γ-Secretase modulators reduce endogenous amyloid β₄₂ levels in human neural progenitor cells without altering neuronal differentiation

Abstract

Soluble γ-secretase modulators (SGSMs) selectively decrease toxic amyloid β (Aβ) peptides (Aβ₄₂). However, their effect on the physiologic functions of γ-secretase has not been tested in human model systems. γ-Secretase regulates fate determination of neural progenitor cells. Thus, we studied the impact of SGSMs on the neuronal differentiation of ReNcell VM (ReN) human neural progenitor cells (hNPCs). Quantitative PCR analysis showed that treatment of neurosphere-like ReN cell aggregate cultures with γ-secretase inhibitors (GSIs), but not SGSMs, induced a 2- to 4-fold increase in the expression of the neuronal markers Tuj1 and doublecortin. GSI treatment also induced neuronal marker protein expression, as shown by Western blot analysis. In the same conditions, SGSM treatment selectively reduced endogenous Aβ₄₂ levels by ∼80%. Mechanistically, we found that Notch target gene expressions were selectively inhibited by a GSI, not by SGSM treatment. We can assert, for the first time, that SGSMs do not affect the neuronal differentiation of hNPCs while selectively decreasing endogenous Aβ₄₂ levels in the same conditions. Our results suggest that our hNPC differentiation system can serve as a useful model to test the impact of GSIs and SGSMs on both endogenous Aβ levels and γ-secretase physiologic functions including endogenous Notch signaling.—D’Avanzo, C., Sliwinski, C., Wagner, S. L., Tanzi, R. E., Kim, D. Y., Kovacs, D. M. γ-Secretase modulators reduce endogenous amyloid β₄₂ levels in human neural progenitor cells without altering neuronal differentiation.

Alzheimer’s disease (AD) is the most common late-onset neurodegenerative disorder and is officially listed as the sixth leading cause of death in the United States (1–3). AD results in the irreversible loss of neurons, particularly in the associative neocortex and hippocampus, important for cognition (1–3). Excess accumulation of toxic amyloid β (Aβ) species, such as the 42-amino acid isoform (Aβ₄₂), is known as a major pathologic event triggering neurodegeneration in AD (4–6). Aβ is produced by sequential cleavages of the amyloid precursor protein (APP), mediated by the secretases β-site APP cleaving enzyme 1 and presenilin (PS)/γ-secretase (7). Therefore, these secretases have been major therapeutic targets for treating AD.

PS/γ-secretase is an enzyme complex composed of the catalytic core PS1 or PS2, nicastrin, anterior pharynx defective 1 homolog A or B, and presenilin enhancer 2 (7, 8). PS/γ-secretase cleaves membrane-tethered C-terminal fragments of APP (APP-CTFs) to generate Aβ isoforms of various lengths, including the pathogenic Aβ₄₂. In addition to APPs, PS/γ-secretase regulates various cellular functions by cleaving at least 90 more cellular substrate proteins (9). Stem cell-based studies, in particular, have shown that PS/γ-secretase regulates the maintenance and the differentiation of embryonic and adult neuron precursor cells (NPCs) in mice (10–12). Pharmacologic inhibition of PS/γ-secretase also modulates the neural differentiation of human neural progenitor cells (hNPCs) derived from human embryonic stem cells (ESCs) and induced pluripotent stem cells, possibly by blocking the cleavage of the differentiation-regulating PS/γ-secretase substrates including Notch1 (13–15). Therefore, neural differentiation of hNPCs can be a good system for monitoring physiologic consequences of altered PS/γ-secretase activity in human cells.

