An exo-cell assay for examining real-time γ-secretase activity and inhibition
© Shelton et al. 2009
Received: 02 April 2009
Accepted: 02 June 2009
Published: 02 June 2009
Skip to main content
© Shelton et al. 2009
Received: 02 April 2009
Accepted: 02 June 2009
Published: 02 June 2009
γ-Secretase is an aspartyl protease that cleaves multiple substrates that are involved in broad biological processes ranging from stem cell development to neurodegeneration. The investigation of γ-secretase has been limited by currently available assays that require genetic or biochemical manipulation in the form of substrate transfection or membrane preparation. Here we report an exo-cell assay that is capable of characterizing γ-secretase activity in any cellular system without limitation. Using a highly active, recombinant substrate this assay can quickly and easily ascertain the status of γ-secretase activity in cell systems and patient samples. We have applied this method to determine the activity of γ-secretase in primary cell samples where transfection and/or membrane isolation are not viable options. Importantly, it allows for the detection of real time γ-secretase activity after inhibitor or drug treatment. The application of this assay to determine the role of γ-secretase in physiological and pathological conditions will greatly facilitate our characterization of this complex protease and help in the development and evaluation of γ-secretase-targeted therapies in Alzheimer's disease or a variety of neoplasms.
γ-Secretase is a multi-subunit protease that executes an extraordinary cleavage of substrates within the lipid bilayer. This process of target hydrolysis within the membrane environment is known as regulated intramembrane proteolysis (RIP)  whereby cleavage by γ-secretase releases a protein fragment from its membrane tether that can then transmit its signal. γ-Secretase was originally identified as the enzyme responsible for cleavage of the amyloid precursor protein (APP) . Cleavage of APP generates β-amyloid peptides that are believed to play a causative role in the neuropathogenesis of Alzheimer's disease  according to the "amyloid cascade hypothesis." Additionally, it has been determined that γ-secretase cleaves a multitude of other substrates that include the Notch receptors , ErbB-4 , CD44 , as well as the Notch ligands Delta-1 and Jagged-2 [7, 8] amongst others. Deregulated Notch signaling has been associated with the development of various cancers, including T-cell Acute Lymphoblastic Leukemia (T-ALL) . Due to the central role of γ-secretase in these pathologies, considerable efforts have been made to characterize this unique protease.
In order to better understand γ-secretase, in vitro assays using purified exogenous recombinant substrate  or assays utilizing isolated membrane from systems overexpressing substrate have been developed and reported [11, 12]. Currently, there are two predominant options to study this protease in a cell line of interest: 1) stably transfect the cell line with plasmids encoding APP, Notch or other substrate fragments and conduct whole-cell based detection assays, or 2) obtain large quantities of the cell line and isolate the membrane fraction in a time-consuming process. This can then be examined using an in vitro assay that employs exogenous recombinant substrate as mentioned previously. Due to these limitations, it is often an extremely challenging task to quickly characterize γ-secretase activity in multiple cell lines and primary cells. Furthermore, it is currently impossible to examine the real-time effect of various treatments on the status of γ-secretase in cell systems without stable transfection. For instance, treatment of a Notch-dependent cell line with γ-secretase inhibitors may have an anti-proliferative effect, but available methods cannot ascertain the extent of real-time γ-secretase inhibition in the system. Therefore, development of an assay that does not require transfection or membrane preparation and is applicable for any cell type has become an urgent issue for defining the relationship of γ-secretase inhibition and its biological responses. This is particularly critical to evaluate γ-secretase inhibitors being used in preclinical and clinical studies because assessment of target inhibition will facilitate the identification and establishment of effective therapies. Recently, we have determined that the use of biotinylated substrate greatly enhanced substrate activity and assay sensitivity over previous versions . This prompted us to apply a similar strategy to the development of a simplified γ-secretase assay capable of quantifying real-time activity in cell-based systems.
