A three-dimensional human neural cell culture model of Alzheimer’s disease

Journal name:
Nature
Volume:
515,
Pages:
274–278
Date published:
DOI:
doi:10.1038/nature13800
Received
Accepted
Published online

Alzheimer’s disease is the most common form of dementia, characterized by two pathological hallmarks: amyloid-β plaques and neurofibrillary tangles1. The amyloid hypothesis of Alzheimer’s disease posits that the excessive accumulation of amyloid-β peptide leads to neurofibrillary tangles composed of aggregated hyperphosphorylated tau2, 3. However, to date, no single disease model has serially linked these two pathological events using human neuronal cells. Mouse models with familial Alzheimer’s disease (FAD) mutations exhibit amyloid-β-induced synaptic and memory deficits but they do not fully recapitulate other key pathological events of Alzheimer’s disease, including distinct neurofibrillary tangle pathology4, 5. Human neurons derived from Alzheimer’s disease patients have shown elevated levels of toxic amyloid-β species and phosphorylated tau but did not demonstrate amyloid-β plaques or neurofibrillary tangles6, 7, 8, 9, 10, 11. Here we report that FAD mutations in β-amyloid precursor protein and presenilin 1 are able to induce robust extracellular deposition of amyloid-β, including amyloid-β plaques, in a human neural stem-cell-derived three-dimensional (3D) culture system. More importantly, the 3D-differentiated neuronal cells expressing FAD mutations exhibited high levels of detergent-resistant, silver-positive aggregates of phosphorylated tau in the soma and neurites, as well as filamentous tau, as detected by immunoelectron microscopy. Inhibition of amyloid-β generation with β- or γ-secretase inhibitors not only decreased amyloid-β pathology, but also attenuated tauopathy. We also found that glycogen synthase kinase 3 (GSK3) regulated amyloid-β-mediated tau phosphorylation. We have successfully recapitulated amyloid-β and tau pathology in a single 3D human neural cell culture system. Our unique strategy for recapitulating Alzheimer’s disease pathology in a 3D neural cell culture model should also serve to facilitate the development of more precise human neural cell models of other neurodegenerative disorders.

At a glance

Figures

  1. Generation of hNPCs with multiple FAD mutations.
    Figure 1: Generation of hNPCs with multiple FAD mutations.

    a, Diagrams showing lentiviral internal ribosome entry sites (IRES) constructs. CMV, cytomegalovirus. b, Increased Aβ40 and Aβ42 levels in 6-week differentiated FAD ReN cells. Amyloid-β levels in conditioned media were normalized to total protein levels. *P < 0.05; **P < 0.01; ***P < 0.001; ANOVA followed by a post hoc Dunnett’s test; n = 3 per each sample. c, Amyloid-β levels are dramatically decreased in FAD ReN cells after treatment with 1 μM β-secretase inhibitor IV or 3.7 nM compound E. Mean ± s.e.m.; *P < 0.05; **P < 0.01; ***P < 0.001; ANOVA followed by a post hoc Dunnett’s test; n = 3 per each sample; ND, not detected.

  2. Robust increases of extracellular amyloid-[bgr] deposits in 3D-differentiated hNPCs with FAD mutations.
    Figure 2: Robust increases of extracellular amyloid-β deposits in 3D-differentiated hNPCs with FAD mutations.

    a, Thin-layer 3D culture protocol. HC, histochemistry; IF, immunofluorescence; IHC, immunohistochemistry. b, Amyloid-β deposits in 6-week differentiated control and FAD ReN cells in 3D Matrigel (green, GFP; blue, 3D6; scale bar, 25 μm; arrowheads, extracellular amyloid-β deposits; right-most panels, 3D6 staining was pseudo-coloured to red). c, Select confocal Z-stack images of 3D6-positive amyloid-β deposits. Z-sections with an interval of 2 μm were captured and sections 1, 3, 4, 6 and 19 are shown (green, GFP; red, 3D6). d, IHC of amyloid-β deposits in ReN-mGAP cells. 3D-differentiated cells were treated with 1 μM β-secretase inhibitor IV, 500 nM DAPT, 500 nM SGSM41 or dimethyl sulphoxide (DMSO). Brown, 3,3′-diaminobenzidine (DAB, BA27); blue, haematoxylin; scale bar, 25 μm; arrowhead, large amyloid-β deposits. e, Detection of amyloid plaques in ReN-mGAP cells with Amylo-Glo (green, GFP; blue, Amylo-Glo; arrows, Amylo-Glo-positive aggregates).

  3. Elevation of amyloid-[bgr] and p-tau levels in TBS-insoluble fractions of 3D-differentiated FAD hNPCs.
    Figure 3: Elevation of amyloid-β and p-tau levels in TBS-insoluble fractions of 3D-differentiated FAD hNPCs.

    a, A diagram showing a thick-layer 3D culture and detergent extraction protocols. b, Western blot of amyloid-β aggregates in 3D-differentiated ReN cells. 6E10-antibody-detected amyloid-β monomers, dimers, trimers and tetramers in SDS-soluble (upper panel) and formic-acid-soluble fractions (lower panel) from the control (ReN-G and -m) and the FAD ReN cells (ReN-GA, ReN-mGAP and HReN-mGAP) after 6 weeks of differentiation. c, Western blot of total and p-tau levels in SDS-soluble and formic-acid-soluble fractions.

  4. Detection of aggregated p-tau in the enriched ReN-mAP cells.
    Figure 4: Detection of aggregated p-tau in the enriched ReN-mAP cells.

    a, Immunofluorescence of p-tau and MAP2 in the enriched ReN-mAP and ReN-m cells after 3D differentiation. Green, p-tau (PHF1); red (pseudo-coloured), MAP2; scale bar, 25 μm; arrows, p-tau-positive neurites; arrowheads, p-tau-positive cell bodies. b, IHC of p-tau in the enriched ReN-mAP and ReN-m cells after 10-week 3D differentiation. The cells were treated with 1 μM β-secretase inhibitor IV, 3.7 nM compound E or DMSO during the final 2 weeks of the 3D differentiation. PHF1 antibody detected the elevated levels of p-tau in soma and neurites. Brown, p-tau; scale bar, 25 μm. c, Western blot of total and p-tau levels in 1% sarkosyl-soluble and -insoluble fractions. d, The modified Gallyas silver staining showed robust increases of strong silver deposits in cell bodies and neurite-like structures in the enriched ReN-mAP cells (lower panel) but not in the enriched ReN-m cells (upper panel). Scale bar, 25 μm; arrows indicate strong silver deposits in soma and neurites. e, Tau filaments were detected in sarkosyl-insoluble fractions from the enriched ReN-mAP cells by transmission electron microscopy after 7-week 3D differentiation. Black dots, anti-tau (tau46) antibodies labelled with immunogold anti-mouse antibodies; scale bar, 100 nm. Lower panel shows a digitally enlarged image (~3.5 fold) (arrowheads indicate filamentous structures).

