Mild Repetitive TBI Reduces Brain-Derived Neurotrophic Factor (BDNF) in the Substantia Nigra and Hippocampus: A Preclinical Model for Testing BDNF-Targeted Therapeutics

Abstract

Clinical studies have consistently shown that neurodegenerative diseases (NDs) such as Parkinson’s disease, Alzheimer’s disease, Amyotrophic Lateral Sclerosis, and Huntington’s disease show absent or low levels of brain-derived neurotrophic factor (BDNF). Despite this relationship between BDNF and ND, only a few ND animal models have been able to recapitulate the low BDNF state, thereby hindering research into the therapeutic targeting of this important neurotrophic factor. In order to address this unmet need, we sought to develop a reproducible model of BDNF reduction by inducing traumatic brain injury (TBI) using a closed head momentum exchange injury model in mature 9-month-old male and female rats. Head impacts were repetitive and varied in intensity from mild to severe. BDNF levels, as assessed by ELISA, were significantly reduced in the hippocampus of both males and females as well as in the substantia nigra of males 12 days after mild TBI. However, we observed significant sexual dimorphism in multiple sequelae, including magnetic resonance imaging-determined vasogenic edema, astrogliosis (GFAP-activation), and microgliosis (Iba1 activation). This study provides an opportunity to investigate the mechanism of BDNF reduction in rodent models and provides a reliable paradigm to test BDNF-targeted therapeutics for the treatment of ND.

1. Introduction

Brain-derived neurotrophic factor (BDNF) and other neurotrophins have been extensively studied in neurodegenerative diseases (NDs). By regulating synaptic plasticity and neurotransmitter levels, BDNF and other neurotrophins enhance the development and survival of dopaminergic neurons (Baquet et al., 2005; Failla et al., 2016). NDs like Parkinson’s disease, Alzheimer’s disease, Amyotrophic Lateral Sclerosis, and Huntington’s disease have repeatedly shown an association between BDNF levels and symptoms of these diseases in postmortem and age-matched studies (Azman and Zakaria, 2022; Baquet et al., 2005; Erickson et al., 2010; Tremolizzo et al., 2016). BDNF mRNA and BDNF protein are reduced in the hippocampus and neocortex of the human brain in Alzheimer’s disease (Azman and Zakaria, 2022). In addition, the serum, stomach, and brain BDNF levels are low in patients with Parkinson’s disease. Similarly, Huntington’s patients have lower BDNF levels in the brainstem, striatum, prefrontal cortex, and thalamic afferents, leading to striatal neurodegeneration (Azman and Zakaria, 2022). BDNF deficiency, therefore, appears to be a common feature of many NDs.

Although decreased BDNF levels have been documented in several clinical NDs, only a few ND animal models have been able to recapitulate the low BDNF state. Furthermore, the extent of BDNF reduction in rodent ND models remains poorly elucidated. Constitutive homogenous knockout (KO) or null BDNF mutation (Bdnf ^−/−) mice show poor postnatal health and premature death within 21 days of birth (Ernfors et al., 1995). This postnatal mortality can be avoided using conditional and heterozygous BDNF KO (Bdnf ^+/−) mice (Ito et al., 2011; Lindholm and Castrén, 2014). Conditional BDNF KO in early postnatal development alters the midbrain dopaminergic system, characterized by a reduction in tyrosine hydroxylase expression in a subset of neurons in the substantiate nigra (SN) and a decrease in motor coordination (Baquet et al., 2005). Inducible BDNF KO in adult mice confined to the forebrain and hippocampus impairs cognitive and emotional behavior (Monteggia et al., 2004). Thus it would appear the neurobiological consequences of reduced BDNF activity are dependent upon different stages of development making it difficult to assign a causative or complimentary role in disease progression for any of the NDs mentioned above. (Chourbaji et al., 2011). Despite the association between BDNF and age-related NDs in humans, there is little data recapitulating the findings of BDNF depletion in rodent ND models. Animal models reliably demonstrating this deficit in BDNF are needed to evaluate the neurodegenerative pathophysiology and the therapeutic efficacy of test candidates. To that end, we developed a method for delivering BDNF to the brain through a nasal portal (Padmakumar et al., 2021). Hence, there is a need for a reproducible, robust model of BDNF depletion.

One promising model is the application of controlled traumatic brain injury (TBI). BDNF-related therapies have shown neuroprotection and increased neuroplasticity, reversing brain injury-related deficits such as those resulting from TBI (Wurzelmann et al., 2017). Repeated mild TBI has been shown to increase the risk of chronic neurodegeneration and neuropsychiatric diseases (Faden and Loane, 2015). With the exception of behavioral studies, most rodent research studies involving TBI and conditional BDNF knockout studies do not exceed more than 3–6 months of age. However, the mortality rate of older adults after severe TBI is higher than that of younger adults (Failla et al., 2016). Therefore the effects of TBI may be particularly severe in the otherwise normally aging brain. Hence, there is a need for a preclinical model of BDNF deficiency that reflects the human experience at any age and under various degrees of head injury. In this study, we hypothesized that two or three head impacts at different velocities classified as mild, moderate or severe would be a reliable method to induce a consistent BDNF decrement in the nine-month-old rat brain. We further analyzed these effects by sex in order to elucidate any potential sexual dimorphism which may exist as a function of sex-determined morphologic or physiologic susceptibilities to head injury.

