INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL METHODS RESULTS DISCUSSION REFERENCES

ANIMAL STUDIES HAVE SUGGESTED that blood-brain barrier (BBB) glucose transporter activity may be upregulated in seizures. This was inferred from observations that the glucose utilization rate determined in seizing rat brain exceeds the BBB glucose transporter maximal velocity (V_max) measured in vivo (24). Other analyses of glucose transporter activity (9, 11, 12) and glucose transporter density (16) in animals also suggest that modulations in BBB glucose transporter are induced by seizures.

The original work of Cremer et al. (11) demonstrated treatment-induced increases in the rates of both glucose phosphorylation and transport, providing a rationale for more detailed analyses of seizure-induced increases in brain glucose transport. These workers (11) saw modest increases in glucose transport rates within certain regions of the rat brain in response to treatment with the tremorgenic agent cismethrin, twofold increases in fasted, cismethrin-treated animals, and they suggested their results could be accounted for by an increase in transporter V_max. In a subsequent study where the convulsant drug kainic acid was administered 30 min before BBB permeability to deoxyglucose was measured, Cremer et al. (12) reported increases in some brain regions (hippocampus, entorhinal cortex, and amygdala) but not in others (visual, somatosensory, or frontal cortex). However, this analysis was somewhat inconclusive, because kinetic analyses of glucose transport into these regions could not be explained in the same terms as those used to express BBB transport in the normal rat (12). In an immunocytochemical study, Gronlund et al. (16) showed significant increases in the neuronal glucose transporter isoform GLUT-3 8-72 h after pentylene tetrazole (PTZ)-induced seizures. The brain capillary glucose transporter isoform GLUT-1 was not significantly increased, however, until 72 h posttreatment.

Thus the evidence for increased BBB glucose transporter activity comes from studies where effects were observed some hours after seizures. It is known that in clinical status epilepticus, seizures must be stopped within 60 min if severe, permanent brain damage is to be avoided. We hypothesized that glucose transporter activity may be acutely upregulated in situations where seizure activity is powerful enough to deplete brain metabolic reserves and brain glucose levels would fall, as shown in neonatal seizures (14). McCandless et al. (21) demonstrated that within minutes of the initiation of seizures by intravenous administration of PTZ, brain glucose levels in the tree shrew are reduced by 50-75% in all forebrain regions of this primate. The objective of the present study was therefore to determine through in vivo analyses whether or not acutely increased BBB glucose transporter activity could be demonstrated within minutes of seizure initiation in previously naive rats.

	EXPERIMENTAL METHODS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL METHODS RESULTS DISCUSSION REFERENCES

Radiochemicals. The [³H]glucose and D-[¹⁴C]glucose and [¹⁴C]sucrose were obtained from New England Nuclear Research Products (NEN-DuPont, Wilmington, DE). The [³H] and [¹⁴C]diazepam, [¹⁴C]inulin, and the tin-indium (In) TFC3 generator were obtained from Amersham (Arlington Heights, IL). ¹⁸F-labeled 2-fluoro-2-deoxy-D-glucose (FDG), synthesized by the method of Hamacher et al. (17), was obtained from the Positron Emission Tomography facility at the Greater Los Angeles Veterans Administration Healthcare System.

Single injection animal methods. The injection solutions were prepared in a mixture containing ~5.0 µCi of [¹⁸F]FDG (0.2 mCi/mg), the test compound, and 2.0 µCi of the reference isotope [³H]diazepam in a total volume of 200 µl of buffered saline. The buffer, HEPES, was obtained from Sigma Chemical and diluted to a final concentration of 10 mM (pH 7.55). A second mixture was prepared for intravenous administration; it contained 1.0 µCi of [¹⁴C]butanol in a total volume of 100 µl of buffered saline.

