Dietary therapy restores glutamatergic input to orexin/hypocretin neurons after traumatic brain injury in mice

Abstract

Study Objectives

In previous work, dietary branched-chain amino acid (BCAA) supplementation, precursors to de novo glutamate and γ-aminobutyric acid (GABA) synthesis, restored impaired sleep–wake regulation and orexin neuronal activity following traumatic brain injury (TBI) in mice. TBI was speculated to reduce orexin neuronal activity through decreased regional excitatory (glutamate) and/or increased inhibitory (GABA) input. Therefore, we hypothesized that TBI would decrease synaptic glutamate and/or increase synaptic GABA in nerve terminals contacting orexin neurons, and BCAA supplementation would restore TBI-induced changes in synaptic glutamate and/or GABA.

Methods

Brain tissue was processed for orexin pre-embed diaminobenzidine labeling and glutamate or GABA postembed immunogold labeling. The density of glutamate and GABA immunogold within presynaptic nerve terminals contacting orexin-positive lateral hypothalamic neurons was quantified using electron microscopy in three groups of mice (n = 8 per group): Sham/noninjured controls, TBI without BCAA supplementation, and TBI with BCAA supplementation (given for 5 days, 48 hr post-TBI). Glutamate and GABA were also quantified within the cortical penumbral region (layer VIb) adjacent to the TBI lesion.

Results

In the hypothalamus and cortex, TBI decreased relative glutamate density in presynaptic terminals making axodendritic contacts. However, BCAA supplementation only restored relative glutamate density within presynaptic terminals contacting orexin-positive hypothalamic neurons. BCAA supplementation did not change relative glutamate density in presynaptic terminals making axosomatic contacts, or relative GABA density in presynaptic terminals making axosomatic or axodendritic contacts, within either the hypothalamus or cortex.

Conclusions

These results suggest TBI compromises orexin neuron function via decreased glutamate density and highlight BCAA supplementation as a potential therapy to restore glutamate density to orexin neurons.

Statement of Significance

Sleep–wake disturbances are common following traumatic brain injury (TBI) and are associated with orexinergic dysfunction, which could result from reduced excitatory (glutamatergic) and/or increased inhibitory (γ-aminobutyric acid [GABA]ergic) neuronal input. The present study investigated whether TBI disrupts hypothalamic excitatory or inhibitory balance, and if so, whether dietary branched-chain amino acid (BCAA; precursors to de novo glutamate and GABA neurotransmitter synthesis) supplementation was restorative. We demonstrate that TBI reduces the density of glutamate immunogold labeling onto orexin neurons, and dietary BCAA supplementation restores this decrease back to preinjury levels. These data provide new evidence for a potential mechanism underlying orexinergic dysfunction following TBI and support prior work suggesting that dietary BCAA supplementation may be a promising therapy to improve sleep quality in humans following TBI.

Introduction

Traumatic brain injury (TBI) is a common (~1.7 million people in the United States per year) and complex injury of varying etiology (e.g., open/closed head injury, diffuse axonal injury) with a broad spectrum of symptoms and disabilities [1]. Sleep–wake disturbances, which inevitably impair functional recovery, are amongst the most prevalent and persistent sequelae of TBI regardless of injury severity [2–4]. Indeed, sleep–wake disturbances (e.g., excessive daytime sleepiness and pleiosomnia [5–7]) have been reported in ~30%–70% of individuals with mild, moderate, or severe TBI up to 3 years postinjury [5, 8–10]. Many studies to date have established pervasive and persistent sleep–wake impairments after TBI (cf. Ref. [11]).

Given the crucial role of orexin/hypocretin (hereinafter referred to as orexin) in regulating sleep and wakefulness [12], compelling evidence suggests that the underlying neurologic mechanism(s) explaining TBI-related sleep–wake disturbances involves orexinergic dysfunction. Orexin producing neurons, located in the perifornical region of the lateral hypothalamus, produce the neuropeptides orexin-A (hypocretin-1) and orexin-B (hypocretin-2) which bind to G-protein–coupled receptors [13–15] and project widely throughout the cerebral cortex, limbic system, and brainstem [12, 16–20] to activate wake-promoting monoaminergic and cholinergic neurons. Indeed, orexin has been shown to be an integral component of sleep–wake regulation [12, 21], with human narcolepsy [22] being associated with reduced orexin cerebrospinal fluid (CSF) levels [23–29] and orexin deficiency causing narcolepsy in dogs [30] and mice [31]. Exogenous orexin administration to animals promotes cortical brain activity, wakefulness, and cognitive performance [32, 33], and interstitial orexin dynamics have been associated with circadian rhythms of wakefulness and locomotor activity [20, 34–36]. In human subjects with TBI, Baumann et al. reported significantly reduced orexin neuropeptide CSF levels in patients with acute moderate-to-severe TBI [37], with this reduction in orexin CSF levels persisting for 6 months postinjury [5]. We have previously shown using in vivo microdialysis in mice that TBI reduced orexin neuropeptide levels [38]. We have also shown decreased orexin neuronal activation in response to sustained wakefulness after TBI in mice [39]. However, the underlying cause of orexinergic dysfunction post-TBI remains unknown.

One straightforward explanation that others have suggested is a reduction in the total number of orexinergic neurons following TBI-related damage to the lateral hypothalamus and/or other deep brain structures involved in the orexin system [40]. A small study examining four human brains on autopsy following lethal, severe TBI revealed a 27% reduction in orexin neurons compared with noninjured control brains [41]. Two recent studies in rodents have also shown that TBI significantly reduced orexin neuron counts [42, 43], although two other studies using mice did not show an effect of TBI on orexin neuron numbers [38, 39]. There may be methodological differences that underlie the discrepant findings, including age, species, and/or differing methods of inducing TBI across these studies. Taken together, the preliminary evidence from human brains supporting a reduction in orexin neuron counts post-TBI is neither conclusively supported, nor refuted, from the available literature in animal-based work, and alternative explanations for orexin dysfunction may still be valid.

