|Home | About | Journals | Submit | Contact Us | Français|
Stem cell technologies have facilitated the development of human cellular disease models that can be used to study pathogenesis and test therapeutic candidates. These models hold promise for complex neurological diseases such as Alzheimer’s disease (AD) because existing animal models have been unable to fully recapitulate all aspects of pathology. We recently reported the characterization of a novel three-dimensional (3D) culture system that exhibits key events in AD pathogenesis, including extracellular aggregation of β-amyloid and accumulation of hyperphosphorylated tau. Here we provide instructions for the generation and analysis of 3D human neural cell cultures, including the production of genetically modified human neural progenitor cells (hNPCs) with familial AD mutations, the differentiation of the hNPCs in a 3D matrix, and the analysis of AD pathogenesis. The 3D culture generation takes 1–2 days. The aggregation of β-amyloid is observed after 6-weeks of differentiation followed by robust tau pathology after 10–14 weeks.
Alzheimer’s disease (AD) is the most common form of age-related dementia and is characterized by progressive memory loss and cognitive impairment. Familial, early-onset (<60 years), autosomal-dominant forms of AD (FAD) can be caused by mutations in the genes amyloid precursor protein (APP), presenilin 1 (PSEN1), and presenilin 2 (PSEN2)1. Sporadic AD typically presents at a later onset and is due to multifactorial genetic and environmental risk factors. The two pathological hallmarks of AD are extracellular amyloid plaques, mainly composed of amyloid β (Aβ) peptides derived from APP via serial cleavage by β- and γ-secretase, and intracellular neurofibrillary tangles, composed of filamentous aggregation of hyperphosphorylated tau (p-tau) proteins1. It is hypothesized that the accumulation and aggregation of Aβ causes the formation of neurofibrillary tangles2–5, but this has not been experimentally proven in any animal models. Only the models that overexpress frontotemporal dementia (FTD)-associated P301L MAPT (tau) mutant forms in addition to human APP and/or PSEN1 with FAD mutations develop both plaques and tangles, albeit in a disconnected manner2–4. Fundamental species-specific differences in genome and protein composition between humans and mice, such as the difference in number of tau isoforms, have precluded an accurate recapitulation of AD pathology.
Advances in the field of stem cell generation have further advanced the prospect of in vitro systems that model AD in human neurons, including the generation of induced pluripotent stem cells (iPSCs) from FAD patient fibroblasts5–10. However, it has been challenging to replicate the aged AD brain environment in the presence of high levels of soluble and insoluble toxic Aβ species and thereby to realize full AD pathology11–13. Recently, we reported that genetically engineered human neural stem cells overexpressing FAD genes combined with a three-dimensional (3D) culture condition induced robust AD pathogenesis, including extracellular aggregation of Aβ and accumulation of hyperphosphorylated/aggregated tau as neurofibrillary tangles14. In this article, we describe our protocol for the generation of these 3D human neural culture models of AD, detail their technical background, describe the techniques we applied for their analysis, and discuss their use and application.
We designed our AD model around two central technologies: human neural progenitor cells (hNPCs) that produce high concentrations of pathogenic Aβ species and a Matrigel-based 3D culture system that provides an environment that favors Aβ deposition.
First, we engineered a FAD cell line that could exhibit significant amyloid pathology, be easily maintained, and survive through multiple passages. We chose the immortalized hNPC cell line ReNcell VM (ReN) as a base for our platform because the cells can be maintained for more than 45 passages, are commercially available, and can differentiate into neurons and glial cells with simple growth-factor deprivation15–27. The ReN cells were then transfected with IRES-mediated polycistronic lentiviral vectors containing FAD genes encoding human APP with both K670N/M671L (Swedish) and V717I (London) mutations (APPSL), PSEN1 with ΔE9 mutation (PSEN1(ΔE9)), and APPSL/PSEN1(ΔE9) with GFP or mCherry as a reporter for viral infection (Fig. 1 and and2).2). Fluorescence-activated cell sorting (FACS) was then employed to enrich the population of cells with the highest expression levels (Fig. 2 and and33).
