Introduction

Top Abstract Introduction Methods Results Discussion References

Many lines of evidence indicate that the A plays a pivotal roles in the progression of brain amyloidgenesis observed in AD (Haass and Selkoe, 1993; Selkoe, 1994, 1997; Games et al., 1995). This provides the rationale that the prevention of A-like deposition may be a plausible approach to halting or slowing the progression of AD (Cordell, 1994). A is derived from larger APP by proteolytic cleavage by yet unidentified endoproteolytic enzyme "secretase(s)" (Selkoe, 1994; Sisodia and Price, 1995).

A major obstacle for elucidating and treating AD has been the lack of experimental animal models that show progressive A-like deposition in the brain. The SAM was established from its progenitor strain AKR/J mice (Takeda et al., 1981, 1991). There are senescence prone strains (SAM-P) and accelerated senescence resistant strains (SAM-R). Each line of SAM-P strain exhibits characteristic pathological phenotype such as systemic senile amyloidosis (Matsumura et al., 1982) or senile cataract (Hosokawa et al., 1984). SAM-P/8 is one line of such senescence-accelerated strains, showing early onset and rapid advancement of senescence (Yagi et al., 1988; Miyamoto et al., 1992). The life expectancy is about 12 mo which is significantly shorter than that of its senescence-resistant counterpart SAM-R/1 (more than 24 mo). Available reports demonstrate that A-LI spontaneously develops in the brain of the SAM (Takemura et al., 1993; Fukunari et al., 1994). Thus, the SAM can be used as an experimental model for testing the possible effectiveness of potential therapeutic agents on senescence-accelerated amyloid formation. Furthermore, we previously reported that deposited structures that are immunostained with an antibody raised against purified PEP were also developed in the hippocampus of the SAM with close temporal and spatial relationships to A-LI (Fukunari et al., 1994), suggesting that PEP may be involved in the formation of A-like granules in SAM hippocampus.

PEP (EC3.4.21.26) is a serine endopeptidase, which is widely distributed in mammalian tissues and specifically cleaves peptide bonds at the carboxylic end of the proline residues of oligopeptides (Yoshimoto et al., 1977, 1978). PEP has unique structural and catalytic features (Rennex et al., 1991; Polg[aa]ar, 1992; Shirasawa et al., 1994; Vanhoof et al., 1994) and has been classified as a new class of serine peptidases (Rawlings et al., 1991). This enzyme can also recognize alanine residues and was identified as a putative A-generating enzyme when assayed with a synthetic model peptide as a substrate (Ishiura et al., 1990).

According to these results, we hypothesized that PEP would be functionally involved in the senescence-accelerated amyloid formation and that long-term inhibition of PEP would suppress the progression of A-like deposition in the SAM hippocampus. We have previously reported that Y-29794, 2-(8-dimethylaminooctylthio)-6-isopropyl-3-pyridyl 2-thienyl ketone citrate, is a non-peptide, potent, selective, and orally active inhibitor of PEP (fig. 1) (Nakajima et al., 1992). Therefore, we desired to determine whether this compound would halt or slow down the progression of A-LI deposition in the hippocampus of the SAM. We report that repeated treatment with Y-29794 prevented the progression of A-like deposition in SAM hippocampus, as assessed by immunohistochemical staining.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Structure of Y-29794, 2-(8-dimethylaminooctylthio)-6-isopropyl-3-pyridyl 2-thienyl ketone citrate.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animals. Male SAMP8/Ta/Sea (SEAC YOSHITOMI, Ltd., Fukuoka, Japan) were used throughout this study. These mice were fed in a ventilated animal rack under the conditions of 12 hr light and 12 hr dark (light period, 6:30 to 18:30). All animal experiments were in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved in advance by the Committee of Animal Experiments in Research Laboratories of Yoshitomi Pharmaceutical Industries Ltd.

Chemical and antibody. Y-29794, 2-(8-dimethylaminooctylthio)-6-isopropyl-3-pyridyl 2-thienyl ketone citrate, was synthesized in our laboratories. Polyclonal antibody to synthetic A(1-16) was a gift from Dr. S. Ishiura (University of Tokyo, Japan).

