Address correspondence and reprint requests to Ken-ichi Hosoya, Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, Toyama 930-0194, Japan. E-mail: firstname.lastname@example.org
An increasing level of prostaglandin (PG) E2 is involved in the progression of neuroinflammation induced by ischemia and bacterial infection. Although an imbalance in the rates of production and clearance of PGE2 under these pathological conditions appears to affect the concentration of PGE2 in the cerebrospinal fluid (CSF), the regulatory system remains incompletely understood. The purpose of this study was to investigate the cellular system of PGE2 production via microsomal PGE synthetase-1 (mPGES-1), the inducible PGE2-generating enzyme, and PGE2 elimination from the CSF via the blood–CSF barrier (BCSFB). Immunohistochemical analysis revealed that mPGES-1 was expressed in the soma and perivascular sheets of astrocytes, pia mater, and brain blood vessel endothelial cells, suggesting that these cells are local production sites of PGE2 in the CSF. The in vivo PGE2 elimination clearance from the CSF was eightfold greater than that of d-mannitol, which is considered to reflect CSF bulk flow. This process was inhibited by the simultaneous injection of unlabeled PGE2 and β-lactam antibiotics, such as benzylpenicillin, cefazolin, and ceftriaxone, which are substrates and/or inhibitors of organic anion transporter 3 (OAT3). The characteristics of PGE2 uptake by the isolated choroid plexus were at least partially consistent with those of OAT3. OAT3 was able to mediate PGE2 transport with a Michaelis–Menten constant of 4.24 μM. These findings indicate that a system regulating the PGE2 level in the CSF involves OAT3-mediated PGE2 uptake by choroid plexus epithelial cells, acting as a cerebral clearance pathway via the BCSFB of locally produced PGE2.
Prostaglandin (PG) E2 is the crucial mediator, which propagates neuroinflammation induced by ischemia and bacterial infection. The PGE2 level is significantly increased in the Cerebrospinal fluid (CSF) of the patients suffering from stroke (0.57–8.5 nM; Carasso et al. 1977) and lipopolysaccharide (LPS)-treated rats (~3.4 nM; Gao et al. 2009), although the PGE2 concentration in normal CSF is below the detection limit in humans (Romero et al. 1984) and 0.15 nM in rats (Gao et al. 2009). As a positive correlation has been found between the PGE2 level in the CSF and the severity and clinical outcome of the stroke (Carasso et al. 1977), the CSF concentration of PGE2 appears to be a key determinant of the progression of neuroinflammation.
The PGE2 accumulation in the CSF appears to be linked to an imbalance in the rates of production and clearance of PGE2 in the CSF, that is, the increased biosynthesis and/or decreased elimination of PGE2 from the CSF. PGE2 is associated with PGH2 produced by the three prostaglandin E synthetase (PGES) isozymes, that is, cytosolic PGES (cPGES), microsomal PGES-1 (mPGES-1), and microsomal PGES-2 (mPGES-2). Although cPGES and mPGES-2 are constitutively expressed in various cells and tissues, the expression of mPGES-1 is predominantly induced by proinflammatory stimuli and transient ischemia, as shown in various models of inflammation (Murakami et al. 2000, 2002; Ikeda-Matsuo et al. 2006). This suggests that the induction of mPGES-1 is a critical step for PGE2 accumulation in the CSF, thereby exacerbating the neuroinflammation. Indeed, the deletion of the mPGES-1 gene results in marked amelioration of the infarction, edema, and behavioral neurological dysfunctions, which are caused by middle cerebral artery occlusion. On the other hand, a clearance system for PGE2 from the CSF is essential for maintaining the low PGE2 level in the CSF. PGE2 seems unlikely to be inactivated by 15-hydroxyprostaglandin dehydrogenase, the rate-limiting enzyme of PG catabolism, in the brain because this enzyme exhibits little expression and activity in the adult brain parenchyma and choroid plexus (Nakano et al. 1972; Alix et al. 2008). Accordingly, it is conceivable that the primary pathway for removing PGE2 from the CSF is the CSF-to-blood vectorial transport across the blood–CSF barrier (BCSFB), which is formed by the complex tight junctions of choroid plexus epithelial cells in the ventricles (Hosoya et al. 2004). It has been reported that the PGE2 level in the CSF falls markedly 5 h after intraperitoneal injection of LPS, in spite of the continuous elevation of mPGES-1 expression, although the initial elevation of PGE2 in the CSF (Inoue et al. 2002) coincides with the elevated expression of mPGES-1 in the brain. These lines of evidence prompted us to hypothesize that the maintenance of the low PGE2 level in the CSF would depend on the rapid clearance of PGE2 across the BCSFB rather than the expression level of mPGES-1.
