|Home | About | Journals | Submit | Contact Us | Français|
Several lines of evidence suggest that aberrant Notch signaling contributes to the development of several types of cancer. Activation of Notch receptor is executed through intramembrane proteolysis by γ-secretase, which is a multimeric membrane-embedded protease comprised of presenilin, nicastrin (NCT), anterior pharynx defective 1 and PEN-2. In this study, we report the neutralization of the γ-secretase activity by a novel monoclonal antibody A5226A against the extracellular domain of NCT, generated by using a recombinant budded baculovirus as an immunogen. This antibody recognized fully glycosylated mature NCT in the active γ-secretase complex on the cell surface, and inhibited the γ-secretase activity by competing with the substrate binding in vitro. Moreover, A5226A abolished the γ-secretase activity-dependent growth of cancer cells in a xenograft model. Our data provide compelling evidence that NCT is a molecular target for the mechanism-based inhibition of γ-secretase, and that targeting NCT might be a novel therapeutic strategy against cancer caused by aberrant γ-secretase activity and Notch signaling.
Notch is a membrane-bound transcription factor that regulates metazoan developmental program (Kopan and Ilagan, 2009). After the interaction with its ligand, Notch is endoproteolyzed by an intramembrane cleaving enzyme, γ-secretase, to liberate the Notch intracellular domain that translocates into the nucleus and activates the transcription of downstream target genes. Recent studies suggest that alterations of Notch signaling contribute to cancer development. Notably, activating mutations of NOTCH1 account for over 50% of human T-cell acute lymphoblastic leukemia (Ferrando, 2009). These mutations result in ligand-independent generation or prolonged existence of Notch intracellular domain, which causes upregulation of Notch signaling. Also, aberrant Notch signaling has been implicated in tumorigenesis of glioma, colon, pancreatic, non-small cell lung and breast cancers (Pannuti et al., 2010; Yin et al., 2010). It has been reported that treatment with γ-secretase inhibitors causes a dramatic decrease in the proliferation of these cancer models (van Es and Clevers, 2005; Fan et al., 2006; Kimura et al., 2007; Osipo et al., 2008; Luistro et al., 2009). Moreover, Notch signaling contributes to angiogenesis in solid tumors as well as drug resistance against chemotherapy. Thus, modulation of Notch signaling could be used in the treatment of these cancers.
Monoclonal antibody (mAb)-based anti-cancer strategy has been developed and now is widely used in clinics (Samaranayake et al., 2009; Weiner et al., 2010). Especially, mAbs exhibit several advantages compared with small compounds; long half-life and multiple immunological mechanisms, including antibody-dependent cell-mediated cytotoxicity, have been implicated in the anti-cancer effects of therapeutic mAbs. Moreover, functional mAbs with antagonistic effects on the target molecule also have been developed as therapeutic mAbs (for example, trastuzumab for HER2, cetuximab for EGFR, tocilizumab for IL-6R). Intriguingly, some of the target molecules are enzymes working within the cell (that is, receptor tyrosine kinase). In such cases, a combination strategy using a therapeutic mAb and a small compound inhibitor against the same target is expected to be therapeutically effective (Samaranayake et al., 2009). However, several difficulties underlie the development of mAbs that react with proteins in a functional state. In general, proper folding and post-translational modifications in extracellular domain (ECD) of the membrane protein are required for its function, while a significant proportion of overexpressed recombinant proteins tends to misfold and aggregate without forming proper tertiary structures. We have developed a novel membrane protein expression and immunization system using budded baculoviral particles, which are able to display biologically active membrane proteins including the γ-secretase complex on viral virions (Masuda et al., 2003; Urano et al., 2003; Hayashi et al., 2004; Saitoh et al., 2006). Using this technology, we have successfully obtained several mAbs that recognize peptide transporters and G-protein-coupled receptors in a functional conformation (Saitoh et al., 2007).
