Biological and Pharmaceutical Bulletin
Online ISSN : 1347-5215
Print ISSN : 0918-6158
ISSN-L : 0918-6158
Current Topics: Reviews
Membrane Transporters Contributing to PGE2 Distribution in Central Nervous System
Yoshinobu NakamuraTakeo NakanishiIkumi Tamai
ジャーナル フリー HTML

2018 年 41 巻 9 号 p. 1337-1347


Prostaglandin (PG) E2 is a well-established lipid mediator that plays a role in diverse functions and diseases of the brain. Cyclooxygenase and PGE synthase have been extensively studied as molecular determinants of extracellular concentration of PGE2 near prostanoid E receptors since the brain has limited capacity of PG metabolism. There is accumulating evidence that several members of the solute carrier (SLC) and ATP-binding cassette (ABC) superfamilies regulate PGE2 distribution in brain capillary endothelial cells, choroid plexus (CP) and arachnoid epithelium, and different parenchyma cells such as neuronal and glial cells. These transporters may mediate entry and exit of PGE2 at blood–brain and blood–cerebrospinal fluid boundaries, resulting in brain distribution of PGE2. However, their roles in neuroinflammation and disease progression remain unclear. In this review, current knowledge on transporters involved in brain distribution of PGE2 is summarized, and especially, potentials of organic anion transporting polypeptide (OATP) and organic anion transporter (OAT) family members are discussed as molecular determinants of PGE2 concentration in the brain.


Eicosanoids are signaling molecules produced from arachidonic acid or other polyunsaturated fatty acids, and released out of cells. They exert diverse physiological and pathophysiological actions through eicosanoid receptor expressed at the plasma1) and nuclear membranes.2,3) Among eicosanoids, prostaglandin (PG) E2 is the most compelling lipid mediator that signals through four distinct G-protein coupled receptors (EP1–4) in autocrine and paracrine manner.4) In the central nervous system (CNS), PGE2 modulates signaling and gene expression to induce various pathophysiological reactions such as brain inflammation,5) fever generation,6) vasodilation7) and vasoconstriction,8) and neurotransmitter release.9,10) Inversely, impaired PGE2 signaling has been associated with brain diseases such as Alzheimer’s disease,11) Parkinson’s disease,12) multiple sclerosis,13) and depression.14) Especially, current evidence suggests that PGE2 concentrations in cerebrospinal fluid (CSF) are correlated with memory impairment.15) Consistently, clinical studies demonstrate therapeutic effect of celecoxib, a cyclooxygenase (COX)-2 selective inhibitor, in major depression.16) Intuitively, PGE2 signaling is regulated by its concentration near the receptors. COX and PGE synthase (PGES) have been extensively studied as important molecular determinants of PGE2 concentration; however, much remains unknown regarding the biological determinants of these processes. To date, there is accumulating evidence that membrane transporters are essential for PGE2 membrane permeability and some have been functionally characterized in brain tissues including the blood–brain barrier (BBB). This review summarizes current knowledge on transporters that recognize PGs, which are involved in membrane permeation of especially PGE2 at the BBB, the blood–CSF barrier (BCSFB), and brain parenchyma cells and describes some pathophysiological roles of such transporters in CNS.


PGs are present in an anionic form under physiological pH, and have been described as poorly permeable to plasma membranes.17,18) They might cross plasma membranes by simple diffusion.19) Indeed, a number of carrier proteins have been indicated to mediate PG transport in humans and rodents as listed in Table 1. Most of them belong to the organic anion transporting polypeptide (OATP/SLCO) and polyspecific organic cation/anion transporter (SLC22A) subfamilies of SLC superfamily. In addition, two members of the ATP-binding cassette subfamily C (ABCC) subfamily have been shown to transport PGs, demonstrating a major role of MRP4/ABCC4 in cellular efflux of PGs, especially PGE2.20) These transporters may serve as regulators for extracellular and intracellular concentration of PGE2; therefore it is important to clarify their expressions and functions in the brain so as to understand complex physiological actions of PGE2.

