Hematopoietic Prostaglandin D Synthase and DP^Sub 1^ Receptor Are Selectively Upregulated in Microgl

Jun 21, 04:05 AM

Current Headlines: By Mohri, Ikuko Kadoyama, Keiichi; Kanekiyo, Takahisa; Sato, Yo; Et al

Abstract Prostaglandin (PG) D^sub 2^ is produced in activated microglia by the action of hematopoietic PGD synthase (HPGDS) and plays important roles in neuroinflammation. Because the fact that neuroinflammation accelerates progression of Alzheimer disease (AD) has been documented, we investigated whether PGD^sub 2^ is also involved in the pathology of AD. Here, we report that the level of the mRNA of the receptor for PGD^sub 2^ (DP^sub 1^) was increased in AD brains compared with the level in non-AD brains. Immunocytochemical analysis showed HPGDS expression to be localized in the microglia surrounding senile plaques. In situ hybridization studies revealed that DP^sub 1^ mRNA was specifically localized in microglia and reactive astrocytes within senile plaques of AD brains. In the brain of Tg2576 mice, a model of AD, HPGDS and DP^sub 1^ proteins were mainly localized immunocytochemically in microglia and astrocytes in the plaques, and the levels of their mRNAs increased in parallel with amyloid beta deposition. These results indicate that PGD^sub 2^ may act as a mediator of plaque-associated inflammation in AD brain and may explain the pharmacologic mechanisms underlying the favorable response of patients with AD to nonsteroidal anti-inflammatory drugs.

Key Words: Alzheimer disease, Amyloid beta, Gliosis, Neuroinflammation, Nonsteroidal anti-inflammatory drugs (NSAIDs), Prostanoid, Tg2576 mouse.

INTRODUCTION

Alzheimer disease (AD) is the most common form of dementia. The cognitive decline in patients with AD is associated with neuronal degeneration, the appearance of neurofibrillary tangles, and the formation of senile plaques (1). It has been reported that inflammatory responses, including the production of proinflammatory cytokines and prostaglandins by activated microglia and astroglia, play important roles in the pathology of AD (2-5). Senile plaques arise from the excessive accumulation, aggregation, and deposition of amyloid beta (Abeta) peptide, which triggers the activation of microglia and astrocytes around the senile plaques. Tg2576 mice overexpressing human APP695 with the "Swedish" mutation (a murine model of AD) develop memory deficits and senile plaques as they age (6). These animals show a rapid increase in Abeta deposition starting around 6 months and amyloid plaques beginning at 9 to 12 months (7).

We recently reported that activated microglia express hematopoietic prostaglandin (PG) D synthase (HPGDS) and that PGD^sub 2^ produced by HPGDS promotes neuroinflammation in a mouse model of Krabbe disease (8). These discoveries prompted us to clarify whether PGD^sub 2^ is also involved in the plaque-associated inflammation in the AD brain. A determination of the mRNA contents of all known brain prostanoid synthases and receptors showed that HPGDS and DP1 mRNAs were selectively expressed to a higher level in AD brains compared with their levels in age-matched control patients. This overexpression was localized to reactive astrocytes and microglia closely associated with senile plaques, suggesting that HPGDS/ PGD^sub 2^/DP^sub 1^ signaling may accelerate chronic local inflammation around plaques in AD. We also found that HPGDS and DP^sub 1^ mRNAs were selectively expressed in reactive astrocytes and microglia closely associated with senile plaques in the AD brains and brains of Tg2576 mouse. Our findings suggest that HPGDS/ PGD^sub 2^/DP^sub 1^ signaling may accelerate chronic local inflammation around senile plaques in the AD brain.

MATERIALS AND METHODS

Human Tissue Source

Human brain tissues were obtained from Tokyo Metropolitan Brain Bank for Aging Research according to the neuropathologic protocols described elsewhere (9). We obtained tissue fresh at autopsy and stored it at -80[degrees]C, and, to avoid RNA degradation, we selected samples with as short a postmortem interval before autopsy as possible. The storage period was 2 to 32 months and 2 to 31 months for AD and control brains, respectively, and the storage period was not significantly different between the 2 groups.

Fifteen brains each from AD patients and control patients without clinicopathologic evidence of AD were the basis of the present work. The diagnosis of AD was based on the following criteria: 1) clinical dementia rating of >/=1 (10); 2) the topographical distribution of senile plaques matching Braak stage C; and 3) of neurofibrillary tangles equal to or above stage IV, as reported elsewhere (9). The selection criterion for control patients was a clinical dementia rating score of 0, senile plaque stage 0 or A, and a neurofibrillary tangle stage lower than stage II. The profiles of AD and control patients are listed in Table 1. Subjects with AD were 70 to 93 years old (82.6 years on average), and the post-mortem delay ranged from 1.5 to 17 hours (average of 8.5 hours). The age of non-AD control patients was 71 to 91 years (79.5 on average), and the postmortem delay was 2.3 to 14.8 hours (average of 12.1 hours).

The tissues were either snap-frozen in liquid nitrogen or powdered dry ice or immediately fixed in 4% paraformaldehyde (Sigma, St. Louis, MO) for 2 days and processed into paraffin blocks. Paraffin or frozen sections of 6-[mu]m thickness from the frontal cortex and the hippocampus were cut with a cryostat (CM1850; Leica Microsystems, Wetzlar, Germany) and microtome (RM 2035; Leica), and used for immunohistochemical studies. Snap-frozen tissues, that were en face to the paraffin sections, were also used for quantitative reverse transcriptase (RT)-polymerase chain reaction (PCR) analysis of mRNA expression. The frontal cortex was selected for this purpose, because the difference in the number of senile plaques was most prominent there between AD and control groups. This study was approved by the institutional review boards of Osaka University Graduate School of Medicine, Tokyo Metropolitan Geriatric Hospital, Tokyo Metropolitan Institute of Gerontology, and Osaka Bioscience Institute.