To confirm this hypothesis, we inflicted a close-head TBI in rats, a common injury found in humans. We then investigated the pathobiochemical effects of repeated TBI on BDNF and associated biomarkers in socially mature, healthy-aged, nine-month-old Long Evans rats. A non-invasive TBI model in rats (Kulkarni et al., 2019) was replicated by simulating the collision mechanics of National Football League players (Viano et al., 2009). The effects of these impacts were assessed using magnetic resonance imaging, biochemical, and immunological biomarkers.

2. Materials and methods

2.1. Materials

Type B Gelatin, with a bloom strength of roughly 225 was purchased from Sigma, St. Louis, MO. Poly (epsilon-caprolactone) of molecular weight of 50 kDa (PCL) was bought from Polysciences Inc. (Warrington, PA). 2,2,2-trifluoroethanol (2,2,2-TFE) was bought from Acro Organics (China). Protein quantification was performed by Pierce BCA assay kit procured from Thermo Fisher Scientific (Waltham, MA). 4% v/v paraformaldehyde in phosphate-buffered saline and optimal cutting temperature (OCT) embedding compound were obtained from Fisher Scientific (Fair Lawn, NJ).

2.2. Methods

2.2.1. Construction of biodegradable osmotic core-shell bi-component implants

The construction of osmotic core-shell implants was executed step-by-step, as described in (Padmakumar et al., 2021) with modification. To fabricate our bi-component core-shell biodegradable implant, the shell and the core were made of polycaprolactone (PCL) and gelatin polymer, respectively.

For dissolution of gelatin, weighed amount of gelatin (16.7% w/v) was first soaked in suitable amounts of cold phosphate-buffered saline, pH 7.4, for 10–15 minutes, followed by warming the hydrated gelatin between 35°C −37°C by intermittent vortex mixing. After complete dissolution, warm gelatin solution was poured on the semi-sphere round silicone mold (food grade, BPA free). The semi-sphere in the silicone mold has a diameter of 4 mm in diameter and a depth of 4 mm. The mold is left at room temperature for about 15 minutes. After cooling, the gelatin beads were carefully separated from the mold. Preparation and storage of the gelatin cores were performed in aseptic conditions. Further assembly of the core-shell implant was completed as described in (Padmakumar et al., 2021), except for the core, wherein a gelatin core was introduced into each shell using forceps.

2.2.2. In vivo studies

All adult male and female Long Evans rats ca. 9 months of age were purchased from Charles River Laboratories (Wilmington, MA, USA) conventionally housed, i.e., 12 h:12 h light/dark cycle (morning lights on 07:00 am) at room temperature (22–24°C) with free access to water and food. For the study, adult male and female rats, nine months of age, were used. All animals were cared for in accordance with the NIH Guide to the Care and Use of Laboratory Animals. Methods and procedures used in this study were pre-approved by the Northeastern University Institutional Animal Care and Use Committee, protocol # 20-0209. The protocols used in this study followed the ARRIVE guidelines for reporting in vivo experiments in animal research (Kilkenny et al., 2010).

2.2.2.1. Minimally Invasive Nasal Depot (MIND) surgical procedure

To establish the TBI-induced BDNF deficient model that would later be used to restore BDNF levels through MIND, it was necessary to prepare all experimental rats with the nasal portal as part of the experimental design. The experimental setup for MIND and placing the implants in the rats was developed in our laboratory (Padmakumar et al., 2021) and is shown in Fig. 1A,B. The rats were briefly subjected to anesthesia with 3% isoflurane in oxygen (Vet Equip, item number 901801), and the rat’s head was securely fixed in a prone position to the stereotactic frame (Kopf stereotaxic frame, Model 963) with the aid of ear bars. The incision area at the snout was shaved and disinfected with isopropyl alcohol, followed by povidone-iodine. A sagittal midline incision was made over the snout, and the skin was pulled back (retracted) to reveal the nasal bones. With a high-speed microdrill (Dremel model 8050-N/18, Mt. Prospect, IL), the nasal bones were removed without damaging the basolateral olfactory mucoperiosteum. In this way, a subcutaneous cavity is created with the basolateral olfactory epithelium as the base. The implant was placed in this cavity such that the implant surface is in direct contact with olfactory epithelium. A 5-O nylon surgical suture (Med-Vet International, Mettawa, IL) was used to close the skin incision.

Fig. 1. A schematic illustration showing different experimental procedures, set-up, and design.