Internal carotid artery perfusions. Seizures can profoundly effect cerebral blood flow rate, and this could effect measurement of PS products. In addition, Cremer et al. (11) demonstrated in trembling rats that in all brain regions where increases in glucose transport were seen, the CBF rate had also significantly increased. Precise control of flow rate under the conditions of perfusion is achieved with in situ brain perfusion (1). Consequently, we used the in situ internal carotid artery perfusion technique of Takasato et al. (27) to complement carotid artery injection studies. The chemical composition, as well as the oxygen tension of the perfusate is precisely controlled, and the CBF (i.e., cerebral perfusion rate) was measured in individual animals in parallel with measurement of BBB PS products for glucose (27). We anesthetized the rats with intraperitoneal ketamine (200 mg/kg)-xylazine (2 mg/kg) and exposed the right common carotid artery. Under a binocular microscope, the occipital, the superior thyroid, and the pterygopalatine arteries were closed by electrocoagulation. The right external carotid artery was catheterized for retrograde perfusion using a PE-10 polyethylene catheter. The common carotid artery was tied with a suture, and the right jugular vein was opened at the time perfusion was started. The perfusate consisted of Krebs-Henseleit buffer (pH 7.4) composed of (in mM) 118 NaCl, 4.7 KCl, 2.5 CaCl₂, 1.2 MgSO₄, 1.2 KH₂PO₄, 25 NaHCO₃, with or without 0.1 g/dl bovine serum albumin and radioisotopes. We used 0.5 µCi/ml of [³H]diazepam, as a CBF marker, and 0.05 µCi/ml of D-[¹⁴C]glucose. [¹⁴C]inulin was used as a blood volume marker in some studies. The perfusate was filtered, bubbled with the desired oxygen concentration, usually O₂-CO₂ (95:5) and warmed to 37°C. The perfusate flow was 3-4 ml/min driven by a peristaltic pump, and other studies indicated that to achieve a 10-s cerebral vasculature perfusion, the pump should be operated for 13 s (26).

Data analyses. The K_m and V_max of the saturable component of transport were estimated as described by Pardridge (26) from nonlinear regression analyses of the clearance data. Under the conditions of this study, clearance (corrected for vascular space) is a good approximation for PS in milliliters per minute per gram. PS is a function of the kinetic constants
and C is the injected (arterial) glucose concentration. It has been suggested that the diffusion component (K_d) of nonsaturable transport may be an artifact of carotid injection (13), and K_d has not been distinguishable from zero in recent analyses (5, 7). Consequently, no attempts were made to quantitatively estimate K_d in the present investigation.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Effect of pentylene tetrazole (PTZ) on 18F-labeled 2-fluoro-2-deoxy-D-glucose (FDG) and flow in conscious, cannulated rats. A: brain extraction of FDG is significantly higher (P < 0.05) at 3 min after PTZ injections. B: simultaneous measurements of cerebral blood flow (CBF) using the external organ method indicate that CBF is also significantly increased by PTZ-induced seizures. C: the permeability times surface area (PS) product of a tracer concentration of FDG is also significantly increased as a result of PTZ-induced seizures. Because PS-to-CBF ratios are also increased in comparing control (0.25) and PTZ treatments (0.32), it seems unlikely that any significant capillary recruitment occurs. These data suggest that PTZ treatment causes increased glucose transporter activity. Vertical bars represent SD. * P < 0.05 vs. respective control.

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Comparison of glucose clearance in frontal (A), parietal (B), and occipital (C) cortex of PTZ-treated and control rats anesthetized with ketamine-xylazine. Note that with the internal carotid arterial perfusion method, the effect seen with single injection (Fig. 1) is confirmed. Over the range of glucose concentrations examined, higher glucose clearances were consistently observed in PTZ-induced seizing rats. Vertical bars represent SD.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Comparison of glucose clearance in hippocampus (A), caudate putamen (B), and midbrain (C) of PTZ-treated and control rats. Note that as in cortical regions (Fig. 2), higher glucose clearances were consistently observed during seizures as opposed to control rats. Vertical bars represent SD.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Inulin distribution (determined after internal carotid artery perfusion) is not significantly different in PTZ-treated vs. control rats for all of the brain regions examined. This suggests that no significant increases in vascularity (or capillary recruitment) are produced by PTZ-induced seizures. Vertical bars represent SD.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL METHODS RESULTS DISCUSSION REFERENCES

Experimental testing was performed in rats that had a cannula implanted in the external carotid artery. Seizures were initiated by the administration of intracarotid PTZ, and brain uptake of [¹⁸F]FDG was measured in the ipsilateral hemisphere using the Oldendorf brain uptake index method. Simultaneously, the rate of CBF was measured using the external organ technique to define both parameters in PTZ-treated and control animals.

