Modulation of Monocarboxylic Acid Transporter-1 Kinetic Function by the cAMP Signaling Pathway in Rat Brain Endothelial Cells*

MCT1 (monocarboxylic acid transporter 1) facilitates bidirectional monocarboxylic acid transport across membranes. MCT1 function and regulation have not been characterized previously in cerebral endothelial cells but may be important during normal cerebral energy metabolism and during brain diseases such as stroke. Here, by using the cytoplasmic pH indicator 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein-acetoxymethyl ester, the initial rates of monocarboxylate-dependent cytoplasmic acidification were measured as an indication of MCT1 kinetic function in vitro using the rat brain endothelial cell (RBE4) model of blood-brain transport. The initial rate of l-lactate-dependent acidification was significantly inhibited by 5–10-min incubations with agonists of intracellular cAMP-dependent cell signaling pathways as follows: dibutyryl cAMP, forskolin, and isoproterenol. Isoproterenol reduced Vmax but did not affect Km values. The effects of forskolin were completely reversed by the protein kinase A inhibitor H89, whereas H89 alone increased transport rates. Cytoplasmic cAMP levels, measured by radioimmunoassay, were increased by forskolin or isoproterenol, and the effect of isoproterenol was inhibited by propranolol. MCT1-independent intracellular pH control mechanisms did not contribute to the forskolin or H89 effects on MCT1 kinetic function as determined with amiloride, monocarboxylate-independent acid loading, or the transport inhibitor α-cyano-4-hydroxycinnamate. The data demonstrate the direct modulation of MCT1 kinetic function in cerebral endothelial cells by agents known to affect the β-adrenergic receptor/adenylyl cyclase/cAMP/protein kinase A intracellular signaling pathway.

MCT1 (monocarboxylic acid transporter 1) facilitates the diffusion of monocarboxylic acids such as L-lactic acid, pyruvic acid, and keto acids across biological membranes. It is an obligatory symporter that carries a dissociated proton-monocarboxylate pair with each transport cycle (for review see Refs. [1][2][3]. Based on sequence homology, MCT1 has been characterized as a member of the monocarboxylic acid transporter family of solute carriers (SLC16 family (4)) that includes up to 14 subtypes (MCT1- 14), although the monocarboxylate substrate specificity has been demonstrated for only 4 (5). MCT1 is the prominent monocarboxylic acid transporter in the cerebral microvascular endothelium, where it is present on the luminal and abluminal membranes of vascular endothelial cells (6,7). Here MCT1 is proposed to be the major regulator of bidirectional monocarboxylic acid transport between the brain and the blood, and may have pharmacologic importance for treating brain diseases such as stroke (2,8).
The vast majority of strokes are associated with severe cerebral lactic acidosis, which is a key factor leading to permanent brain cell damage (for review see Ref. 9). Decreased activity of MCT1 in the cerebral microvasculature may be a critical determinant of acid-related cell damage during stroke, and this is evident because lactic acid is not cleared from brain during cerebral ischemia (10). Similarly, overexpression of MCT1 mRNA in brain is correlated with a reduction in cell damage following experimental transient focal cerebral ischemia in rodents. 3 Therefore, factors that regulate the kinetic function of MCT1 in the cerebral microvascular endothelium may be important in mediating the extent of lactic acidosis and therefore brain damage during stroke.
The regulation of MCT1 function is suggested by its shared structural similarity with other transporters that are kinetically regulated by specific intracellular signaling mechanisms. These include sodium/hydrogen exchangers (11,12), sodium/calcium exchangers (13), and glucose transporters (14,15). Regulation of MCT1 by phosphorylation is suggested by the presence of sequence motifs that have a high phosphorylation potential in its intracellular protein domain. Recently, hormoneinduced translocation of MCT1 from cytoplasm to plasma membrane has been shown to regulate MCT1-dependent butyrate uptake in a cell line from human intestine, suggesting the possibility of such a regulatory mechanism for cerebral vascular MCT1 (16). Although the Michaelis-Menten kinetics and inhibition of MCT1 have been clearly defined by a number of studies in other cells (17)(18)(19), neither its basic kinetics nor its functional regulation in cerebral microvascular cells has been investigated.
