissue in more detail (See figure on next page.) Fig. 3 PML forms container-like structures within the complex and precipitates with APP-CT. (A) A direct co-localization of APP-CT (blue, BFP) with PML (green, EGFP) was only observed to some extent (cells with nuclear aggregates), but most cells revealed a uniform staining pattern of APP-CT50 and PML in the cytosol and nucleus. (B) Additional co-expression of FE65 (red, mcherry) enriched the aggregation of PML in the nucleus, but a strong co-localization to APP-CT50/FE65 was not observable in the imaging study. (C) Additional transfections showed that p53 (red, mCherry) is part of the PML aggregates in the nucleus that also contained APP-CT50, as indicated by the visible co-localization (for example see white arrows). (D) Notably, another suspected binding partners for PML like the DNA helicase BLM (blue, BFP), which was identified as binding protein in the APP-CT50 complex, is not co-localized with the PML aggregates. (E) Confocal 3D imaging validated the co-localization of PML (green, EGFP) and APP-CT50 (red, mCherry) in the nucleus (FE65/TIP60 were co-transfected w/o fluorophore). (F) Interaction of APP-CT with PML was shown using co-immunoprecipitation assay. Precipitation using anti-GFP antibody revealed detection of APP-CT-EGFP (as expected) as well as PML (PML1 isoform was used, white arrow). This was true for two different APP-CT isoforms (APP-CT50 and CT57), whereas control conditions revealed no unspecific co-precipitation. The before mentioned isoforms of APP’s c-terminal domain differ in their respective amino acid length generated through ε-cleavage. While APP-CT50 represents the most common form, APP-CT57 is less common.Results were the same for two different PML1 tags: in the left panel of blots PML1-HA was used, whereas PML1-myc was used in the right panel. (G) High-resolution STED imaging specified the localization of APP-CT50 (green, EGFP) within the PML bodies (red, mCherry). Different phenotypes were evident, either with a uniform localization within the bodies or with APP-CT50 signal at the inside wall of the PML body Marks et al. acta neuropathol commun (2021) 9:66 Page 10 of 16 M arks et al. acta neuropathol commun (2021) 9:66 Page 11 of 16 (Fig.  5). As the human hippocampus belongs to those PML bodies compared to all cells counted (cells with- brain areas revealing early pathological AD features, out any PML body, cells with more than 5 PML bod- we studied the extent of PML bodies in the Cornu ies), revealed a significant (p < 0.05) reduced percentage Ammonis areas 1 or 3 (CA1, CA3). CA regions were of PML positive nuclei in areas with high plaque load not evident for all human brain samples due to differ - (Fig. 5F, plaque) compared to plaque free areas (Fig. 5F, ent quality (different post-mortem times, preparation no plaque). Tile-scans from individuals without any artefacts), thus we limited our analysis to CA1 or CA3 neurodegenerative pathology (controls) demonstrated areas, which were distinctly assignable, e.g. Figure  5A higher percentage of PML positive nuclei compared to corresponds to a sample with CA1 assignment, but not the “no plaque” group (p < 0.05). More detailed analyses CA3. Haematoxylin Eosin (HE) staining of human hip- demonstrated that the observed significance is particu - pocampal frozen sections (samples from 15 AD indi- larly caused by cells containing only one or two PML −5 viduals with different Braak stages) was used to identify bodies per nucleus (Fig. 5G, *p < 0.05, **p < 10 ). the specific areas. Parallel sections (from the same indi - vidual) were used for low-resolution tile-scale imag- Discussion ing (DAPI channel, Fig.  5A) to allocate hippocampal The amyloid precursor protein is a ubiquitously areas according to HE staining. For subsequent detailed expressed protein, thus it is not surprising that many analysis, confocal tile-scan imaging was used includ- different cell types are capable to induce APP-CT ing 5 × 5 tiles and 5 z-stacks (Fig.  5B). Afterwards, (APP-CT50) signalling and to set up nuclear aggre- maximum projection algorithm was applied result- gates. FE65 was initially reported to be brain tissue- ing in high-resolution imaging enabling identification specific, and indeed, expression analysis suggested of the number of PML bodies in each nucleus over an FE65 to be a neuronal protein [24–26]. TIP60 is also area of 800 × 800  µm within CA1 or CA3 (Fig.  5C). ubiquitously expressed with the highest amounts in This imaging pipeline was further extended to identify testis and placenta. Thus, we conclude that APP-CT50/ areas of high plaque load (within CA1 or CA3) using FE65/TIP60 signalling is a ubiquitous pathway with a Thioflavin co-staining (Fig.  5D). In total, we scanned preference in neuronal cells due to neuron-specific 68 areas using this approach. All tile-scans were sub- FE65 expression. Notably, APP-CT50 dissociation and sequently processed using CellProfiler software in nuclear translocation was described to predominantly order to identify and extract nuclei (Fig.  5E, upper occur through the amyloidogenic processing pathway row) and to detect and count the number of PML bod- [27]. This hypothesis is further supported by findings ies within the extracted nuclei (Fig.  5E, bottom row). in APP-CT over-expressing mice revealing neuronal Data analysis of all cells containing between 1 and 5 network abnormalities [28]. APP-CT has an impact (See figure on next page.) Fig. 4 Highly mobile APP-CT50-depending complexes that are also present in the aged human brain drive PML complex generation. (A) Expression of APP-CT50/ FE65/TIP60-HA in HEK293 cells reveals a highly mobile complex moving three-dimensionally in the cellular nucleus. Ring-like structures were coloured (orange to yellow) according to z-level (confocal microscopy). The indicated structure (white arrowhead) revealed time-dependent movement. The corresponding video is given in the supplement. (B) Movement of the individual aggregates was tracked using Huygens object tracker software. Transfection in HEK293 cells included APP-CT50/FE65-mCherry/TIP60-HA with and without EGFP-PML co-expression. The FE65-mCherry signal was used for tracking, revealing lower speed in cells co-expressing PML (first diagram). In addition, the distance from the track origin (at time point 0) was analysed. Co-expression of PML revealed significantly lower distances pointing to mutual trapping of both complexes. The mean speed was 0.38 in PML versus 0.76 µm/s in non-PML co-expressing cells (last diagram). (C) This part reveals a representative image demonstrating the complex generation of APP-CT50/FE65/TIP60 (red) and PML (green aggregates). (D) Time-dependent generation of nuclear APP-CT50/PML aggregates. PML (green, EGFP) expression revealed a uniform distribution within the nucleus and cytosol (first row). Co-expression of FE65 (red, mCherry) and APP-CT50 (blue, BFP) caused initial aggregate formation after 48 h (middle row). Additional expression of TIP60-HA (w/o fluorophore) (last row) showed early generation of nuclear aggregates after 24 h. 48 h after transfection large nuclear aggregates were observed (white arrow). (E) PML (green, FITC) and APP-CT50 (red, TRITC) co-localization was studied in human brain sections. In total, 15 human hippocampal sections were analysed (different Braak stages). Confocal tile-scan imaging (5 z-stacks, then fused by maximum projection algorithm) revealed strong co-localization of PML with APP-CT50. As in cell culture experiments, nuclei containing many small aggregates (arrow) as well as nuclei with larger aggregates (arrowhead) were evident. Co-localization is further shown by intensity tracking of both fluorescent channels in the diagram for a nucleus along the dotted white arrow. (F) Similarly, co-staining of PML (green, FITC) and FE65 (red, TRITC), which confirmed the association of both proteins in the nuclei of the human brain, was performed. (G) In order to address the question whether co-localization also occurs in non-aged tissue, we differentiated human cerebral organoids from induced pluripotent stem cells. Embryonic bodies were embedded in Matrigel at day 11 followed by neuronal induction to generate organoids, which were analysed after 30 days in culture (seeding at day 0). Staining of cryosections failed to demonstrate co-localization of APP-CT50 (red, TRITC) and PML (green, FITC) Marks et al. acta neuropathol commun (2021) 9:66 Page 12 of 16 APP M arks et al. acta neuropathol commun (2021) 9:66 Page 13 of 16 on gene expression [9–12], and our findings of posi - secondary antibody used for PML staining, suggest- tive histone 3 K9 acetylation in the core of the APP-CT ing a localization of APP-CT50 and FE65 within PML aggregates further support this hypothesis. The donut- spherical aggregates, which is in good agreement to like structure of the nuclear aggregates fits to this func - other reports [29]. tion as well, assuming that DNA is incorporated in (or Relevance of APP-CT50/FE65/PML aggregates for associated to) the spherical aggregates as already shown the pathophysiology of AD was finally demonstrated for cells in G2 phase [29]. Thus, APP-CT50 aggregates by tile-scan imaging of hippocampal CA1 and CA3 might correspond to DNA incubation containers possi- areas. Our analysis revealed highly significant results bly modifying DNA in a way to change expression. demonstrating a reduction in the number of PML bod- In order to better understand the nuclear APP-CT50- ies in nuclei close to AD relevant hot spots with high depending aggregates, we studied their composition plaque load. Assuming that these areas also corre- and identified the two tumour suppressor proteins p53 spond to elevated APP cleavage, we conclude that APP and PML as additional components. According to our nuclear signalling involving the adapter protein FE65 is results in HEK293 cells, this interplay is strongly driven correlated to AD pathology. Expression of APP-CT50, by the APP-CT50 nuclear translocation, as pure PML FE65, TIP60 and PML in HEK293T cells caused a time- expression revealed a rather homogenous cellular dis- dependent fusion of the nuclear aggregates. Thus, APP- tribution. The complex generation follows a temporally driven fusion of PML aggregates may occur in the AD organized scheme with generation of small aggregates brain