Another possible explanation for the TBI-related reduction in orexin CSF levels is circuit-level dysfunction affecting excitation or inhibition ratios of inputs to orexin neurons. Indeed, previous work in rats [42] and mice [38, 39], using both lateral fluid percussion brain injury (FPI) and controlled cortical impact injury to induce TBI, has demonstrated a reduction of orexin neuronal activation after TBI. Our previous work identified dietary supplementation with branched-chain amino acids (BCAA), precursors to de novo glutamate and γ-aminobutyric acid (GABA) synthesis in the brain, restores network excitability and ameliorates hippocampal-dependent cognitive deficits in mice with TBI [44]. Furthermore, our previous work has shown that the same dietary BCAA supplementation protocol also restores wakefulness and orexin neuron activity in mice with TBI [39]. However, based on this prior work [39], it is still unknown whether glutamatergic and/or GABAergic input to orexin neurons is reduced following TBI, and if so, whether dietary BCAA supplementation restores glutamate and/or GABA input to orexin neurons to preinjury levels.

Thus, the purpose of the present study was to address this gap in the literature and thereby extend the findings by Lim et al., through a neuroanatomical examination of TBI-related changes in glutamate and GABA following an identical experimental paradigm of TBI and dietary BCAA supplementation as before [39]. The present study employs a novel, sequential orexin pre-embed immunolabeling, and glutamate/GABA postembed immunogold labeling technique to examine the relative density of nerve terminal glutamate and GABA immunogold labeling in axosomatic and axodendritic synapses via electron microscopy (EM) within the lateral hypothalamus, as well as within the penumbral area of injury in the cortex. The lateral hypothalamus was targeted due to being the sole locale of orexin producing neurons, and the penumbral area of injury in the cortex at layer VIb was specifically examined because of known orexinergic projections to this cortical layer [45, 46]. We hypothesized that following TBI, an overall increase in cortical inhibition would result from a decrease in the relative density of nerve terminal glutamate and/or an increase in the relative density of GABA immunogold labeling, within presynaptic nerve terminals contacting orexin-positive neurons in the hypothalamus. Secondarily, we hypothesized that dietary BCAA supplementation would restore the normal density of glutamate and GABA immunogold labeling within terminals contacting orexin-positive neurons.

Experimental Procedures

Animals

All experiments were performed on 5- to 7-week-old, 20–25 g, male C57BL/6J mice (Jackson Laboratory: https://www.jax.org/strain/000664). The animals were housed in an accredited animal facility maintained at an ambient temperature of 23 ± 1°C with a relative humidity of 25 ± 5% on an automatically controlled 12 hr light/12 hr dark cycle (light on at 07:00 hr, illumination intensity ~100 lux). The animals had free access to food and water. Every effort was made to minimize the number of animals used and any pain and discomfort experienced by the subjects. Animal experiments were performed in accordance with the guidelines published in the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the VA Portland Health Care System Animal Care and Use Committee in accordance with international guidelines on the ethical use of animals.

The experimental design (Figure 1) replicated the exact protocol previously published [39] and therefore utilized the same TBI model (see Fluid percussion injury) and dietary BCAA supplementation protocol (see Branched-chain amino acid supplementation) in the same three experimental groups of mice: (1) Sham/noninjured (Control), (2) TBI without BCAA supplementation (TBI), and (3) TBI with BCAA supplementation (TBI + BCAA).

Figure 1.

Schematic of our experimental design. Brains were extracted from three experimental groups (Sham/Control, TBI, and TBI + BCAA), sectioned, and processed for double-labeling of orexin and postembed immunogold labeling of glutamate or GABA using electron microscopy. Presynaptic terminals in both the lateral hypothalamus and ipsilateral penumbral injured cortex were analyzed for glutamate and GABA immunogold density. Synapses in the lateral hypothalamus were categorized into orexin-positive axosomatic (terminals contacting cell bodies) or orexin-positive axodendritic (terminals contacting dendrites). Presynaptic terminals in the ipsilateral penumbral injured cortex were analyzed for glutamate and GABA immuno-gold density. Synapses in the cortex were all orexin-negative and axodendritic (i.e., terminals contacting dendrites). In both the lateral hypothalamus and the cortex, glutamatergic synapses were found to be primarily asymmetrical, whereas GABAergic synapses were found to be primarily symmetrical. Hypothal = lateral hypothalamus; Glu = glutamate.

Fluid percussion injury

The mild FPI protocol was carried out over 2 days, as described previously [39]. On the first day, the animal was anesthetized using inhaled isoflurane (3%–5% for induction, and 1%–2% for maintenance) and placed in a mouse stereotactic frame (Stoelting). The scalp was incised and reflected. A craniotomy was performed with a trephine (3 mm outer diameter) over the right parietal area between bregma and lambda, lateral to the sagittal suture, and lateral to the lateral cranial ridge (coordinates relative to bregma: AP −2.0 mm, ML 2.7 mm). The dura remained intact throughout the craniotomy procedure. A rigid Luer-loc needle hub (3 mm inside diameter) was secured to the skull over the opening with Loctite adhesive and subsequently cyanoacrylate plus dental acrylic. The skull sutures were sealed with the cyanoacrylate during this process to ensure that the fluid bolus from the injury remained within the cranial cavity. The Luer-loc needle hub was filled with isotonic sterile saline and the hub was capped. The mouse was then placed on a circulating water heating pad and returned to the home cage once ambulatory. On the second day, the animal was briefly placed under isoflurane anesthesia (500 mL/min) via nose cone, and respiration was visually monitored. When the animal was breathing once every 2 s, the nose cone was removed, the cap over the hub removed, and dural integrity visually confirmed. The hub was topped off with isotonic sterile saline, and a 32-cm section of high-pressure tubing extending from the FPI device attached to the Luer-loc fitting of the hub (Department of Biomedical Engineering, Virginia Commonwealth University, Richmond, VA). The animal was then placed on its left side and observed. Once normal breathing resumed and just as the animal regained its toe pinch withdrawal reflex, a 20 ms pulse of saline onto the dura was delivered. A pressure gauge attached to an oscilloscope was used to ensure delivered pressures between 1.4 and 2.1 atmospheres. Importantly, based on a significant body of work over the past 3 decades [47], this fluid pressure pulse has previously been shown to generate a mild brain injury, given the short latency to righting reflex [48, 49] and lack significant cell death and cavitation of the ventricles [50]. Immediately after injury, the hub was removed from the skull and the animal was placed in a supine position. The animal was then reanesthetized with isoflurane for scalp closure. Control (sham/noninjured) animals received all of the above, with the exception of the fluid pulse. The animal was returned to a circulating water heating pad until ambulatory and then returned to its home cage.