As indicated in Fig. 1A, the extraction of FDG was significantly greater in the PTZ-treated forebrain than in sham-injected control animals. CBF rates were also significantly increased in response to convulsant treatment (Fig. 1B). The PS of FDG was also significantly increased (Fig. 1C). Increased PS could be a function of either increased glucose permeability (transport) or increased capillary recruitment, whereby a greater number of central nervous system capillaries is perfused as a result of the seizure. However, if the increases in PS were solely a function of capillary perfusion, then the PS-to-CBF ratio in PTZ-treated (0.32) and control (0.25) situations would remain constant. These differences suggest that PS-to-CBF ratios for FDG are increased in response to PTZ treatment, consistent with data obtained when glucose PS is determined (see Table 1).

Confirmation of this effect was sought using the internal carotid artery perfusion technique of Takasato et al. (27), wherein the external carotid artery is cannulated, the common carotid artery is ligated, and the ipsilateral hemisphere is perfused with an isotopic buffered saline at a constant perfusion rate. From these studies D-[¹⁴C]glucose influx was measured over a range of substrate concentrations to define transporter V_max and K_m in treatment (PTZ seizures) and control states. Because perfusion flow rate is externally controlled, glucose permeability was measured in the absence of CBF changes. The clearance of glucose was relatively greater in brain regions of PTZ-treated rats than control rats, and the effect was demonstrated over a wide range (0.5-75 mM) of physiological glucose concentrations. Minor regional variations are seen in frontal (Fig. 2A) parietal (Fig. 2B), and occipital cortex (Fig. 2C), as well as in the hippocampus (Fig. 3A), caudate-putamen (Fig. 3B), and midbrain (Fig. 3C). Inulin distribution was similarly compared in the same brain regions of PTZ-treated and control rats and found to be unchanged as a result of convulsant treatment (Fig. 4). The fact that inulin distribution was unchanged in PTZ-treated and control rats suggests that capillary surface areas were also unchanged in seizures and also that the BBB was not compromised by capillary damage during the short-term experimental seizure period. This observation further implies that the increases in PS seen with PTZ treatment (Tables 1 and 2) are attributable to increased BBB glucose permeability.

In each of the brain regions dissected for kinetic analysis, the affinity of the transporter for glucose appears to be unchanged as a result of PTZ treatment. Because the coefficient of variation of the regression estimate K_m averaged 27%, a considerable portion of the variation in K_m across regions may just be due to estimation imprecision. K_m was ~10 mM in both control and PTZ-treated brain over the six regions examined. There may be greater variability of K_m between different brain regions than between PTZ-treated and control groups. K_m estimates ranged from 5.7 to 12.6 mM in controls, compared with 8.7-11.9 mM in PTZ treatment (Table 1). However, in PTZ-treated animals the transporter V_max is consistently greater than in controls. The V_max was increased by 45% in midbrain, 10% in caudate putamen, 58% in hippocampus, 79% in occipital cortex, 38% in parietal cortex, and 54% in the frontal cortex (Table 1). Further analyses of the data were consequently performed, assuming that the K_m for glucose transport was unchanged in PTZ-treated and control situations. Also, because PTZ produces generalized (as opposed to focal) seizures, data were pooled into cortical (frontal cortex, parietal cortex, and occipital cortex) and subcortical (hippocampus, caudate putamen, and midbrain) regions.

Kinetic constants derived for the cortical brain regions indicated that the K_m was estimated to be 9.5 ± 1.6 mM. Transporter V_max determined 3 min after initiation of seizures by the intracarotid administration of PTZ (2.46 ± 0.34 µmol · min¹ · g¹) was significantly (42 ± 12%) greater than glucose transporter V_max in controls (1.74 ± 0.26 µmol · min¹ · g¹) (P < 0.05). A similar effect was observed in the subcortical forebrain. The K_m was determined to be 9.4 ± 1.5 mM. Glucose transporter V_max determined 3 min after initiation of seizures by PTZ administration (2.35 ± 0.30 µmol · min¹ · g¹) was also significantly greater (31 ± 10%) than the transporter V_max in controls (1.80 ± 0.25 µmol · min¹ · g¹) (P < 0.05).

