Calpain Inhibitor I Increases β-Amyloid Peptide Production by Inhibiting the Degradation of the Substrate of γ-Secretase

The calpain inhibitorN-acetyl-leucyl-leucyl-norleucinal (ALLN) has been reported to have complex effects on the production of the β-amyloid peptide (Aβ). In this study, the effects of ALLN on the processing of the amyloid precursor protein (APP) to Aβ were examined in 293 cells expressing APP or the C-terminal 100 amino acids of APP (C100). In cells expressing APP or low levels of C100, ALLN increased Aβ40 and Aβ42 secretion at low concentrations, decreased Aβ40 and Aβ42 secretion at high concentrations, and increased cellular levels of C100 in a concentration-dependent manner by inhibiting C100 degradation. Low concentrations of ALLN increased Aβ42 secretion more dramatically than Aβ40 secretion. ALLN treatment of cells expressing high levels of C100 did not alter cellular C100 levels and inhibited Aβ40 and Aβ42 secretion with similar IC50 values. These results suggest that C100 can be processed both by γ-secretase and by a degradation pathway that is inhibited by low concentrations of ALLN. The data are consistent with inhibition of γ-secretase by high concentrations of ALLN but do not support previous assertions that ALLN is a selective inhibitor of the γ-secretase producing Aβ40. Rather, Aβ42 secretion may be more dependent on C100 substrate concentration than Aβ40 secretion.

The ␤-amyloid peptide (A␤) 1 is the major protein component of the senile plaques found in the brain of Alzheimer's disease (AD) patients. A␤ is produced by proteolysis of a single transmembrane domain protein known as the amyloid precursor protein (APP) (reviewed in Refs. 1 and 2). The first step in A␤ production involves the cleavage of APP by an uncharacterized protease termed ␤-secretase. Cleavage of APP by ␤-secretase produces a large ectodomain protein known as APPs␤, which is ultimately secreted, and a C-terminal 14-kDa membrane-bound fragment known as C100 (also termed C99 in some references). C100 is subsequently cleaved by ␥-secretase, another uncharacterized protease that cleaves within the transmembrane domain of C100 and produces the 39 -43-amino acid A␤ peptides. A␤40 is the dominant species of A␤ secreted from cultured cells and is also more abundant in cerebrospinal fluid of normal and AD patients. A␤42, which comprises about 5-10% of total A␤ secreted from cultured cells, is more amyloidogenic and is the major species of A␤ that is deposited at the early stage of senile plaques formation.
Familial AD has thus far been associated with autosomal dominant mutations in the genes encoding APP, presenilin 1 (PS1), and presenilin 2 (2,3). Multiple mutations in these three genes are associated with increased A␤42 production (4 -6). Collectively, these data suggest that excessive A␤42 production is critical for the development of AD. Whereas the locations of mutations in the APP gene suggest that the mutations lead to increased A␤42 production by increasing cleavage of APP by ␤or ␥-secretase, the mechanism by which presenilin mutations increase A␤42 production remains unclear. Primary neuronal cultures derived from PS1 knock-out mice exhibit marked reduction of A␤ secretion (7), suggesting an essential role of PS1 in generating A␤. Understanding the cellular mechanisms that regulate A␤ production will be a key step to unraveling the pathogenesis of AD.
A␤ production can also be modulated by peptide aldehyde protease inhibitors such as N-acetyl-leucyl-leucyl-norleucinal (ALLN, also known as calpain inhibitor I or LLnL) (5, 8 -11). ALLN was first identified as a cysteine protease inhibitor (12), but at high concentrations it can also inhibit proteasome-associated activities (13). It has been reported that ALLN inhibits A␤40 production at concentrations that have little effect on or even increase A␤42 production (5,9). These data are interpreted as evidence suggesting that A␤40 and A␤42 are produced by distinct ␥-secretases. In contrast, a recent study demonstrates that ALLN increased A␤40 and A␤42 production at low concentrations and decreased A␤40 and A␤42 production at higher concentrations (10). Thus, the reported effects of ALLN on A␤ production are conflicting, and the mechanism(s) by which ALLN modulates A␤ production are not clear. Nevertheless, ALLN may serve as an important tool to investigate the regulation of A␤ biosynthesis.