Because PGE2 (pKa = ~5) exists predominantly in charged form at physiological pH, a carrier-mediated process rather than passive diffusion is likely to produce the PGE2 efflux transport across the BCSFB in the CSF-to-blood direction. Cellular transport of PGE2 is mediated by a variety of transporters, that is, organic anion transporters (OATs) (Sekine et al. 1997, 1998; Kimura et al. 2002; Nilwarangkoon et al. 2007; Shiraya et al. 2010), organic cation transporters (OCTs) (Kimura et al. 2002), and organic anion-transporting polypeptides (oatps) (Kanai et al. 1995; Masuda et al. 1999; Cattori et al. 2001). Among these transporters for PGE2, OAT3 (SLC22A3) (Nagata et al. 2002), prostaglandin transporter (PGT/SLCO2A1) (Adachi et al. 2003; Kis et al. 2006), and oatp1a5 (SLCO1A5) (Kusuhara et al. 2003; Ohtsuki et al. 2004) are expressed in the choroid plexus. These transporters are localized on the brush-border membrane of choroid plexus epithelial cells (Nagata et al. 2002; Ohtsuki et al. 2003, 2004; Tachikawa et al. 2012, in press). The elimination process via transporter(s) at the BCSFB should be responsible for the accumulation of PGE2 in the CSF, being a therapeutic target for drugs which can prevent PGE2 accumulation in the CSF.
The purpose of this study was (i) to clarify the cellular expression of the inducible PGE2-generating enzyme, mPGES-1, by immunohistochemistry, (ii) to investigate PGE2 elimination from the CSF across the BCSFB by means of the intracerebroventricular administration method and uptake by the isolated choroid plexus, and (iii) to identify which transporter is responsible for PGE2 transport using a Xenopus laevis oocyte expression system.
Materials and methods
Adult male ddY mice (25–30 g), Wistar rats (260–280 g), and female Hartley guinea pigs (300–350 g) were purchased from Japan SLC (Hamamatsu, Japan). Mature female Xenopus laevis were purchased from Kato-S-Science (Chiba, Japan). A mouse model of peripheral inflammation was obtained by intraperitoneal administration of 3.0 mg/kg lipopolysaccharide (LPS) of Escherichia coli 0111:B4 (Sigma Aldrich, St Louis, MO, USA) 24 h prior to experiments. They were maintained in a controlled environment and all experiments were approved by the Animal Care Committee, University of Toyama.
[5,6,8,11,12,14,15-3H(N)]Prostaglandin E2 ([3H]PGE2, 200 Ci/mmol) was obtained from Perkin-Elmer Life and Analytical Sciences (Boston, MA, USA). n-[1-14C]Butanol ([14C]butanol, 5.0 mCi/mmol) and d-[1-3H(N)]mannitol ([14C]d-mannitol, 55 mCi/mmol) were obtained from American Radiolabeled Chemicals (St Louis, MO, USA). All other chemicals were commercial products of analytical grade.