  5. Generation of FACS-sorted ReN cells with FAD mutations.
    Extended Data Fig. 1: Generation of FACS-sorted ReN cells with FAD mutations.

    a, FACS sorting of ReNcell VM human neural stem (ReN) cells that were stably transfected with polycistronic GFP and/or mCherry lentiviral vector(s). The cells were then enriched based on GFP and/or mCherry signals by FACS (red-dotted boxes, the selected ranges of cells for the experiments; mCherry-A, area intensity of mCherry signal; GFP-A, area intensity of GFP signal). b, ReN cells stably expressing GFP alone (ReN-G), APPSL-GFP (ReN-GA), APPSL-GFP-PSEN1(ΔE9)-mCherry (ReN-mGAP), mCherry alone (ReN-m), APPSL-PSEN1(ΔE9)-mCherry (ReN-mAP) or GFP-APPSL-PSEN1(ΔE9)-mCherry (HReN-mGAP). Green, GFP; red, mCherry; scale bar, 25 μm. c, The representative fluorescence microscope images of ReN cells that were differentiated by growth-factor deprivation for 3 weeks (green, GFP; red, mCherry; scale bar, 25 μm). d, Immunofluorescence of neuronal (Tuj1) and glial markers (GFAP) in 3-week differentiated control and FAD ReN cells. Scale bar, 25 µm. e, Western blot of APPSL and PSEN1(ΔE9) expression in control (ReN-G and ReN-m) and FAD ReN (ReN-GA, ReN-mGAP and HReN-mGAP) cells. APP C-terminal fragment (CTF) levels were greatly increased by 500 nM DAPT treatments for 24 h. BACE1, β-secretase 1, F.L. APP, full-length APP. f, A table summarizing the control and FAD ReN cells generated for this study.

  6. Characterization of the control and FAD ReN cells.
    Extended Data Fig. 2: Characterization of the control and FAD ReN cells.

    a, Western blot of neuronal (MAP2, Tuj1, NCAM, synapsin 1) and glial (GFAP) markers in undifferentiated and 3-week differentiated control and FAD ReN cells. b, Confocal immunofluorescence of presynaptic (VGluT1, green) and dendritic (MAP2, red (pseudo-coloured)) markers in 6-week differentiated control ReN-m cells. Top-left, top-right and bottom-left panels, ×100 magnification; bottom-right panel represents a digitally magnified image of the respective outlined region for better visualization of punctate structures. c, qPCR array analysis of neuronal and glial markers of 7-week differentiated control ReN-G cells. Gene expression levels were normalized against β-actin levels in each sample and the fold changes were calculated by setting the expression levels of each gene in undifferentiated control ReN-G cells as 1 (n = 3 for ReN-G whereas n = 5 for ReN-G in 3D differentiation). FAD ReN cells (HReN-mGAP and ReN-mAP) showed a similar pattern of increases in neuronal and glial markers (data not shown). d, Analysis of 4-repeat (4R) or 3-repeat (3R) tau isoforms in 7-week differentiated control ReN cells (ReN-G and ReN-m) and FAD ReN cells (ReN-GA, ReN-mGAP, HReN-mGAP and ReN-mAP). Complementary DNA samples prepared from undifferentiated control ReN-G (1st lane) and human adult brains (9th lane) were used as controls. e, f, Electrophysiological properties of differentiated control ReN-G cells. The currents were elicited by 10 mV voltage steps from −100 to +60 mV in external solution (e) without (left panel in f) or with 500 nM tetrodotoxin (TTX, right panel in f). TTX treatment specifically blocked voltage-gated sodium currents. ReN-G cells were differentiated for 29 days by the previously described ‘preD’ method12. g, Sodium currents are shown as subtracted currents. h, Premature ReN-G cells (<16 days) mostly showed voltage-gated potassium currents without TTX-sensitive sodium currents. i, Western blot of amyloid-β (Aβ) levels in the conditioned media collected from 6-week differentiated control (ReN-m) and FAD ReN (ReN-mAP and HReN-mGAP) cells. j, A table summarizing APOE genotypes of control (ReN-m) and FAD ReN (ReN-mAP) cells used in this study. Two APOE SNP markers, rs429358 (minor allele = C) and rs7412 (minor allele = T), were used to determine APOE ε2/3/4 genotypes.

  7. Characterization of differentiated ReN cells in 3D cultures.
    Extended Data Fig. 3: Characterization of differentiated ReN cells in 3D cultures.

    a, Haematoxylin staining of a representative paraffin section (10 μm) from 9-week differentiated ReN-m cells in thick-layer 3D Matrigel. The sections were vertically cut to show the top and bottom of the 3D cultures. The pictures (10) were serially taken from top to bottom and digitally combined together. Bottom panel represents a digitally magnified image of the respective outlined region for better visualization; scale bar, 50 µm. b, The paraffin sections from control (ReN-G and ReN-m) and FAD ReN (ReN-GA and ReN-mAP (enriched)) were immunofluorescence stained with antibodies against the neuronal markers Tuj1 and MAP2. Blue, DAPI; scale bar, 25 μm. c, Immunofluorescence of thick-layer 3D-differentiated ReN-m cells stained with antibodies against additional neuronal markers tau, VGluT1 or GluR2 (glutamate receptor, ionotropic, AMPA 2); scale bar, 25 μm. d, Immunofluorescence of 3D-differentiated ReN-m cells (7 weeks) stained with antibodies against mature neuronal markers TH (tyrosine hydroxylase), NR2B (NMDA receptor 2B) or GABA(B)R2 (GABA-B-receptor 2); blue, DAPI; scale bar, 20 μm.