Branched-chain amino acid supplementation

As described previously [39], 48 hr following the mild FPI procedure, mice were randomly assigned to receive either BCAA-supplemented drinking water (TBI + BCAA) or untreated drinking water (TBI) for 5 days. BCAA-supplemented drinking water consisted of a combination of L-leucine, L-isoleucine, and L-valine (obtained individually from Sigma-Aldrich, St. Louis, MO, USA) at 100 mM each, to be consumed ad libitum. The amount of drinking water remaining in the bottle was measured each day, and fresh BCAA or control drinking water was replenished as needed. Mice drank on average 3–5 mL of solution per day, and we have previously shown that a similar dosage of BCAA supplementation does not affect body weight [44]. The original rationale for starting BCAA supplementation 48 hr after injury, and not sooner, was due to an effort to avoid significant dietary alterations in the immediate/acute period postinjury time period when transient pathological alterations are occurring, such as minor swelling, cell death, and localized blood–brain barrier breakdown. Seven days after sham or TBI injury (i.e., 5 days after start of BCAA treatment in the TBI + BCAA group), mice were deeply anesthetized with ketamine/xylazine (100 mg/kg ketamine, 10 mg/kg xylazine) and transcardially perfused with 1000 units/mL of heparin in 0.1 M phosphate buffered saline (PBS), followed immediately with 50 mL of EM fixative (0.5% paraformaldehyde, 1% glutaraldehyde, and 0.1% picric acid in 0.1 M PBS, pH 7.3, at room temperature). Brains were extracted and further postfixed in EM fixative for 90 min using a novel microwave procedure with modifications [51–54].

Pre-embed immunohistochemistry

Orexin immunohistochemistry was performed on the lateral hypothalamus and a portion of the ipsilateral penumbral injured cortex in such a way as to be compatible with EM and postembed immunohistochemical reactions for glutamate and GABA. This protocol for orexin pre-embed labeling using a modified microwave technique is novel and has not been previously published.

The Pelco BioWave (Ted Pella, Redding, CA, USA) microwave was used to process the tissue for immunohistochemistry and EM processing. Brains were sectioned at 60 µm and a 1:6 series was processed for diaminobenzidine (DAB) immunolabeling for localization of orexin-A (#8070, goat polyclonal antibody, Santa Cruz Biotechnology, Inc., Dallas, TX, USA). The primary antibody concentration was a 1:50 dilution using the microwave procedure (further details below). In our experience, the Pelco BioWave microwave technique significantly expedites the tissue processing time, enabling greater depth of antibody penetration and minimizing the leaching of glutamate and GABA out of the brain sections into the surrounding solution. Using this technique, all n = 24 brains were sectioned within a single day, and all tissue was processed for pre-embed immunohistochemistry within 48 hr of sectioning.

Tissue was incubated in a microwave processor at 35°C for 5 min at 550 W with the vacuum on in 10 mM Tris base and 1 mM EDTA, pH 9.0 (antigen retrieval), rinsed in 0.1 M PBS for 2 × 1 min at 150 W with the vacuum off, exposed to 3% hydrogen peroxide for 1 min at 150 W with the vacuum off, rinsed in 0.1 M PBS for 2 × 1 min at 150 W with the vacuum off, exposed to 0.5% Triton X-100 for 5 min at 550 W with the vacuum on, and then treated with the primary antibody (e.g., orexin goat polyclonal) for 36 min at 200 W four times using the following microwave cycle: 2 min on, 2 min off, 2 min on, and 5 min off, all with the vacuum on at a constant 20 mm Hg. The tissue was then rinsed in 0.1 M PBS for 2 × 1 min at 150 W with the vacuum off and then exposed to the secondary antibody (#705-065-003, biotinylated donkey antigoat; Jackson ImmunoResearch; West Grove, PA, USA) at a dilution of 1:100 for 16 min at 200 W for two cycles of the following microwave cycle: 4 min on, 3 min off, 4 min on, and 5 min off, all with the vacuum on at a constant 20 mm Hg. The tissue was then rinsed in 0.1 M PBS for 2 × 1 min at 150 W with the vacuum off and then exposed to ABC (Vector Elite Kit, 1 μL/mL solution A and B in working imidazole) for 16 min at 150 W with the vacuum on using the following microwave cycle: 4 min on, 3 min off, and 4 min on at a constant 20 mm Hg. The tissue was then rinsed in 0.1 M imidazole buffer for 2 × 1 min at 150 W with the vacuum off and then exposed to DAB (0.5 μg/mL + 1.5% hydrogen peroxide in 0.1 M imidazole buffer) for 10 min in the microwave at 200 W. Finally, the tissue was rinsed in 0.1 M imidazole buffer for 1 min in the microwave at 150 W with the vacuum off and then rinsed in 0.1 M PBS for 1 min in the microwave at 150 W with the vacuum off.