Mice

All animal experiments were performed in accordance with Japanese law for the protection of experimental animals and conformed to the regulations issued by the National Institutes of Health and the Society for Neuroscience. Homozygous Tg2576 mice were purchased from Taconic (Hudson, NY) and maintained by interbreeding.

Quantitative Reverse Transcriptase-Polymerase Chain Reaction Analysis of mRNA

By quantitative RT-PCR analysis, we determined the mRNA contents of enzymes involved in prostanoid synthesis, including cyclooxygenase (COX)-1, COX-2, HPGDS, lipocalin-type PGD synthase (L- PGDS), microsomal PGE synthases (mPGES-1 and mPGES-2), cytosolic PGE synthase (cPGES), PGF synthase, prostacyclin synthase, thromboxane (TX) A synthase, and those of prostanoid receptors for PGD^sub 2^ (DP^sub 1^ and DP^sub 2^), PGE^sub 2^ (EP^sub 1^, EP^sub 2^, EP^sub 3^, and EP^sub 4^), PGF^sub 2alpha^ (FP), PGI^sub 2^ (IP), and TXA^sub 2^ (TP), as well as the mRNA content of cytosolic phospholipase A^sub 2^ (cPLA^sub 2^) and peroxisome proliferator- activated receptor (PPAR)-gamma. After removal of the pia-arachnoid membrane, total RNA was prepared from the frontal cortex of 17 AD patients and 12 control patients by using ISOGEN (Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions. In the mouse model, 3 each of Tg2576 and age-matched wild-type mice at ages 6 months, 1 year, and 2 years were subjected to quantitative RT-PCR for determination of COX-1, COX-2, HPGDS, DP^sub 1^, and DP^sub 2^ mRNA levels. All primers were synthesized by Sigma-Aldrich Japan (Tokyo, Japan). First-strand cDNAs were synthesized from 1 [mu]g of total RNA by using avian myeloblastosis virus reverse transcriptase (Takara, Kyoto, Japan) and oligo dT-adaptor primer (Takara) at 50[degrees]C for 40 minutes after denaturation at 72[degrees]C for 3 minutes. The cDNA was amplified by use of a real-time PCR LightCycler system (Roche Diagnostics, Indianapolis, IN), a LightCycler-FastStart DNA Master SYBR Green I (Roche Diagnostics), and gene-specific primers under the following conditions: an initial denaturation at 94[degrees]C for 10 minutes, followed by 40 cycles of denaturation at 94[degrees]C for 15 seconds, annealing for 5 seconds, and extension at 72[degrees]C for 10 seconds. The oligonucleotide primers and annealing temperatures used are shown in Table 2. All oligonucleotide primers were synthesized by Sigma Genosys Japan (Tokyo, Japan). PCR products were evaluated by melting- curve analysis following the manufacturer's instructions, checked after agarose gel electrophoresis, and sequenced. All values were corrected with reference to the value for glyceraldehyde-3- phosphate dehydrogenase (G3PDH), used as an internal standard.

Immunocytochemistry

Rabbit polyclonal antibodies against human HPGDS, L-PGDS, and mPGES-1 (11), and rabbit polyclonal (0.1 [mu]g/mL) and rat monoclonal (0.2 [mu]g/mL) anti-mouse HPGDS antibodies (12) were raised and purified in Osaka Bioscience Institute. The specificity of each antiserum (1:1000 dilution) was confirmed by the disappearance of immunoreactivity after incubation of the antiserum with an excess amount of the corresponding purified recombinant immunogen. The other primary antibodies used in this study were as follows: anti-cow glial fibrillary acidic protein (GFAP) (1:5000 dilution; DakoCytomation, Glostrup, Denmark) for astrocytes; anti- human CD68 antibody (1:100 dilution; DakoCytomation) and anti-Ibal (a generous gift of Dr. Shinichi Kousaka, National Center of Neurology and Psychiatry, Japan) for microglia/macrophages; and anti- human Abeta (11-28) antibody (1:100 dilution; IBL, Gunma, Japan) for detecting amyloid plaques. TABLE 1. Clinicopathologic Features of the Patients

Tg2576 and control mice were used for immunocytochemical analysis. Under deep ether anesthesia, the mice were perfused via the heart with physiologic saline, followed by 4% paraformaldehyde in 0.1 M sodium phosphate (pH 7.4) for 10 minutes. The brains were removed and postfixed in the same fixative overnight. Coronal slices were taken and routinely embedded in paraffin. Additionally, 2 each of Tg2576 and control mice were perfused with physiologic saline only and processed for the preparation of fresh-frozen sections. Both paraffin and frozen sections (5-[mu]m thickness) were mounted on 3-aminopropyltriethoxysilane-coated slides.

Deparaffinized sections were preincubated with 0.3% H^sub 2^O^sub 2^ (Wako, Osaka, Japan) in methanol followed by PBS containing 0.2% Triton X-100 (Nakarai Tesque, Kyoto, Japan). After pretreatment with 0.1% trypsin (Sigma) at 37[degrees]C for 15 minutes, they were sequentially incubated with primary antibody, biotinylated secondary antibody (2 [mu]g/mL; Vector Laboratories, Burlingame, CA), and avidin-biotin-complex (ABC) by using an ABC elite system (Vector Laboratories) according to the manufacturer's protocol.

For double immunostaining, deparaffinized sections were incubated at 4[degrees]C overnight with either anti-GFAP or anti-CD68 antibody together with rabbit anti-human HPGDS antiserum. The sections were then reacted with alkaline phosphatase-conjugated anti-mouse IgG antibody (2 [mu]g/mL; Aurora, Cambridge, UK) and biotin-conjugated anti-rabbit IgG antibody (Vector Laboratories), followed by ABC. The reaction products of horseradish peroxidase and alkaline phosphatase activities were visualized with diaminobenzidine (Dotite, Kumamoto, Japan) and naphthol AS-BI phosphate (Sigma) coupled to hexazotized new fuchsin (Merck, Darmstadt, Germany), respectively, as substrates.