Repeated treatment for immunostaining. Doses used in our studies refer to bases. The 2.5-mo-old mice were given oral doses of Y-29794 (1, 10, 20 mg eq/kg/day) for 5.5 mo in drinking water ad libitum (n = 6 for the control and n = 5 for the drug-treated group). Because the life expectancy of the SAMP8 is about 12 mo, which is significantly shorter than that of the nonsenescent lines, we initiated the repeated treatment of the animals with each n > 6, allowing for the expected attrition of the animals. In a preliminary repeated treatment study, the daily intake of water and the levels of Y-29794 in the plasma and brain of SAMP8/Ta/Sea were compared between repeated direct oral dosing and repeated oral dosing via drinking water ad libitum. This experiment gave us desirable drug concentrations in drinking water which gives drug levels equivalent to those of direct oral dosing. According to this result, a concentrated solution (1.46 mg/ml) of Y-29794 was diluted to give the desired concentrations of drug solutions. The solutions were made every Monday and Friday. An aqueous solution of Y-29794 was stable for at least 2 wk at room temperature. We started the treatment of SAMP8/Ta/Sea from 2.5 mo of age, because A-LI appeared in 4-mo-old or older SAMP8/Ta (Takemura et al., 1993; Fukunari et al., 1994). The animals of an age-matched control group were given vehicle solution throughout the dosing period.

Immunostaining. After the repeated dosing, the mice were deeply anesthetized with an i.p. injection of sodium pentobarbital and perfused transcardially with 50 ml of ice-cold physiological saline, followed by a fixative of 50 ml of 4% formaldehyde in 0.1 M phosphate buffer (pH 7.4). The brains were removed and immersed in the same fixative for 16 hr at 4°C. Each tissue block was dehydrated, embedded in paraffin and cut into 3-µm-thick coronal sections. Ten sections were selected that covered equivalent volumes of the hippocampus with equal brain anatomy for each mouse. Deparaffinized sections (10 sections from each animal) were immunohistochemically stained by the labeled streptavidin-biotin complex method (Palacios et al., 1992), using an LSAB-kit (DAKO Japan; product number, K682). Sections were kept in 0.3% H₂O₂ in 0.05 M TBS (pH 7.5) for 5 min at room temperature, rinsed with TBS containing 0.1% Tween 80 and incubated in a primary antibody diluted in Tris buffer for 24 hr at 4°C. A rabbit polyclonal antibody raised against synthetic A peptide (1-16) was used (1: 300). The sections were rinsed with Tris buffer and incubated in biotinylated anti-rabbit IgG diluted (1:200) in TBS for 20 min at room temperature. After washing with TBS, the sections were incubated in a streptavidin-biotin peroxidase complex for 15 min at room temperature. They were then washed with TBS and developed with amino-ethyl carbazol and 0.006% H₂O₂ in 0.2% sodium acetate buffer (pH 5.3). The absorption experiment with synthetic A(1-40) proved that the positive staining was specific to A. Furthermore, the A antibody used in this study recognizes amyloid A(1-16) and has high specificity for the A peptide (Tagawa et al., 1993).

Digital image analysis. Light microscopy was performed using a light microscope (VANOX-S AH-2, Olympus, Tokyo, Japan) with a video camera system (ITC-370 M, Olympus-Ikegami, Tokyo, Japan). Photographs of the hippocampus were digitized with a flatbed image scanner (HP Scan Jet IIcx, Hewlett-Packard, Downer's Grove, IL) at a resolution of 288 dots per inch in a black-and-white-photo mode. Particle analyses of the granules were performed on a Macintosh Quadra 840AV computer with the public domain NIH Image program in 256-gray-scale mode (white = 0, black = 255). NIH Image was written by Wayne Rasband at the U.S. National Institutes of Health (Bethesda, MD) and is available from the Internet by anonymous ftp from zippy.nimh.nih.bov or on floppy disks from NTIS (5285 Port Royal Rd., Springfield, VA 22161, part number PB93-504868). Immunostaining, taking photographs and capturing digital images were blindly assigned to three investigators (A.K., A.F. and Y.S.).