  8. Reconstitution of amyloid-[bgr] aggregates in 3D-cultured FAD ReN cells.
    Extended Data Fig. 4: Reconstitution of amyloid-β aggregates in 3D-cultured FAD ReN cells.

    a, IHC of amyloid-β deposits in control (ReN-G) and FAD ReN (HReN-mGAP) cells differentiated in 3D Matrigel. The control and FAD ReN cells were 3D-differentiated for 6 weeks and then treated with 1 μM BACE1 inhibitor IV (β-secretase inhibitor), 500 nM DAPT (γ-secretase inhibitor), 500 nM SGSM41 (γ-secretase modulator) or DMSO for an additional 3 weeks. The cultures were then fixed and immunostained with HRP-conjugated BA27 anti-Aβ40 antibodies (brown, DAB (BA27); blue, haematoxylin; scale bars, 25 μm; arrowhead, large amyloid-β deposits). b, Enlarged images of amyloid-β deposits in control (ReN-G) and FAD ReN (ReN-mGAP, HReN-mGAP) cell pictures shown in Fig. 2d and Extended Data Fig. 4a. Scale bar, 25 µm. c, IHC of amyloid-β deposits in control (ReN-G, left panels) and FAD ReN (ReN-mGAP, middle panels; HReN-mGAP, right panels) cells differentiated in 3D thin-layer Matrigels for 9 weeks. The fixed thin-layer 3D cultures were immunostained with HRP-conjugated BA27 anti-Aβ40 antibodies (brown, DAB (BA27); scale bar, 25 mm; arrows, large amyloid-β deposits). d, Amylo-Glo staining of ReN-G 3D cultures. Green, GFP; blue, Amylo-Glo; ×10 magnification.

  9. Accumulation of p-tau in FAD ReN cells.
    Extended Data Fig. 5: Accumulation of p-tau in FAD ReN cells.

    a, Elevated p-tau levels were significantly decreased by 1 μM DAPT (γ-secretase inhibitor) treatment in 6-week differentiated HReN-mGAP cells. The antibody against human specific mitochondrial marker (h-mito) was used to show an equal loading of the samples. b, Quantification of p-tau levels in control (DMSO)- and DAPT-treated HReN-mGAP cells (**P < 0.01; t-test; n = 3 per each sample). y axis values represent relative signal intensities (% adjusted volumes) of each p-tau band. c, p-tau IHC showed p-tau-positive cells in 3D-differentiated FAD ReN cells. Two p-tau antibodies, AT8 (pSer 199/Ser 202/Thr 205) and PHF1 (pSer 396/Ser 404), against different phosphorylation sites were used. Brown, p-tau; scale bar, 25 μm; arrows indicate cells with high levels of p-tau. d, High magnification (×100) images of p-tau-positive neurons (brown, AT8 p-tau). IHC of AT8 p-tau staining showed neurons with high levels of p-tau accumulation in soma and neurite-like structures (arrowheads). e, IHC of p-tau (AT8) staining showed the cells with p-tau accumulations in soma and neurite-like structures (arrow and arrowhead, respectively); scale bar, 25 µm. f, A digitally enlarged image of a dotted box in e. Brown, p-tau (AT8); blue, haematoxylin; scale bar, 25 μm. g, Total number of cells with high levels of p-tau in a single well of a 96-well plate was counted in control (ReN-G and ReN-m) and FAD ReN (ReN-GA, ReN-mGAP and HReN-mGAP) cells (*P < 0.05; t-test; n = 5 for control ReN cells and n = 12 for FAD ReN cells). h, BACE1 inhibitor IV (1 μM) or compound E (3.7 nM) treatments markedly decreased the number of cells with high p-tau accumulation in ReN-mAP cell populations (***P < 0.001; ANOVA followed by a post hoc Dunnett’s test; n = 3 for control ReN-m and ReN-mAP cells).

  10. FACS enrichment of ReN-mAP and ReN-m cells for higher expressions of APP and PS1.
    Extended Data Fig. 6: FACS enrichment of ReN-mAP and ReN-m cells for higher expressions of APP and PS1.

    a, FACS sorting of ReN-mAP cells with top 1–2% mCherry signal. mCherry-A, area intensity of mCherry signal. b, Representative images of the mCherry-labelled (red), enriched control ReN-m and ReN-mAP cells before or after the differentiation under growth-factor free conditions (×200 magnification); 1- and 3-week differentiated cells show increased neurite outgrowth. c, The enriched ReN-mAP cells secreted high levels of Aβ40 and Aβ42 after 9-week 3D differentiation. The secreted Aβ38/40/42 levels were measured by a multi-array Meso Scale electrochemiluminescence (Meso Scale SQ 120 system). Relative levels of amyloid-β (fold increases) were calculated by setting amyloid-β levels of the control ReN-m as 1. Amyloid-β levels were changed after treating 1 μM BACE1 inhibitor IV, 3.7 nM compound E or 500 nM SGSM41 (**P < 0.01; ***P < 0.001; ANOVA followed by a post hoc Dunnett’s test; n = 3 for the enriched ReN-m and ReN-mAP cells).

  11. Increased p-tau levels in FAD ReN cells.
    Extended Data Fig. 7: Increased p-tau levels in FAD ReN cells.

    a, Immunofluorescence of AT8 p-tau and Tuj1 in the enriched ReN-mAP and control ReN-m cells after 9 weeks of 3D differentiation. BACE1 inhibitor IV treatment for 3 weeks dramatically reduced AT8 p-tau staining (green, AT8 p-tau; red (pseudo-coloured), Tuj1; scale bar, 25 μm). b, Western blot of total and p-tau levels in control (enriched ReN-m) and FAD ReN (enriched ReN-mAP) cells. The cells were 3D differentiated for 9 weeks. Three weeks of BACE1 inhibitor treatments significantly decreased p-tau levels without changing total tau levels. HSP70 heat shock protein levels are shown to demonstrate equal loading of each sample.