Table 1. Transporters That Recognize PGE2 as Substrate
SpeciesGene symbolProtein nameAliasesAffinity to PGE2 (nM)Protein expression*
SLC TransporterHumanSLCO1A2OATP1A2OATP-A, SLC21A3N.D.40)Br,39,46) Eye,46,99) Ki and Li39)
SLCO1B1OATP1B1OATP-C, LST-1, OATP2, SLC21A6N.D.40,44,49)Li100,101)
SLCO1B3OATP1B3OATP8, LST-2, SLC21A8N.D.49)Li100103)
SLCO1C1OATP1C1OATP-F, OATP14, SLC21A14N.D.49)Br,104,105) Eye,99) Te106)
SLCO2A1OATP2A1PGT, MATR1, SLC21A2, PHOAR2100b) 55), 331a) 56)Br,30) Eye,107) Ov,108) St109)
SLCO2B1OATP2B1OATP-B, SLC21A9N.D.44,49)Br,46) Eye,46,99,107) Li,40) Si110)
SLCO3A1OATP3A1OATP-D, OATP3A1_v1, SLC21A1156a) 75), 218a) 81)Eye,99) Br and Te81)
SLCO3A1OATP3A1_v2OATP-D, SLC21A11372a) 81)Br and Te81)
SLCO4A1OATP4A1OATP-E, SLC21A12N.D.44,48,49)Eye99)
SLC22A1OCT1HOCT1657a) 57)Br,64) Li,111) Lu112)
SLC22A2OCT2-28.9a) 57)Br,64,113) Ki,103,111) Lu,112) Si103)
SLC22A6OAT1HOAT1, PAHT, ROAT1970a) 57)Br,114) Ki,115,116) Sm117)
SLC22A7OAT2NLT, hOAT11713a) 57)Ki,118) Li101)
SLC22A8OAT3-345a) 57)Br,114) Ki,116,119) Sm117)
SLC22A11OAT4hOAT4154a) 57)Ki,116,120) Pl121)
SLC51AOSTαOSTAN.D.122)Ki, Li and Si123)
SLC51BOSTβOSTBKi, Li and Si123)
RatSlco1a1Oatp1a1Oatp1, Slc21a1, Slc21a3N.D.49)Li124,125)
Slco1a2Oatp1a4Oatp2, Slco1a4, Slc21a5N.D.49)Br,47) Eye,99,126) Ki,127) Li125,126)
Slco1a5Oatp1a5Oatp3, Slc21a75680a) 49), 3500a) 41)Br,128) Eye,99,126) Li,126) Si129)
Slco1b2Oatp1b2Oatp4, Slc21a109570a) 49), 1300a) 41)Eye,99) Li41,130)
Slco2a1Oatp2a1Matr1, Slc21a294a) 53)Br,29,75,76) Dp,131) Lu,132) Ki, Ov, Ht, and Te75)
Slco2b1Oatp2b1Moat1, Slc21a9N.D.49)Br,47) Si130)
Slco3a1Oatp3a1OATP-D, Slc21a11N.D.75)Br, Ht, Ki, Lu, Ov, Te, and Ut75)
Slc22a7Oat2-N.D.133)Ki,134) Li101)
Slc22a8Oat3Roct4240a) 58)Br,47,51,79) Ki79)
Slc22a22OAT-PG-143a) 135)Ki135)
MouseSlco1a1Oatp1a1Oatp1, Slc21a1N.D.136)Ki137)
Slco2a1Oatp2a1Pgt, Slc21a2110c) 54)Br,59,138) Ki,139) Lu,132,139) Sk34)
Slc22a8Oat3Roct1480a) 140)Br and Ki141)
Slc22a22OAT-PG-118a) 142)Ki142)
Slc51aOstαOstaN.D.122)Ki, Li and Si123)
Slc51bOstβOstbKi, Li and Si123)
ABC TransporterHumanABCC2MRP2CMOAT, cMRP, ABC30, DJSN.D.143)Br,144) Li103)
ABCC4MRP4MOAT-B3400a) 20)Br,20,66) Ht,145) Ki,146) Li147)
MouseAbcc4Mrp4MOATBN.D.72)Br,20,66) Ki,148) Pl149)

a); Km value, b); K1/2 value, c); Ki value *Br; brain, Dp; dental pulp, Ht; heart, Ki; kidney, Li; liver, Lu; lung, Ov; ovary, Pl; placenta, Si; small intestine, Sk; skin, Sm; skeletal muscle, St; stomach, Te; testis, Ut; uterus. N.D. represents not determined.