Here we report the modulation of MCT1 kinetic function in the RBE4 cell model of rat cerebral endothelial cells by agents known to affect the ␤-adrenergic/adenylyl cyclase/protein kinase A-intracellular signal transudation pathway. The elucidation of this regulation improves our understanding of the roles that monocarboxylate and MCT1 have in normal brain energy metabolism. It also suggests the pharmacological modulation of MCT1 kinetic function as a novel target for drug-based therapies to reduce the damaging effects of lactic acidosis during cerebral ischemia and stroke.
Gene Expression by Real Time PCR-Primer pairs for real time PCR were designed using Oligo 6.4 software (Molecular Biology Insights, Cascade, CO) ( Table 1). Total RNA was extracted from RBE4 cells using an RNeasy mini kit (Qiagen). cDNA was reverse-transcribed from total RNA using an Omniscript TM kit (Qiagen) with oligo(dT) and random hexamer primers. Reactions for quantitative reverse transcriptase PCR were performed on a LightCycler TM instrument using the Light-Cycler TM DNA Master SYBR Green I kit (Roche Applied Science). Before quantitative measurements of cDNA were performed, PCR conditions for each primer set were optimized, and a melting curve analysis was used to confirm that each primer pair produced a single product. Standard curves for candidate cDNAs were prepared from a series of five 10-fold serial dilutions of target cDNA. The quantity of DNA in each sample was normalized to 18 S rRNA or glyceraldehyde phosphate dehydrogenase.
Immunocytochemical Analysis of MCT1 Protein Expression-A polyclonal antibody was raised against a peptide coding the 15 carboxylterminal amino acids of rat monocarboxylic acid transporter 1 (MCT1, LQNSSGDPAEEESPV), isolated from chicken egg yolks, affinity-purified, and used in immunocytochemical studies (6). The antibody recognizes a protein migrating with an apparent molecular mass of 46 -50 kDa on immunoblots of brain membrane proteins from adult and suckling rats (data not shown) (6). The avidin-biotin-peroxidase complex immunocytochemistry elite kit (Vector Laboratories, Burlingame, CA) was employed with color development for 1-4 min in 3,3Ј-diaminobenzidine (Dako, Carpinteria, CA). To provide controls for nonspecific staining, the antigenic peptide was added to the primary antibody at 20 g/ml before use.
Determination of pH-dependent Fluorescence-BCECF fluorescence was determined using an inverted Nikon Eclipse TE300 fluorescence microscope with a 10ϫ objective, a 490 Ϯ 20-nm excitation filter combined with a 1.0 neutral density filter, a 440 Ϯ 20-nm excitation filter, and a 535 Ϯ 25-nm emission filter cube (Omega Optical, Inc., Brattleboro, VT). A Sutter 10-position filter wheel with a LUDL controller was used to control the excitation filtering. Fluorescence images were obtained using a SIT68 video camera (Dage-MTI, Inc.) and computerized using a MuTech MV1000 frame grabber and MetaFluor Imaging System software, version 4.6 (Universal Imaging Corp., West Chester, PA). Cells were excited for 100 ms at each wavelength, with a sample rate of ϳ1 image pair per 1.8 s. Background fluorescence was subtracted from a cell-free region of each image during acquisition. The image pairs were stored digitally for off-line analysis using MetaFluor Imaging System software (above) and Microcal TM Origin TM graphical analysis software (Microcal Software, Inc., Northampton, MA).
Calibration of BCECF Fluorescence to pH i -Ratiometric images were calibrated to indicate the approximate pH i from image regions corresponding to individual cells by using the method of Thomas et al. (21). Briefly, cells were perfused with HBS substituted with 135 mM KCl and 5 mM NaCl, containing the K ϩ /H ϩ -exchange ionophore nigericin (10 -20 M), in solutions of different pH values (adjusted with NaOH). Nigericin equilibrates the intracellular and extracellular pH in the presence of high K ϩ (21). At the end of each experiment, the fluorescence ratio from nigericin-treated cells was determined at eight pH values ranging from pH 6.2 to 8.2. The calibration data were fitted to a sigmoidal dose-response curve using Microcal Origin TM software to determine the maximum (R max ) and minimum (R min ) fluorescence ratios and the pK of the dye in each cell. The 495/440 excitation ratio data (R) was transformed to determine the approximate pH i at each time point and for each individual cell in all experiments, using the following equation: pH i ϭ pK ϩ log((R Ϫ R min )/(R max Ϫ R)). As an example, the sigmoidal fit for the average R versus extracellular pH in nigericin from 20 cells (ϮS.E.) is illustrated; however, in practice these data were determined for each individual cell (Fig. 3A).