  12. Immunoelectron microscopy analysis of sarkosyl-insoluble fraction from FAD and control ReN cells.
    Extended Data Fig. 8: Immunoelectron microscopy analysis of sarkosyl-insoluble fraction from FAD and control ReN cells.

    a, Sarkosyl-insoluble fractions prepared from 3D-differentiated ReN-mAP (enriched, 7-week differentiated) were placed on carbon grids, labelled with tau46 and anti-mouse 10 nm gold antibodies and imaged using a JEOL JEM 1011 transmission electron microscope (scale bar, 500 nm). b, Sarkosyl-insoluble fractions from 3D-differentiated control ReN-G cells (7-week differentiated). No immunogold-labelled filamentous structures were detected in these samples (scale bar, 500 nm).

  13. Treatment with 1-azakenpaullone, a GSK3[bgr] inhibitor, decreased amyloid-[bgr]-induced tau phosphorylation without changing total amyloid-[bgr] levels.
    Extended Data Fig. 9: Treatment with 1-azakenpaullone, a GSK3β inhibitor, decreased amyloid-β-induced tau phosphorylation without changing total amyloid-β levels.

    a, Immunofluorescence of p-tau and MAP2 in the enriched ReN-mAP and control ReN-m cells with or without treatment with 1-azakenpaullone, a GSK3β inhibitor. The differentiated cells were treated with 2.5 μM 1-azakenpaullone or DMSO for the last 5 days of the 3D differentiation (green, p-tau (PHF1); red (pseudo-coloured), MAP2; scale bar, 25 μm). b, Western blot of total and p-tau levels in control (enriched ReN-m) and FAD ReN (enriched ReN-mAP) cells. The cells were 3D differentiated for 4 weeks followed by additional 5-day treatments of DMSO or 2.5 μM 1-azakenpaullone. c, Analysis of Aβ40 and Aβ42 levels in the enriched ReN-mAP cells treated with either DMSO or 2.5 μM 1-azakenpaullone (1-Aza) under the same conditions.

Tables

  1. The structure and the properties of SGSM41, a novel soluble [ggr]-secretase modulator
    Extended Data Table 1: The structure and the properties of SGSM41, a novel soluble γ-secretase modulator

Videos

  1. 3D reconstitution of ReN-G cells after 2-week differentiation in 3D Matrigel (green, GFP; 400x magnification)
    Video 1: 3D reconstitution of ReN-G cells after 2-week differentiation in 3D Matrigel (green, GFP; 400x magnification)
    Due to the technical limitation of spinning disk confocal microscopy (Olympus), 3D reconstituted images represent only a part of the whole 3D culture.
  2. 3D reconstitution of ReN-m (enriched) cells after 4-week differentiation in 3D Matrigel (red, mCherry; 400x magnification)
    Video 2: 3D reconstitution of ReN-m (enriched) cells after 4-week differentiation in 3D Matrigel (red, mCherry; 400x magnification)
    3D reconstitution of ReN-m (enriched) cells after 4-week differentiation in 3D Matrigel (red, mCherry; 400x magnification)
  3. 3D reconstitution of 3D-differentiated ReN-m cells for 6 weeks (green, Synapsin I; red, mCherry; 1,000x magnification).
    Video 3: 3D reconstitution of 3D-differentiated ReN-m cells for 6 weeks (green, Synapsin I; red, mCherry; 1,000x magnification).
    3D reconstitution of 3D-differentiated ReN-m cells for 6 weeks (green, Synapsin I; red, mCherry; 1,000x magnification).
  4. Z-section sequences of 3D6-stained Aβ aggregates in 6-week differentiated ReN-mGAP cells (green, GFP; 400x magnification).
    Video 4: Z-section sequences of 3D6-stained Aβ aggregates in 6-week differentiated ReN-mGAP cells (green, GFP; 400x magnification).
    Z-section sequences of 3D6-stained Aβ aggregates in 6-week differentiated ReN-mGAP cells (green, GFP; 400x magnification).