in a similar fashion. Certainly, further studies are at an early phase and few large nuclear complexes at pivotal to understand the consequences of APP to PML later time points. As these aggregates were evident in body generation, fusion, and, in particular, their impact the human brain of aged patients but not in cerebral on neurodegeneration and dependence on different organoids, we conclude that the APP-CT50 nuclear sig- Braak stages. nalling is age-dependent and potentially of relevance for the pathophysiology of Alzheimer’s disease. Stain- ing of APP-CT50 and FE65 was only successful using a Conclusion sophisticated protocol with defined order of antibody APP-CT50 signal transduction is of high relevance for incubation, meaning that incubation with the second- AD and causes a reduction in the number of PML bod- ary antibody for PML detection precluded the iden- ies in nuclei close to AD relevant hot spots. The com - tification of APP-CT50 or FE65 in the complex. In position of the nuclear aggregates includes APP-CT50, contrast, usage of the secondary antibody as the last and tumor suppressor proteins PML and p53. Aggre- step (after application of the Avidin-TRITC complex gates are associated to gene expression changes with to identify APP-CT50 of FE65) was successful to reveal putative impact in neurodegeneration. co-localization. Thus, accessibility of the epitopes of APP-CT50 or FE65 was presumably masked by the (See figure on next page.) Fig. 5 Significant enrichment of PML bodies in human brain areas with high plaque load. (A) Haemotoxylin eosin (HE) staining of human hippocampal sections was used to define Cornu Ammonis (CA) 1–3, Gyrus dentatus (GD), and Plexus areal. Parallel sections were used for PML immunofluorescence staining. DAPI co-staining was used for low-resolution tile-scan imaging and allocation of specific CA areas according to the initial HE staining. (B) Hippocampal CA1 or CA3 areas were used for high-resolution tile scan imaging. A single scan included 5 × 5 images with 5 z-stacks, which were subsequently combined using maximum projection. (C) A representative high-resolution (100 × objective) confocal tile-scan image demonstrates identification of PML bodies in DAPI counter-stained nuclei. (D) Tile-scan imaging was then established with Thioflavin co-staining to identify areas with high plaque load in the human hippocampus (CA1 or CA3). (E) CellProfiler software was used to automatically annotate and extract nuclei (DAPI channel in grey scale) from the image (upper row). Subsequently, PML body identification and quantification was done in the extracted nuclei. (F) All nuclei containing 1 to 5 PML bodies were used to determine the percentage of PML positive nuclei (y axis). Hippocampal areas with high plaque load (plaque, average = 36.2%) revealed a significant lower percentage of PML positive nuclei compared to areas without plaque (no plaque, average = 44.1%) (p < 0.05; every dot (left to each bar) indicates a single tile-scan experiment; for areas with high plaque load 31 tile-scans were analysed and quantified, 28 for no plaque, 9 for control). Tile-scans of control sections (individuals without any plaque pathology, average = 53.8%) revealed again higher percentage of PML positive nuclei compared to “no plaque” areas (p < 0.05). (G) Detailed analysis of cells containing one up to ten PML bodies per nucleus revealed significant differences. Nuclei with a single PML body were underrepresented (p < 0.05) in areas with high plaque load versus areas without plaques. Difference to control tile scans were highly significant −5 (p < 10 ). In addition, nuclei with two PML bodies (red bars) also revealed highly significant differences to the control. Representative images for nuclei containing one, two, or three PML bodies are given Marks et al. acta neuropathol commun (2021) 9:66 Page 14 of 16 g M arks et al. acta neuropathol commun (2021) 9:66 Page 15 of 16 Abbreviations Received: 4 March 2021 Accepted: 29 March 2021 AD: Alzheimer’s disease; APP-CT: Amyloid precursor protein c-terminal domain; CA: Cornu ammonis; CMLE: Classic Maximum Likelihood Estimation; CO: Cerebral organoids; FITC: Fluorescein isothiocyanate; GFP: Green fluo - rescent protein; HE: Haematoxylin Eosin; HyD: Hybrid detector; iPS: Induced pluripotent stem (cell); LDA: Linear discrimination analysis; PBS: Phosphate References buffered saline; PML: Promyelocytic leukemia protein; PVDF: Polyvinylidene 1. Hardy J, Selkoe DJ (2002) The amyloid hypothesis of Alzheimer’s difluoride; ROI: Region of interest; STED: Stimulated emission depletion; TRITC: disease: progress and problems on the road to therapeutics. Science Tetramethylrhodamine. 297(5580):353–356. https:// doi. org/ 10. 1126/ scien ce. 10729 94 2. Bertram L, Lill CM, Tanzi RE (2010) The genetics of Alzheimer disease: back to the future. Neuron 68(2):270–281. https:// doi. org/ 10. 1016/j. neuron. Supplementary Information 2010. 10. 013 The online version contains supplementary material available at https:// doi. 3. Lott IT, Head E, Doran E, Busciglio J (2006) Beta-amyloid, oxidative stress org/ 10. 1186/ s40478- 021- 01174-x. and down syndrome. Curr Alzheimer Res 3(5):521–528 4. Chasseigneaux S, Allinquant B (2012) Functions of Aβ, sAPPα and sAPPβ : similarities and differences. J Neurochem 120(Suppl 1):99–108. https:// Additional file1. This file contains supplementary Figures and legends doi. org/ 10. 1111/j. 1471- 4159. 2011. 07584.x S1–S7. 5. Hardy JA, Higgins GA (1992) Alzheimer’s disease: the amyloid cascade hypothesis. Science 256(5054):184–185 Additional file2. This file contains a supplementary movie to Figure 4 A. 6. Bukhari H, Glotzbach A, Kolbe K, Leonhardt G, Loosse C, Müller T (2017) Small things matter: implications of APP intracellular domain AICD Acknowledgements nuclear signaling in the progression and pathogenesis of Alzheimer’s We would like to thank the Electron Microscopy Unit (EMU), Imaging Center disease. Prog Neurobiol 156:189–213. https:// doi. org/ 10. 1016/j. pneur Essen (IMCES), University Hospital Essen as well as the Imaging Centre Campus obio. 2017. 05. 005 Essen (ICCE) for their support. We would like to thank Dr. Mavrommatis and 7. Radzimanowski J, Simon B, Sattler M, Beyreuther K, Sinning I, Wild K PD Zaehres, Ruhr-University Bochum, Medical Faculty, Institute of Anatomy for (2008) Structure of the intracellular domain of the amyloid precursor providing CD34 iPSC and help with iPSC culture. Instrument Leica TCS SP8X protein in complex with Fe65-PTB2. EMBO Rep 9(11):1134–1140. https:// FALCON was obtained through DFG funding (Major Research Instrumentation doi. org/ 10. 1038/ embor. 2008. 188 Programme as per Art. 91b GG, INST 20876/294-1 FUGG). Instrument Leica 8. Schrötter A, Mastalski T, Nensa FM, Neumann M, Loosse C, Pfeiffer K, TCS SP8 was supported by an instrument grant from the German Research Magraoui FE, Platta HW, Erdmann R, Theiss C, Uszkoreit J, Eisenacher M, Foundation (INST 213/886-1 FUGG). Meyer HE, Marcus K, Müller T (2013) FE65 regulates and interacts with the Bloom syndrome protein in dynamic nuclear spheres - potential Authors’ contribution relevance to Alzheimer’s disease. J Cell Sci 126(Pt 11):2480–2492. https:// DM, NH, LB, SM, MW, LG, PH, VB, ES, TM did the experiments. KW corrected the doi. org/ 10. 1242/ jcs. 121004 manuscript. The study was designed by SKK and TM. Supervision of the whole 9. Grimm MO, Mett J, Stahlmann CP, Grösgen S, Haupenthal VJ, Blümel T, project was done by TM. All authors read and approved the final manuscript. Hundsdörfer B, Zimmer VC, Mylonas NT, Tanila H, Müller U, Grimm HS, Hartmann T (2015) APP intracellular domain derived from amyloidogenic Funding β- and γ-secretase cleavage regulates neprilysin expression. Front Aging Open Access funding enabled and organized by Projekt DEAL. This work was Neurosci 7:77. https:// doi. org/ 10. 3389/ fnagi. 2015. 00077 supported by funding from German Research Foundation (DFG, Germany 10. Kim HS, Kim EM, Lee JP, Park CH, Kim S, Seo JH, Chang KA, Yu E, Jeong MU3525/3-2), an instrument grant from DFG (INST 213/886-1 FUGG) and SJ, Chong YH, Suh YH (2003) C-terminal fragments of amyloid precursor Mercator Research Center Ruhr (Mercur, Germany, Pr-2016-0010). protein exert neurotoxicity by inducing glycogen synthase kinase-3beta expression. FASEB J 17(13):1951–1953. https:// doi. org/ 10. 1096/. fj 03- 0106f je Declarations 11. von Rotz RC, Kohli BM, Bosset J, Meier M, Suzuki T, Nitsch RM, Konietzko U (2004) The APP intracellular domain forms nuclear multiprotein Ethical approval and consent to participate complexes and regulates the transcription of its own precursor. J Cell Sci An ethical vote for the current work is available (17-6138, Ethical Commission 117(Pt 19):4435–4448. https:// doi. org/ 10. 1242/ jcs. 01323 of the Ruhr- University Bochum). 12. Zhou F, Gong K, van Laar T, Gong Y, Zhang L (2011) Wnt/β-catenin signal pathway stabilizes APP intracellular domain (AICD) and promotes its tran- Consent for publication scriptional activity. Biochem Biophys Res Commun 412(1):68–73. https:// All authors agreed to publish the data given in the manuscript and to provide doi. org/ 10. 1016/j. bbrc. 2011. 07. 040 data and material to the scientific community. 13. Lallemand-Breitenbach V, de Thé H (2018) PML nuclear bodies: from architecture to function. Curr Opin Cell Biol 52:154–161. https:// doi. org/ Availability of supporting data 10. 1016/j. ceb. 2018. 03. 011 In case of acceptance, supporting data can be accessed at the Cell image 14. Van Damme E, Laukens K, Dang TH, Van Ostade X (2010) A manually library (http:// www. celli magel ibrary. org/ home). curated network of the PML nuclear body interactome reveals an impor- tant role for PML-NBs in SUMOylation dynamics. Int J Biol Sci 6(1):51–67. Competing interests https:// doi. org/ 10. 7150/ ijbs.6. 51 There are no competing interests. 15. Dellaire G, Bazett-Jones DP (2004) PML nuclear bodies: dynamic sensors of DNA damage and cellular stress. BioEssays 26(9):963–977. https:// doi. Author details org/ 10. 1002/ bies. 20089 Department of Molecular Biochemistry, Cell Signalling, Faculty of Chemistry 16. Chang HR, Munkhjargal A, Kim MJ, Park SY, Jung E, Ryu JH, Yang Y, Lim and Biochemistry, Ruhr University Bochum, Bochum, Germany. Institute JS, Kim Y (2018) The functional roles of PML nuclear bodies in genome of Psychiatric Phenomics and Genomics (IPPG), University Hospital, LMU maintenance. Mutat Res 809:99–107. https:// doi. org/ 10. 