Changes in Y-29794 levels after repeated treatment. In a separate experiment, 11-mo-old male SAMP8/Ta/Sea (n = 5 for each treatment group) were given oral doses of Y-29794 in drinking water ad libitum for 3 wk. The concentration of the drug solution was determined as described above. At the end of the dosing period, blood was collected transcardially into a heparinized syringe under deep pentobarbital anesthesia. Sampling time was at 1, 4, 7 and 10 A.M., and 1, 4, 7, 10 P.M. The blood samples were centrifuged at 800 × g for 15 min to collect the plasma. The brains, excluding the cerebellum, were dissected immediately after the blood collection. The plasma and brain samples were stored frozen until Y-29794 was measured by reversed-phase high-performance liquid chromatography. Frozen plasma (0.4 ml) was thawed and mixed with 0.5 ml of 0.5 M Na₂CO₃ and 4 ml of 1,2-dichloroethane: n-hexane (1:9, v/v) in a 10-ml glass test tube with a sealed cap. The mixture was vigorously shaken for 10 min and centrifuged at 800 × g for 5 min. A 3.7-ml portion of the upper organic phase was pipetted and mixed with 0.3 ml of MeOH: 0.1 N HCl (1:1, v/v) in a 5-ml glass test tube with a sealed cap. The mixture was vigorously shaken for 10 min and centrifuged at 800 × g for 5 min. After the centrifugation, a 0.15-ml portion of the lower aqueous phase was collected for high-performance liquid chromatography injection. Samples were not stored but were immediately subjected to high-performance liquid chromatography after the extraction to avoid unforeseen decomposition. The frozen brain tissue (0.3-0.4 g wet weight) was thawed, mixed with 0.5 ml of distilled water, and homogenized with a Potter-Elvehjem glass-Teflon homogenizer. To this homogenate, 0.1 ml of distilled water, 0.5 ml of 0.5 M Na₂CO₃, and 4 ml of 1,2-dichloroethane: n-hexane (1:9, v/v) were added. The mixture was processed as described above.

Data analysis. The statistical differences among the control and drug-treated groups were calculated by the Dunnett method. The distribution of the control group was simulated with RS/1, by assuming that it was the sum of multiple normal distributions.

	Results

Top Abstract Introduction Methods Results Discussion References

Although a small number of deaths occurred in both control and drug-treated groups due to the shorter life expectancy than that of nonsenescent lines, mortality was not significantly different among these groups. In addition, body weight steadily increased in both control and drug-treated groups throughout the dosing period.

Immunostaining of SAM brain with specific antibody to A. Granular structures of A-LI were predominantly immunostained in the hippocampus of the senescence-prone strain (fig. 2A). Granular structures of A-LI were also distributed in other regions; both gray and white matters such as the lateral olfactory tract, medial septum, cerebral cortex, thalamus, caudate putamen, pyramidal tract, cerebellar peduncle, cerebellum and some cranial nerve roots and nuclei.

View larger version (K): [in this window] [in a new window]   Fig. 2.   Typical immunohistochemical detection of A-LI in the CA1 field of the dorsal hippocampus and cerebral microvessel in the thalamus of the SAM brain. A, Dorsal hippocampus. Coronal sections from SAMP8/Ta/Sea (8-mo-old) were immunostained with anti-A. B, Cerebral microvessels in the thalamus. Photograph was taken with a Nomarski differential interference module. Bar represents 100 µm in (A) and 10 µm in (B). Immunostaining procedures are described in "Methods."

View larger version (58K): [in this window] [in a new window]   Fig. 3.   Outline of the procedures of digital image analyses of A-LI granules after repeated oral dosing with Y-29794. Photographs of immunostained coronal sections of the hippocampus were digitized. Particle analyses of the granules were performed with NIH Image program. Detailed conditions are described in "Methods."

Effect of repeated treatment of Y-29794 on the number, area and density of A-like deposition. Repeated treatment of the SAM with Y-29794 at 20 mg eq/kg/day for 5.5 mo significantly reduced the total number of the A-positive granules compared with the control group (P < .05, fig. 4A). Lower doses of Y-29794 (1 and 10 mg eq/kg/day) did not produce a significant decrease in the number of granules, although the effect of Y-29794 was dose dependent. Similarly, treatment with Y-29794 at 20 mg eq/kg/day significantly reduced the total area of the A-positive granules over which they were distributed (P < .05, fig. 4B). The granules were 8.1 ± 0.17 µm in diameter; this did not significantly vary among the control and drug-treated groups. The suppressive effect of Y-29794 was more evident on the mean density of the granules (fig. 4C). Chronic treatment with Y-29794 at 1, 10 and 20 mg eq/kg/day significantly decreased the mean density of the granules compared with the control group (P < .01).

