Transient pharmacologic lowering of Aβ production prior to deposition results in sustained reduction of amyloid plaque pathology

Ly-411,575 is potent orally bioavailable γ-secretase inhibitor [12]. We synthesized large quantities for these studies and then performed both in vitro and in vivo studies to validate Ly-411,575 potency. The IC50 in cell culture for LY-411,575 was between 1–3 nm (see Additional file 1: Figure S1A). Six hours after a single IP dose, Aβ40 levels are reduced by ~75% in the brain of non-depositing Tg2576 mice (Additional file 1: Figure S1B). Oral administration (10 mg/kg/day) administered as a single dose reduces brain Aβ40 levels 24 hours after the final dose (7 days) by ~65% (Additional file 1: Figure S1C) and plasma Aβ40 by ~95% (Additional file 1: Figure S1D). Plasma Aβ levels are reduced more quickly than brain levels following oral dosing with maximal inhibition seen after 3 days (Additional file 1: Figure S1D). For long term in vivo mouse studies, we tested administration of Ly-411,575 using various dosing strategies and time points (Table 1). Oral administration (formulated in rodent chow) continually for 2 weeks with either - 1 mg/kg/day or 2.5 5 mg/kg/day reduced brain Aβ40 levels by ~50% and plasma Aβ40 by ~80%, without overt toxicities ( Table 1). However, continuous oral dosing (>2 weeks) using the higher dose of 5 mg/kg/day or 10 mg/kg/day resulted in overt toxicities in mice (Table 1). Therefore, in all long term studies presented in this manuscript, we have used the 2.5 mg/kg/day dose formulated in rodent chow.

For these experiments, we transiently dosed Tg2576 mice with LY from 4-7M, 7-10M or 12-15M and aged treated mice to 15M (see Figure 1B for experimental design). To confirm efficacy of LY in all groups, we measured plasma Aβ levels immediately after the last day of treatment in sentinel mice and showed a similar reduction in plasma Aβ levels in all cohorts (Table 2). Following sacrifice at 15M, levels of Aβ in the brain were examined. The 4-7M LY treatment significantly reduced Aβ deposition; plaque burden in the frontal cortex and hippocampus was decreased by ~68% (Figure 1C) and FA-solubilized brain Aβ levels were reduced by ~60% (Figure 1D). The 7-10M LY treatment non-significantly decreased Aβ plaque burden by ~19% compared to controls (Figure 1C), whereas FA-solubilized Aβ levels were reduced by ~34% (Figure 1D). The 12-15M GSI treatment group had no significant effect on either plaque burden or FA-solubilized Aβ levels (Figure 1C, D). We further evaluated effect of the 4-7M LY treatment on cored amyloid plaques and cerebral amyloid angiopathy (CAA). Compared to untreated control mice, there were significant reductions both in cored plaques in the frontal cortex and hippocampus (~47% reduction) and CAA in the leptomeninges (~41% reduction, Figure 2A, B). To further determine whether the magnitude of the effect diminished with further aging, we performed an additional experiment, where we aged 4-7M LY treated mice to 18M, and again we observed significantly reduced Aβ plaque burden and FA-solubilized Aβ levels (Figure 2C), demonstrating that the suppression of deposition was maintained even up 11 months after treatment was halted.

Because Aβ deposition in mice follows an exponential course, we sought to determine how the observed reductions in Aβ accumulation correlate with Aβ accumulation to untreated mice. We compared Aβ levels in the brains of untreated Tg2576 mice from various ages ranging from 10–18 months and plotted the natural logarithm transformed value of these against Aβ levels from the 4-7M LY treated mice, aged to 15M and 18M (Figure 2D). To minimize variance, we assayed all of the samples in a single set of ELISA studies. This analysis showed that the 4 -7M GSI treatment time frame “shifts” amyloid depositing by approximately 3 months, essentially delaying Aβ accumulation by a time roughly equivalent to the treatment window.

We next attempted to determine whether this 4-7M narrow therapeutic “window” can be further refined. To do this, we transiently dosed Tg2576 mice with LY for 1, 2 or 3M intervals between the 4-7M age time frame and again aged the mice to 15M. Aβ levels were again significantly reduced in this second 4-7M LY cohort (~59% reduction) at 15M of age compared to controls (Figure 2E). Similarly, i) the 4-6M LY cohort showed a significant reduction in Aβ levels (~51% reduction) and; ii) the 4-5M LY group showed a modest but non-significant reduction in Aβ levels (~ 24% reduction) (Figure 2E). However, in the remaining 5-6M, 6-7M or 5-7M LY cohorts tested, none showed any significant reductions in Aβ levels at 15M compared to controls (Figure 2E). Since we show that the 4-7M LY “shifts” amyloid deposition equivalently to the duration of treatment, we wondered whether we will see a similar shift in amyloid deposition in the shorter treatment intervals examined that showed efficacy, e.g., in the 4-5M and 4-6M cohorts. Again, we compared Aβ levels in the brains of untreated Tg2576 mice from various ages ranging from 10–15 months against Aβ levels from the 4-5M, 4-6M, and 4-7M LY treated cohorts aged to 15M (Figure 2F). This analysis seems to show that the 4-5M GSI treatment “shifts” amyloid depositing by approximately 1 month, however, both the 4-6M and 4-7M treatment intervals essentially show a similar shift in Aβ accumulation, namely by roughly 3 months (Figure 2F). These data would suggest, at least within the 4-7M time frame that we have studied, that the 4-6M age interval may play a critical role in the early seeding phases of amyloid deposition. Additional studies, using either earlier treatment times (e.g. starting at 3M of age) or longer treatment intervals may be necessary to fully appreciate the usefulness of this early “therapeutic window” to achieve the best possible outcome on subsequent amyloid deposition.

