In Vivo Characterization of Aβ(40) Changes in Brain and Cerebrospinal Fluid Using the Novel γ-Secretase Inhibitor N-[cis-4-[(4-Chlorophenyl)sulfonyl]-4-(2,5-difluorophenyl)cyclohexyl]-1,1,1-trifluoromethanesulfonamide (MRK-560) in the Rat

The neuropathology of AD is characterized by extracellular protein deposits in the brain parenchyma known as plaques, along with intracellular neurofibrillary tangles, which are comprised of hyperphosphorylated tau protein. The plaques are mainly made up of the Aβ proteins Aβ(1-40) and (1-42), cleavage products of the APP of which the majority is the more fibrillogenic (1-42) form (Selkoe, 2001). These observations, along with a number of mutations in the APP gene that lead to early onset familial AD, give evidence of the direct involvement of aberrant APP processing in AD (Hardy, 1997).

Consequently, much effort has focused on ways to inhibit the production of Aβ. The enzymes responsible for processing APP into Aβ are the aspartyl proteases, β-site APP cleaving enzyme (β-secretase) and γ-secretase (Churcher and Beher, 2005). Although an attractive target for drug discovery, inhibition of β-secretase has proved challenging in terms of identification of small molecules for therapeutic use (Selkoe and Schenk, 2003; Middendorp et al., 2004). An alternative strategy has been to inhibit γ-secretase, an enzyme complex composed of at least four different protein subunits: presenilin (an aspartyl protease, mutations of which are associated with familial AD), nicastrin, APH-1, and PEN-2. γ-Secretase is responsible for the intramembraneous proteolytic cleavage of the C-terminal fragment of APP, resulting in mainly Aβ(40) or Aβ(42) production (for review, see Selkoe and Schenk, 2003; Haass, 2004).

There have been a number of reports of novel, bioavailable γ-secretase inhibitors that have been tested in both transgenic and nontransgenic animal models (Dovey et al., 2001; Lanz et al., 2003, 2004; Anderson et al., 2005; Barten et al., 2005; Best et al., 2005; Grimwood et al., 2005). From these studies, it was established that γ-secretase inhibitors are able to reduce Aβ in both overexpressing and physiological models. The γ-secretase inhibitor LY-411575 was able to reduce Aβ in both the transgenic Tg2576 mice (which harbor the Swedish familial AD APP mutation) (Lanz et al., 2004) and the physiological rat model (Best et al., 2005). Furthermore, BMS-299897 was reported to be able to reduce Aβ in Tg2576 mice (Barten et al., 2005), APP-YAC mice (which overexpress the normal human APP gene), and guinea pigs (Anderson et al., 2005).

We have previously reported a potent novel class of bioavailable γ-secretase inhibitor, MRK-560 (compound 32; Churcher et al., 2006). In SH-SY5Y neuroblastoma cells, this compound inhibited the production of Aβ(40) and Aβ(42) with similar in vitro IC₅₀ values in the range of 0.65 nM and reduced Aβ(40) in the APP-YAC mouse model with an ED₅₀ of 1.2 mg/kg. The aim of the present study was to further characterize the in vivo effects of MRK-560 in the rat and to extend these investigations to examine the effects of MRK-560 on CSF Aβ(40) concentrations to determine whether changes in CSF Aβ(40) reflect changes in the brain Aβ(40).

Materials and Methods

Chemicals. Diethylamine, phosphate-buffered saline, bovine serum albumin, and Tween 20 were purchased from Sigma Chemical (Poole, Dorset, UK). Protease inhibitors were obtained from Roche Diagnostics (Mannheim, Germany). Biotinylated antibody 4G8 was purchased from Signet Laboratories (Dedham, MA).

Chemistry. MRK-560 (Fig. 1) was prepared by methods described previously (Churcher et al., 2006).

Animals. All procedures were conducted in accordance with the Animals (Scientific Procedures) Act of 1986 and its associated guidelines. All animals were maintained on a 12:12-h light/dark cycle with unrestricted access to food and water until use.

Pharmacokinetic Analysis of MRK-560. Male Sprague-Dawley rats (350 g; Charles River, Manston, Kent, UK) had their food withdrawn overnight. MRK-560 was administered as a bolus injection of 1 mg/kg intravenously [1 ml/kg as a solution in 3:1 polyethylene glycol 300/water (v/v)] and orally (5 ml/kg as a suspension in 0.5% methylcellulose in water). Serial blood samples were collected at specified time points up to 24 h after dosing. Plasma was separated by centrifugation, and the samples were stored at -80°C before analysis. The plasma samples were thawed, processed, and analyzed as described previously (Best et al., 2005).