TABLE 2. Primer Sequences and Annealing Temperatures Used for Quantitative Reverse Transcriptase-Polymerase Chain Reaction Analysis

All sections were analyzed, and images were obtained with an Olympus BX51 microscope (Olympus, Tokyo, Japan) equipped with a DP50 digital camera. The digital images were adjusted to an appropriate figure size required but were otherwise not processed.

FIGURE 1. (A-H) Expression of prostaglandin synthase mRNA in the frontal cortex of AD (n = 17) and control brains (n = 12), as evaluated by quantitative RT-PCR analysis. The numbers of copies relative to those of G3PDH are shown. The mean values and SE are indicated by the crossbars. AD, Alzheimer disease; COX, cyclooxygenase; cPGES, cytosolic PGES; HPGDS, hematopoietic prostaglandin (PG) D synthase; G3PDH, glyceraldehyde-3-phosphate dehydrogenase; L-PGDS, lipocalin-type PGD synthase; mPGES, microsomal PGE synthase; PGFS, PGF synthase.

In Situ Hybridization

In situ hybridization was performed by using a digoxigenin (DIG)- anti-DIG technique. For preparation of the human DP^sub 1^ riboprobe, a 484-base pair fragment was amplified by use of FastStart Taq DNA polymerase (Roche Diagnostics) with human DP^sub 1^-specific primers (forward, 5'-GGGGTACTCTGTGCTCTACTCCAG-3' and reverse, 5'-ACTGGATTCCATGTTAGTGGAATTG-3'), subcloned into pGEM-T Easy vector (Promega, Madison, WI), and converted into the corresponding RNA. Five-micrometer-thick cryosections were fixed in 10% formalin (Wako) for 1 hour and hybridized at 58[degrees]C for 16 hours with the DIG-labeled riboprobe for DP^sub 1^ in 50% formamide (Wako), 5 x saline sodium citrate, 5 x Denhardt's solution, 0.25 mg/mL yeast tRNA, and 0.5 mg/mL herring sperm DNA. After the sections had been washed with PBS, the DIG-labeled RNA was detected by using a Genius DNA labeling and detection kit (Roche Diagnostics) according to the manufacturer's protocol.

For identification of DP^sub 1^-expressing cells, the sections were reacted with the DP^sub 1^-specific riboprobe and then immunostained with either anti-human GFAP antibody, anti-human CD68 antibody, or anti-human Abeta (11 -28) antibody detected by using either immunofluorescence or immunoperoxidase. For fluorescence labeling, a biotinylated riboprobe was used in place of the DIG- labeled riboprobe. For double immunofluorescence, we confirmed the colocalization by omitting 1 of the first antibodies during the process.

Primary Cultures of Mouse Microglia and Astrocytes

We prepared primary cultures of microglia from wildtype mouse brains at postnatal day 1. Cerebral cortices were dissected, and the leptomeninges were completely removed. The tissues were minced, suspended in PBS (GIBCO, Grand Island, NY) containing 0.05% trypsin (GIBCO) and 0.01% DNase I (Sigma), and then incubated for 10 minutes at 37[degrees]C. After incubation and centrifugation, the cell pellets were washed 3 times with PBS. The cells were then filtered through a 75-[mu]m nylon mesh, centrifuged, suspended in Dulbecco's modified Eagle's medium (Nakalai Tesque) containing 10% fetal bovine serum (JRH Biosciences, Lenexa, KS), 100 IU/mL penicillin (GIBCO), and 100 [mu]g/mL streptomycin (GIBCO), and transferred to culture dishes. For microglial cultures, the medium was changed to Dulbecco's modified Eagle's medium containing 10% fetal bovine serum after a 24-hour culture period. This medium was exchanged for fresh medium twice weekly thereafter. The supernatants including microglial cells were collected and subcultured at 1 x 10^sup 5^ cells/well (6-well plate). After incubation in Dulbecco's modified Eagle's medium without fetal bovine serum for 6 hours, the microglia were stimulated by the addition of Abeta (1-40) (Peptide Institute, Osaka, Japan) for 24 hours. After the cells had been washed with PBS, RNA was extracted from them and subjected to quantitative RT- PCR.

FIGURE 2. (A-G) mRNA expression of prostanoid receptors in the frontal cortex of Alzheimer disease (AD) (n = 17) and control brains (n = 12), as determined by quantitative RT-PCR analysis. The numbers of copies relative to glyceraldehyde-3-phosphate dehydrogenase (G3PDH) are shown. The mean values and SE are indicated by the crossbars. *, p < 0.05.

Data Analysis

The difference of means among groups of data was analyzed by analysis of variance. Statistical significance was established at the level of p < 0.05 or 0.01.

RESULTS

Expression of Prostanoid-Producing Enzymes in Alzheimer Disease Brains

Using the frontal cortex obtained from AD and control patients, we examined the mRNA expression level of all enzymes involved in prostanoid synthesis. Quantitative RT-PCR analysis revealed the mRNA levels of HPGDS (Fig. 1A) and L-PGDS (Fig. 1B) to be elevated in AD brains compared with those of the control brains. However, the mRNA levels of other prostanoid-producing enzymes including COX-1 and -2 (Fig. 1C, D), PGESs including mPGES-1, mPGES-2, and cPGES (Fig. 1E- G), PGF synthase (Fig. 1H), and TXA synthase (data not shown), as well as the mRNA level of cPLA^sub 2^ (data not shown), were similar between control and AD brains. Prostacyclin synthase mRNA was not detected in either AD or control samples (data not shown).

