Address correspondence and reprint requests to Dr. R. F. Keep at Department of Surgery (Neurosurgery), University of Michigan, R5550 Kresge I, Ann Arbor, MI 48109-0532, U.S.A. E-mail rkeep@ umich.edu
Abstract : 5-Aminolevulinic acid (5-ALA) is a precursor of porphyrins and heme that has been implicated in the neuropsychiatric symptoms associated with porphyrias. It is also being used clinically to delineate malignant gliomas. The blood-CSF barrier may be an important interface for 5-ALA transport between blood and brain as in vivo studies have indicated 5-ALA is taken up by the choroid plexuses whereas the normal blood-brain barrier appears to be relatively impermeable. This study examines the mechanisms of 5-[3H]ALA uptake into isolated rat lateral ventricle choroid plexuses. Results suggest that there are two uptake mechanisms. The first was a Na+-independent uptake system that was pH dependent (being stimulated at low pH). Uptake was inhibited by the dipeptide Gly-Gly and by cefadroxil, an α-amino-containing cephalosporin. These properties are the same as the proton-dependent peptide transporters PEPT1 and PEPT2, which have recently been shown to transport 5-ALA in frog oocyte expression experiments. Choroid plexus uptake was not inhibited by captopril, a PEPT1 inhibitor, suggesting PEPT2-mediated uptake. The presence of PEPT2 and absence of PEPT1 in the choroid plexus were confirmed by western blotting. The second potential mechanism was both Na+ and HCO3- dependent and appears to be an organic anion transporter, although it is possible that removal of Na+ and HCO3- may indirectly affect PEPT2 by affecting intracellular pH. The presence of PEPT2 and a putative Na+/HCO3--dependent organic anion transporter is important not only for an understanding of 5-ALA movement between blood and brain but also because these transporters may affect the distribution of a number of drugs between blood and CSF.
5-Aminolevulinic acid (δ-aminolevulinic acid ; 5-ALA) is a precursor of porphyrins and heme. Thus, an understanding of 5-ALA production, metabolism, and transport is important in relation to the production of hemecontaining proteins. In addition, 5-ALA may be important in some CNS diseases. An overproduction of 5-ALA is associated with hereditary hepatic porphyrias. These diseases are accompanied by a variety of neuropsychiatric symptoms (Brennan and Cantrill, 1979), and it has been suggested that these symptoms arise from the accumulation of the porphyrin precursors 5-ALA and/or porphobilinogen (Yeung et al., 1987). 5-ALA may be directly neurotoxic (Percy et al., 1981 ; Russell et al., 1983), and it may modulate CNS signal transduction by being an agonist for presynaptic GABA autoreceptors (Brennan and Cantrill, 1979). 5-ALA has also been used clinically to delineate malignant gliomas (Stummer et al., 1998c). This relies on the fact that oral administration of 5-ALA results in a highly selective accumulation of protoporphyrin IX (PpIX) in the glioma, which can then be detected by fluorescence microscopy in the operating theater (Stummer et al., 1998 a, c). Studies on C6 glioma cells in rats also indicate selective accumulation of PpIX into the glioma after systemic administration of 5-ALA, although the choroid plexuses also exhibited a strong PpIX fluorescence after 5-ALA administration (Stummer et al., 1998b).
Most (Terr and Weiner, 1983 ; Garcia et al., 1998) but not all (McGillion et al., 1974) studies have suggested that there is little blood-to-brain 5-ALA transport. An exception is at the choroid plexus, the site of the blood-CSF barrier, where autoradiography (Terr and Weiner, 1983) and PpIX fluorescence (Stummer et al., 1998b) studies indicate selective uptake. This apparently high uptake suggests that there may be 5-ALA transport at the basolateral (blood-facing) membrane of the choroid plexus epithelium that could control the movement of 5-ALA between blood and brain. However, CSF concentrations of 5-ALA are much lower than those found in plasma (Percy and Shanley, 1977 ; Gorchein and Webber, 1987), suggesting that there may also be an active efflux transporter for 5-ALA at the apical (CSF-facing) membrane of the choroid plexus epithelium.