Immunoblotting and Immunohistochemistry
A polyclonal antibody to mPGES-1 was raised against amino acid residues 38–76 of mouse mPGES-1 (GenBank accession number: NP_071860) as described previously (Tachikawa et al. 2004), and its specificity was examined by immunoblot for mouse tissues and Chinese hamster ovary (CHO) cells over-expressing mouse mPGES-1. The detailed methods of immunoblot and construction of the mPGES-1 over-expressing CHO cells are included as supporting online material (supporting methods). For immunohistochemistry, under deep pentobarbital anesthesia (100 mg/kg body weight, i.p.), adult mice were killed after transcardial fixation with 4% paraformaldehyde in 0.1 M sodium phosphate buffer (PB, pH 7.4). The brains were removed and immersed in 30% sucrose in 0.1 M PB solution. Frozen sections (40 μm in thickness) were prepared on a cryostat (CM1900; Leica, Nussloch, Germany). The sections were immunoreacted overnight with guinea pig antibody to mPGES-1 (2 μg/mL) singly or in combination with rabbit glucose transporter 1 (GLUT1) antibody [0.5 μg/mL, (Sakai et al. 2003)], rabbit glutamine synthetase (GS) antibody [1 μg/mL, (Nakashima et al. 2005)], and rabbit glial fibrillary acidic protein (GFAP) antibody (1 : 50 dilution, DAKO, Carpinteria, CA, USA). Subsequently, they were incubated with species-specific Alexa Fluor 488- (Invitrogen, Carlsbad, CA, USA) and Cy3-conjugated secondary antibodies for 2 h (Jackson ImmunoResearch, West Grove, PA, USA). Photographs were taken using a confocal laser scanning microscope (TCS-SP5; Leica).
In vivo PGE2 efflux from the CSF after intracerebroventricular administration
[3H]PGE2 elimination from the CSF after intracerebroventricular administration was studied using the method described previously (Tachikawa et al. 2008a). The details of the methods and the data analyses are included as supporting online material (supporting methods).
PGE2 uptake using freshly isolated rat choroid plexus
The uptake of [3H]PGE2 by rat choroid plexus was examined using the centrifugal filtration method described previously (Tachikawa et al. 2008a,b). The rats were decapitated and the choroid plexus was isolated from the lateral ventricles and incubated at 37°C for 1 min in 300 μL ECF buffer. Incubation medium containing [3H]PGE2 in the presence or absence of inhibitors was added to initiate uptake. The final concentration of the incubation medium was [3H]PGE2 (50 nM) and [14C]butanol (300 μM), respectively. The radioactivity in the specimens was determined using a liquid scintillation spectrophotometer (LSC-5000; Aloka, Tokyo, Japan). The tissue-to-medium (tissue/medium) concentration ratio was calculated from the following eqn 1 using [14C]butanol to determine the cell volume of the choroid plexus (μL choroid plexus).
Uptake using rat OAT3-expressing Xenopus laevis oocytes
Rat OAT3-mediated uptake of [3H]PGE2 was examined as described previously (Mori et al. 2003). The detailed methods of the uptake study are included as supporting online material (supporting methods).
The kinetic parameters for PGE2 uptake by the choroid plexus were obtained from the following eqn 2:
The kinetic parameters for PGE2 uptake by OAT3-expressing Xenopus oocytes were obtained from the following eqn 3:
where V is the uptake rate of PGE2, S is the PGE2 concentration in the medium, Km is the Michaelis–Menten constant, Vmax is the maximum uptake rate, and Kd is the non-saturable transport clearance. To obtain kinetic parameters, the equation was fitted using the iterative non-linear least-squares regression analysis program, MULTI (Yamaoka et al. 1986).
Unless otherwise indicated, all data are presented as the mean ± SEM. The kinetic parameters are presented as the mean ± SD. An unpaired, two-tailed Student's t-test was used to determine the significance of differences (p < 0.01) between two group means. One-way analysis of variance followed by the modified Fisher's least-squares difference method was used to assess the statistical significance of differences (p < 0.01) among means of more than two groups.