(A) Rat set up for MIND surgery (Created with BioRender.com). (B) Implantation of biodegradable gelatin core-PCL shell implant into MIND-generated subcutaneous cavity at the nasal site of rats. (C) Closed head, momentum exchange model used for inducing TBI in rats. The rats receiving multiple hits were spaced 48 h apart. Rats were sacrificed either 24 h after the last brain injury, i.e., on the 6^th day (D, F), or after one week, i.e., on the 12^th day (E, G) after the first brain injury. Rats were scanned for brain neuroanatomy by MRI, and harvested brains were used either for BDNF analysis by ELISA or immunofluorescent staining of biomarkers.

2.2.2.2. Traumatic Brain Injury (TBI) induction

Animals that had MIND surgery were randomized into groups to receive repeated TBI and sham controls. A closed head momentum exchange model of injury was followed to induce head impacts for TBI, as described previously (Cai et al., 2021; Kulkarni et al., 2020; Kulkarni et al., 2019). Briefly, a lightly anesthetized animal (2% isoflurane, until nonresponsive to toe pinch) was placed prone on a Teflon board with the head facing the impactor (Fig. 1C). The compactor piston was directed to the midline on the top of the skull near the Bregma. Using a pneumatic barrel, a 50 g compactor was accelerated at 7.4-, 9.3- and 11.2 m/s. These velocities were described as concussions (mild), medium (average), and severe (elite) TBI, respectively (Viano et al., 2009). The compactor piston theoretically produced an impact of 1.37, 2.16, and 3.14 joules, respectively, at the site of impact and rotated the rat slightly in a linear motion. The rats were subject to multiple hits, as shown in Fig. 1D. Animals with skull fractures were excluded. All head-impacted rats were returned to their cage after regaining consciousness.

2.2.3. Determination of BDNF concentrations by ELISA

For ELISA, harvested brains were immediately stored at −80°C to determine BDNF. Olfactory bulb (OB), striatum (STR), hippocampus (HC), frontal cortex (FC), substantia nigra (SN), and cerebellum (CB) regions were retrieved from the frozen brain using biopsy punches with a diameter of 3mm. Different brain regions were homogenized in ice-cold phosphate buffer saline (pH 7.4) supplemented with 1% EDTA-free protease and phosphatase inhibitor cocktail (100X, Thermo Scientific Cat # 78441) using BeadBlaster^™ refrigerated microtube homogenizer having grinding balls (Zirconium, 1.5mm), (Southern Labware, Cat # 24R, D2400-R) for 2 min at 30s on/30s off. The supernatants were collected carefully after centrifugation of the homogenates for 30 min at 20,000g at 4°C and stored at −80°C. Collected supernatants were used for analysis within one month of extraction. The protein contents in the supernatants were measured using Pierce^™ BCA protein assay kit (Thermo Scientific, Cat # 23225).

We used the Rat BDNF Sandwich ELISA kits (MyBioSource, Cat # MBS355345) following the manufacturer’s instructions to quantify BDNF levels in brain regions. BDNF protein concentration values were subsequently normalized based on the total protein content in each sample and expressed as pg of BDNF per mg protein.

2.2.4. Image acquisition by Magnetic Resonance Imaging (MRI)

MRI was performed with a Bruker Biospec 7.0T/20-cm USR horizontal bore magnet (Bruker, Billerica, MA, USA) with Para Vision ^™ 6.0.1 software. The detailed description of the methodology is explained in (Kulkarni et al., 2015). Radiofrequency signals were both transmitted and received through a quadrature volume coil integrated into an animal restrainer provided by Ekam Imaging Inc. (Boston, MA). This restrainer system featured a padded headrest, eliminating the need for ear bars, reducing animal discomfort, and minimizing motion artifacts. During imaging sessions, all rats were maintained under 1–2% isoflurane anesthesia while their respiratory rate was kept at 40–50 breaths per min. At the start of each imaging session, a high-resolution anatomical dataset was acquired using a T2-weighted RARE pulse sequence with the following parameters: 30 slices, each 1 mm thick; a field of view (FOV) measuring 3 cm × 3 cm; matrix size of 256 × 256; repetition time (TR) of 3000 ms; effective echo time (TE) of 36 ms; Number of averages (NEX) set at 4; and an acquisition time of 6 min 24s.

2.2.5. Diffusion Weighted Imaging (DWI) for quantitative anisotropy

DWI was performed utilizing a spin-echo echo-planar-imaging (EPI) pulse sequence with the following parameters: Repetition time (TR)/Echo time (TE) = 500/24.47 ms, eight EPI segments, and 10 non-collinear gradient directions employing a single B-value shell at 1000 s/mm² along with one image possessing a B-value of 0 s/mm² (commonly referred to as B0). The geometrical specifications consisted of 34 coronal slices, each with a thickness of 0.344 mm (comprising the brain volume) and with an in-plane resolution of 0.344 × 0.344 mm² (matrix size 96 × 96; FOV 33 mm²). Each Diffusion Tensor Imaging (DTI) imaging took approximately 50 min, contributing to an overall MRI protocol duration of approximately 1 h 15 min. To mitigate potential motion artifacts stemming from sporadic breathing during DWI acquisition, each image (for each slice and each gradient direction) underwent scrutiny for motion artifacts before DWI analysis. Any images affected by motion artifacts were excluded from further analysis.

