DNA Damage Increases Secreted Aβ 40 and Aβ 42 in Neuronal Progenitor Cells: Relevance to Alzheimer’s Disease

Cell culture and neuronal differentiation

The human ReN VM GA2 immortalized neuronal progenitor cell line was a kind gift of R.E. Tanzi (Harvard University, Cambridge, MA). ReN VM GA2 stably expresses a GFP reporter and contains Familial AD-related genes encoding human AβPP with both the K670N/M671L (Swedish) and V717L (London) mutations. These mutations lead to overexpression of AβPP, altered subcellular localization of AβPP, and increased Aβ₄₂ generation, and were generated as previously described [37]. We used these cells to provide a basal level of Aβ detectable by ELISA; this would not have been possible without these ReN GA2 cells. NPCs were cultured in 2D Matrigel coated T75 flasks or 6-well plates (Fisher Scientific, 354230) in proliferation media: Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (Thermo Fisher, 10565042) supplemented with heparin (Sigma-Aldrich, H3393), B-27 (Gibco, 17504044), human fibroblast growth factor basic protein (bFGF) (Fisher Scientific, GF003AF100UG), and human recombinant epidermal growth factor protein (hEGF) (Millipore Sigma, GF144) at 37°C, 5% CO₂, and 5% O₂. For differentiation into neurons, NPCs were passaged at 500,000 cells/well in 2D Matrigel coated 6-well plates or 3.5 cm dishes in proliferation media for 24 h at 37°C, 5% CO₂, and 5% O₂. Proliferation media was then exchanged for Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 and B-27 with ½ media changes every 3 days for 20 days.

DNA damage assays

Ren GA2 NPCs were passaged at 500,000–700,000 cells/mL on 2D Matrigel coated 6-well plates or 3.5 cm dishes and allowed to grow for 48 h at 37°C, 5% CO₂, and 5% O₂. (Neurons were differentiated as outlined above.) For peroxide experiments, cells were treated with or without α-tocopherol, an oxygen radical scavenger. Where the scavenger was used, cells were pre-treated with 5μM α-tocopherol (Sigma-Aldrich, T3251), or 0.5% DMSO control. 5μM α-tocopherol was then maintained throughout peroxide damage treatment and recovery time courses. Cells were then treated with 2.5μM peroxide for 0.5 h, washed with 1 X PBS, the media was replaced, and the cells were allowed to recover for noted time courses at 37°C, 5% CO₂, and 5% O₂. 5μM α-tocopherol was maintained throughout recovery time courses. For etoposide treatments, cells were treated with 0.5μM, 1.0μM, 5.0μM, 10μM, and 20μM etoposide or DMSO control (0.02% DMSO for etoposide treatment alone) for 6 h at 37°C. Cells were then washed with 1 X PBS, the media was replaced, and the cells were allowed to recover for noted time courses at 37°C, 5% CO₂, and 5% O₂. Both α-tocopherol and etoposide were re-suspended in 100% DMSO prior to dilutions.

Immunostaining

ReN GA2 NPCs or neurons were grown on 2D Matrigel coated 12 mm diameter coverslips in 6-well plates or 3.5 cm dishes for immunostaining. For endogenous GFAP, MAP2, Tuj1, and Ki67 staining, coverslips were washed with 1 X PBS and fixed with 4% paraformaldehyde with 4% sucrose for 15 min at room temperature. The fixed cells were rinsed three times with 1 X PBS, permeabilized with 0.2% Triton for 10 min at room temperature, then rinsed three times with 1 X PBS. Cells were blocked with 10% goat serum in 1 X PBS for 1 h at room temperature. For primary antibody staining, coverslips were incubated with primary antibody in 10% goat serum overnight at 4°C in a humidity chamber. Coverslips were then washed three times with 1 X PBS and incubated with 1:1000 DAPI and secondary antibody Alexa Fluor 694/555 goat anti-mouse -rabbit immunoglobulin G conjugate (Invitrogen) in 10% goat serum for 1 h at room temperature. Coverslips were washed three times with 1 X PBS, once with Millipore water, then mounted with ProLong Diamond Antifade Mountant (Fisher Scientific, P36961) onto microscope slides. Primary antibodies used were anti-GFAP (1:1000, Fisher Scientific, 01-670-261), anti-MAP2 (1:1000, Millipore, MAB378), anti-Tuj1 (1:1000, Abcam, ab18207), and anti-Ki67 (1:1000, Fisher Scientific, PA5-19462). Fluorescence images were captured with a Nikon Eclipse Ti2 confocal laser scanning microscope (Nikon Instruments, Inc.).

