Altered Expression of Brain Monocarboxylate Transporter 1 in Models of Temporal Lobe Epilepsy

Chemicals and animals

Chemicals were obtained from Sigma Chemical Co. (St. Louis, Mo.) unless otherwise noted. Male Sprague Dawley rats were used in this study (methionine sulfoximine [MSO] model, 200 to 250 g; perforant pathway stimulation models, 350 to 450 g; Charles River Laboratories, Wilmington, Mass.). Food and water were provided ad libitum. The animal care and use procedures were approved by the Institutional Animal Care and Use Committee of Yale University. All studies were performed in accordance with current guidelines.

MSO model

Chronic infusion of the glutamine synthetase inhibitor MSO unilaterally into the hippocampus in rats leads to recurrent seizures and neuropathological changes similar to those observed in human TLE. The procedure for creating the MSO model is described in detail by (Eid et al., 2008; Wang et al., 2009). Under Isoflurane anesthesia, a drug delivery cannula was introduced into the right hippocampus and connected to a subcutaneously implanted Alzet mini-osmotic pump (Model 2004, Durect Corp., Cupertino, Calif.). In one set of animals, the pump was filled with MSO to achieve a delivery of 0.625 g of MSO per h for ~28 days (as per manufacturer’s specifications). In another set, the pump was filled with 0.9% NaCl (saline). The infusion of MSO or saline started immediately postsurgery. Two unipolar depth electrodes (E363/2/SPC stainless steel electrode, Plastics One, Roanoke, Va.) with a bare diameter of 0.2 mm and an insulated diameter of 0.23 mm were introduced into the left and right dorsal hippocampal formations to record bilateral intrahippocampal EEG activity. Some animals had two stainless steel screw electrodes (Plastics One) implanted instead of depth electrodes to record EEG activity. This was done to minimize detachment of the electrodes from the animals’ skull (see (Wang et al., 2009) for further explanation). The electrodes were positioned in the epidural space overlying the parietal neocortex of each hemisphere. A third electrode was positioned in the epidural space near lambda to serve as the reference. A fourth electrode was positioned over the occipital bone (not touching the dura) to serve as the ground.

Perforant pathway stimulation models

Under Isoflurane anesthesia, rats were implanted bilaterally with two stimulating electrodes (NEX-200 × 3.5 mm, Rhodes Medical Instruments, Summerland, Calif.) and two custom made tungsten recording electrodes with a bare diameter of 0.076 mm and an insulated diameter of 0.11 mm (A-M Systems, Carlsborg, Wash.). The electrodes were placed in the angular bundle of the perforant pathway and in the dentate granule cell layer, respectively, as previously described (Harvey and Sloviter, 2005). The optimal locations of the stimulating electrodes were determined by maximizing the evoked responses to perforant pathway stimulation (Andersen et al., 1966). Two screws serving as ground and reference electrodes were positioned anterior to the bregma, one over each hemisphere. The female socket contacts on the ends of the electrodes were inserted into two plastic pedestals (Plastics One), and the entire setup was secured by UV curing acrylic adhesive (Loctite 3106 light cure adhesive, Henkel Loctite Corp., Rocky Hill, Conn.).

At least two weeks postsurgery, awake animals were stimulated bilaterally for 3 or 8 h. In the 3-h paradigm, 15- to 20-V paired-pulse stimuli were delivered at 2 Hz, with a 40-ms interpulse interval. These stimuli were coupled with 10-s-long 20-Hz stimulus trains of single 15- to 20-V stimuli delivered every minute. In the 8-h paradigm, 10-s-long 20-Hz stimulus trains of single 30-V stimuli with 0.1-ms duration were delivered once per minute. These paradigms are designed to evoke hippocampal granule and pyramidal cell discharges (Sloviter et al., 1996). Isoflurane gas was administered to the rats at the end of the 3- or 8-h stimulation paradigms to terminate residual seizure activity. Rats appeared to be fully recovered behaviorally and resumed normal eating and drinking habits approximately one day after stimulation.