FIGURE 3. Expression of hematopoietic prostaglandin synthase (HPGDS) in Alzheimer disease (AD) and control brains. In the following panels, asterisks indicate the presence of senile plaques. (A, B) Immunohistochemical analysis of HPGDS in control (A) and AD (B) brains. Arrows point to HPGDS-positive cells. Scale bar = 50 [mu]m. (C, D) Immunostaining for HPGDS (C) and amyloid beta (Abeta) (D) associated with the senile plaques in an AD brain. The same cells in the serial sections are indicated by arrows. Scale bar = 20 [mu]m. (E, F) Double immunostaining for HPGDS (purple) and glial fibrillary acidic protein (GFAP) (pink, E) or CD68 (pink, F) in an AD brain. HPGDS-positive/GFAP- (E) or CD68-positive (F) and HPGDS- negative/GFAP- (E) or CD68-positive (F) cells are indicated by arrows and arrowheads, respectively. Scale bar = 20 [mu]m.

Increased DP^sub 1^ Expression in Alzheimer Disease Brains

Next, we examined the mRNA expression level of all prostanoid receptors thus far identified (Fig. 2) and found that the expression of DP^sub 1^ was significantly upregulated in AD brains as compared with that in control brains (Fig. 2A). In 6 of the 17 AD patients, the mRNA level was higher than 2 SD from the mean of control brains, and these values were 2.5- to 16-fold greater than this mean. The mRNA level of DP^sub 2^, which is another PGD^sub 2^ receptor, was also increased, but not significantly, in AD brains (Fig. 2B). In contrast, the mRNA levels of other prostanoid receptors were almost the same between the 2 groups (Fig. 2C-G) except for the approximately 2-fold increases in EP^sub 2^ (Fig. 2D) and FP (Fig. 2G) in AD brains. In addition, the mRNA level of PPARgamma, which is a nuclear receptor for PGD^sub 2^ metabolites (13), was not significantly different between AD and control brains (data not shown).

Expression of Hematopoietic Prostaglandin D Synthase and DP^sub 1^ Is Associated with Senile Plaques in Alzheimer Disease Brains

Because the mRNA levels of HPGDS and DP^sub 1^ mRNA were increased in AD brains, as judged from the results of RT-PCR analysis, we investigated the cellular localization of these messages.

Immunocytochemical analysis revealed that anti-HPGDS- immunoreactive cells were greater in number in the AD brain than in the control brain (Fig. 3A, B). Serial irnmunostaining for HPGDS and Abeta revealed that HPGDS-positive cells were closely associated with Abeta-positive senile plaques (Fig. 3C, D) in the frontal cortex and the hippocampus. HPGDS-positive cells were frequently recognized in relatively small-sized plaques accompanied by activated microglia (Fig. 3B, C) but not in the large "burned out class" of plaques (data not shown). Double immunocytochemical staining revealed that HPGDS was localized in GFAP-positive reactive astrocytes (Fig. 3E) and in CD68-positive microglia (Fig. 3F) associated with the senile plaques, suggesting that HPGDS expression was upregulated in reactive astrocytes and microglia within the senile plaques. Immunocytochemical analysis revealed that L-PGDS- expressing cells having the same morphologic characteristics as oligodendrocytes, as reported previously (14), were increased in number and were more intensely stained in AD than in control brains; however, the distribution of L-PGDS-positive oligodendrocytes was almost ubiquitous and seemingly unrelated to that of senile plaques (data not shown). mPGES was localized in some neurons of both AD and control brains, but there was no significant difference in intensity of staining or in the number of immunoreactive neurons between the 2 (data not shown). FIGURE 4. (A) In situ hybridization for DP^sub 1^ in control and Alzheimer disease (AD) brains. Control (left) and AD brain (middle) sections were incubated with antisense riboprobe. Another AD brain section was incubated with sense riboprobe (right). Scale bar = 100 [mu]m. (B, C) In situ hybridization for DP^sub 1^ (purple) and immunocytochemistry for amyloid beta (Abeta) (pink) in an AD brain. The DP^sub 1^-positive cells are uniformly intermingled with Abeta immunoreactivity in the senile plaque (B), and a large- sized DP^sub 1^-negative senile plaque is shown (C). Scale bar = 20 [mu]m. (D, E) Double staining by in situ hybridization for DP^sub 1^ (purple) and immunocytochemistry for glial fibrillary acidic protein (GFAP) (brown, D) or CD68 (brown, E). Scale bar = 50 [mu]m. (F, G) Double fluorescence labeling for DP^sub 1^ (green) and GFAP (red, F) or CD68 (red, G). Double-positive cells are indicated by arrows. Scale bar = 30 [mu]m.

Upregulation of DP^sub 1^ mRNA in AD brains was confirmed by in situ hybridization with a DP^sub 1^-specific antisense riboprobe, because antibody specific for human DP^sub 1^ is still unavailable. DP^sub 1^ mRNA was hardly detected in the control brain (Fig. 4A); however, high signal density was detected in AD brains (Fig. 4A). When we hybridized the sections with the DIG-labeled antisense riboprobe in the presence of a 500-fold excess of nonlabeled antisense riboprobe or with sense riboprobe, no signal was observed in either control or AD brain sections (Fig. 4A, sense). Double labeling for Abeta and DP^sub 1^ hybridization revealed that the distribution of the DP^sub 1^ signal was closely associated with Abeta immunoreactivity (Fig. 4B). However, the intensity of the DP^sub 1^ signal differed among the senile plaques; and the signal was less preferentially observed in large-sized senile plaques (Fig. 4C). Furthermore, double staining with antibodies against glial markers revealed that the DP^sub 1^ signal was localized in GFAP- positive reactive astrocytes (Fig. 4D, F) and CD68-positive microglia (Fig. 4E, G) within senile plaques, similar to the distribution of HPGDS shown in Figure 3E and F. These results indicate that expression of HPGDS and DP^sub 1^ was increased in reactive astrocytes and microglia within the senile plaques in AD brain and suggest that PGD^sub 2^ produced by HPGDS exerts its function through binding to DP^sub 1^ in a paracrine or autocrine manner. The preferential expression of these PGD2-related molecules in the small-sized senile plaques suggests the contribution of PGD^sub 2^ to neuroinflammation in the early phase of plaque evolution.