PS/γ-secretase inhibitors (GSIs) block the generation of Aβ and therefore were expected to prevent the progression of AD. However, human clinical trials have been discontinued on account of severe side effects, possibly due to the impact of GSIs on the cleavage of essential γ-secretase substrates other than APPs (7, 16–19). As opposed to GSIs, PS/γ-secretase modulators (GSMs) selectively decrease pathogenic Aβ₄₂ generation without affecting other essential PS/γ-secretase substrate cleavages. Soluble γ-secretase modulators (SGSMs) are newly developed, highly soluble GSMs that selectively lower Aβ₄₂ levels without inhibiting proteolysis of other PS/γ-secretase substrates or causing accumulation of APP-CTFs (20). Therefore, SGSMs are not expected to have a major effect on the overall physiologic functions of PS/γ-secretase. However, this has not been rigorously tested in human neuronal systems.

In the present study, we used ReNcell VM (ReN) hNPCs (21, 22) to investigate the impact of SGSMs and GSIs on neuronal differentiation. ReN cell hNPCs are immortalized hNPCs derived from human fetal brains (22). Previous studies showed that these cells differentiate into mature neurons, astrocytes, and oligodendrocytes after 1–3 weeks under growth factor-deprived conditions (22–24). Our results clearly demonstrate that SGSMs, at a concentration that abolishes toxic Aβ generation, do not significantly affect hNPC differentiation into neurons.

MATERIALS AND METHODS

Cell lines, media, and reagents

ReN human neural precursor cells (EMD Millipore, Billerica, MA, USA) were plated onto BD Matrigel (catalog #356234; BD Biosciences, San Jose, CA, USA)-coated T25 cell culture flasks (BD Biosciences) and maintained in ReN cell maintenance medium containing DMEM/F12 (Life Technologies, Grand Island, NY, USA) supplemented with 2% (v/v) B27 neural supplement (Life Technologies), 2 μg/ml heparin (Stemcell Technologies, Vancouver, BC, Canada), 20 μg/ml epidermal growth factor (EGF; Sigma-Aldrich, St. Louis, MO, USA), 20 μg/ml basic fibroblast growth factor (bFGF; Stemgent, Cambridge, MA, USA), and 1% (v/v) penicillin-streptomycin-amphotericin B solution (Lonza, Hopkinton, MA, USA) under CO₂ cell culture incubator. The cell culture media were changed every 3 days until the cells were confluent. For some of the studies, we used ReN cell stably expressing green fluorescent protein (GFP) (ReN-G), which was described previously (25). Preaggregation neurosphere differentiation (PreD) was performed as described previously with some modifications (22). Confluent single-layer ReN cells were washed with PBS, dissociated into single cells by Accutase (Life Technologies) treatment, plated onto noncoated 96- or 24-well plates (83,000 cells per well), and incubated with ReN cell maintenance medium. Neurosphere-like aggregates were observed 1 day after plating, and GSIs or SGSMs were directly added to the medium. The same volumes of media were added after 3 days. For standard monolayer differentiation, the dissociated ReN cells were plated onto BD Matrigel-coated 6- or 12-well plates with the ReN cell differentiation medium containing DMEM/F12 supplemented with 2% (v/v) B27 neural supplement, 2 μg/ml heparin, 20 μg/ml EGF, 20 μg/ml bFGF, and 1% (v/v) penicillin-streptomycin-amphotericin B solution. Cells were treated with GSIs or SGSMs after 1 day. GSIs, (3,5-difluorophenylacetyl)-l-alanyl-l-2-phenylglycine t-butyl ester (DAPT), and compound E (CpdE) were purchased from EMD Millipore. SGSMs 36, 41, 46, and 49 were described previously (20, 25).