Dosing and Tissue Collection. Male Sprague-Dawley rats (250-300 g; Charles River) were dosed orally as a suspension in 0.5% methylcellulose at 1 ml/kg. In the first study, a dose response was conducted at 8 h after administration of MRK-560 (1, 3, 10, 30, and 100 mg/kg p.o.). In the second study, animals received a single dose or were dosed once a day for 2 weeks with 45 mg/kg, and tissue was collected at 4, 8, 16, and 24 h after the final dose. Animals were anesthetized using isoflurane. CSF was removed by puncturing the cisterna magna with a 21-gauge butterfly cannula, and then they were euthanized by decapitation. Brains were removed and, along with the CSF, immediately frozen on dry ice and stored at -80°C until use. CSF samples with visible blood contaminants were discarded. After decapitation, blood was collected into EDTA-coated Vacutainers (Becton Dickinson UK Ltd., Oxford, UK) and spun at 3000 rpm for 10 min; plasma was collected, and both samples were frozen at -80°C until use. The plasma and brain samples were thawed, processed, and analyzed for drug concentrations as described previously (Best et al., 2005).

Tissue Sample Preparation for Aβ(40) Measurement. The frozen brains were homogenized in 10 volumes (w/v) of 0.2% Diethylamine containing 50 mM NaCl, pH 10, and protease inhibitors, and then centrifuged at 355,000g, 4°C, for 30 min (Optima MAX series ultracentrifuge; Beckman Coulter, Fullerton, CA). The resulting supernatant was retained as the soluble fraction and neutralized by addition of 10% 0.5 M Tris-HCl, pH 6.8. Samples were frozen at -80°C awaiting analysis. There was no significant difference in Aβ levels between this extraction method and the guanidinium-HCl extraction method, as might be expected in normal physiological, plaque-free animals (data not shown; Lanz et al., 2004). Before analysis, the CSF was thawed and centrifuged at 2300g, and the supernatant was diluted 1:4 with phosphate-buffered saline, 2% bovine serum albumin, and 0.5% Tween 20 plus protease inhibitors.

Measurement of Aβ(40). The monoclonal antibody G2-10 (Ida et al., 1996) was used with biotinylated antibody 4G8 (Kim et al., 1988) to detect Aβ peptides in solution ending at residue 40, with negligible cross-reactivity. These species, referred to as Aβ(40), reflect subpopulations of peptides with heterogeneous N termini encompassing at least the 4G8 epitope at residues 17 to 24. Analysis of the samples was performed using the Meso Scale Discovery Sector Imager 6000 (Meso Scale Discovery, Gaithersburg, MD), as described previously (Best et al., 2005).

Data and Statistical Analyses. Groups were analyzed using two-way and one-way analysis of variance and, where appropriate, post hoc Dunnett's t test with vehicle/control group or Bonferroni's t test between experimental groups applied (Prism 3.03; GraphPad Software Inc., San Diego, CA). For time courses, Aβ(40) level reductions were integrated using the area under the curve trapezoid rule.

Results

Pharmacokinetics of MRK-560 in Rat. MRK-560 (1 mg/kg) was administered intravenously and orally. Plasma drug concentrations were measured out to 24 h postdose. The i.v. profile (Fig. 2A) of MRK-560 revealed a low plasma clearance of <5 ml/min/kg with a volume of distribution of approximately 6 l/kg, which translated to a long half-life of >15 h, which demonstrated the compound was suitable for once-a-day dosing. The T_max after the oral dose (Fig. 2B) was 12 h, and bioavailability was 70 to 90%.

Dose-Related Effects of MRK-560 on Brain Aβ(40). To determine the effect of MRK-560 in the rat, a dose-response experiment was performed 8 h after oral dosing, which was deemed to be the most favorable time point from the oral pharmacokinetic profile (Fig. 2B). The plasma and brain concentrations of MRK-560 increased in a dose-dependent manner over the dose range tested (1-100 mg/kg; Fig. 3, A and B). Moreover, the plasma drug concentrations in this particular study (0.21 ± 0.025 μM) were comparable with those observed in the initial study (0.18 ± 0.036 μM; Fig. 2B), suggesting that oral dosing as a methyl cellulose suspension gives reproducible exposure at the lower doses.