Orexin-specific labeling was clearly visualized within coronal brain sections using light microscopy at 10× and 20× magnification and showed DAB precipitate within individual cell bodies in the lateral and posterior hypothalamus (relative to bregma, −1.06 to 0.14 mm) [55]. Using our previously published methods of orexin immunohistochemistry and light microscopy (i.e., fixation in 4% paraformaldehyde in 0.1 M PBS), orexin labeling along projection fibers and fibers of passage was clearly visualized extending throughout the rest of the forebrain, including terminals in layer VIb of the cortex as previously described (data not shown) [18, 45, 46, 56]. However, in brain tissue that was EM-fixed using the above protocol (i.e., 0.5% paraformaldehyde, 1% glutaraldehyde, and 0.1% picric acid), orexin labeling of fibers within cortical layer VIb was substantially reduced and only sporadically observed.

Electron microscopy processing

After orexin pre-embed labeling was performed, sections were then embedded for EM processing using the microwave procedure as described previously [57]. Sections were incubated in 1% osmium tetroxide and 1.5% potassium ferricyanide, rinsed three times in deionized water, incubated in 0.5% uranyl acetate for 6 min, and then dehydrated in 50%, 75%, 95%, 100% ethanol, and finally a 1:1 Epon-Spurs with propylene oxide incubations. Sections were then incubated in Epon-Spurs Resin and then placed in an oven overnight at 60°C.

Following overnight incubation, regions of interest (ROIs; 1 × 1 mm in area) from both the lateral hypothalamus and the penumbral area of the ipsilateral-injured cortex (somatosensory S1 cortex, layer VIb—where orexin projection fibers were observed to be most dense [45, 46]) were microdissected out and glued onto two separate resin blocks for sectioning. Layer VIb below the penumbral area of injury in the cortex is deep enough to avoid direct physical damage from the injury, yet still superficial enough to be potentially afflicted postinjury. The same coordinates for the ROI from both the lateral hypothalamus and cortex were used across all animals. These microdissected brain regions were then sectioned at a thickness of 60 nm using an ultra-microtome (Leica Biosystems, EM UC7, Buffalo Grove, IL, USA). Sections were cut leaving a leading edge of the tissue for both brain regions using a diamond knife (Diatome, Hatfield, PA, USA).

Postembed immunogold labeling

Postembedding immunogold EM was performed using a glutamate (#G6642, nonaffinity-purified, rabbit polyclonal; Sigma-Aldrich) and GABA antibody (#A2052, affinity-purified, rabbit polyclonal; Sigma-Aldrich) for both the lateral hypothalamus and cortex on sections placed on 50-slot grids. The primary glutamate antibody was diluted 1:250 in TBST (0.05 M Tris, pH 7.6, with 0.9% NaCl and 0.1% Triton X-100), and aspartate (1 mM; factored into the 1:250 dilution of TBST) was added to the glutamate antibody mixture 24 hr before incubation to prevent any cross-reactivity with aspartate within the tissue, as described previously [58]. The primary GABA antibody was diluted 1:250 in TBST (0.05 M Tris, pH 7.6, with 0.9% NaCl and 0.1% Triton X-100). The secondary glutamate and GABA antibodies were goat antirabbit IgG (Jackson ImmunoResearch; in TBST, pH 8.2), tagged with 12 nm gold particles, incubated at a 1:50 dilution for 90 min. Tissue was then counterstained with 5% uranyl acetate for 30 min and 5% lead citrate for 30 s. Water rinses (3 × 5 min) were used to wash off solutions in between incubations. Photographs (10–20 per synapse type: cell or dendrite) were taken using a digital camera (Advanced Microscopy Techniques, Danvers, MA, USA) on a JEOL 1400 transmission electron microscope (JEOL, Peabody, MA, USA) of orexin-labeled dendrites and neuronal cell bodies from a single 50-slot grid throughout the lateral hypothalamus and cortex for a final magnification of 40,000×. Images were taken in a systematically random and unbiased fashion by an individual blinded to the experimental groups. The density of immunogold labeling was determined within each presynaptic nerve terminal synapsing onto an orexin-positive cell body or dendrite as further described below.

Electron microscopy quantification

For quantification of both glutamate and GABA labeling, the area of the nerve terminal was first determined by drawing a ROI in ImagePro Plus software (Media Cybernetics, Rockville, MD, USA). Next, the number of immunogold particles located within or at least touching the synaptic vesicle membrane (i.e., vesicular pool, including both the readily releasable, recycling pool and reserve pool) was automatically counted using ImagePro Plus. Quantification of glutamate and GABA immunogold density was done by an individual blinded to the experimental groups. The entire area of the nerve terminal within a single image was measured, since all three pools of synaptic vesicles (readily releasable, recycling, and reserve) cover the entire area of the nerve terminal. Synapses were classified as axosomatic versus axodendritic, as well as symmetrical and asymmetrical based upon established morphological criteria.

Throughout the lateral hypothalamus and cortex, nerve terminals making a symmetrical synaptic contact were observed to contain some glutamate immunogold particles, but very few nerve terminals making an asymmetrical synaptic contact were observed to contain labeling for GABA immunogold particles. Our previous work has reported that nerve terminals making a symmetrical contact contain GABA [59], the precursor for which is glutamate. Therefore, nerve terminals making a symmetrical contact will naturally contain some glutamate immunolabeling and cannot be considered immuno-negative as a way of determining a ratio between glutamatergic and GABAergic terminals [59, 60]. Thus, the data presented here reflect only the nerve terminals making an asymmetrical synaptic contact labeling for glutamate immunogold, and the nerve terminals making a symmetrical synaptic contact labeling for GABA immunogold.