Neutral comet assay

The neutral comet assay was performed using the Comet Assay Kit (Abcam, ab238544) according to the manufacturer’s protocol. DNA damage resolves as a comet that can be quantified by calculating the “% tail DNA”, or percentage of fragmented DNA in the tail relative to the total amount of DNA [42, 43]. In brief, after treatment, ReN GA2 NPCs or 20-day neurons were trypsinized, centrifuged at 700 g for 2 min, and washed once with ice cold 1 X PBS. After re-centrifugation, the pellet was suspended at 1×10⁵ cells/ml in ice-cold 1 X PBS. Cell samples were mixed with comet agarose in a 1/10 ratio (v/v) and immediately transferred onto the slides previously coated with comet agarose base layer. After incubating with pre-chilled lysis buffer for 2 h at 4°C, then pre-chilled alkaline buffer for 0.5 h at 4°C, the slides were subjected to electrophoresis in TBE Electrophoresis Solution for 40 min at approximately 35 V on ice. After electrophoresis, the slides were incubated with Vista Green DNA dye. Images were obtained with a Nikon Eclipse Ti2 confocal laser scanning microscope (Nikon Instruments, Inc.) using the FITC filter. Approximately 3–6 images (5–20 cells per image) were randomly taken, and the % tail DNA (100 x Tail DNA intensity/Cell DNA intensity) of 30 cells was calculated using the OpenComet plugin of ImageJ analysis software ver. 2.0.0.

Lysate preparation and western blot

ReN GA2 NPCs or 20-day neurons were grown on 2D Matrigel coated 6-well plates or 3.5 cm dishes and treated as described in the DNA damage assays. After treatments and recovery time courses, media was removed and spun at 13000 rpm for 10 min at 4°C to pellet dead cells and debris. Supernatant was then removed. For sAβPPβ detection, 30μL of supernatant with 4% β-meracaptoethanol in 4 X Non-Reducing LDS Sample Buffer (Thermo Scientific, 84788) was used. Cells were washed with pre-warmed 1 X PBS, then detached from the plate using diluted trypsin (0.05%). Cell pellets were spun at 13000 rpm for 10 min at 4°C. Cell lysates were prepared with 100μL of ice cold RIPA lysis buffer per sample (25 mM Tris HCL pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS) with 1 X protease inhibitor (Sigma-Aldrich, 04693116001) and 25–30 U benzonase (Sigma-Aldrich, 70746-3). Total protein concentrations of the lysates were determined by Bicinchoninic Acid (BCA) assays. Samples were prepared for western blot using a lysate dilution that would yield 25–40μg total protein (NPCs) or 20μg total protein (differentiated cells) with 4% β-meracaptoethanol in 2 X Novex Tris-Glycine SDS Sample Buffer (Life Technologies, LC2676). Samples were heated to 70°C for 10 min while shaking at 800 rpm, then run on 4%–20% Novex Wedge Well Tris-Glycine gels (Thermo Fisher, XP04205BOX) in 1 X Tris/Glycine/SDS running buffer (Thermo Fisher, BP13414) at 80 V for 10 min, then 100 V for 2 h. The protein was transferred to a nitrocellulose membrane using the Invitrogen Power Blotter System (Thermo Fisher, PB0012) and 80% pre-made transfer buffer stock (5.85 g glycine, 11.6 g Tris, 0.375 g SDS, milliQ water to 1 L) with 20% ethanol at a constant 25 V for 45 min. Blots were stained with Ponceau (5% acetic acid, 0.1% Ponceau stain) for 5 min to quantify total protein. Blots were then blocked in TBST with 5% milk, and immunoblotted overnight at 4°C using the following antibodies at a 1:1000 dilution in 5% milk and TBST: anti-γH2AX (Abcam, ab26350), anti-6E10 (Biolegend, 803001), anti-BACE1 (Abcam, ab108394); at 2μg/mL dilution in 5% milk and TBST: anti-sAβPPβ-sw (6A1) (IBL, 10321); and at 1:10000 dilution in 5% milk and TBST: anti-β-Actin (Sigma-Aldrich, A5441), and anti-GAPDH (Sigma-Aldrich, G9545). Blots were washed three times with TBST for 5 min each, then probed with appropriate secondary antibodies (anti-Mouse IgG Horseradish Peroxidase HRP or anti-Rabbit IgG HRP) in a 1:5000 dilution in 5% milk and TBST for 1 h at room temperature. Following secondary antibody probing, blots were rinsed twice, then washed for 10 min in TBST, then rinsed twice and washed for 10 min in TBS. Blots were imaged using ECL substrate or SuperSignal West Femto Maximum Sensitivity Substrate in the Bio-Rad ChemiDoc. Bands were quantified utilizing ImageJ analysis software ver. 2.0.0.