Second, we differentiated and maintained the FACS-sorted ReN cells expressing high levels of FAD genes in a 3D Matrigel culture system to promote extracellular deposition of Aβ (Fig. 4 and Supplementary Fig. 1). We posited that, in two-dimensional models, secreted Aβ may diffuse into the cell culture medium, disrupting aggregation, whereas the 3D Matrigel may prevent this diffusion of Aβ, allowing for high local concentrations that are sufficient to initiate aggregation. We chose Matrigel specifically as a 3D culture matrix because it can be easily solidified with ReN cells through moderate thermal change and because it provides a brain-like environment rich in structural proteins such as laminin, entactin, collagen, and heparan sulfate proteoglycans28. We confirmed that the 3D Matrigel serves as an excellent 3D seeding structure for Aβ aggregation once the aggregation is initiated although it is still permeable to Aβ species. Moreover, previous studies suggested that 3D conditions that more closely mimic in vivo environments can accelerate neuronal differentiation and neural network formation23,29–34. Indeed, we found that 3D differentiated FAD-gene ReN cells expressed higher levels of specific neural markers and elevated 4-repeat adult tau (4R tau) isoforms versus 2D14.
While the characteristics of the 3D culture system are essential for fully recapitulating AD, the same properties cause several technical difficulties in classical biochemical and imaging analyses. For immunofluorescence (IF) and fluorescence/confocal microscopy, we developed thin layer cultures (~100–300 μm) that could be imaged at high magnification. For biochemical analyses that required higher concentrations of biomolecules, we differentiated cells into thick layer cultures (~4 mm) (Fig. 2). We confirmed that both thin layer and thick layer 3D cultures robustly differentiate into neurons and glia14.
Most iPSC-derived human neurons with either APP or PSEN1 FAD mutations exhibit significant increases in Aβ42 levels as compared to the control cells, consistent with previous findings in other model systems5–10. Interestingly, increases in p-tau levels were reported in some of these models. Israel et al. showed a ~2-fold increases of p-tau/total tau ratio in AD neurons6. Recently, Muratore et al. showed that AD neurons harboring APPV717I mutation also showed increases in p-tau and total tau levels and furthermore, demonstrated that abnormal elevation in tau levels could be decreased by treating with anti-Aβ antibodies10. Another recent study determined that differentiated human neurons from Down syndrome (DS) patients generated high levels of Aβ40 and 42 due to an extra copy of the APP gene7. Interestingly, DS neuronal cultures revealed intracellular and extracellular aggregates of Aβ42 as well as increases of p-tau levels, even in 2D culture conditions after extended differentiation periods (>90 days). However, these studies could not recapitulate either robust extracellular aggregation of Aβ or the aggregated p-tau pathology that are evident in our 3D culture study. Limitations of 2D culture protocols and/or low Aβ42 levels may be responsible for the lack of robust AD pathology in the previously described studies using stem cell-derived neuronal cultures.
As this model is the first human cellular model to comprehensively show Aβ-driven tauopathy this protocol can now be used to examine many other central parts of AD pathogenesis in vitro, including the molecular mechanisms underlying the production of high concentrations of Aβ, the accumulation of extracellular Aβ, the deposition of Aβ aggregates, the hyperphosphorylation of tau, and p-tau aggregation. These paths may lead to new diagnostic and prognostic biomarkers of AD. The model can also be used to test other genetic or environmental factors associated with AD, either in conjunction with the FAD-causing mutations used in this study or in place of them. The flexible scalability of the system also makes it ideal for use in larger scale testing and drug screening. More generally, the following protocols may also be applicable to other neurodegenerative diseases with genetic variations and can be especially suitable for diseases with abnormal aggregation of misfolded proteins.
However, it is important to note that while the 3D-differentiated ReN cells with FAD mutations have reproduced key events in the pathogenic cascade of AD including Aβ deposition and p-tau accumulation/aggregation, they are not surrogate human brain systems. First, our 3D culture model lacks human microglia cells (data not shown) that play a crucial role in Aβ clearance, brain inflammation, and synaptic and neuronal damage. Our 3D differentiation protocol is also not designed to model the specific brain regions that are most affected in AD, i.e. the hippocampus and specific cortices. Indeed, we found that 3D-differentiated ReN cells are composed of a heterogeneous cell population, expressing markers for GABAergic, glutamatergic and dopaminergic neurons as well as astrocytes and oligodendrocytes14. The overexpression of AD genes with multiple FAD mutations may generally restrict the use of our model for the study of the early stages of the disease, in which Aβ gradually accumulates without robust increases in tau phosphorylation. Moreover, our model also characterizes the late stages of the disease, where hyperphosphorylated tau aggregation develops.