References

  1. Tanzi, R. E. & Bertram, L. Twenty years of the Alzheimer’s disease amyloid hypothesis: a genetic perspective. Cell 120, 545555 (2005)
  2. Hardy, J. & Selkoe, D. J. The amyloid hypothesis of Alzheimer’s disease: progress and problems on the road to therapeutics. Science 297, 353356 (2002)
  3. Selkoe, D. Alzheimer’s disease is a synaptic failure. Science 298, 789791 (2002)
  4. Duff, K. Transgenic mouse models of Alzheimer’s disease: phenotype and mechanisms of pathogenesis. Biochem. Soc. Symp. 67, 195202 (2001)
  5. Chin, J. Selecting a mouse model of Alzheimer’s disease. Methods Mol. Biol. 670, 169189 (2011)
  6. Choi, S. H. & Tanzi, R. E. iPSCs to the rescue in Alzheimer’s research. Cell Stem Cell 10, 235236 (2012)
  7. Yagi, T. et al. Modeling familial Alzheimer’s disease with induced pluripotent stem cells. Hum. Mol. Genet. 20, 45304539 (2011)
  8. Israel, M. A. et al. Probing sporadic and familial Alzheimer’s disease using induced pluripotent stem cells. Nature 482, 216220 (2012)
  9. Kondo, T. et al. Modeling Alzheimer’s disease with iPSCs reveals stress phenotypes associated with intracellular Aβ and differential drug responsiveness. Cell Stem Cell 12, 487496 (2013)
  10. Muratore, C. R. et al. The familial Alzheimer’s disease APPV717I mutation alters APP processing and Tau expression in iPSC-derived neurons. Hum. Mol. Genet. 23, 35233536 (2014)
  11. Sproul, A. A. et al. Characterization and molecular profiling of PSEN1 familial Alzheimer’s disease iPSC-derived neural progenitors. PLoS ONE 9, e84547 (2014)
  12. Donato, R. et al. Differential development of neuronal physiological responsiveness in two human neural stem cell lines. BMC Neurosci. 8, 36 (2007)
  13. Morgan, P. J. et al. Protection of neurons derived from human neural progenitor cells by veratridine. Neuroreport 20, 12251229 (2009)
  14. Mazemondet, O. et al. Quantitative and kinetic profile of Wnt/β-catenin signaling components during human neural progenitor cell differentiation. Cell. Mol. Biol. Lett. 16, 515538 (2011)
  15. Li, H., Wijekoon, A. & Leipzig, N. D. 3D differentiation of neural stem cells in macroporous photopolymerizable hydrogel scaffolds. PLoS ONE 7, e48824 (2012)
  16. Lancaster, M. A. et al. Cerebral organoids model human brain development and microcephaly. Nature 501, 373379 (2013)
  17. Tang-Schomer, M. D. et al. Bioengineered functional brain-like cortical tissue. Proc. Natl Acad. Sci. USA http://dx.doi.org/10.1073/pnas.1324214111 (August 11, 2014)
  18. Higuchi, M., Trojanowski, J. Q. & Lee, V. M.-Y. Tau protein and tauopathy. Neuropsychopharmacology: The Fifth Generation of Progress (eds Davis, K. L., Charney, D., Coyle, J. T. & Nemeroff, C.) 13391354 (Lippincott, Williams & Wilkins, 2002)
  19. Wagner, S. L. et al. Soluble γ-secretase modulators selectively inhibit the production of the 42-amino acid amyloid β peptide variant and augment the production of multiple carboxy-truncated amyloid β species. Biochemistry 53, 702713 (2014)
  20. Trojanowski, J. Q. & Lee, V. M.-Y. The role of tau in Alzheimer’s disease. Med. Clin. North Am. 86, 615627 (2002)
  21. Cras, P. et al. Neuronal and microglial involvement in β-amyloid protein deposition in Alzheimer’s disease. Am. J. Pathol. 137, 241246 (1990)
  22. Nadler, J. V. & Evenson, D. A. Use of excitatory amino acids to make axon-sparing lesions of hypothalamus. Methods Enzymol. 103, 393400 (1983)
  23. Kunick, C., Lauenroth, K., Leost, M., Meijer, L. & Lemcke, T. 1-Azakenpaullone is a selective inhibitor of glycogen synthase kinase-3β. Bioorg. Med. Chem. Lett. 14, 413416 (2004)
  24. Kirby, L. A. et al. Glycogen synthase kinase 3 (GSK3) inhibitor, SB-216763, promotes pluripotency in mouse embryonic stem cells. PLoS ONE 7, e39329 (2012)
  25. Phiel, C. J., Wilson, C. A., Lee, V. M.-Y. & Klein, P. S. GSK-3α regulates production of Alzheimer’s disease amyloid-β peptides. Nature 423, 435439 (2003)
  26. Jaworski, T. et al. GSK-3α/β kinases and amyloid production in vivo. Nature 423, 435439 (2003)
  27. Cohen, R. M. et al. A transgenic Alzheimer rat with plaques, tau pathology, behavioral impairment, oligomeric Aβ, and frank neuronal loss. J. Neurosci. 33, 62456256 (2013)
  28. Planel, E. et al. Acceleration and persistence of neurofibrillary pathology in a mouse model of tauopathy following anesthesia. FASEB J. 23, 25952604 (2009)
  29. Julien, C., Bretteville, A. & Planel, E. Biochemical isolation of insoluble tau in transgenic mouse models of tauopathies. Methods Mol. Biol. 849, 473491 (2012)
  30. Iovino, M., Patani, R., Watts, C., Chandran, S. & Spillantini, M. G. Human stem cell-derived neurons: a system to study human tau function and dysfunction. PLoS ONE 5, e13947 (2010)

Download references

Author information

  1. These authors contributed equally to this work.

    • Se Hoon Choi &
    • Young Hye Kim

Affiliations

  1. Genetics and Aging Research Unit, MassGeneral Institute for Neurodegenerative Disease, Massachusetts General Hospital, Harvard Medical School, Charlestown, Massachusetts 02129, USA

    • Se Hoon Choi,
    • Young Hye Kim,
    • Matthias Hebisch,
    • Christopher Sliwinski,
    • Carla D’Avanzo,
    • Hechao Chen,
    • Basavaraj Hooli,
    • Caroline Asselin,
    • Justin B. Klee,
    • Can Zhang,
    • Dora M. Kovacs,
    • Rudolph E. Tanzi &
    • Doo Yeon Kim
  2. Division of Mass Spectrometry Research, Korea Basic Science Institute, Cheongju-si, Chungbuk 363-883, South Korea

    • Young Hye Kim
  3. Institute of Reconstructive Neurobiology, Life and Brain Center, University of Bonn and Hertie Foundation, 53127 Bonn, Germany

    • Matthias Hebisch &
    • Michael Peitz
  4. FM Kirby Neurobiology Center, Boston Children’s Hospital and Harvard Stem Cell Institute, Boston, Massachusetts 02115, USA

    • Seungkyu Lee,
    • Brian J. Wainger &
    • Clifford J. Woolf
  5. The Whitehead Institute for Biomedical Research, Cambridge, Massachusetts 02142, USA

    • Julien Muffat
  6. Department of Neurosciences, University of California, San Diego, La Jolla, California 92093, USA

    • Steven L. Wagner

Contributions

D.Y.K. and R.E.T. were equally responsible for experimental design and data interpretation. S.H.C., Y.H.K. and D.Y.K. mainly contributed to writing and revising the manuscript. D.Y.K., Y.H.K., S.H.C., M.H., S.L., C.D., H.C., C.S., B.H., J.B.K., C.A. and C.Z. conducted the experiments. S.L.W. synthesized SGSM41 and S.L.W. and C.Z. characterized SGSM41. J.M., B.J.W., M.P., C.J.W. and D.M.K. contributed to data interpretation.

Competing financial interests

The authors declare no competing financial interests.