Aβ enzyme-linked immunosorbent assay (ELISA assay)

Secreted Aβ₄₀ and Aβ₄₂ levels were determined by standard sandwich ELISA using end-specific antibodies provided by Janssen Pharmaceutica as previously described [44]. Briefly, media from cultures was centrifuged at 15000 rpm for 10 min at 4°C to pellet dead cells and debris. Supernatant was then removed. 96-well plates were coated with 1.5 ug/mL Aβ₄₀ and Aβ₄₂ monoclonal antibodies (JRFcAb₄₀/28 and JRFcAb₄₂/26), which recognize the C terminus of Aβ species terminating at amino acid 40 or 42, respectively, in coating buffer (10 mM Tris HCl, 10 mM NaCl, 10 mM NaN₃) overnight at 4°C. Plates were washed five times with 1 X PBS and 0.05% Tween and blocked with 0.1% casein buffer for 4 h, then washed five times with 1 X PBS and 0.5% Tween. Diluted standards were made with Aβ₄₀ and Aβ₄₂ antibodies (rPeptide, A-1163-1) (rPeptide, A-1153-1) in experimental media. Supernatant samples were diluted accordingly (1:2, 1:3, 1:6, and/or 1:10) in experimental media. Standards and samples were mixed with detection antibody horseradish peroxidase (HRP)-conjugated JRFAbN25 (1:6000 for Aβ₄₀, 1:4000 for Aβ₄₂) and incubated overnight at 4°C. Plates were washed five times with 1 X PBS and 0.5% Tween, then developed with 100 mM NaAc buffer with 3, 3’, 5, 5’-tetramethylbenzidine (TMB) (Sigma-Aldrich, T2885) and 30% peroxide for 5–10 min in the dark. The reaction was stopped with an equal volume of 2N H₂SO₄, and plates were read at 450 nm using a BioTek Synergy Neo 2 microplate reader. Assays were performed in triplicate with two technical replicates each.

Measurement of cellular apoptosis

Cell death of ReN GA2 NPCs and differentiated neurons was measured using a Click-IT TUNEL Alexa Fluor 647 imaging assay (C10247, Thermo Fisher) following the manufacturer’s instructions. Briefly, cells seeded on coverslips were treated with experimental conditions. Cells were then fixed with 4% paraformaldehyde in 1 X PBS for 15 min and rinsed three times with 1 x PBS. Cells were permeabilized with 0.25% Triton-X 100 in 1 X PBS for 20 min, then rinsed twice with DI. Each coverslip was incubated with 100μL of TdT reaction buffer for 10 min at room temperature, then removed. Cells were incubated in 100μL TdT reaction cocktail for 1 h at 37°C in a humidity chamber. Cells were washed twice in 3% BSA in 1 X PBS for 2 min, then incubated with Click-IT reaction buffer with additive for 30 min in the humidity chamber protected from the light. Cells were rinsed with DI and counterstained with Hoechst 33342, and mounted with ProLong Diamond Antifade Mountant (Fisher Scientific, P36961) onto microscope slides. Fluorescence images were captured with a Nikon Eclipse Ti2 confocal laser scanning microscope (Nikon Instruments, Inc.) at 20 X magnification. Experiments were performed in duplicate with at least 200 cells counted manually per experiment for each condition. ImageJ (2.0.0-rc-69/1.52p) software was used for image analysis.

