Prostaglandin E2 metabolism in rat brain: Role of the blood-brain interfaces

Background Prostaglandin E2 (PGE2) is involved in the regulation of synaptic activity and plasticity, and in brain maturation. It is also an important mediator of the central response to inflammatory challenges. The aim of this study was to evaluate the ability of the tissues forming the blood-brain interfaces to act as signal termination sites for PGE2 by metabolic inactivation. Methods The specific activity of 15-hydroxyprostaglandin dehydrogenase was measured in homogenates of microvessels, choroid plexuses and cerebral cortex isolated from postnatal and adult rat brain, and compared to the activity measured in peripheral organs which are established signal termination sites for prostaglandins. PGE2 metabolites produced ex vivo by choroid plexuses were identified and quantified by HPLC coupled to radiochemical detection. Results The data confirmed the absence of metabolic activity in brain parenchyma, and showed that no detectable activity was associated with brain microvessels forming the blood-brain barrier. By contrast, 15-hydroxyprostaglandin dehydrogenase activity was measured in both fourth and lateral ventricle choroid plexuses from 2-day-old rats, albeit at a lower level than in lung or kidney. The activity was barely detectable in adult choroidal tissue. Metabolic profiles indicated that isolated choroid plexus has the ability to metabolize PGE2, mainly into 13,14-dihydro-15-keto-PGE2. In short-term incubations, this metabolite distributed in the tissue rather than in the external medium, suggesting its release in the choroidal stroma. Conclusion The rat choroidal tissue has a significant ability to metabolize PGE2 during early postnatal life. This metabolic activity may participate in signal termination of centrally released PGE2 in the brain, or function as an enzymatic barrier acting to maintain PGE2 homeostasis in CSF during the critical early postnatal period of brain development.


Background
Prostaglandin E 2 (PGE 2 ) is a main product of the cyclooxygenase (Cox) pathway. Two Cox isoenzymes, Cox-1 and Cox-2, convert arachidonic acid released by phospholipases A 2 to PGH 2 , which in turn is metabolized by terminal prostaglandin E synthases into PGE 2 [1]. While Cox-1 is constitutively expressed in most tissues where it finetunes physiological processes [2], Cox-2 expression is very limited in normal conditions in peripheral organs. Yet it is induced by inflammatory stimuli, and then, being functionally coupled to microsomal prostaglandin E synthase 1, it plays a major role in the response to inflammation via PGE 2 production [3].
In the brain, our current understanding of PGE 2 metabolic cascade indicates that Cox-2 is constitutively expressed in several neuronal cell populations, especially in hippocampal and cortical glutamatergic neurons. The enzyme participates in synaptic activity, hippocampal long-term synaptic plasticity, and brain maturation (reviewed by Chen and Bazan, and Minghetti [4,5]). Inflammatory challenges trigger the cerebral upregulation of Cox 2, particularly in venule endothelial cells, and the subsequent production of PGE 2 . The prostaglandin acts as a modulator of the sickness behavior syndrome, and specifically induces fever via hypothalamic EP3 receptor activation [6,7]. The efficiency and length of the biological response to PGE 2 is dependent upon the balance between its production and its inactivation. In peripheral organs, the first step of the inactivation process is mediated by NAD +dependent 15-hydroxyprostaglandin dehydrogenase (PGDH) [8]. This enzyme is particularly active in the lung or the kidney. By contrast, it is considered absent in the brain of rodent and other mammalian species, from late gestation throughout postnatal life [1,9,10]. PGE 2 catabolism has however been reported specifically in the choroid plexus of sheep during postnatal development and to some extent in adulthood [11]. Immunohistochemical evidence for the presence of PGDH in lamb choroid plexus exists also [12]. The choroidal tissue constitutes a major interface between the cerebrospinal fluid (CSF) and the blood, and in conjunction with the cerebral capillaries regulates the exchanges between the blood and the central nervous system. The mechanisms responsible for this crucial regulation are multiple and involve barrier and transport, as well as metabolic processes towards biologically active endogenous compounds.
We investigated in rats, whether the cells forming the blood-brain interfaces are a site of PGE 2 metabolism into inactive compounds and as such, of signal termination. We isolated cerebral capillaries and choroid plexuses from developing and adult rat brain, measured PGDH activity in these tissues, and identified the metabolites actually produced from PGE 2 .

