D-serine mitigates cell loss associated with temporal lobe epilepsy

Abstract

Temporal lobe epilepsy (TLE) is the most common type of drug-resistant epilepsy in adults, with an unknown etiology. A hallmark of TLE is the characteristic loss of layer 3 neurons in the medial entorhinal area (MEA) that underlies seizure development. One approach to intervention is preventing loss of these neurons through better understanding of underlying pathophysiological mechanisms. Here, we show that both neurons and glia together give rise to the pathology that is mitigated by the amino acid D-serine whose levels are potentially diminished under epileptic conditions. Focal administration of D-serine to the MEA attenuates neuronal loss in this region thereby preventing epileptogenesis in an animal model of TLE. Additionally, treatment with D-serine reduces astrocyte counts in the MEA, alters their reactive status, and attenuates proliferation and/or infiltration of microglia to the region thereby curtailing the deleterious consequences of neuroinflammation. Given the paucity of compounds that reduce hyperexcitability and neuron loss, have anti-inflammatory properties, and are well tolerated by the brain, D-serine, an endogenous amino acid, offers new hope as a therapeutic agent for refractory TLE.

Fig. 1. Assaying D-serine’s efficacy in preventing temporal lobe epilepsy.

a Schematic and details of the route and mode of D-serine administration. Note changes in stereotaxic coordinates required to accurately target the right MEA with a cannula (confirmed through injection of India ink, inset) as a function of age and body weight of the rat (gray area indicates range of animals used). The non-cannulated cohort of animals includes those utilized in past studies of TLE in the laboratory^,. The MEA is also referred to as area entorhinalis pars medialis in the literature. b Tree diagram showing behaviorally observed outcomes (status epilepticus, frank seizures, epilepsy or death) of pilocarpine treatment (insult) under conditions identified in a. The boxed numbers indicate percent of total animals (red) or percent of animals with status epilepticus (status, blue), under various regimens (changes are flagged to encourage comparison). All video recordings of behavior are available upon request.

Fig. 2. D-serine reduces number and severity of frank seizures and attenuates neuronal loss in the MEA.

a Timeline of experimental manipulations and behavioral outcomes (color coded). Pie charts indicate percent of post-status animals infused with aCSF (vehicle, left) or D-serine (right) that were categorized, based on video monitoring, as nonepileptic, pre-epileptic, or epileptic at the end of the latent period (subsequently used for stereological counting of neurons within MEA). Inset, the modified Racine scale used in this study for assessing seizure development and scoring severity. b Profile of changes in animal weight following initial insult (n = 15, 14, 11 animals for non-status aCSF or D-serine, post-status aCSF, and post-status D-serine, respectively). Error bars represent SEM. ***p < 0.001, two-way ANOVA with a Tukey’s post-hoc correction. c Cumulative distribution functions for changes in days to first seizure following initial insult with and without D-serine on board. d Severity of spontaneous seizures scored on a modified Racine scale (3–5) for animals infused with aCSF or D-serine. The number of animals used (n, numerator) is delineated from the total number of seizures recorded in them (denominator) within the bar plots. Error bars represent SEM. **p < 0.01, Student’s t test. e An example of a complete one-in-six series of brain sections (50 μm thick, covering the dorsal-ventral extent of the MEA; R right, L left) used for stereological counting of neurons. Note the ventral most location of the cannula (black arrowhead, section no. 12). f Brain sections from control and post-status rats treated D-serine or aCSF (vehicle), stained with thionin, and viewed at different magnifications (each set of panels, from L to R, are zoomed images of boxed regions, red (right), green (left), in the previous set). Note the conspicuous absence of Nissl-stained neurons (arrowheads) in layer (L) 3, but not L2, of MEA in aCSF-infused post-status rats. ld is lamina dissecans. g Images of sampling sites (at an optical depth interval of 5 μm) in wet-mounted brain sections from animals in various groups used for stereological counting of neurons. Neurons were counted based on visualization of their nucleoli (arrowheads). h Neuron counts estimated using stereology for non-status (controls) and post-status rats treated with D-serine or aCSF (left panel). Raw data and histograms of averaged neuronal counts from both hemispheres (middle panel) and a breakdown based on epileptogenic status (right panel; data presented in the interest of full disclosure and not used for any statistical comparison). The total number of animals used in this part of the study are indicated in the bar plot to the left of the panel). The numbers in the bars indicate the total number of animals (n) used. ***p < 0.001, ****p < 0.0001, one-way ANOVA with a Dunnett’s post-hoc correction. Error bars represent SEM.

Fig. 3. Focal application of D-serine minimizes astrocytic density in layer 3 of the MEA.

a Immunofluorescence images of the MEA (inset, bottom left panel) in non-status (controls) and post-status rats treated with D-serine or aCSF (vehicle). Neurons immunoassayed with fluorescently tagged antibodies against NeuN (red), astrocytes with antibodies against GFAP (green), and nuclei with DAPI (blue) shown merged (leftmost panels) and separately, as high magnification images of the boxed areas in panels to the left (yellow). Raw data and histograms of astrocytic (b) and microglia (c, immunoassayed with the anti-CD11b antibody, OX42, and shown in Fig. 4a) densities in layers 2 and 3 of MEA in animals under the conditions indicated (color codes show segregation of cohorts based on treatment regimen and final outcomes). Data within bar plots indicates number of animals used (numerator) and the total number of sections assayed for each condition (denominator). Astrocyte dendritic complexity (quantified using Sholl analysis d, f) and volume (e) in animals under the conditions indicated. Note differences in dendritic morphology of representative cells (left; arrowheads in d point to somata, purple and dendrites, yellow) and their tracings (middle), color coded for changes in the number of branches at concentric circles of increasing radii (right). Raw data and histogram of astrocytic cell volumes (e) and plot of the number of crossings as a function of radial distance (f), averaged across 10–14 cells from three animals in each cohort. g Scatter plot of volume versus maximum crossings for astrocytes in epileptic (pink), nonepileptic (green), and D-serine infused animals (blue). ***p < 0.001, ****p < 0.0001, one-way ANOVA with a Dunnett’s post-hoc correction. Error bars represent SEM.