Similar values for these parameters were estimated when the regression analysis was performed without the assumption that the K_m is unchanged in control PTZ-treated rats.

In the cortex, the estimated K_m for controls was 8.58 ± 2.63 mM, and K_m for PTZ-treated rats was 9.98 ± 2.13 mM. The consistency of these estimates with the predicted single K_m = 9.53 ± 1.62 mM for treatment and control groups suggests this assumption is consistent with the data. In the subcortex, the estimated K_m for controls was 8.90 ± 2.43 mM and K_m for PTZ-treated rats was 9.62 ± 1.94 mM. These estimates for this brain region also compare favorably with the predicted single K_m = 9.36 ± 1.48 mM for treatment and control groups. Estimates of other parameters describing transporter kinetics were also relatively unchanged as a result of assuming K_m to be the same in treatment and control groups, as indicated in Table 2.

Our data also demonstrate that BBB glucose permeability was significantly increased in PTZ-treated rats (Table 2). This is indicated by the demonstration that the ratio of PTZ to control unsaturated PS (PS = V_max/K_m) was >1.0, and this difference was significant at the 5% level in both cortical and subcortical brain regions.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL METHODS RESULTS DISCUSSION REFERENCES

The present study demonstrates that in previously naive animals, a single convulsant-induced seizure causes a significant increase in BBB glucose transporter activity. There is a general expectation that seizures should stimulate unidirectional BBB glucose transfer (9, 12, 16, 18, 24), but this has yet to be demonstrated using kinetic characterization. Such estimates are considered to be a reliable and conservative indicator of altered BBB glucose transport (11). To our knowledge, the present study represents the first in vivo demonstration of increased BBB transporter V_max in response to acute seizures. An important aspect of this observation is the remarkably acute nature of upregulated brain glucose transport: within 3 min after seizures are initiated, a 31 ± 9% (subcortical forebrain) to 42 ± 12% (cortical forebrain) increase in V_max occurs. This rapid response is an indicator of just how urgently the BBB GLUT-1 transporter recognizes the brain's sudden requirement for additional glucose during seizures.

Capillary recruitment (an increase in size and/or number of capillaries being perfused) is a well-recognized phenomenon in tissues such as muscle but seems unlikely from our single injection studies of experimental seizures (Fig. 1). The additional observation that inulin distribution volumes were unchanged in all brain regions examined as a result of PTZ treatment (Fig. 4) is important because it demonstrates that the volume (and surface area) of capillaries undergoing perfusion does not significantly change during the experimental seizure period. Thus increases in PS seen are neither due to capillary recruitment nor increased endothelial surface area. Therefore, glucose permeability is increased. The glucose transporter V_max estimations reflect the number (density) and intrinsic activity of capillary GLUT-1 transporters. Changes in estimated V_max may be attributable to more or fewer transporter proteins on the membranes or changes in the intrinsic rate at which glucose molecules are shuttled (or translocated) across brain capillary membranes. Electron microscopic immunogold methods have been employed to provide quantitative estimates of membrane transporter density (8). However, the presence of both high- and low-GLUT-1-expressing capillaries seen in (interictally resected) human seizure tissue (6) does not provide information about rapid ictally induced changes in GLUT-1 expression. The fact that transporter V_max is increased so rapidly (by 30-40%) in response to seizures possibly indicates that increased intrinsic activity of the GLUT-1 occurs; i.e., an increased shuttling rate of glucose by transporter proteins is induced. Regulation of glucose transport into cells is generally thought to occur through transporter expression at the cell surface, and the extent to which intrinsic properties of glucose transporters are regulated is at present controversial (22). However, Kuroda et al. (19) observed that insulin-induced alterations in adipocyte glucose transport occurred in the absence of concomitantly measurable changes in the quantity of membrane transporter proteins, and they attributed these results to altered intrinsic activity of the transporter protein. Furthermore, interleukin-3 (IL-3) treatment causes a rapid increase of glucose transport into cultured bone marrow cells, without alterations in membrane expression of GLUT-1 or GLUT-3 transporter proteins. Consequently, McCoy et al. (22) concluded that IL-3 produced increases in the intrinsic shuttling activity of the glucose transporter proteins. Thus in vitro studies support the proposal that increased intrinsic activity of the BBB GLUT-1 transporter might explain the rapid increases in glucose transport brought on by seizures.