Expression of mPGES-1 in the adult brain
Using immunoblotting with the crude membrane fraction from CHO cells transiently expressing mouse mPGES-1, the antibody recognized a single band at 19 kDa, whereas no band was seen in non-transfected CHO cells (Fig. 1a). A single band at 17 kDa was detected in the brain as in the kidney used as a positive control for mPGES-1 expression (Fig. 1a), which is consistent with the previous result (Yamagata et al. 2001). Because Xpress epitope-tag (~1 kDa) and Histidine-tag (~1 kDa) were added to mPGES-1 protein expressed in CHO cells, the size of the band from CHO cells transiently expressing mPGES-1 was approximately 2 kDa higher than that from the brain and kidney. Moreover, by pre-absorbing anti-mPGES-1 antibody with the antigen peptides, these bands were abolished completely (Fig. 1b). The intensity of the band from the brain of LPS-treated mice was increased compared with that of control mice (Fig. 1c), indicating the induced expression of mPGES-1 under proinflammatory conditions. From these results, the guinea pig mPGES-1 antibody was judged to be specific.
The cellular expression of mPGES-1 in the adult brain was characterized using immunohistochemistry. The guinea pig mPGES-1 antibody strongly labeled small cells (arrowheads; Fig. 2a–c) which were dispersed throughout the cerebral cortex, blood vessel-like structures (arrows; Fig. 2a–d), and the pia mater (double arrowheads; Fig. 2a and b). The blood vessel-like signals were most likely confined to the outer side of crescent-shaped blood vessel endothelium nuclei (arrows, Fig. 2c inset), although they were occasionally found both on the inner and outer sides of the endothelium nuclei (arrows, Fig. 2d). These signals were completely abolished by pre-absorbing anti-mPGES-1 antibody with the antigen peptides (Fig. 2e). Using double immunofluorescence for GS, a marker of astrocytes, the mPGES-1 immunoreactivities showed extensive overlapping with those of GS (arrowheads, Fig. 2f). The expression of mPGES-1 was mostly negative in GLUT1-positive blood vessel endothelial cells (arrows, Fig. 2g), although mPGES-1 was found in a small number of GLUT1-positive blood vessels (double arrowheads, Fig. 2h). The intense immunostaining of mPGES-1 was present in perivasucular sheets of astrocytes (arrows, Fig. 2g). These results indicate that mPGES-1 is preferentially localized on cell bodies and perivascular sheets of astrocytes under normal conditions. At 24 h after intraperitoneal administration of LPS, the mPGES-1 immunostaining was also found in GFAP-positive soma (arrowheads, Fig. 2i and l) and perivascular sheets (arrows, Fig. 2j–l) of astrocytes. The mPGES-1 expression was up-regulated in GLUT1-positive blood vessels (double arrowheads, Fig. 2j–k) by LPS administration, although not all blood vessels were positive for mPGES-1 (arrows, Fig. 2l).
Elimination of PGE2 from rat CSF after intra-cerebroventricular administration
Figure 3a shows the residual CSF concentration of [3H]PGE2 and [14C]d-mannitol after intracerebroventricular administration as a function of time. [3H]PGE2 was eliminated from the CSF with a higher rate constant of 0.204 ± 0.012/min than that of [14C]d-mannitol (0.030 ± 0.012/min), a reference compound for CSF turnover and diffusion into the brain interstitial space through the ependymal space. The half-life of PGE2 elimination from the CSF was estimated to be 3.40 min. The elimination clearance of PGE2 from the CSF (39.5 μL/min) was approximately eightfold higher than that of d-mannitol (4.79 μL/min). The elimination clearance of d-mannitol was close to the CSF bulk flow rate (2.9 μL/min) obtained by Suzuki et al. (1985). As shown in Fig. 3b and c, the simultaneous injection of unlabeled PGE2 (0.2 mM as a concentration in the CSF), benzylpenicillin (1 mM), cefazolin (1 mM), and ceftriaxone (5 mM) with [3H]PGE2 resulted in a 1.5- to 2.6-fold increase in the residual concentration of [3H]PGE2 compared with that in the absence of inhibitors at 5 min.