Initial Rates of Acidification-All experiments were conducted at room temperature. Media were exchanged within 2 s by aspirating the imaging chamber (Warner Instruments RC-21BRW) and flushing it with 10 -20 chamber volumes (5 ml) of fresh medium. The initial rates of cytoplasmic acidification (v i ) were determined using Microcal Origin TM software by line fitting the first four or more points of pH i time data occurring in the earliest, most linear region of a response to bathapplied substrate and were expressed as pH units/s (Fig. 3B). Michaelis-Menten kinetic parameters (V max and K m ) were determined by plotting Cytoplasmic cAMP Concentrations-Cytoplasmic cAMP levels were measured by radioimmunoassay. Briefly, subconfluent RBE4 cells growing in 35-mm culture dishes (Corning) were treated at room temperature with drugs in HBS for 5-10 min in the presence or absence of the phosphodiesterase inhibitor, 3-isobutyl 1-methylxanthine (25 mM). The media were aspirated and replaced with 1 ml of absolute ethanol, and the cells were scraped, sonicated, and pelleted, and the level of cAMP in the supernatant was measured using a TRK.432 kit (Amersham Biosciences) following the manufacturers' recommended protocol. cAMP levels were expressed as picomoles of cAMP per 35-mm culture dish.

TABLE 1 RT-PCR primers
The primers were specific for isoforms of monocarboxylic acid transporter cDNA templates.

MCT1
Forward, 5Ј-ATG TAT GCC GGA GGT CCT ATC-3Ј  Reverse, 5Ј-CCA ATG GTC GCT TCT TGT AGA-3Ј  MCT2  Forward, 5Ј-CTG GCT GTC ATG TAC GCA GGA- Drugs and Solutions-All drugs, unless otherwise noted, were purchased from Sigma. Drugs were dissolved in ethanol, dimethyl sulfoxide, HBS, or water as recommended by the manufacturer and diluted to their final concentrations in HBS, and the pH was adjusted if necessary (KOH). In no case did any vehicle at its final concentration cause measurable effects on RBE4 cell pH i (data not shown). The [Cl Ϫ ] and osmolarity of media containing transporter substrates were kept constant by substituting the substrates for an equivalent portion of NaCl.
Statistics-All statistics and curve fitting were performed using Microcal Origin TM software and applied as indicated in the figure legends or text.

Functional MCT1 Is Expressed in RBE4
Cells-Strong immunoreactivity against MCT1 protein was present in RBE4 cell cultures using a polyclonal antibody that is specific for MCT1 (6). Staining was evident throughout the plasma membrane of all cells and was completely inhibited by the antigenic peptide, confirming that the RBE4 cells expressed robust levels of MCT1 protein (Fig. 1). mRNA for MCT1 was the predominant monocarboxylic acid transporter transcript expressed in these cells, as shown by quantitative RT-PCR analysis (Fig. 2). The MCT2 RT-PCR product was minimally detected at less than 15% of the MCT1 product. It was not surprising that MCT2 mRNA would be low in these cells because MCT2 protein is not detectable in the rat cerebral endothelium from which RBE4 cells are derived (22,23). MCT1 is, however, the only detectable monocarboxylate transporter in rat brain endothelium in situ (6). These data suggested that MCT2 is relatively insignificant in RBE4 cells and was further supported by our kinetic analysis (see below). MCT4, -7, and -8 PCR product levels were very low in RBE4 cells, between 0.16 and 4% of the MCT1 product (Fig. 2). The combined data suggested that RBE4 cells would be an excellent model for the study of MCT1 protein function.
To measure the function of MCT1, we used ratiometric fluorescence video microscopy of cells loaded with the pH i indicator BCECF-AM, and we measured the responses to extracellularly applied monocarboxylate substrates of MCT1. All experimentation was conducted in bicarbonate-free buffer to minimize the activity of bicarbonate-dependent pH i regulatory mechanisms. This method has been shown previously to give very reliable estimates of MCT1 kinetic parameters and is comparable in efficacy to methods using intracellular pH-sensitive electrodes and uptake of L-[ 14 C]lactate (17)(18)(19)24).