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Suppressive effects of Y-29794 on the deposition of A-like deposits in SAM hippocampus after repeated treatment. A, Changes in total number of the A-LI granules per mouse. B, Changes in total area of the A-LI granules per mouse. C, Changes in mean density (gray scale value) of the A-LI granules. Data are mean ± S.E. (n = 6 for control, n = 5 for drug-treated group). Particle analyses of the granules were performed as described in figure 3. *P < .05; **P < .01 vs. control (Dunnett method).

Effect of repeated treatment of Y-29794 on the distribution of gray scale values of the density of A-positive granules. There was no evident denser core in the granules in both the control and Y-29794 treated mice. However, the granules were classified into two groups, the lighter and the denser; the latter was markedly decreased in the Y-29794 treated groups. The gray scale values of A-positive granules of the control group showed characteristic biphasic distribution (fig. 5A). From a simulation by fitting the distribution to a normal distribution function, it was calculated that two normal distributions, with m = 74,  = 20.5 and m = 127,  = 26.6, gave the best fit, where m and  represented the mean and the S.D., respectively. The calculated relative abundance of the two curves was 35.4 and 64.6%, respectively. A similar calculation demonstrated that the distribution of drug-treated groups could be simulated by only one curve (fig. 5B-D). The statistical parameters for drug-treated groups were as follows: m = 40,  = 17.9 for 1 mg/kg; m = 33,  = 15.1 for 10 mg/kg; m = 27,  = 10.5 for 20 mg/kg.

View larger version (42K): [in this window] [in a new window]   Fig. 5.   Distribution of gray scale values of the digitized images of the granules. Curves overlapped on the histograms represent the summed and individual normal distribution obtained by the simulation as described in "Methods."

Changes in Y-29794 levels after repeated treatment. When SAM were repeatedly administered Y-29794 (10 and 20 mg eq/kg/day) in a separate experiment (fig. 6), the levels of Y-29794 in the SAM brain throughout the final day of the dosing period were above 10 ng/gram of tissue (7.7 nM). The levels in the brain were constantly several times higher than those in the plasma.

View larger version (70K): [in this window] [in a new window]   Fig. 6.   Y-29794 levels after repeated oral administration. Shadowed area represents dark period (P.M. 6:30 - A.M. 6:30). Hatched bar shows the concentration corresponding to the IC50 value of Y-29794 to PEP. Each point represents mean ± S.E. from five mice.

	Discussion

Top Abstract Introduction Methods Results Discussion References

In our study, granular structures of A-LI were predominantly immunostained in the hippocampus with progression of age. Repeated treatment of the SAM with Y-29794, a specific inhibitor of PEP, significantly reduced the number and density of A-positive granular structures in the hippocampus of the SAM. This is the first report, to our knowledge, demonstrating the inhibition of A-like deposition in an animal model.

A major histopathological hallmark of AD is the presence of amyloid deposits in the hippocampus, as well as in the parenchyma of the amygdala and neocortex. Although the structure of A immunoreactive deposits in the SAM hippocampus is different from that in the brain of AD patients, we also observed A-positive deposits around the cerebral microvessels of the SAM (fig. 2B). Cerebral amyloidosis is another prominent and characteristic pathological change found in the brains of AD patients (Mountjoy et al., 1982; Kawai et al., 1993; Selkoe, 1994), non-human primates (Martin et al., 1991; Bons et al., 1992; Bons and Mestre, 1993; Bons et al., 1994) and canines (Cummings et al., 1993). The presence of A immunoreactive deposits and cerebral amyloidosis in the SAM brain suggests that A-like granular structures in the SAM brain may pathologically related to human A deposits.