The mechanism(s) involved in choroid plexus transport have not been examined. In a number of other tissues, 5-ALA uptake has been shown to be energy dependent, and there is evidence that 5-ALA can interact with a number of transport systems, including those for GABA (McGeer et al., 1961 ; Becker et al., 1974 ; Garcia et al., 1998), glutamate (McLoughlin and Cantrill, 1984), p-aminohippurate (PAH) (Cheeks and Wedeen, 1986), and the di-/tripeptide transporters PEPT1 and PEPT2 (Doring et al., 1998), although for glutamate there is question over whether 5-ALA is merely an inhibitor rather than a substrate for the transporter (Brennan and Cantrill, 1979).
The current study addresses the issue of the mechanisms of 5-ALA transport at the choroid plexus in vitro as a first step toward understanding the movement of 5-ALA between blood and brain in health and disease. Results indicate that there may be two transport systems for 5-ALA at the choroid plexus, one of which is the peptide transporter PEPT2. This latter finding has implications for peptide as well as 5-ALA movement between blood and brain.
MATERIALS AND METHODS
Experiments were performed on lateral ventricle choroid plexuses isolated from pentobarbital (50 mg/kg i.p.)-anesthetized male Sprague-Dawley rats aged 30-50 days. The methods used have been described in detail previously for studies on glutamine and 86Rb uptake (Keep et al., 1994 ; Keep and Xiang, 1995). The lateral ventricle plexuses were weighed and transferred to control buffer at 37°C. There was a 5-min recovery period prior to the beginning of any experiment.
5-[3H]ALA hydrochloride (3 Ci/mmol) and 5-[14C]ALA hydrochloride (47.6 mCi/mmol) were supplied by New England Nuclear Life Science Products (Boston, MA, U.S.A.). [14C]Mannitol (53 mCi/mmol) and [3H]mannitol (15 Ci/mmol) came from American Radiolabeled Chemicals (St. Louis, MO, U.S.A.). Hyamine hydroxide and Cytoscint were purchased from ICN (Costa Mesa, CA, U.S.A.), and other reagents came from Sigma (St. Louis, MO, U.S.A.).
Most experiments were performed in bicarbonate buffers that were continuously bubbled with 5% CO2/95% O2 and contained 127 mM NaCl, 20 mM NaHCO3, 2.4 mM KCl, 0.5 mM KH2PO4, 1.1 mM CaCl2, 0.85 mM MgCl2, 0.5 mM Na2SO4, and 5 mM glucose (pH 7.3) at 37°C. In experiments where a low-Na+ buffer was used to investigate the role of sodium in 5-[3H]ALA uptake, buffer NaCl and NaHCO3 were replaced by choline chloride and choline bicarbonate, producing a 1 mM Na+ solution due to the presence of Na2SO4.
The effect of pH and bicarbonate removal on 5-ALA uptake was examined in medium buffered with Bis-Tris instead of bicarbonate. This Bis-Tris-based artificial CSF contained 147 mM NaCl, 2.4 mM KCl, 0.5 mM KH2PO4, 1.1 mM CaCl2, 0.85 mM MgCl2, 0.5 mM Na2SO4, 5 mM glucose, and 10 mM Bis-Tris (pH varied from 6.0 to 7.75) and was bubbled with 100% O2. Choline chloride was used to replace NaCl to produce a low-Na+ (1 mM) medium. The bicarbonate-dependent uptake was assessed by the addition of 20 mM NaHCO3 or choline bicarbonate to the Bis-Tris-biffered medium.
After the recovery period, the plexuses were transferred to 0.95 ml of bicarbonate buffer with or without drug for 0.5 min. Uptake was started by addition of 0.05 ml of buffer with ~0.2 μCi of 5-[3H]ALA and 0.1 μCi of [14C]mannitol (an extracellular marker). Unless otherwise stated, this was the only source of 5-ALA (65 nM) in the solutions. Uptake was terminated after 2 min by transferring the plexus to iced control buffer and filtering under reduced pressure. The filters (made from 118-μm nylon mesh ; Tetko, Kansas City, MO, U.S.A.) were washed three times with the same buffer. The filters and choroid plexuses were then soaked in 0.33 ml of 1 M hyamine hydroxide (a tissue solubilizer) for 30 min prior to the addition of scintillation cocktail (Cytoscint) and counting with a dual-channel liquid scintillation counter (Beckman LS 3801, Fullerton, CA, U.S.A.).