Within the hypothalamus, only glutamate or GABA immunogold labeled terminals contacting orexin positive neurons were photographed, which was the main purpose of this study. Within the penumbral cortical layer VIb, because orexin labeling of fibers was minimally and only sporadically observed due to limitations from the EM-fixative (see Pre-embed immunohistochemistry), photographs were randomly taken within the neuropil of glutamate and GABA immunogold labeled terminals making synaptic contact with unlabeled dendrites. Again, consistent with the quantification method for lateral hypothalamus above, only the nerve terminals making an asymmetrical synaptic contact were quantified for glutamate immunogold labeling, and only the nerve terminals making a symmetrical synaptic contact were quantified for GABA immunogold labeling.

From these data, we determined the relative density of glutamate immunogold labeling within nerve terminals (number of gold particles/µm² of terminal area). Previous work has reported that the cytoplasmic pool accounts for less than 10% of the entire nerve terminal pool of glutamate, so the two pools were combined in the final analysis [59, 60]. The specificity of the glutamate and GABA antibodies has previously been reported, with incubation of the antibody with 3 mM glutamate or GABA resulting in no immunogold labeling for either neurotransmitter [60]. Background glutamate and GABA immunogold labeling was determined within glial cell processes and was found to be 10 immunogold-labeled particles/µm². This was subtracted from the density of presynaptic immunogold-labeled glutamate and GABA within the nerve terminals.

Statistical analyses

All statistical analyses were performed using R (A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/). For all tests, p < 0.05 was considered statistically significant. Data are reported as the mean ± SEM unless otherwise noted. An a priori power analysis revealed the minimum sample size required to see a statistically significant effect (power = 0.8, alpha = 0.05) to be n = 6 per group. Thus, our experimental design here includes an n = 6–8 per group: (1) Sham/noninjured (Control), (2) TBI without BCAA supplementation (TBI), and (3) TBI with BCAA supplementation (TBI + BCAA). The density of gold particles per square micrometer of nerve terminal area was calculated for each individual animal based on the nerve terminal area and number of gold particles within the ROI, which was compared using a one-way analysis of variance (ANOVA) followed by Tukey’s honestly significant difference (HSD) post hoc test for multiple comparisons (Control vs. TBI, Control vs. TBI + BCAA, and TBI vs. TBI + BCAA).

Results

Hypothalamus

The effect of TBI and dietary BCAA supplementation on the density of glutamate immunogold labeling within nerve terminals contacting orexin-positive neuronal cell bodies (axosomatic) or dendrites (axodendritic) in the lateral hypothalamus was quantified and compared across groups. The specificity of the glutamate and GABA antibodies has previously been reported, with incubation of the antibody with 3 mM glutamate or GABA resulting in no immunogold labeling for either neurotransmitter [60].

The density of glutamate immunogold labeling within presynaptic terminals making an axosomatic contact with an orexin-positive neuron was unchanged after TBI, and after TBI with dietary BCAA supplementation (one-way ANOVA, F(2,18) = 0.095, p = 0.91; Figure 2, A–D). In contrast, the density of glutamate immunogold labeling within presynaptic terminals making an axodendritic contact with an orexin-positive neuron was significantly decreased following TBI and significantly increased following TBI with dietary BCAA supplementation (one-way ANOVA, F(2,19) = 5.3314, p = 0.0145. Tukey HSD post hoc test, Control vs. TBI, p = 0.0173; and TBI vs. TBI + BCAA, p = 0.0462; Figure 3, A–D). There was no statistically significant change in the area of the glutamate immunogold containing nerve terminals between any of the treatment groups (data not shown). All glutamate data reported within lateral hypothalamus reflect asymmetrical synapses, as discussed in Methods.

Figure 2.

Glutamate immunogold density within orexin-positive axosomatic synapses in the lateral hypothalamus. Representative electron photomicrograph showing a nerve terminal making an axosomatic synaptic contacting onto an orexin-DAB–labeled soma (cell body) from a vehicle-treated control animal (A; Control, n = 8), a vehicle-treated TBI animal (B; TBI, n = 6), and a BCAA-treated TBI animal (C; TBI+BCAA, n = 7). Groups did not significantly differ in the density of glutamate within terminals making an axosomatic synaptic contact onto orexin-DAB positive neurons (D). Data are normalized such that TBI and TBI + BCAA are expressed as a percentage of Control. The whiskers reflect the minimum and maximum values, whereas the box is constructed of the median of the first and third quartile. The line in the box represents the median of the data set and the plus sign represents the mean of the data set. Note that the data here reflect only asymmetrical synapses. NT = presynaptic nerve terminal; Soma = the cell body; white arrows without stem are pointing towards gold particles; white arrows with stem are pointing towards synapses; asterisk denotes DAB labeling; Hypothal = lateral hypothalamus. Scale bar = 500 nm.

Figure 3.

Glutamate immunogold density within orexin-positive axodendritic synapses in the lateral hypothalamus. Representative electron photomicrograph showing a nerve terminal making an axodendritic synapse contacting onto an orexin-DAB–labeled dendrite from a vehicle-treated control animal (A; Control n = 8), a vehicle-treated TBI animal (B; TBI, n = 6), and a BCAA-treated TBI animal (C; TBI + BCAA, n = 7). Relative nerve terminal glutamate density is decreased across axodendritic synapses in TBI animals by approximately 30%, compared with controls, and restored to control levels with dietary BCAA supplementation (D). Data are normalized such that TBI and TBI + BCAA are expressed as a percentage of Control. The whiskers reflect the minimum and maximum values, whereas the box is constructed of the median of the first and third quartile. The line in the box represents the median of the data set and the plus sign represents the mean of the data set. Note that the data here reflect only asymmetrical synapses. NT = pre-synaptic nerve terminal; Dend = the dendrite; white arrows without stem are pointing towards gold particles; white arrows with stem are pointing towards synapses; asterisk denotes DAB labeling; Hypothal = lateral hypothalamus. Scale bar = 500 nm.