Similar granular structures were seen in the hippocampi of aged C57/B1 mice (Jucker et al., 1992, 1994). Those deposits were stained with many polyclonal antibodies including A and neuropeptides. However, the staining was not blocked even when preabsorbed antibodies were used. However, the positive staining of the deposited structures in SAM in our study was completely blocked by preabsorbed anti-A(1-40), indicating that the staining was specific to A.

Y-29794 has been demonstrated to be highly selective to mammalian prolyl endopeptidase among serine-, thiol-, metallo-proteases and bacterial homologues of mammalian prolyl endopeptidase in the previous study. Y-29794 also lacks significant activity on various ion channels or receptors for neurotransmitters (Nakajima et al., 1992). Furthermore, when the SAM were repeatedly given Y-29794 (fig. 6), the levels of Y-29794 in the SAM brain at 10 and 20 mg eq/kg/day throughout the dosing period were above 10 ng/g of tissue (7.7 nM) which was high enough to produce an effect on the basis of the IC₅₀ value of Y-29794 (0.3 nM). Therefore, the reported effect in the SAM hippocampus can be attributed to the inhibitory action of Y-29794 on prolyl endopeptidase.

The daily change in drug levels (fig. 6) is mainly due to that of spontaneous motor activity of the SAMP8, because the patterns of these correspond well to each other. In rats, dogs and monkeys, a change in the brain levels of Y-29794 closely resembles that of the plasma, as in the case of SAM. These and previous data (Nakajima et al., 1992) suggest that Y-29794 can readily penetrate from the blood to the brain.

Although a small number of death cases occurred in both control and drug-treated groups, mortality and change in body weight were not significantly different among these groups. Furthermore, the doses we used in this study were nontoxic doses, based on the results of toxicology studies of repeated dosing for 3 mo in rats and dogs that we completed in our laboratories (H. Kashima, S. Oda and M. Setoguchi, unpublished results). Therefore, the observed effects of Y-29794 did not results from the toxic effect of this compound.

Although Y-29794 is known to undergo first-pass transformation to inactive metabolites in rats, dogs and monkeys, unchanged Y-29794 was found in the brain tissue at a concentration several times higher than that in plasma. In addition, the brain levels of Y-29794 in monkey and dogs were higher than in rodents (K. Sagara and Y. Tsunekawa, unpublished observation). The levels of Y-29794 in the SAM brain were also higher than those in the plasma. The rapid inactivation of Y-29794 will be advantageous to avoid possible undesirable effects on peripheral tissues where PEP is present.

There are two major molecular species in A, namely A(1-40) which is a 40-residue form of A, and A(1-42/43) which has a two- to three-residue extension at the C-terminus. A(1-42/43) appears to be the predominant A species in senile plaques (Masters et al., 1985; Roher et al., 1986; Iwatsubo et al., 1994, 1995; Fukumoto et al., 1996). A(1-42/43), which deposits in plaques with varying digresses of congophilicity, is critically important in AD (Gravina et al., 1995; Shinkai et al., 1995). These reports support the hypothesis that soluble A(1-40) could be seeded by a small amount of A(1-42/43) to initiate the formation of aggregates or granules (Jarrett and Lansbury., 1993). One can speculate that the number of granules observed in our study reflects the number of seeds, if the granules of A-LI in the SAM hippocampus were formed by a similar nucleation mechanism as proposed (Jarrett and Lansbury, 1993). Reduction of the number of granules by Y-29794 suggests that Y-29794 might slow down the spread of A. Although the quantitative relationship between the gray scale values and the amount of antigen has not yet been clarified, the intensity of immunohistochemical staining correlated with the amount of A peptide (Tamaoka et al., 1995). Therefore, the result shown in figure 4C suggests that the amount of A-antigen could be decreased in the SAM by chronic treatment with Y-29794.

In our study, only a 20-mg dose showed effects on the granule number and total area, whereas all three doses affected the mean density. This difference could be attributed to a difference in the sensitivity of these parameters. We do not set any criteria on the size of deposits in digital image analysis. For example, we counted both a deposit with a size of 1 pixel and the one of 100 pixels as "one" deposit. Therefore, we expected that a change in the number and total area of the deposit was lesser than that of mean density. The results in figure 4 meet this expectation.