BCECF-AM-loaded RBE4 cells showed cytoplasmic fluorescence that was even and bright. Fluorescence ratios of BCECF-loaded cells, excited at 495/440 nm, were calibrated to indicate the approximate pH i using the method of Thomas et al. (21). In these experiments, the data exhibited minimal variation and were tightly fit by the theoretical parameters used in the calibration, further indicating the high efficacy of this method (Fig. 3) (21). Resting pH i was stable for periods longer than 1 ⁄ 2 h, whereas experiments were typically completed within 10 min. pH i   presence of 20 M nigericin, and HBS was substituted with 135 mM K ϩ and 5 Na ϩ . The data were fit with a sigmoidal dose-response curve to determine the pK of the dye, and the maximum and minimum 495/440 excitation ratios (R max and R min ) for these cells. The data in this figure are given as an example of the data quality and goodness of the curve fit for 20 cells; however, similar data from each individual cell was determined for calibrating all experiments in this report (see "Experimental Procedures"). B, to determine the lactate-dependent v i , the extracellular medium was rapidly exchanged at time ϭ 23 s with medium containing 20 mM L-lactate and again at time ϭ 105 s with lactate-free HBS. Cells responded immediately with a reversible cytoplasmic acidification of about 0.6 pH units. The first four data points of the acidification were selected for linear regression analysis, as indicated by the vertical line intersecting the points. The average v i determined by that fit was 0.0307 Ϯ 0.0035 (S.E.) pH units/s for these 10 cells.
was unaffected by repetitive exchanges of the bath with fresh HBS. Rapid exchange of the bathing medium with HBS containing 20 mM L-lactate induced an immediate acidification of the cytoplasm of all cells that was completely reversed upon reperfusion with lactate-free HBS (Fig. 3B). As a measure of MCT1 kinetic function, the initial rate of v i was determined by linear regression of the earliest, most linear portion of this response in each cell and was expressed in pH units/s (Fig. 3B).
Plots of the average v i measured in groups of 19 -40 cells, Ϯ S.E., were determined at different concentrations of MCT1 substrates and fit to the Michaelis-Menten equation (Fig. 4). A summary of kinetic data ( Table 2) indicates that RBE4 cells responded to the MCT1 substrates, L-lactate, D-3-hydroxybutyrate, and pyruvate very consistently with previously published kinetic studies that characterized the Michaelis-Menten parameters for MCT1 function using both BCECF and pH-sensitive intracellular electrodes (17)(18)(19). These data supported our RT-PCR data suggesting that MCT2 would not be significant in RBE4 cells, because K m values for MCT2, as determined previously with these substrates, are only between 8 and 21% of K m values for MCT1 (25). In the presence of 5 mM ␣-cyano-4-hydroxycinnamate (CHC), a specific MCT1 inhibitor, v i was linear with increasing concentrations of L-lactate (not shown). CHC blocked nearly all of the lactate-dependent acidification at the highest lactate concentrations, further suggesting the specificity of these studies for measuring MCT1 functionality (Fig. 4). Combined, our immunocytochemistry, quantitative RT-PCR, Michaelis-Menten kinetics, and CHC inhibition of lactate-dependent v i were very consistent with other studies that have defined MCT1 kinetic func-tion, indicating the specificity of our experimental system for measuring MCT1 function.
MCT1 Kinetic Function Is Modulated by Intracellular cAMP Signaling Pathways in RBE4 Cells-Identical RBE4 cultures always responded to 25 mM L-lactate consistently, without statistically significant variation of v i , when imaged on the same day. K m was also very consistent over time, producing values very near 4 mM L-lactate when tested on at least four separate occasions. In contrast, V max and the 25 mM L-lactate-dependent v i varied daily, depending on the particular batch of cells used and their length of time in culture. Therefore, v i values were compared only among identical cultures from the same day and were internally consistent in their responses to L-lactate concentrations (data not shown).