The role of circulating PGE2 in the brain has been studied in relation to the febrile response. A febrile response is a major host defense in response to infection. Since mice lacking Cox-2/Ptgs2,21) mPges-1/Ptges,22) or EP3 receptor (Ptger3)23) fail to develop fever in response to pyrogenic immunomodulators, PGE2 has been accepted as a neuronal transmitter essential for the febrile response. Because blood PGE2 concentration simultaneously rises with body temperature, PGE2 produced by macrophages of peripheral tissues (mostly lung and liver) is considered a trigger of fever generation other than PGE2 produced in the preoptic/anterior hypothalamic area (POA/AH).24) This notion is supported by several lines of experimental evidence. For instance, fever is blocked by neutralization of circulating PGE2 with anti-PGE2 antibody and inhibition of PG synthesis by nonsteroidal anti-inflammatory drugs that unlikely cross the BBB.24,25) Moreover, intravenous or intracarotid PGE2 or PGE1 causes fever without affecting vagal afferent signaling.26) Accordingly, current evidence at least suggests that circulating PGE2 produced outside the brain penetrates the BBB and induces fever in POA. The BBB is a diffusion barrier and consists of tight junctions present between cerebral endothelial cells and selectively excludes most compounds from entering the brain. Considering the poor membrane permeability of PGE2, membrane transporters are required for this signaling molecule to cross the plasma membranes. Some transporters that recognize PGE2 are expressed in brain capillary endothelial cells; however, there is limited information on PGE2 transport across the BBB. Current evidence about expression of transporters is illustrated in Fig. 1.

Fig. 1. Current Knowledge on Transporters for PGE2 at the BBB

Current knowledge on transporters for PGE2 at the BBB is described according to information on transporters’ recognition to PGE2 and localization of transporters by histological examinations. Blue and red circles indicate SLC and ABC transporters, respectively. In rodents, the characters “m” or “r” written at the beginning of transporters name represent “mouse” and “rat,” respectively. Arrows indicate the directions of PGE2 transport across the luminal or abluminal membranes of brain vascular endothelial cells. (Color figure can be accessed in the online version.)

Taogoshi et al. reported that PGE1 transport across the BBB is inhibited in the presence of bromocresol green, digoxin, or taurocholate by in situ rat brain perfusion technique.27) Bromocresol green is a well-known inhibitor of OATP2A1. RT-PCR confirmed mRNA expression of rat Oatp2a1 in brain capillary-rich fraction as well as whole brain.27) Our laboratory reported Oatp2a1 immunoreactivity in mouse brain capillaries,28) and Oatp2a1 was immunolocalized predominantly at the luminal membranes of rat primary cultured cerebral endothelial cells.29) OATP2A1 protein is also expressed in blood vessels at the middle frontal gyrus of human brains.30) Furthermore, attenuation of lipopolysaccharide (LPS)-induced febrile response was associated with lower PGE2 concentration in brain interstitial fluid in Slco2a1 global knockout mice, compared with wildtype counterparts.28) This evidence implies that OATP2A1 plays a role in PGE2 supply from central circulation across the BBB. On the other hand, recent protein expression profiling assay experiments suggest negligible expression of OATP2A1 in human and mouse brain capillary-enriched fractions.31,32) Therefore, the role of OATP2A1 remains controversial. Since OATP2A1 is also important in PGE2 signaling in endothelial cells,3335) its pathophysiological significance in brain capillary should be clarified in future.