Upregulation of Hematopoietic Prostaglandin D Synthase and DP^sub 1^ Expression in Tg2576 Mouse Brain

Next we examined the Tg2576 mouse brain to clarify the relevancy of our findings on human AD brains. In 2-year-old mice, there were numerous amyloid plaques in the cortex of Tg2576 brains (Fig. 5A) but not in that of the wild-type mouse brain (Fig. 5B). The results of the combination of Congo red staining and immunocytochemistry for Ibal, a marker for activated microglia, are shown in Figure 5C and D. Surrounding the amyloid plaques were many Ibal-positive activated microglia (Fig. 5D, arrowheads), which were not seen in the wild- type control. Double labeling with Congo red and anti-GFAP showed that fibers from activated astrocytes surrounded and enclosed amyloid plaques (Fig. 5E, F). These findings indicate that inflammatory responses such as microglial activation and astrogliosis were remarkable around early amyloid plaques in the Tg2576 mouse brain.

Next we investigated the localization of HPGDS and DP^sub 1^ in the Tg2576 mouse brain. As in the case of the human AD brains, HPGDS- positive microglia were common in relatively small plaques (Fig. 5G) but not in the large burned out class of plaques (Fig. 5H). HPGDS and Iba1 double immunofluorescence revealed that HPGDS was expressed in a certain number of Iba1-positive activated microglia (Fig. 51). DP^sub 1^ and GFAP double immunofluorescence revealed that DP^sub 1^ was expressed on the astrocytic processes close to the amyloid plaques (Fig. 5J).

FIGURE 5. (A, B) Amyloid beta (Abeta) immunostaining of primary motor cortex of Tg2576 (A) and wild-type (B) brains from 2-year-old mice. Scale bar = 200 [mu]m. (C, D) Combination of Congo red staining (pink) and Iba1 immunocytochemistry (brown) in a Tg2576 mouse brain. (D) The arrowheads and asterisk indicate Iba1-positive activated microglia and an amyloid plaque, respectively. Scale bars = (C) 200 [mu]m; (D) 20 [mu]m. (E, F) Combination of Congo red staining (pink) and glial fibrillary acidic protein (GFAP) immunocytochemistry (brown) in a Tg2576 mouse brain. (F) Arrowheads and asterisk indicate processes of GFAP-positive astrocytes and an amyloid plaque, respectively. Scale bars = (E) 100 [mu]m; (F) 20 [mu]m. (G, H) Hematopoietic prostaglandin D synthase (HPGDS) immunocytochemistry on a Tg2576 mouse brain. HPGDS-positive cells surround a small reactive amyloid plaque (G) but not the large burned out class of plaques (H). Asterisks indicate amyloid plaques. Scale bar = 20 [mu]m. (I) Double immunofluorescence for HPGDS (pink) and Iba1 (green). Some of the Iba1-positive microglia express HPGDS. Scale bar = 50 [mu]m. Inset shows a high magnification view of double-positive cells showing morphologic characteristics of ameboid microglia. (J) Double immunofluorescence for DP^sub 1^ (red) and GFAP (green). DP^sub 1^ is expressed on the processes of reactive astrocytes that have extended toward the center of this amyloid plaque. Scale bar =10 [mu]m.

Quantitative RT-PCR was performed to evaluate the level of PGD^sub 2^-related molecules in Tg2576 mouse brains and age-matched controls at the age of 6 months, 1 year, and 2 years. The expression of COX-1 mRNA was not different between Tg2576 and control mice (Fig. 6A), but that for COX-2 was higher in the Tg2578 mouse starting from the age of 6 months (Fig. 6B). The level of HPGDS mRNA was the same between Tg2576 and wild-type controls until they reached 2 years of age, at which time its level was 1.5 times higher in the Tg2576 mice than in the controls (Fig. 6C). The level of DP^sub 1^ mRNA progressively increased in the Tg2576 mouse brain and reached a level 5-times higher than that in the control mice at the age of 2 years (Fig. 6D). In the case of DP^sub 2^, its mRNA level progressively increased starting before the age of 1 year compared with that of the age-matched controls (Fig. 6E). These lines of evidence show that HPGDS and DP^sub 1^ expression was associated with the small senile plaques and that the increase in DP^sub 2^ receptors was more marked than that of HPGDS in the mouse model as in human AD.

To examine whether Abeta could directly stimulate the production of HPGDS in microglia, we performed an in vitro study using primary cultures of microglia obtained from normal neonatal mouse brains. We measured the level of HPGDS mRNA in these cultures before and after stimulating the cells with Abeta(1-42) at concentrations from 0 to 1000 nM. As shown in Figure 6F, the level of HPGDS mRNA did not increase when a 1 nM concentration of Abeta(1-42) was tested but did significantly increase when a 10 nM concentration or higher was used. However, no further increase was observed at concentrations higher than 10 nM (Fig. 6F).

FIGURE 6. (A-E) Number of copies of mRNA for cyclooxygenase (COX)- 1, COX-2, hematopoietic prostaglandin D synthase (HPGDS), DP^sub 1^, and DP^sub 2^ relative to that of glyceraldehyde-3-phosphate dehydrogenase (G3PDH) in brains from Tg2576 mouse (closed circles) and wild-type mouse (open circles) taken at the age of 6 months, 1 year, and 2 years (n = 3). *, p < 0.05; **, p < 0.01. (F) Number of copies of HPGDS mRNA relative to that of G3PDH from primary cultures of microglia before and after amyloid beta (Abeta) (1-42) stimulation (n = 3) *, p < 0.05.