In this study, the effects of ALLN on A␤40 and A␤42 secretion are examined in detail and the mechanism by which ALLN modulates A␤ secretion is further defined. The data provide the novel insights that substrate availability plays a major role in regulating A␤40 and A␤42 production by ␥-secretase. many). W02 recognizes an epitope at amino acids 5-8 of the A␤ peptide, and G2-10 and G2-11 specifically recognize the C terminus of A␤40 and A␤42, respectively (14). Antibody 54 was obtained from Dr. Barry Greenberg (Cephalon, Inc., West Chester, PA) and recognizes the secreted APP ectoprotein formed after ␤-secretase cleavage (APPs␤) (15). Antibody 14, which recognizes an N-terminal domain of PS1 (16), was obtained from Dr. Samuel Gandy (Cornell University, New York). N-Acetyl-leucyl-leucyl-norleucinal (ALLN) was purchased from Boehringer Mannheim. All tissue culture reagents used in this work were from Life Technologies, Inc.
cDNA Constructs, Cell Culture, and Transfection of Cultured Cells-A human APP695 cDNA clone with the Swedish mutation (APPsw) and a cDNA encoding the the C-terminal 99 amino acids of APP plus an N-terminal methionine (hereafter referred to as C100) were obtained from Dr. Barry Greenberg. C100 was cloned into the expression vector pcDNA3.1 (Invitrogen, San Diego, CA). The SPC100 construct consists of the N-terminal 18 amino acids of APP appended to the N terminus of C99 as described by others (17). To prepare the SPC100 construct, residues 19 -596 of the APP695 cDNA were deleted by using the Seamless Cloning Kit (Stratagene, La Jolla, CA). The resulting SPC100 construct was cloned into the pcDNA3.1 vector. The APP London mutation (18) was introduced into C100 and SPC100 constructs using the QuickChange TM site-directed mutagenesis kit (Stratagene). The human cDNAs encoding wild type PS1 and mutant PS1 with the exon 9 deletion (PS1⌬E9) (19) were obtained from Dr.
Peter St. George-Hyslop (University of Toronto, Toronto, Canada). PS1⌬E9 sometimes is referred to as exon 10 deletion (20), as an alternate 5Ј-untranslated region exon was missed in the initial characterization of PS1 genomic structure (19).
Human embryonic kidney 293 cells were purchased from American Type Culture Collection (Rockville, MD) and were grown in Dulbecco's modified Eagle's media (DMEM) supplemented with 10% fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin. For transient expression of C100 and SPC100, 293 cells were seeded into 6-well plates, and transfection was conducted 2 days later when the cells reached 60 -70% confluence. Cells were transfected by means of Lipo-fectAMINE Plus (Life Technologies, Inc.) according to the manufacturer's instructions. To prepare 293 cells stably expressing APPsw, 293 cells were transfected as described above. About 24 h after transfection, cells were passed to media containing 0.4 mg/ml G418, and G418resistant clones were analyzed for A␤ secretion by ELISA (see below). Clones secreting high levels of A␤ were expanded and maintained in media supplemented with 0.2 mg/ml G418.
ALLN Treatment-APPsw cells and cells transiently expressing C100 or SPC100 were treated with various concentrations of ALLN for 16 h. The conditioned media were then collected, centrifuged at 10,000 ϫ g for 5 min to remove cell debris, and stored at Ϫ20°C prior to ELISA and Western blot analysis. The cell monolayers were washed with cold phosphate-buffered saline and stored frozen at Ϫ20°C prior to Western blot analysis.
FIG. 1. Protein expression and A␤ secretion in cells expressing C100 and SPC100. A, schematic diagram of the C100 and SPC100 expression constructs. B, Western blot analysis of lysates prepared from cells expressing C100 and SPC100. The Western blot was probed with antibody W02 as described under "Experimental Procedures." C and D, concentration of A␤40 (C) and A␤42 (D) in the conditioned media from cells expressing C100 and SPC100. A␤40 and A␤42 were quantitated by ELISA assay as described under "Experimental Procedures." The data shown are from one transfection and are representative of more than five transfections.
Western Blot-APPs␤ was detected in conditioned media by Western blot analysis with antibody 54. C100 and SPC100 were identified in cell lysates with antibody W02. Visualization was performed with an ECL kit (Amersham Pharmacia Biotech) according to the manufacturer's procedure.