The DW-3D-EPI images were subjected to image analysis to generate apparent diffusion coefficient (ADC) maps. The DWI analysis was executed using Matlab^© (Mathworks, USA) and MedINRIA software (1.9.0; http://www-sop.inria.fr/asclepios/software/MedINRIA/index.php) following the procedures described in (Kulkarni et al., 2015). Subsequently, each brain volume, including the B0 image, was registered to the 3D MRI Rat Brain Atlas^© which features annotated brain regions and 171 segments utilizing EVA (Ekam Visualization and Analysis, Ekam Solutions LLC, Boston, MA) software. This registration employed a nine-parameter affine transformation, enabling voxel- and region-based statistical analyses.

2.2.6. Immunofluorescence staining

On day 12, deeply anesthetized rats from the sham control and hit impact injury groups were subjected to transcardial perfusion, first with PBS and subsequently with a 4% paraformaldehyde solution in PBS (PFA, Thermo Scientific Chemicals, Cat # J19943.K2). Brains were harvested from animals and post-fixed for 48 h in PFA at 4°C. After fixation, the brains were cryoprotected in a series of sucrose gradients, starting with 15% and then transitioning to 30% sucrose. Subsequently, the cryoprotected brains were embedded in OCT (optimal cutting temperature) embedding compound, and the block was snap-frozen in liquid nitrogen. OCT blocks were cut into 35-μm coronal sections and fixed on pre-coated slides.

For immunofluorescence staining, the brain sections were first hydrated in wash buffer, followed by blocking with 6% goat serum in PBS for one hour at room temperature. Following the blocking step, the sections were incubated overnight at 4°C with primary antibodies. Specifically, primary antibodies used included anti- Ionized calcium Binding Adaptor molecule 1 (Iba1) (1:100, Invitrogen, Cat # MA5-36257) and anti-glial Fibrillary Acidic Protein (GFAP) (1:100, Invitrogen, Cat # MA5-35237), both in a solution of 1% goat serum and 0.3% Triton X-100 in 0.1% Tween infused PBS (PBS-T). After overnight incubation, the sections were thoroughly washed with PBS-T. Following washing, the sections were exposed to a secondary antibody conjugated to Alexa Fluor 488/594 (1:5000, goat anti-rabbit IgG, Alexa Fluor 488) at room temperature for one hour, followed by another round of PBS-T washing. Negative controls were also prepared simultaneously to detect the non-specific binding of the secondary antibody. Finally, the immunostained sections were mounted with Vectashield Mounting Medium containing 4′,6-diamidino-2-phenylindole (DAPI) for nuclear staining (Vector Laboratories). All images were captured using a Zeiss LSM 800 inverted fluorescence microscopy (Carl Zeiss Microscopy GmbH, Germany) with a 20 × objective. For each group, multiple sections from 3 rats were analyzed. For quantitative analysis, tissue sections were evaluated at 40 × magnification. The regions of interest (ROIs) were defined with the help of rat atlas (Paxinos and Watson, 1997). The immunoreactive protein expression was calculated by the mean immunofluorescence signal intensity using ImageJ software (ImageJ 1.4, NIH, Bethesda, MD, USA) as described in (Shihan et al., 2021). Data was collected from 2 slices per rat, and each slice was observed for a minimum of 10 regions. Each measurement represented the average of all region measurements for the three rats in the group.

2.2.7. Statistical analysis

All calculations related to ADC and anisotropy were carried out using MATLAB (MathWorks, Natick, MA, USA). Data for BDNF levels and ADC were collected from individual rats and subjected to analysis using GraphPad Prism^® software Version 8.0 (GraphPad Software Inc., San Diego, CA, USA). One-way ANOVA, post hoc Sidak’s multiple comparison tests or Kruskal-Wallis test and uncorrected Dunn’s post-hoc were used to compare the groups for immunofluorescence and BDNF levels, respectively. Statistical significance was established at p < 0.05.

3. Results

3.1. Female and male rats respond differently to repeated TBI

T2-weighted anatomical sections of the rat brains encompassing the hit region are presented in Fig. 2A–L. The anatomical information facilitated the visualization of anomalies, cortical lesions, and concussions/contusions in the injured rats subjected to multiple TBIs.

Fig. 2. TBI induction in 9-month-old Long Evans rats.

Coronal MRI images following head impacts with representative T2-weighted anatomical brain images collected using RARE pulse sequences in males of sham-operated (A, B) and hit injured rats exposed to different velocity impacts at 7.4 m/s (E, F) and 9.3 m/s (I, J) and age-matched females (C, D, G, H, K, L) with 50 g impactor on day 6 and day 12 after the first injury respectively. The impact site is with a high-intensity signal identified by a red arrow (n=3).

No structural brain damage or skull fracture was observed in rats subjected to repeated 3x of mild TBI or 2x of moderate hits. The increased signal intensity on the tissue above the skull (impact site) represents edema and is visible in most images on day 6. However, any vascular damage to the skin decreased with time, and thus, the high-intensity signal decreased or was almost absent with 12 days of survival.