To test the hypothesis that elevated cytoplasmic concentrations of cAMP might lead to a modulation of MCT1 kinetic function, we incubated RBE4 cells in the membrane-permeant cAMP analog Bt 2 cAMP or the adenylyl cyclase activator forskolin for 5-10 min prior to measuring L-lactate-dependent v i values. Neither drug caused a perceptible change in the resting pH i during the preincubation (not shown); however, the initial rate of acidification with 25 mM L-lactate was significantly reduced by both drugs (Fig. 5A). This result was repeated on more than four separate occasions.
A role for cAMP-dependent protein kinase A (PKA) in slowing lactatedependent v i was suggested because the PKA activator, Bt 2 cAMP, inhibited MCT1 function. To determine whether the forskolin-dependent reduction in lactate-dependent v i was mediated by PKA, we pretreated cells with the specific PKA antagonist 20 M H89, with and without forskolin, for 5 min before measuring responses to 25 mM L-lactate. H89 did not cause a perceptible change in the basal pH i during the preincubation (not shown). However, inclusion of H89 in the forskolin preincubation completely reversed the effects of forskolin (Fig. 5B). Interestingly, H89 alone caused a significant increase in L-lactate-dependent v i , suggesting that PKA may be exerting a tonic slowing effect on MCT1 kinetic function under control conditions (Fig. 5B). The above results were repeated on at least three separate occasions and suggest a role for PKA in modulating MCT1 kinetic function in RBE4 cells.
Because PKA activation is downstream of cell surface receptor-mediated mechanisms (26,27), and because cerebral endothelium is known to express ␤-adrenergic receptors that activate PKA, we tested the hypothesis that a ␤-adrenergic receptor agonist would lead to a reduction in the lactate-dependent v i . Isoproterenol caused a dose-dependent slowing of v i consistent with a ␤-adrenergic receptor-mediated process having an IC 50 near 4 M when tested with 25 mM L-lactate (Fig.  6A). The transport-slowing effect of 100 M isoproterenol was similar in magnitude to that observed with Bt 2 cAMP and forskolin (Figs. 5 and 6). Inclusion of the ␤-adrenergic receptor antagonist, propranolol (10 nM), in the preincubation with isoproterenol, led to a statistically significant reversal of the isoproterenol-dependent reduction in v i . However, the  reversal of inhibition was not complete (data not shown). Isoproterenol or propranolol alone did not cause a perceptible change in the basal pH i during the preincubation (not shown). Comparison of Michaelis-Menten plots constructed at varying L-lactate concentrations in the presence and absence of 100 M isoproterenol showed that drug treatment caused a 14% reduction in V max but only a 2% change in K m (Fig. 6B).
The combined data demonstrated that MCT1 kinetic function in RBE4 cells is modulated by agents known to affect the ␤-adrenergic receptor/ adenylyl cyclase/cAMP/PKA intracellular signal transduction pathway. Forskolin and Isoproterenol Pretreatment Increased cAMP Levels in RBE4 Cells-Cytoplasmic cAMP levels were measured by radioimmunoassay in identical cultures of RBE4 cells. cAMP was not detectable in control cultures but increased when cells were incubated for 5 min in the phosphodiesterase inhibitor 3-isobutyl 1-methylxanthine ( Table 3).
The data demonstrated that cAMP was produced in RBE4 cells under resting conditions. Five-to 10-min pretreatments in 100 M isoproterenol or 100 M forskolin in 3-isobutyl 1-methylxanthine caused cAMP levels to become elevated. This indicated that both agents, at the concentrations used in the fluorescence imaging studies, led to enhanced production of cAMP in the RBE4 cells. The rise in cAMP induced by isoproterenol was greatly reduced by the addition of 10 nM propranolol in the preincubation ( Table 3). The combined data showed that forskolin and isoproterenol-mediated slowing of lactate-dependent v i were associated with elevated cytoplasmic cAMP in RBE4 cells.