Quantitative RT-PCR analysis

Total RNAs were purified by using RNeasy Mini Kits (Qiagen, Germantown, MD, USA), and cDNAs were synthesized by the SuperScript III First-Strand Synthesis Kit (Life Technologies). The prevalidated primer sets for human neural markers [Tuj1, doublecortin (DCX), neural cell adhesion molecule 1 (NCAM1), and MAP2], Notch1 target genes [CDKN1A (human cyclin-dependent kinase inhibitor 1A), HEY-1 (human hairy/enhancer-of-split related with YRPW motif 1), and HES-1 (human hairy and enhancer of split 1)], and the control β-actin were purchased from Real Time Primers, Limited Liability Company (Elkins Park, PA, USA). The amplification was done in a final volume of 20 μl under the following conditions: 15 minutes at 95°C and then 55 cycles at 95°C for 10 seconds, 58°C for 45 seconds, and 72°C for 45 seconds. The sizes of qPCR products were confirmed by agarose gel electrophoresis. The Bio-Rad iCycler (Hercules, CA, USA) was used to determine cycle threshold (Ct) values for each sample. 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 DMSO control as 1. The following are the neuronal gene names and PCR product sizes: NCAM1 (174 bp), Tuj1 (human β-tubulin type III; 107 bp), DCX (human DCX; 245 bp), MAP2 (microtubule-associated protein 2; 266 bp), β-actin (233 bp), CDKN1A (220 bp), HES-1 (182 bp), and HEY-1 (197 bp).

Western blot analysis

Protein samples (15–75 μg) were analyzed on 4–12% gradient Bis/Tris gels (Life Technologies) and electrotransferred onto PVDF membranes (Bio-Rad). Membranes were blocked with 5% nonfat dry milk in 0.1% Tween 20 [Tris-buffered saline and Tween 20 (TBS-T)], 2 mM Tris-HCl, and 50 mM NaCl (pH 7.5) for 1 hour at room temperature (RT) and subsequently incubated overnight at 4°C in blocking buffer with the antibody for Tuj1, β-actin, NCAM1, or heat shock protein (HSP)-70.

The membranes were washed with 0.1% Tween 20 and incubated with the secondary antibodies (1:1000; Amersham, Pittsburgh, PA, USA) for 1 hour. Immunoreactive bands were detected with the ECL (Amersham). The optical density of the bands (normalized with those of actin) was captured with the VersaDoc imaging system (Bio-Rad). The images were quantified by using Quantity One software (Bio-Rad). Primary antibodies were used at the following dilutions: 1:5000, anti-Tuj1 (Abcam, Cambridge, MA, USA); 1:1000, anti-β-actin (Cell Signaling Technology, Danvers, MA, USA); 1:1000, anti-HSP-70 (Enzo Life Sciences, Farmingdale, NY, USA); and 1:1000, anti-NCAM (Cell Signaling Technology).

Immunofluorescence staining

ReN was fixed with 4% paraformaldehyde at RT for 24 hours. The fixed cells were then permeabilized and blocked by incubating with a solution containing 50 mM Tris (pH 7.4), 0.1% Tween 20, 4% donkey serum, 1% bovine serum albumin, and 0.3 M glycine at 4°C for 12 hours. ReN cells were then incubated with primary antibodies in the blocking solution at 4°C overnight. After washing 3 times with TBS-T, the cells were incubated in TBS-T overnight with gentle rocking at 4°C and incubated with Alexa Fluor secondary antibodies (Life Technologies) the following day overnight at 4°C. To avoid fluorescence quenching, a drop of Antifade Gold (Life Technologies) was added on top of the fixed/stained cells before imaging. The fluorescence images were captured by an Olympus DSU confocal microscope (Olympus USA, Center Valley, PA, USA), and image analyses were performed with MetaMorph software (Molecular Devices, Sunnyvale, CA, USA). Antibodies were used at the following dilutions: 1:200, anti-Tuj1; 1:200, anti-DCX antibody (Cell Signaling Technology); and 1:200, Alexa Fluor 350/488 anti-mouse and rabbit secondary antibodies (Life Technologies).

Aβ ELISA

Aβ₄₀ and Aβ₄₂ levels were measured by the Human β Amyloid ELISA Kit (Wako Pure Chemicals, Osaka, Japan). The conditioned media from neurosphere-like ReN cell aggregates or differentiated ReN cells were collected and diluted by 1:10 or 1:100 with a dilution buffer provided by the company. The Synergy 2 ELISA plate reader (BioTek Instruments, Winooski, VT, USA) was used to quantify Aβ₄₀ and Aβ₄₂ ELISA signals.