1016/j. mrfmmm. Munich, Munich, Germany. Leibniz Institute On Aging Research, Jena, 2017. 05. 002 Germany. Department of Molecular Cell Biology, Institute of Biochemistry and Pathobiochemistry, Medical Faculty, Ruhr-University Bochum, Bochum, Germany. Department of Molecular Biology II, Center of Medical Biotechnol- ogy (ZMB), University of Duisburg-Essen, Essen, Germany. Marks et al. acta neuropathol commun (2021) 9:66 Page 16 of 16 17. Stadler M, Chelbi-Alix MK, Koken MH, Venturini L, Lee C, Saïb A, Quignon 25. Kesavapany S, Banner SJ, Lau KF, Shaw CE, Miller CC, Cooper JD, F, Pelicano L, Guillemin MC, Schindler C (1995) Transcriptional induction McLoughlin DM (2002) Expression of the Fe65 adapter protein in adult of the PML growth suppressor gene by interferons is mediated through and developing mouse brain. Neuroscience 115(3):951–960 an ISRE and a GAS element. Oncogene 11(12):2565–2573 26. Simeone A, Duilio A, Fiore F, Acampora D, De Felice C, Faraonio R, Paolocci 18. Scherer M, Stamminger T (2016) Emerging role of PML nuclear bodies in F, Cimino F, Russo T (1994) Expression of the neuron-specific FE65 gene innate immune signaling. J Virol 90(13):5850–5854. https:// doi. org/ 10. marks the development of embryo ganglionic derivatives. Dev Neurosci 1128/ JVI. 01979- 15 16(1–2):53–60. https:// doi. org/ 10. 1159/ 00011 2088 19. Lancaster MA, Knoblich JA (2014) Generation of cerebral organoids from 27. Goodger ZV, Rajendran L, Trutzel A, Kohli BM, Nitsch RM, Konietzko U human pluripotent stem cells. Nat Protoc 9(10):2329–2340. https:// doi. (2009) Nuclear signaling by the APP intracellular domain occurs predomi- org/ 10. 1038/ nprot. 2014. 158 nantly through the amyloidogenic processing pathway. J Cell Sci 122(Pt 20. Okada Y, Nakagawa S (2015) Super-resolution imaging of nuclear bodies 20):3703–3714. https:// doi. org/ 10. 1242/ jcs. 048090 by STED microscopy. Methods Mol Biol 1262:21–35. https:// doi. org/ 10. 28. Vogt DL, Thomas D, Galvan V, Bredesen DE, Lamb BT, Pimplikar SW (2011) 1007/ 978-1- 4939- 2253-6_2 Abnormal neuronal networks and seizure susceptibility in mice overex- 21. Fogal V, Gostissa M, Sandy P, Zacchi P, Sternsdorf T, Jensen K, Pandolfi PP, pressing the APP intracellular domain. Neurobiol Aging 32(9):1725–1729. Will H, Schneider C, Del Sal G (2000) Regulation of p53 activity in nuclear https:// doi. org/ 10. 1016/j. neuro biola ging. 2009. 09. 002 bodies by a specific PML isoform. EMBO J 19(22):6185–6195. https:// doi. 29. Luciani JJ, Depetris D, Usson Y, Metzler-Guillemain C, Mignon-Ravix C, org/ 10. 1093/ emboj/ 19. 22. 6185 Mitchell MJ, Megarbane A, Sarda P, Sirma H, Moncla A, Feunteun J, Mattei 22. Wang XW, Tseng A, Ellis NA, Spillare EA, Linke SP, Robles AI, Seker H, Yang MG (2006) PML nuclear bodies are highly organised DNA-protein struc- Q, Hu P, Beresten S, Bemmels NA, Garfield S, Harris CC (2001) Functional tures with a function in heterochromatin remodelling at the G2 phase. J interaction of p53 and BLM DNA helicase in apoptosis. J Biol Chem Cell Sci 119(Pt 12):2518–2531. https:// doi. org/ 10. 1242/ jcs. 02965 276(35):32948–32955. https:// doi. org/ 10. 1074/ jbc. M1032 98200 23. Tang Y, Luo J, Zhang W, Gu W (2006) Tip60-dependent acetylation of p53 Publisher’s Note modulates the decision between cell-cycle arrest and apoptosis. Mol Cell Springer Nature remains neutral with regard to jurisdictional claims in pub- 24(6):827–839. https:// doi. org/ 10. 1016/j. molcel. 2006. 11. 021 lished maps and institutional affiliations. 24. Duilio A, Zambrano N, Mogavero AR, Ammendola R, Cimino F, Russo T (1991) A rat brain mRNA encoding a transcriptional activator homolo- gous to the DNA binding domain of retroviral integrases. Nucleic Acids Res 19(19):5269–5274. https:// doi. org/ 10. 1093/ nar/ 19. 19. 5269 Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Acta Neuropathologica Communications Springer Journals

Amyloid precursor protein elevates fusion of promyelocytic leukemia nuclear bodies in human hippocampal areas with high plaque load

Download PDF
16 pages

Loading next page...
 
/lp/springer-journals/amyloid-precursor-protein-elevates-fusion-of-promyelocytic-leukemia-YwlkQA5aGD

References (30)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
eISSN
2051-5960
DOI
10.1186/s40478-021-01174-x
Publisher site
See Article on Publisher Site

Abstract

The amyloid precursor protein (APP) is a type I transmembrane protein with unknown physiological function but potential impact in neurodegeneration. The current study demonstrates that APP signals to the nucleus causing the generation of aggregates consisting of its adapter protein FE65, the histone acetyltransferase TIP60 and the tumour suppressor proteins p53 and PML. APP C-terminal (APP-CT50) complexes co-localize and co-precipitate with p53 and PML. The PML nuclear body generation is induced and fusion occurs over time depending on APP signalling and STED imaging revealed active gene expression within the complex. We further show that the nuclear aggregates of APP-CT50 fragments together with PML and FE65 are present in the aged human brain but not in cerebral organoids differentiated from iPS cells. Notably, human Alzheimer’s disease brains reveal a highly significant reduction of these nuclear aggregates in areas with high plaque load compared to plaque-free areas of the same individual. Based on these results we conclude that APP-CT50 signalling to the nucleus takes place in the aged human brain and is involved in the pathophysiology of AD. Keywords: Alzheimer’s disease, APP-CT50, PML, IPSC-derived cerebral organoids, 3D culture, Nuclear complexes, Human brain, Amyloidogenic plaques, HSV, Viral defence Introduction the generation of three fragments: (i) the secreted Increased amyloidogenic processing of the amyloid extracellular domain (sAPPβ), (ii) the β-amyloid pep- precursor protein (APP) occurs in sporadic Alzhei- tide (Aβ), and (iii) the APP C-terminal fragment mer’s disease (AD) [1], in familial AD with mutations (APP-CT). The secreted fragment (i) was reported to in APP or in its processing enzymes [2], and in trisomy provoke neurotrophic effects [4], Aβ (ii) is the main 21 patients [3]. Amyloidogenic APP processing causes component of amyloidogenic plaques [5] and APP- CT (iii) was suggested to play an important role in a nuclear signal transduction pathway [6]. APP-CT, which has been reported to exist in different isoforms *Correspondence: thorsten.t.mueller@rub.de David Marks, Natalie Heinen, Lisa Bachmann: shared first authors with 50aa in length being the most stable one (APP- Department of Molecular Biochemistry, Cell Signalling, Faculty CT50), indeed is a remarkable protein fragment as it of Chemistry and Biochemistry, Ruhr University Bochum, Bochum, is intrinsically unstructured [7]. Though, this changes Germany Full list of author information is available at the end of the article upon interaction with FE65, causing APP-CT50 to fold © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Marks et al. acta neuropathol commun (2021) 9:66 Page 2 of 16 into a three-dimensional conformation that can be Geltrex (Gibco) according to the manufacturer’s analysed by x-ray crystallography. APP-CT50 is capa- protocol, and split before reaching 70% confluency. ble to enter the nucleus establishing a protein complex HEK293T cells were seeded and incubated in DMEM consisting of additional proteins like FE65, TIP60, and (Gibco) with 10% heat inactivated FBS (Gibco), 1% BLM [8]. The presence of the histone acetyl trans- Penicillin/Streptomycin and 1% L-glutamine (Gibco), ferase (TIP60) and the DNA helicase (BLM) in the to a confluency of 70%. For overexpression assays, ster - complex points to a functional role in essential bio- ile precision cover glasses (1.5 H Marienfeld Superior) logical mechanisms such as gene expression, DNA were placed into a 24-well cell culture plate (Sarstedt) replication/damage/repair or chromatin modification. and coated with 0.01% poly-l-ornithine solution (Sigma Indeed, a variety of target genes like GSK3β, IDE, and Aldrich). The respective plasmids were transfected APP have been proposed to be APP-CT50 dependently via the K4 Transfection Kit (Biontex) according to regulated [9–12]. manufacturer’s recommendations for 24-well culture Promyelocytic leukemia nuclear bodies (PML-NBs) plates. After 24 or 48  h, cells were briefly washed with are multiprotein complexes with PML as the main DPBS (Gibco) and fixed in 4% paraformaldehyde in building component [13]. A diverse set of nuclear pro- PBS. Cover glasses were mounted with the Shandon teins have been identified as permanent or transient Immu-Mount solution (Thermo Scientific) on glass PML-NB-binding partners [14]. PML-NBs are highly slides and dried overnight at RT. For immunofluores - dynamic structures with respect to mobility, composi- cence staining, HEK293T cells were seeded in 8-well- tion, architecture, and function [15]. While their pre- µ-slide-ibi Treat (ibidi , Martinsried, Germany) and cise biochemical functions have not been elucidated transfected using calcium phosphate transfection after yet, they have been linked to many aspects of chroma- 24 h. Cells were fixed with Roti -Histofix 4% (4% phos - tin biology, including transcription, histone modifica- phate buffered formaldehyde solution; Roth, Karlsruhe, tion, repair and recombination, degradation, hence DE) for 20 min at 37 °C, and permeabilized and blocked genome maintenance [16]. Transcription of PML is with 5% normal goat serum (NGS) in 0.3% (w/v) Triton strongly upregulated by interferons and p53 [17], caus- X-100/PBS for 30  min at RT. The cells were incubated ing a significant increase in the number and size of the with primary antibodies diluted in 1% BSA/0.3% Tri- bodies. Recent studies revealed an emerging role of ton X-100/DPBS (mouse anti-HA (BioLegend, 901501; PML-NBs as coregulatory structures of both type I and 1:1000), mouse anti-myc (NEB/Cell Signalling, 2276; type II interferon responses [18]. Within this work, we 1:1500) o/n at 4  °C. For the secondary antibody stain- demonstrate that PML nuclear bodies interact with ing and the cell staining, goat-anti-mouse AF568 (Inv- highly mobile APP-CT complexes and progressively itrogen, A11004, 