Statistical analysis

Data are presented as mean±SD. Statistical analysis was performed using GraphPad Prism version 9.0.0. Differences between means were assessed by Students T-test, one-way analysis of variance (ANOVA), or two-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test. All experiments were completed with a minimum of two to three replicates unless indicated in the figure legends. p < 0.05 was considered significant.

Global peroxide damage induces increased Aβ₄₀ and Aβ₄₂ secretion in ReN GA2 NPCs 24 h after treatment

A direct mechanistic link between oxidative damage and the induction of Aβ secretion has yet to be elucidated. Therefore, we sought to investigate whether global peroxide damage induces secretion of Aβ₄₀ and Aβ₄₂ in human NPCs. We pre-treated ReN GA2 NPCs with and without 5μM of the antioxidant α-tocopherol for 0.5 h. α-Tocopherol, the most biologically active form of vitamin E, is a lipid soluble antioxidant that acts as a chain-breaking antioxidant by scavenging lipid peroxyl and alkoxyl radicals [45, 46]. Cells were then treated with and without 2.5μM hydrogen peroxide for 0.5 h. Cells were allowed to recover with and without α-tocopherol for 2 h, 8 h, and 24 h (Fig. 1A). We found that treatment with hydrogen peroxide significantly reduced cell viability 24 h after treatment (Fig. 1B). Secretion of both Aβ₄₀ and Aβ₄₂ significantly increased 24 h following hydrogen peroxide treatment compared to controls. This increase in Aβ₄₀ and Aβ₄₂ secretion was mitigated by α-tocopherol pre-treatment (Fig. 1C). Hydrogen peroxide treatment did not alter the ratio of Aβ₄₂/Aβ₄₀ (Fig. 1D). Together, these results indicate that peroxide damage directly leads to increased Aβ₄₀ and Aβ₄₂ generation in ReN GA2 NPCs, and that anti-oxidant treatment directly prevents the elevation in Aβ generation. (For all timepoints, see Supplementary Figure 1A-C.)

OPEN IN VIEWER

Fig. 1

Hydrogen peroxide damage increases Aβ₄₀ and Aβ₄₂ secretion in ReN GA2 NPCs, an effect that is mitigated by α-tocopherol treatment. A) Schematic of ReN GA2 NPC experimental design. ReN GA2 NPCs were treated with and without 5μM α-tocopherol prior to treatment with and without 2.5μM hydrogen peroxide for 0.5 h and allowed to recover for 24 h. (Schematic created in BioRender, 2021.) B) Cell viability was determined after indicated treatment and 24 h recovery time. C) Percent secreted Aβ₄₀ and Aβ₄₂ relative to controls after designated treatments and 24 h recovery time. D) Aβ₄₂/Aβ₄₀ ratio calculated from raw values. Error bars represent means±SD of three separate experiments, and the p values were determined using 2-way ANOVA and Tukey’s multiple comparisons test. ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001.

Etoposide damage induces increased Aβ₄₀ and Aβ₄₂ secretion in ReN GA2 NPCs 2 h and 4 h after treatment

The two major endogenous sources of direct DNA damage in the human brain arise with ROS from metabolic activity and activity-dependent transcription that induces DSBs [15, 47]. Peroxide alone does not damage DNA but can give rise to highly reactive OH^- radicals that damage DNA in a process known

as the Haber-Weiss and Fenton reaction [48]. Due to the global and complex nature of peroxide damage, we questioned whether peroxide-induced DSBs could be the contributing factor to increased secretion of Aβ₄₀ and Aβ₄₂ in ReN GA2 NPCs. To determine whether peroxide damage induces DNA DSBs in ReN GA2 NPCs, we performed a neutral comet assay, which can be used to measure single strand breaks (SSBs) and DSBs in single cells (Supplementary Figure 2). In our study, the neutral comet assay indicates that peroxide damage induces significant DNA damage including, SSBs and DSBs, as compared to untreated controls, and that α-tocopherol treatment mitigates this damage (Supplementary Figure 2A).