Statistics

All statistical analyses were performed using 1-way ANOVA followed by a post hoc Dunnett test or post hoc Newman-Keuls test. Error bars showed in graphs in the figures denote the SEM.

RESULTS

GSIs, but not SGSMs, dramatically increased the attachment of neurosphere-like ReN cell aggregates

Previous studies showed that preaggregation of ReN cells into neurosphere-like aggregates (PreD protocol) further promotes neuronal differentiation (22–24). To easily monitor the proliferation and differentiation of the neurosphere-like aggregates, we used ReN-G, as described previously (25). As reported, 24-hour incubation of dissociated ReN-G cells in uncoated plates promoted the aggregation into spheres with sizes of 100–800 μm in diameter (Fig. 1A, B). The aggregated ReN-G cells did not attach to the bottom of the culture plates because these plates had not been precoated with Matrigel, a matrix protein mixture normally used to promote the attachment of ReN cells into the plates (25). Preaggregated ReN-G cells were then treated with GSIs, SGSMs, or control DMSO for 7 days at a concentration that was 5-fold higher than the half-maximal inhibitory concentration (IC₅₀; Aβ₄₂) of each drug (Fig. 1A and Table 1). It is surprising that we found that GSIs, including CpdE (Fig. 1B, second column of the panels) and DAPT (data not shown), largely increased the attachment of the ReN-G aggregates to the plastic bottom of the noncoated plates. In addition, ReN-G aggregates developed a robust network of neurite-like projections at the periphery of the neurospheres, even in the presence of the growth factors in cell culture medium, which normally block the neuronal differentiation of the ReN cells (Fig. 1B). However, the cells treated with SGSMs or DMSO did not show any sign of increased aggregate attachment or neurite-like projections (Fig. 1B). Quantitation of the relative number of attached ReN-G neurospheres is shown in Fig. 1C.

Figure 1.

GSIs, but not SGSMs, induce differentiation of neurosphere-like ReN cell aggregates. A) Schematic representation of GSI and SGSM treatments in neurosphere-like ReN cell aggregates (PreD protocol). B) Confocal immunofluorescence (first row) and phase-contrast (second and third rows) images of representative neurosphere-like ReN-G. After allowing the formation of neurosphere-like aggregates, cells were treated for 1 week with either SGSMs or the GSI CpdE in the presence of growth factors. Scale bars, 60 µm. C) Quantitative analysis of spheres showing projections. Pictures were quantified in a blinded manner (mean ± sem). ***P < 0.0001, ANOVA followed by a post hoc Dunnett test (SGSMs, n = 4; DMSO and CpdE, n = 6).

TABLE 1.

Final drug concentrations used in this study

Compound name	Treatment concentration (nM)
SGSM36	500
SGSM41	575
SGSM46	600
SGSM49	150
CpdE	1.85
DAPT	1000

The concentrations of SGSMs and GSIs were chosen based on the Aβ₄₂ IC₅₀ of each drug in previous in vitro studies (20, 25). Cell treatment concentrations were designed to be 5-fold the Aβ₄₂ IC₅₀ of each compound.

GSIs, but not SGSMs, induced neuronal differentiation of ReN cell aggregates

To test whether the increased attachment/projections of ReN-G aggregates were a result of an accelerated neuronal differentiation, we analyzed the expression levels of early neuronal markers by quantitative RT-PCR (qRT-PCR) (Fig. 2A). As Fig. 1B, C indicates, GSI-treated cells showed a dramatic increase in early neuronal marker expression including DCX and Tuj1 (β-tubulin 3) compared with DMSO controls (Fig. 2A). However, no significant changes were detected in SGSM- or DMSO-treated samples (Fig. 2A). To further characterize neuronal differentiation, DCX expression following treatment with DMSO, CpdE, or SGSM49 was additionally analyzed by immunofluorescence staining (IF) in fixed ReN-G aggregates (Fig. 2B and Supplemental Fig. 1A, B). As predicted, we detected a dramatic increase of DCX-positive neurite-like projections in CpdE-treated samples, but not in DMSO- or SGSM49-treated samples (Fig. 2B and Supplemental Fig. 1A). Interestingly, we found that DCX-stained projections were also increased inside the neurosphere-like aggregates treated with CpdE (Supplemental Fig. 1B). Increased expression of the neuronal marker NCAM1 was also detected by Western blot analysis (Supplemental Fig. 1C).