Video-intracranial EEG monitoring

The experimental setup for recording video-EEG was adapted from the work of Bertram et al. (Bertram et al., 1997). The rats were housed individually in custom-made Plexiglas cages, and video-intracranial EEG monitoring was initiated immediately after stimulation or after the onset of MSO or saline infusion. A spring-covered, 6-channel cable connected the electrode pedestal to a commutator (Plastics One), while a second cable connected the commutator to the digital EEG recording unit (Ceegraph VISION LTM, Natus Medical, Mundelein, Ill.). Digital cameras with infrared capacity were used to record animal behavior. The digital video signal was encoded and synchronized to the digital EEG signals. Seizures were identified by visual inspection of the EEG record. Interictal activity is essentially transient in nature and can be distinguished by its duration; subclinical seizures can be distinguished by examination of the video record. To establish the start and stop points of a seizure, we determined a point that was unequivocally within the seizure and moved backward and forward in time to the points where the EEG was characterized by normal baseline activity. The video record was examined to score the seizures on a modified Racine scale (Racine et al., 1973) as follows: Subclinical, no remarkable behavior; Stage I, immobilization, eye blinking, twitching of vibrissae and mouth movements; Stage II, head nodding, often accompanied by facial clonus; Stage III, forelimb clonus; Stage IV, rearing; Stage V, rearing, falling and generalized convulsions. Recurrent seizures were defined as ≥2 seizures of Stage III or higher at least 1 h apart.

Tissue preparation

After continuous video-EEG recordings for several days, the rats were reanesthetized and perfused transcardially either with saline followed by 4% paraformaldehyde (PFA)-0.1% glutaraldehyde in phosphate buffer (PB; 0.1 M, pH 7.4) (perforant pathway stimulation groups) or with saline followed by 4% PFA in PB (MSO group). In the MSO group, the brains were subsequently removed, postfixed in the same solution overnight at 4°C and stored at −80°C. In the perforant pathway stimulation groups, the brains were left in situ for 4 h at 4°C before being removed from the skulls. A Vibratome and sliding microtome were used to cut 80-μm- and 50-μm-thick brain sections from the MSO and perforant pathway stimulation models, respectively. Every fifth section was transferred to a gelatin-coated slide and stained with cresyl violet. Subsets of the remaining sections were used for immunohistochemistry.

Immunohistochemistry

For single-labeling immunohistochemistry, sections from the MSO model and perforant pathway stimulation models were incubated free-floating in solutions of chicken anti-MCT1 antibody (AB1286, Millipore, Billerica, Mass.) diluted 1:10,000 (MSO model) or 1:1,000 (perforant pathway stimulation models), or mouse anti-RECA-1 antibody (MCA970R, AbD Serotec, Dusseldorf, Germany) diluted 1:1,000. All antibodies were incubated with tissue sections for 72 h at 4°C and processed according to the avidin biotin peroxidase method (Hsu et al., 1981), using the Vectastain Elite Kit (Vector Laboratories, Burlingame, Calif.) with diaminobenzidine as the chromogen. Sections labeled with RECA-1 were counterstained with cresyl violet. The immunostained sections were mounted on gelatin-coated glass slides, cover-slipped and examined by light microscopy. For double-labeling fluorescence immunohistochemistry, Vibratome sections were incubated at room temperature for 24 h (20 to 24°C), free-floating in a mixture of chicken anti-MCT1 (1:1,000) and either mouse anti-RECA-1 (1:2,500), mouse anti-GFAP (MAB360, Millipore; 1:2,000), rabbit anti-AQP4 (AB3594, Millipore; 1:2,000), rabbit anti-synaptophysin 1 (101002, Synaptic Systems, Goettingen, Germany; 1:500), rabbit anti-Vglut1 (135302, Synaptic Systems; 1:10,000), mouse anti-PSD95 (MAB1598, Millipore; 1:1,000) or mouse anti-MAP 2A/2B (MAB378, Millipore; 1:1,500). After washing, sections were incubated for 2 h in a mixture of Alexa Fluor 488 goat anti-chicken IgG and either Alexa Fluor 555 goat anti-mouse IgG or Alexa Fluor 633 goat anti-rabbit IgG. All fluorescent secondary antibodies were diluted 1:1,000. Double-labeled sections were examined by a confocal laser-scanning microscope (Zeiss LSM META, Carl Zeiss MicroImaging, Thornwood, N.Y.).