During neuronal development, neuronal activity triggers the activation of numerous transcription factors and early response genes near promoter regions [49]. To prevent topological constraints in these regions, topoisomerase IIβ resolves DNA tangles and supercoils that may occur during DNA transcription and replication by generating transient DSBs, and then re-ligating the broken DNA strands [50]. We incubated ReN GA2 NPCs with etoposide, a topoisomerase II inhibitor, which maintains these induced DSBs by binding DNA at topoisomerase II-DNA cleavage sites and inhibiting the re-ligation process [51]. After a 6 h treatment with 10μM etoposide, ReN GA2 NPCs were allowed to recover for 2 h, 4 h, and 8 h (Fig. 2A). We found that after treatment with etoposide, cell viability was significantly reduced after all recovery times (Fig. 2B). However, a dose response curve showed no significant increase in apoptosis as indicated by TUNEL staining (Supplementary Figure 3B). ReN GA2 NPCs also accumulated significant quantities of DNA DSBs as indicated by the neutral comet assay and increased expression of the DSB-associated histone variant γ-H2AX (Supplementary Figures 2B and 3A). Secretion of Aβ₄₀ significantly increased after 2 h, 4 h, and 8 h of recovery following etoposide treatment, while secretion of Aβ₄₂ increased after 2 h and 4 h of recovery compared to controls (Fig. 2C). In this system, etoposide treatment significantly alters the ratio of Aβ₄₂/Aβ₄₀ only when compared to controls in ReN GA2 NPCs (Fig. 2D).

OPEN IN VIEWER

Fig. 2

Etoposide damage increases Aβ₄₀ and Aβ₄₂ secretion after 2 h and 4 h recovery in ReN GA2 NPCs. A) Treatment schematic for ReN GA2 NPCs treated with and without 10μM etoposide for 6 h and allowed to recover for designated times. (Schematic created in BioRender, 2021.) B) Percent cell viability after 6 h etoposide treatment and designated recovery times. C) Percent secreted Aβ₄₀ and Aβ₄₂ relative to controls after designated treatments and recovery times. D) Aβ₄₂/Aβ₄₀ ratio calculated from raw values. Error bars represent means±SD of three separate experiments, and the p values were determined using 2-way ANOVA and Tukey’s multiple comparisons test. ^*p < 0.05, ^**p < 0.01, ^****p < 0.0001, ns: not significant.

Previous reports in neurons have shown that treatment with hydrogen peroxide or 4-Hydroxenonenal (4-HNE), an aldehydic product of lipid peroxidation, leads to upregulation of the expression of BACE1, the rate limiting enzyme in the amyloidogenic process that cleaves AβPP, which results in an increase in intracellular and secreted Aβ, as well as activation of the apoptotic pathway [52–54]. Given our observed reduction in cell viability at all timepoints of recovery following 6 h treatment with etoposide (Fig. 2B), we therefore asked whether the increase in Aβ₄₀ and Aβ₄₂ secretion after etoposide treatment was due to the secretion of Aβ from living cells or the release of intracellular Aβ upon cell death (Fig. 3). We probed GA2 NPC cell lysates with antibodies specific for AβPP and BACE1 and probed the supernatant with antibodies specific for sAβPPβ, the cleavage product of BACE1. We found no significant increase in BACE1 expression after etoposide treatment compared to controls (Fig. 3A). However, sAβPPβ levels were increased 2 h after recovery following etoposide treatment in ReN GA2 NPCs (Fig. 3C). Total AβPP expression did not change at any recovery timepoint in ReN GA2 NPCs despite the increase in sAβPPβ and differential expression of mature and immature AβPP (upper and lower bands, respectively), which may indicate changes in intracellular AβPP trafficking (Fig. 3B). These results indicate that etoposide treatment does not affect BACE1 expression or total AβPP expression throughout the recovery timecourse. However, DNA damage, specifically DSBs, increases BACE1’s production of sAβPPβ and consequent increase in Aβ₄₀ and Aβ₄₂ secretion in ReN GA2 NPCs. This result may reflect increased processing of mature AβPP by BACE1 with a depletion of the pool of immature AβPP. In any case, the absence of an effect upon total AβPP supports our conclusion that this increase in secreted Aβ which we have found is not due to the release of Aβ upon cell death. Further investigation into the specific activity of BACE1 and other secretases involved in the production of Aβ₄₀ and Aβ₄₂ such as γ-secretase, as well as the intracellular trafficking of AβPP after etoposide damage, is necessary in order to fully define the mechanism of DSB induction of Aβ₄₀ and Aβ₄₂ secretion.