Corresponding authors

Correspondence to:

Author details

Extended data figures and tables

Extended Data Figures

  1. Extended Data Figure 1: Generation of FACS-sorted ReN cells with FAD mutations. (990 KB)

    a, FACS sorting of ReNcell VM human neural stem (ReN) cells that were stably transfected with polycistronic GFP and/or mCherry lentiviral vector(s). The cells were then enriched based on GFP and/or mCherry signals by FACS (red-dotted boxes, the selected ranges of cells for the experiments; mCherry-A, area intensity of mCherry signal; GFP-A, area intensity of GFP signal). b, ReN cells stably expressing GFP alone (ReN-G), APPSL-GFP (ReN-GA), APPSL-GFP-PSEN1(ΔE9)-mCherry (ReN-mGAP), mCherry alone (ReN-m), APPSL-PSEN1(ΔE9)-mCherry (ReN-mAP) or GFP-APPSL-PSEN1(ΔE9)-mCherry (HReN-mGAP). Green, GFP; red, mCherry; scale bar, 25 μm. c, The representative fluorescence microscope images of ReN cells that were differentiated by growth-factor deprivation for 3 weeks (green, GFP; red, mCherry; scale bar, 25 μm). d, Immunofluorescence of neuronal (Tuj1) and glial markers (GFAP) in 3-week differentiated control and FAD ReN cells. Scale bar, 25 µm. e, Western blot of APPSL and PSEN1(ΔE9) expression in control (ReN-G and ReN-m) and FAD ReN (ReN-GA, ReN-mGAP and HReN-mGAP) cells. APP C-terminal fragment (CTF) levels were greatly increased by 500 nM DAPT treatments for 24 h. BACE1, β-secretase 1, F.L. APP, full-length APP. f, A table summarizing the control and FAD ReN cells generated for this study.

  2. Extended Data Figure 2: Characterization of the control and FAD ReN cells. (500 KB)

    a, Western blot of neuronal (MAP2, Tuj1, NCAM, synapsin 1) and glial (GFAP) markers in undifferentiated and 3-week differentiated control and FAD ReN cells. b, Confocal immunofluorescence of presynaptic (VGluT1, green) and dendritic (MAP2, red (pseudo-coloured)) markers in 6-week differentiated control ReN-m cells. Top-left, top-right and bottom-left panels, ×100 magnification; bottom-right panel represents a digitally magnified image of the respective outlined region for better visualization of punctate structures. c, qPCR array analysis of neuronal and glial markers of 7-week differentiated control ReN-G cells. Gene expression levels were normalized against β-actin levels in each sample and the fold changes were calculated by setting the expression levels of each gene in undifferentiated control ReN-G cells as 1 (n = 3 for ReN-G whereas n = 5 for ReN-G in 3D differentiation). FAD ReN cells (HReN-mGAP and ReN-mAP) showed a similar pattern of increases in neuronal and glial markers (data not shown). d, Analysis of 4-repeat (4R) or 3-repeat (3R) tau isoforms in 7-week differentiated control ReN cells (ReN-G and ReN-m) and FAD ReN cells (ReN-GA, ReN-mGAP, HReN-mGAP and ReN-mAP). Complementary DNA samples prepared from undifferentiated control ReN-G (1st lane) and human adult brains (9th lane) were used as controls. e, f, Electrophysiological properties of differentiated control ReN-G cells. The currents were elicited by 10 mV voltage steps from −100 to +60 mV in external solution (e) without (left panel in f) or with 500 nM tetrodotoxin (TTX, right panel in f). TTX treatment specifically blocked voltage-gated sodium currents. ReN-G cells were differentiated for 29 days by the previously described ‘preD’ method12. g, Sodium currents are shown as subtracted currents. h, Premature ReN-G cells (<16 days) mostly showed voltage-gated potassium currents without TTX-sensitive sodium currents. i, Western blot of amyloid-β (Aβ) levels in the conditioned media collected from 6-week differentiated control (ReN-m) and FAD ReN (ReN-mAP and HReN-mGAP) cells. j, A table summarizing APOE genotypes of control (ReN-m) and FAD ReN (ReN-mAP) cells used in this study. Two APOE SNP markers, rs429358 (minor allele = C) and rs7412 (minor allele = T), were used to determine APOE ε2/3/4 genotypes.

  3. Extended Data Figure 3: Characterization of differentiated ReN cells in 3D cultures. (1,243 KB)

    a, Haematoxylin staining of a representative paraffin section (10 μm) from 9-week differentiated ReN-m cells in thick-layer 3D Matrigel. The sections were vertically cut to show the top and bottom of the 3D cultures. The pictures (10) were serially taken from top to bottom and digitally combined together. Bottom panel represents a digitally magnified image of the respective outlined region for better visualization; scale bar, 50 µm. b, The paraffin sections from control (ReN-G and ReN-m) and FAD ReN (ReN-GA and ReN-mAP (enriched)) were immunofluorescence stained with antibodies against the neuronal markers Tuj1 and MAP2. Blue, DAPI; scale bar, 25 μm. c, Immunofluorescence of thick-layer 3D-differentiated ReN-m cells stained with antibodies against additional neuronal markers tau, VGluT1 or GluR2 (glutamate receptor, ionotropic, AMPA 2); scale bar, 25 μm. d, Immunofluorescence of 3D-differentiated ReN-m cells (7 weeks) stained with antibodies against mature neuronal markers TH (tyrosine hydroxylase), NR2B (NMDA receptor 2B) or GABA(B)R2 (GABA-B-receptor 2); blue, DAPI; scale bar, 20 μm.

  4. Extended Data Figure 4: Reconstitution of amyloid-β aggregates in 3D-cultured FAD ReN cells. (2,139 KB)

    a, IHC of amyloid-β deposits in control (ReN-G) and FAD ReN (HReN-mGAP) cells differentiated in 3D Matrigel. The control and FAD ReN cells were 3D-differentiated for 6 weeks and then treated with 1 μM BACE1 inhibitor IV (β-secretase inhibitor), 500 nM DAPT (γ-secretase inhibitor), 500 nM SGSM41 (γ-secretase modulator) or DMSO for an additional 3 weeks. The cultures were then fixed and immunostained with HRP-conjugated BA27 anti-Aβ40 antibodies (brown, DAB (BA27); blue, haematoxylin; scale bars, 25 μm; arrowhead, large amyloid-β deposits). b, Enlarged images of amyloid-β deposits in control (ReN-G) and FAD ReN (ReN-mGAP, HReN-mGAP) cell pictures shown in Fig. 2d and Extended Data Fig. 4a. Scale bar, 25 µm. c, IHC of amyloid-β deposits in control (ReN-G, left panels) and FAD ReN (ReN-mGAP, middle panels; HReN-mGAP, right panels) cells differentiated in 3D thin-layer Matrigels for 9 weeks. The fixed thin-layer 3D cultures were immunostained with HRP-conjugated BA27 anti-Aβ40 antibodies (brown, DAB (BA27); scale bar, 25 mm; arrows, large amyloid-β deposits). d, Amylo-Glo staining of ReN-G 3D cultures. Green, GFP; blue, Amylo-Glo; ×10 magnification.