γ-Secretase is responsible for the intramembrane proteolysis of various type I transmembrane proteins including Notch as well as amyloid precursor protein, the latter being a precursor for amyloid-b peptide that is implicated in the pathogenesis of Alzheimer disease (Tomita, 2009; De Strooper et al., 2010). To date, however, mAbs against the γ-secretase complex that inhibit the proteolytic activity have not been reported, while several potent γ-secretase inhibitors have been identified. γ-secretase is a multimeric membrane protein complex comprised of presenilin (PS), nicastrin (NCT), anterior pharynx defective 1 (APH-1) and PEN-2. PS is the catalytic subunit that forms the hydrophilic catalytic pore buried within the membrane (Takasugi et al., 2003). Importantly, several γ-secretase inhibitors identified so far directly target PS, while other components are also indispensable for the physiological γ-secretase activity (Tomita, 2009). Among these subunits, NCT is a type I membrane glycoprotein with a large ECD. During the proteolytic process, NCT ECD captures the N terminus of the substrate as a primary substrate receptor (that is exosite) for γ-secretase (Shah et al., 2005). In the present study, we identified a mAb A5226A that recognizes NCT ECD as a neutralizing mAb against the intramembrane proteolysis by the γ-secretase. We also showed a therapeutic potential of A5226A against proliferation of cancer cells, in which aberrant activation of the γ-secretase activity-dependent signaling is involved.
To generate functional mAbs that modulate the γ-secretase activities, we chose NCT as a target protein among the four known components for the following reasons: (i) NCT has a large ECD that is expected to be displayed on the surface of budded baculovirus, (ii) NCT has been reported to function as a substrate receptor (Shah et al., 2005). We immunized gp64 transgenic mice with NCT-expressing budded baculo-virus and obtained a dozen of positive hybridoma clones. We further characterized two anti-NCT mAbs (A5201A and A5226A), which showed a high titer against NCT ECD (Figure 1a). NCT undergoes maturation process including endoglycosidase H-resistant, complex-type N-glycosylation and acquiring a trypsin resistance during the assembly of the γ-secretase complex (Yu et al., 2000; Esler et al., 2002; Tomita et al., 2002; Shirotani et al., 2003; Hayashi et al., 2009). Only the mature form of NCT is detected in the functionally active γ-secretase complex at late secretory pathway including cell surface (Kaether et al., 2002), while the immature form of NCT forms a subcomplex with APH-1 at endoplasmic reticulum (LaVoie et al., 2003). On immunoblot analysis, both mAbs specifically reacted with both mature and immature forms of human NCT overexpressed in fibroblasts from Ncstn knockout mouse (NKO cells) (Figure 1b). Intriguingly, both mAbs reacted exclusively with human NCT but not with that of other species (that is mouse, rat, chicken and fruit fly) (Supplementary Figure 1A–C). These mAbs recognized endoglycosidase H- as well as PNGase-treated deglycosylated NCT, the latter lacking all N-glycosylation, although the reactivity against deglycosylated NCT was considerably reduced (Figures 1c and d). Collectively, these data suggest that the major epitopes for these mAbs are located on NCT polypeptide, and that the sugar chains attached to NCT might be involved in the recognition by the mAbs.
To investigate the reactivities of these mAbs against NCT under a condition where the integrity of the γ-secretase complex is preserved, immunoprecipitation experiments were performed. Surprisingly, A5201A precipitated only the immature form of NCT, whereas A5226A bound to both mature and immature NCT (Figures 2a and b). In accordance with this observation, A5226A precipitated all the other components of the γ-secretase complex, that is, PS1, APH-1aL and PEN-2. In contrast, A5201A precipitated only APH-1aL, indicating that A5201A specifically binds immature NCT as well as NCT–APH-1 subcomplex (LaVoie et al., 2003). Using a set of deletion mutants (Shirotani et al., 2003) (Figure 2c), we found that epitopes for these mAbs reside within NCT ECD in a different manner: A5201A showed no binding to any mutants except for NCT/Δ1. In contrast, A5226A failed to react with NCT/Δ4 lacking amino acid residues 361–516 of NCT, while the reactivity against the other deletion mutants was significantly reduced (Figure 2d). Considering previous reports that NCT ECD undergoes a conformational change during the formation of the active γ-secretase complex, these data suggest that A5201A recognizes a broad region or structure in NCT ECD that is hindered in the mature NCT-containing complex by the conformational change; in contrast, the recognition by A5226A requires a presence of the amino acid residues 361–516, which is accessible in the folded/mature NCT ECD. To further narrow down the epitope of A5226A, we analyzed the reactivity of human and mouse NCT chimeric polypeptides using sandwich enzyme-linked immunosorbent assay. Preliminary experiments using various human/mouse chimeric NCT revealed that reactivity of A5226A critically depended on the region encompassing residues 333–393. As only six residues (that is K349, Q355, V359, E375, V383 and R391) within this region differed between human and mouse NCT (Figure 2e), we mutated each residues and tested their reactivity with A5226A using the same assay. As shown in Figure 2f, human to mouse mutations K349N, Q355R, V359I, E375D and V383M did not affect the binding of A5226A, while mutation of Arg391 to Lys completely abolished the binding. Conversely, mouse NCT gained full reactivity toward A5226A when K391 was mutated to R. These results clearly indicate that R391 lies at the heart of the binding epitope of A5226A.