Figure 2.

qRT-PCR and immunofluorescence analyses of neuronal marker expression in GSI- or SGSM-treated cells. A) qRT-PCR analysis of neural markers (DCX and Tuj1) of neurosphere-like ReN cell aggregates after 1 week of SGSM/GSI treatment. Gene expression levels were normalized against β-actin levels in each sample. The Ct value for each gene was determined by iCycler software (Bio-Rad), and fold changes were calculated using the ΔΔCt method. Expression levels of DCX and Tuj1 were set to 1 in DMSO-treated control (mean ± sem). ***P < 0.0001 for DCX and **P < 0.001 for Tuj1, ANOVA followed by a post hoc Newman-Keuls test (n = 3). B) Confocal IF of the neural marker DCX (red) in 1-week-treated (SGSM49 and CpdE) neurosphere-like aggregates. Scale bar, 50 µm.

To further explore the underlying mechanism, we investigated alterations in Notch signaling, which has been previously proposed to regulate differentiation of hNPCs (26). Previous studies have shown that SGSMs do not affect cleavage of a recombinant Notch1 CTF in cell culture models (20). Here, we analyzed the downstream targets of the endogenous Notch signaling, including CDKN1A, HES-1, and HEY-1, in the neurosphere-like ReN cell aggregates. qRT-PCR analysis showed that treatments of ReN cells with GSIs, but not SGSMs, dramatically decreased CDKN1A, HEY-1, and HES-1 levels as compared to DMSO controls (Supplemental Fig. 2). However, no significant changes were detected in SGSMs or DMSO-treated samples (Supplemental Fig. 2). These data clearly demonstrate that treatment with SGSMs does not affect endogenous Notch signaling in ReN cell hNPCs.

GSI, but not SGSM, treatment promoted neuronal differentiation induced by growth factor deprivation

Next, we tested the GSI/SGSM effect on neuronal differentiation in standard 2-dimensional monolayer cultures of ReN-G cells in the presence of growth factors, but we did not find any significant changes of neuronal marker expressions in this condition (data not shown), suggesting that preaggregated neurosphere-like cultures of ReN cells can serve as a unique model system to test the impact of altered γ-secretase activity in hNPCs. Thus, we asked whether GSI/SGSM treatment could further promote ReN cell differentiation induced by growth factor deprivation, a standard differentiation protocol (22, 24). ReN cells were differentiated by growth factor deprivation for 5–7 days in the presence or absence of GSI/SGSMs and analyzed by Western blot and IF (Fig. 3). As previously reported, we confirmed that neuronal/glial marker levels were increased in ReN cells under growth factor deprivation. Similar to Figs. 1 and 2, Western blot analysis showed that Tuj1 neuronal marker levels were largely increased by CpdE (GSI) treatment but were unaffected by SGSMs (Fig. 3B). We found similar changes in the early-stage neuronal marker DCX (data not shown). Quantitative analysis showed that Tuj1 levels were increased ∼2-fold in the CpdE-treated samples as compared to DMSO controls (significance levels are indicated in the histograms in Fig. 3C). Immunofluorescence analysis also confirmed that CpdE, but not SGSM46 or SGSM49 treatments, up-regulated Tuj1 levels (Fig. 3D). Interestingly, we also found that CpdE treatment led to an abnormal cytosolic accumulation of Tuj1 (Fig. 3D, upper-right panel). Together, we have confirmed that GSI treatment increases expression of neuronal markers, as previously reported in human ESC-derived hNPCs (13–15). However, our data clearly indicate that SGSM treatment does not induce neuronal differentiation or promote ReN cell differentiation induced by the growth factor deprivation.