1:1000) was used, together with HCS form immobile large nuclear structures with relevance CellMask Deep Red Stain (ThermoFisher Scientific, for AD pathophysiology. H32721, 1:5000), Hoechst33342 (10  mg/mL in H2O, Applichem, A0741, 1:1000) in DPBS (1% BSA, 0.3% Tri- Methods ton) and incubated for 1 h at RT. Vector constructs The plasmids encoding the fusion proteins (FE65- Immunoprecipitation EGFP, FE65-mCherry, APP-CT-GFP, TIP60-EGFP, HEK 293  T cells were seeded in 10  cm dishes and co- TIP60-HA, TIP60-BMP, EGFP-PML, PML-HA, transfection was performed 24  h after seeding. Whole PML-myc, p53-EGFP, Daxx-EGFP, H2A-mTurquoise, cell extracts were prepared 24  h after transfection by HIPK2-EGFP, HP1ß-EGFP, UBE2D2-mCherry, and scraping the cells from the dish with a cell scraper, WRN-EGFP) described in this paper, were gener- washing the cell pellet in ice-cold PBS, extracting with ated using the In-Fusion HD cloning kit (Takara Bio) 1  ml interaction buffer (50  mM Tris pH 8, 150  mM according to manufacturer’s instructions or were pur- NaCl, 5  mM EDTA, 0.5% NP40, 1  mM DTT, 1  mM chased (Addgene). Amplification and purification of PMSF, 1× complete protease inhibitor cocktail), fol- the plasmids were done according to standard proto- lowed by sonication (15  s at 95% amplitude) using a cols. An overview of all constructs used in this study is Sonopuls mini20 device (Bandelin, Berlin, Germany). given in Additional file 1: Fig. S7. The lysates were centrifuged (15,000g , 15 min, 4 °C) and the supernatant was transferred to a new reaction tube. Cell culture, transfection, and immunofluorescence Input samples of the lysates were stored separately. Stem cells (iPS CD34) were cultured in StemFlex Immunoprecipitation (IP) was carried out with the (Gibco) on 35  mm dishes, coated with Matrigel or µMACS isolation kits for tagged proteins from Miltenyi M arks et al. acta neuropathol commun (2021) 9:66 Page 3 of 16 Biotec (Bergisch-Gladbach, Germany). The eluates, as attachment plate (Corning) with 500  µL Neural Induc- well as input samples of the lysates were subjected to tion (NI) -medium. Every day, half of the media was SDS-PAGE and immunoblotting. exchanged with 500  µL fresh NI-medium. On day 12, the embryoid bodies were embedded in droplets of Immunoblotting Matrigel (Corning) and incubated for 25  min at 37  °C Protein concentrations were determined using the for Matrigel polymerization. Afterwards, the droplets Bio-Rad protein assay system (Bio-Rad Laborato- were transferred to a 50  mm dish with differentiation ries, Richmond, CA). Equal amounts of protein were medium without vitamin A (DM-A) medium for fur- resolved by SDS-PAGE using a 10% acrylamide gel ther incubation at 37 °C in a 5% CO atmosphere. Four and subsequently transferred onto polyvinylidene dif- days later, the medium was changed to DM+A and luoride (PVDF) membranes (Amersham Hybond, GE the developing cerebral organoids (COs) were main- Healthcare) via the PerfectBlue tank electro blotter tained at 37  °C with 5% C O until experiments were (Peqlab, Erlangen, Germany) with 350 mA for 90 min. performed. To minimize unspecific binding, the membranes were blocked in 5% (w/v) non-fat dried milk powder Histology and immunohistochemistry in TBST for 30  min at RT. Membranes were probed The COs were removed from the media, washed with with primary antibodies against GFP (rabbit, poly- PBS, and fixed with 4% paraformaldehyde in PBS for clonal, 1:2000, Santa Cruz, sc-8334), HA tag (mouse, 90 min at 4 °C. After washing with PBS, organoids were monoclonal, 1:1000, BioLegend, 901501) and p53 incubated in 30% sucrose solution for cryoprotection at (mouse, monoclonal, 1:200, Novus Biologicals, NBP2- 4  °C overnight. The next day, the COs were embedded 29419) diluted in blocking solution overnight at 4  °C. in a 1:1 mixture of 30% sucrose and Tissue-Tek O.C.T. The membranes were washed three times with TBST, embedding medium (Science Services, SA62550-01), before they were incubated with HRP-conjugated sec- snap-frozen on dry-ice, and then stored at − 80 °C until ondary antibodies (1:10,000, NXA931 and NA934, cryosectioning. Frozen COs were sliced into 15 µm sec- GE Healthcare Europe, Freiburg, DE) also diluted in tions using a cryostat (Leica CM3050S), mounted on ™ ™ blocking solution for 1 h at RT. Visualization of bound SuperFrost slides (ThermoScientific ), and stored at antibodies occurred via enhanced chemiluminescence − 80 °C until further use. (ECL) with the ECLplus Western Blotting Substrate For immunohistochemistry, COs and brain tis- from Pierce (Rockford, IL, USA) according to the sue sections were thawed for 2  min in PBS. To apply manufacturer’s instructions. After incubation with the the biotin-avidin system used for the enhancement of substrate, the detection of the generated signal was fluorescence, the sections were first blocked with avi - carried out with the ChemiDoc MP Imaging System din for 10  min and, after washing twice with PBS for (Bio-Rad Laboratories GmbH, Feldkirchen, Germany). 4  min, blocked with biotin for 10  min. After washing twice with PBS for 4  min, sections were blocked and permeabilized in 0.1% Triton X-100, 5% goat serum in Cerebral organoids PBS for 1 h at RT, followed by incubation with primary Cerebral organoids were generated according to the antibodies in a humidified chamber overnight at 4  °C. protocol from Lancaster and Knoblich [19] with minor Primary antibodies were diluted in 0.1% Triton X-100 modifications, all media compositions remained in PBS as follows: APP-CT (mouse, Millipore MAB343, unchanged. Briefly, at day 0, iPS CD34 positive cells ™ 1:100), PML (rabbit, Novus Biologicals NB100-59787, were detached and harvested using TrypLE (Ther - 1:400), PML (mouse, Abcam ab96051, 1:200), TIP60 moFisher, Germany). Afterwards, DMEM/F12 was (mouse, Abcam Ab54277, 1:400), FE65 (mouse, Acris added to the detached cells and the cell number was AM32556SU-N, 1:400), FE65 (rabbit, Santa Cruz calculated using a Neubauer chamber. Next, 9000 cells/ sc-33155, 1:400), β-tubulin III (mouse, StemCell 01409, well were seeded into a 96-well ultra-low attachment 1:100), p53 (mouse, Novus Biologicals NB200-103, plate (Corning) with a total amount of 150  µL hESC- 1:100). Sections were incubated with the biotinylated medium (containing 4 ng/mL bFGF and 50 µM ROCK- secondary antibody (goat anti-mouse IgG Biotin, Life Inhibitor) per well. On day 3, half of the media was Technologies B-2763, 1:100) diluted in 0.1% Triton exchanged with 150 µL of hESC-medium without bFGF X-100 in PBS for 1  h at RT in a humidified chamber. and ROCK-Inhibitor. Subsequently (day 6), the embry- Following washing twice with PBS for 4  min, sections oid bodies were transferred to a 24-well ultra-low Marks et al. acta neuropathol commun (2021) 9:66 Page 4 of 16 were incubated with Avidin-TRITC (1:1000) and a non- nuclear bodies. The detection threshold was adjusted to biotinylated secondary antibody (donkey anti-rabbit measure objects with a positive generated score, com- FITC, Santa Cruz sc-2090, 1:100) diluted in 0.1% Triton puted by the software, to further discriminate against X-100 in PBS for 45 min in a humidified chamber pro - the background. The bodies were tracked within the cells tected from light at RT, and subsequently washed twice over a time span of 300  s and the speed was calculated with PBS for 4 min. using the integrated software. In case of thioflavin-S counterstaining of amyloid plaques, sections were incubated with 0.1% aqueous thioflavin-S solution, washed twice with PBS for 4 min, STED microscopy washed with 30% ethanol followed by 50% ethanol for For STED, GFP fusion proteins in fixed cells were 5  min each, and finally washed twice for 4  min with labelled with Alexa Fluor 647-coupled GFP nano- PBS. bodies (GFP-booster gb2AF647-50, Chromotek, For counterstaining of nuclei, DAPI solution Germany) at 1:100 dilution. Endogenous and overex- (0.001  mg/mL) was added to the sections for 15  min pressed PML was immunofluorescently labelled with while protected from light, then slides were washed anti-PML antibody (rabbit, ABD-030, Jena Bioscience, twice with PBS for 4 min and mounted. Germany, 1:500), followed by secondary antibody coupled with STAR 580 STED dye (goat-anti-rabbit, Imaging and tracking ST580-1002-500UG, Abberior, Göttingen, Germany, Cells were either imaged after fixation and mount- 1:100). Stained cells were embedded in ProlongGold ™ ™ ing (Shandon Immu Mount solution, Thermo Sci- with DAPI (Thermo Fisher Scientific, Germany) and entific) on glass slides or for life cell imaging directly covered with 12  mm round cover glasses (Thickness using the integrated incubation chamber of the Leica 0.17 ± 0.01  mm). Gated STED images were acquired (Mannheim, Germany) TCS SP8 microscope system on a Leica TCS SP8 STED microscope equipped with (37 °C and 5% CO ). Samples were imaged using a 63× a 100× oil objective (HC PL APO CS2 100×/1.40 Oil) water (1.2 NA) or 100× oil objective (1.4 NA). Fluo- according to protocols established for nuclear bod- rophores were excited with 405/488/514/561  nm laser ies by Okada and Nakagawa [20]. Pixel size in STED lines performing a sequential scan beginning with the acquisition was applied automatically in LAS-X soft- most red-shifted wavelength. Images were recorded ware (Leica, Mannheim, Germany) for the most red- into 1024 × 1024 images at a scan speed of 200  Hz shifted dye (AF 647), usually resulting in a pixel size with HyD detectors. Tile scans were imaged through of less than 20 × 20  nm. STED beam alignment was the selection of 800 × 800  µm areas (5 × 5 tiles) in x- performed before each imaging session between the and y-direction. Additionally, z-stacks (n = 5) of 2  µm pulsed white light laser and the 592 nm depletion laser. between each plane (8 µm in total) were recorded and DAPI, Alexa Fluor 488, Star 580 and Alexa Fluor 647 merged via the maximum projection tool in the LASX- were excited with laser lines 405  nm, 488  nm, 580  nm software (Leica, Mannheim, Germany). The fluores- and 635 nm of the white light laser, respectively. Emis- cence intensity curves were measured along the cell sion was captured through band pass settings 430– nucleus within a region of interest (ROI) and the chro- 470  nm, 505–550  nm, 590–620  nm and 648–720  nm, matogram was normalized using the quantitative