ELISA Analysis of A␤ Peptides-Sandwich ELISA assays were developed to measure A␤40 and A␤42 using the combination of antibodies G2-10/W02 and G2-11/W02, respectively. Both antibody G2-10 and G2-11 are more than 100-fold selective for A␤40 and A␤42, respectively (14), and the sensitivity of these assays are about 50 -100 pg/ml. Briefly, Nunc MaxiSorb immunoassay plates were coated overnight at 4°C with 0.4 g/well G2-10 in 100 mM NaHCO 3 (pH 9.5) or with 1 g/well G2-11 in 100 mM Tris-HCl (pH 7.4). Subsequently, the antibody solution was removed, and the wells were incubated overnight at 4°C with 5% bovine serum albumin in 20 mM Tris-HCl (pH 7.4), 150 mM NaCl (TBS). The wells were then washed with TBS plus 0.05% Tween 20 (TTBS) and were stored at 4°C for up to 6 months. Conditioned media were diluted with 10% bovine serum albumin in TBS to yield a final concentration of 2% bovine serum albumin, and 100 l of diluted media was added to each well along with 40 ng/well biotinylated W02. Biotinylation of W02 was performed with the EZ-Link TM Sulfo-NHS-LC-Biotinylation kit (Pierce) according to the manufacturer's instructions. The plate was incubated at 4°C with gentle shaking either overnight (for A␤40 measurement) or for at least 24 h (for A␤42 measurement). The plate was then washed five times with TTBS, and 100 l of 0.5 g/ml horseradish peroxidase-conjugated NeutrAvidin (Pierce) was added to each well and incubated at room temperature for 1 h. The color was developed with the TMB-H 2 O 2 system (Kirkegaard & Perry Laboratories, Gaithersburg, MD) according to the manufacturer's instructions, and absorption at 450 nm was measured on a plate reader.
Metabolic Labeling of C100 with [ 35 S]Methionine-293 cells were seeded in 60-mm dishes and transiently transfected with C100 as described above. About 40 h after transfection, cells were incubated in methionine-and cysteine-free DMEM medium for 1 h and were then labeled for 1 h with 150 Ci/ml [ 35 S]methionine. The cells were subsequently washed and either kept frozen at Ϫ20°C (pulse) or incubated with fresh complete DMEM (chase) in the absence and presence of 25 M ALLN. The chase media and cell monolayers were collected at different time points and kept frozen until further analysis. C100 and A␤ peptide were immunoprecipitated with antibody W02 from radiolabeled cells and chased media, respectively. The cells were solubilized with 0.6 ml/dish of 1ϫ RIPA buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% deoxycholate, 1% Triton X-100, 0.1% SDS), and 0.5 ml of cell lysate was used for each immunoprecipitation assay with 3 g of anti-

FIG. 2. Modulation of A␤ production by APP and PS1 FAD mutations in cells expressing C100 and SPC100.
Expression vectors containing cDNAs encoding C100, 100-lon, SPC100, or SPC100-lon) were transiently transfected into 293 cells together with either an empty expression vector or an expression vector containing cDNAs encoding PS1wt or PS1⌬E9. About 24 h after transfection, the cells were fed with fresh media, and conditioned media were collected the next day. The concentrations of A␤40 and A␤42 in the conditioned media were determined by ELISA assay as described under "Experimental Procedures." The cell monolayers were washed with icecold PBS and stored at Ϫ20°C before analysis of protein expression by Western blot as described under "Experimental Procedures." A, A␤42:A␤40 ratio in the conditioned media. B, Western blot analysis of PS1 expression with antibody 14. C, Western blot analysis of C100 expression in the cell lysates with antibody W02. The data shown here are from one experiment and are representative of three independent transfections. body W02. For immunoprecipitation of A␤ from chased media, 0.2 ml of 5ϫ RIPA was added to 1 ml of media together with 3 g of antibody W02. Forty microliters of Protein G plus Protein A-agarose (Calbiochem) was added to each immunoprecipitation reaction, and the mixtures were rocked overnight at 4°C before being centrifuged at 10,000 ϫ g for 2 min. The pellets were then washed twice with 1ϫ RIPA buffer and once with 10 mM Tris-HCl (pH 7.5). The immunocomplexes were denatured in SDS-polyacrylamide gel electrophoresis sample buffer and boiled for 5 min before being processed by electrophoresis and autoradiography.