Figure 3.

SGSMs do not induce further differentiation of growth factor-deprived ReN cells. A) Schematic representation of GSI and SGSM treatments of ReN cells under growth factor-deprivation conditions. B) Western blot analysis shows Tuj1 expression after 5-day treatment with CpdE (GSI) and SGSMs in ReN cells without growth factors. DMSO represents the negative control for the GSM/GSI treatments and β-actin was used as a protein loading control. C) Histograms show relative Tuj1 levels after GSI and SGSMs treatment (n = 4). Student’s t test, *P < 0.05. D) Immunofluorescence analysis of ReN cells after 7-day differentiation without growth factors, in the presence of CpdE (GSI) and GSMs, stained with an antibody against Tuj1. DAPI labels nuclei in blue.

SGSMs selectively decrease endogenous Aβ₄₂ levels in ReN cell hNPCs

SGSMs were developed to potently inhibit the production of Aβ₄₂ and to a lesser degree Aβ₄₀, whereas there is increased generation of shorter Aβ peptide species, such as Aβ₃₈ and Aβ₃₇ (20). We have shown that SGSMs do not significantly alter the neuronal differentiation of ReN cell hNPCs at concentrations that were previously reported to block Aβ₄₂ generation (Figs. 1, 2, and 3 and Table 1) (20–25, 27). Thus, we asked whether SGSMs blocked endogenous toxic Aβ species in ReN cell hNPCs under the conditions used for this study. We collected cell culture media from the experiments shown in Figs. 1 and 2 and analyzed soluble Aβ₄₀ and Aβ₄₂ levels by ELISA (Fig. 4). Consistent with previous reports, SGSM41, SGSM46, and SGSM49 dramatically decreased Aβ₄₂ levels similarly to the CpdE-treated samples (significance levels are indicated in the histograms in Fig. 4A). As expected, SGSMs did not decrease Aβ₄₀ to the same extent as Aβ₄₂ (Fig. 4). These data clearly demonstrate that SGSMs significantly blocked pathogenic Aβ₄₂ levels without affecting neuronal differentiation patterns in ReN cell hNPCs.

Figure 4.

In contrast to GSIs, SGSMs selectively decrease pathogenic Aβ₄₂ levels in neurosphere-like ReN cell aggregates. A) Secreted Aβ₄₂ levels as measured by a sandwich ELISA (n = 3). ***P < 0.0001, ANOVA followed by a post hoc Dunnett test. B) The secreted Aβ₄₀ levels as measured by a sandwich ELISA (n = 3). ***P < 0.0001, ANOVA followed by a post hoc Dunnett test; **P < 0.001 for SGSM49.

DISCUSSION

Human clinical trials of GSIs have not been successful because of unexpected side effects and failure to improve cognitive function in patients with AD (16, 18, 19, 28). A recent phase III clinical trial of semagacestat (LY450139) was discontinued because of various side effects, including gastrointestinal symptoms, infection, skin cancer, weight loss, and even worsening of cognitive function (16, 17). Recently, GSMs have drawn attention for their potential safety as compared to GSIs, but they have never been rigorously tested in human neural systems (20, 27). Here, we showed for the first time that when used in clinically relevant concentrations that can block toxic Aβ species in a human cell culture system, SGSMs do not affect neuronal differentiation of hNPCs. Meanwhile, we have confirmed that GSIs induced neuronal differentiation of ReN hNPCs even in the presence of growth factors, including EGF and bFGF (Figs. 1 and 2 and Supplemental Figs. 1 and 2) (13–15). Mechanistically, we found that Notch target gene expressions were selectively inhibited by a GSI, not by SGSMs treatment, suggesting that SGSM treatment does not affect the endogenous human Notch signaling cascade. Together, our data clearly suggest that our neurosphere-like cultures of ReN hNPCs can serve as a unique model system to test the impact of altered PS/γ-secretase function, including endogenous Notch signaling.