tools respectively. Depletion of STAR 580 and AF 647 was of the LASX-software tool (Leica, Mannheim, Ger- performed with the 775 nm depletion laser. The power many). For the 3-dimensional imaging, several z-stacks of the depletion laser was optimized for each dye to (n = 10) of 1  µm step size were recorded and the obtain highest resolution while minimizing bleach- 3-dimensional image was generated using the LASX- ing. Imaging conditions were fine-tuned on several software tool. cells before application of the optimized settings for The track analyser of the Hyugens object tracker wiz - final images. Each dye was imaged in sequential scans ard was used to study the 3-dimensional motion of the to avoid spectral overlaps. While hybrid detector gain nuclear bodies of cells that were previously transfected was set to 100%, excitation laser intensity was set such with and without PML. Therefore, ROIs containing to prevent pixel saturation. Images were obtained nuclear bodies only or background only were selected for using a pixel dwell time of 100  ns. Photon time gating the tuning of the detection filters via linear discrimina - was employed by collecting lifetimes between 0.3 and tion analysis (LDA) and the subsequently tracking of the 6.0  ns. To compensate for inevitable signal intensity M arks et al. acta neuropathol commun (2021) 9:66 Page 5 of 16 loss during STED acquisition, the excitation laser All analysed cell types exhibited the same pheno- power was set three–fivefold higher than in conven- types—from cells with many tiny aggregates (Fig.  1B, tional confocal mode. When using STED channels, the arrow) to those with a few large speckle-like structures pinhole was set 1.0 Airy Units. In non-STED channels (Fig.  1B, arrowhead; an overview image in Additional the pinhole was set to 0.49 Airy Units to allow for sub- file  1: Fig. S1). Live cell imaging demonstrated a phe- Airy super-resolution confocal microscopy according notype of the complexes resembling a highly dynamic to the HyVolution II mode of the Leica SP8 microscope circular structure, suggesting being membrane coated system. All images were deconvolved with Huygens aggregates. However, electron microscopy analy- Professional Software (Scientific Volume Imaging B.V., sis (Fig.  1C; Additional file  1: Fig. S2) as well as Cell- Hilversum, The Netherlands) using the deconvolu- Mask membrane stain (Additional file  1: Fig. S3) tion pre-settings in Huygens software applying Classic argued against this hypothesis and rather revealed a Maximum Likelihood Estimation (CMLE) algorithms. donut-like shape of the intranuclear aggregates with an electron-dense border and an electron-poor cen- Cell profiler data analysis tre. According to present literature [9–11], a potential Brain tissue slides acquired from patients with Alz- function of the APP-dependent nuclear aggregates is heimer’s disease in different stages of severity, were the modulation of gene expression in dependence of stained with Thioflavin, DAPI and anti-PML antibod- yet unknown stimuli. In order to test this hypothesis, ies (as described before). The slides were imaged using transfected cells were fixed and stained with an anti- the Leica TCS SP8 confocal microscope system with a Histone3-K9ac antibody recognizing transcriptional 100× oil objective (HC PL APO CS2 100×/1.40 Oil) active loci. Indeed, high resolution STED indicated and tile scans (10 × 10 tiles) containing amyloidogenic active gene expression (positive staining in red) within plaques were recorded. Additionally, several plaque- large ring-like structures (Fig.  1D, arrows; Additional free regions were captured as control (n = 9 from 3 file 1: Fig. S4). individuals). PML bodies and cell nuclei were identi- fied and counted using the CellProfiler Software. In The nuclear APP‑CT complex associates to two tumour the tile scans, rectangular areas around the plaques suppressor proteins were defined as ‘plaque near’ (n = 31 from 8 indi- Next, we aimed to identify the protein composi- viduals) and the surrounding area as ‘plaque distant’ tion of these structures in more detail. Therefore, we (n = 28 from 7 individuals), not including cells close to extracted a set of proteins from the literature revealing the edge of the tile scans. The acquired data from Cell- a nuclear phenotype similar to the APP-CT50-depend- Profiler were exported to Excel, sorted in those two ent aggregates, which resulted in a list of the follow- groups and compared regarding the amount of PML ing proteins: Daxx, H2A, HIPK2, HP1ß, p53, PML, bodies inside the cells and the percentage of cells con- UBE2D2, and WRN. For all of these proteins, vectors taining those aggregates. encoding the respective candidate DNA sequence fused to a fluorescent protein cassette were cloned and Results co-transfected with expression constructs for the APP- APP‑CT induces gene expression‑active aggregates CT50 complex (APP-CT50/FE65/TIP60). The vast with donut‑like shape in a variety of cells majority of our candidates showed no co-localization Our study was initiated by studying the presence of with two exceptions. One protein has been described the potential nuclear APP signalling pathway in vari- before to be part of nuclear bodies [21] and to bind to ous cell lines including primary neurons (Fig.  1). As BLM [8, 22], as well as to reveal an unstable interac- published earlier, co-expression of FE65/TIP60 (omit- tion to TIP60 [23]: the tumour suppressor protein p53. ting APP-CT50) is sufficient to establish the nuclear Indeed, co-expression of p53-EGFP, FE65-mCherry dot-like phenotype [6, 8]. In order to not overwhelm and TIP60-BFP demonstrated a strong co-localization the cell with unnecessary expression constructs, we in nuclear aggregates (Fig.  2A). This was validated passed on APP-CT50 expression for some of the sub- by profiling of the individual fluorescence intensi- sequent experiments. Indeed, many different cell lines ties (Fig.  2B), all three fluorescent signals revealed the including cancer cells, fibroblasts and primary neu- same intensity course along the dotted line with peaks rons revealed the typical dot-like phenotype upon at position I, II, and III (Fig.  2B). Omitting FE65 co- co-expression of the complex components (Fig.  1A). expression demonstrated p53 co-localization to TIP60 Marks et al. acta neuropathol commun (2021) 9:66 Page 6 of 16 Fig. 1 Dynamic nuclear aggregates are present in various cells, lack a membrane coating and are transcriptionally active. (A) Upon co-expression of FE65-EGFP (green) and TIP60-HA (w/o fluorophore, stained using anti-HA tag antibody (red)), nuclear aggregates in various sizes are generated in multiple cell lines including neurons (for neurons no co-staining was done as a mCherry vector was additionally co-transfected to identify neuronal cell structure). Transfected vectors with respective fluorophore (EGFP, mCherry) are indicated. (B) Every cell type used in A demonstrated cells with many tiny (arrow) or few large spheres (arrowhead) (or transition states) supporting the hypothesis of sphere fusion over time (blue, APP-CT50; red, FE65-mCherry, TIP60-HA un-stained; a different overview image is also given in Additional file 1: Fig. S1). (C) Transmission electron microscopy of FE65/TIP60-HA transfected cells revealed an electron-dense ring structure (additional image in Figure S2). However, there was no evidence for a membrane sheath. Additionally, these results were also confirmed by a CellMask staining (Additional file 1: Fig. S3). (D) High-resolution STED imaging revealed that the inner core of the aggregates is positive for anti K9 acetylation histone 3 antibody staining (red, RFP) supporting the hypothesis of active gene expression within the aggregates M arks et al. acta neuropathol commun (2021) 9:66 Page 7 of 16 speckles alone as well (Fig.  2C), which is in line with dynamics of the nuclear complexes was compelling earlier results [23]. To further confirm the interac- and pointed to a spatial-temporally highly organized tion of p53 with APP-CT50-depending complexes, co- mechanism. Direct co-localization of PML with the immunoprecipitation assays (co-IPs) were performed APP-CT50 fragment (w/o FE65/TIP60 co-expression) (Fig. 2D). Sample conditions were selected as indicated was also observed in a minor percentage of trans- in the input blot (Fig.  2D, left). Co-IP (Fig.  2D, right) fected cells (not shown), however, the main phenotype using anti-HA tag antibody (TIP60-HA) revealed pre- revealed a uniform APP-CT50 signal without accu- cipitation of FE65 (as expected, Fig.  2D, white arrow) mulation in nuclear aggregates (Fig.  3A). Neverthe- and of p53 (red arrow), which validated its participa- less, cytosolic PML might also be bound to APP-CT50. tion in the complex. The TIP60/p53 interaction (inde- Additional expression of FE65 did not change the main pendent of FE65, in agreement to Fig.  2C) could be phenotype of uniformly localized APP-CT50 (Fig.  3B). confirmed by precipitating TIP60 via a GFP tag anti- Omitting TIP60 co-expression caused the generation body (white arrowhead). In addition to the p53-EGFP of a PML/APP-CT50/p53 complex (Fig.  3C). In con- signal, the presence of co-precipitated endogenous trast, the DNA helicase BLM is not co-localized to p53 was observed (3rd land, red arrowhead). Endog- PML (Fig.  3D). A more detailed analysis, utilizing 3D enous p53 is also detectable in conditions with TIP60 confocal imaging, revealed that APP-CT50 and PML co-expression (e.g. lane 7, blue arrow) suggesting an are associated with each other in all nuclear aggre- interaction of the (endogenous) tumour suppressor gates (Fig.  3E). Finally, the PML/APP-CT50 interac- protein p53 with the histone acetyltransferase TIP60 tion was confirmed by a co-immunoprecipitation assay in nuclear aggregates. (Fig. 3F). Co-expression of APP-CT with either PML1- The second candidate was a tumour suppressor pro- HA or PML1-myc demonstrated precipitation of PML tein as well, the promyelocytic leukemia protein (PML) (detection via anti-myc or anti-HA antibody, Fig.  3F, [13]. Co-expression of PML-EGFP, BFP-APP-CT50, white arrow) upon anti-GFP IP (independent experi- FE65-mCherry and TIP60 (w/o fluorophore) revealed ment given in Additional file  1: Fig. S5). This was true a close association of PML bodies to one or two APP- for two different APP-CT isoforms with 50 and 57  aa CT50 aggregates (Fig.  2E, zoom 1), which was also in length. High-resolution STED imaging further spec- confirmed by the fluorescence intensity line scan given ified the co-localization of PML and APP-CT50 within in Fig.  2E on the right. Indeed, the green PML peak is the PML bodies. APP-CT50 was either uniformly accompanied by two peaks of APP-CT50 (blue) and distributed or concentrated at the inner wall of the FE65 (red). Other cells of the same condition (zoom nuclear body (Fig. 3G). 