RESULTS
Expression of C100 and SPC100 -In addition to full-length APP with the Swedish mutation, two truncated APP constructs were used in this study to examine the processing of APP by ␥-secretase (see Fig. 1A). The construct designated as C100 consists of the methionine initiation codon plus the C-terminal 99 amino acids of APP. The N terminus of C100 corresponds to the ␤-secretase cleavage site, i.e. the N terminus of the A␤ peptide. The construct designated SPC100 consists of the methionine initiation codon, the 16 amino acid signal peptide of APP, and the first two amino acids of APP appended to the N terminus of C100. C100 and SPC100 are similar to the constructs previously reported by Dykes et al. (17) except that the expression vector pcDNA3.1 was used instead of pCEP4. Western blot analysis of extracts from 293 cells transiently expressing SPC100 detected a 14-kDa band that co-migrated with native C100 (Fig. 1B). These data confirm that the signal peptide of SPC100 was removed. As reported previously (21), the level of expression of C100 was much higher in cells transiently expressing SPC100 than in cells transiently expressing C100 (Fig. 1B).
293 cells transiently expressing SPC100 secreted 2-3-fold more A␤40 and A␤42 than cells transiently expressing C100 (Fig. 1, C and D), consistent with a previous report (21). The increased A␤ secretion from cells expressing SPC100 relative to cells expressing C100 is probably due to the higher level of C100 in these cells (Fig. 1B). Despite the quantitative differences in A␤ secretion from cells expressing C100 and SPC100, the relative amounts of A␤42 and A␤40 secreted by cells expressing the two constructs (i.e. the A␤42:A␤40 ratios) were similar ( Fig. 2A). The effects of APP and PS1 FAD mutations on A␤ secretion from cells expressing C100 and SPC100 were also tested. As reported previously (16), extracts from 293 cells expressing wild type PS1 displayed a 48-kDa product corresponding to full-length PS1 as well as a 33-kDa N-terminal fragment of PS1 (Fig. 2B). Extracts from 293 cells expressing the PS1 mutant PS1⌬E9 displayed a full-length PS1 protein with slightly greater mobility than that of full-length wild type PS1 due to the deletion of exon 9 in this mutant (Fig. 2B). The 33-kDa N-terminal fragment of PS1, which is derived primarily from endogenous PS1, was not increased significantly by overexpression of PS1wt or PS1⌬E9. As previously reported for full-length APP (4), the secretion of A␤42 from cells expressing either C100 or SPC100 was selectively increased by co-expression of PS1⌬E9 or by introduction of the London mutation into C100 or SPC100 ( Fig. 2A). Co-expression of PS1⌬E9 did not alter A␤40 secretion from cells expressing either C100 or SPC100, nor did co-expression of wild type PS1 affect A␤ secretion from cells expressing either construct (data not shown). Co-expression of wild type PS1 or PS1⌬E9 did not affect the levels of C100 protein in cells expressing either C100, SPC100, C100-London, or SPC100-London ( Fig. 2C and data not shown). Thus, ␥-secretase processing of C100 derived from either the C100 or SPC100 constructs is qualitatively similar as evidenced by the similar relative secretion of A␤40 and A␤42 (i.e. the A␤42:A␤40 ratio) and the similar effect of APP and PS1 FAD mutations on A␤ secretion.

Modulation of A␤ Production by ALLN-ALLN and related
peptide aldehyde protease inhibitors have multiple effects on APP processing. In addition to modulating A␤ production, it has been observed that these inhibitors can also potentiate the ␣-secretase pathway and enhance the secretion of APPs␣ (5,8). The altered ␣-secretase processing may indirectly affect the ␤-secretase pathway as the two proteases compete for a common substrate. Consequently, the effect of ALLN on A␤ production in cells expressing APP is complex, and it is very difficult to distinguish the specific action of ALLN on the ␥-secretase cleavage step. Since C100 is the product of ␤-secretase cleavage and is the native substrate for ␥-secretase (1), A␤ production by cells expressing C100 or SPC100 reflects only ␥-secretase processing. We therefore utilized cells expressing C100 and SPC100 to study the effect of ALLN on the ␥-secretase cleavage reaction without the influence of this inhibitor on ␣and ␤-secretase processing. In cells expressing C100 or APPsw, both A␤40 and A␤42 secretions were increased by treatment with low concentrations of ALLN and decreased by treatment with high concentrations of ALLN (Fig. 3, A and C). A␤42 secretion from cells expressing C100 or APPsw was increased much more dramatically by low concentrations of ALLN than was A␤40 secretion (Fig. 3, A and C). In contrast, ALLN had only concentration-dependent inhibitory effects on A␤40 and A␤42 secretion from cells expressing SPC100 (Fig. 3B).