Regardless, we believe that the model and techniques described in this protocol offer an exciting model for the AD field, especially given that plaques and tangles formation are the two major hallmarks of this disease. Moreover, the model provides a methodology for the potential development of human neural cell cultures for the study of other neurodegenerative diseases and a robust platform for finding new AD disease targets and therapeutics.
Although the 3D culture described in here is optimized for ReN cells, we have also confirmed that the same 3D culture condition can be applied for differentiating human iPSC-derived neural stem cells including It-NES cells35,36 (data not shown).
This protocol is optimized for a commercially available human neural progenitor cell line (ReNcell VM, EMD Millipore), which is immortalized, well characterized, and genetically stable after prolonged passaging (>45 passages)15–27. ReN cells were transfected with either CSCW-APP-GFP, CSCW-PSEN1(ΔE9)-mCherry or CSCW-APP-IRES-PSEN1(ΔE9)-IRES-mCherry alone or together (Fig. 1–4). The FAD mutations APP (K670N/M671L (Swedish) and V717I (London)) and PSEN1(ΔE9) were chosen specifically because they cause high production of Aβ42 and Aβ40 and a high ratio of Aβ42/Aβ40. ReN cells with CSCW-IRES-GFP or CSCW-IRES-mCherry alone were also generated as a control. Other FAD mutations may be selected for further characterization or comparison. Other gene editing and overexpression strategies including electroporation, Zinc-finger nucleases (ZFNs), transcription activator like effector nucleases (TALENs), or clustered regulatory interspaced short palindromic repeat (CRISPR)/Cas based RNA-guided DNA endonucleases could also be utilized37–39.
We describe how to develop thin-layer (~100~300 μm) and thick-layer (~4 mm) 3D cultures by mixing different concentrations of Matrigel with cells. Thin-layer 3D cultures can be used for immunocytochemistry and thick-layer 3D cultures can be used for molecular, biochemical, and ELISA analyses (Fig. 2, ,55–9). Also differentiate cells transfected with the GFP or mCherry control lentiviral particles. The presence of Aβ aggregates and p-tau aggregates described herein may depend on the hNPC line, the passage number, the type of FAD mutations, the number of cells seeded, and the proportion of enriched cells by FACS. Generally, we observed extracellular Aβ aggregates starting at 6 weeks of differentiation and robust p-tau accumulation after 10 weeks. We also observed that both Aβ aggregates and p-tau pathology were further elevated through 17 weeks, depending on the cell lines (Fig. 7–9).
For the thin-layer 3D culture, we describe how to analyze Aβ and abnormal p-tau accumulation using immunofluorescence staining and immunohistochemistry (Fig. 7–8), as previously shown14. Since most of our FAD ReN cells express high levels of GFP and mCherry, the UV-compatible amyloid dye Amylo-Glo can be used for detecting β-amyloid aggregates in the thin layer 3D culture (Supplementary Fig. 2). For thick-layer 3D culture, we describe biochemical extraction to show accumulation of Aβ and p-tau species in both soluble and insoluble fractions. Specifically, for detecting Aβ aggregates, we recommend using TBS/2%SDS/Formic acid serial extraction, while for detecting tau aggregation, we recommend using a 1% Sarkosyl extraction method, as shown in Fig. 9. Accumulation of SDS-resistant Aβ oligomers, including dimers, trimers and tetramers, can also be used as indirect markers for Aβ aggregation. To monitor the proper 3D differentiation, analyze expression of adult neural markers, including 4-repeat tau, by using RT-PCR, qRT-PCR14 and immunofluorescence staining (as shown Fig. 6). LDH levels in 3D culture media should also be measured routinely to detect the cell death rate in these cultures.
We obtained the lentiviral polycistronic vectors (CSCW-IRES-GFP and CSCW-IRES-mCherry) from Massachusetts General Hospital (MGH) viral core42, which are available through the MGH viral core (https://vectorcore.mgh.harvard.edu). CSCW-APP-GFP, CSCW-PSEN1(ΔE9)-mCherry or CSCW-APP-IRES-PSEN1(ΔE9)-IRES-mCherry vectors (Fig. 1) were constructed by inserting full length APP with FAD mutations APP (K670N/M671L (Swedish) and V717I (London)), PSEN1 with a FAD mutation, or both APP and PSEN1 with the FAD mutations 43. We have our lentiviral vectors packaged by our viral core, the by the MGH viral core 42,43. To package lentiviral vectors, first generate lentiviral particles by co-transfecting 293T cells with plasmids, the lentivirus packaging genome CMVRΔ8.91 and envelope coding plasmid (pVSV-G). 48 to 72 h post-transfection, harvest the lentiviral particles, concentrate by ultracentrifugation and titer [transducing units (t.u.)/ml] by real-time PCR using specific primers against WPRE elements as previously described42. ▲CRITICAL The aliquots of packaged viral particles can be stored at −80 °C up to one year without a significant loss of titer.