Characteristics of PGE2 uptake by isolated rat choroid plexus
To examine the contribution of the BCSFB to PGE2 eliminated from the CSF, the study of [3H]PGE2 uptake by the isolated rat choroid plexus was performed. [3H]PGE2 uptake by the choroid plexus exhibited a time-dependent increase linearly for up to a 2-min incubation with an initial uptake rate of 1.32 ± 0.35 μL/(min μL choroid plexus) (Fig. 4a). The [3H]PGE2 uptake was concentration dependent and composed of saturable and non-saturable components (Fig. 4b). The apparent Km and Vmax values of the saturable component were found to be 23.0 ± 8.0 μM and 58.7 ± 20.4 pmol/(min μL choroid plexus), respectively. The uptake clearance (Kd) of the non-saturable component was 0.852 ± 0.122 μL/(min μL choroid plexus). To characterize the transporter(s) involved in [3H]PGE2 uptake by the isolated choroid plexus, the inhibitory effects of various compounds on the [3H]PGE2 uptake were investigated (Table 1). PGE2, PGB1, diclofenac, and bromocresolgreen produced a marked inhibition (more than 60%) at a concentration of 0.16–1 mM. Benzylpenicillin, taurocholate, cefazolin, and cefmetazole inhibited the [3H]PGE2 uptake by 50–60% at a concentration of 1 mM. p-Aminohippuric acid (PAH) had a weaker effect at a concentration of 1 mM. Digoxin had no significant effect on the uptake at a concentration of 0.1 mM. These results indicate an involvement of an organic anion-preferring carrier-mediated process in the [3H]PGE2 uptake by the choroid plexus.
Table 1. Effect of several compounds on [3H]PGE2 uptake by isolated rat choroid plexus
Concentration of inhibitor (mM)
% of control
Choroid plexus was incubated with [3H]PGE2 (50 nM) in the absence (control) or presence of inhibitors at 37°C for 1 min. Each value represents the mean ± SEM (n = 3). *p < 0.01, significantly different from control.
Characteristics of OAT3-mediated [3H]PGE2 uptake in rat OAT3-expressing Xenopus oocytes
The characteristics of OAT3-mediated PGE2 transport were investigated using OAT3-expressing oocytes (OAT3/oocytes). [3H]PGE2 was taken up in a time-dependent manner by OAT3/oocytes (Fig. 5a), and the uptake in these oocytes was several-fold higher than in water-injected oocytes. [3H]PGE2 uptake by OAT3/oocytes was inhibited by unlabeled PGE2 at a concentration of 10 μM (Table 2). As shown in Fig. 5b, PGE2 uptake by OAT3/oocytes exhibited saturable kinetics with a Km of 4.24 ± 0.79 μM and a Vmax of 1.25 ± 0.16 pmol/(h oocyte). The inhibitory effects of OAT3 substrates and inhibitors on the [3H]PGE2 uptake were further examined (Table 2). PGB1, diclofenac, and indomethacin exhibited a marked inhibition (over 77%) at a concentration of 0.1 mM. Benzylpenicillin, bromocresolgreen, bromosulfophthalein, and dehydroepiandrosterone-3-sulfate (DHEAS) all inhibited the [3H]PGE2 uptake by more than 66% at a concentration of 1 mM, whereas PAH had a much weaker effect.
Table 2. Effect of several compounds on [3H]PGE2 uptake by Xenopus oocytes expressing rat organic anion transporter 3 (OAT3/oocytes)
Concentration of inhibitor (mM)
% of control
[3H]PGE2 uptake (1 nM) by OAT3/oocytes was measured at 20°C for 1 h in the absence (control) or presence of inhibitors at the indicated concentration. Each value represents the mean ± SEM (n = 5–8). *p < 0.01, significantly different from control.
This study demonstrates that a regulatory system for the PGE2 level in the CSF involves OAT3-mediated PGE2 uptake by choroid plexus epithelial cells, acting as a clearance pathway via the BCSFB of PGE2 produced in the brain.