Taogoshi et al.27) reported that both digoxin and taurocholate are substrates of rat Oatp1a4/Slco1a2 (a.k.a. Oatp2), implying a contribution of rOatp1a4 to PGE2 entry to the brain from blood across the BBB. It has been indicated that members of the OATP1A/SLCO1A subfamily mediate transport of solutes and opioid peptides across the BBB.36) The protein encoded by rat Slco1a2 has been designated as Oatp1a4, and is the rat orthologue of mouse Oatp1a4/Slco1a4. Rat and mouse Oatp1a4 are expressed in luminal and abluminal membranes of brain capillary endothelial cells.37,38) Abundant protein expression of OATP1A2, which is only one OATP1A family member in humans, was detected in brain microvessels and capillaries but not in astrocytes and neurons.36,39) In addition, PGE2 is an established substrate of hOATP1A2,40) although there is a report showing that PGE2 is not transported by rOatp1a4.41) There are two more additional OATP1A members in rodents: Oatp1a5/Slco1a5 and Oatp1a6/Slco1a6 (a.k.a. Oatp5). Both Oatp1a5 and 1a6 were detected in brain capillary-rich fraction from mice, and immunohistochemistry confirmed that Oatp1a5 is expressed in brain capillaries and cerebral cortex.42) Since Oatp1a5 accepts PGE2 as a substrate,41) this carrier is an alternative candidate for PGE2 transport at the BBB. Localization of Oatp1a6 is not confirmed yet, and little is known on substrate specificity of Oatp1a6 except that it transports taurocholate. This implies a role of OATP1A subfamily in PGE2 transport at the BBB, although there are known cases that no straightforward orthologue is shared between humans and rodents. Because of discrepancy between species, their roles in PGE2 transport at the BBB remain elusive.

OATP2B1 protein expression was confirmed in human brain capillary cells by means of ultra-performance LC-multiple reaction monitoring mass spectrometry.43) PGE2 was demonstrated a transportable substrate of OATP2B1.44) OATP2B1-immunoreactivity was localized to the luminal surface of capillaries in human gliomas,45) and its expression at the plasma membranes of the endothelial cells in the cortex and hippocampus in humans implies a role in cellular uptake of peptide neurotransmitters.46) It is worthwhile to note that the expression of rOatp2b1 was indicated at the abluminal membranes of microvessels and pericytes associated with microvessels.47) Therefore, OATP2B1/SLCO2B1 is another candidate carrier that contributes to PGE2 transport at the BBB. However, to date, there is no direct evidence that OATP2B1 is involved in PGE2 transport across the BBB.

Our laboratory has shown that another member of OATP family OATP4A1/SLCO4A1 recognizes PGE2 as a substrate,44) although Fujiwara et al.48) characterized this transporter for thyroid hormones rather than PGs. OATP4A1 is abundantly expressed in placenta, lung, and liver. Little OATP4A1 is expressed in the brain, from where it was isolated,48) and its localization in the brain remains undetermined. Besides, OATP1C1/SLCO1C1 is a well-characterized OATP family member at the BBB. This gene was originally isolated as the BBB-specific anion transporter type 1 from mice, and is now classified as Oatp1c1/Slco1c1 (a.k.a. Oatp14). Oatp1c1 is specifically expressed along the plasma membranes of brain capillary endothelial cells, and has been shown to contribute to brain uptake of triiodothyronine and thyroxine. Since hOATP1C1 accepts PGE2 as a substrate,49) and OATP1C1-mediated transport is inhibited by PGE2,50) this suggests that OATP1C1 functions as a PGE2 importer at the BBB.

OAT3/SLC22A8 is expressed at the BBB and may transfer a solute in the direction from brain to blood. Oat3/Slc22a8 immunoreactivity was detected specifically at the abluminal membranes of brain capillary endothelium of rats,47,51,52) suggesting its role in export of various organic anions including homovallinic acid,52) p-aminohippuric acid, and benzylpenicillin51) from brain to blood across the BBB. Akanuma et al.32) found that decrease in Oat3 and Oatp1a4 protein expression was associated with reduced elimination rate of PGE2 from the brain of LPS-treated mice, implying that these transporters are involved in PGE2 elimination from the brain. PGs are well-established substrates of OAT3, and their affinities to OAT3 are lower than those to OATP2A1.5356) Km values of PGE2 to OAT3 range from 345 to 4240 nM, whereas those to OATP2A1 are approximately 100 to 300 nM5356) (Table 1); therefore, OAT3 is regarded as a low-affinity/high-capacity transporter for PGs.5759) Indeed, efflux rate of externally pre-injected PGE2 into the rat brain was significantly reduced in the presence of an OAT3 substrate benzylpenicillin.60) This suggests that OAT3 is likely a candidate carrier for PGE2 efflux at the BBB. In addition, Oatp1a4 is considered a part of efflux system for wastes and unwanted materials from the brain to central circulation across the BBB because it is expressed at both the luminal and abluminal membranes of brain capillary endothelial cells and mediates bidirectional transport of solutes. In contrast to OAT3, the remaining amount of externally pre-injected PGE2 in the mouse brain (the S2 region) neither affected intracerebral pre-administration of digoxin, a substrate of Oatp1a4, nor intravenous (i.v.) injection of its inhibitor, amiodarone; therefore Oatp1a4 is less likely involved in PGE2 elimination across the BBB. Thus cerebral concentration of PGE2 can be affected by transporter in addition to PGE2-synthesizing enzymes, such as COX-2 and microsomal PGES.