In this study we have developed a novel γ-secretase assay that does not require membrane preparation and/or substrate plasmid transfection. This γ-secretase assay that we refer to as an "exo-cell" assay applies a highly active, biotinylated recombinant substrate (Sb4) of γ-secretase exogenously to cells in the presence of a small quantity of CHAPSO detergent. We have found that this 96-well assay format can detect γ-secretase activity from as little as a few thousand cells. Furthermore, we can easily detect γ-secretase activity from primary B-cell Chronic Lymphocytic Leukemia (B-CLL) cells isolated from patients. More importantly, this assay can monitor the real-time γ-secretase activity in a 96-well format after inhibitor treatment and has allowed us to establish a correlation between the anti-proliferative effect of γ-secretase inhibitors against lymphoma cells and real-time reduction in γ-secretase activity. Taken together, the development of this novel assay allows for the characterization of real-time γ-secretase activity directly in cell lines as well as primary patient samples. This assay will simplify the study of γ-secretase and provide new tools in the characterization of this enzyme as well as facilitate the development of therapies against Alzheimer's disease and Notch-dependent neoplasms. Furthermore, the application of this simplified method will greatly enhance our ability to examine this unique enzyme and advance our understanding of γ-secretase biology.
We recently demonstrated that it was possible to directly biotinylate a γ-secretase peptide substrate to be utilized in an in vitro assay . Here, we have designed a truncated, recombinant APP protein that is directly biotinylated during overproduction in Eschericia coli. This substrate is highly active and offers an advantage to develop an easy and sensitive γ-secretase assay. This allows for the elimination of stable transfection of γ-secretase substrate into the cell line of interest or isolation of membrane from large numbers of cells that can then be examined using an in vitro γ-secretase assay.
γ-Secretase inhibitors are currently being used as molecular probes and are being evaluated as potential therapeutics in Alzheimer's disease as well as for various neoplasms like T-cell acute lymphoblastic leukemia where γ-secretase-mediated Notch signaling is tumorigenic. We next modified the exo-cell assay so that our system can monitor real-time γ-secretase activity and inhibition. After HeLa cells were incubated with varying concentrations of GSI-34 or Compound E for 24 hours, media was removed and the cells were washed to remove excess unbound inhibitor. Fresh media containing only CHAPSO detergent and Sb4 substrate were placed back onto the cells and the exo-cell assay was then conducted as previously described. HeLa cells that were treated in this manner with GSI-34 show a dose-dependent inhibition of γ-secretase activity (Fig. 2c). This modified, extended treatment exo-cell assay is capable of quantifying remaining γ-secretase activity following drug treatment on virtually any cell type. The IC50 values for Compound E and GSI-34 were calculated using the extended exo-cell assay (Fig. 2d and Fig. 3). Comparing the potency of these unique GSIs in currently established in vitro and cell-based assays reveals that the extended exo-cell assay more closely mimics that witnessed in a cell-based γ-secretase assay that uses N2A mouse neuroblastoma cells stably expressing the APP substrate (N2A-APP) (Fig. 2d and Fig. 3). The IC50 values for Compound E and GSI-34 in the extended exo-cell assay were 2.83 nM and 38.7 nM respectively (Fig. 2d and Fig. 3) as compared to 4.6 nM and 32.4 nM, respectively, in the cell-based assay (Fig. 3) – both GSIs exhibited decreased potency in the cell-based and extended exo-cell assays as compared to their respective in vitro values. Regardless, the trend of decreasing potencies of GSIs in the extended exo-cell assay is similar to that witnessed in the stable N2A-APP cell-based system and this is likely due to the GSIs being incubated for 24 hours in the presence of a cellular environment that can affect compound half-life amongst other factors. These data show that our exo-cell assay can be used to evaluate the real-time status of γ-secretase activity in cell lines in a simple and sensitive manner.
Lastly, it was not known whether this exo-cell method could be used to measure activity in primary samples from patients. Peripheral primary B-cell chronic lymphocytic leukemia cells (B-CLL) are arrested in the G0 phase of the cell cycle  and this condition makes it very difficult to assay γ-secretase activity in these primary cell samples. Additionally, it has been shown that Notch2 plays a role in the overexpression of CD23 in B-CLL and this may be related to the development of this neoplasm . Therefore, the study of γ-secretase and Notch with regard to B-CLL biology has recently become an urgent issue. Stable transfection of substrate into a non-proliferating cell line is not a practical option and isolating enough B-CLL cells from a patient to prepare membrane fractions for use in in vitro assays is not feasible. However, the exo-cell assay now allows for the determination of protease activity quite easily. We have quantitated activity from three separate B-CLL patient samples and defined background activity for the assay in the presence of 1 μM Compound E (Fig. 4b) from 50,000 total B-CLL cells. Using the exo-cell assay we have been able to characterize γ-secretase activity in B-CLL patient samples for the first time. This previously would have been nearly impossible, but this assay makes it simple to detect protease activity in troublesome B-CLL patient samples in a few hours.