MRK-560 caused a dose-dependent reduction in both brain and CSF Aβ(40) levels (Fig. 3C), with essentially complete inhibition of the production of both peptides being observed at a dose of 100 mg/kg. The ED₅₀ values generated for brain and CSF Aβ(40) were 6 and 10 mg/kg, respectively, demonstrating that the effect of MRK-560 on the inhibition of the production of Aβ(40) in brain and CSF were similar.

By plotting Aβ(40) levels as a function of plasma and brain drug concentrations, the drug concentrations required to inhibit brain and CSF Aβ(40) levels by 50% (EC₅₀) were calculated (Fig. 4). For plasma, these values were found to be 1.2 μM for brain and 1.7 μM for CSF (Fig. 4A), and for brain concentrations, these values were 0.13 μM for brain and 0.19 μM for CSF (Fig. 4B), again demonstrating that inhibition of CSF Aβ(40) paralleled brain Aβ(40).

Changes in Brain and CSF Aβ(40) Concentrations upon Acute and Chronic Dosing of MRK-560. To establish whether the pharmacodynamic response [i.e., Aβ(40) reductions] followed plasma pharmacokinetics and to see whether reductions in the CSF and brain Aβ(40) were maintained after chronic dosing, the time courses of CSF and brain Aβ(40) reduction were studied after a single, acute dose as well as after chronic (2-week) dosing (Fig. 5). In keeping with the oral pharmacokinetic profile of MRK-560, the brain Aβ(40) levels in the acute dose study (Fig. 5A) gradually decreased with the maximal reduction taking place between 16 and 24 h. The CSF levels of Aβ(40) initially seemed to be reduced more rapidly at 4 and 8 h than the brain levels, but because a lag was not seen in the other studies conducted here, it was not deemed significant. After these time points, the brain and CSF Aβ(40) levels followed each other closely. The overall integrated reductions across the 24-h period were very similar, with brain Aβ(40) levels being reduced by 35% and CSF levels by 42%.

The pharmacokinetic data (Fig. 2) suggested that the concentrations of MRK-560 should accumulate on repeated dosing resulting in a greater reduction in Aβ(40) levels. This effect was observed experimentally after 2-week dosing in that the levels of both CSF and brain Aβ(40) were reduced more than in the single dose study. The overall reduction of brain Aβ(40) was 88% in the 24-h period after the last dose. The overall reduction of CSF Aβ(40) was 55% in the period after the last dose.

Correlation of Brain and CSF Aβ(40). Qualitatively, changes in CSF Aβ(40) mirror those seen in brain Aβ(40) in both dose-response and time-course experiments (Figs. 3 and 5, respectively). To more specifically determine how well the reduction of Aβ(40) in the CSF paralleled the brain Aβ(40) reduction, the values for brain and CSF from the dose-response and time-course experiments were correlated using linear regression analysis (Fig. 6). The data demonstrated a significant correlation (F = 123.7; p < 0.0001).

Discussion

The involvement of Aβ in the pathogenesis of Alzheimer's disease has been demonstrated in several studies; as a result, there are currently a number of approaches to lowering Aβ levels in the central nervous system, with one of the main focuses being γ-secretase inhibition (for review, see Harrison et al., 2004). Recent studies have reported bioavailable γ-secretase inhibitors that reduce Aβ levels in both transgenic and nontransgenic species (Lanz et al., 2003, 2004; Anderson et al., 2005; Barten et al., 2005; Best et al., 2005).

We have previously reported a novel series of cyclohexyl sulfone γ-secretase inhibitors that were able to reduce Aβ levels in SH-SY5Y cells with IC₅₀ values in the nanomolar range (Churcher et al., 2006). Of these, compound 32 (MRK-560) caused significant reductions of brain Aβ(40) levels in the APP-YAC mouse model, with an ED₅₀ of 1.2 mg/kg 4 h postdosing, which compared very favorably with BMS-299897, which lowered brain Aβ(40), with an ED₅₀ of 30 mg/kg at 3 h in the APP-YAC mouse (Anderson et al., 2005). We have previously demonstrated that the rat is a good physiological model for assessing γ-secretase inhibitors (Best et al., 2005). Thus, the current study was conducted to characterize further MRK-560 in the rat and assess its effects on Aβ(40) in the brain and CSF and to investigate the relationship between Aβ(40) in the brain and CSF.