Mortality was observed when female rats were subjected to repeated injury 3x of moderate hits. Similarly, female rats could only withstand a 1x TBI when the impact speed was increased to 11.2 m/s (severe hits), and the number of deaths increased with the number of hits and skull fractures. However, no mortality was observed in male rats up to 2x of severe hits. Higher impact velocities were more likely to cause skull damage in females (Supplementary Fig. 1). At 3x of moderate hits impacts, 33% of mortality was observed in females and reached 50% at 2x of severe hits. Because of the mortality observed at high-impact velocity and the increased number of hits, we limited the study to 3x of mild hits, and 2x of moderate hits in both sexes.

In general, any head can withstand physical stress or force without damaging the brain. However, if the force increases beyond the head’s capacity, a TBI occurs (Semple and Panagiotopoulou, 2023). Adult Long Evans males had larger body sizes (p < 0.05) than age-matched females. Males were larger in most parameters of the skull, such as thickness in the parietal, frontal, and occipital regions (Supplementary Fig. 2). A thicker skull increases brain protection and reduces bone deformation on impact (Semple and Panagiotopoulou, 2023), thereby changing the mechanics of the TBI. A morphometric analysis of our study rats confirmed that the heterogeneity of the skull was also dependent on age and sex. Thus, male rats were able to withstand repeated collisions at higher-impact velocities compared to their female counterparts. Despite the significant degree of brain maturation in rats of this age, aging can also be accompanied by altered cranial bone stiffness. The sexual dimorphism of bone resorption further increases the risk of skull fractures in female rats during the estrous cycle (de Bakker et al., 2018). Dubal et al., 2001 and Frick et al., 2009 reported irregular/declining reproductive cycles with persistent estrus in middle-aged nine-month female rats (Dubal and Wise, 2001; Frick, 2009).

3.2. BDNF protein expression after repeated TBI

Next, we evaluated the impact of repeated TBI on BDNF protein levels in rats. Changes in BDNF levels were quantified by ELISA, as depicted in Fig. 3A, B, Supplementary Fig. 3–4 for male and female rats. BDNF was widely distributed across various brain regions.

Fig. 3. BDNF expression in various brain regions following TBI.

Histograms showing BDNF levels in distinct brain regions on day 12 for both sham control and injured male and female rats. BDNF levels in the striatum, frontal cortex, hippocampus, substantia nigra (Sub nigra), and cerebellum were measured by ELISA. Each vertical bar represents the mean ± standard deviation, calculated from data obtained from 3 rats per group. Each dot represents one rat. Significance denoted as *p < 0.05, **p < 0.001 determined using Kruskal-Wallis and uncorrected Dunn’s post-hoc test.

Hippocampus

Among various brain regions, the intact hippocampus is rich in BDNF (Katoh-Semba et al., 1997; Zhou et al., 1996). Innervation of the hippocampus by serotonin fibers originates from the raphe nuclei of the midbrain. In the mild TBI group, BDNF levels decreased on day 12 in both males and females (p < 0.05). Specifically, compared to the sham injured, the BDNF reduction was almost 2.6-fold in males and in females, 2.3-fold. Moreover, a significant reduction in hippocampal BDNF expression was observed in both sexes subjected to moderate hits (p < 0.05).

Dopaminergic Regions

BDNF protein seemed to decrease by day 12 in the substantia nigra of both males and females after mild hits. However, the reduction was significant in males, with a 1.8-fold decrease (p < 0.05) and non-significant in females, although it showed a decreasing trend.

Mild hits did not alter the striatal BDNF levels in males and females. However, a decrease in striatal BDNF protein levels could be observed in females with moderate hits.

Cerebellum and Olfactory Bulb

Cerebellar granule cells synthesize BDNF (Carter et al., 2002). However, no changes in BDNF levels were observed in the cerebellar and frontal cortex of both male and female rats with mild to moderate head injury. TBIs are usually correlated with olfactory impairment, mainly due to rupture or tearing of the olfactory nerve. BDNF gene expression is also high in the olfactory bulb, which is part of the meso-telencephalic projection neurons that support dopaminergic neurons (Zhou et al., 1996). The regeneration potential of olfactory neurons (Gustafsson et al., 2021) can be attributed to no change in BDNF levels (Supplementary Fig. 4).

3.3. TBI alters the vasogenic and cytogenic edema in local hit areas.

Quantitative anisotropy was utilized to detect alterations in the microarchitecture of gray matter. The apparent diffusion coefficient (ADC) measures water dynamics in hydrated tissues, thus reflecting cytotoxic swelling, vasogenic swelling, and white matter integrity. Cytotoxic edema is the influx of water into the intracellular compartment from the extracellular compartment, which reduces ADC signals. Conversely, vasogenic edema extravasates fluid into the extracellular space, increasing ADC signals (Van Putten et al.). TBI typically leads to a mixture of both vasogenic and cytotoxic edema (Citton et al., 2012).