PGE2 is recognized by not only organic anion transporters but also organic cation transporters. To date, two polyspecific organic cation transporters, human OCT1/SLC22A1 and OCT2/SLC22A2, were described to transport PGE2,57) although capability of OCT2 to transport PGs is in dispute.61) Previous literature showed lack of expression of Oct1/Slc22a1 in the brain in rats,62) whereas mRNA expression of rOct2/Slc22a2 was noted at detectable levels in various regions of the brain.63) More recent analysis of OCT1 and OCT2 showed their expression at both the luminal and abluminal sides of isolated brain microvessel endothelial cells from humans, rats, and mice. A typical substrate for OCTs, N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine uptake by rat brain capillary endothelial cells was competitively inhibited by amantadine, a pan-inhibitor of OCTs, suggesting a role of OCT1 and OCT2 in PGE2 transport across the BBB.64) Although the contribution of OCTs to PGE2 transport through the BBB is not known, these cationic transporters cannot be excluded.

Not only solute carriers but also ABC transporters are involved in PGE2 disposition at the BBB. An ABC transporter MRP4/ABCC4 has been characterized as a major transporter for PGE2 export.20) Besides, in vitro experiment with Mrp4-expressing vesicles suggests that mouse Mrp4 also transports PGs.65) MRP4 expression has been confirmed in brain capillaries in mice,66) cows,67) and humans.68,69) MRP4 immunoreactivity was clearly shown at the luminal wall of endothelial cells of brain capillaries in mice66) and humans,69) whereas it was found additionally at the abluminal membranes in cows.67) Previous literature consistently demonstrates that MRP4 limits brain distribution of its substrate drugs including topotecan and oseltamivir.66,70,71) In this context, the role of this carrier in PGE2 elimination from the brain has been studied by means of brain efflux index method. Intracerebral injection of benzylpenicillin and Ro64-0802, which are substrates of OAT3 and MRP4, reduced elimination rate of PGE2 from rat brain.60) Similarly, Akanuma et al. found that i.v. injection of MRP4 substrate antibiotics, cefmetazol and cefazolin, reduced elimination rate of pre-injected PGE2 to the brain.72) However, other MRP4 substrate drugs such as cefotaxime and ketoprofen did not affect the elimination rate. Because cefotaxime and ketoprofen are poorly permeable to the BBB, their intracellular concentration might not be sufficient to inhibit MRP4 function in brain endothelial cells.72) In humans, misoprostol, a PGE1 analogue, which can cause high fevers, is a substrate of MRP4. A non-synonymous single nucleotide polymorphism of MRP4 (rs11568658, G187W), which shows reduced transport activity, has been reported associated with misoprostol-induced fever.73) Accordingly, current experimental evidence suggests that MRP4 facilitates elimination of produced PGE2 from the brain, which is most likely taken up by OAT3 expressed at the abluminal membranes of endothelial cells, and limits entry of PGE2 into the brain from blood.


Choroid plexus (CP) is located in all components of ventricular system, where CSF is produced. CP epithelium is composed of epithelial cells that continuously connect to ependymal cell layers, and surrounds capillaries without connective tissues therebetween. Tight junctions are well developed in the epithelial layer on the side facing the ventricle (apical side) and limit free paracellular diffusion of substances from blood to the ventricle.74) Unnecessary solutes and waste materials in CSF are also absorbed at the CP and eliminated to blood. Accordingly, the CP epithelium serves as a barrier that separates peripheral blood from CSF. CSF enters the cerebral blood at the arachnoid granulation from subarachnoid space, and tightly adherent epithelial cells of arachnoid mater function as a barrier at the cerebral blood–CSF boundary. These barrier structures are so-called blood–CSF-barrier (BCSFB), where membrane transporters play a role in selective permeability of membranes, namely, function of transporters at the BCSFB contributes to maintain homeostasis in the brain. In the BCSFB, some SLC and ABC transporters are implied responsible for PGE2 transport across the BCSFB; however, their roles are not fully understood. Localization of transporters for PGE2 at the CP epithelium is illustrated in Fig. 2.