MRK-560 had a long half-life of >15 h in the rat, suggesting its suitability for once-a-day dosing. Furthermore, the fact that the levels of MRK-560 had not returned to zero after 24 h suggested that the compound would accumulate and therefore comparatively higher efficacy could be attained for a given dose when the compound reached steady-state levels. This profile contrasts with LY-411575, which had a very short half-life (2 h) in the rat (Best et al., 2005).

The oral pharmacokinetic profile suggested that the optimum time point for investigating Aβ reductions was 8 to 12 h. Thus, a dose-response experiment was conducted at 8 h and demonstrated a robust dose-dependent reduction of brain Aβ(40), with 30 mg/kg causing a complete reduction.

A number of previous studies in rodents demonstrated that changes in CSF Aβ(40) levels correlate with brain Aβ(40) levels (Lanz et al., 2004; Anderson et al., 2005; Barten et al., 2005; Best et al., 2005). As well as being able to study the response of γ-secretase inhibitors on a physiological system, another advantage of the rat is the relative ease with which CSF can be obtained, thereby allowing investigation of the relationship between changes in CSF Aβ(40) levels and the brain Aβ(40). The dose-response experiments revealed a similar reduction of CSF Aβ(40) compared with brain Aβ(40) with both the ED₅₀ and EC₅₀ being within 2-fold of each other. To determine whether this relationship was maintained upon repeated dosing of MRK-560, acute and subchronic time courses were conducted. In the acute dose study, maximal brain Aβ(40) reductions were reached at 8 h and remained fairly constant up to 24 h, reflecting the pharmacokinetics of the compound. The reduction of Aβ(40) observed at 8 h with 45 mg/kg was less than the single dose of 30 mg/kg in the dose response and is probably due to lower drug concentrations, indicating oral dosing with MRK-560 is less reproducible at higher doses. In the 2-week study, as expected, the brain and CSF Aβ(40) levels were further reduced because of the expected accumulation of MRK-560 as a result of its long half-life. The brain Aβ(40) levels were constant across the 24-h period, demonstrating MRK-560 suitability for once-a-day dosing regimes. Although not directly paralleling the brain levels, the profile of CSF Aβ(40) across the 24 h was reflective of the brain levels. When all the data from the studies were correlated, a significant relationship was seen between the brain and CSF Aβ(40) levels, confirming that CSF Aβ(40) is a good marker of brain Aβ(40) levels.

In addition to CSF-brain transport mechanisms, it has been demonstrated that rapid efflux of Aβ from brain to plasma is mediated by low-density lipoprotein receptor-related protein, which also seems to be a substrate for γ-secretase (Shibata et al., 2000; Lleó et al., 2005). A recent study in normal human volunteers demonstrated changes in plasma Aβ without effects on CSF Aβ (Siemers et al., 2005). The importance of these findings reinforces the need for an assay able to measure picomolar levels of Aβ expected in the rat plasma.

In addition to processing APP, γ-secretase mediates a number of other regulatory functions. One the best characterized is the Notch pathway, which has been implicated in peripheral organ toxicity (Searfoss et al., 2003; Wong et al., 2004). In a study of organ toxicity, it was demonstrated that LY-411575 caused thymus atrophy and deterioration of intestinal epithelium with accompanying weight loss after dosing for 2 weeks in transgenic mice (Wong et al., 2004). Furthermore, a 4-day dosing study with 10 mg/kg compound X revealed gross changes in the ileum, which correlated with abnormal histological changes in cell morphology (Searfoss et al., 2003). Preliminary visual inspection of the digestive system of the animals dosed for 2 weeks revealed no profound effects of MRK-560 on the ileum or effect on weight gain (data not shown). These initial findings suggest a more detailed analysis of effects of MRK-560 on the peripheral organ system is needed, because it seems that MRK-560 is able to attain substantial brain Aβ(40) reductions without confounding side effects in the periphery.

In summary, the current study demonstrates MRK-560 as a novel, orally bioavailable γ-secretase inhibitor able to significantly reduce the levels of Aβ(40) in the rat brain. Furthermore, these results confirm the use of CSF as a dependable biomarker for monitoring brain Aβ(40) levels and further substantiate the rat as a reliable physiological model for measuring changes of Aβ(40).