Preliminary experiments, where uptake time was varied, showed that 5-[3H]ALA uptake was only unidirectional for the first 2 min, and therefore experiments were limited to this time. Other preliminary experiments utilizing 5-[14C]ALA showed similar uptake rates to that found with 5-[3H]ALA. The uptake of 5-[3H]ALA into choroid plexus, in microliters per milligram of wet weight, was calculated correcting for filter binding and extracellular space as described by Keep and Xiang (1995) and the influx rate constant determined by dividing by the duration of the uptake experiment (2 min). This was then converted to a flux by multiplying by the medium ALA concentration. Under control conditions (2-min uptake, 65 nM ALA, 37°C), filter binding was <1% of total tissue uptake and the extracellular space correction was 3-5%.
To determine the kinetic constants for Na+-independent uptake, data were fit to a two-component model using a non-linear curve-fitting program (Systat 5.2 ; Systat, Evanston, IL, U.S.A.) : v = [Vmax× [S])/(Km + [S])] + (KD× [S]), where [S] is the 5-ALA concentration, Vmax and Km are the maximum velocity and Michaelis constant for saturable uptake, and KD is a diffusion constant for nonsaturable transport. Whether there are multiple Na+-independent transporters was examined using a Woolf-Augustinsson-Hofstee plot.
Pentobarbital (50 mg/kg)-anesthetized male Sprague-Dawley rats aged 30-50 days were killed by decapitation, and the lateral ventricle choroid plexuses were isolated from the brain. The obtained tissues were added to 0.5 ml of western blot loading buffer (62.5 mM Tris, 2% sodium dodecyl sulfate, 10% glycerol, and 0.5% 2-mercaptoethanol) and sonicated. The samples were centrifuged at 14,000 g for 10 min at 4°C. The supernatant was used for protein assay (Bio-Rad, Richmond, CA, U.S.A.) and to load sodium dodecyl sulfate-polyacrylamide gels (7.5% ; Bio-Rad) for electrophoresis. From 50 to 80 μg of protein was then loaded on 7.5% polyacrylamide gels (Bio-Rad).
After electrophoresis, samples were transferred to nitrocellulose membranes (Amersham) and immunoblotted with either rabbit anti-rat PEPT1 or PEPT2 antibodies followed by peroxidase-linked goat anti-rabbit IgG antibody. The derivation and specificity of these antibodies have been described previously (Shen et al., 1999) ; they were raised against synthetic peptides based on the COOH-terminal region of rat PEPT1 (YSSLEPVSQTNM, amino acids 699-710) and PEPT2 (NMINLETKNTRL, amino acids 718-729). The primary antibodies were used at a dilution of 1 : 1,000 (PEPT1) and 1 : 500 (PEPT2), respectively. The secondary antibodies were used at a 1 : 2,500 dilution. Membranes were developed with the ECL system (Amersham, Arlington Heights, IL, U.S.A.) and the images were transferred to x-ray film. To determine specificity, western blots were also performed with antibodies preincubated with the synthetic PEPT1 and PEPT2 peptides.
Comparisons were made using analysis of variance and post hoc multiple comparison tests. For comparisons between multiple experimental groups and a single control group, a Dunnet's test was used. A Scheffé's test was used for multiple comparisons between several groups. Differences were considered significant at the p < 0.05 level (two-tailed test). Data presented in the text are given as means ± SE.
Choroid plexus 5-[3H]ALA uptake in bicarbonate-buffered CSF
Figure 1 depicts the time course of 5-[3H]ALA uptake into isolated rat lateral ventricle choroid plexus in bicarbonate-containing buffer. Uptake was linear for the first 2 min and reached a tissue-to-medium ratio of ~8 : 1 by 30 min. In all of the following experiments, uptake was measured for 2 min. The same graph also shows uptake measured at 4°C at 2 and 30 min. At both time points, there was a marked inhibition (98-99%) of uptake compared with experiments at 37°C.