2nd line) demonstrated association of three APP- CT50 complexes, whereas another phenotype (zoom 3rd line) demonstrated large APP-CT50 complexes with incorporated PML bodies. Collectively, the high (See figure on next page.) Fig. 2 The nuclear APP-CT50 complex associates to p53 and PML. (A) FE65 (red, mCherry) and TIP60 (blue, BFP), which are known to co-localize and form a dot-like structure, were transfected into HEK293T cells. With the addition of the tumour suppressor protein p53 (green, EGFP), the same pattern could be revealed (for example see white arrows), proving further co-localization and interactions of the proteins. (B) Co-localization of the protein complex is proven by tracking of the fluorescence intensity along the indicated arrow, which revealed peak intensities of each component (green: p53-EGFP, red: FE65-mcherry, blue: TIP60-BFP) supporting the association of all components within one complex. (C) Omitting FE65 and transfecting only p53 and TIP60 also revealed a co-localization (for example see white arrows), proving that the complex is independent of FE65. (D) P53 interaction with the APP-CT50 complex (FE65-EGFP/ TIP60-HA co-transfected to p53-EGFP versus control (EGFP-NLS, nuclear localization sequence)) was validated by co-immunoprecipitation (left side: input blot, right side: elution blot). IP using anti-HA tag antibody (against TIP60-HA) revealed precipitation of FE65 (white arrow) as well as of p53 (red arrow). Respective controls did not show a co-precipitation. TIP60-HA precipitation also occurred using anti-GFP as bait in well agreement to results obtained in part C (white arrowhead). Notably, high levels of endogenous p53 co-eluted in the same condition (red arrowhead), whereas a moderate endogenous p53 signal was observable in control conditions (red arrow). (E) A second tumour suppressor protein was identified to associate with the APP-CT50 complex: the promyelocytic leukemia protein PML. Different phenotypes of association were observed, e.g. one or two APP-CT50 (blue, BFP)/FE65 (red, mCherry) dots ( TIP60-HA was co-transfected w/o fluorophore) associated with a single PML (green, EGFP) aggregate (first and second zoom-in row, compare fluorescence intensities). Alternatively, large APP-CT50/FE65 complexes with enclosed PML-dots were found (third row) Marks et al. acta neuropathol commun (2021) 9:66 Page 8 of 16 e M arks et al. acta neuropathol commun (2021) 9:66 Page 9 of 16 APP‑CT50 depending complexes drive PML complex 96  h (middle row), whereas expression of PML/FE65/ generation that are also present in the aged human brain APP-CT50/TIP60 demonstrated nuclear body forma- To further investigate this interaction, live cell imag- tion already after 24  h (below row). After 48  h, some ing experiments were performed in HEK293 cells cells revealed formation of super-aggregates within with ectopic expression of the aggregate components. the nucleus (white arrow). We subsequently aimed to Expression of FE65-EGFP, TIP60 (w/o fluorophore) investigate whether these complexes are present in the revealed highly dynamic ring-like structures moving human brain as well (Fig. 4E). Human hippocampal fro- throughout the nucleoplasm (Fig.  4A, arrowhead, con- zen brain samples from 15 AD patients were used to focal image, structures coloured according to z-level; study the co-localization of APP-CT50 and PML. We full video given in the Additional file  2). To investigate observed a strong co-localization of PML and APP- the influence of PML in these dynamics, the complexes CT50 in the human brain (Additional file  1: Fig. S6, an (based on the FE65-mCherry signal) were tracked in antibody raised against APP 643–695 was used in order APP-CT50/FE65/TIP60-transfected cells with versus to identify APP-CT50 in the nucleus). Tracking of the without PML co-expression (Fig.  4B). The velocity of fluorescence intensity along the white arrow (includ - the nuclear structures in the presence of PML (Fig. 4B, ing three aggregates) validated the strong co-localiza- diagram  1, red graph) was significantly reduced com - tion. We also investigated the localization of FE65 in pared to the condition without PML (black graph). the same manner (Fig. 4F) and demonstrated that FE65 Moreover, the average distance from the track ori- co-localized with PML in the human brain as well. gin was significantly less in PML co-expressing cells Both co-localizations were evident in the brains of AD (Fig.  4B, diagram 2). The mean speed was 0.38 in PML patients with different Braak stages. As all samples were versus 0.76  µm/s in non-PML co-expressing cells obtained from individuals older than 65 years, we next (Fig.  4B, diagram  3). These results suggest a mutual aimed to study whether this phenotype is age-depend- trapping function of APP-CT50 aggregates and PML, ent. To study this question in a human 3D model, we for which Fig.  4C shows the typical phenotype of up differentiated iPS cells to cerebral organoids (Fig.  4G). to three APP-CT50 aggregates (red) bound to a single In contrast to the aged human brain, immunofluores - PML body (green). In order to understand the condi- cence analysis demonstrated no co-localization of APP- tions for aggregation in more detail, we monitored CT50 with PML in cerebral organoid sections. transfected cells over time (Fig. 4D). Pure PML expres- sion (upper row) revealed a homogenous distribution Reduction of PML bodies occurs in human hippocampal of the protein within the whole cell (cytoplasm and brain areas with high plaque load nucleus). Co-transfection of APP-CT50 (blue) and In order to study a potential pathophysiological rel- FE65 (red) caused late aggregate formation mostly after evance, we examined human brain tissue in more detail (See figure on next page.) Fig. 3 PML forms container-like structures within the complex and precipitates with APP-CT. (A) A direct co-localization of APP-CT (blue, BFP) with PML (green, EGFP) was only observed to some extent (cells with nuclear aggregates), but most cells revealed a uniform staining pattern of APP-CT50 and PML in the cytosol and nucleus. (B) Additional co-expression of FE65 (red, mcherry) enriched the aggregation of PML in the nucleus, but a strong co-localization to APP-CT50/FE65 was not observable in the imaging study. (C) Additional transfections showed that p53 (red, mCherry) is part of the PML aggregates in the nucleus that also contained APP-CT50, as indicated by the visible co-localization (for example see white arrows). (D) Notably, another suspected binding partners for PML like the DNA helicase BLM (blue, BFP), which was identified as binding protein in the APP-CT50 complex, is not co-localized with the PML aggregates. (E) Confocal 3D imaging validated the co-localization of PML (green, EGFP) and APP-CT50 (red, mCherry) in the nucleus (FE65/TIP60 were co-transfected w/o fluorophore). (F) Interaction of APP-CT with PML was shown using co-immunoprecipitation assay. Precipitation using anti-GFP antibody revealed detection of APP-CT-EGFP (as expected) as well as PML (PML1 isoform was used, white arrow). This was true for two different APP-CT isoforms (APP-CT50 and CT57), whereas control conditions revealed no unspecific co-precipitation. The before mentioned isoforms of APP’s c-terminal domain differ in their respective amino acid length generated through ε-cleavage. While APP-CT50 represents the most common form, APP-CT57 is less common.Results were the same for two different PML1 tags: in the left panel of blots PML1-HA was used, whereas PML1-myc was used in the right panel. (G) High-resolution STED imaging specified the localization of APP-CT50 (green, EGFP) within the PML bodies (red, mCherry). Different phenotypes were evident, either with a uniform localization within the bodies or with APP-CT50 signal at the inside wall of the PML body Marks et al. acta neuropathol commun (2021) 9:66 Page 10 of 16 M arks et al. acta neuropathol commun (2021) 9:66 Page 11 of 16 (Fig.  5). As the human hippocampus belongs to those PML bodies compared to all cells counted (cells with- brain areas revealing early pathological AD features, out any PML body, cells with more than 5 PML bod- we studied the extent of PML bodies in the Cornu ies), revealed a significant (p < 0.05) reduced percentage Ammonis areas 1 or 3 (CA1, CA3). CA regions were of PML positive nuclei in areas with high plaque load not evident for all human brain samples due to differ - (Fig. 5F, plaque) compared to plaque free areas (Fig. 5F, ent quality (different post-mortem times, preparation no plaque). Tile-scans from individuals without any artefacts), thus we limited our analysis to CA1 or CA3 neurodegenerative pathology (controls) demonstrated areas, which were distinctly assignable, e.g. Figure  5A higher percentage of PML positive nuclei compared to corresponds to a sample with CA1 assignment, but not the “no plaque” group (p < 0.05). More detailed analyses CA3. Haematoxylin Eosin (HE) staining of human hip- demonstrated that the observed significance is particu - pocampal frozen sections (samples from 15 AD indi- larly caused by cells containing only one or two PML −5 viduals with different Braak stages) was used to identify bodies per nucleus (Fig. 5G, *p < 0.05, **p < 10 ). the specific areas. Parallel sections (from the same indi - vidual) were used for low-resolution tile-scale imag- Discussion ing (DAPI channel, Fig.  5A) to allocate hippocampal The amyloid precursor protein is a ubiquitously areas according to HE staining. For subsequent detailed expressed protein, thus it is not surprising that many analysis, confocal tile-scan imaging was used includ- different cell types are capable to induce APP-CT ing 5 × 5 tiles and 5 z-stacks (Fig.  5B). Afterwards, (APP-CT50) signalling and to set up nuclear aggre- maximum projection algorithm was