TBI did not significantly affect the density of GABA immunogold labeling within presynaptic terminals making an axosomatic (one-way ANOVA, F(2,13) = 0.2392, p = 0.79; Figure 4D) or axodendritic (one-way ANOVA, F(2,14) = 1.8813, p = 0.19; Figure 4E) contact with an orexin-positive neuron. There was no change in the area of the GABA immunogold containing nerve terminals between any of the treatment groups (data not shown). All GABA data reported within lateral hypothalamus reflect symmetrical synapses, as discussed in Methods.

Figure 4.

GABA immuno-gold density within axosomatic and axodendritic synapses in the lateral hypothalamus. Representative electron photomicrographs showing a nerve terminal making an axodendritic synaptic contact from a vehicle-treated control animal (A; Control, n = 8), a vehicle-treated TBI animal (B; TBI, n = 6), and a BCAA-treated TBI animal (C; TBI+BCAA, n = 6). In the lateral hypothalamus (LH), TBI does not significantly affect the density of GABA immunogold labeling within presynaptic terminals contacting either an orexin-positive neuronal cell body (D) or dendrite (E). The whiskers reflect the minimum and maximum values, whereas the box is constructed of the median of the first and third quartile. The line in the box represents the median of the data set and the plus sign represents the mean of the data set. Note that in the lateral hypothalamus, very few nerve terminals making an asymmetrical synaptic contact were observed containing GABA immunogold particles. Therefore, the data here reflect only symmetrical synapses. NT = presynaptic nerve terminal; Dend = the dendrite; white arrows without stem are pointing towards gold particles; white arrows with stem are pointing towards synapses; asterisk denotes DAB labeling; Hypothal = lateral hypothalamus. Scale bar = 500 nm.

Penumbral cortex

To determine the effect of TBI and dietary BCAA supplementation on the density of glutamate and GABA immunogold labeling within terminals in the injured cortex (layer VIb), presynaptic nerve terminal immunogold density was quantified and compared between groups. Very few axosomatic synaptic contacts were observed surrounding the neurons within layer VIb of the somatosensory cortex; therefore, only axodendritic synaptic contacts were analyzed. Therefore, these cortical data likely reflect a variety of other cell types, such as pyramidal, fusiform, cells of Martinotti, and interneurons. Again, consistent with the quantification method for the lateral hypothalamus above, only the nerve terminals making an asymmetrical synaptic contact were quantified for glutamate immunogold labeling, and only the nerve terminals making a symmetrical synaptic contact were quantified for GABA immunogold labeling.

TBI mice showed a significant decrease in the density of glutamate immunogold labeling within presynaptic terminals making an axodendritic contact (one-way ANOVA, F(2,23) = 4.2848, p = 0.0262. Tukey HSD post hoc test, Control vs. TBI, p = 0.0387; Figure 5A–D). This effect was not rescued by BCAA therapy (Tukey HSD post hoc test, TBI vs. TBI + BCAA, p = 0.8096). In contrast, TBI did not significantly affect the density of GABA immunogold labeling within presynaptic terminals making axodendritic symmetrical contacts (one-way ANOVA, F(2,16) = 1.1781, p = 0.33; Figure 6A–D). These data indicate that TBI decreases the density of glutamate immunogold labeling within terminals making an excitatory, but not inhibitory, synaptic contact in the injured cortex. However, this effect was not rescued by BCAA therapy. There was no change in the area of the GABA or glutamate immunogold containing nerve terminals between any of the treatment groups (data not shown).

Figure 5.

Glutamate immunogold density within axodendritic synapses in the penumbral injured cortex. Representative electron photomicrographs showing a nerve terminal making an axospinous synaptic contact from a vehicle-treated control animal (A; Control, n = 8), a vehicle-treated TBI animal (B; TBI, n = 6), and a BCAA-treated TBI animal (C; TBI + BCAA, n = 8). In layer VIb of the penumbral injured cortex (CTX), TBI mice show significantly decreased the density of glutamate immunogold labeling within presynaptic terminals making asymmetrical synaptic contacts onto postsynaptic dendritic spines compared to the control group (D), but this effect was not rescued by BCAA supplementation. Data are normalized such that TBI and TBI + BCAA are expressed as a percentage of Control. The whiskers reflect the minimum and maximum values, whereas the box is constructed of the median of the first and third quartile. The line in the box represents the median of the data set and the plus sign represents the mean of the data set. Note that in the cortex, very few nerve terminals making a symmetrical synaptic contact were observed labeled for glutamate; therefore, glutamate data presented here reflect only asymmetrical synapses. NT = presynaptic nerve terminal; SP = dendritic spine; white arrows without stem are pointing towards gold particles; white arrows with stem are pointing towards synapses. Scale bar = 500 nm.

Figure 6.

GABA immunogold density within axodendritic synapses in the penumbral-injured cortex. Representative electron photomicrographs showing a nerve terminal making an axodendritic synaptic contact from a vehicle-treated control animal (A; Control, n = 6), a vehicle-treated TBI animal (B; TBI, n = 5), and a BCAA-treated TBI animal (C; TBI + BCAA, n = 8). In layer VIb of the penumbral injured cortex (CTX), TBI mice showed a small, but not statistically significant, decrease in the density of GABA immuno-gold labeling within presynaptic terminals making symmetrical synaptic contact onto the postsynaptic dendrite compared to the control group (D). The whiskers reflect the minimum and maximum value, while the box is constructed of the median of the first and third quartile. The line in the box represents the median of the data set and the plus sign represents the mean of the data set. Note that in the cortex, very few nerve terminals making an asymmetrical synaptic contact were observed labeled for GABA; therefore, GABA data reflect only symmetrical synapses. NT = presynaptic nerve terminal; Dend = the dendrite; white arrows without stem are pointing towards gold particles; white arrows with stem are pointing towards synapses. Scale bar = 500 nm.