Examination of cell lysates revealed that ALLN increased the cellular level of C100 in a concentration-dependent manner in cells expressing C100 and APPsw, whereas it did not affect C100 protein levels in cells expressing SPC100 (Figs. 4 and 5A). The ALLN-induced increase in C100 in APPsw cells was not due to increased ␤-secretase activity since ALLN did not potentiate the secretion of APPs␤ from these cells (Fig. 5B). The increase in C100 protein level induced by ALLN in cells expressing C100 or APPsw was not a result of inhibition of ␥-secretase since concentrations of ALLN that increased cellular C100 protein levels also increased both A␤40 and A␤42 production. In addition, pulse-chase experiments demonstrated that at concentrations of ALLN that increased A␤ production (data not shown), this compound decreased the rate of C100 turnover in cells expressing C100 (Fig. 6), suggesting that ALLN increases C100 protein levels by inhibiting C100 degradation. DISCUSSION In cells expressing APP, the protease inhibitor ALLN has recently been shown to inhibit selectively the production of A␤40 at concentrations that have little effect on the production of A␤42 (5,9). These data were interpreted as indicating that FIG. 4. Effect of ALLN on cellular C100 levels in cells expressing C100 and SPC100. 293 cells transiently expressing C100 or SPC100 were incubated with various concentrations of ALLN as described in Fig. 3. After conditioned media were collected, cell monolayers were washed with cold PBS and lysed with SDS sample buffer. C100 levels in the cell lysates were determined by Western blot analysis with antibody W02 as described under "Experimental Procedures." Lanes 1-6, cell lysates from cells expressing C100 treated with the indicated concentrations of ALLN; lanes 7-12, cell lysates from cells expressing SPC100 treated with the indicated concentrations of ALLN. distinct ␥-secretases are responsible for A␤40 and A␤42 production. Although these data are intriguing, a recent study suggests that the effects of ALLN on A␤ secretion from cells expressing APP are complex (10). To elucidate the mechanisms by which ALLN modulates A␤ production and, by inference, ␥-secretase activity, this study examined the effects of ALLN on A␤ production in more detail.
The ␥-secretase reaction was studied in 293 cells expressing either APPsw or various amounts of C100, the C-terminal fragment of APP that represents the immediate substrate of this enzyme. C100 was produced in 293 cells by expression of two constructs designated as C100 and SPC100 (Fig. 1). As reported previously, much higher levels of C100 were produced in cells expressing SPC100 than in cells producing C100 (21), presumably because the signal peptide present in SPC100 permits more efficient processing or sorting of the protein. The characteristics of A␤ production in cells expressing APPsw, C100 or SPC100 cells were similar. The relative amounts of A␤40 and A␤42 secreted by cells expressing C100 or SPC100 were similar as reflected by the similar A␤42:A␤40 ratios, and like the APP expression cells (4,18), the secretion of A␤42 was specifically increased by co-expression of mutant PS1 or by introduction of the London mutation into the constructs (Fig.  2). These observations argue that the same ␥-secretase(s) is(are) responsible for A␤ production in cells expressing all three constructs.
Treatment of cells expressing C100 or APPsw with low concentrations of ALLN resulted in increased secretion of both A␤40 and A␤42 (Fig. 3, A and D), which is consistent with results reported by others (10). Interestingly, treatment of cells expressing either C100 or APPsw with low concentrations of ALLN also increased cellular C100 protein accumulation (Figs. 4 and 5A). Pulse-chase experiments in cells expressing C100 demonstrated that a low concentration of ALLN decreased the rate of C100 turnover (Fig. 6), whereas A␤ secretion was increased during the same period (data not shown). Taken together, these data suggest that low concentrations of ALLN increase A␤ production by inhibiting C100 turnover and, hence, increasing the amount of C100 substrate available for ␥-secretase cleavage (Fig. 7). This suggestion is supported by the observation that ALLN did not affect cellular C100 levels in cells expressing SPC100 and correspondingly did not increase A␤ secretion from these cells at any concentration. One implication of these results is that in addition to cleavage by ␥-secretase, C100 is normally degraded by a distinct ALLN-sensitive pathway (Fig. 7). Channeling of C100 into this alternative, ALLN-sensitive degradation pathway would prevent A␤ production (Fig. 7). The fact that ALLN did not increase cellular C100 levels in cells expressing SPC100 may be due to the fact that the ALLN-sensitive degradation pathways is overwhelmed by the much higher levels of cellular levels of C100 present in these cells. Low concentrations of ALLN also increase A␤ secretion from primary hippocampal cultures where only endogenous APP is expressed, 2 suggesting that the ALLNsensitive degradation of this APP intermediate represents a normal metabolic process and is not merely an artifact of overexpressing C100 in cultured cells. The presence of an ALLNsensitive catabolic pathway for C100 may provide a mechanism by which cells regulate substrate availability for ␥-secretase and thus regulate cellular A␤ production. In this regard, the regulation of C100 metabolism may play an important role in AD pathogenesis.