Considering bregma as the impact location, we focused on regions proximal to the impact to obtain anisotropic values, including the hippocampus, cortex, and basal ganglia. The combined ADC values regions after DWI computational analysis were summed up to obtain an average ADC for the entire region. The ROIs selected for each region are shown in Supplementary Table 1.

On day 6, significant differences were observed between hippocampal and basal ganglia ADCs in females exposed to mild hits and their age- and sex-matched shams (p < 0.05) (Fig. 4B). Conversely, males with mild hits showed no changes in ADC for any of the regions compared to their controls (Fig. 4A). As the impact velocity increased (e.g., 2x, 9.3 m/s), no change in ADC was observed in any regions of injured male and female rats on day 6.

Fig. 4. Anisotropy changes in TBI hit rats.

Apparent diffusion coefficient (ADC) values for male and female rats subjected to TBI and sham-operated on day 6 (A, B) and day 12 (C, D), respectively. The impact intensities were 3x, 7.4 m/s or 2x, 9.3 m/s hits, encompassing regions such as the hippocampus, basal ganglia, and cortex. Values are shown as mean ± standard deviation, calculated from data obtained from 3 rats per group. Each dot represents one rat. Significance is denoted as *p < 0.05 versus sham-injured determined using one-way ANOVA post-hoc Sidak’s multiple comparison.

On day 12, there were no differences in ADC values for any regions in the injured and sham controls (Fig. 4C, D) for males and females subjected to mild hits. Further female rats subjected to moderate impacts showed increased hippocampal ADC values compared to sham controls. However, the males had a reduced ADC trend in the hippocampus region compared to the sham control.

3.4. Biochemical markers of TBI and inflammation

Reactive astrocytes stained by GFAP

Reactive astrocytes were evaluated through the staining of the intermediate filament protein, glial fibrillary acidic protein (GFAP), which serves as an activation marker (Okonkwo et al., 2013). Reactive gliosis, particularly astrogliosis, represents an early occurrence following TBI. Immunofluorescence staining was performed on day 12 because the BDNF levels were significantly reduced on day 12 compared to day 6.

GFAP immunostaining in astrocytes observed in brain regions of injured rats are shown in Fig. 5A,B,C. Weakly stained cells and fine processes were observed in sham-control rats. In females, mild hits appeared to affect distant regions, midbrain, and deep brain regions, such as the striatum, hippocampus, substantia nigra and cerebellum with strong GFAP immunoreactivity fluorescence (p < 0.001) compared to their sham controls. At an increased projectile velocity of 2x, 9.3 m/s, there was high GFAP immunoreactivity in all the regions (p < 0.001) compared to sham-injured. Activated astrocytes displayed hypertrophy, characterized by an increase in the area of astrocyte bodies and the thickness of their processes). Conversely, only an increase in swelling of the processes was observed with an increase in projectile velocity (indicated by white arrows in Fig. 5A). Diffuse TBI has been reported to induce mild to moderate gliosis with hypertrophied astrocytes (Cikriklar et al., 2016). Hypertrophic astrocytes are associated with interneuronal damage and abnormal neurogenesis (Hamberger et al., 2009) or brain edema (Li et al., 2009). Hypertrophic astrocytes are sought after as healers after tissue injury, usually surrounding the primary lesions.

Fig. 5. Glial fibrillary acidic protein (GFAP) immunostaining for determining astrocyte damage following repeated TBI.

(A) Representative photomicrographs captured using a fluorescent microscope illustrating Glial fibrillary acidic protein (GFAP) immunoreactivity in various brain regions of males and females, both sham control and injured rats. GFAP-positive astrocytes are shown in green (Alexa Fluor 488), with nuclei in blue (DAPI); scale bar, 50 μm. White arrows identify hypertrophic astrocytes. Graphical representation of mean immunofluorescence intensity counts in each brain region of (B) males and (C) females, both sham control and injured rats. Each vertical bar represents the mean immunofluorescence signal intensity/mm² ± standard deviation obtained from different regions and slices from 3 rats per group. Significance is depicted as **p < 0.001, ***p < 0.005 compared to the control group, determined using an ordinary one-way ANOVA post-hoc Sidak’s multiple comparison.

Interestingly, increased gliosis could also be observed in male hit rats in hippocampus, substantia nigra and cerebellum (p < 0.001) of those subjected to mild hits compared to sham control. Activated astrocytes were also hypertrophic in mild impacts. Conversely, an increase in projectile velocity did not show an in fluorescence of GFAP immunoreactivity with a reduced intensity of hypertrophic astrocytes and a more amoeboid-like activated cell shape. However, neither diffused GFAP activation nor gliotic scarring was observed in either sex.

Reactive microglia stained by Iba1

During inflammation, microglia migrate to the injury site and remove damaged cells to restore a normal and protective environment. Highly activated microglia release reactive oxygen species, proinflammatory cytokines, proinflammatory cytokines, and excitatory glutamate neurotransmitters, which are harmful and neurotoxic (Donat et al., 2017). The presence of microglia activation, as shown by Iba1 immunoreactive staining in various brain regions after TBI, indicates the presence of inflammation (Fig. 6A).