Discussion

The present study addresses a gap in the literature by extending the findings by Lim et al. [39], to determine a potential mechanism by which TBI causes a decreased orexin neuron function, and by which a promising dietary therapy may restore TBI-induced sleep–wake disturbances. These results demonstrate that TBI decreased the density of glutamate immunogold labeling within presynaptic nerve terminals making contact onto orexin-labeled neurons in the lateral hypothalamus, and this effect was restored by dietary BCAA supplementation. Additionally, TBI had no effect on the density of GABA immunogold labeling within presynaptic nerve terminals onto orexin neurons in the lateral hypothalamus, suggesting that, at least in our mouse model, the main effect of TBI is reduction of excitation, rather than increased inhibition, onto orexin neurons. In addition to findings in the lateral hypothalamus, we also found that TBI decreased the density of glutamate immunogold labeling in the presynaptic terminals within the deeper layers of the injured somatosensory cortex. However, this cortical effect was not restored by dietary BCAA supplementation. Although unknown, it is possible that the direct physical damage imposed to the cortex underlies why synapses in the penumbral cortical layer VIb of the injured cortex are not amenable to the neurorestorative effect of dietary BCAA supplementation. Taken together, these results suggest the mechanism by which TBI compromises orexin neuron function and causes sleep–wake disturbances may be through a reduction in the density of glutamate contacting orexin-positive neurons in the lateral hypothalamus. Furthermore, these results highlight the potential for dietary supplementation with BCAAs following TBI as a restorative therapy for glutamate deficits within presynaptic nerve terminals that contact orexin neurons. In this way, TBI-related sleep–wake disturbances could potentially be ameliorated through normalizing the density of glutamate content within excitatory inputs to the lateral hypothalamus.

Sleep–wake disturbances (e.g., excessive daytime sleepiness and pleiosomnia [5–7]) are amongst the most prevalent and persistent sequelae of TBI regardless of injury severity [2–5, 8–10]. Many studies to date have established pervasive and persistent sleep–wake impairments after TBI (cf. Ref. 11) both in humans and animals. Collectively, these studies support there being an increase in sleep, less time spent in wakefulness, and an overall increase in sleep–wake fragmentation post-TBI (cf. Ref. 11). Furthermore, although additional work utilizing quantitative electroencephalography (EEG) techniques is needed, previous work has reported an increase in δ power during NREM sleep [61] (suggesting an increase in sleep pressure), which has been shown to be reduced following BCAA supplementation [39]. Power spectral analyses have also shown reduced θ power, and a shift of θ peaks to slower frequencies post-TBI which can also be ameliorated with BCAA supplementation. An increase in total sleep time, less time spent in wakefulness, increased sleep–wake fragmentation, and increased δ power during sleep are all suggestive of a reduction in orexinergic activity. Although the mechanistic basis for this reduction in orexinergic activity post-TBI remains unknown, the present study provides neuroanatomical evidence that it may stem from a reduction in excitatory (glutamatergic) input to orexin neurons. Indeed, glutamatergic input has been established as an important mediator of orexin neuronal activity [62].

Of interest is the fact that we found an effect of TBI on the density of glutamate immunogold labeling within axodendritic, but not axosomatic synapses within the lateral hypothalamus. The significance of this distinction has yet to be elucidated, but it is possible that axodendritic synapses reflect a different neuronal input population compared with axosomatic synapses. This topographical organization of different inputs making contacts onto cell bodies versus dendrites has been reported in other brain regions such as the striatum [63] and cortex (e.g., vesicular glutamate transporter 1 [cortical] input primarily to the apical dendrites in layer II vs. vesicular glutamate transporter 2 [thalamic] input to primarily the basal dendrites in layer V [C. Moore and C. Meshul, unpublished observations]). Less has been described about the topographical organization of inputs within the lateral hypothalamus, but it is known that approximately 14% of inputs to orexin neurons are from the cortex, 12% are from the lateral septum, 10% are from the bed nucleus of the stria terminalis, and 35% of the inputs are from local interneurons (some of which are also orexin/glutamate containing neurons) [20, 64]. It remains possible that TBI differentially affects some of these inputs more than others due to the proximity to the direct injury or other reasons.

The BCAAs L-Leucine, L-isoleucine, and L-Valine cannot be synthesized endogenously by humans and, therefore, are acquired solely through dietary intake. BCAAs are involved in a multitude of physiologic functions, including the de novo synthesis of glutamate and GABA in the brain, which are the primary excitatory and inhibitory, respectively, neurotransmitters in the central nervous system [65–67]. Radiolabeling studies suggest that ~50% of dietary BCAAs comprise de novo glutamate and subsequently GABA synthesis via the nitrogen donation cycle, and that 40% of BCAAs end up in small clear vesicles containing glutamate at the nerve terminal [68]. Thus, dietary BCAA supplementation may function to re-establish network excitability after TBI by restoring synaptic pools of vesicular glutamate and GABA. However, it is also possible that BCAA supplementation may restore orexin neuronal function via an alternate mechanism besides restoring synaptic pools of vesicular glutamate and GABA. One such possibility, albeit speculative, is through “anti-inflammatory” effects such as mitigation of microglial activation. Previous work has reported TBI to acutely increase proinflammatory cytokines (i.e., IL-1β) as well as promote microglial activation (determined via IBA-1 staining) in the injured cortex [69]. This mechanism remains to be explored.