The uptake of 5-[3H]ALA was markedly (90%) inhibited by the presence of 1 mM 5-ALA in the medium (Fig. 2). Uptake was also dependent on oxidative metabolism [95% reduction in the presence of 1 mM dinitrophenol (DNP)] and partially inhibited by a reduction in medium Na+ from 148 to 1 mM, suggesting that may be Na+-dependent and Na+-independent transporters (Fig. 2A). Evidence, presented below, indicates that the Na+-independent transporters is the proton-coupled peptide transporter PEPT2. Whether the apparent Na+ dependency of 5-[3H]ALA uptake was secondary to an effect on Na+/H+ exchange was therefore examined by determining whether a reduction in medium Na+ concentration would affect 5-[3H]ALA uptake in the presence of a Na+/H+ exchange inhibitor, dimethyl amiloride (100 μM ; Fig. 2B). Dimethyl amiloride markedly reduced the Na+-independent transport (by 90%), suggesting that a Na+/H+ exchange-generated pH gradient is important for the Na+-independent transport. However, although there was some inhibition of Na+-dependent transport by dimethyl amiloride (53%), reduction in medium Na+ from 148 to 1 mM still caused a significant reduction in uptake (p < 0.01 ; Fig. 2B). Thus, it appears that Na+ removal inhibits 5-ALA transport by inhibiting both Na+/H+ exchange and another mechanism, potentially a transporter directly coupled to Na+.
The saturation kinetics of 5-ALA transport were examined (Fig. 3). Kinetic constants were determined for Na+-independent transport, not Na+-dependent transport, because, as described above, the latter probably reflects effects on two transporters. Saturable Na+-independent uptake had a Km of 0.26 ± 0.14 mM and a Vmax of 53 ± 12 pmol/mg/min (Fig. 3). A Woolf-Augustinsson-Hofstee plot indicated that there is a single Na+-independent transporter. In control medium (65 nM ALA), diffusion-mediated uptake was estimated as 0.7 ± 0.1 fmol/mg/min, a rate similar to the uptake of 5-ALA at 4°C (1.1 ± 0.4 fmol/mg/min ; Fig. 1). At physiological (or pathophysiological) concentrations of 5-ALA found in plasma or CSF plasma (Percy and Shanley, 1977 ; Gorchein and Webber, 1987), uptake via diffusion would represent <5% of total uptake.
The effect of addition of potential uptake inhibitors (at 1 mM) to control incubation medium is depicted in Fig. 4. The amino acids GABA, glutamate, histidine, and glycine had no significant effect on uptake. In contrast, the dipeptide Gly-Gly almost abolished uptake. Taurocholate and bromosulfophthalein (BSP), two organic anion transport substrates, reduced 5-[3H]ALA by approximately half, but PAH had no significant effect. The effects of taurocholate and BSP on 5-[3H]ALA uptake were not specific to either of the putative Na+-dependent or Na+-independent transport systems (Fig. 5).
Choroid plexus 5-ALA uptake in Bis-Tris-buffered CSF
Compared with uptake in bicarbonate-buffered CSF, 5-[3H]ALA uptake in Bis-Tris-buffered CSF was reduced (0.37 ± 0.02 vs. 0.68 ± 0.04 μl/mg/min ; p < 0.001). Unlike in bicarbonate-buffered CSF, Na+ removal did not reduce uptake in Bis-Tris-buffered CSF (Fig. 6). However, addition of bicarbonate (and bubbling with O2/CO2) to the Bis-Tris-buffered CSF resulted in a restoration of Na+-dependent transport (Fig. 6) ; that is, it appears that there may be a transport system that is both Na+ and HCO3- dependent present in choroid plexus.
In Bis-Tris-buffered (no bicarbonate) artificial CSF, reducing pH resulted in a stimulation in 5-[3H]ALA uptake, reaching a peak at pH 6.5 (Fig. 7). As with bicarbonate-buffered CSF, 5-[3H]ALA uptake was nearly abolished by the dipeptide Gly-Gly in Bis-Tris CSF (Fig. 8), suggesting the involvement of a peptide transporter. Cefadroxil, a compound that inhibits PEPT1 and PEPT2 transport, also almost abolished 5-[3H]ALA uptake, whereas captopril, an angiotensin-converting enzyme inhibitor with a higher specificity for PEPT1 than for PEPT2, did not reduce 5-[3H]ALA uptake (Fig. 8).