OPEN IN VIEWER

Fig. 3

DSBs increase sAβPPβ, but not total AβPP in ReN GA2 NPCs. A) Western blot of endogenous BACE1 expression in ReN GA2 NPCs and 20 day differentiated neurons treated with and without designated concentrations of etoposide for 6 h and allowed to recover for 2 h. B) Western blot of endogenous AβPP expression in ReN GA2 NPCs treated with and without 10μM etoposide for 6 h and allowed to recover for designated times. C) Western blot of sAβPPβ in ReN GA2 NPCs and 20 day differentiated neurons with and without 10μM of etoposide treatment for 6 h and allowed to recover for 2 h. Error bars represent means±SD of two to three separate experiments, and the p values were determined using two-way ANOVA and Tukey’s multiple comparisons test. ^***p < 0.001, ^****p < 0.0001.

Etoposide damage does not induce increased Aβ₄₀ and Aβ₄₂ secretion in differentiated ReN GA2 neurons

Mitotic and post-mitotic neurons maintain genomic integrity after induction of DSBs through different DDR mechanisms due primarily to the lack of a sister chromatid template for canonical homologous recombination (HR) in post-mitotic neurons [55]. Research has also shown that DNA DSBs are increased in human AD and 5xFAD mouse hippocampal astrocytes and neurons [25]. We therefore sought to determine whether DNA DSBs affect Aβ₄₀ and Aβ₄₂ generation in post-mitotic ReN GA2 neurons as well as NPCs. First, we differentiated the ReN GA2 NPCs into neurons using a previously established protocol for 20 days [37] (Fig. 4A). The ReN GA2 neurons were then fixed and immunolabeled with antibodies specific for the neuronal markers MAP2 and Tuj1, the astrocyte marker GFAP, and the nuclear protein proliferation marker Ki67 (Fig. 4B and Supplementary Figure 4A, B). The 20-day-differentiated ReN GA2 neurons showed distinct neuronal morphology, expressed the neuronal marker protein MAP2, did not express the astrocyte marker GFAP, and were non-proliferating (Ki67 negative). We then treated the ReN GA2 neuronal cultures with 10μM etoposide for 6 h and allowed them to recover for 2 h, 8 h, and 24 h (Fig. 4C). We confirmed increased DSBs in ReN GA2 neurons after etoposide treatment via a neutral comet assay as well as increased expression of the DSB-associated histone variant γ-H2AX (Supplementary Figures 2B and 3A). A dose response curve showed no significant increase in apoptosis after etoposide treatments compared to basal levels of controls as indicated by TUNEL staining (Supplementary Figure 3B). We found that after induction of DSBs, the post-mitotic 20-day ReN GA2 neurons do not secrete significantly increased amounts of Aβ₄₀ and Aβ₄₂. The ratio of Aβ₄₂/Aβ₄₀ also remained unchanged after etoposide treatment and recovery (Fig. 4D). BACE1 and sAβPPβ expression levels did not increase after etoposide treatment compared to controls (Fig. 3A). These results indicate that etoposide-induced DNA DSBs do not lead to increased secretion of Aβ₄₀ and Aβ₄₂ in 20-day differentiated ReN GA2 cells.