Animals and tissue isolation
Animal care and procedures have been conducted according to the guidelines approved by the French Ethical Committee (decree 87-848) and by the European Community directive 86-609-EEC. Rats, 200-240 g, Sprague-Dawley males or timed pregnant females were obtained from Harlan, Gannat, France. Following halothane anesthesia and decapitation of the animals, rat brains were removed and the choroid plexuses were sampled intact under a stereomicroscope, briefly rinsed in Ringer-Hepes (RH) buffer [13], and kept at -80°C until used for enzymatic measurement. Kidney, lung and meninges-free brain cortex were also sampled. In some experiments freshly isolated intact choroid plexuses from both adult and 2-day-old rats were kept in RH buffer at 37°C for metabolic analysis. Microvessels from 9-day-old and adult brain cortices were isolated at 4°C in oxygenated buffers according to a previously described procedure [14], except that the capillaries were collected on a 40 µm-mesh nylon filter instead of glass beads. The purity of each preparation was controlled by phase contrast microscopy and by measuring the γglutamyl transferase specific activity as a capillary marker [14].

Prostaglandin dehydrogenase activity measurement
Pools of choroid plexuses from at least eight 2-day-old or four adult animals, pools of isolated brain microvessels from twelve 9-day-old animals or four adults, brain cortex, kidney or lung tissue were homogenized in 50 mM Tris, 1 mM EDTA, 2 mM DTT buffer, pH 7.4, using a glassglass homogeniser. The homogenates were centrifuged for 30 min at 14 000 rpm at 4°C, and the resulting supernatant assayed for PGDH activity. This measurement was performed at 37°C by kinetic analysis on a VARIAN Carry 100 double-beam spectrophometer (Mulgrave, VIC, Australia) set at 340 nm as follows: the supernatant was added to the Tris-EDTA-DTT buffer in both reference and sample cuvettes. After baseline stabilisation, NAD + (1 mM) was added to both cuvettes and the baseline further recorded until it stabilized again. PGE 2 (20 µM) was then added to the sample cuvette and the optical density was recorded to follow the appearance of the reduced nucleotide NADH. The specific activity was calculated using the extinction coefficient of 6.22 × 10 -3 M -1 .cm -1 . An aliquot of kidney supernatant was run in each set of measurements as an internal control. The total protein content of the supernatants was determined by the method of Peterson [15] with bovine serum albumin as the standard.

PGE 2 metabolism by isolated choroid plexuses
Choroid plexuses from lateral and fourth ventricles were treated separately. Choroid plexuses from four 2-day-old animals or two adult animals were pooled and incubated on a rotating shaker in 160 µl of RH at 37°C for 5 or 45 min in the presence of 100 nCi of [ 3 H]PGE 2 . Incubation medium without tissue was run in parallel with each set of measurements. At the end of the incubation, the choroidal tissue was removed and added to 40 µl of distilled water, homogenized in the presence of an additional 40 µl of acetonitrile, and centrifuged at 14,000 rpm for 30 min. The resulting supernatant and the incubation medium were then analyzed by HPLC. Choroid plexus protein content was evaluated separately on pools of choroid plexus tissue from the same litter (2-day-old animals), or from the same batch of animals (adults).

HPLC analysis, and expression of the results
Incubation media and homogenate supernatants were analysed by reverse phase HPLC performed on a LC10 Shimadzu system (Duisburg, Germany) as follows: Samples (20 or 40 µl) were loaded with a mix of unlabelled PGE 2 and its metabolites to allow UV detection, applied onto an Ultrasphere ODS RP-18 analytical column (5 µm, 46 mm × 150 mm, Beckman, Fullerton, California, USA), and eluted using a mobile phase of 35% acetonitrile/0.1% acetic acid/water pumped at 1 ml/min. Absorbance of the effluent was monitored at 210 nm. The effluent was collected for radiochemical analysis by liquid scintillation counting. Retention times of PGE 2 , 15-keto-PGE 2 , 13,14dihydro-15-keto-PGE 2 and bicyclo-PGE 2 were 8.5, 12, 16 and 43 min, respectively. The purity of radiolabelled PGE 2 was estimated from the incubation medium without choroidal tissue, as the ratio of radioactivity associated with PGE 2 to the total radioactivity recovered and was taken into account in further calculations. Amounts of remaining PGE 2 and of the metabolites produced during the incubation with the choroidal tissue were expressed as percentage values of the initial PGE 2 -associated radioactivity. The radioactive profile obtained from the incubation medium without tissue was used as a background profile. The background radioactivity eluted within the time-frame of collection for each metabolite was subtracted from the corresponding radioactivity measured in incubation medium following tissue metabolism. The amount of radioactivity (nCi) associated with PGE 2 and each metabolite was calculated separately for the medium and the choroidal tissue, and then summed to generate the total % of PGE 2 remaining, or metabolite produced at the end of the incubation period. To establish the tissuemedium distribution of PGE 2 and 13,14-dihydro-15keto-PGE 2 , the amount of each species present in the incu-bation medium and in the choroidal tissue at the end of the incubation was expressed as % of the total amount (medium plus tissue).