PS/γ-secretase activity regulates NPC differentiation in adult brains (adult neurogenesis) as well as in early neuronal development (10, 29). In addition to Notch signaling, EGF receptor and β-catenin signalings were proposed to regulate NPC differentiation in adult brains (10, 30). In Supplemental Fig. 2, we showed that GSI, not SGSMs, dramatically decreased the expression levels of Notch target genes in ReN hNPCs, resulting in the altering of their neuronal differentiation. It is not clear whether GSI treatment directly affects EGF receptor and β-catenin signaling, but a recent study showed that the EGF receptor is cleaved by PS/γ-secretase (29). Previous studies also showed that PS1 interacts with β-catenin, forming a complex (31–33). Therefore, it is possible that inhibition of PS/γ-secretase activity by GSIs may interfere with adult neurogenesis through multiple mechanisms in mice (10, 33) and possibly in humans. This additional activity may have contributed to the side effects observed in the previous human clinical trials with GSIs. It will be interesting to test the impact of SGSMs as well as GSIs on adult neurogenesis in future studies.

GSI treatments induced neuronal differentiation of neurosphere-like ReN cell aggregates, even in the presence of growth factors including EGF and bFGF (Figs. 1 and 2 and Supplemental Figs. 1 and 2). We also tested whether GSI treatments accelerated neuronal differentiation in a standard monolayer culture condition with those growth factors. Interestingly, we did not find any significant changes in neuronal marker expression nor increased neurite-like projections in the same ReN cells (data not shown). We did observe that GSI treatment only accelerated neuronal differentiation in monolayer culture conditions when paired with growth factor deprivation, the condition that strongly induces neuronal differentiation (Fig. 3). A 3-dimensional neurosphere-like environment or limited access to growth factors in the packed/aggregated culture may explain the differential impact of GSIs on neuronal differentiation. An advantage of the ReN aggregate cell culture system is that changes of the ReN cell aggregates can be easily monitored in a high-throughput format. As shown in Fig. 1, attachment of ReN cell aggregates and the increased projections can be monitored by a conventional phase-contrast or fluorescence microscope. Indeed, we have already tested the impact of 5 additional SGSMs on neuronal differentiations using the ReN aggregate cell culture system in a 96-well format (data not shown). In those conditions, we were also able to detect endogenous Aβ levels in the same wells. We propose to use our hNPC-based model systems to prescreen novel GSMs for their efficacy on Aβ generation and for potential safety issues before clinical trials.

Glossary

Aβ

amyloid β

AD

Alzheimer’s disease

APP

amyloid precursor protein

APP-CTF

C-terminal fragment of amyloid precursor protein

bFGF

basic fibroblast growth factor

CDKN1A

human cyclin-dependent kinase inhibitor 1A

CpdE

compound E

Ct

cycle threshold

CTF

C-terminal fragment

DAPT

(3,5-difluorophenylacetyl)- l-alanyl-l-2-phenylglycine t-butyl ester

DCX

doublecortin

EGF

epidermal growth factor

ESC

embryonic stem cell

GFP

green fluorescent protein

GSI

γ-secretase inhibitor

HES-1

human hairy and enhancer of split 1

HEY-1

human hairy/enhancer-of-split related with YRPW motif 1

hNPC

human neural progenitor cell

HSP

heat shock protein

IC₅₀

half-maximal inhibitory concentration

IF

immunofluorescence staining

MAP2

microtubule-associated protein 2

NCAM1

neural cell adhesion molecule 1

NPC

neuron precursor protein

PreD

preaggregation neurosphere differentiation

PS

presenilin

qRT-PCR

quantitative RT-PCR

ReN

ReNcell VM

ReN-G

ReNcell VM cell stably expressing green fluorescent protein

RT

room temperature

SGSM

soluble γ-secretase modulator

TBS-T

Tris-buffered saline and Tween 20