DISCUSSION

HPGDS/PGD^sub 2^/DP^sub 1^ Signaling Pathway Plays Important Roles in Inflammatory Reactions Within the Senile Plaques

We demonstrated that expression of HPGDS and DP^sub 1^ was increased in reactive astrocytes and microglia encircling senile plaques in the AD brain. In contrast, the mRNA levels of all other prostanoid synthases and receptors were almost the same between AD and control groups. A previous study by Iwamoto et al (15) demonstrated that the amounts of PGD^sub 2^ and TXB^sub 2^ were significantly increased in the brains of Alzheimer-type dementia patients, although other prostaglandin metabolites in their AD group showed no significant changes from normal patients. In this study, we used a more specific and sensitive criterion for AD (9) and determined the levels of prostaglandin synthases and receptors instead of direct prostaglandin determination, because the latter is subject to a large postmortem artifact (16). In addition, we chose the most recent cases with the least postmortem interval for more accurate mRNA analysis. Ages and the postmortem interval of the patients at autopsy were not significantly different between AD and control groups. Senile plaques have different characteristics depending on their stages (17); for example, there are primitive, neuritic (classic), and burned out class of plaques in various stage of human diseases. alpha-Macroglobulin and anti-chymotrypsin are expressed in the neuritic-type plaques but not in the burned out ones (18, 19). Although plaques homologous with neuritic plaques were not recognized in Tg2576 mouse brains, we observed that HPGDS expression was limited mainly to the relatively small plaques with surrounding activated microglia and was diminished in the burned out class of plaques in both human and mouse AD brains. These observations may explain why some patients did not show HPGDS mRNA upregulation and may suggest that HPGDS is responsible for early inflammation caused by Abeta accumulation in the evolution of amyloid plaques.

The question as to which is first, plaques or inflammation, still remains to be clarified. We previously demonstrated that secondary demyelination was suppressed by inhibition of PGD^sub 2^ production (8). We consider, therefore, that once the plaque is formed, microglia and astrocytes are activated and produce cytokines, which cause further neuronal injury and plaque formation. To clarify this, additional investigation is needed.

New Potential Therapies for Alzheimer Disease Based on Inhibition of the Hematopoietic Prostaglandin D Synthase/Prostaglandin D^sub 2^/ DP^sub 1^ Signaling Pathway

Many epidemiologic and animal studies have revealed that nonsteroidal anti-inflammatory drugs (NSAIDs) are beneficial for AD patients (20-24). For example, a large, prospective and population- based cohort study confirmed that the relative risk of developing AD was significantly reduced in long-term users of NSAIDs compared with nonusers (25). Furthermore, orally administered ibuprofen produced significant diminution of the ultimate number of amyloid deposits as well as significantly reduced gliosis in Tg2576 mice (24).

PGD^sub 2^ is formed from arachidonic acid, which is released from the lipid portion of membranes by cPLA^sub 2^, by successive enzyme reactions mediated by COX and PGD synthases. COXs, expression of which is upregulated in the AD brain (26, 27), are the major targets of NSAIDs. NSAIDs may reduce the plaque pathology through either their antiinflammatory actions or suppression of gamma- secretase activity (28, 29) or both. The latter reduces the level of the 42-amino acid form of Abeta protein (Abeta42), the most toxic form of the protein. NSAIDs inhibit the constitutively expressed COX- 1 as well as the inducible COX-2, both of which produce PGH^sub 2^, a common precursor of PGD^sub 2^, PGE^sub 2^, PGF^sub 2alpha^, prostacyclin (PGI^sub 2^), and TXA^sub 2^ (30). PGD^sub 2^ is well known as an inflammatory mediator: it augments vascular permeability (31), regulates chemotaxis (32) and antigen presentation (33), and inhibits platelet aggregation (34). The biologic actions of PGD^sub 2^ are elicited through binding to its receptors, DP^sub 1^ (35) or DP^sub 2^ (32). Regarding the anti-inflammatory property of NSAIDs, 3 molecular targets of these drugs have been identified so far in AD: COX-1, COX-2, and PPARgamma. NSAIDs not only inhibit the activity of COXs but also activate the nuclear PPARgamma (36), which inhibits the production of pro-inflammatory cytokines. All 3 of these NSAID targets seem to be expressed by activated microglia (5). In this study, the level of PPARgamma was not changed in AD brains, although this finding does not exclude the possibility of PPARgamma- mediated neuroprotection by NSAIDs.

COX-2 has been the major target of anti-inflammatory therapy and a previous study showed an approximately 25% increase in the COX-2 level in the frontal cortex of AD brains (n = 17) compared with that in control brains (n = 12), as determined by densitometric analysis of Northern blots (37). We detected upregulation of COX-2 mRNA in Tg2576 mouse brains but not in the 17 AD frontal cortices by quantitative RT-PCR analysis. Ueno et al (38) reported that HPGDS preferentially utilized COX-1 in the A23187-induced immediate response for PGD^sub 2^ synthesis. Therefore, which COX was upregulated along with HPGDS induction needs to be clarified. Although an inflammatory response can have either a potentially beneficial or detrimental outcome, the administration of an HPGDS inhibitor, HQL-79, to the twitcher mouse, which is an animal model of Krabbe disease, remarkably ameliorated their astrogliosis as well as their clinical symptoms (8). HPGDS inhibitors and DP^sub 1^ antagonists have already been developed as antiallergic drugs (39), and HQL-79 is an orally active inhibitor of HPGDS (40, 41) that can penetrate through the blood-brain barrier. These drugs will specifically inhibit the PGD^sub 2^ signaling pathway, whereas NSAIDs may inhibit the production of all prostanoids, including the neuroprotective PGE^sub 2^ (42, 43).

In conclusion, we propose that PGD^sub 2^ is a novel inflammatory mediator in the AD brain and that the efficacy of HPGDS inhibitors and DP^sub 1^ antagonists should be investigated for halting the devastating course of AD.

ACKNOWLEDGMENTS

We thank Dr. Shinichi Kousaka, National Center of Neurology and Psychiatry, Japan, for the generous gift of the anti-Iba1 antibody. We thank Ms. Shigeko Matsumoto for performing the immunocytochemical analysis and Ms. Megumi Yamaguchi, Ms. Megumi Yamada, and Ms. Taeko Nishimoto for their secretarial assistance.