Under a biosafety culture hood, add 2 ml of 0.2 μm-filtered 10 mM acetic acid to 2 mg lyophilized EGF (Sigma) (1 mg/ml), mix, make further dilution with 0.2 μm-filtered 0.1% bovine serum albumin solution to 20 μg/ml final concentration, and store 1 ml aliquots at −80 °C.
Under a biosafety culture hood, add 2 ml of 0.2 μm-filtered 10 mM Tris (pH 7.6) to 50 μg lyophilized bFGF (Stemgent), mix, make 0.2 ml aliquots and store them at −80 °C. ▲CRITICAL To avoid the activity loss by multiple freezing/thawing cycle, bFGF or EGF stocks should be kept at 4 °C once they are thawed. 4 °C stocks should be used within 2–3 weeks. The bFGF and EGF stock aliquots can be stored up to one year at −80 °C.
Add 2% (v/v) of B27 supplement and 2% (v/v) of KnockOut serum replacement to 50 ml of D-PBS. The cell sorting media can be stored at 4 °C for 6 weeks.
Aliquot in 1 ml stocks on ice and keep the stocks at −80 °C. ▲CRITICAL Thaw Matrigel stock overnight at 4 °C before use. Matrigel tends to solidify above 10 °C. The frozen Matrigel stock can be stored at −80 °C up to one year.
Mix 0.5 ml of Matrigel and 50 ml cold DMEM/F12 medium (1:100 dilution) on ice under a biosafety culture hood. ▲CRITICAL Take out a Matrigel stock (1 ml) from −80 °C freezer and place it in 4 °C refrigerator one day before the experiments. Prepare fresh Matrigel working solution.
Add 3 ml of a cold Matrigel working solution into each T25 flasks, shake gently to cover all the surface area, incubate under 37 °C CO2 incubator at least for 1 h, and remove the remaining solution. Matrigel-coated culture dishes can be stored at 4 °C for up to 2 months.
To prepare 500 ml of medium, combine 484.5 ml of DMEM/F12 (Gibco/Life Technologies) with 0.5 ml of Heparin (2 mg/ml stock, StemCell Technology), 10 ml of B27 (Life Technologies) and 5 ml of 100x Penicillin/Streptomycin/Amphotericin (Lonza). The cell sorting media can be stored at 4 °C for 2–3 months.
To prepare 100 ml of medium, combine 100 ml of ReN differentiation medium with 80 μL bFGF stock and 100 μL EGF stock. Filter the medium after adding all the reagents. ▲CRITICAL Medium with growth factors can be kept at 4 °C for several weeks, but use a fresh one if there is any noticeable decrease in cell growth rate.
Add 108 g Tris, 55 g Boric Acid and 7.425 g EDTA into a beaker, fill up to 1 L with distilled-deionized H2O (ddH2O) and mix to dissolve. The 10x TBE buffer is stable at least for 6 months at room temperature.
To prepare 100 ml buffer, dissolve 6.05 g Tris and 8.76 g NaCl in 80 ml of H2O. Adjust pH to 7.5 with 1 M HCl and make volume up to 100 ml with H2O. The 10x TBS buffer is stable at least for 6 months at room temperature.
To prepare 1 L buffer, add 100 ml 10x TBS buffer and 2 ml of Tween 20 and make volume up to 1 L with H2O. The 1x TBST buffer is stable at least for 3 months at room temperature.
To prepare 250 ml blocking/dilution solution, add 2.5 g bovine serum albumin (Sigma-Aldrich), 5.63g glycine, 0.25 g gelatin in 200 ml TBST, heat at 55 °C for ~10 min to dissolve gelatin, add 10 ml of donkey serum (Sigma-Aldrich) and add TBST to make the final volume to 250 ml. Filter the blocking/dilution solution with 0.4 μm filter unit (Gibco) and store the solution at 4 °C. The Blocking/dilution solution can be stored at 4 °C for 2–3 months.
Add 2 μl Amylo-Glo (100x) to 10 ml 0.9% (w/v) NaCl. Prepare fresh every time.