Fig. 2. Current Knowledge on Transporters for PGE2 at the BCSFB

Current knowledge on transporters for PGE2 at the BCSFB is described according to information on transporters’ recognition to PGE2 and localization of transporters by histological examinations. Blue and red circles indicate SLC and ABC transporters, respectively. In rodents, the characters “m” or “r” written at the beginning of transporters name represent “mouse” and “rat,” respectively. Arrows indicate the directions of PGE2 transport across the apical or basolateral membranes of the epithelial cells of the CP. (Color figure can be accessed in the online version.)

To understand a mechanism of PGE2 elimination from CSF, PGE2 uptake by isolated rat CP was characterized by Tachikawa et al.58) They described that PGE2 uptake consists of carrier-mediated and non-saturable components with an apparent Km value of 23.0 µM. The uptake was inhibited by various organic anions, especially bromocresol green (an inhibitor for OATP2A1), and to a lesser extent benzylpenicillin and p-aminohippuric acid, whereas digoxin did not affect it. This observation agrees with the facts that PGD2 was actively uptaken by rat CP,59) implying a role of organic anion carriers in PGE2 clearance from CSF at the CP.

Oatp2a1 is expressed at the CP in rats,29,75) and more recently its immunoreactivity was confirmed at the apical membranes of epithelial cells at the CP in mice and rats.59,76) Currently, there is no information on OATP2A1 expression in the CP in humans. Moreover, sophisticated in situ hybridization with immunohistochemical approach revealed Oatp2a1 expression at the apical side (facing the subarachnoidal space) of arachnoid membranes in rats.76) Our laboratory recently showed that PGE2 concentration in CSF in LPS—intraperitoneally (i.p.) injected Slco2a1−/− mice was significantly greater than that in wildtype counterparts.28) Elimination clearance of radiolabeled PGD2, which is another substrate of OATP2A1 with similar affinity to PGE2,5355) was approximately 15 times greater than inulin, a reference compound for CSF turnover and diffusion into the brain interstitial space, and significantly inhibited by co-injection of unlabeled PGD2.59) In addition, Hosotani et al.76) suggested that elevated expression of Oatp2a1 in blood vessels of subarachnoid space facilitates elimination of PGE2 from CSF. Thus OATP2A1 expressed at the CP is considered one of the molecular determinants of PGE2 in CSF.

Other OATP family members may be involved in PGE2 transport at the BCSFB. Rat Oatp1a5 was demonstrated expressed at the apical side of the CP epithelium,77) implying its contribution to CP uptake of PGE2. Oatp1a1 immunofluorescence was detected at the apical membranes of CP epithelial cells.37,78) However, Oatp1a1 may not be involved in PG absorption from the CP epithelium because Oatp1a1 accepts PGE2 only under acidic conditions.41,49) In contrast, Oatp1a4 and Oatp2b1 (a.k.a. Moat1) are likely expressed at the basolateral side (facing the capillaries) of the CP epithelium.37,47,78) Their roles in PGE2 transport across the BCSFB remained undetermined. Compared with these Oatp family members, there is more certain information for OAT3 at the CP. OAT3 expression has been confirmed at the apical membranes of rat CP epithelial cells by immunohistochemistry.79) Na+-dependent apical uptake of p-aminohippuric acid by the CP of Oat3-null mice was almost completely abolished.80) Furthermore, inhibitory effect of OAT3 substrates on elimination rate of PGE2 and PGD2 in CSF suggests a contribution of OAT3 to PGE2 elimination from CSF as well as OATP2A1.