REFERENCES

1. Selkoe DJ. Alzheimer's disease: Genes, proteins, and therapy. Physiol Rev 2001;81:741-66

2. Griffin WS, Stanley LC, Ling C, et al. Brain interleukin 1 and S-100 immunoreactivity are elevated in Down syndrome and Alzheimer disease. Proc Natl Acad Sci U S A 1989;86:7611-15

3. Rogers J, Luber-Narod J, Styren SD, et al. Expression of immune system-associated antigens by cells of the human central nervous system: Relationship to the pathology of Alzheimer's disease. Neurobiol Aging 1988;9:339-49

4. Pike CJ, Cummings BJ, Cotman CW. Early association of reactive astrocytes with senile plaques in Alzheimer's disease. Exp Neurol 1995; 132:172-79

5. Akiyama H, Barger S, Barnum S, et al. Inflammation and Alzheimer's disease. Neurobiol Aging 2000;21:383-421

6. Hsiao K, Chapman P, Nilsen S, et al. Correlative memory deficits, Abeta elevation, and amyloid plaques in transgenic mice. Science 1996;274: 99-102

7. Kawarabayashi T, Younkin LH, Saido TC, et al. Age-dependent changes in brain, CSF, and plasma amyloid (beta) protein in the Tg2576 transgenic mouse model of Alzheimer's disease. J Neurosci 2001;21:372-81

8. Mohri I, Taniike M, Taniguchi H, et al. Prostaglandin D^sub 2^- mediated microglia/astrocyte interaction enhances astrogliosis and demyelination in twitcher. J Neurosci 2006;26:4383-93

9. Murayama S, Saitoh Y. Neuropathological diagnostic criteria for Alzheimer disease. Neuropathology 2004;24:254-60

10. Hughes CP, Berg L, Danziger WL, et al. A new clinical scale for the staging of dementia. Br J Psychiatry 1982;140:566-72

11. Lazarus M, Kubata BK, Eguchi N, et al. Biochemical characterization of mouse microsomal prostaglandin E synthase-1 and its colocalization with cyclooxygenase-2 in peritoneal macrophages. Arch Biochem Biophys 2002;397:336-41

12. Mohri I, Eguchi N, Suzuki K, et al. Hematopoietic prostaglandin D synthase is expressed in microglia in the developing postnatal mouse brain. Glia 2003;42:263-74

13. Forman BM, Tontonoz P, Chen J, et al. 15-Deoxy-delta 12,14- prostaglandin J^sub 2^ is a ligand for the adipocyte determination factor PPARgamma. Cell 1995; 83:803-12

14. Kagitani-Shimono K, Mohri I, Oda H, et al. Lipocalin-type prostaglandin D synthase (beta-trace) is upregulated in the alphaB- crystallin-positive oligodendrocytes and astrocytes in the chronic multiple sclerosis. Neuropathol Appl Neurobiol 2006;32:64-73

15. Iwamoto N, Kobayashi K, Kosaka K. The formation of prostaglandins in the postmortem cerebral cortex of Alzheimer-type dementia patients. J Neurol 1989;236:80-84

16. Narumiya S, Ogorochi T, Nakao K, et al. Prostaglandin D^sub 2^ in rat brain, spinal cord and pituitary: Basal level and regional distribution. Life Sci 1982;31:2093-2103

17. Probst A, Brunnschweiler H, Lautenschlager C, et al. A special type of senile plaque, possibly an initial stage. Acta Neuropathol (Berl) 1987; 74:133-41

18. Van Gool D, De Strooper B, Van Leuven F, et al. alpha^sub 2^- Macroglobulin expression in neuritic-type plaques in patients with Alzheimer's disease. Neurobiol Aging 1993;14:233-37

19. Van Gool D, Carmeliet G, Triau E, et al. Appearance of localized immunoreactivity for the alpha4 integrin subunit and for fibronectin in brains from Alzheimer's, Lewy body dementia patients and aged controls. Neurosci Lett 1994;170:71-73

20. Jenkinson ML, Bliss MR, Brain AT, et al. Rheumatoid arthritis and senile dementia of the Alzheimer's type. Br J Rheumatol 1989;28:86-88

21. Rogers J, Kirby LC, Hempelman SR, et al. Clinical trial of indomethacin in Alzheimer's disease. Neurology 1993;43:1609-11

22. Rich JB, Rasmusson DX, Folstein MF, et al. Nonsteroidal anti- inflammatory drugs in Alzheimer's disease. Neurology 1995;45:51-55

23. McGeer PL, Schulzer M, McGeer EG. Arthritis and anti- inflammatory agents as possible protective factors for Alzheimer's disease: A review of 17 epidemiologic studies. Neurology 1996;47:425- 32

24. Lim GP, Yang F, Chu T, et al. Ibuprofen suppresses plaque pathology and inflammation in a mouse model for Alzheimer's disease. J Neurosci 2000;20:5709-14

25. in t' Veld BA, Ruitenberg A, Hofman A, et al. Nonsteroidal antiinflammatory drugs and the risk of Alzheimer's disease. N Engl J Med 2001;345:1515-21 26. Fiala M, Liu QN, Sayre J, et al. Cyclooxygenase-2-positive macrophages infiltrate the Alzheimer's disease brain and damage the blood-brain barrier. Eur J Clin Invest 2002;32:360-71

27. Hoozemans JJ, Bruckner MK, Rozemuller AJ, et al. Cyclin D1 and cyclin E are co-localized with cyclo-oxygenase 2 (COX-2) in pyramidal neurons in Alzheimer disease temporal cortex. J Neuropathol Exp Neurol 2002;61;678-88

28. Eriksen JL, Sagi SA, Smith TE, et al. NSAIDs and enantiomers of flurbiprofen target gamma-secretase and lower Abeta42 in vivo. J Clin Invest 2003;112:440-49