The application of our new exo-cell assay will have wide-ranging implications for the study of γ-secretase. While extensive efforts to characterize this protease have already been put forth due to its connection to Alzheimer's disease as well as numerous cancers, technical difficulties associated with the study of membrane enzymology has hindered much progress. The presented exo-cell assay will simplify the investigation of γ-secretase across different cell lines and tissue types as well as following diverse treatment conditions. Furthermore, this assay is able to efficiently evaluate the real-time status of γ-secretase from primary patient samples and completely eliminates the need for stable transfection of substrate or isolation of membrane. Not only will the quantification of protease activity in various cell lines now be simple, but this novel assay also provides a rapid means for evaluating the efficacy of potential therapeutics that affect γ-secretase in their target environment. For instance, γ-secretase inhibitors could previously be screened using in vitro assays against membrane isolated from the target cell line or tissue they were being used to treat, however, the real-time γ-secretase activity remaining in this system could not readily be evaluated following treatment. This is now an easy process that can be applied to nearly any cell line or tissue type and will require a nominal number of cells to do so. We are able to detect significant γ-secretase activity from as few as 2,500 HeLa adenocarcinoma cells using 0.25% CHAPSO detergent concentration. The detection of real-time activity is a significant development due to the central role of γ-secretase in numerous biological signaling pathways as well as in various disease states. Previously, it was impractical to evaluate γ-secretase activity in a system following cellular manipulation, however the exo-cell assay provides a simple means to assess the effect of GSI-application or other treatments on γ-secretase activity. Interestingly, we have shown that in a γ-secretase-dependent lymphoma line, cellular proliferation can be maintained by a minimal fraction of remaining γ-secretase activity. This finding may help to clarify why therapeutic concentrations of GSIs against lymphoma and other systems sensitive to inhibition of γ-secretase far exceed their calculated in vitro IC50 values. The real-time exo-cell assay simplifies the study of γ-secretase in diverse systems and will greatly facilitate the evaluation and development of novel therapies against Alzheimer's disease and cancers where γ-secretase plays a tumorigenic role as well as allow for the investigation of γ-secretase biology in systems where examination was not previously possible.
Briefly, a DNA fragment encompassing amino acids 620–695 of the 695-aa isoform of APP as well as a maltose binding protein tag was cloned into the pIAD16 prokaryotic vector . Additionally, there was an Avitag also incorporated into this vector. Avitag, a specific 15-residue peptide, is recognized by biotin ligase that specifically catalyzes an attachment of biotin to the lysine residue within the Avitag. The protein was then co-expressed in Eschericia coli with the pACYC184 biotin ligase plasmid. IPTG at 0.1 mM was used to induce expression of Sb4 as well as biotin ligase at 20°C for 5 hours in the presence of 50 μM biotin. Biotin ligase directly biotinylates the avitag during protein expression. Sb4 was ultimately purified using an amylose resin column, eluted with excess maltose and thrombin-cleaved to remove maltose-binding protein from the purified substrate. The biotinylation of recombinant substrate was confirmed by LC-MS.
Cells were seeded at their indicated concentration in 96-well plates and allowed to attach overnight (Fig. 1c). The next day, media was removed and cells were washed once with PBS. Fresh media was then added containing 0.25% CHAPSO detergent, Sb4 substrate to a final concentration of 1 μM, and 1% DMSO or γ-secretase inhibitor. This was incubated for 2.5 hours at 37°C. Media was removed and cell debris was pelleted from this media for 5 min. at 3,500 rpm's. Supernatant was then added to ruthenylated G2-10* antibody that recognizes cleaved product, but not uncleaved substrate. This was incubated for an additional 2 hours at room temperature. Finally, magnetic streptavidin beads were added to a final concentration of 80 μg/ml and incubated for 30 min. at room temperature. Assay buffer was added to the samples and γ-secretase-mediated cleavage of substrate was monitored using electrochemiluminescence (ECL) . Protocol for HTRF detection can be found below.