The immunohistochemical data reveal that mPGES-1 is expressed in the pia mater which faces the CSF (Fig. 2a and b), suggesting a local production of PGE2 in the CSF. Because cPGES and mPGES-2 are constitutively expressed in brain parenchymal cells (Vazquez-Tello et al. 2004; Chaudhry et al. 2010), cPGES and mPGES-2 can also contribute to the local PGE2 production in the brain. [3H]PGE2, after intracerebroventricular administration, was eliminated from the CSF at a rate eightfold higher than that of [14C]d-mannitol, a CSF bulk-flow marker (Fig. 3a). This is in good agreement with a previous report demonstrating an active elimination of PGF2α from rabbit CSF using ventriculo-cisternal perfusion (Bito et al. 1976). The rapid elimination of [3H]PGE2 from the CSF was inhibited by simultaneous injection of unlabeled PGE2 and benzylpenicillin (Fig. 3b). Furthermore, [3H]PGE2 undergoes concentrative uptake by the isolated choroid plexus (Fig. 4a). Considering that a carrier-mediated transport for organic anions is involved in benzylpenicillin uptake by the choroid plexus (Suzuki et al. 1987), the same uptake system appears to be involved in the elimination of PGE2 and benzylpenicillin from the CSF across the BCSFB. It has been reported that the PGE2 level in the CSF falls markedly 5 h after intraperitonal injection of LPS, in spite of the continuous elevation of mPGES-1 expression (Inoue et al. 2002). This report supports our hypothesis that the BCSFB plays a pivotal role in the efficient removal of PGE2 from the CSF even in the presence of elevated mPGES-1 expression.
The elimination clearance of PGE2 from the CSF via the BCSFB was estimated from the initial uptake rate of PGE2 by isolated choroid plexus to be 7.92 μL/min per rat [1.32 μL/(min μL choroid plexus) (Fig. 4a) × 6 μL (the total rat choroid plexus volume per rat; Ogawa et al. 1994)]. The PGE2 efflux transport via the BCSFB makes a 20.1% contribution to the total PGE2 elimination clearance from the CSF in vivo (39.5 μL/min per rat). As the elimination clearance of d-mannitol from the CSF was 4.79 μL/min per rat, 12.1% of the total PGE2 elimination clearance from the CSF would reflect the CSF bulk flow and diffusion into the brain interstitial space through the ependymal space. The remainder might correspond to active and rapid transport/binding of PGE2 in ependymal cells and/or brain parenchymal cells. This notion is supported by an earlier finding that intracerebroventricularly administered PGE2 induces fever by acting on the anterior hypothalamic pre-optic area (Splawinski et al. 1978).
The inhibition profiles of PGB1, diclofenac, bromocresolgreen, and PAH on [3H]PGE2 uptake by the isolated choroid plexus were almost identical to those of OAT3-mediated [3H]PGE2 uptake (Tables 1 and 2). However, it should be noted that the inhibitory effect of benzylpenicillin (1 mM) on the [3H]PGE2 uptake by the choroid plexus (50%; Table 1) is lower than that on OAT3-mediated [3H]PGE2 uptake (84%; Table 2). This discrepancy may be explained by the presence of an additional transporter involved in [3H]PGE2 uptake by the choroid plexus. It has been demonstrated that benzylpenicillin has little or no effect on oatp1a5-mediated [3H]17β-estradiol-d-17β-glucuronide (E217βG) uptake (Kusuhara et al. 2003) and PGT-mediated [3H]PGD2 uptake (Tachikawa et al. 2012; in press). This raises one possibility that oatp1a5 and/or PGT would be responsible for the benzylpenicillin-insensitive [3H]PGE2 uptake by the choroid plexus. On the other hand, taurocholate, a transportable substrate for OAT3 (Sweet et al. 2002) and oatp1a5 (Walters et al. 2000), inhibits the [3H]PGE2 uptake by 49% at a concentration of 1 mM. This is still less than the degree of inhibition by bromocresolgreen (84%; Table 1). Because it has been reported that bromocresolgreen at a concentration of 1 mM inhibits PGT-mediated [3H]PGD2 uptake more potently than OAT3-mediated [3H]PGD2 