Fig. 4. Focal application of D-serine limits microglia proliferation and/or infiltration into layer 3 of the MEA.

a Immunofluorescence images of the MEA in non-status (controls) and post-status rats treated with D-serine or aCSF (vehicle). Neurons immunoassayed with fluorescently tagged antibodies against NeuN (blue) and microglia with the anti-CD11b antibody, OX42 (green), shown merged (leftmost panels) and separately as high magnification images of the boxed areas in panels to the left (yellow). b Quadruple immunostaining of neurons (blue), astrocytes (green), microglia (magenta), and nuclei (blue) highlighting complete neuroglia pathology within MEA of post-status rats treated with D-serine or aCSF (panels on right are high magnification images of the boxed regions in panels on left). c Section of the brain slice containing the MEA that was micro dissected out for use in immunoblotting (DG dentate gurus, Hip hippocampus). Roughly four sections (600 μm thick) per hemisphere were harvested from each animal and pooled. Cerebellar tissue from these brains was used as control. d Representative immunoblots for NeuN (neurons), GFAP (astrocytes), and Iba1 (microglia) for MEA and cerebellum harvested from non-status (ns) and post-status (ps) rats 1, 5, 12, and 29 days post insult. GAPDH was used as loading control. Relative positions of standard molecular weight markers indicated to the right of each panel (for full scans see Supplementary Fig. 5). e Time course of changes in relative abundance of NeuN (neurons), GFAP (astrocytes), and Iba1 (microglia) in MEA and cerebellum quantified from the immunoblots (n = 3). Error bars represent SEM.

Fig. 5. Serine racemase is localized in both neurons and astrocytes in the MEA.

a Representative images of epileptic tissue acquired on a confocal microscope showing anti-NeuN (N), anti-serine racemase (SR), and anti-GFAP (A) immunoreactivity in neurons and astrocytes within different regions of the MEA (see schematic, row 6, column 3). Fluorophores for N (red), A (red), and SR (green) have been pseudo colored to show colocalization (yellow) in the merged (M) images. Panels in rows 3 through 6 are high magnification images of the lettered boxed regions identified above or within the same row. Text (top-right) in the high magnification panels identifies the immunoreactivity shown. For example, expression of serine racemase in neurons (box A) is shown in high mag through panels in row 3 (columns 1–3) and in astrocytes through panels in row 4 (columns 1–2). Panels in row 5 showcase racemase-positive (*) and racemase-negative neurons (red arrows) within the same section. bv blood vessels. b Representative images from control tissue showing colocalization of serine racemase with neurons and astrocytes. The panel in row 3, column 2 is a composite high mag image of N, A, and SR in the boxed region E (row 3, column 1). c A secondary-only control for SR with N and A (in control tissue) highlighting the specificity of the antibody used. In the absence of the SR antibody there is neither a signal nor colocalization with N or A. Each experiment shown in a–c was repeated at least twice independently with similar results.

Fig. 6. Assessment of D-serine levels in MEA and summary of neuroglia changes in TLE.

a Representative electropherograms of the secreted fractions from MEA (top row), LEA (middle row), and the cerebellum (bottom row) of non-status control and epileptic rats. Peak identification: (1) L-Ser; (2) D-Ser; (3) L-Asn; (4) L-Thr; (5) L-Gln; (6) L/D-Glu + D-Gln + D-Thr; (7) β-HSer (internal standard); (8) L/D-Asp; (9) L-His; (10) Gly; (11) L/D-Ala; (12) α-ABA (IS); (13) L-Tyr; (14) D-Tyr; (15) L-Met; (16) D-Met; (17) β-Ala; (18) L-Val; (19) D-Val; (20) Tau; (21) GABA; (22) L-Ile + L/D-Leu; (23) L/D-Phe; (24) L-Trp; (25) D-Trp; (26) β-Phe (IS); and (27) L/D-Arg. Insets (rightmost panel) are overlays of zoomed-in images of electropherograms from control (black) and epileptic (red) animals highlighting differences in the normalized fluorescence intensities of the L- and D-serine peaks (time-shifted to facilitate viewing). The bars atop peaks facilitate viewing of the amplitude differences. b Raw data and histograms of average D-serine levels in brain tissue (supernatant or secreted fraction, left panel; cell homogenate fraction, right panel) from the indicated regions of non-status control and epileptic rats expressed in μg/g of brain tissue (top panels) and as a percentage of total serine (D + L, bottom panels) within the samples. Numbers within bar plots indicate animals used (n) and the inset identifies location of LEA, presubiculum (PrS), and parasubiculum (Par) relative to the MEA. *p < 0.05, ***p < 0.001, Student’s t test. Error bars represent SEM. c Summary of neuroglia changes characterizing TLE pathology and two possible avenues through which D-serine might mitigate cell loss within the MEA.