To quantify the density of microvessels labeled with MCT1 or RECA-1 in the hippocampal formation, stereological point count analysis was carried out (Gundersen, 1988). Hippocampal formations from MSO-infused (MSO R, n=6 and n=5), contralateral untreated hippocampal formations of MSO-infused rats (MSO L, n=6 and n= 6), saline infused (n=6 and n=6), sham-operated (n=5 and n=4), 3h-stimulated (n=9, n=9) and 8h-stimulated (n=4 and n=4) rats were included in the analysis of MCT1 and RECA-1 respectively. The densities of MCT1 or RECA-1 positive microvessels in each hippocampal formation, i.e. an area comprising of the dentate gyrus, the dentate hilus, CA3, CA2, CA1 and subiculum, were calculated by the following procedure. (1) One image of the dorsal part of the hippocampal formation was taken from each rat using a light microscope. (2) A test system with a set of regularly spaced points was superimposed on to each image. (3) Points hitting MCT1-, or RECA-1 positive structures within the hippocampal formation were counted and divided by the total number of points within the hippocampal profile. The investigator performing the counts was blinded to the experimental identity of the tissue. The density of blood vessels was expressed as number of counts hitting the blood vessels per 100 total counts.

Statistical Analysis

Analysis of repeated measurements was performed using a one way ANOVA with a Tukey-Kramer multiple comparisons post test. The significance level was set to a P value of <0.05. Unless otherwise noted, the data are presented as mean ± standard deviation.

Nissl stain

Using Nissl-stained tissue sections, we first assessed the general histopathology of the hippocampal formations from each of the five groups of rats included in this study (Fig. 1). Animals continuously infused with saline into the right hippocampal formation (MSO control group; n = 6) displayed either a narrow zone of neuronal loss and gliosis around the saline infusion cannula or no significant neuronal loss or gliosis in the hippocampal formation (Fig. 1A1-4). Five of the six animals continuously infused with MSO into the right hippocampal formation (n = 6) did not exhibit significant neuronal loss or glial proliferation in the left contralateral hippocampal formation (Fig. 1B1-4), whereas all MSO-treated animals exhibited a range of morphological changes in the right hippocampal formation, from minimal to significant neuronal loss and including a sclerosis-like pathology (Fig. 1C1-4), as previously reported (Eid et al., 2008; Wang et al., 2009). One MSO-treated rat had glial proliferation and moderate loss of pyramidal neurons in CA3 of the contralateral (left) untreated hippocampus (not shown). Sham-operated, nonstimulated rats (perforant pathway stimulation control group; n = 5) did not exhibit significant neuronal loss or proliferation of glial cells in the hippocampal formation (Fig. 1D1-4). Rats subject to perforant pathway stimulation for 3 h (n = 10, Fig. 1E1-4) or 8 h (n = 5, Fig. 1F1-4) exhibited a range of neuropathological changes. The changes in the 3-h stimulation group were distributed as follows: (a) no visual neuronal loss or gliosis (3 of 10 rats), (b) neuronal loss and gliosis mainly in the hilus (1 of 10 rats), (c) neuronal loss and gliosis in the hilus and CA3 (2 of 10 rats) and (d) hippocampal sclerosis (4 of 10 rats) (Fig. 1E1-4). Eight hours of perforant pathway stimulation resulted in a hippocampal sclerosis-like picture in 4 of 5 rats (Fig 1F1-4) and 1 rat with extensive neuronal loss throughout the hippocampal formation. Two 8-h-stimulated rats with hippocampal sclerosis also experienced a moderate loss of the dentate granule cells (data not shown).

Continuous video-intracranial EEG monitoring

All rats were monitored by video-intracranial EEG for continuous periods up to a maximum of 73 days. None of the saline-infused (n = 6) or sham-operated, nonstimulated (n = 5) rats exhibited recurrent seizures (Fig. 2A). In contrast, all of the MSO-treated (n = 6, Fig. 2B) and 8-h-stimulated (n = 5, Fig. 2D) rats developed recurrent seizures. All but one 3-h stimulated rat (n = 10) had recurrent seizures (Fig. 2C). This rat’s electrode implantation fell off on the third day of recording. Had it stayed on for a longer period, we probably would have captured recurrent seizures in this animal as well.