uptake (Tachikawa et al. 2012; in press), PGT would be responsible for the benzylpenicillin-insensitive and bromocresolgreen-sensitive [3H]PGE2 uptake by the choroid plexus. However, the estimated Km value of [3H]PGE2 uptake by the isolated choroid plexus (23 μM; Fig. 4b) is between the Km values of OAT3-mediated (4.24 μM; Fig. 5b) and oatp1a5-mediated [3H]PGE2 uptake (35 μM; Cattori et al. 2001), whereas it is very different from the Km value of PGT-mediated PGE2 transport (94 nM; Kanai et al. 1995). This may be because the Km values of OAT3-, oatp1a5-, and PGT-mediated [3H]PGE2 uptake were difficult to be distinguished by kinetic analysis. Taking these findings into consideration, OAT3 on the brush-border membrane of choroid plexus epithelial cells would be at least partly responsible for the PGE2 uptake by choroid plexus, although a contribution by oatp1a5 and PGT cannot be ruled out. The contribution of OAT3 is supported by the fact that cephalosporin antibiotics, such as cefazolin and ceftriaxone, which are inhibitors of rat and human OAT3 (Jung et al. 2002; Takeda et al. 2002), produce a marked inhibition of in vivo [3H]PGE2 elimination from the CSF (Fig. 3c) and/or [3H]PGE2 uptake by the choroid plexus (Table 1). It is necessary to clarify whether these cephalosporin antibiotics could affect oatp1a5- and/or PGT-mediated [3H]PGE2 uptake by the choroid plexus.
It has been reported that the polarized choroid plexus epithelial cells are involved in the transcellular transport of PGE2 in the apical-to-basolateral direction without the inactivation of PGE2 in the cells (Khuth et al. 2005). To excrete PGE2, which undergoes OAT3-mediated uptake by choroid plexus epithelial cells from the CSF into the circulating blood, an efflux transporter should be present on the basolateral membrane of the choroid plexus epithelial cells. Multidrug resistance-associated protein 1 and 4/ATP-binding cassette transporter C1 and C4 (MRP1 and 4/ABCC1 and 4) are localized on the basolateral membrane of choroid plexus epithelial cells (Wijnholds et al. 2000; Leggas et al. 2004) and mediate PGE2 transport (Reid et al. 2003; de Waart et al. 2006). Thus, MRP1 and 4 would be excellent candidate transporters for mediating PGE2 efflux from the choroid plexus epithelial cells into the circulating blood.
The Km value of PGE2 uptake by the choroid plexus (23.0 μM; Fig. 4b) is almost four orders of magnitude greater than the PGE2 concentrations in the rat CSF under normal (1.2 nM) and inflammatory (~3.4 nM) conditions (Gao et al. 2009), suggesting that the PGE2 uptake without saturation produces continuous removal of PGE2 from the CSF. This notion is supported by the following facts (Gao et al. 2009): (i) the normal CSF concentration of PGE2 in rats (0.15 nM) is 55-fold lower than the normal plasma concentration of PGE2 (8.2 nM) and (ii) even although the plasma concentration of PGE2 is increased by intraperitoneal injection of LPS in rats, the PGE2 concentration in the CSF is still lower than that in plasma. However, the chronic inhibition of PGE2 clearance at the BCSFB may cause the accumulation of PGE2 in the CSF, thus exacerbating neuroinflammation in the brain. Indeed, exposure of the epithelial cells with T lymphocytes activated by a retroviral infection reduces the transcellular transport of PGE2 in the apical-to-basolateral direction (Khuth et al. 2005). Reduced uptake of benzylpenicillin by the choroid plexus has been reported in rats that received intracisternal LPS, an experimental model of bacterial meningitis (Han et al. 2002). As rat OAT3 is the transporter responsible for the uptake of benzylpenicillin (Nagata et al. 2002) and PGE2 in the choroid plexus, the OAT3-mediated PGE2 clearance from the CSF would be reduced as a result of inflammation, leading to the continuous accumulation of PGE2 in the CSF. Therefore, OAT3 at the BCSFB is an important factor governing the CSF concentration of PGE2.