To determine if the 4-7M LY treatment had long-lasting effects on Aβ production or APP levels and processing, we conducted a number of studies. We measured plasma Aβ levels at termination of GSI treatment and then determined how long before steady state levels of Aβ were normalized. Plasma Aβ levels were significantly lower immediately after treatment, but returned to control levels between with 1–2 weeks after treatment was halted (Figure 3A). We also measured levels of full-length APP and APP C-terminal fragments (CTFs) in the brain at the end of the 4-7M treatment and at 15M. Immediately following discontinuation of treatment, APP levels were unchanged but APP CTFs were increased (Figure 3B,D). However, at 15 months, there were no significant differences in full length APP or CTFs (Figure 3C,D). Finally, as GSI treatment has been shown to affect peripheral lymphocytes [13], we performed FACS analysis and quantified both B and T cell numbers in the spleen after the 4–7M LY transient treatment and at 15M. B and T cell numbers were not affected (Figure 4A). However, when we analyzed Th1 and Th2 cytokine profiles, splenic CD4+ T cells from the 4-7M LY treated cohort showed a bias towards Th1 polarization with increased IFN-γ secretion (a canonical Th1 cytokine) and concurrent decreases in IL-4 levels (a signature Th2 cytokine), which persisted even after the treatment was halted for 8 months (Figure 4B, C,). The significance of this long-lasting effect on peripheral CD4 + T cell immune responses is not clear, but certainly warrants further investigation.

APP Tg2576 mice were generated, maintained and genotyped as described previously [7]. All animal husbandry procedures performed were approved by the Mayo Clinic Institutional Animal Care and Use Committee in accordance with National Institutes of Health guidelines. All animals were housed three to five to a cage and maintained on ad libitum food and water with a 12 h light/dark cycle.

LY-411, 575 (LY) was synthesized at the Mayo Clinic Chemical Core as previously described [15]. LY at a concentration of 16.75 mg per kg of diet was homogenously incorporated into Harlan Teklad 7012 kibble chow by Research Diets, Inc, New Jersey). Based upon dietary consumption at this age, these diets were designed to deliver 2.5 mg of LY on average per day. Tg2575 mice (n = 7-15/group/experiment) were transiently treated with LY chow at 1, 2 or 3 months interval (between 4–7 months of age), from 7–10 months of age and from 12–15 months of age. Untreated control mice from each age group were as aged and used for analysis. Groups of mice (n =3) were sacrificed immediately after the end of each treatment time-point to access efficacy of LY inhibition. After treatment was halted, mice were then aged till 15 months or 18 months of age and then sacrificed for Aβ analysis. Chow consumption, general health and body weight were monitored on a weekly basis in all treatment groups.

Snap-frozen hemi-forebrains from Tg2576 mice were two-step sequentially extracted in 2% SDS buffer followed by 70% formic acid (FA) as described previously [16]. Aβ levels from brain lysates or plasma were the measured using sandwich ELISA techniques as described previously [16] with end specific mAbs 2.1.3 (human Aβx-42 specific, Mayo) and mAb 13.1.1 (human Aβx-40 specific, Mayo) for capture and HRP-conjugated mAb Ab5 (human Aβ1–16 specific, Mayo) for detection.

Sagittal sections of paraffin embedded sections were used for analysis. Immunohistochemical staining was done using pan-Aβ antibody (mAb 33.1.1; Mayo), mAb 13.1.1 (human Aβx-40 specific, binds cored plaques and CAA; Mayo) as previously described [17]. Immunohistochemically stained sections were captured using the Scanscope XT image scanner (Aperio, Vista, CA, USA) and analyzed using the ImageScope program. Aβ plaque burden was then calculated using the Positive Pixel Count program (Aperio). At least 5 sections/brain 30 μm apart, were used and averaged by a blinded observer to calculate plaque burden.

2% SDS solubilized brain lysate samples were separated on Bis-Tris 12% XT gels (Bio-Rad, Hercules, CA, USA) and probed with the antibody CT20 (anti-APP C-terminal 20 amino acid; 1:1000; P.D) and anti-actin (1:1000; Sigma, St. Louis, MO, USA). Band intensity was quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Single cell suspensions were prepared aseptically from spleens from control or GSI -treated mice. CD4+ T cells were positively selected using the BD-IMAG system and CD4 + DM particles (BD Biosciences, San Diego, CA), according to the manufacturer’s directions. Twelve-well tissue culture plates were precoated with anti-hamster Ig (Sigma), then with antibodies specific for CD3ε and CD28. Equal numbers of cells (3x10⁶) were added to each well. For TH1 polarizing conditions, anti-IL4 (10μg/ml) plus IL-12 (1 ng/ml) was added to culture wells; for TH2 polarizing conditions, anti-IFNγ (10μg/ml) plus IL-4 (1 ng/ml) was added to culture wells. No additional reagents were added to cells cultured under neutral conditions. After 24 hours, all wells were pulsed with IL-2 (10 ng/ml). After 96 hours, supernatants were collected and frozen at -80C until analyzed by ELISA for cytokine secretion. IFNγ and IL-4 cytokine secretion was measured using ELISA kits per manufactures instructions (BD Pharmingen). For cell surface staining, approximately 1x10⁶ cells (each antibody) were stained with the following antibodies individually: IgD-FITC, biotinylated anti-CD19 plus strep-PE, CD4-FITC, CD8-PE (BD Pharmingen). Data were analyzed using an LSRII flow cytometer (BD Biosciences) and FACSDiva software for acquisition and analysis.