Fig. 6. Ionized calcium-binding adaptor molecule-1 (Iba-1) immunostaining for determining microglia activation following repeated TBI.

(A) Representative photomicrographs captured with a fluorescent microscope illustrating Iba-1 immunoreactivity in various brain regions of males and females, both sham control and injured rats. Iba-1-positive astrocytes are visualized in green (Alexa Fluor 488), with nuclei in blue (DAPI); scale bar, 50 μm. White arrows identify hyper-ramified microglia. Graphical representation of mean immunofluorescence intensity counts in each brain region of (B) males and (C) females, both sham control and injured rats. Each vertical bar represents the mean immunofluorescence signal intensity/mm² ± standard deviation obtained from different regions and slices from 3 rats per group. Significance is depicted as *p < 0.05, **p < 0.001, ***p < 0.005 compared to the control group, determined using an ordinary one-way ANOVA post-hoc Sidak’s multiple comparison.

In our study, injured females had more microglia Iba1-immunoreactivity than males (Fig. 6B and C). Females exposed to mild hits showed increased numbers of activated microglial cells in the hippocampus, striatum, substantia nigra and cerebellum compared to the region-specific sham controls (p < 0.05). Microglia from mild hit female rats showed more hyper-ramification (increased branching points of processes indicated by white arrows in Fig. 6A) in the striatum, hippocampus and substantia nigra. However, males showed more toward a primed ramification in the striatum and hippocampus with changes in Iba1-immunoreactivity.

Females exposed to moderate hits also showed increased Iba1 immunoreactivity and a hyper-ramified morphology in the hippocampus, substantia nigra, and cerebellum (Fig. 6A,C). Microglia from males exposed to higher velocity also increased Iba1 immunoreactivity in the striatum, substantia nigra and cerebellum but were stained with amoeboid morphology. Some areas in sham animals showed fluorescence of resting microglia (bodies showing little branching of their processes). This observation supports that age is associated with increased activated glia (Zakoliukina et al., 2019).

4. Discussion

We have developed and characterized a rat model of reduced BDNF levels after closed-head mild TBI. The purpose of creating this model is to establish a reliable rodent model of BDNF reduction to enable more robust studies of both the mechanism of BDNF decrement in ND disease and the benefits of potential BDNF-targeted neurotherapeutics. Our research shows that males and females of the same age exhibit sexually dimorphic sequelae as a function of impact velocity and number (Fig. 2). Many studies have reported that females have greater severity, number, and/or duration of post-concussion traumatic symptoms than males in humans (Bazarian et al., 1999; Eyolfson et al., 2022). In addition to differences in neuroprotective effects between males and females, there are other plausible reasons for this difference that are not directly related to neuroprotection. Gender and age can influence morphometry and physical injury (Supplementary Fig. 1). The cranial mass of a female rat skull appears slightly smaller than that of male rats according to bone morphometry, which complicates biomechanical studies. Such differences make females more susceptible to brain injury than males. In the absence of structural brain damage, critical neuropathological outcomes remained similar, but their magnitude differed by gender. Lesions detected by immunoreactivity were evident at a greater rate in females than in males.

The early and transient increase in endogenous BDNF at day 6 (Supplementary Fig. 3, upward trend but not significant) is likely a compensatory response to acute stress or brain injury to protect neurons (Khalin et al., 2016). However, the increased susceptibility of females to brain injury resulted in an increase in vasogenic swelling near the affected brain regions (hippocampus and basal ganglia) compared to males, where there is no change in ADC in nearby affected brain regions on day 6 (Fig. 4A, B). The increase in ADC is due to the increased diffusion of water to the site of injury, which is probably due to partial compromise in blood-brain barrier integrity, hypoxia, cell loss, and cystic formation. Surprisingly, subjecting the rats to moderate hits did not significantly affect the ADC of females and males by day 6.

On day 12, mildly concussed rats showed significantly reduced BDNF levels in the hippocampus of both sexes and additionally in the substantia nigra of males (Fig. 3A, B). BDNF levels were low even in moderately concussed rats, mainly in the hippocampus of males and females (Fig. 3A) and additionally in the striatum of injured females (Fig. 3B). Decreased BDNF levels in the hippocampus compared to sham-injured, a region often damaged in TBI, likely indicates neuronal loss. Astrogliosis and microgliosis were observed in the hippocampus of the injured males and females with mild and moderate hits; however, at moderate hits, microgliosis was persistent only in the hippocampus of the females. Males failed to show microgliosis at moderate hits. The low BDNF levels in the substantia nigra of mildly concussed males may have been due to persistent astrogliosis on day 12. On day 12, astrogliosis and microgliosis were observed in the substantia nigra of females with mild hits; the fall in BDNF levels was not significant. The substantia nigral neurons in the substantia nigra pars compacta express BDNF mRNA (Altar et al., 1997). Astrogliosis may have caused the decreased BDNF secretion. Thus, increased astrogliosis in the hippocampus and substantia nigra coincides with BDNF levels on day 12 rather than microgliosis. The ADC value of rats with mild hits returned to sham levels or showed reduced vasogenic edema in both sexes by day 12 (Fig. 4C, D). The reduction of vasogenic edema in males can also be supported by the fact that no significant microgliosis or inflammation was observed on day 12 compared to sham control (Fig. 5B, 6B). The females, though, showed a decrease in ADC on day 12, astrogliosis was visible in all brain regions in both mild and moderate hits (except striatum from moderate hits), but microgliosis was limited to the hippocampus, substantia nigra and cerebellum in those with moderate hits. A reduction in ADC also occurs in cytogenic edema, with the possibility of focal lesions due to local devascularization. That may lead to increased astrogliosis and microgliosis despite reduced ADC in females with mild hits. In contrast, in females with moderate hits, the secondary pathological causes may have developed over time with vasogenic edema resulting in increased ADC levels.