  5. Extended Data Figure 5: Accumulation of p-tau in FAD ReN cells. (980 KB)

    a, Elevated p-tau levels were significantly decreased by 1 μM DAPT (γ-secretase inhibitor) treatment in 6-week differentiated HReN-mGAP cells. The antibody against human specific mitochondrial marker (h-mito) was used to show an equal loading of the samples. b, Quantification of p-tau levels in control (DMSO)- and DAPT-treated HReN-mGAP cells (**P < 0.01; t-test; n = 3 per each sample). y axis values represent relative signal intensities (% adjusted volumes) of each p-tau band. c, p-tau IHC showed p-tau-positive cells in 3D-differentiated FAD ReN cells. Two p-tau antibodies, AT8 (pSer 199/Ser 202/Thr 205) and PHF1 (pSer 396/Ser 404), against different phosphorylation sites were used. Brown, p-tau; scale bar, 25 μm; arrows indicate cells with high levels of p-tau. d, High magnification (×100) images of p-tau-positive neurons (brown, AT8 p-tau). IHC of AT8 p-tau staining showed neurons with high levels of p-tau accumulation in soma and neurite-like structures (arrowheads). e, IHC of p-tau (AT8) staining showed the cells with p-tau accumulations in soma and neurite-like structures (arrow and arrowhead, respectively); scale bar, 25 µm. f, A digitally enlarged image of a dotted box in e. Brown, p-tau (AT8); blue, haematoxylin; scale bar, 25 μm. g, Total number of cells with high levels of p-tau in a single well of a 96-well plate was counted in control (ReN-G and ReN-m) and FAD ReN (ReN-GA, ReN-mGAP and HReN-mGAP) cells (*P < 0.05; t-test; n = 5 for control ReN cells and n = 12 for FAD ReN cells). h, BACE1 inhibitor IV (1 μM) or compound E (3.7 nM) treatments markedly decreased the number of cells with high p-tau accumulation in ReN-mAP cell populations (***P < 0.001; ANOVA followed by a post hoc Dunnett’s test; n = 3 for control ReN-m and ReN-mAP cells).

  6. Extended Data Figure 6: FACS enrichment of ReN-mAP and ReN-m cells for higher expressions of APP and PS1. (475 KB)

    a, FACS sorting of ReN-mAP cells with top 1–2% mCherry signal. mCherry-A, area intensity of mCherry signal. b, Representative images of the mCherry-labelled (red), enriched control ReN-m and ReN-mAP cells before or after the differentiation under growth-factor free conditions (×200 magnification); 1- and 3-week differentiated cells show increased neurite outgrowth. c, The enriched ReN-mAP cells secreted high levels of Aβ40 and Aβ42 after 9-week 3D differentiation. The secreted Aβ38/40/42 levels were measured by a multi-array Meso Scale electrochemiluminescence (Meso Scale SQ 120 system). Relative levels of amyloid-β (fold increases) were calculated by setting amyloid-β levels of the control ReN-m as 1. Amyloid-β levels were changed after treating 1 μM BACE1 inhibitor IV, 3.7 nM compound E or 500 nM SGSM41 (**P < 0.01; ***P < 0.001; ANOVA followed by a post hoc Dunnett’s test; n = 3 for the enriched ReN-m and ReN-mAP cells).

  7. Extended Data Figure 7: Increased p-tau levels in FAD ReN cells. (454 KB)

    a, Immunofluorescence of AT8 p-tau and Tuj1 in the enriched ReN-mAP and control ReN-m cells after 9 weeks of 3D differentiation. BACE1 inhibitor IV treatment for 3 weeks dramatically reduced AT8 p-tau staining (green, AT8 p-tau; red (pseudo-coloured), Tuj1; scale bar, 25 μm). b, Western blot of total and p-tau levels in control (enriched ReN-m) and FAD ReN (enriched ReN-mAP) cells. The cells were 3D differentiated for 9 weeks. Three weeks of BACE1 inhibitor treatments significantly decreased p-tau levels without changing total tau levels. HSP70 heat shock protein levels are shown to demonstrate equal loading of each sample.

  8. Extended Data Figure 8: Immunoelectron microscopy analysis of sarkosyl-insoluble fraction from FAD and control ReN cells. (227 KB)

    a, Sarkosyl-insoluble fractions prepared from 3D-differentiated ReN-mAP (enriched, 7-week differentiated) were placed on carbon grids, labelled with tau46 and anti-mouse 10 nm gold antibodies and imaged using a JEOL JEM 1011 transmission electron microscope (scale bar, 500 nm). b, Sarkosyl-insoluble fractions from 3D-differentiated control ReN-G cells (7-week differentiated). No immunogold-labelled filamentous structures were detected in these samples (scale bar, 500 nm).

  9. Extended Data Figure 9: Treatment with 1-azakenpaullone, a GSK3β inhibitor, decreased amyloid-β-induced tau phosphorylation without changing total amyloid-β levels. (981 KB)

    a, Immunofluorescence of p-tau and MAP2 in the enriched ReN-mAP and control ReN-m cells with or without treatment with 1-azakenpaullone, a GSK3β inhibitor. The differentiated cells were treated with 2.5 μM 1-azakenpaullone or DMSO for the last 5 days of the 3D differentiation (green, p-tau (PHF1); red (pseudo-coloured), MAP2; scale bar, 25 μm). b, Western blot of total and p-tau levels in control (enriched ReN-m) and FAD ReN (enriched ReN-mAP) cells. The cells were 3D differentiated for 4 weeks followed by additional 5-day treatments of DMSO or 2.5 μM 1-azakenpaullone. c, Analysis of Aβ40 and Aβ42 levels in the enriched ReN-mAP cells treated with either DMSO or 2.5 μM 1-azakenpaullone (1-Aza) under the same conditions.