Glutamate plays an integral role in the regulation of orexin neuronal excitability. Indeed, glutamatergic neuronal input to orexin neurons regulate wakefulness, and glutamatergic interneurons have been suggested to play a role in a positive feedback recruitment of orexin to orexin neurons [62]. More than 50% of orexin neurons in the lateral hypothalamus corelease glutamate and also express vesicular glutamate transporters 1 and 2 [70]. Additionally, asymmetric synaptic contacts, which are generally glutamatergic (i.e., excitatory), have been shown to dominate over symmetric synapses of orexin neurons in the lateral hypothalamus [71]. Consistent with this report, in our present study, the majority of synapses visualized were also asymmetrical.

How might TBI decrease relative glutamate immunogold density in the hypothalamus and penumbral cortex? Previous studies have demonstrated that immediately following TBI (i.e., <60 min), there is an initial increase in glutamate release in rodents [72–75] and in humans 24 hr post-TBI [76]. However, in the present study and that which was previously published by Lim et al. [39], the focus was on persistent sleep impairment after TBI, and thus, the animals in the current study were examined 7 days post-TBI. Little is known regarding the temporal changes in glutamate release through this longer time period postinjury. Recent work by Cantu et al. investigated long-term (2–4 weeks) changes in glutamate signaling using a FRET-biosensor in a mouse model of severe TBI (via controlled cortical impact) and demonstrated an increase in glutamate signaling in specific cortical regions [77]. Additionally, in a rat model of TBI (via controlled cortical impact), it has been shown that for up to 5 days post-TBI, cortical glutamate measured using high-pressure liquid chromatography is increased following a severe injury and unchanged following a mild injury [78]. Our EM study results indicate a decrease in both cortical and hypothalamic glutamate immunogold density in presynaptic nerve terminals following mild TBI.

Previous studies have demonstrated modest improvement in disability and cognition in human subjects with severe TBI after the administration of intravenous BCAAs (~20 g/day) for 15 days [79, 80]. This neurorestorative effect has also been demonstrated in C57BL/6J mice with TBI after dietary supplementation of BCAAs via drinking water, which improved memory (e.g., hippocampal-dependent contextual fear conditioning) and sleep–wake disturbances [39, 44, 81]. The potential neuroprotective effect of BCAA supplementation given prior to TBI has not yet been examined. One hypothesis as to the mechanism of action of dietary BCAA in the improvement of hippocampal excitation:inhibition ratio and memory is the replenishment of decreased glutamate synthesis at the synapse. In support of a region-specific “glutamate sink” after TBI, our recent study showed that BCAA administration needed to be “on board” in order to have a neurorestorative effect on hippocampal memory function, and withdrawal of BCAA caused reversion to the injured phenotype [81]. This observation supports the hypothesis that BCAAs may act to replenish vesicular glutamate after TBI. Additional work that bridges our ultrastructural findings with the more global physiological changes reported for glutamate/GABA signaling after TBI is still needed.

Our study has several important limitations. EM, while a powerful tool providing a neuroanatomical window into the relative density of neurotransmitter within the presynaptic nerve terminal, does not provide the functional perspective that is gained with in vivo microdialysis, glutamate biosensors, and/or synaptic electrophysiology. Previous studies may allow some extrapolation to correlate ultrastructural EM results with functional studies. We have previously reported an inverse relationship between the density of both glutamate and GABA immunogold labeling and changes in the extracellular levels of both of these neurotransmitters using in vivo microdialysis in the striatum [59, 82–85]. In these studies, an increase or decrease in the relative density of nerve terminal glutamate immunogold labeling was associated with a decrease or increase, respectively, in the extracellular levels of glutamate. Therefore, in the current study, we hypothesize that initially following the TBI, there is increased release of glutamate, eventually resulting in a long-term depletion of this neurotransmitter that is maintained out to 7 days post-TBI, the time point used in this study. Future studies using microdialysis or biosensors, followed by ultrastructural analysis of the glutamate-labeled nerve terminals, are warranted to confirm that real-time fluctuations in glutamate immediately after TBI with and without BCAA supplementation reflect our ultrastructural findings. Furthermore, our results do not preclude the possibility that there may be other relevant physiological changes, such as alterations in the intrinsic excitability of orexin neurons after TBI. In addition, our study, while adequately powered to detect significant group differences, still only sampled at a single time point 1 week postinjury. This time point was chosen due to the experimental design which mirrored previous work demonstrating EEG-based sleep–wake disruption at this time point [39]. Finally, although the present study does not include a sham/noninjured control group with BCAA supplementation, we have previously observed that BCAA administration in noninjured sham controls has no effect on sleep–wake staging (Lim et al., unpublished observations), nor does it affect contextual fear conditioning [39, 44, 81, 86]; therefore, this additional control group was omitted from the present study design. Mechanistic implications are limited by the above points.

Conclusions

TBI decreases the density of glutamate immunogold labeling within nerve terminals making an excitatory/asymmetric synaptic contact onto orexin neurons, and this is restored back to preinjury levels by dietary BCAA supplementation. These results extend the findings by Lim et al. [39] and suggest a possible mechanism by which TBI compromises orexin neuron function and causes sleep–wake disturbances, while highlighting a potential therapy that restores immunogold labeled glutamate to orexin neurons.

Funding

This material is the result of work supported with resources and the use of facilities at the VA Portland Health Care System, the VA Career Development Award #IK2 BX002712, the American Sleep Medicine Foundation (#105-JF-14, #110-BK-14), and the Portland VA Research Foundation (to M.M.L.); VA Merit Review #I01 BX001643A (to C.K.M.); and National Institutes of Health T32 AT 002688 (to J.E.E.) The contents do not represent the views of the U.S. Department of Veterans Affairs or the U.S. Government.

Notes

Conflicts of interest statement. None declared.