Western blots examining whether PEPT2 and PEPT1 are present in choroid plexus are shown in Fig. 9. For PEPT2, protein extracted from renal brush border membrane vesicles was used as a positive control (Shen et al., 1999), and PEPT2 polypeptide was detected as a broad band at ~85 kDa (Fig. 9A). A band at the same location was also detected in choroid plexus. Both the choroid plexus and the renal bands were greatly reduced by preincubation of the PEPT2 antibody with the synthetic PEPT2 polypeptide (data not shown). For PEPT1, protein from the small intestine was used as a positive control (Shen et al., 1999), and for that tissue PEPT1 was detected as a band at ~90 kDa (Fig. 9B). Choroid plexus tissue was, however, negative for PEPT1.
These experiments indicate that there may be two apparent transport systems for 5-ALA at the choroid plexus. There is a Na+-independent system that both transport data and western blotting indicate is PEPT2. There may also be an Na+-dependent system that is also bicarbonate dependent, which has not yet been identified.
Compared with other in vitro measurements of 5-ALA transport in brain tissues, the choroid plexus has a high rate of uptake. Thus, at 10 μM the choroid plexus uptake of 5-ALA was 5.4 pmol/mg of wet wt/min or ~27 pmol/mg of protein/min. In contrast, in cultured neurons and glia, Percy et al. (1981) reported uptakes of ~2-5 pmol/mg/min, whereas Brennan and Cantrill (1979) and Garcia et al. (1998) could not detect any 5-[14C]ALA uptake into synaptosomes and cerebral capillaries, respectively. We have also examined 5-ALA uptake into C6 glioma cells in culture (A. Novotny, W. Stummer, and R. F. Keep, unpublished observations) and that uptake is 20- to 30-fold less than described here for choroid plexus.
Several pieces of evidence indicate that the Na+-independent uptake of 5-ALA is mediated by PEPT2. PEPT2 is a Na+-independent transporter that transports di- and tripeptides (Fei et al., 1994 ; Boll et al., 1996) ; 5-ALA uptake was blocked by the dipeptide Gly-Gly, whereas the single amino acid glycine had no effect. PEPT2-mediated transport is proton coupled and thus pH dependent (Fei et al., 1994 ; Chen et al., 1999) ; like peptide uptake mediated by PEPT2, 5-ALA uptake was pH dependent, being stimulated at low pH. PEPT2 transports cephalosporins with α-amino carbons, such as cefadroxil (Terada et al., 1997), and cefadroxil inhibited choroid plexus 5-ALA uptake. PEPT2, unlike PEPT1, has reduced specificity for the angiotensin-converting enzyme inhibitor captopril (Boll et al., 1996), and our results indicate that captopril does not significantly inhibit 5-ALA uptake. In addition, we have evidence from western blots indicating that PEPT2 is present at choroid plexus but PEPT1 is not. The mRNA for PEPT2 has also been detected in choroid plexus (Berger and Hediger, 1999). That PEPT2 can mediate 5-ALA transport is indicated by an expression study in frog oocytes (Doring et al., 1998). The Km for PEPT2-mediated 5-ALA transport in that study (0.23 mM) was similar to the Km determined for Na+-independent transport in this choroid plexus study (0.26 mM).
The mRNA of another Na+-independent peptide transporter, PHT-1, has been identified in choroid plexus (Yamashita et al., 1997). That transporter, however, does not appear to have a major involvement in choroid plexus 5-ALA transport as L-histidine, a PHT-1 substrate, did not significantly inhibit uptake at saturating concentrations.