15-hydroxyprostaglandin dehydrogenase specific activity at blood-brain interfaces
PGDH activity was measured in homogenates of choroid plexuses, brain microvessels and cortical tissue ( Table 1). The enzymatic analysis confirmed the absence of a detectable PGE 2 metabolism in cerebral cortex of both adult and 2-day-old animals, and indicated that adult cerebral microvessel preparations also lack PGDH activity. The earliest postnatal period allowing pure capillary isolation is day 9 which marks the end of the sprouting period during which the basal lamina of primitive vessels remains thin and uneven [16]. The activity measured in microvessel preparations from 9-day-old animals was also undetectable (not listed in Table 1). By contrast, this enzymatic activity was readily detected in choroidal material sampled from lateral or fourth ventricle of 2-day-old rat brain. In both types of choroid plexus the activity was lower than in kidney and lung, two peripheral organs involved in prostaglandin signal termination (p < 0.01 and p < 0.05, respectively, one-way ANOVA followed by Tukey-Kramer's test). It strongly decreased in adult choroid plexus (p < 0.05, one-tailed student's t-test for unequal variance).

PGE 2 metabolism in choroid plexuses
To investigate further the ability of the choroidal tissue to inactivate PGE 2 , we analyzed the metabolites produced upon exposure of intact isolated choroid plexus to the prostaglandin. Radiolabelled PGE 2 was used as substrate and the metabolites were quantified by HPLC coupled to radiochemical detection. Figures 1a and 1b show the spectrophotometric profiles obtained using incubation medium supplemented or not, with a mix of unlabelled PGE 2 and its main metabolites. The chromatographic conditions were adequate to separate the two potential PGE 2 metabolites, i.e. 15-keto-PGE 2 and 13,14-dihydro-15- keto-PGE 2 , as well as a non enzymatic cyclization breakdown product of 13,14-dihydro-15-keto-PGE 2 known as bicyclo-PGE 2 [11]. Typical radioactive profiles obtained by incubating [ 3 H]PGE 2 in the absence or presence of isolated intact lateral ventricle choroid plexuses of 2-day-old rats are shown in Figure 1c and 1d, respectively. In addition to the above-mentioned metabolites, an additional peak appeared with a short retention time (Peak 1; 1.75 min). Because the biotransformation of PGE 2 into 15keto-metabolites led to the release, as tritiated water, of one of the seven tritiated hydrogen atom on the prostaglandin, it is likely that tritiated water contributed to the radioactivity in this peak.
The quantitative analysis of the data indicates that after 5 min incubation with lateral ventricle choroid plexuses from four 2-day-old animals, 12% of initial PGE 2 has been metabolized, (representing a metabolic clearance of 18 µl, Figure 2). The most abundant metabolite, 13,14dihydro-15-keto-PGE 2 represents 9% of the radioactivity initially associated with PGE 2 . When the incubation time was extended to 45 min to allow substrate recycling, thus resulting in the biotransformation of 66% of PGE 2 , 13,14dihydro-15-keto-PGE 2 remained the main metabolite. The radioactivity associated with 15-keto-PGE 2 and bicyclo-PGE 2 at 45 min represented only 0.13 and 1% of the total radioactivity initially associated with PGE 2 ( Figure  2). At both 5 and 45 min, 75% of the amount of radioactivity forming Peak 1 may be accounted for by tritiated water formed during the transformation of PGE 2 into keto-metabolites.
Very similar data were obtained using choroid plexuses sampled from the cerebral fourth ventricle of 2-day-old animals (not shown). The ability of intact isolated choroid plexuses from adult animals to inactivate PGE 2 , was also studied and compared to that of 2-day-old animals. The total amounts of PGE 2 metabolized by lateral and fourth ventricle choroid plexuses of adults were respectively 10.5 and 6% of those biotransformed by the corresponding choroidal tissue of 2-day-old rats, following normalization for total protein content. As in 2-dayold material, the main metabolite identified was 13,14dihydro-15-keto-PGE 2 (data not shown).