Cephalosporin antibiotics and non-steroidal anti-inflammatory drugs (NSAIDs) are clinically used to treat infections and autoimmune responses such as fever (Molavi 1991; Kim et al. 2009). This study suggests that cefazolin, ceftriaxone, cefmetazole, and diclofenac inhibit PGE2 elimination from the CSF and/or PGE2 uptake by the choroid plexus (Fig. 3c, Table 1). We have also found that cefazolin and cefmetazole, which are administered intravenously at a clinically relevant blood concentration in mice, inhibit the brain-to-blood [3H]PGE2 efflux transport across the blood–brain barrier (BBB) because of the inhibition of MRP4 (Akanuma et al. 2010). Thus, cephalosporins and NSAIDs may attenuate PGE2 efflux transport at the BCSFB as well as the BBB, exacerbating neuroinflammation. This may explain the mechanism of adverse effects such as encephalitis, induced by some cephalosporins and NSAIDs (Schliamser et al. 1991; Sunden et al. 2003). On the other hand, although intravenous administration of ceftriaxone is clinically useful for the treatment of bacterial meningitis (Craig 1984; Scheld 1984), the Ki value of ceftriaxone for human OAT3-mediated [3H]estron-3-sulfate transport is estimated to be 4.39 mM (Takeda et al. 2002). Therefore, ceftriaxone does not inhibit PGE2 uptake by the choroid plexus at a clinically relevant concentration. From this viewpoint, the interaction between these drugs and PGE2 efflux transport at the BCSFB should be taken into consideration when choosing therapeutic drugs as well as in the development of new drugs.
Immunohistochemical analysis reveals that mPGES-1 is localized in the soma and perivascular sheets of astrocytes (Fig. 2). The astrocytic expression of mPGES-1 is also seen in the human brain (Chaudhry et al. 2008), whereas cPGES and mPGES-2 are not detected in astrocytes (Chaudhry and Dore 2009; Chaudhry et al. 2010). Gordon et al. (2008) have proposed that the increased Ca2+ concentration in astrocytes facilitates the production and release of PGE2 under low O2 conditions, leading to the accumulation of extracellular PGE2 and subsequent vasodilation (Gordon et al. 2008). Because the mPGES-1 expression is selectively induced in the brain after transient ischemia (Ikeda-Matsuo et al. 2006), mPGES-1 appears to be involved in astrocytic control of the cerebrovascular diameter. We have shown that [3H]PGE2 microinjected into the cerebral cortex undergoes brain-to-blood efflux transport across the BBB (Akanuma et al. 2010). In this regard, the interplay of mPGES-1-mediated PGE2 production in perivascular sheets of astrocytes and PGE2 clearance at the BBB seems to be an efficient way of terminating the vasodilation reaction in a neurovascular unit. The mPGES-1 expression is up-regulated in blood vessels by intraperitoneal injection of LPS (Fig. 2), which is consistent with previous results (Inoue et al. 2002). Inoue et al. (2002) have proposed that the mPGES-1-mediated production of PGE2 in blood vessels causes fever, an acute neuroinflammatory response. It thus appears that the increased production of PGE2 in the endothelial cells might affect the BBB function under inflammatory conditions.
In conclusion, OAT3 at the BCSFB is involved in the continuous removal of PGE2 from the CSF, thereby maintaining low PGE2 levels in the CSF. The present findings provide a novel insight into the production and clearance of PGE2 in the CSF and may be helpful in the development of new therapeutic targets for the treatment of neuroinflammation.
We would like to thank Drs T. Abe and T. Terasaki for supplying the pGEM-HEN vector for protein expression in Xenopus laevis oocytes. This study was supported, in part, by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture, and from the Japan Society for the Promotion of Science, Japan.