Target molecules of the therapeutic mAbs should be located at late secretory pathway including plasma membrane. However, precise localization of the active γ-secretase complex still remains unclear, while chemical biology and genetic approaches revealed that Notch cleavage by γ-secretase is occurred at plasma membrane and/or endocytic compartment (Tarassishin et al., 2004; Kopan and Ilagan, 2009). To investigate whether anti-NCT mAbs recognizes endogenous NCT at cell surface, we performed immunocytochemical analysis of non-permeabilized HeLa cells. Intriguingly, immunostaining with A5226A gave a ‘scattered patchy’ pattern on cell surface unlike that of A5201A, which showed no staining (Figure 3a). We found that this patchy staining was well merged with the staining by CTB, which binds to the pentasaccharide chain of ganglioside GM1 (Boesze-Battaglia, 2006). We further examined subcellular localization of endogenous NCT using permeabilized HeLa cells (Figure 3b). Immunostaining by A5201A resulted in a reticular pattern that merged well with that by an anti-protein disulfide isomerase antibody, suggesting that A5201A specifically recognizes the endoplasmic reticulum-resident, immature NCT under the native condition. However, the CTB-positive patchy dots stained by A5226A were not merged with endoplasmic reticulum or Golgi markers, protein disulfide isomerase or TGN46, respectively, suggesting that A5226A recognizes NCT in CTB-positive micro-domain. To confirm that the observed patchy pattern corresponds to the localization of bona fide NCT, NKO cells were analyzed (Li et al., 2003) (Figure 3c). Neither of the mAbs showed immunoreactivities in NKO cells. In contrast, in NKO cells expressing human NCT, A5201A and A5226A stained the reticular and patchy structure, respectively, in a similar fashion to those observed in HeLa cells. Moreover, A5226A-positive patch was well merged with the immunoreactivity of anti-PS1 mAb, suggesting that the functional γ-secretase complex resides at GM1-positive membrane microdo-main (Figure 3d). In good accordance with these results, several biochemical analyses revealed that the mature form of NCT as well as the γ-secretase activity was fractionated in lipid rafts, where sphingolipids were enriched (Urano et al., 2005; Vetrivel and Thinakaran, 2010). Finally, we examined the localization of a direct γ-secretase substrate, NEXT fragment, proteolytically generated by site 2 cleavage from Notch holoprotein as well as a recombinant Notch-based substrate NΔE. Treatment with γ-secretase inhibitor, N-[N-(3,5-difluorophenacetyl)-L-alanyl]-(S)-phenylglycine t-butyl ester (DAPT), caused a significant accumulation of NEXT fragment that was specifically recognized by a Notch1 Val1711 antibody (van Tetering et al., 2009). Intriguingly, NEXT fragments were located in proximity to, and partially co-localized, with the A5226A-positive patch (Figure 3e). Taken together, these data strongly suggest that A5226A specifically recognizes the active γ-secretase complex containing mature NCT localized in CTB-positive microdomain at the plasma membrane.