29. Weggen S, Eriksen JL, Sagi SA, et al. Abeta42-lowering nonsteroidal anti-inflammatory drugs preserve intramembrane cleavage of the amyloid precursor protein (APP) and ErbB-4 receptor and signaling through the APP intracellular domain. J Biol Chem 2003;278:30748-54

30. Smith WL, DeWitt DL, Garavito RM. Cyclooxygenases: Structural, cellular, and molecular biology. Annu Rev Biochem 2000;69:145-82

31. Flower RJ, Harvey EA, Kingston WP. Inflammatory effects of prostaglandin D^sub 2^ in rat and human skin. Br J Pharmacol 1976;56:229-33

32. Hirai H, Tanaka K, Yoshie O, et al. Prostaglandin D^sub 2^ selectively induces chemotaxis in T helper type 2 cells, eosinophils, and basophils via seven-transmembrane receptor CRTH2. J Exp Med 2001;193:255-61

33. Herve M, Angeli V, Pinzar E, et al. Pivotal roles of the parasite PGD^sub 2^ synthase and of the host D prostanoid receptor 1 in schistosome immune evasion. Eur J Immunol 2003;33:2764-72

34. Whittle BJ, Moncada S, Vane JR. Comparison of the effects of prostacyclin (PGI^sub 2^), prostaglandin E^sub 1^ and D^sub 2^ on platelet aggregation in different species. Prostaglandins 1978;16:373-88

35. Hirata M, Kakizuka A, Aizawa M, et al. Molecular characterization of a mouse prostaglandin D receptor and functional expression of the cloned gene. Proc Natl Acad Sci U S A 1994;91:11192-96

36. Lehmann JM, Lenhard JM, Oliver BB, et al. Peroxisome proliferator-activated receptors alpha and gamma are activated by indomethacin and other non-steroidal anti-inflammatory drugs. J Biol Chem 1997;272:3406-10

37. Pasinetti GM, Aisen PS. Cyclooxygenase-2 expression is increased in frontal cortex of Alzheimer's disease brain. Neuroscience 1998;87: 319-24

38. Ueno N, Murakami M, Tanioka T, et al. Coupling between cyclooxygenase, terminal prostanoid synthase, and phospholipase A^sub 2^. J Biol Chem 2001;276:34918-27

39. Matsuoka T, Hirata M, Tanaka H, et al. Prostaglandin D^sub 2^ as a mediator of allergic asthma. Science 2000;287:2013-17

40. Matsushita N, Aritake K, Takada A, et al. Pharmacological studies on the novel antiallergic drug HQL-79: II. Elucidation of mechanisms for antiallergic and antiasthmatic effects. Jpn J Pharmacol 1998;78:11-22

41. Aritake K, Kado Y, Inoue T, et al. Structural and functional characterization of HQL-79, an orally active, selective inhibitor for human hematopoietic prostaglandin D synthase. J Biol Chem 2006;281:15277-86

42. McCullough L, Wu L, Haughey N, et al. Neuroprotective function of the PGE^sub 2^ EP2 receptor in cerebral ischemia. J Neurosci 2004;24:257-68

43. Bilak M, Wu L, Wang Q, et al. PGE^sub 2^ receptors rescue motor neurons in a model of amyotrophic lateral sclerosis. Ann Neurol 2004;56:240-48

Ikuko Mohri, MD, PhD, Keiichi Kadoyama, PhD, Takahisa Kanekiyo, MD, Yo Sato, PhD, Kuriko Kagitani-Shimono, MD, PhD, Yuko Saito, MD, PhD, Kinuko Suzuki, MD, Takashi Kudo, MD, PhD, Masatoshi Takeda, MD, PhD, Yoshihiro Urade, PhD, Shigeo Murayama, MD, PhD, and Masako Taniike, MD, PhD

From Department of Mental Health and Environmental Effects Research (IM, M. Taniike), The Research Center for Child Mental Development, Osaka University Graduate School of Medicine, Suita, Osaka, Japan; Department of Developmental Medicine (Pediatrics) (IM, TK, KK-S, M. Taniike), Osaka University Graduate School of Medicine, Suita, Osaka, Japan; Department of Molecular Behavioral Biology (IM, KK, Y. Sato, YU), Osaka Bioscience Institute, Suita, Osaka, Japan; Department of Pathology and Laboratory Medicine (IM, KS), School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina; Departments of Neuropathology and Pathology (Y. Saito, SM), Tokyo Metropolitan Geriatric Hospital, Itabashi-ku, Tokyo, Japan; and Division of Psychiatry and Behavioral Proteomics (TK, M. Takeda), Department of Post-Genomics and Diseases, Course of Advanced Medicine, Osaka University, Graduate School of Medicine, Suita, Osaka, Japan.

Drs. Mohri, Kadoyama, and Kanekiyo contributed equally to this work.

Drs. Urade, Murayama, and Taniike contributed equally in supervising this project.

Send correspondence and reprint requests to: Masako Taniike, MD, PhD, Department of Mental Health and Environmental Effects Research, The Research Center for Child Mental Development, Osaka University Graduate School of Medicine, 2-2, D-5, Yamadaoka, Suita, Osaka 565- 0871, Japan; E-mail: masako@ped.med.osaka-u.ac.jp

This study was supported in part by research grants from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (M. Taniike); the Osaka Medical Research Foundation for Incurable Diseases (M. Taniike), Japan Science and Technology Corporation (YU); Takeda Science Foundation (YU); Aid for Scientific Research on Priority Areas Advanced Brain Science Project from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (SM, Y. Saito); Aid for Longevity Sciences from the Ministry of Health, Labor, and Welfare (SM); and Aid for Long-Term Comprehensive Research on Age-Associated Dementia from Tokyo Metropolitan Institute of Gerontology (SM), and Osaka City.

Copyright Lippincott Williams & Wilkins Jun 2007

(c) 2007 Journal of Neuropathology and Experimental Neurology. Provided by ProQuest Information and Learning. All rights Reserved.