Although astrogliosis and microglial inflammation were observed in the cerebellum of males and females (Fig. 5B,5C,6B,6C), these could not be correlated with the observed BDNF levels (Fig. 3B). There is a strong possibility that the cerebellum responds to acute stress to repair and protect neurons. Cerebellar granule cells synthesize BDNF (Carter et al., 2002) with high expression of the TrkB receptor and BDNF (Camuso et al., 2022) and, thus, the variation. Likewise, BDNF levels in females were also unchanged in the striatum at mild hits. However, astrogliosis and microgliosis were elevated in the striatum of the injured females (Fig. 5C,6C). The striatum generally relies on substantia nigra and corticostriatal input (cerebral cortex, amygdala, and thalamus) for the supply of BDNF (Baydyuk and Xu, 2014; Li et al., 2012), which may have been attributed to unchanged BDNF levels. The females showed decreased BDNF levels at moderate hits which may be consistent with the observation of astogliosis and absence of microgliosis.

During secondary inflammation, it is somewhat difficult to interpret at which stage gliosis produces proinflammatory and anti-inflammatory factors and BDNF (Zakoliukina et al., 2019). The ratio of healing glia to destructive glia varies with age. Zakoliukina et al., 2019, observed strong BDNF secretion in young rats compared to the old rats, which had a shift towards free-radical generation, toxic to neurons (Zakoliukina et al., 2019). The group also demonstrated that BDNF secretion was more associated with astrocytes than with microglia in the lipopolysaccharide aged-rat model of neuroinflammation. In our study, the astrogliosis in either sex and their BDNF levels on day 12 were also somewhat related at mild hits but not at moderate velocity hits. This hypothesizes that activated astrocytes appear to play an active role in BDNF secretion under neuroinflammatory conditions in aged rats. We hypothesize that the interference of female hormones may be minimal in our study due to the continous estrous cycle and reduced reproductive cycles typically observed in middle-aged female rats (Dubal and Wise, 2001; Frick, 2009). However, we have not tracked the female reproductive cycles.

In this study, we thus observed a decrease in BDNF levels in male and female Long Evan rats after noninvasive repetitive mild and moderate TBI. Among the different regions, the hippocampus was the first to show altered BDNF levels in either sex, followed by effects on the nigrostriatal pathway. Rats subjected to 3x, 7.4 m/s mild TBI were better and more conclusive on BDNF levels, edema, astrogliosis, and inflammatory patterns in males and females. However, the extent of pathogenesis of inflammation and edema differs between males and females. Females are more susceptible to brain injury due to skull morphometry and may thus exacerbate injuries and respond to minor insults. Repeated mild TBI may be necessary to maintain lower BDNF levels. Rats subjected to moderate impact velocities (9.3 m/s) had somewhat unclear and unpredictable trends, likely due to a secondary cascade of inflammation, further complicating the mechanism of brain injury. However, there is no disagreement regarding gender differences in BDNF levels affecting the hippocampus and the nigrostriatal regions. Furthermore, neurodegenerative disease not only involves neuronal dysfunction or death but also involves dysfunctional non-neuronal cells as drivers or initiators of disease progression (Lobsiger and Cleveland, 2007). Neurodegenerative diseases of old age have reported microgliosis and astrogliosis (Zakoliukina et al., 2019), which were characterized in our rodent model at both velocities.

5. Conclusion

In conclusion, we developed a simple rat model demonstrating reliable decrements in BDNF levels using a TBI stimulus. In the future, this model can be used together with genetic models of ND to study the pathophysiologic contributions and therapeutic implications of neurotrophin derangement in Parkinson’s, Alzheimer’s, and other central nervous system-related diseases. The evident differences we found as a function of sex suggest not only that further investigation is needed into the mechanisms of sex-dependent TBI sequelae but that use of such models to study neurotrophin-mediating ND therapeutics will also need to account for sexual dimorphism when analyzing pharmacokinetic and pharmacodynamic readouts.

Data availability

Data will be available on request.

Highlights.

• BDNF levels reduce by day 12 in a rat model following mild TBI

• Rats exhibited sexually dimorphic pathogenesis in inflammation and MRI

• TBI provides a reliable rodent model to study BDNF-related pathogenesis

• This model is useful in for studying the therapeutic benefits of BDNF in CNS disease