Modulation of L-Lactate-dependent v i Was Not Because of MCT1independent pH i Regulatory Mechanisms-The cAMP signaling pathway is well characterized as a modulator of other pH i regulatory proteins such as the Na ϩ /H ϩ exchanger (11,12). Therefore, we investigated the possibility that forskolin-and Bt 2 cAMP-mediated effects on v i might be due to enhanced proton efflux through Na ϩ /H ϩ exchangers, rather than by slowed proton influx through MCT1. When tested with 25 mM L-lactate, neither the control v i nor the slowing of v i induced by 5-10-min preincubations in Bt 2 cAMP or forskolin were changed by 10-min preincubations with the well characterized Na ϩ /H ϩ exchanger inhibitor 100 M amiloride (28) (Fig. 7). These data suggest that amiloride-sensitive proton transporters, such as the Na ϩ /H ϩ exchanger, were not mediating the forskolin-or Bt 2 cAMP-dependent slowing of lactate-dependent v i .

TABLE 3 Cytoplasmic cAMP levels from 35-mm cultures of subconfluent RBE4 cells
In the presence of 25 mM IBMX, a 5-min incubation with 100 M forskolin or a 10-min incubation in 100 M isoproterenol led to increased levels of cAMP. The isoproterenol-mediated elevation in cAMP was blocked when 10 nM propranolol was included in the incubation. cAMP was not detectable in the absence of 3-isobutyl 1-methylxanthine (IBMX). ND indicates none detected.  Amiloride alone did not effect v i . Ten-minute preincubation with 500 M Bt 2 cAMP (dbcAMP) or 100 M forskolin alone caused a significant inhibition of v i that was not significantly changed when amiloride was included in the preincubation mixture. For each bar, v i was normalized to the control rate determined in the absence of drugs. n ϭ 20 cells for all data points, Ϯ S.E., and statistical significance was determined by comparing control to drug-treated groups with a paired Student's t test, p ϭ 0.005.
To examine further the potential role of MCT1 and lactate-independent pH i regulatory processes in modulating lactate-dependent v i , the rates of intracellular alkalinization were measured in RBE4 cells following acid loading by the NH 4 Cl pre-pulse method (29). NH 4 ϩ caused a rapid and sustained elevation of pH i , and subsequent perfusion with Na ϩ -free HBS led to hyperacidification of the cytoplasm. Following this acid loading, the pH i recovered very slowly relative to v i in Na ϩ -free HBS, indicating the presence of a slow Na ϩ -independent pH i regulatory mechanism in the cells (Fig. 8A). During Na ϩ reperfusion, pH i alkalinized slowly compared with v i , suggesting the presence of functional Na ϩ -dependent pH i regulatory mechanisms, such as Na ϩ /H ϩ exchangers. However, because the fastest alkalinization rates were less than 10% of the lactate-dependent v i , lactateindependent pH i regulatory mechanisms measured by this method could not have made a significant contribution to lactate-dependent v i (Fig. 8A). More importantly, we investigated whether the rates of these lactateindependent pH i regulatory mechanisms were significantly affected by 10-min preincubations in either H89 or forskolin. Following NH 4 Cl acid loading, the rates of intracellular alkalinization, measured in Na ϩ -free and Na ϩ -containing HBS, were not significantly affected by H89. Moreover, the rates of alkalinization decreased somewhat after pretreatment in forskolin (Fig. 8B). Because forskolin decreased the rate of proton efflux from acid-loaded cells, the effect of this drug on lactate-independent pH i regulatory mechanisms could only have resulted in a small underestimation of v i (approximately less than 2%). Thus, the effects of H89 and forskolin on lactate-independent pH i regulatory mechanisms such as Na ϩ /H ϩ exchangers could not have been the underlying reason for the effects of these drugs in modulating MCT1 function.
To examine more directly the specificity of H89-and forskolin-mediated slowing of lactate-dependent v i , MCT1 function was inhibited by a 2-min pretreatment with 5 mM CHC (see above), and the effects of H89 and forskolin were measured. Lactate-dependent v i was not signif-icantly affected by either H89 or forskolin in the presence of CHC (Fig.  9). These data demonstrate that modulation of lactate-dependent v i by H89 and forskolin, as shown in Fig. 5B, does not occur when MCT1 is inhibited (Fig. 9). Combined, the data in Figs. 7-9, strongly suggest that MCT1-independent pH i regulatory mechanisms were not responsible for the modulation of lactate-dependent v i observed in RBE4 cells.