Immunocytochemical findings that A5226A labels mature NCT, which constitutes an active γ-secretase complex, at CTB-positive patches on cell surface prompted us to examine whether this mAb affects the γ-secretase activity in living cells. Strikingly, A5226A treatment significantly inhibited the γ-secretase-mediated cleavage in a dose-dependent manner in HEK293 cells stably expressing amyloid precursor protein- or Notch-derived g-substrate fused with GAL4/UAS system (Imamura et al., 2009) (Figure 4a). Intriguingly, the levels of PS1 fragments were slightly decreased in A5226A-treated cells (Figure 4b). However, the endocytosis rate of mature NCT, PS1 and amyloid precursor protein at cell surface was unchanged by A5226A treatment (Figure 4c). Moreover, the localization of the substrate as well as NCT was not significantly altered by A5226A treatment (data not shown). Nevertheless, further investigation for the molecular mechanism whereby PS1 fragment was decreased by A5226A incubation would be required. Next, we analyzed the effect of A5226A in in vitro γ-secretase assay using recombinant substrate. Notably, A5226A decreased intrinsic γ-secretase activity in an in vitro assay (Figure 5a). These data indicate that A5226A functions as a neutralizing mAb against the γ-secretase activity. Decreased inhibitory effect of A5226A in in vitro assay might be due to difference in requirement of NCT function in proteolytic mechanism of solubilized enzyme, while a slight decrease in the levels of PS/γ-secretase complex might additionally contribute to the inhibitory effect. NCT harbors a dual function for the active γ-secretase as a stabilization factor and a regulator for intrinsic enzyme activity (Shirotani et al., 2003; Takasugi et al., 2003; Shah et al., 2005; Chavez-Gutierrez et al., 2008; Dries et al., 2009; Hayashi et al., 2009; Zhao et al., 2010). Therefore, we hypothesized that A5226A inhibits the proteolytic activity by interfering with the interaction between NCT and the substrates. To examine this possibility, we performed co-immunoprecipitation experiments using recombinant NCT and Notch-based substrate, N100 (Figure 5b) (Shah et al., 2005). We observed specific precipitation of NCT with the substrate, which was abolished by SDS-pretreatment of NCT, as previously described (Shah et al., 2005). Addition of A5201A or bovine serum albumin showed no effect on the amount of co-precipitated NCT and N100. In contrast, A5226A reduced the precipitation of the substrate with NCT, suggesting that A5226A interfered with the interaction between NCT and the substrates, thereby reducing the γ-secretase activity.
Anti-tumor activity of γ-secretase inhibitors has been examined in clinical trials (Pannuti et al., 2010; Yin et al., 2010). The results above that A5226A exhibits an inhibitory activity against γ-secretase, led us to examine the effect on cancer cell proliferation. Among several types of cancer cells, we selected A549 cells derived from non-small cell lung cancer because of its sensitivity to γ-secretase inhibitors (Luistro et al., 2009). Cell viability of A549 cells was reduced by DAPT treatment, whereas that of HeLa cells was not affected. Further, knockdown of endogenous NCT by small interference RNA in A549 cells also resulted in reduced cell viability, suggesting that proliferation of A549 cells is dependent on the γ-secretase activity (Supplementary Figure 2A-C). We next treated A549 and HeLa cells with A5226A. As observed with DAPT, A5226A reduced cell viability of A549 cells but not that of HeLa cells, suggesting that the γ-secretase-dependent proliferation of A549 cells is inhibited by A5226A (Figure 6a). Notably, A5226A treatment further reduced the viability of DAPT-treated A549 cells, suggesting that A5226A is applicable for a combination therapy targeting γ-secretase with small compounds (Figure 6b). Further, we tested the effect of A5226A on the proliferation of lymphoblastic leukemia DND-41 cells carrying activating mutations in NOTCH1 gene (that is, L1594P and 2444stop) (Weng et al., 2004; Ferrando, 2009; Masuda et al., 2009). Cell survival of DND-41 was sensitive to the treatment by a potent γ-secretase inhibitor dibenzazepine (DBZ) in a similar manner to that of A549 cells (Supplementary Figure 2D). As expected, A5226A treatment, but not control IgG fraction, inhibited the proliferation of DND-41 (Figure 6c) as well as Notch intracellular domain generation (Figure 6d). These data implicated the possibility that γ-secretase-dependent cell survival is also inhibited by A5226A in DND-41. Finally, we examined the anti-tumor activity of A5226A in a xenograft mouse model, in which severe combined immunodeficiency mice were inoculated subcutaneously with DND-41. A5226A, A5201A or control IgG was intraperitoneally administered daily for 4–5 days after inoculation. The size of subcutaneous tumors of the A5226A-treated groups was significantly reduced compared with those of control groups (that is 55.2% reduction in tumor volume at day 38) (Figure 6e). Taken together, we concluded that A5226A functions as an inhibitory mAb for the γ-secretase activity-dependent cancer cell growth in vivo.