Conversely, an extended exo-cell assay was performed whereby cells were seeded and allowed to attach overnight as previously described. However, the next day media was removed and replaced with fresh media containing 1% DMSO or γ-secretase inhibitor. These were incubated for 24 hours at 37°C. Following this incubation, the cells were washed once with PBS and fresh media containing detergent and substrate were added. From this point on, the exo-cell assay as described above was followed.
Homogeneous time-resolved fluorescence (HTRF) was also incorporated for the exo-cell assay detection of cleaved substrate. The exo-cell assay proceeded as previously described except that media supernatent was collected and 10 μl of supernatent was added to 10 μl of HTRF detection mix. The HTRF detection mix contained 1 nM G2-10 antibody, 2 nM anti-mouse IgG cryptate entity and 15 nM Streptavidin-XL665 flourophore. These were all dissolved in HTRF detection buffer which was 50 mM phosphate buffer, pH 7.0, 0.2% BSA and 0.8 M potassium fluoride. Exo-cell supernatent and HTRF detection mix were incubated together for 5 hours in a 384-well plate and then activity was measured by stimulating at 337 nm and reading at 665 nm.
The in vitro and cell-based γ-secretase assays were performed similarly as previously described . Briefly, the recombinant Sb4 substrate was incubated for 2.5 hours at 37°C in pH 7.0 PIPES buffer in the presence of 0.25% CHAPSO detergent and solubilized γ-secretase at a final concentration of 40 ng/ul. The detection of cleaved substrate was determined using ruthenylated G2-10* antibody. Due to biotinylation of the Sb4 substrate, this eliminated the additional use of 6E10 biotinylated antibody from the original protocol.
The cell-based assay utilized N2A mouse neuroblastoma cells that stably overexpress the amyloid precursor protein. These were incubated with γ-secretase inhibitors in a final concentration of 1% DMSO for 24 hours. Following incubation, the supernatant was removed from the cells and assayed for Aβ 40-site cleaved APP product using ruthenylated G2-10* antibody as well as biotinylated 4G8 antibody.
A20 mouse lymphoma cells were seeded in a 96-well plate at a concentration of 5 × 105 cells/ml in 100 μl RPMI media containing 2% fetal bovine serum. An additional 100 μl of media containing DMSO or γ-secretase inhibitor was added and incubated for 48 hours at 37°C. After this incubation, the cells were incubated for 5 hours with 2 μCi/ml [3H]thymidine at 37°C. Proliferative response was then evaluated by harvesting the tritiated DNA from cells using a Skatron cell harvester and proliferation assessed as a function of [3H]thymidine incorporation measured on a β-counter.
Primary samples were obtained from patients diagnosed with B-cell chronic lymphocytic leukemia who were untreated. Written informed consent was obtained from each patient in accordance with the guidelines of the Institutional Review Board of Memorial Sloan Kettering Cancer Center and the Declaration of Helsinki. Peripheral blood mononuclear cells (PBMCs) were isolated using standard Ficoll-Hypaque density gradient and subsequently stored in liquid nitrogen. Prior to use in assay, samples were thawed and resuspended in RPMI media and allowed to attach overnight at 37°C. The extended exo-cell assay was then performed as described above.
The authors would like to express thanks to L. Placanica for critical analysis of the manuscript. We are also grateful to D. Scheinberg for assistance in obtaining B-CLL patient samples as well as helpful discussion regarding experimental design. C. Shelton is supported by an NIH NRSA predoctoral fellowship 5F31NS053218-02. Additional support was provided by NIH Grant AG026660 (Y.M.L.) and the Alzheimer's Association (Y.M.L.), Mr. William H. Goodwin and Mrs. Alice Goodwin and the Commonwealth Foundation for Cancer Research, the Experimental Therapeutics Center of MSKCC, and the William Randolph Hearst Fund in Experimental Therapeutics.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.