DISCUSSION
The ␤-Adrenergic Receptor/cAMP/PKA Signaling Pathway-regulated MCT1 Kinetic Function in Rat Brain Endothelial Cells-The data presented here support the hypothesis that stimulation of ␤-adrenergic receptors causes an adenylyl cyclase-mediated increase in intracellular cAMP levels and an associated PKA-dependent regulation of MCT1 FIGURE 8. Rates of pH i recovery from NH 4 Cl-mediated acid loading were much slower than lactate-dependent acidification in RBE4 cells. A, as a control, perfusion with extracellular medium containing 20 mM NH 4 Cl (substituted for NaCl) caused the pH i to become more alkaline. Upon removal of NH 4 ϩ , by perfusing with NaCl-free HBS (substituted with choline-Cl), cells became hyperacidic. pH i recovered during Na ϩ -free perfusion with a rate that was less than 10% as fast as the rate of lactate-dependent v i in these cells (not shown). Recovery to the resting pH i continued at a slower pace during reperfusion in Na ϩ -containing HBS. B, a short preincubation in 20 M H89 did not significantly affect either of the rates of alkalinization following acid loading. Preincubation in 100 M forskolin (Forsk) caused both the Na ϩ -free and Na ϩ -containing rates of alkalinization to be slowed as compared with control. Rates of alkalinization, following acid load, were measured by linear regression and expressed as the absolute percent of the 25 mM L-lactate-dependent v i measured in these same cells before drug application (lactate responses not shown). n ϭ 19 or 20 cells for all data points, Ϯ S.E. . Proposed regulation of MCT1-mediated transport by the ␤-adrenergic/ PKA signaling pathway. Extracellular agonists such as isoproterenol bind to guanine nucleotide-binding protein-coupled ␤-adrenergic receptors (␤-ad) and activate adenylyl cyclase (AC) causing intracellular cAMP production. Elevated cAMP stimulates PKA activity and phosphorylation of downstream target proteins. One of the downstream effects of PKA is decreased MCT1 kinetic function, although the details of this mechanism are unknown. MCT1 internalization is a possibility as suggested by the shift in its V max with isoproterenol stimulating PKA. This may be mediated via the MCT1 chaperone protein CD-147; however, direct phosphorylation of MCT1 or other protein-MCT1 interactions must also be considered. kinetic function in cerebral endothelial cells (Fig. 10). As discussed below, this finding contributes importantly to our understanding of the role that monocarboxylates and their transporters fulfill during brain homeostasis in both health and disease.
The ␤-adrenergic receptor-dependent inhibition of MCT1 kinetic function in RBE4 cells is consistent with a number of studies showing that isoproterenol (30) and cAMP have the general effect of reducing endothelial cell permeability (30 -32). Most interesting is that this effect of cAMP is hypothesized to provide protection for brain cells from various forms of stress or disease, albeit by unknown mechanisms. Thus, the inhibition of MCT1 function in cerebral microvascular cells may be part of such neuroprotective mechanisms mediated at the blood-brain barrier (31).
Based on the evidence presented, the signaling pathway between the ␤-adrenergic receptor and PKA is relatively defined in the RBE4 model (Fig. 10). However, the pathway downstream of PKA that links protein phosphorylation to altered MCT1 activity remains to be elucidated. A reduction in the number of functional transporters and not a more subtle modification of the basic activity of the transporter is a likely mechanism for the kinetic modulation shown here, because the V max , but not K m , value for L-lactate transport was affected by isoproterenol (Fig. 6B). This finding compliments a recent study in a human intestinal cell line showing a hormone-mediated increase in MCT1 function through enhanced recruitment of the MCT1 protein to the plasma membrane (16). Whether this mechanism also applies to MCT1 in RBE4 cells or whether another mechanism such as direct phosphorylation of MCT1, modification of an accessory protein such as CD147, or other protein-protein interaction is involved will require further investigations. It is, however, unlikely that gene expression and de novo protein synthesis were important in these studies, because all drug treatments and pretreatments were limited to less than 10 min. Although a forskolin-dependent up-regulation of MCT1/CD147 mRNA and protein synthesis in thyroid cell lines has been reported (33), this required a 60 -72-h time course and more likely reflects divergent effects of cAMP on gene expression as compared with other short term biochemical events (34,35).