Targeted inhibition of γ-secretase activity has a potential to affect cancer cell proliferation/differentiation and angiogenesis in solid tumors (Ferrando, 2009; Kopan and Ilagan, 2009; Pannuti et al., 2010; Yin et al., 2010). Moreover, the effectiveness of antibodies in treating patients with cancer has been increasingly recognized (Samaranayake et al., 2009; Weiner et al., 2010). In the present study, we have identified a novel antibody A5226A against NCT ECD generated by a novel budded baculovirus immunization technology. Biochemical and cell biological analyses showed that A5226A recognizes the active γ-secretase complex containing mature NCT located at CTB-positive microdomain on the plasma membrane. Moreover, A5226A inhibited the γ-secretase activity by perturbing the NCT/substrate interaction, and also showed the inhibitory activity for proliferation of cancer cells both in vitro and in vivo. These data provided us with the compelling evidence that NCT is a molecular target for the modulation of γ-secretase activity, and that the anti-NCT mAb A5226A is applicable to cancer immunotherapy targeting γ-secretase and Notch.
It has been reported that functionally active, properly folded recombinant proteins can be displayed on baculoviral virion membranes released from infected insect cells (Masuda et al., 2003; Urano et al., 2003; Hayashi et al., 2004; Saitoh et al., 2006). Thus, utilizing budded baculovirus as immunogen is a versatile approach to generate mAbs against membrane proteins of interest. Using this technology, we have generated A5201A and A5226A that specifically interact with recombinant NCT ECD with different binding properties. Using deletion and swap mutants of NCT, we found that the epitope of A5226A locates around the residues R391 in human NCT ECD. To gain a structural insight on the significance of R391, we built a homology model of NCT ECD using the structure of transferrin receptor (TfR; PDB ID: 1CX8) as a template (Supplementary Figure 3) (Lawrence et al., 1999). In this model, E333 and the DYIGS motif that are required for the folding and activity of NCT (Yu et al., 2000; Shirotani et al., 2004; Shah et al., 2005; Chavez-Gutierrez et al., 2008; Dries et al., 2009) are situated at the back of a deep channel made by the peptidase-like domain and the apical domain. They correspond to, or lie close to, the active site residues in homologous aminopeptidases and may constitute a putative substrate binding pocket (Shah et al., 2005). Intriguingly, R391 is located in a loop that constitutes an entrance of this channel, suggesting that the binding of A5226A may obstruct the access of substrates. Although the precise function of NCT still remains controversial (Shah et al., 2005; Chavez-Gutierrez et al., 2008; Dries et al., 2009; Hayashi et al., 2009; Zhao et al., 2010), the neutralizing mAb A5226A inhibited the binding between NCT and substrates in vitro, supporting the notion that one of the NCT functions is a substrate receptor. Also, we observed that A5226A treatment slightly decreased the levels of PS1 fragments; a slight reduction in the active enzyme might have potentiated the inhibitory effect of A5226A in vivo. Taken together, as the major mechanism of action of A5226A, one can speculate that the binding of mAb to NCT located on the cell surface might mask a region around E333 and the DYIGS motif and thereby inhibit the substrate recognition by NCT. Further structural studies at an atomic level using X-ray crystallography, nuclear magnetic resonance or other techniques would help us to understand the mode of action of A5226A.
To date, the precise localization of the endogenous γ-secretase complex remains unclear. By immunocyto-chemical analyses, we found that A5226A-positive endogenous NCT is co-localized with CTB-positive patches on the cell surface. CTB is a ligand for GM1 ganglioside that is enriched on the lipid raft micro-domain (Boesze-Battaglia, 2006). In accordance with our results, several lines of evidence indicate that γ-secretase activity is biochemically detected in lipid raft fractions (Urano et al., 2005; Vetrivel and Thinakaran, 2010), suggesting that A5226A-positive patches represent the localization of the endogenous active gsecretase complex at the cell surface. Recently, biochemical co-purification of tetraspanins (that is CD9 and CD81) with active γ-secretase complex was reported (Wakabayashi et al., 2009), which form cholesterol-rich lipid raft-like microdomains termed tetraspanin-enriched microdomains. However, CTB-positive patches were not co-localized with CD9 nor CD81 on NKO/ hNCT cell surface (Supplementary Figure 4), suggesting that A5226A-positive patches are distinct from tetraspanin-enriched microdomains containing CD9 or CD81. However, several tetraspanins form different microdomains in a cell-type or subcellular localization-specific manner. In particular, TSPAN5 and TSPAN33 are implicated in γ-secretase cleavage of Notch (Dunn et al., 2010). It would be of interest to examine if CTB-and A5226A-positive patches contain tetraspanins. Further molecular and biochemical analyses would be needed to clarify the nature of CTB- and A5226A-patches on the cell surface.