Location of PGE 2 metabolite secretion by choroid plexuses
The medium and the choroidal tissue were analyzed separately to provide insight into the site of metabolite excretion from choroid plexus cells. The data are shown for lateral ventricle choroid plexuses (Figure 3). PGE 2 remaining at the end of the incubation period of 5 min was mostly found in the incubation medium. By contrast, 13,14-dihydro-15-keto-PGE 2 was mostly associated with the choroidal tissue, i.e. remained intracellular or concentrated within the stroma of the choroid plexus, rather than Radiochemical reverse phase HPLC analysis of PGE 2 metabo-lism by isolated choroid plexuses

Discussion
In this paper, we explored PGE 2 catabolism in rat brain, focusing more specifically on the involvement of the blood-brain interfaces, at both postnatal and adult stages. PGDH is considered as a key oxidizing enzyme in PGE 2 inactivation cascade, as the primary metabolite 15-keto-PGE 2 has a greatly reduced biological activity. 13,14-dihydro-15-keto-PGE 2 is a secondary metabolite without biological activity generated by ∆13-15-keto-prostaglandin reductase [17].
In agreement with other groups, we observed no PGDH activity in brain cortical tissue in either young or adult animals, and we showed the absence of detectable enzyme in microvessels, i.e. at the blood-brain barrier. By contrast, and in accordance with data presented in sheep [12], we gathered evidence for PGE 2 catabolic activity at the choroid plexus, which is the main site of the blood-CSF barrier. First, a significant specific activity of PGDH was measured in choroidal tissue homogenates prepared from 2-day-old rats. Second, the incubation of isolated whole choroid plexuses with PGE 2 , coupled to HPLC analysis, demonstrated the production of PGE 2 metabolites, in particular 13,14-dihydro-15-keto-PGE 2 , thereby revealing the functional coupling of ∆13-15-keto-prostaglandin reductase to PGDH in the choroidal tissue.
The functional significance of this choroidal metabolic pathway may relate either to the termination of CSFborne PGE 2 signal, or to the prevention of blood-borne PGE 2 penetration into the CSF. In our experimental setting for ex-vivo choroid plexus incubation, the isolated tissues were kept entire, which maximally limits rapid transfer between the external medium and the choroidal stromal core, and allows us to assume that most PGE 2 was presented apically, i.e. at the CSF-side of the choroidal epithelium. Our results therefore suggest that in vivo, centrally released PGE 2 that circulates in CSF will be metabolized to some extent by choroidal metabolizing enzymes. In line with this, an apical uptake of PGE 2 mediated by an inwardly-directed probenecid-sensitive transport system has been reported in choroidal epithelium of different species [11,18,19]. Both metabolic and transport affinity PGE 2 metabolism by isolated choroid plexuses from lateral ventricles of 2-day-old rat brain Figure 2 PGE 2 metabolism by isolated choroid plexuses from lateral ventricles of 2-day-old rat brain. Data are expressed as mean ± SD of three different experiments, and represent the amount of radioactivity associated to each molecular species, relative to the initial radioactivity associated to PGE 2 (see method), after either 5 or 45 minutes of incubation. * and **: different from PGE 2 amount in incubation without choroidal tissue, p < 0.05 and 0.01, respectively, paired student's t-test. Abbreviations: KPGE 2 : 15-keto-PGE 2 , DHKPGE 2 : 13,14-dihydro-15-keto-PGE 2 , BCPGE 2 : bicyclo-PGE 2 .  Tissue Medium constants have been determined in the micromolar range [8,20]. Given that PGE 2 levels in the CSF remain below this concentration in physiological as well as pathological conditions [21][22][23], neither the enzymatic nor the transport process will reach saturation. The extent to which choroidal PGE 2 metabolism can impact on the concentration of CSF-borne PGE 2 remains however to be evaluated. Although PGDH activity is readily detected in choroidal tissue from young pups, it is 10 to 20 times lower than in kidney, or lung which is the main organ involved in peripheral signal termination of circulating prostaglandins [1]. In ex-vivo tissues from 2-day-old animals, after 5min incubations, a significant amount of untransformed PGE 2 , similar to the total amount of metabolites produced, was found associated with the choroidal tissue.