The alternate possibility requires that increased numbers of transporter proteins are recruited to the membranes. Because <25% of the total brain endothelial cell GLUT-1 is found in the cytoplasm (4, 6), a 30-40% increase in membrane transport would require rapid generation (<3 min) of membrane GLUT-1 transporter. This recycling of cytoplasmic GLUT-1 to the capillary membranes cannot be discounted, however, because 35-45% of the endothelial GLUT-1 may be located in the endothelial cytoplasm of rats (15) and rabbits (8). Also, the possibility exists that brain endothelial cells may be able to recruit transporter proteins from the abluminal to luminal membranes to meet the metabolic demands inherent in seizure activity. Thus it appears that further quantitative GLUT-1-immunogold analyses of membrane transporter concentrations will be required to address this issue.

The kinetic constants for glucose transport derived in the present study (Table 2) appear to be consistent with previously published data. In vivo analyses of BBB glucose transporter V_max have resulted in estimates of 1.42 ± 0.14 µmol · min¹ · g¹ in normal rats (24) and 1.6 µmol · min¹ · g¹ in the gerbil forebrain (2). Estimates of the glucose transporter K_m are on the order of 9.6 ± 0.9 to 11.0 ± 1.4 mM in normal rats (24), and we estimated similar K_m (Table 2).

In addition to the data presented above (see RESULTS), three lines of evidence support our assumption that modulations in BBB glucose transporter activity occur without a change in the K_m. First, other researchers have typically elected to assume that the only (or primary change) was in transport velocity, on the basis that this assumption was the most conservative (11, 12). Second, studies of glucose influx in the developing rabbit have demonstrated that significant increases in BBB glucose transporter V_max occur in comparisons of neonatal, suckling, weanling, and adult animals. In contrast, developmental variations in the K_m were not significant (7). Third, PTZ-induced increases in the FDG PS (Fig. 1) as well as the glucose PS (Table 1) suggest that either a relative increase in V_max has occurred or a relative reduction in K_m. The fact that the estimated values for K_m were essentially unchanged or slightly increased for five of the six regions examined (Table 1) is not consistent with a reduction in K_m. These data thus support the assumption that K_m be considered unchanged (Table 2) as a result of PTZ treatment.

Present data suggest that in response to a single seizure (ictal), BBB glucose transporter activity is acutely upregulated by 30-40% within a span of 3 min. In contrast, studies of the epileptogenic focus of patients with chronic partial seizures of temporal and frontal lobe origin show that tracer FDG influx is interictally reduced in the seizure focus (3). Electron microscopic analyses of brain capillary glucose transporters from such patients exhibit both high- and low-GLUT-1 expression, perhaps suggesting that the bimodal GLUT-1 configurations permit either up- or downregulation in tissues characterized by repeated seizures (6). Resolution of these two possibly conflicting observations, however, appears possible if physiological accommodation occurs in chronic seizures. Perhaps as a result of the repeated upregulation during ictal events (seen in the present study), there is a subsequent accommodation of BBB glucose transporter activity in the seizure focus. This accommodation results in a lowering of the basal (resting state) activity of glucose transporter protein densities in a subpopulation (but not all) of the BBB endothelial cells in human epilepsy (6). How this effect is controlled is not understood. It is not known if the endothelial cells that exhibit low glucose transporter (6) are able to upregulate, for example, in response to a seizure, but it is now known that functional transporter activity does indeed increase. But it is not known whether this seizure-induced induction of glucose transporter is persistent or reversible. With the recognition that BBB glucose transporter activity is so dynamically altered due to seizures, a more complete understanding of these events and the responsible mechanisms will be a focus of future research.