MCT1 immunohistochemistry

A widespread network of MCT1-positive blood vessels was present in the hippocampal formation in the saline-infused rats (Fig. 3A1-4), nonstimulated controls (Fig. 3D1-4) and left (noninfused) hippocampal formation in the MSO-treated animals (Fig. 3B1-4). In contrast, markedly fewer MCT1-immunoreactive blood vessels were present in the right hippocampal formation in the MSO-infused rats (Fig. 3C1-4) and in both hippocampal formations in the two stimulation models (Fig. 3E, F). Specifically, MCT1 was lost on microvessels in the dentate hilus in all of the epileptic rats, independent of the model studied (Fig. 3C₂, E₂, F₂). MCT1 was also lost on microvessels in CA3 and CA1 in MSO-treated (Fig. 3C_{3, 4};) and electrically stimulated rats (Fig. 3E_{3, 4} and 3F_{3, 4}). A large number of the epileptic rats, but none of the control animals, also revealed areas of increased MCT1 immunoreactivity in the hippocampal formation. The increased immunoreactivity was distributed diffusely in a granular pattern in the hilus, CA3 and CA1 in 85% of the MSO-treated animals (Fig. 3C), 70% of the 3-h-stimulated animals (Fig. 3E) and all of the 8-h-stimulated animals (Fig. 3F).

A comprehensive quantitative analysis revealed no significant difference in the density of MCT1-positive blood vessels between the right saline-treated hippocampal formation and the left (untreated) hippocampal formation in the MSO-infused rats (Fig. 4A). In contrast, the density of MCT1-positive microvessels was reduced by approximately 70 % in the right MSO-infused hippocampal formation vs. the left (untreated) hippocampal formation in the same animals and vs. the hippocampal formation in the saline treated controls (Fig. 4A). The density of MCT1-positive microvessels was reduced by approximately 80% in the hippocampal formation in the 3h and 8h stimulated rats vs. the non-stimulated control rats (Fig. 4B).

RECA-1 immunohistochemistry

We next analyzed the density of blood vessels in the hippocampal formations in all animals using the endothelial cell marker RECA-1. This was done to assess whether the loss of MCT1 on microvessels in TLE is merely due to a reduced vascular density in this disease. Visual evaluations of the RECA-1 positive blood vessels revealed dense networks of blood vessels in all subfields of the hippocampal formation in all animal groups (Fig. 5A-F). Quantitative analysis revealed no difference in blood vessel densities among hippocampal formations in the saline/MSO groups (Fig. 5G, P = 0.147) and in the sham/stimulated groups (Fig. 5H, P = 0.961). As expected, the ratio of MCT1 to RECA-1 labeling tended to be lower for MSO R, 3h and 8h-stim compared to in saline, MSO L and Sham-control (Fig. 5I).

Double-labeling confocal microscopy

To further assess the cellular localization of MCT1 in the hippocampal formations of the epileptic animals, double-labeling experiments were performed using MCT1 and one of the following markers for, astrocytes (GFAP, Fig. 6A; AQP4, Fig. 6B) or neurons (MAP 2A/2B, Vglut1, synaptophysin 1 or PSD95, Fig. 6C, D). Because the different epilepsy models showed similar patterns of MCT1 labeling, the study was limited to analysis of 8-h-stimulated rats and sham-operated controls. In the hippocampal formations in non-stimulated control rats, MCT1 was predominantly found on circular structures (Fig. 6A₁, B₁, C₁, D₁) which co-localized with the endothelial cell marker RECA-1 (not shown), confirming that MCT1 is mainly localized to endothelial cells under normal conditions. Only minimal labeling for MCT1 was present on non-endothelial cells in non-stimuated control rats (Fig. 6A-D). While MCT1 did not co-localize with GFAP in hippocampal formations in control rats (Fig. 6A_1-3), such co-localization was strikingly present in epileptic animals (Fig. 6A_4-6). The overall density of GFAP labeling was also greatly increased in neuron-depleted (sclerotic) areas of the hippocampal formation in epileptic animals (Fig. 6A_5,6) vs. the same areas in controls (Fig. 6A_2,3). In the epileptic animals, GFAP not only labeled discrete bundles of astrocyte filaments but also often labeled the entire cytoplasm of cells. Such diffuse cytoplasmic labeling with GFAP was not seen in control animals.