Chemical biology studies revealed that all γ-secretase inhibitors investigated so far target PS (Tomita, 2009). Our results clearly indicate that targeting NCT function by A5226A is effective against cancer cell growth in vitro as well as in vivo. Moreover, a combination approach with DAPT increased the efficacy, suggesting that targeting NCT ECD is a novel and versatile approach against γ-secretase activity-dependent cancers. Supporting this notion, it has been recently reported that the expression levels of NCT were upregulated in breast cancer, and that a loss-of-function of NCT decreased the cell proliferation of breast cancer cells (Filipovic et al., 2011). However, a remaining issue on targeting γ-secretase activity is the adverse effects by total Notch signaling inhibition (Tomita, 2009; De Strooper et al., 2010). In this context, targeted inhibition of Notch1 or Notch2 by specific mAbs showed low toxicity (Wu et al., 2010), supporting the notion that mechanism-based adverse effects associated with Notch signaling inhibition can be avoided by using a substrate-specific inhibition. Although A5226A showed similar levels of inhibition against amyloid precursor protein and Notch cleavages in a cell-based assay, it remains to be determined whether A5226A inhibits all four Notch family members and what other substrates might be preferentially inhibited. Intriguingly, some point mutations at NCT ECD affected the γ-secretase activity in a substrate-specific manner (Chavez-Gutierrez et al., 2008), suggesting that NCT ECD is involved in the substrate selectivity of γ-secretase. Thus, substrate-specific inhibitory mAbs might be obtained by targeting NCT ECD. In addition, subcellular localization of the substrates might also affect the substrate preference of γ-secretase (Vetrivel and Thinakaran, 2010). Understanding possible substrates within the A5226A-positive patches on the cell surface will be of considerable interest. Therapeutic mAbs provide clinical benefit by target specificity, low toxicity and longer half-lives in vivo. Importantly, mAbs can be engineered on a more rational basis to enhance the efficacy and the specificity (Samaranayake et al., 2009; Weiner et al., 2010). Nevertheless, more in depth understandings of the complex biology of γ-secretase regarding its substrate selectivity will be needed to predict liabilities of γ-secretase inhibitors and inhibitory mAbs such as A5226A.
In summary, we have established anti-NCT mAb A5226A that neutralizes the γ-secretase activity, thereby affecting cancer cell proliferation in vivo. Development of therapeutic mAbs will not only provide clinical benefits to patients but also lead to discoveries of functions of unknown cell surface antigens (Samaranayake et al., 2009; Weiner et al., 2010). Our results support the notion that one of the NCT functions is a substrate binding, which was inhibited by A5226A, at GM1-positive domain on the cell surface. Importantly, to our knowledge, A5226A is the first mAb that regulates the proteolytic activity of the intramembrane proteases, which execute emerging biological and pathological functions, by the inhibition of substrate recognition. Nevertheless, our present study also suggests that small molecules or proteins that regulate the NCT activity might provide a novel therapeutic strategy for modulating the γ-secretase activity against cancers.