Although the precise localization of PGDH (epithelial and/or stromal) within the CP is unclear (see infra), the latter observation suggests that the metabolism capacity of the tissue towards PGE 2 is lower than its uptake capacity. Transepithelial flux of PGE 2 has been demonstrated in an in vitro model of the choroidal epithelium, implying that following its apical uptake from the CSF, native PGE 2 can be exported across the basolateral stroma-facing membrane [19]. The relative capacity of this basolateral efflux mechanism, favoring PGE 2 elimination in blood and thus supplementing the enzymatic signal termination mechanism needs to be established by comparison to uptake and metabolism in order to delineate how CSFborne PGE 2 concentration is controlled in developing animals. PGE 2 catabolism in the choroidal tissue may also be relevant in preventing blood-borne PGE 2 from entering the CSF during early postnatal life. During this period, PGE 2 is associated with hypothalamic maturation processes [24] and an increase in CSF PGE 2 induces respiratory depression [reviewed in [11,12]]. Therefore abnormal blood PGE 2 concentrations following infection or inflammation, were they to disrupt physiological PGE 2 levels in CSF, could possibly lead to cerebral dysfunction. Based on the directionality and membrane distribution of the epithelial organic anion transporters that are likely candidates for membrane transfer of the prostaglandin [20], blood-to-CSF permeability to PGE 2 is expected to be much lower than CSF-to-blood permeability. In 2-day-old rats, the metabolic activity of the choroid plexus tissue towards the prostaglandin will add an enzymatic barrier component to the transporter-mediated barrier properties of the epithelium, thereby contributing to buffer blood perturbations and maintain PGE 2 homeostasis in CSF during this critical period of life. Of note, ontogenic maturation of the choroid plexuses is precocious and this tissue appears to play key functions in the control of brain homeostasis when the cerebral vasculature is still developing. [25,26]. The metabolic capacity displayed by the choroid plexuses towards PGE 2 , during the postnatal period highlights the early functional maturity of the interface.
In the adult, PGDH enzymatic activity is strongly decreased in choroidal tissue, a finding confirmed by the age-dependent decreased metabolic capacity observed in isolated choroid plexus. This leaves peripheral organs such as lung as the most likely sites of catabolism for the prostaglandin following its clearance from adult brain [1].
In ex vivo studies using short duration incubation, PGE 2 metabolites were mostly associated with the choroidal tissue, indicating that they were produced by the epithelial cells and then preferentially released in the stroma, and/ or produced by the fibroblasts or other stromal cells.
When the incubation was prolonged, 13,14-dihydro-15keto-PGE 2 reached the external medium, probably as a result of diffusion from stroma. We previously showed, using a polarized cellular model of the blood-CSF barrier, that 13,14-dihydro-15-keto-PGE 2 was produced and excreted at the basolateral membrane of the epithelial cells [19]. The latter step may involve the multidrug resistance associated protein abcc4, which transports organic anions such as prostaglandins [27], although its affinity for keto-metabolites remains to be established. This transporter has been immunodetected at the basolateral membrane of choroidal epithelium in several species [28]. PGE 2 metabolism was however limited in the epithelial cells and did not significantly impede the transcellular flux of the prostaglandin [19]. Alternatively, the stromal hypothesis of PGE 2 metabolism is supported by the immunohistochemical description in sheep of a switch in PGDH localisation from the epithelium to the stromal cells at birth [12]. In rat, attempts to locate PGDH in choroidal tissue by immunohistochemistry in our laboratory have been so far inconclusive (not shown). Regardless of the cellular site of PGE 2 metabolism, the stromal i.e. blood side of metabolite excretion adds to the efficiency of the metabolic barrier by driving the clearance of the metabolite towards the blood circulation.

Conclusion
Metabolism of PGE 2 occurs in the rat choroid plexus tissue at early postnatal stage, leading to the production of the inactive 13,14-dihydro-15-keto-PGE 2 metabolite. This function of the choroidal tissue, which disappears in adulthood, may be involved in maintaining CSF PGE 2 homeostasis during the critical early phase of postnatal brain development.