The distribution of gray scale values in the control group was biphasic (fig. 4), whereas the distribution in the drug-treated groups did not meet this criterion. In drug-treated groups, the denser part was markedly decreased, and the lighter part shifted to the left in a dose-dependent manner with a change in the mean value of the simulated normal distribution from 40 to 27. The peak height of the distribution of the lighter component was not significantly enhanced, suggesting that the formation of, rather than the accumulation of, the denser component was preferentially inhibited by Y-29794.

Although our study suggests that Y-29794 prevents the formation of A-like deposits in the SAM hippocampus, several issues remain to be clarified. First, we failed to detect significant neuronal degeneration in the SAM hippocampus. However, this could be attributed to the short life span of this senescence-accelerated strain. Because the SAMP8 strain has a significantly shorter life span (average 12 mo) than control SAMR1 mice (average 24 mo or more), death may occur before neurodegeneration can develop as a consequence of A production. Furthermore, transgenic mice also lack neuronal degeneration, although these mice express very high levels of amyloid peptides and amyloid deposits (Games et al., 1995).

Second, further characterization of the content and metabolism of the deposited structures in the SAM hippocampus remains to be investigated. We should try to stain the SAM hippocampus with antibodies to isoforms of A with a different truncated amino-terminus, such as AN3(pE), to compare the abundance of these isoforms after chronic treatment with Y-29794 (Saido et al., 1995). Furthermore, recent evidence indicates that some of the N-terminal amino acids of amyloid peptides are also modified (Saido et al., 1995). Other components, such as -antichymotrypsin, C1q, and apolipoprotein E, are also reported to colocalize in the human senile plaques (Abraham et al., 1988; Ma et al., 1994; Afagh et al., 1996). Furthermore, the effects of Y-29794 on amyloidosis in peripheral tissues including blood vessels and ideally in transgenic animals are no doubt very important issues to be examined in future studies.

Transgenic mice that express high levels of the human APP gene with the promoter region of PDGF (PD-APP) (Games et al., 1995) or hamster prions (Hsiao et al., 1996) were found to develop progressively Alzheimer-like neuropathology, including formation of congophilic plaques, A deposition, gliosis and loss of synaptic density, as well as memory deficits (Hsiao et al., 1996). These transgenic mice could be suitable for assessing the pathogenicity of APP and the therapeutic potential of novel compounds, because these animals have a defined genetic background. SAM, in contrast, spontaneously develop A-immunoreactive deposition, although whether their genetic features are altered has not yet been determined. However, there are reports indicating that SAMP8 exhibits age-related deficits in memory and learning ability (Miyamoto et al., 1986; Yagi et al., 1988). Because SAMP8 spontaneously exhibits amyloid-like deposition, as well as memory deficits in an age-related manner, and AD is genetically heterogeneous (Kang et al., 1987; Tanzi et al., 1987, 1993; Schellenberg et al., 1992; Levy Lahad et al., 1995; Schellenberg, 1995), SAM could be another animal model that researchers could employ, in addition to APP-transgenic mice, for assessing the therapeutic potential of compounds aimed at preventing the development of AD.

The seed hypothesis is not the only way to explain the mechanisms of A deposition. The failure of a mechanism of removal of A or an imbalance in amyloidgenic and nonamyloidgenic processing could be another possible mechanism. Heparan sulfate proteoglycan (Kisilevski, 1990; Snow et al., 1994), apolipoprotein E4 (Saunders et al., 1993; Strittmatter and Roses, 1995), complement C1q and inflammatory processes (Jiang et al., 1994; Afagh et al., 1996) are closely related to neurodegeneration in the brains of Alzheimer's patients, and therefore, early association of A with these components could also accelerate A deposition. Immunohistochemistry of these components in the SAM hippocampus as well as cerebral vessels at both light and electron microscopic levels would give us further clues to the mechanisms of the deposition and the action of Y-29794.

In conclusion, our results indicate that a non-peptide PEP inhibitor, Y-29794, can significantly inhibit the senescence-accelerated formation and deposition of A-like substance in the SAM hippocampus. The deposited structures in the SAM brain may be similar to some very early stages of human brain amyloidgenesis, and the prevention of this process by the PEP inhibitor Y-29794 may be a novel, potentially therapeutic strategy to curb early events leading ultimately to amyloid formation and to AD.