Extended Data Tables

  1. Extended Data Table 1: The structure and the properties of SGSM41, a novel soluble γ-secretase modulator (46 KB)

Supplementary information

Video

  1. Video 1: 3D reconstitution of ReN-G cells after 2-week differentiation in 3D Matrigel (green, GFP; 400x magnification) (8.16 MB, Download)
    Due to the technical limitation of spinning disk confocal microscopy (Olympus), 3D reconstituted images represent only a part of the whole 3D culture.
  2. Video 2: 3D reconstitution of ReN-m (enriched) cells after 4-week differentiation in 3D Matrigel (red, mCherry; 400x magnification) (2.89 MB, Download)
    3D reconstitution of ReN-m (enriched) cells after 4-week differentiation in 3D Matrigel (red, mCherry; 400x magnification)
  3. Video 3: 3D reconstitution of 3D-differentiated ReN-m cells for 6 weeks (green, Synapsin I; red, mCherry; 1,000x magnification). (25.22 MB, Download)
    3D reconstitution of 3D-differentiated ReN-m cells for 6 weeks (green, Synapsin I; red, mCherry; 1,000x magnification).
  4. Video 4: Z-section sequences of 3D6-stained Aβ aggregates in 6-week differentiated ReN-mGAP cells (green, GFP; 400x magnification). (3.75 MB, Download)
    Z-section sequences of 3D6-stained Aβ aggregates in 6-week differentiated ReN-mGAP cells (green, GFP; 400x magnification).

Comments

  1. Report this comment #64385

    Chris Exley said:

    Is there something fundamental missing from this study? As far as I have been able to ascertain the authors do not demonstrate the assembly of Abeta into beta sheet structures. They use a 'new' product called Amylo-Glo to identify deposits of Abeta. However, I have not been able to find any unequivocal demonstration in the scientific literature that this product only gives fluorescence when bound to amyloid beta sheets. One wonders why the authors did not use ThT fluorescence microscopy combined with TEM to confirm that the Abeta released formed beta pleated sheets. As it stands I do not think that this has been shown and this could be a significant flaw in this model?

  2. Report this comment #64401

    Chris Exley said:

    Following up my previous comment. I am unable to find how the amyloid beta levels given in Figure 1 were measured? What were the absolute concentrations? Expressing per mg of protein is unhelpful. How can the protein content of an assay medium be measured without also including the content of amyloid beta in the total?
    The question arises as to what were the total concentrations in nM (nanomolar) of amyloid beta 40 and amyloid beta 42 in the conditioned media? Presumably they exceeded the solubility of these peptides but this should be demonstrated both in the 2D system and later in the 3D system.

  3. Report this comment #64409

    Doo Yeon Kim said:

    Thank you very much for your comments on our study. We agree that use of ThT fluorescence would be useful as a more traditional alternative to the relatively new Amylo-Glo products. But unfortunately, our AD cell line models (ReN-GA, ReN-mGAP, HReN-mGAP) used for Abeta aggregation, express high levels of eGFP, which overlaps with ThT fluorescence. Thus, we had no choice. Data was not shown in the current paper due to the space limitation, we also performed Congo red staining, a well-known amyloid beta sheets dye, and detected Congo red stained aggregates only in 3D cultures from AD model cells. We also performed biochemical extraction and showed accumulation of Abeta species including SDS-resistant dimer. trimer and even tetramer species in TBS-insoluble, SDS-insoluble fractions (Fig. 3b), which is another convincing proof of Abeta aggregation in our system. At least, based on current publications (Schmued et al., 2012; Sarkar et al., 2014), Amylo-Glo successfully stained amyloid plaques in brain tissue sections. We also have communications with Biosensis, a distributing company of Amylo-Glo, and they reconfirmed that this dye can specifically detect amyloid plaques in AD animal models and human AD brain sections. They also stressed that Schmued et al. showed a nice comparison of Amylo-glo staining in AD transgenic mice and compared it directly to all of the other major AD dyes includling ThT and congo red, and even antibody staining. Together with our immunostaining with widely used Abeta peptide antibodies, we believe that our current data is strong enough to convince that amyloid aggregates/plaque can be formed in our 3D culture conditions. We are currently working on TEM of Amyloid aggregates in the same 3D culture system and we hope that we can provide an additional proof for the presence of amyloid plaque in our 3D culture system.
    Doo Yeon Kim, Ph.D.
    Rudolph E. Tanzi, Ph.D.

  4. Report this comment #64415

    Doo Yeon Kim said:

    Again, we appreciate Dr. Exley?s comments on our paper. Regarding amyloid beta levels shown in Figure 1, we first measured Abeta 40 and 42 CONCENTRATIONS (pmol/ml) in the conditioned media collected from 2D-differentiated ReN cells, using amyloid-b ELISA Kit from Wako (Osaka, Japan). The secreted Abeta40 and 42 levels in the media were then further normalized by the total cellular protein levels of the corresponding cell pellets, which is a common practice in AD laboratories measuring soluble Abeta levels (check Israel et al., 2012, Koch et al., 2012). Without this normalization, soluble Abeta concentration in the media will be different in every experiments depending on the cell numbers plated in each cell culture wells. If needed, we will be happy to upload original Abeta concentration (pmol/ml) data before normalization in a public domain.

    Regarding Dr. Exley?s assumption that Abeta concentration in our ReN 2D culture media, exceed the solubility limit of Abeta peptide, we have never seen any sign of Abeta aggregation in the conditioned media (Please see our Extended Data Fig. 2i showing that only Abeta monomer bands in the 2D ReN cell media; In the 3D culture, you can appreciate the presence of SDS-resistant Abeta aggregates including Abeta dimer, trimer and tetramers (Fig. 3b)). We think that 3D culture conditions provide local environment with higher pathogenic Abeta concentration, which can lead to Abeta aggregation we have shown in our paper. We believe that this is one of the interesting findings of our paper.

    Doo Yeon Kim
    Rudolph E Tanzi

  5. Report this comment #64431

    Chris Exley said:

    Thanks for replying. I did, of course, read these papers on AG before writing my comments. Reading carefully you will see that AG definitely stains 'amyloid' deposits. However, in neither paper is it demonstrated that AG is specific for beta sheets. For example, Congo red stains red for amyloid BUT the presence of beta sheets can only be confirmed under polarised light when beta sheets give apple-green birefringence. AG could well be very useful for fluorescence microscopy of amyloid but it will require much better validation before we can be confident that it also is indicative of beta sheets.

Subscribe to comments

Additional data