Information is limited on transporter gene expression at the CP in humans. Several research groups including ours characterized OATP3A1/SLCO3A1 (a.k.a. OATP-D) as an importer for PG with a relatively broad tissue distribution in humans and rats.44,75) Huber et al.81) identified an alternative splice variant of SLCO3A1 and showed that the encoded protein has a similar PG transport activity (with a Km for PGE2 of 218 nM) to original OATP3A1 (Km for PGE2 of 101 nM). This splice variant is named OATP3A1_v2 to distinguish from the original OATP3A1 (which is renamed as OATP3A1_v1). mRNA of both variants is expressed ubiquitously; however, their cellular and subcellular localization differ in the testis, CP, and brain frontal cortex in humans. In the CP epithelium, OATP3A1_v1 expression was localized to the basolateral membranes of the epithelial cells, whereas OATP3A1_v2 expression was more polarized to apical membranes and/or subapical intracellular vesicular compartment.81) Each variant functions as an importer for PGE2 from CSF and blood; therefore how these transporters affect PGE2 distribution in CSF should be clarified.

Experiments with the anti-MRP4 monoclonal antibody, M4I10, which cross-reacts with rodent orthologues, clearly suggest that MRP4 is expressed at the basolateral membranes of CP epithelial cells in humans, mice, and rats.66,69) Considering a contribution of MRP4 in barrier function in blood–CSF boundary, MRP4 may serve as an efflux carrier for PGE2 from CSF to blood, which is consistent with the observation that the elimination rate of pre-injected radio-labeled PGE2 from CSF was slowed by simultaneous injection of MRP4 substrate antibiotics, ceftriaxone and cefazolin.58)

Arachnoid mater epithelial cells are part of the BCSFB; however, molecular entities in the barrier function remain unknown. More recently, Zhang et al.82) applied quantitative targeted absolute proteomics for plasma membrane fraction of isolated rat leptomeninges, and found that 15 solute carriers (SLC transporters) and 10 ABC transporters are expressed at significant level. In their study, high expression of two characteristic organic anion transporters, Oat1/Slc22a6 and Oat3/Slc22a8, in the arachnoid mater epithelial membranes was associated with elimination of p-aminohippuric acid from CSF in rats. Since Oat1 and Oat3 accept PGE2, PGE2 concentration in CSF could be regulated by these transporters at the cerebral blood–CSF boundary. Other SLC transporters, which transport PGs, detected in arachnoid mater epithelial cells include Oatp1a4, Oatp1a5, and Oatp1c1 in the descending order of expression. In addition, Mrp4 was also detected. However, cellular and subcellular localization of these transporters in arachnoid mater is unknown.


In the brain, multiple stressors may occur at the same time and cause PGE2 release/secretion from different cell types to adapt to various stresses. This is so-called sickness responses, which are evoked by peripheral inflammation, including fever, pain, and neuroendocrine response as well as behavioral responses.83) Peripheral inflammation caused by bacterial infection or immunomodulators first activates peripheral immune cells to release blood–borne pro-inflammatory cytokines (e.g., tumor necrosis factor (TNF)-α, interleukin (IL)-1β), which in turn induce COX-2 expression in vascular endothelial cells and/or perivascular cells.84) On the other hand, psychological stress is generated by a feeling of strain and pressure, and involves neuronal cells to produce PGE2 via COX-2 pathway.85) Since COX-1 is constitutively expressed in microglial cells, they are considered an important source of PGE2 in response to psychological stress.14,86,87) Thus PGE2 released/secreted from various brain parenchyma cells may be one of the critical stress adaptation factors. In this context, transporters responsible for cellular PG disposition may play a role in stress response because they are involved in adjusting local PGE2 concentration (e.g. in the vicinity of EP receptors). Nonetheless, physiological relevance of transporters in brain parenchyma cells is poorly understood. To date, several research groups have suggested expression of some transporters in glial and neuronal cells. Of those, MRP4 as an efflux system has been described in glial cells. A substantial level of Mrp4 mRNA expression was found in astroglia- and microglia-rich cultures, whereas lower expression was reported in neuron- and oligodendroglia-rich cell culture from rat brain.88) Expression of Mrp4 mRNA was found approximately 55 times greater in mouse BV-2 microglial cells than liver tissue.89) Plasma membrane expression of MRP4 was confirmed by immunocytochemistry in BV-2 cells,89) and its function was demonstrated by means of an MRP4 substrate antiviral drug, 9-(2-phosphonylmethoxyethyl) adenine, in rat MLS-9 microglia cell line.90) It is also worth to note that decreased Mrp4 protein expression was associated with increased accumulation of calcein AM, a prototypic substrate for MRPs, in LPS-treated BV-2 cells. Indeed, MRP4 mediates cellular efflux of PGE220) and cyclic nucleosides (e.g. cAMP and cGMP),91) which serve as second messenger in PGE2 signaling via EP; therefore alteration in activated microglia may influence signaling with other neuronal cells.