Other pH i -regulatory Mechanisms Did Not Contribute to the Effects of Drugs in Slowing MCT1 Kinetics-If MCT1-mediated cytoplasmic acidification was countered by Na ϩ /H ϩ exchanger-mediated proton efflux from RBE4 cells, then any drug-induced enhancement of the Na ϩ /H ϩ exchanger function might have given the false impression of slowed MCT1 functional activity. Indeed, studies of the Na ϩ /H ϩ exchanger NHE-1 showed that forskolin and cAMP can lead to increased Na ϩ /H ϩ exchanger activity (12). Conversely, these same agents reportedly inhibit the NHE-3 isoform (11,36). In the present studies, amiloride, a well characterized inhibitor of Na ϩ /H ϩ exchangers (28), did not significantly affect the lactate-dependent v i nor its slowing by forskolin or Bt 2 cAMP. Therefore, Na ϩ /H ϩ exchangers are unlikely to have contributed significantly to the observed drug effects (Fig. 7).
The presence of additional non-MCT1-mediated pH i regulatory mechanisms was also evaluated under lactate-independent conditions using the NH 4 Cl pre-pulse technique (Fig. 8) (29). These experiments revealed that the PKA inhibitor, H89, had a negligible effect on proton efflux from RBE4 cells following monocarboxylate-independent acid loading, whereas forskolin lowered the rates. Therefore, the effects of H89 and forskolin on lactate-independent pH i regulatory mechanisms could not have been the underlying reason for the reduced lactate-dependent v i observed with these agents (Figs. 5 and 8). This was further confirmed by the absence of an effect of forskolin or H89 on rates of lactate-dependent acidification in the presence of the MCT1 inhibitor CHC (Fig. 9). Therefore, the combined data (Figs. 7-9), and our use of bicarbonate-free buffer, point to a direct effect of the agents on MCT1 kinetic function, rather than acting through other secondary pH i regulatory proteins such as Na ϩ /H ϩ exchangers.
The Importance of Modulation of Cerebral Endothelial MCT1 Kinetic Function in the Normal and Diseased Brain-Unlike monosaccharides, the monocarboxylates are components of every major metabolic pathway in mammalian cells and serve as either a substrate or product of energy metabolism. Movement of short chain monocarboxylates is greatly restricted unless facilitated by a membrane carrier. Therefore, it is not surprising that nearly every mammalian cell possesses membrane transporters for monocarboxylates (37). Clearly, the basic function of MCT1 in the cerebral microvascular endothelium is to facilitate bloodbrain transport of monocarboxylic acids; however, the possible regulation of this transport function has not been examined previously at the molecular level in cells of the central nervous system. The dynamic control of MCT1 function in the cerebral microvascular endothelium would be very important in brain energy metabolism because it would regulate the transport of monocarboxylates into and out of the brain, controlling their availability as brain energy substrates. For example, activation of receptors on the cerebral vascular endothelial cells by elevation of stress-induced hormones may restrict MCT1 function. This would lead to increased brain retention of monocarboxylates that are normally exported and enhance energy production during times of brain stress or high metabolic demand by providing additional substrates for the mitochondrial citric acid cycle and oxidative phosphorylation. Therefore, it is hypothesized that the regulation of MCT1 kinetic function in the cerebral microvasculature is an important part of maintaining energy homeostasis for the healthy brain as it works to meet a dynamic energy demand.
During disease conditions such as stroke, MCT1 function in the cerebral microvascular endothelium may play an important role in determining the severity of tissue damage and extent of cell recovery. During ischemia or anoxia, lactic acid levels rise dramatically in the adult brain. Therefore, it is evident that the normal function of MCT1 to transport lactic acid down its concentration gradient is rate-limiting, and regulation of MCT1 activity has important implications and potential consequences. Pharmacological enhancement of MCT1 transport activity in the cerebral microvascular endothelium may be a therapy to reduce the severity of lactic acidosis, and its downstream and damaging effects in stroke patients. This hypothesis is supported by data showing that overexpression of MCT1 protein is neuroprotective under conditions of focal cerebral ischemia in rats. 3 Our understanding of the importance of cerebral microvascular MCT1 is only beginning to emerge but has strong potential to help resolve many important questions, including the role of monocarboxylates, and their transporters, in brain energy metabolism, stroke, and other pathological processes. By demonstrating for the first time the regulation of MCT1 kinetic function in rat brain endothelial cells, we have set a precedent for future development.