OPEN IN VIEWER

Fig. 4

Etoposide damage does not increase Aβ₄₀ and Aβ₄₂ secretion in differentiated ReN GA2 neurons. A) Treatment schematic for differentiating ReN GA2 neurons and treatment with and without 10μM etoposide for 6 h and designated recovery times. (Schematic created in BioRender, 2021.) B) Immunofluorescent staining of ReN GA2 NPCs and 20-day differentiated neurons with antibodies for anti-MAP2 (neuron) and anti-Ki67 (proliferation marker). C) Percent secreted Aβ₄₀ and Aβ₄₂ relative to controls after 6 h etoposide treatment and designated recovery times. D) Aβ₄₂/Aβ₄₀ ratio calculated from raw values. Error bars represent means±SD of three separate experiments.

DNA susceptibility, damage, and secretion of Aβ₄₀ and Aβ₄₂

Neuronal progenitors and post-mitotic cells are differentially susceptible to DNA damage and therefore secrete varying amounts of Aβ₄₀ and Aβ₄₂ in response to this damage as evidenced by novel data from this study. In the first section, we showed that global peroxide damage and etoposide treatments induce DNA DSBs in NPCs and neurons (Supplementary Figure 2A, B). Peroxide treatment results in the production of free radicals and oxidative stress that can affect proteins, lipids, and membranes, as well as resulting in numerous DNA lesions such as SSBs and DSBs.

Etoposide, which specifically inhibits topoisomerase II from re-ligating induced DSBs, differentially affects mitotic and post-mitotic cells. In mitotic cells, topoisomerase II also acts to prevent supercoiling of DNA during replication (reviewed in [57]). Supported by our comet assay data, progenitor cells experience DSBs induced by etoposide due to vulnerability at both replication sites and active transcription sites (Supplementary Figure 2B). However, topoisomerase II is only active in post-mitotic cells near the promoters at active transcription sites [58]. This leaves progenitor cells hypothetically more vulnerable to etoposide-induced DSBs. Moreover, in contrast to our results showing that mature ReN GA2 neurons do not accumulate increased Aβ despite having etoposide-induced DSBs, the brain in sporadic AD is marked by DSBs and Aβ, with most neurons being post-mitotic. It may be that the cellular milieu of ReN GA2 NPCs is more supportive of BACE1 increased processing of AβPP and Aβ accumulation than the cellular milieu of mature ReN GA2 neurons.

The DNA damage repair pathways also differ between the two cell types. Post-mitotic cells are unable to utilize the more error-free DSB repair pathway of HR due to the lack of a sister chromatid, but, like progenitor cells, they are capable of utilizing the RNA templated repair pathway TC-HR for DSB repair [26, 59]. Our prior research has shown that in terminally differentiated rat cortical neurons, high concentrations of extracellular Aβ₄₂ oligomers reduce the concentration of the transcription-coupled homologous recombination (TC-HR)-associated repair protein RAD52, and also inhibit its recruitment to DNA damage sites [26]. A recent characterization of the neuronal transcriptome has also shown that persistent and unrepaired DNA damage leads to the downregulation of RAD52 [60]. Other studies have shown that extracellular Aβ₄₂ oligomers reduce the HR-associated protein BRCA1 and increase the DSB-associated histone modification y-H2AX [27]. For neuronal progenitors that differentiate into post-mitotic neurons, our model hypothetically reflects the human brain where neural progenitors in the hippocampus and the cerebral cortex actively replicate early in life [61, 62]. Taken together, these observations strongly suggest that once Aβ accumulates, DNA damage repair is compromised. Additionally, with an increase in neuronal activity, activity-induced DSBs unrepaired by topoisomerase II must be repaired through either the HR or TC-HR pathways. In AD, these pathways would be inhibited by Aβ₄₂ oligomers, which would be increased by DSBs as found in this study. Given the fact that AD appears to be initiated 10–20 years before memory and cognition decline, when NPCs may still be present in vulnerable regions of the brain, our NPC data is of particular interest. With neuronal progenitors secreting increased Aβ after DNA damage and the resultant increase in the more toxic effects of Aβ such as increased ROS, this could lead to a feedback loop of further DNA SSB and DSB accumulation, especially if acute or chronic DNA damage is experienced early in life or in regions of the brain where the progenitor cell population is more robust. This consequent feedback loop would contribute to further DNA damage, more production of Aβ, inhibition of repair, and finally the collapse of neuronal genomic integrity. Thus, it may be that the Aβ seen in symptomatic AD had its origin in NPCs years before symptoms, when Aβ accumulation and AD began.