Construction of expression plasmid, cell culture, transfection/ infection and immunological methods are shown in Supple mentary Methods. For generation of mAbs using NCT expressing budded baculovirus as an immunogen, transgenic mice expressing gp64, a major viral envelope protein, were immunized five times (Saitoh et al., 2007). Hybridoma cells were generated by a standard polyethylene glycol mediated method using splenocytes and NS 1 myeloma cells. Hybrido ma cells were maintained in RPMI 1640 medium containing 15% (v/v) fetal bovine serum, penicillin/streptomycin, 2 mM of L glutamine (Invitrogen, Carlsbad, CA, USA), 1 mM of sodium pyruvate (SIGMA, St Louis, MO, USA) and 45 μg/ml of gentamicin (Schering Plough, Kenilworth, NJ, USA). En zyme linked immunosorbent assay screenings were performed using mock or NCT expressing budded baculovirus as a capture antigen. For the preparation of ascites, hybridoma cells (1 3 × 107 cells) were injected intraperitoneally to 6 week old male BALB/c mice, which were treated with pristane (SIGMA) 3 7 days before the injection. After 10 14 days, the ascites were collected and purified with MAbTrap kit (GE Healthcare, Piscataway, NJ, USA). Purified ascites were stored at 80 °C until used. For Scatchard plot, 10 μg/ml of mAbs were incubated with 0, 0.19, 0.38, 0.75 or 1.5 mM of NCT ECD to form antibody/antigen complex and were applied to NCT ECD coated plates. Subsequently, anti mouse IgG antibody conjugated with horseradish peroxidase was incubated, and the binding of free mAb was quantitated by measuring OD450 using peroxidase substrate.
Microsome fraction was obtained from infected Sf9 cells expressing NCT V5/His or N100 FLAG/His. After solubiliza tion by 1% 3 [(3 cholamidopropyl)dimethylammonio] 2 hy droxypropanesulfonate (CHAPSO), the lysates were mixed with or without antibodies, and pulled down by anti V5 or FLAG antibody conjugated agarose beads (SIGMA). The precipitated proteins were eluted by sample buffer and analyzed by immunoblot.
Cells were plated in 96 well plates (2.5 5 × 103 cells per well) 24 h before the addition of designated concentrations of γ secretase inhibitors or mAbs. Equal concentration of the vehicle carrier (dimethyl sulfoxide or phosphate buffered saline, respectively) was always present in the control wells. After incubation for indicated time, 20 μl of 2.5 mg/ml of 3 (4,5 dimethylthiazol 2 yl) 2,5 diphenyltetrazolium bromide (final concentration: 400 μg/ml) or 10 μl of Alamar Blue (Serotec, Oxford, UK) was added to each well and incubated for 3 4 h at 37 degrees. In the case of 3 (4,5 dimethylthiazol 2 yl) 2,5 diphenyltetrazolium bromide, solubilization of the precipitate was performed with 200 μl of stop solution (10% (w/v) SDS and 0.01 M HCl). Cell viability was calculated from absorption or fluorescence values for 3 (4,5 dimethylthiazol 2 yl) 2,5 diphenyltetrazolium bromide or Alamar Blue, respec tively.
Severe combined immunodeficiency mice (C.B 17/Icr scid/scid Jcl; 6 weeks old female) were purchased from CLEA Japan (Tokyo, Japan) and maintained under specific pathogen free conditions. All experimental procedures were performed in accordance with the guidelines for animal experiments of the University of Tokyo. For xenograft mouse model (Masuda et al., 2009), severe combined immunodeficiency mice at 6 8 weeks of age were inoculated subcutaneously in the right flank with 3 × 107 cells in 300 μl of phosphate buffered saline, and were randomly assigned to control or mAb treated groups the day after tumor inoculation. The antibodies were administered intraperitoneally every 4 5 days for at least 30 days at a dose of 50 mg/kg/day. Tumor size was measured at the greatest length and width. The volume was calculated as 1/2×(tumor length) × (tumor width)2.
We are grateful to Drs T Fukuyama, S Yokoshima (The University of Tokyo), C Haass (Ludwig Maximilians Univer sity H Natsugari (Teikyo University), R Kopan (Washington University in St Louis), G Thinakaran (The University of Chicago), M Vooijs (University Medical Center Utrecht Cancer Center) and G Yu (The University of Texas Southwestern Medical Center) for valuable reagents and our current and previous laboratory members for helpful discussions. We also would like to thank Keiko Tamura Kawakami and Maiko Nampo for their excellent technical support. This work is supported in part by Grants in Aid for Young Scientists (S) (for TT) from Japan Society for the Promotion of Science (JSPS), by the Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation (for TT, TI), Scientific Research on Priority Areas ‘Research on Pathomechanisms of Brain Disorders’ from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) (for TT, TI), by Targeted Proteins Research Program grant from the MEXT (for TT, TI, JT), by Core Research for Evolutional Science and Technology grant from the MEXT (for TT, TI), Japan. IH and ST were research fellows of JSPS.
Conflict of interest
The authors declare no conflict of interest.