An alternative mechanism has been proposed to explain the ability of ALLN to increase A␤ secretion (22). Several protease inhibitors, including ALLN, are known to prevent the proteasome-mediated degradation of PS1 (22,23). Based on this observation and the fact that presenilin mutations are associated with increases in A␤42 secretion, it has been suggested that the stabilization of presenilins by protease inhibitors like ALLN may directly potentiate A␤42 production in cells (22). Our data do not support this model for two reasons. First, the effects of ALLN and PS1 mutations are additive, i.e. low concentrations ALLN further increase the A␤42:A␤40 ratio in cells expressing mutant PS1 (22). 3 Second, the fact that A␤42 production in cells expressing SPC100 can be increased by co-expression of a mutant PS1 ( Fig. 2A), but not by ALLN (Fig. 3B), provides additional evidence that PS1 mutations and ALLN regulate A␤42 production by independent mechanisms.
High concentrations of ALLN inhibited A␤40 and A␤42 production by cells expressing either APPsw or C100. High concentrations of ALLN may directly inhibit ␥-secretase and, hence, decrease A␤ production, but other effects of ALLN on cellular metabolism could also be responsible for this effect. In cells expressing SPC100, which gives rise to much higher cellular C100 levels than in cells expressing APPsw or C100, ALLN inhibited A␤40 and A␤42 production at all concentrations and the IC 50 values of ALLN for inhibition of A␤40 and A␤42 production were very similar. This observation does not support the conclusion that ALLN selectively inhibits A␤40 production (5,9), a finding that was interpreted as implying the existence of distinct ␥-secretases responsible for A␤40 and A␤42 production. Since ALLN modulates A␤ production by multiple mechanisms, as documented in this study and Ref. 10, it is difficult to use this compound to pharmacologically distinguish multiple ␥-secretases.
Whereas low concentrations of ALLN and other calpain/ proteasome inhibitors increase A␤40 and A␤42 production, the increase in A␤42 production induced by these protease inhibitors is much more pronounced (Ref. 10, Fig. 3, A and D). Although a definitive explanation for this phenomenon cannot be discerned from the existing data, a reasonable model can be proposed. As discussed above and illustrated in the model diagrammed in Fig. 7, ALLN increases A␤ production by increasing the availability of the ␥-secretase substrate C100. Since ALLN increases A␤42 secretion more dramatically than A␤40 secretion, this model implies that the ␥-secretase cleavage reaction producing A␤42 is more dependent on substrate concentration than the reaction producing A␤40. In other words, the ␥-secretase that cleaves C100 to produce A␤42 has a higher K m for the substrate C100 than the ␥-secretase that 2 P. Fraser and L. Zhang, unpublished data. 3 L. Zhang and L. Song, unpublished data. cleaves C100 to produce A␤40. Thus, the ALLN-induced increase in the cellular concentration of the ␥-secretase substrate C100 will increase A␤42 secretion more than A␤40 secretion. The fact that A␤40 secretion was only slightly increased by ALLN would suggest that the ␥-secretase that produces A␤40 is nearly saturated with C100 under the conditions used in these experiments. This model would explain why A␤40 is the dominant A␤ species produced during normal physiologic processing of APP. It should be pointed out that this model and the experimental data that support it do not necessarily require the existence of two distinct ␥-secretase enzymes that independently produce A␤40 and A␤42. The major weakness of this model is the observation that the higher cellular levels of C100 seen in cells expressing SPC100 relative to cells expressing C100 is associated with similar increases in A␤40 and A␤42 secretion rather than a more selective increase in A␤42 secretion as would be predicted by the model. It is possible that the model is fundamentally correct but that increasing cellular C100 levels by expressing SPC100 rather than C100 is somehow biochemically or mechanistically different from increasing cellular C100 levels with ALLN. In any case, the hypothesis that substrate concentration is a more important determinant of A␤42 production than A␤40 production provides a novel framework for further experiments aimed at understanding the mechanisms regulating A␤ production.