Choi et al.30) first reported that OATP2A1 immunoreactivity was evident in neurons, microglia, and astrocytes in humans. Although total protein expression of OATP2A1 was decreased in the brains of the Alzheimer’s disease patients, OATP2A1 immunostaining became more intense in cytoplasmic compartments of their microglia, compared with those in control human subjects.30) Microglia and astrocytes activated by neuroinflammation secrete pro-inflammatory mediators including cytokines, PGs, and reactive oxygen species, and contribute to neurodegenerative changes, which exacerbate the disease.92,93) Current evidence suggests that OATP2A1 works for uptake and secretion of PGs, depending on the intracellular expression.9496) We recently reported that OATP2A1 was primarily expressed in cytoplasm of murine macrophages and mediates PGE2 uptake by intracellular vesicles containing light lysosomes, showing that extracellular PGE2 was significantly lowered in Slco2a1−/− mice-derived peritoneal macrophages compared with Slco2a1+/+ cells.97) Furthermore, macrophage/monocyte-specific Slco2a1-knockout (Slco2a1Fl/Fl/LysMCre/+) mice exhibited significantly lower PGE2 concentrations in brain interstitial fluid and incomplete febrile response, compared with their control counterparts (Slco2a1Fl/Fl mice), when they were treated with LPS.28) Accordingly, OATP2A1 may be a potentially important player in glial–neuronal interaction, by regulating PGE2 concentration in brain interstitial fluid. Future study is warranted to clarify the pathophysiological significance of OATP2A1 in brain disorders including Alzheimer’s disease.

In brain neuronal cells, some reports describe cell type-specific expression and distinct distribution of OATP family members. OATP1A2 was recently found abundantly expressed in neurons in the frontal cortex and hippocampus, whereas immunolocalization of OATP2B1 was more likely limited to brain capillary endothelial cells in humans.46) In addition to differential expression of OATP3A1 variants at the CP, OATP3A1_v1 staining was observed in neuroglial cells, but not in neurons, of the gray matter, while OATP3A1_v2 expression was found in neurons of the gray and white matter of human frontal cortex.81) Besides, current evidence suggests that OCT2 is localized in cholinergic neurons in the mouse brain and contributes to clearance of neurotransmitters such as norepinephrine and serotonin.98) These transporters are involved in mediating transport of diverse endogenous substrates including PGs, neuropeptides, and steroids across the plasma membranes of neurons and capillary endothelial cells, implying important roles in cell–cell interactions. PG can be one of the important mediators for these interactions. As evidence accumulates on the expression of PG carriers in the brain, understanding their pathophysiological significance may provide clues to designing novel treatments for refractory brain diseases.


There are several ways in which transporters contribute to brain disposition of PGE2; they serve: 1) to facilitate entry of PGE2 into the brain across the BBB especially in early phase of fever generation; 2) reduce efficiently PGE2 from CSF at the blood–CSF boundary; and 3) regulate interstitial PGE2 concentration in response to sickness and psychological stressors. In each case, current information is so far not conclusive. For PGE2 entry to brain, OATP1A subfamily members are likely candidates; however, their lack of orthologues makes it difficult to determine the major contributor to permeation of PGE2 across the BBB. At the BCSFB, OATPs and OATs are considered important to regulate PGE2 concentration in CSF. Finally, little is known on function of transporters expressed in brain parenchyma cells including microglia and neuronal cells at this point. Future challenge is needed to clarify those roles, providing us with clues to understand molecular mechanisms underlying neuroinflammation and glial–neuronal interaction.


This research was carried out with the support of a Grant-in-Aid for Scientific Research (KAKENHI, 15H04755) from the Japan Society for the Promotion of Science, and the Smoking Research Foundation (Tokyo, Japan) to T. N.

Conflict of Interest

The authors declare no conflict of interest.

© 2018 The Pharmaceutical Society of Japan