Add 10 μl Amylo-Glo (100x) to 10 ml 0.9% (w/v) NaCl. Prepare fresh every time.
To prepare 100 ml buffer, dissolve 6.05 g Tris and 8.76 g NaCl in 80 ml of H2O. Add 4 ml of 0.4 M EDTA (pH 7.4) and adjust pH to 7.6 with 1 M HCl and make volume up to 100 ml with H2O. The 10x Tris/EDTA buffer is stable at least for 6 months at room temperature.
To prepare 10 ml buffer, mix 8.65 ml of distilled water (HPLC grade), 1 ml of 10xTBS buffer (pH 7.4), 1 tablet of protease inhibitor cocktail, 0.1 ml of NaVO3 (1 M), 0.1 ml of NaF (1 M), 0.05 ml of PNT (200 mM), 0.05 ml of PMSF (200 mM) and 0.05 ml of phosphatase inhibitor. ▲CRITICAL Prepare the fresh working solution before use. 1 ml aliquots of the extraction buffer without phosphatase inhibitors can be stored at −80 °C for several months.
To prepare 10 ml buffer, mix 7.5 ml of distilled water (HPLC grade), 2 ml of 10xTBS buffer (pH 7.4), 2 tablets of protease inhibitor cocktail, 0.1 ml of NaVO3 (1 M), 0.1 ml of NaF (1 M), 0.1 ml of PNT (200 mM), 0.1 ml of PMSF (200 mM) and 0.1 ml of phosphatase inhibitor. ▲CRITICAL Prepare the fresh working solution before use. 1 ml aliquots of the extraction buffer without phosphatase inhibitors can be stored at −80 °C for several months.
To prepare 10 ml buffer, mix 1 ml of SDS solution (20% stock), 1 ml of 1% Triton X-100 (10% stock), 4.5 ml of distilled water (HPLC grade), 1 ml of 10xTBS/EDTA buffer (pH 7.4), 2 tablets of protease inhibitor cocktail, 0.1 ml of NaVO3 (1 M), 0.1 ml of NaF (1 M), 0.1 ml of PNT (200 mM), 0.1 ml of PMSF (200 mM) and 0.1 ml of phosphatase inhibitor (100x). ▲CRITICAL Prepare the fresh working solution before use. 1 ml aliquots of the extraction buffer without phosphatase inhibitors can be stored at −80 °C for several months.
To prepare 10 ml buffer, mix 2 ml of SDS solution (20% stock), 2 ml of 1% Triton X-100 (10% stock), 4.5 ml of distilled water (HPLC grade), 1 ml of 10xTBS/EDTA buffer (pH 7.4), 2 tablets of protease inhibitor cocktail, 0.1 ml of NaVO3 (1 M), 0.1 ml of NaF (1 M), 0.1 ml of PNT (200 mM), 0.1 ml of PMSF (200 mM) and 0.1 ml of phosphatase inhibitor (100x). ▲CRITICAL Prepare the fresh working solution before use. 1 ml aliquots of the extraction buffer without phosphatase inhibitors can be stored at −80 °C for several months.
Add 0.1 g N-laurylsarcosine with total 0.5 ml of distilled water (HPLC grade). The 20 % Sarkosyl solution should be made fresh before the experiments.
To prepare 10 ml buffer, mix 0.5 ml of sodium deoxylcholate solution (10% stock), 0.2 ml of 2% NP-40 (ICABAL substitute), 7.8 ml of distilled water (HPLC grade), 1 ml of 10xTBS/EDTA buffer (pH 7.4), 02 tablets of protease inhibitor cocktail, 0.1 ml of NaVO3 (1 M), 0.1 ml of NaF (1 M), 0.1 ml of PNT (200 mM), 0.1 ml of PMSF (200 mM) and 0.1 ml of phosphatase inhibitor (100x). ▲CRITICAL Prepare fresh working solution before use. 1 ml aliquots of the extraction buffer without phosphatase inhibitors can be stored at −20 °C for several months.
To prepare 10 ml buffer, mix 1 ml of 10xTBS, 0.47g NaCl, 0.01 ml EGTA (1 M), 1 g sucrose, 1 tablet of protease inhibitor cocktail, 0.1 ml of NaVO3 (1 M), 0.1 ml of NaF (1 M), 0.1 ml of PNT (200 mM), 0.1 ml of PMSF (200 mM) and 0.1 ml of phosphatase inhibitor (100x). Make 10 ml with distilled water (HPLC grade). ▲CRITICAL Prepare fresh working solution before use. 1 ml aliquots of the extraction buffer without phosphatase inhibitors can be stored at −20 °C for several months.