Although MCT1 was located near AQP4 in control animals, the two molecules did not completely overlap (Fig. 6B_1-3). The lack of overlap can be explained by the fact that, under normal conditions, MCT1 preferentially labels endothelial cells, whereas AQP4 primarily labels perivascular astrocyte endfeet (Eid et al., 2005; Frigeri et al., 1995; Lauritzen et al., 2011; Lee et al., 2004a; Nielsen et al., 1997). However, in the hippocampal formation in epileptic rats, increased labeling for AQP4 was present in the extravascular neuropil relative to the perivascular compartment (Fig. 6B_5,6). Such increase in extravascular AQP4 is also present in human TLE (Eid et al., 2005; Lee et al., 2004b). In the hippocampal formation in epileptic rats, extravascular MCT1 showed pronounced overlap with AQP4 (Fig. 6B_4-6), whereas such overlap was not seen in control rats (Fig. 6B_1-3). This observation indicates that MCT1 is strongly expressed on the plasma membrane of astrocytes in the extravascular neuropil in the epileptogenic hippocampal formation.

Finally, an assessment of MCT1 and a variety of neuronal markers did not reveal any colocalization between MCT1 and MAP 2A/2B (Fig. 6C_1,2), Vglut1 (Fig. 6D_1,2), synaptophysin 1 (not shown) or PSD95 (not shown) in the hippocampal formation in epileptic or control rats.

Loss of MCT1 on brain microvessels

A deficiency in MCT1 on microvessels in the hippocampal formation in TLE was first reported in patients undergoing resection of the anteromedial temporal lobe for the treatment of medication-refractory TLE (Lauritzen et al., 2011). In this patient population, MCT1 was severely deficient on the endothelial cell membrane of microvessels in several areas of the hippocampal formation, especially CA1, but also CA3 and the hilus of the dentate gyrus. The loss of microvessel-associated MCT1 was most prominent in TLE patients with concomitant hippocampal sclerosis, although a smaller reduction in MCT1 was also evident in TLE patients without hippocampal sclerosis compared with that in autopsy specimens from nonepileptic control subjects. The observation that MCT1 is lost on brain microvessels in human TLE as well as in three different animal models of the disease indicates that a lack of MCT1 on microvessels is of critical importance for the pathophysiology of the disease.

MCT1 is a key carrier molecule for monocarboxylates, amino acids and acidic drugs across the blood-brain barrier, and a deficiency in MCT1 is expected to impair the uptake of such compounds by the brain. The potential significance of such impaired uptake has been discussed in detail in our previous paper on the loss of MCT1 in human TLE (Lauritzen et al., 2011); thus, only a brief summary of the most pertinent points of this discussion will be provided here.

Firstly, the loss of MCT1 may result in decreased concentrations of blood-derived monocarboxylate fuels, especially ketone bodies, in the epileptogenic hippocampus. The presence of ketone bodies in the brain has been proposed to increase the brain energy stores (Pan et al., 1999), stabilize the neuronal membrane potential (Bough, 2008), enhance GABA-mediated inhibition (Cantello et al., 2007; Yudkoff et al., 2004), inhibit the transport of glutamate into synaptic vesicles (Juge et al., 2010) and attenuate the release of vesicular glutamate by neurons (Juge et al., 2010). Low basal levels of brain ketone bodies may therefore contribute to the seizure susceptibility of the hippocampal formation in TLE. Furthermore, treatment with a high-fat, low-carbohydrate – i.e., ketogenic – diet may reverse the deleterious effects of the MCT1 deficiency by enhancing the uptake of ketone bodies via existing MCT1 due to an increased concentration gradient between the blood and the brain, or by induction of the synthesis of additional MCT1 on brain microvessels, as demonstrated by Leino and colleagues (Leino et al., 2001).

Secondly, because the transport of acidic drugs such as valproic acid across the blood-brain barrier is facilitated by MCT1 (Fischer et al., 2008; Kang et al., 1990; Utoguchi and Audus, 2000), the tissue concentration of key antiepileptic drugs may be remarkably low in the MCT1-deficient, epileptogenic hippocampus. Some patients with TLE are resistant to valproic acid and other antiepileptic drugs, and the lack of MCT1 represents an attractive possible explanation for such pharmacoresistance.

Thirdly, the loss of MCT1 on brain microvessels in TLE may be accompanied by a concomitant upregulation of multidrug-resistance-associated proteins (MDPs) on the same blood vessels (Aronica et al., 2004; Kubota et al., 2006). MDPs facilitate the efflux of antiepileptic drugs from brain tissue, and a combination of decreased MCT1 and increased MDPs is expected to effectively prevent accumulation of antiepileptic drugs to therapeutic levels in the epileptogenic hippocampus (Cascorbi, 2010; Loscher, 2007; Marchi et al., 2010).