The putative Na+ - and HCO3--dependent transport system has not been fully identified. There is still the possibility that the effects of Na+ and HCO3- removal might be due to pH-mediated inhibition of PEPT2, although the fact that Na+ removal still reduced 5-ALA transport in the presence of a Na+/H+ exchange inhibitor suggests that this is not the case. Other forms of experimentation (e.g., PEPT2 knockout mice) are needed to verify whether there are indeed two transporters. Previous studies on other tissues have suggested that 5-ALA interacts with GABA, glutamate, and PAH transporters (McGeer et al., 1961 ; Becker et al., 1974 ; McLoughlin and Cantrill, 1984 ; Cheeks and Wedeen, 1986 ; Garcia et al., 1998). Evidence (effects of Na+ removal or ouabain) indicates that there are Na+-dependent transporters for these three compounds at the choroid plexus (Holloway and Cassin, 1972 ; Kim et al., 1996 ; Ramanathan et al., 1997 ; J. Xiang and R. F. Keep, unpublished observations on glutamate and PAH). However, none of those compounds significantly competed with 5-ALA uptake into choroid plexus even when present at 1 mM, a concentration that should have significantly saturated each of the transporters (Kim et al., 1996 ; Ramanathan et al., 1997 ; X. Xiang and R. F. Keep, unpublished observations on glutamate and PAH). Although PAH did not effect 5-ALA uptake, two other organic anions, taurocholate and BSP, did significantly reduce Na+-dependent transport. Taurocholate and BSP are substrates for a variety of organic anion-transporting polypeptides (oatp), of which at least oatp1 and oatp2 are present in the choroid plexus (Angeletti et al., 1997 ; Gao et al., 1999). However, although oatp1-mediated transport may be bicarbonate dependent (Satlin et al., 1997), none of the oatp transporters are Na+ dependent. Another superfamily of organic anion transporters, OAT1, OAT2, and OAT3, have been cloned (Sekine et al., 1997, 1998 ; Kusuhara et al., 1999). They are not directly coupled to Na+, but OAT1, at least, is indirectly coupled to Na+ as it mediates organic anion/dicarboxylic acid exchange where the dicarboxylic acid gradient is maintained by Na+/dicarboxylic acid transport (Roch-Ramel, 1998). OAT3 appears to be the predominant form found in brain, and preliminary data suggest it is present at the choroid plexus (Kusuhara et al., 1999), but its affinity for PAH [Km = 65 μM (Kusuhara et al., 1999)] is such that 1 mM PAH should completely saturate the transporter, whereas this concentration had no effect on 5-ALA uptake in choroid plexus. Sodium-dependent taurocholate transporters are present in liver (Hagenbuch et al., 1991) and ileum/kidney (Wong et al., 1994). However, the mRNA for both of those transporters is not present in brain (Hagenbuch et al., 1991 ; Craddock et al., 1998). Thus, our results suggest that there may be yet another form of Na+-dependent organic anion transport present at the choroid plexus.
The Na+-independent transport system had a Km value of 0.26 mM. In comparison, Gorchein and Webber (1987) found that the physiological concentrations of 5-ALA in humans are ~92 nM in plasma and 19 nM in CSF. In that study, the plasma concentration in 89 subjects never exceeded 270 nM, whereas the CSF concentration never exceeded 36 nM. Percy and Shanley (1977) examined patients in the acute phase of variegate porphyria, but even there the maximum plasma concentration was only ~9 μM. Thus, plasma and CSF 5-ALA concentrations, even in disease states, always appear to be much lower than the Km values for the Na+-independent system. It appears, therefore, very unlikely that the neurotoxic effects of 5-ALA in porphyrias could be mediated by competition effects at the transporter (e.g., 5-ALA-reducing transport of neuroactive peptides by PEPT2).
These results on isolated choroid plexuses do not address the issue of whether the transporters are on the apical or basolateral membrane of the choroid plexus epithelial cell. However, the uptake experiments in this study were of short duration (2 min), suggesting that the measured uptake may be primarily apical and, as noted above, 5-ALA concentrations in CSF are considerably lower than in plasma (Percy and Shanley, 1977 ; Gorchein and Webber, 1987), suggesting that there is an active mechanism clearing 5-ALA from CSF. This suggestion is supported by studies on the blood-to-brain transport of dipeptides, which indicate that there is a very restricted entry into brain (Zlokovic, 1995). 5-ALA injected into the bloodstream does appear in choroid plexus (Terr and Weiner, 1983) and is converted to PpIX (Stummer et al., 1998b), but whether that uptake is directly through the basolateral membrane or results from diffusion into CSF followed by uptake at the apical membrane is as yet uncertain. It is also possible that the Na+ -independent PEPT2 transporter and the putative Na+ -dependent transporter are on different membranes. An elucidation of the location and, for the Na+ -dependent transporter, also the substrate specificity of these transporters is important not only for an understanding of 5-ALA movement between blood and brain (for example, in porphyrias) but also because they affect the distribution of a number of drugs between blood and CSF (and thus brain).