Using a CytoTox-ONE Assay Kit, add 11 ml of Assay Buffer to each vial of Substrate Mix to prepare CytoTox-ONE Reagent. The CytoTox-One reagent is stable for 6–8 weeks at −20 °C.
Sterilize the forceps under the biosafety hood using ethanol and UV radiations. Place 24-well plates (Falcon), open the lids and carefully place cell culture inserts into each well with the sterilized forceps. ▲CRITICAL Always use the companion plates (Falcon) for setting tissue culture inserts (Falcon).
Place the coverglass slide plates (Nunc) or the glass-bottomed 35 mm dishes (MatTek) inside 150 ml culture dishes to reduce the chance of contamination.
RNA work should be done in a dedicated area or bench. Keep a separate set of equipment only used in this area. Wipe the work surfaces with RNaseZap® or other RNase removal reagents before RNA extraction. Solutions and reagents should be stored in small aliquots to minimize contaminations with RNases. Gloves should be worn when handling RNA and should be changed frequently. The temperature in the benchtop centrifuge should be set-up to 4 °C or placed in a fridge.
▲CRITICAL Extraction from one thick layer culture insert usually yields enough RNA for quantitative RT-PCR analyses.
▲CRITICAL While extraction from one thick layer culture insert usually yields enough RNA for quantitative RT-PCR analyses. If needed, pool together 2–3 wells from thin layer cultures (96-well format).
|3R4R tau forward primer (10 μM)||1 μL||500 nM|
|3R4R tau reverse primer (10 μM)||1 μL||500 nM|
|SYBR® Select Master Mix (2X)||10 μL||1X|
|1||95 °C, 15 min|
|2–31||94 °C, 30 s||60 °C, 30s||74 °C, 90 s|
|32||74 °C, 10 min|
|Neural marker primers F+R (5 μM)||0.4 μL||0.1 μM|
|SYBR® Select Master Mix (2x)||10 μL||1X|
|1||95 °C, 10 min|
|2–62||95 °C, 10 s||58 °C, 45s||72 °C, 30 s|
|Melting curve||95 °C, 60s||55 °C, 60s followed by 55 + 0.5 °C, 10s each|
|Tau (3R or 4R) forward primer (10 μM)||1 μL||500 nM|
|Tau (3R or 4R) reverse primer (10 μM)||1 μL||500 nM|
|SYBR® Select Master Mix (2x)||10 μL||1X|
|1||95 °C, 10 min|
|2–62||95 °C, 10 s||58 °C, 45s||72 °C, 30 s|
|Melting curve||95 °C, 60s||55 °C, 60s followed by 55 + 0.5 °C, 10s each|
▲CRITICAL Dot blot analysis can be a useful for quickly testing p-tau levels before EM analysis.
▲CRITICAL Users of vacuum-based systems will need to modify these steps.
▲CRITICAL In the control ReN cell 3D cultures, we found that levels of LDH release are below 3% of total LDH levels.
For reproducible output, we found that it is very important to plate a high number of cells in order to achieve early Aβ aggregation. In addition, the passage number and the condition of the cells are also important. Especially for the highly enriched cell lines expressing APPSL and PSEN1ΔE9, we strongly recommend using only cells with very early passage numbers (up to 4) to ensure that a robust Aβ and tau pathology is consistently observed. FAD ReN cell lines with high Aβ42/40 ratio will show denser amyloid deposits in 3D culture (e.g. ReN-mGAP cells that express both GFP:APPSL and mCherry:PSEN1ΔE9 as shown in Fig. 5d and and7;7; ReN-mAP cells that express mCherry:APPSL:PSEN1ΔE9 in Fig. 8; HReN-mGAP (enriched) cells that express both GFP and mCherry:APPSL:PSEN1ΔE9 in Fig. 9 and supplementary fig. 2) while the cell lines with mid/low Aβ42/40 ratio will develop mostly diffuse or no obvious deposits despite these cells having secreted high levels of total Aβ (e.g. ReN-GA2 cells in Fig. 5b–d). In general, we observed that the ReN cells secreting high levels of Aβ, preferentially Aβ42, develop more robust p-tau pathology.