applied result- gates. FE65 was initially reported to be brain tissue- ing in high-resolution imaging enabling identification specific, and indeed, expression analysis suggested of the number of PML bodies in each nucleus over an FE65 to be a neuronal protein [24–26]. TIP60 is also area of 800 × 800  µm within CA1 or CA3 (Fig.  5C). ubiquitously expressed with the highest amounts in This imaging pipeline was further extended to identify testis and placenta. Thus, we conclude that APP-CT50/ areas of high plaque load (within CA1 or CA3) using FE65/TIP60 signalling is a ubiquitous pathway with a Thioflavin co-staining (Fig.  5D). In total, we scanned preference in neuronal cells due to neuron-specific 68 areas using this approach. All tile-scans were sub- FE65 expression. Notably, APP-CT50 dissociation and sequently processed using CellProfiler software in nuclear translocation was described to predominantly order to identify and extract nuclei (Fig.  5E, upper occur through the amyloidogenic processing pathway row) and to detect and count the number of PML bod- [27]. This hypothesis is further supported by findings ies within the extracted nuclei (Fig.  5E, bottom row). in APP-CT over-expressing mice revealing neuronal Data analysis of all cells containing between 1 and 5 network abnormalities [28]. APP-CT has an impact (See figure on next page.) Fig. 4 Highly mobile APP-CT50-depending complexes that are also present in the aged human brain drive PML complex generation. (A) Expression of APP-CT50/ FE65/TIP60-HA in HEK293 cells reveals a highly mobile complex moving three-dimensionally in the cellular nucleus. Ring-like structures were coloured (orange to yellow) according to z-level (confocal microscopy). The indicated structure (white arrowhead) revealed time-dependent movement. The corresponding video is given in the supplement. (B) Movement of the individual aggregates was tracked using Huygens object tracker software. Transfection in HEK293 cells included APP-CT50/FE65-mCherry/TIP60-HA with and without EGFP-PML co-expression. The FE65-mCherry signal was used for tracking, revealing lower speed in cells co-expressing PML (first diagram). In addition, the distance from the track origin (at time point 0) was analysed. Co-expression of PML revealed significantly lower distances pointing to mutual trapping of both complexes. The mean speed was 0.38 in PML versus 0.76 µm/s in non-PML co-expressing cells (last diagram). (C) This part reveals a representative image demonstrating the complex generation of APP-CT50/FE65/TIP60 (red) and PML (green aggregates). (D) Time-dependent generation of nuclear APP-CT50/PML aggregates. PML (green, EGFP) expression revealed a uniform distribution within the nucleus and cytosol (first row). Co-expression of FE65 (red, mCherry) and APP-CT50 (blue, BFP) caused initial aggregate formation after 48 h (middle row). Additional expression of TIP60-HA (w/o fluorophore) (last row) showed early generation of nuclear aggregates after 24 h. 48 h after transfection large nuclear aggregates were observed (white arrow). (E) PML (green, FITC) and APP-CT50 (red, TRITC) co-localization was studied in human brain sections. In total, 15 human hippocampal sections were analysed (different Braak stages). Confocal tile-scan imaging (5 z-stacks, then fused by maximum projection algorithm) revealed strong co-localization of PML with APP-CT50. As in cell culture experiments, nuclei containing many small aggregates (arrow) as well as nuclei with larger aggregates (arrowhead) were evident. Co-localization is further shown by intensity tracking of both fluorescent channels in the diagram for a nucleus along the dotted white arrow. (F) Similarly, co-staining of PML (green, FITC) and FE65 (red, TRITC), which confirmed the association of both proteins in the nuclei of the human brain, was performed. (G) In order to address the question whether co-localization also occurs in non-aged tissue, we differentiated human cerebral organoids from induced pluripotent stem cells. Embryonic bodies were embedded in Matrigel at day 11 followed by neuronal induction to generate organoids, which were analysed after 30 days in culture (seeding at day 0). Staining of cryosections failed to demonstrate co-localization of APP-CT50 (red, TRITC) and PML (green, FITC) Marks et al. acta neuropathol commun (2021) 9:66 Page 12 of 16 APP M arks et al. acta neuropathol commun (2021) 9:66 Page 13 of 16 on gene expression [9–12], and our findings of posi - secondary antibody used for PML staining, suggest- tive histone 3 K9 acetylation in the core of the APP-CT ing a localization of APP-CT50 and FE65 within PML aggregates further support this hypothesis. The donut- spherical aggregates, which is in good agreement to like structure of the nuclear aggregates fits to this func - other reports [29]. tion as well, assuming that DNA is incorporated in (or Relevance of APP-CT50/FE65/PML aggregates for associated to) the spherical aggregates as already shown the pathophysiology of AD was finally demonstrated for cells in G2 phase [29]. Thus, APP-CT50 aggregates by tile-scan imaging of hippocampal CA1 and CA3 might correspond to DNA incubation containers possi- areas. Our analysis revealed highly significant results bly modifying DNA in a way to change expression. demonstrating a reduction in the number of PML bod- In order to better understand the nuclear APP-CT50- ies in nuclei close to AD relevant hot spots with high depending aggregates, we studied their composition plaque load. Assuming that these areas also corre- and identified the two tumour suppressor proteins p53 spond to elevated APP cleavage, we conclude that APP and PML as additional components. According to our nuclear signalling involving the adapter protein FE65 is results in HEK293 cells, this interplay is strongly driven correlated to AD pathology. Expression of APP-CT50, by the APP-CT50 nuclear translocation, as pure PML FE65, TIP60 and PML in HEK293T cells caused a time- expression revealed a rather homogenous cellular dis- dependent fusion of the nuclear aggregates. Thus, APP- tribution. The complex generation follows a temporally driven fusion of PML aggregates may occur in the AD organized scheme with generation of small aggregates brain in a similar fashion. Certainly, further studies are at an early phase and few large nuclear complexes at pivotal to understand the consequences of APP to PML later time points. As these aggregates were evident in body generation, fusion, and, in particular, their impact the human brain of aged patients but not in cerebral on neurodegeneration and dependence on different organoids, we conclude that the APP-CT50 nuclear sig- Braak stages. nalling is age-dependent and potentially of relevance for the pathophysiology of Alzheimer’s disease. Stain- ing of APP-CT50 and FE65 was only successful using a Conclusion sophisticated protocol with defined order of antibody APP-CT50 signal transduction is of high relevance for incubation, meaning that incubation with the second- AD and causes a reduction in the number of PML bod- ary antibody for PML detection precluded the iden- ies in nuclei close to AD relevant hot spots. The com - tification of APP-CT50 or FE65 in the complex. In position of the nuclear aggregates includes APP-CT50, contrast, usage of the secondary antibody as the last and tumor suppressor proteins PML and p53. Aggre- step (after application of the Avidin-TRITC complex gates are associated to gene expression changes with to identify APP-CT50 of FE65) was successful to reveal putative impact in neurodegeneration. co-localization. Thus, accessibility of the epitopes of APP-CT50 or FE65 was presumably masked by the (See figure on next page.) Fig. 5 Significant enrichment of PML bodies in human brain areas with high plaque load. (A) Haemotoxylin eosin (HE) staining of human hippocampal sections was used to define Cornu Ammonis (CA) 1–3, Gyrus dentatus (GD), and Plexus areal. Parallel sections were used for PML immunofluorescence staining. DAPI co-staining was used for low-resolution tile-scan imaging and allocation of specific CA areas according to the initial HE staining. (B) Hippocampal CA1 or CA3 areas were used for high-resolution tile scan imaging. A single scan included 5 × 5 images with 5 z-stacks, which were subsequently combined using maximum projection. (C) A representative high-resolution (100 × objective) confocal tile-scan image demonstrates identification of PML bodies in DAPI counter-stained nuclei. (D) Tile-scan imaging was then established with Thioflavin co-staining to identify areas with high plaque load in the human hippocampus (CA1 or CA3). (E) CellProfiler software was used to automatically annotate and extract nuclei (DAPI channel in grey scale) from the image (upper row). Subsequently, PML body identification and quantification was done in the extracted nuclei. (F) All nuclei containing 1 to 5 PML bodies were used to determine the percentage of PML positive nuclei (y axis). Hippocampal areas with high plaque load (plaque, average = 36.2%) revealed a significant lower percentage of PML positive nuclei compared to areas without plaque (no plaque, average = 44.1%) (p < 0.05; every dot (left to each bar) indicates a single tile-scan experiment; for areas with high plaque load 31 tile-scans were analysed and quantified, 28 for no plaque, 9 for control). Tile-scans of control sections (individuals without any plaque pathology, average = 53.8%) revealed again higher percentage of PML positive nuclei compared to “no plaque” areas (p < 0.05). (G) Detailed analysis of cells containing one up to ten PML bodies per nucleus revealed significant differences. Nuclei with a single PML body were underrepresented (p < 0.05) in areas with high plaque load versus areas without plaques. Difference to control tile scans were highly significant −5 (p < 10 ). In addition, nuclei with two PML bodies (red bars) also revealed highly significant differences to the control. Representative images for nuclei containing one, two, or three PML bodies are given Marks et al. acta neuropathol commun (2021) 9:66 Page 14 of 16 g M arks et al. acta neuropathol commun (2021) 9:66 Page 15 of 16 Abbreviations Received: 4 March 2021 Accepted: 29 March 2021 AD: Alzheimer’s disease; APP-CT: Amyloid precursor protein c-terminal domain; CA: Cornu ammonis; CMLE: Classic Maximum Likelihood Estimation; CO: Cerebral organoids; FITC: Fluorescein isothiocyanate; GFP: Green fluo - rescent protein; HE: Haematoxylin Eosin; HyD: Hybrid detector; iPS: Induced pluripotent stem (cell); LDA: Linear discrimination analysis; PBS: Phosphate References buffered saline; PML: Promyelocytic leukemia protein; PVDF: Polyvinylidene 1. Hardy J, Selkoe DJ (2002) The amyloid hypothesis of Alzheimer’s difluoride; ROI: Region of interest; STED: Stimulated emission depletion; TRITC: disease: progress and problems on the road to therapeutics. Science Tetramethylrhodamine. 