Aβ₄₀ and Aβ₄₂ toxicity

The two most abundant forms of Aβ, Aβ₄₀ and Aβ₄₂, are typically found in the human brain in the ratio Aβ₄₂/Aβ₄₀ of approximately 1:9 [63]. In the inherited form of early onset familial AD for which there are multiple known mutations, this ratio can shift to a higher percentage of Aβ₄₂. The increased ratio is associated with a more aggressive form of neurodegeneration than sporadic AD and negatively affects synaptic activity, neuronal viability, and memory formation in mouse models [64–66]. Aβ₄₂ is reported to be more fibrillogenic and forms more neurotoxic assemblies than Aβ₄₀ [67, 68]. Late-stage AD cerebrospinal fluid samples typically indicate lower Aβ₄₂/Aβ₄₀ ratios compared to non-AD patient samples. This is potentially due to neuronal loss, aggregated Aβ₄₂ retained in plaques in the brain, clearance rates of Aβ₄₀ and Aβ₄₂, and variance among individuals (reviewed in [69]). Initially, ReN GA2 NPCs cells secrete Aβ₄₀ up to three times higher in concentration than Aβ₄₂ [62]. In our study utilizing ReN GA2 NPCs, etoposide-induced DNA DSBs led to increased secretion of both Aβ₄₀ and Aβ₄₂ species, altering the overall ratio (Fig. 2D). From a biological standpoint, it is difficult to ascertain if the shift in the Aβ₄₂/Aβ₄₀ ratio is relevant due to the difference in experimental technique (cell culture versus human cerebrospinal fluid sampling).

Summary and limitations

In summary, while many authors have hypothesized that neuronal DNA DSBs might induce Aβ, our study is the first to demonstrate this experimentally in a cell-based model. The increased secretion of Aβ₄₀ and Aβ₄₂ after induction of DNA DSBs in neuronal progenitors points to the need to focus on the mechanism of development of AD many years before extensive neuronal loss and the accumulation of Aβ plaques typically seen once patients present with AD-associated symptoms at diagnosis. While our results indicate that DNA damage induces the production of Aβ, and Aβ likely contributes to AD pathology, other etiologies might also induce Aβ production, whether or not Aβ is the primary cause of AD, a contributing cause, or an epiphenomenon. Further, the interpretation of data generated in our model, based on DNA damage in ReN GA2 NPCs and neurons, may be limited by the fact that these cells have mutations that lead to high concentrations of Aβ before any extrinsic DNA damage. However, in our study, we would not have had a detectable Aβ baseline by ELISA without these cells. Moreover, our study does not establish whether DNA damage induces neuronal hyper-secretion of Aβ in the brain itself, given the physiologic complexity of this organ, and this question might be addressed in brain organoids [70]. If, in fact, Aβ accumulation is primarily caused by damage to DNA, the age-dependence of most cases of AD would be consistent with a lifetime of oxidative metabolic activity, especially in people with a genetic susceptibility to damage or to the lack of damage repair. This would be especially relevant in NPCs and neural stem cells which differentiate into post-mitotic neurons. If the accumulation of Aβ begins at early stages in development when NPCs are more populous, then mitigating the effects of acute or chronic DNA damage at those stages would be an effective approach to AD prevention. Moreover, future research that clarifies the role of Aβ in the DDR may provide preventive therapeutic strategies to combat the progression of AD before its irreversible pathologies emerge.