To enhance Aβ and p-tau pathologies, we used FACS sorting to enrich ReN cells expressing high levels of transgene. As shown in Fig. 3, FACS sorting of the top 2, 10 or 50% levels of GFP signals (Fig. 3a) lead to the generation of ReN-GA cells (APPSL/GFP) expressing high, mid and low APPSL/GFP (ReN-GA2, -GA10 and -GA50; Fig. 3b). The control ReN-G cells with comparable levels of GFP expression were also prepared (ReN-G2, -G10 and -G50; Fig. 3a–b). As expected, ReN-GA2 and the control ReN-G2 cells differentiated comparably in the thin-layer 3D culture condition into neurons and glial cells (Fig. 6). Western blot analysis (Fig. 5a) and Aβ ELISA (Fig. 5b–d) showed robustly elevated Aβ levels in ReN-GA2 cells as compared to the control ReN-G2 cells. Aβ ELISA also detected an elevated level of Aβ in conditioned medium collected from 1-week differentiated thick-layer 3D cultures of ReN-GA2 cells (Fig. 5c). However, the Aβ42/40 ratio was moderately increased in these cells because they do not express PSEN1ΔE9, which alters γ-secretase structures to produce more Aβ42. As a result, we observed that ReN-mGAP (enriched) cell lines displayed much more robust Aβ deposits as compared to ReN-GA (enriched) cells (Fig. 5d). Similar results were also observed in ReN-mAP (enriched) cell lines (data not shown).
The quality of the ReN cell 3D cultures can be monitored by LDH releases assay, qRT-PCR/RT-PCR analyses of adult neural marker expression and immunofluorescence staining (as shown in Fig. 6). Based on immunofluorescence staining, we estimate that ~68% ReN cells differentiate into MAP2-positive neurons and ~30% into GFAP-positive astrocytes following 2–4 weeks of 3D differentiation. We also found a robust increase in adult neuronal markers, particularly 4-repeat adult tau isoforms in 3D culture conditions, which are essential for replicating tauopathy in human neuronal cultures14.
With the proper cell density and passage number, extracellular β-amyloid deposits start developing after 6 weeks in FAD ReN cell 3D cultures with high levels of Aβ42 (Fig. 7) and the number and size of β-amyloid aggregations increases with longer culture. Amylo-Glo45, a fluorescent amyloid dye can also be used to detect Aβ deposits, particularly after 8-weeks differentiation (Supplementary Fig. 2). Accumulation of p-tau in cell bodies and neurites can be detected by immunostaining with p-tau antibodies, including AT-8 and PHF-1, after 8 weeks (Fig. 8) while robust accumulation of p-tau aggregates are detected after 10–14 weeks14. PHF-like tau filaments can be detected after 14 weeks in Sarkosyl-insoluble fractions of FAD ReN cells with highly robust Aβ levels (preferentially Aβ42) by electron microscopy.
This work is supported by the grants from the Cure Alzheimer’s fund to D. Y. K. and R. E. T. and National Institute of Health grants 1RF1AG048080-01 (D.Y.K. and R.E.T.), 5P01AG15379 (D.Y.K. and R.E.T.), 2R01AG014713 (D.Y.K.) and 5R37MH060009 (R.E.T.). We appreciate Drs. Bradley T. Hyman, Oksana Berezovska, Dora M. Kovacs (Massachusetts General Hospital, Boston, USA), John Hardy (NIH, Bethesda, MD, USA) and Peter Davies (Albert Einstein College of Medicine, Bronx, USA) for providing cDNAs and antibodies. We also would like to acknowledge Dr. Bakhos A. Tannos at MGH Viral Vector Core (supported by NIH/NINDS P30NS04776), Mr. Michael Waring at Ragon Institute’s Imaging Core facility (part of the Harvard CFAR Immunology Core), Ms. Mary L. McKee and Ms. Diane Capen at MGH Microscopy Core of the Center for Systems Biology for providing the detailed protocols and revised the manuscript.
DECLARATION OF INTEREST
The authors report no competing financial interest.
AUTHOR CONTRIBUTIONSD.Y.K. and R.E.T. were equally responsible for experimental design and supervising the whole project. Y.H.K., S.H.C., C.D. and D.Y.K., mainly contributed to writing and revising the manuscript. C.D., E.B., K.J.W., M.H. and D.Y.K. performed the experiments. M.H., O.B., J.B.K. and C.S. contribute to writing the manuscript.