297(5580):353–356. https:// doi. org/ 10. 1126/ scien ce. 10729 94 2. Bertram L, Lill CM, Tanzi RE (2010) The genetics of Alzheimer disease: back to the future. Neuron 68(2):270–281. https:// doi. org/ 10. 1016/j. neuron. Supplementary Information 2010. 10. 013 The online version contains supplementary material available at https:// doi. 3. Lott IT, Head E, Doran E, Busciglio J (2006) Beta-amyloid, oxidative stress org/ 10. 1186/ s40478- 021- 01174-x. and down syndrome. Curr Alzheimer Res 3(5):521–528 4. Chasseigneaux S, Allinquant B (2012) Functions of Aβ, sAPPα and sAPPβ : similarities and differences. J Neurochem 120(Suppl 1):99–108. https:// Additional file1. This file contains supplementary Figures and legends doi. org/ 10. 1111/j. 1471- 4159. 2011. 07584.x S1–S7. 5. Hardy JA, Higgins GA (1992) Alzheimer’s disease: the amyloid cascade hypothesis. Science 256(5054):184–185 Additional file2. This file contains a supplementary movie to Figure 4 A. 6. Bukhari H, Glotzbach A, Kolbe K, Leonhardt G, Loosse C, Müller T (2017) Small things matter: implications of APP intracellular domain AICD Acknowledgements nuclear signaling in the progression and pathogenesis of Alzheimer’s We would like to thank the Electron Microscopy Unit (EMU), Imaging Center disease. Prog Neurobiol 156:189–213. https:// doi. org/ 10. 1016/j. pneur Essen (IMCES), University Hospital Essen as well as the Imaging Centre Campus obio. 2017. 05. 005 Essen (ICCE) for their support. We would like to thank Dr. Mavrommatis and 7. Radzimanowski J, Simon B, Sattler M, Beyreuther K, Sinning I, Wild K PD Zaehres, Ruhr-University Bochum, Medical Faculty, Institute of Anatomy for (2008) Structure of the intracellular domain of the amyloid precursor providing CD34 iPSC and help with iPSC culture. Instrument Leica TCS SP8X protein in complex with Fe65-PTB2. EMBO Rep 9(11):1134–1140. https:// FALCON was obtained through DFG funding (Major Research Instrumentation doi. org/ 10. 1038/ embor. 2008. 188 Programme as per Art. 91b GG, INST 20876/294-1 FUGG). Instrument Leica 8. Schrötter A, Mastalski T, Nensa FM, Neumann M, Loosse C, Pfeiffer K, TCS SP8 was supported by an instrument grant from the German Research Magraoui FE, Platta HW, Erdmann R, Theiss C, Uszkoreit J, Eisenacher M, Foundation (INST 213/886-1 FUGG). Meyer HE, Marcus K, Müller T (2013) FE65 regulates and interacts with the Bloom syndrome protein in dynamic nuclear spheres - potential Authors’ contribution relevance to Alzheimer’s disease. J Cell Sci 126(Pt 11):2480–2492. https:// DM, NH, LB, SM, MW, LG, PH, VB, ES, TM did the experiments. KW corrected the doi. org/ 10. 1242/ jcs. 121004 manuscript. The study was designed by SKK and TM. Supervision of the whole 9. Grimm MO, Mett J, Stahlmann CP, Grösgen S, Haupenthal VJ, Blümel T, project was done by TM. All authors read and approved the final manuscript. Hundsdörfer B, Zimmer VC, Mylonas NT, Tanila H, Müller U, Grimm HS, Hartmann T (2015) APP intracellular domain derived from amyloidogenic Funding β- and γ-secretase cleavage regulates neprilysin expression. Front Aging Open Access funding enabled and organized by Projekt DEAL. This work was Neurosci 7:77. https:// doi. org/ 10. 3389/ fnagi. 2015. 00077 supported by funding from German Research Foundation (DFG, Germany 10. Kim HS, Kim EM, Lee JP, Park CH, Kim S, Seo JH, Chang KA, Yu E, Jeong MU3525/3-2), an instrument grant from DFG (INST 213/886-1 FUGG) and SJ, Chong YH, Suh YH (2003) C-terminal fragments of amyloid precursor Mercator Research Center Ruhr (Mercur, Germany, Pr-2016-0010). protein exert neurotoxicity by inducing glycogen synthase kinase-3beta expression. FASEB J 17(13):1951–1953. https:// doi. org/ 10. 1096/. fj 03- 0106f je Declarations 11. von Rotz RC, Kohli BM, Bosset J, Meier M, Suzuki T, Nitsch RM, Konietzko U (2004) The APP intracellular domain forms nuclear multiprotein Ethical approval and consent to participate complexes and regulates the transcription of its own precursor. J Cell Sci An ethical vote for the current work is available (17-6138, Ethical Commission 117(Pt 19):4435–4448. https:// doi. org/ 10. 1242/ jcs. 01323 of the Ruhr- University Bochum). 12. Zhou F, Gong K, van Laar T, Gong Y, Zhang L (2011) Wnt/β-catenin signal pathway stabilizes APP intracellular domain (AICD) and promotes its tran- Consent for publication scriptional activity. Biochem Biophys Res Commun 412(1):68–73. https:// All authors agreed to publish the data given in the manuscript and to provide doi. org/ 10. 1016/j. bbrc. 2011. 07. 040 data and material to the scientific community. 13. Lallemand-Breitenbach V, de Thé H (2018) PML nuclear bodies: from architecture to function. Curr Opin Cell Biol 52:154–161. https:// doi. org/ Availability of supporting data 10. 1016/j. ceb. 2018. 03. 011 In case of acceptance, supporting data can be accessed at the Cell image 14. Van Damme E, Laukens K, Dang TH, Van Ostade X (2010) A manually library (http:// www. celli magel ibrary. org/ home). curated network of the PML nuclear body interactome reveals an impor- tant role for PML-NBs in SUMOylation dynamics. Int J Biol Sci 6(1):51–67. Competing interests https:// doi. org/ 10. 7150/ ijbs.6. 51 There are no competing interests. 15. Dellaire G, Bazett-Jones DP (2004) PML nuclear bodies: dynamic sensors of DNA damage and cellular stress. BioEssays 26(9):963–977. https:// doi. Author details org/ 10. 1002/ bies. 20089 Department of Molecular Biochemistry, Cell Signalling, Faculty of Chemistry 16. Chang HR, Munkhjargal A, Kim MJ, Park SY, Jung E, Ryu JH, Yang Y, Lim and Biochemistry, Ruhr University Bochum, Bochum, Germany. Institute JS, Kim Y (2018) The functional roles of PML nuclear bodies in genome of Psychiatric Phenomics and Genomics (IPPG), University Hospital, LMU maintenance. Mutat Res 809:99–107. https:// doi. org/ 10. 1016/j. mrfmmm. Munich, Munich, Germany. Leibniz Institute On Aging Research, Jena, 2017. 05. 002 Germany. Department of Molecular Cell Biology, Institute of Biochemistry and Pathobiochemistry, Medical Faculty, Ruhr-University Bochum, Bochum, Germany. Department of Molecular Biology II, Center of Medical Biotechnol- ogy (ZMB), University of Duisburg-Essen, Essen, Germany. Marks et al. acta neuropathol commun (2021) 9:66 Page 16 of 16 17. Stadler M, Chelbi-Alix MK, Koken MH, Venturini L, Lee C, Saïb A, Quignon 25. Kesavapany S, Banner SJ, Lau KF, Shaw CE, Miller CC, Cooper JD, F, Pelicano L, Guillemin MC, Schindler C (1995) Transcriptional induction McLoughlin DM (2002) Expression of the Fe65 adapter protein in adult of the PML growth suppressor gene by interferons is mediated through and developing mouse brain. Neuroscience 115(3):951–960 an ISRE and a GAS element. Oncogene 11(12):2565–2573 26. Simeone A, Duilio A, Fiore F, Acampora D, De Felice C, Faraonio R, Paolocci 18. Scherer M, Stamminger T (2016) Emerging role of PML nuclear bodies in F, Cimino F, Russo T (1994) Expression of the neuron-specific FE65 gene innate immune signaling. J Virol 90(13):5850–5854. https:// doi. org/ 10. marks the development of embryo ganglionic derivatives. Dev Neurosci 1128/ JVI. 01979- 15 16(1–2):53–60. https:// doi. org/ 10. 1159/ 00011 2088 19. Lancaster MA, Knoblich JA (2014) Generation of cerebral organoids from 27. Goodger ZV, Rajendran L, Trutzel A, Kohli BM, Nitsch RM, Konietzko U human pluripotent stem cells. Nat Protoc 9(10):2329–2340. https:// doi. (2009) Nuclear signaling by the APP intracellular domain occurs predomi- org/ 10. 1038/ nprot. 2014. 158 nantly through the amyloidogenic processing pathway. J Cell Sci 122(Pt 20. Okada Y, Nakagawa S (2015) Super-resolution imaging of nuclear bodies 20):3703–3714. https:// doi. org/ 10. 1242/ jcs. 048090 by STED microscopy. Methods Mol Biol 1262:21–35. https:// doi. org/ 10. 28. Vogt DL, Thomas D, Galvan V, Bredesen DE, Lamb BT, Pimplikar SW (2011) 1007/ 978-1- 4939- 2253-6_2 Abnormal neuronal networks and seizure susceptibility in mice overex- 21. Fogal V, Gostissa M, Sandy P, Zacchi P, Sternsdorf T, Jensen K, Pandolfi PP, pressing the APP intracellular domain. Neurobiol Aging 32(9):1725–1729. Will H, Schneider C, Del Sal G (2000) Regulation of p53 activity in nuclear https:// doi. org/ 10. 1016/j. neuro biola ging. 2009. 09. 002 bodies by a specific PML isoform. EMBO J 19(22):6185–6195. https:// doi. 29. Luciani JJ, Depetris D, Usson Y, Metzler-Guillemain C, Mignon-Ravix C, org/ 10. 1093/ emboj/ 19. 22. 6185 Mitchell MJ, Megarbane A, Sarda P, Sirma H, Moncla A, Feunteun J, Mattei 22. Wang XW, Tseng A, Ellis NA, Spillare EA, Linke SP, Robles AI, Seker H, Yang MG (2006) PML nuclear bodies are highly organised DNA-protein struc- Q, Hu P, Beresten S, Bemmels NA, Garfield S, Harris CC (2001) Functional tures with a function in heterochromatin remodelling at the G2 phase. J interaction of p53 and BLM DNA helicase in apoptosis. J Biol Chem Cell Sci 119(Pt 12):2518–2531. https:// doi. org/ 10. 1242/ jcs. 02965 276(35):32948–32955. https:// doi. org/ 10. 1074/ jbc. M1032 98200 23. Tang Y, Luo J, Zhang W, Gu W (2006) Tip60-dependent acetylation of p53 Publisher’s Note modulates the decision between cell-cycle arrest and apoptosis. Mol Cell Springer Nature remains neutral with regard to jurisdictional claims in pub- 24(6):827–839. https:// doi. org/ 10. 1016/j. molcel. 2006. 11. 021 lished maps and institutional affiliations. 24. Duilio A, Zambrano N, Mogavero AR, Ammendola R, Cimino F, Russo T (1991) A rat brain mRNA encoding a transcriptional activator homolo- gous to the DNA binding domain of retroviral integrases. Nucleic Acids Res 19(19):5269–5274. https:// doi. org/ 10. 1093/ nar/ 19. 19. 5269 Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions

Journal

Acta Neuropathologica Communications – Springer Journals

Published: Apr 13, 2021

There are no references for this article.