An NSAID-like compound, FT-9, preferentially inhibits gamma-secretase cleavage of the amyloid precursor protein compared to its effect on amyloid precursor-like protein 1

Abstract

Inhibition of gamma-secretase cleavage of the amyloid precursor protein (APP) is a prime target for the development of therapeutics for treating Alzheimer's disease; however, complete inhibition of this activity would also impair the processing of many other proteins, including the APP homologues, amyloid precursor-like protein (APLP) 1 and 2. To prevent unwanted side effects, therapeutically useful gamma-secretase inhibitors should specifically target APP processing while sparing cleavage of other gamma-substrates. Thus, since APLP1 and APLP2 are more similar to APP than any of the other known gamma-secretase substrates and have important physiological roles in their own right, we reasoned that comparison of the effect of gamma-secretase inhibitors on APLP processing should provide a sensitive indicator of the selectivity of putative inhibitors. To address this issue, we have optimized microsome and cell culture assays to monitor the gamma-secretase proteolysis of APP and APLPs. Production of the gamma-secretase-generated intracellular domain (ICD) occurs more rapidly from APLP1 than from either APLP2 or APP, suggesting that APLP1 is a better gamma-substrate and that substrate recognition is not restricted to the highly conserved amino acid sequences surrounding the epsilon-site. As expected, the well-characterized gamma-secretase modulator, fenofibrate, did not inhibit ICD release, whereas a related compound, FT-9, inhibited gamma-secretase both in microsomes and in whole cells. Importantly, FT-9 displayed a preferential effect, inhibiting cleavage of APP much more effectively than cleavage of APLP1. These findings suggest that selective inhibitors can be developed and that screening of compounds against APP and APLPs should assist in this process.

Figure 1

C-Terminal six-His-tagged APP, APLP1, and APLP2 are expressed and processed in a manner highly similar to that of untagged proteins. Lysates from selected clones of cell lines expressing APP-His, APLP1-His, and APLP2-His (clones D4G, D7G, and D7F, respectively) were compared with lysates of cell lines expressing untagged versions of the same proteins (clones 1B1,4H9, and 4B6 for APP, APLP1, and APLP2, respectively) and with a lysate of naive CHO cells. Lysates were analyzed with a panel of antibodies recognizing either an epitope lying in the N-terminal ectodomain (22C11 for the APP N-terminus, W1NT for the APLP1 N-terminus, and D2-II for the APLP2 N-terminus), an epitope located at the very C-terminus (C8 for the APP C-terminus, W1CT for the APLP1 C-terminus, and W2CT for the APLP2 C-terminus), or an antibody specific for the six-His C-terminal tag (Tetra-His). (a–c) Full-length proteins (top panels) are indicated by arrows (m, mature form; i, immature form), while CTFs (bottom panels) are indicated by arrowheads.

Figure 2

Characterization of the ICDivg assay. Microsomes derived from APP-His-, APLPl-His-, and APLP2-His-expressing cells were incubated for different amounts of time and ICD levels measured by immunoblotting. Representative Western blots and calculated curves describing the time dependency of ICD production for each cell line are shown (a—c). Values are given as the percentage of ICD production, where 100% is set to respective mean plateau levels. Each experimental point is the mean of at least six independent experiments with duplicate samples in each experiment. APLP1 ICDs (dashed–dotted line with gray squares) are produced faster than both APLP2 ICDs (solid line with black diamonds) and APP ICDs (dashed line with open circles) (d). The initial velocity (v₀) for the γ-secretase activity of the three substrates and the time required to achieve half of the maximal amount of ICD produced (½t_max) are indicated (d). Lysates of microsomes at time zero and after an incubation for the respective ½t_max were analyzed for the γ-secretase components PS1 CTF and nicastrin (e), and for the levels of the respective CTFs using TetraHis antibody (f). Neither the levels of γ-secretase nor CTFs changed significantly during the course of the reaction.

Figure 3

Dose-dependent inhibition of γ-secretase cleavage of APP, APLP1, and APLP2. γ-Secretase processing of APP (dashed line with open circles), APLP1 (dashed–dotted line with gray squares), and APLP2 (solid line with black diamonds) are similarly inhibited by compound E (a) and DAPT (b). Microsomes of each cell line were incubated for the appropriate ½t_max with different concentrations of inhibitor, and ICD levels were measured. For each compound, the molecular structure, representative Western blots, and calculated inhibition curves are shown. Each experimental point is expressed as the mean of at least four independent experiments with duplicate samples at each dose ± the standard error.

Figure 4

Aβ₄₂-lowering GSMs differentially inhibit γ-secretase cleavage of some substrates. Microsomes of each cell line were incubated for the appropriate ½t_max, with different concentrations of test compound, and ICD levels were measured. For each compound, the molecular structure, representative Western blots, and calculated inhibition curves are shown. Each experimental point is expressed as the mean of at least four independent experiments with duplicate samples at each dose ± the standard error. All compounds were tested in the range between 100 and 500 µM; higher concentrations were not possible because of solubility issues. Fenofibrate did not inhibit ICD production in the concentration range examined (a). R-Flurbiprofen showed a weak inhibitory effect on APP and APLP2 but did not alter APLP1 ICD production (b). FT-9 acts as a γ-secretase inhibitor, causing similar dose-dependent inhibition of the γ-secretase cleavage of APP and APLP2, but has a less potent effect on γ-secretase processing of APLP1 (c). Key: dashed line with open circles for APP, dashed–dotted line with gray squares for APLP1, and solid line with black diamonds for APLP2.

Figure 5

FT-9 preferentially inhibits γ-processing of APP in intact cells. Cells were treated overnight with subtoxic concentrations of FT-9 or compound E, and lysates were analyzed by Western blotting with TetraHis antibody (a and b). A dose-dependent accumulation of CTFs can be seen in the two cell lines for both compounds (a and b). Quantitation of the CTFs levels demonstrates that FT-9 preferentially inhibits γ-secretase processing of APP (c), whereas compound E has no preferential effect (d). Results are presented as the means of at least three measurements ± the standard error, with the fold increase in CTFs levels determined by comparison of the CTFs detected in the vehicle control. An asterisk indicates a significant difference in the fold of increased APP and APLP1 CTFs (P ≥ 0.05).