Acute γ-Secretase Inhibition of Nonhuman Primate CNS Shifts Amyloid Precursor Protein (APP) Metabolism from Amyloid-β Production to Alternative APP Fragments without Amyloid-β Rebound

The duration of ¹³C₆-labeling does not affect Aβ production estimates, and plasma better reflects the ¹³C₆-leucine precursor for Aβ compared with CSF. A, B, Rhesus monkeys (n = 3) were infused with ¹³C₆-leucine (8 mg/kg/h) for either 21 h (black, solid circle) or 12 h (red, solid triangle). The ratio of ¹³C-Aβ to ¹²C-Aβ from CSF was measured by LC-MS, and normalized to the ¹³C₆-leucine precursor enrichment measured in either CSF (A) or plasma (B). Values >1 indicate incorrect estimates. Error bars represent SEM.

Rhesus monkey CNS Aβ metabolism is similar to results observed in humans (Bateman et al., 2006). A, Optimized ¹³C₆-leucine labeling protocol was 4 mg/kg/h for 12 h, and CSF collected for 39 h (n = 12). B, Rates of Aβ production and clearance are balanced in the rhesus monkey CNS. Combining results from three studies using identical protocols, the production and clearance rates were found to be consistent across n = 12 rhesus monkeys. There was no significant difference between the means observed for FSR and FCR, as measured by a Student's t test. Error bars indicate SEM.

Plasma levels of Aβ during and after γ-secretase inhibition indicate peripheral Aβ rebound in rhesus monkeys treated with a CNS-penetrant GSI. A, Concentration of GSI in rhesus monkey plasma after dosing with GSI. B, Aβ levels in rhesus monkeys (n = 6), measured by ELISA are shown normalized to baseline. A dose–response effect on Aβ levels was observed with 60 mg/kg and 240 mg/kg. Plasma Aβ levels in the 60 mg/kg dose group recovered to baseline after 24 h, and then increased above baseline. Means between vehicle and 60 mg/kg from hours 33–48 are significantly different (paired t test, ***p < 0.0001). C, D, In an extended study (n = 6, 3/group), animals were treated with either vehicle or 240 mg/kg GSI. C, Concentration of GSI in rhesus monkey plasma after dosing with 240 mg/kg GSI. D, Plasma Aβ levels overshoot baseline after 48 h in 240 mg/kg GSI-treated animals (unpaired t test, ***p < 0.0001, h48–240). (vehicle, black circle; 60 mg/kg, blue square; 240 mg/kg, red triangle). Error bars indicate SEM.

No evidence of central Aβ rebound in rhesus monkeys treated with a CNS-penetrant GSI. A, Concentration of GSI in rhesus monkey CSF after dosing. B, C, Aβ levels in rhesus monkeys (n = 6), measured by ELISA, were normalized to baseline. A dose–response effect on Aβ levels was observed with GSI treatment. A rebound effect was not observed in levels of CSF Aβ40 (B) or Aβ42 (C). After 48 h, the Aβ levels with 240 mg/kg treatment only recovered to 50% of baseline, while the 60 mg/kg treatment group reached baseline at 30 h without overshoot. D–F, In an extended study (n = 6, 3/group) Aβ levels returned to baseline within 72 h of treatment (240 mg/kg), without rebound. (vehicle, black circle; 60 mg/kg, blue square; 240 mg/kg, red triangle). Error bars indicate SEM.

Newly generated Aβ was reduced in response to GSI treatment in the CNS of rhesus monkeys (n = 6). A, C, D, In a crossover study, male rhesus monkeys (n = 6) were infused with ¹³C₆-leucine (4 mg/kg/h) for 12 h, and treated with vehicle (black circle), 60 mg/kg GSI (blue square), or 240 mg/kg GSI (red triangle). B, In an extended study (n = 6, 3/group), animals were treated with vehicle or 240 mg/kg GSI. A, B, Generation of new Aβ was partially blocked with administration of 60 mg/kg, and nearly completely blocked at the 240 mg/kg dose as indicated by the dose-dependent decrease in the amount of ¹³C₆-leucine-labeled Aβ (LC-MS). C, D, AUC analysis of newly generated Aβ indicates that the means of newly synthesized Aβ are significantly reduced by GSI treatment during the leucine infusion (0–12 h) (C) and levels do not recover within 48 h (D), as assessed by a repeated measures 1-way ANOVA (p < 0.0001) and post hoc analysis (Tukey) (*p = 0.01 to 0.05;**p = 0.001 to 0.01; ***p < 0.001). Error bars indicate SEM.