Overexpression of MCT1 on astrocytes

The second major finding, that MCT1 is overexpressed on astrocytes in the hippocampal neuropil in the three models of TLE, is also likely to be true for the epileptogenic hippocampal formation in human TLE. We previously reported a pattern of diffuse labeling with MCT1 in the neuropil of the hippocampal formation in patients with TLE and concomitant hippocampal sclerosis (Lauritzen et al., 2011). Such labeling was most prominent in areas of neuronal loss and astroglial proliferation such as the CA1. Although double-labeling experiments with MCT1 and astroglial markers were not performed in this study, the labeling pattern is strikingly similar to what we have now identified as an overexpression of MCT1 on astrocytes. A key question is whether the overexpression of MCT1 on astrocytes indicates increased efflux or influx of monocarboxylates in these cells.

Several arguments support the hypothesis that the overexpression of MCT1 causes enhanced efflux of monocarboxylates from astrocytes. Due to the increased release of extracellular glutamate during seizures (During and Spencer, 1993), the clearance of the neurotransmitter by astrocytes is likely to be enhanced. Because the cellular uptake of glutamate by glutamate transporters is energy-dependent (reviewed by (Danbolt, 2001), more ATP is required during conditions of excessive glutamate uptake. A highly active glycolytic pathway, which is an important source of ATP in astrocytes (Bergersen, 2007; Hertz and Dienel, 2005; Pellerin and Magistretti, 2003), will result in large amounts of lactate and protons, which must be removed from the cytosol to continue the synthesis of ATP. An increased expression of MCT1 on astrocytes could therefore boost the efflux of lactate and protons from these cells. The lactate efflux hypothesis presented here is also supported by the observation in human TLE that the interictal (basal) extracellular concentration of lactate is approximately 48% higher in the epileptogenic hippocampal formation (6.8 mM) than in the nonepileptogenic hippocampal formation (4.6 mM) (Cavus et al., 2005) and 90% higher during, and in the first hour (60 to 90 min) after, spontaneous complex partial seizures with secondary generalization (During et al., 1994). The biological effects of high extracellular brain lactate are not clearly understood, but one possibility is that increased lactate levels lead to decreased pH in the epileptic focus with arrest of seizure activity and postictal refractoriness. The antiseizure effect of low pH is probably mediated by reduced currents through the n-methyl-D-aspartate (NMDA) receptor (During et al., 1994).

It can also be argued that the upregulation of astrocytic MCT1 facilitates an increased influx of monocarboxylates into these cells, due to alterations in the astrocyte metabolism. Such upregulation of MCT1 is evident in skeletal muscle cells when the metabolism changes from glycolytic to oxidative after experimental reinnervation by a nerve that causes a continuous activation pattern (Bergersen et al., 2006). Hosoi and colleagues have shown that the monocarboxylate acetate is taken up by cultured rat astrocytes via MCT1 and that the uptake is enhanced when glutamate is added to the medium (Hosoi et al., 2009). The same group also discovered a 130 to 240% increase in the brain uptake of acetate 2 h after systemic injection of the chemoconvulsant pilocarpine (Hosoi et al., 2010). Thus, increased flux of monocarboxylates into astrocytes is likely to occur during glutamate stimulation and seizures. The upregulation of MCT1 on astrocytes in TLE may therefore reflect an increased metabolic demand for acetate and other monocarboxylates in epileptogenic areas of the brain.

Finally, the overexpression of MCT1 on astrocytes also suggests a lactate “siphoning” potential of the astrocyte syncytium in TLE. For example, neurons can release lactate into the extracellular space during periods of excitation and high glycolytic activity (Dienel and Cruz, 2008). The released lactate may then be taken up by perisynaptic astrocytes via MCT1 and “siphoned” via the astrocytic syncytium away from regions of high neuronal activity and toward areas of lower activity where the lactate can be more effectively metabolized. The lactate taken up this way may also be “siphoned” toward blood vessels and released to the blood for metabolism by peripheral organs such as liver, heart and oxidative skeletal muscle fibers; however, such a scenario is less likely in the hippocampal formation in TLE, where MCT1 is deficient on endothelial cells.