Longitudinal Biochemical and Behavioral Alterations in a Gyrencephalic Model of Blast-Related Mild Traumatic Brain Injury

Blast-related traumatic brain injury (bTBI) is a major cause of neurological disorders in the U.S. military that can adversely impact some civilian populations as well and can lead to lifelong deficits and diminished quality of life. Among these types of injuries, the long-term sequelae are poorly understood because of variability in intensity and number of the blast exposure, as well as the range of subsequent symptoms that can overlap with those resulting from other traumatic events (e.g., post-traumatic stress disorder). Despite the valuable insights that rodent models have provided, there is a growing interest in using injury models using species with neuroanatomical features that more closely resemble the human brain. With this purpose, we established a gyrencephalic model of blast injury in ferrets, which underwent blast exposure applying conditions that closely mimic those associated with primary blast injuries to warfighters. In this study, we evaluated brain biochemical, microstructural, and behavioral profiles after blast exposure using in vivo longitudinal magnetic resonance imaging, histology, and behavioral assessments. In ferrets subjected to blast, the following alterations were found: 1) heightened impulsivity in decision making associated with pre-frontal cortex/amygdalar axis dysfunction; 2) transiently increased glutamate levels that are consistent with earlier findings during subacute stages post-TBI and may be involved in concomitant behavioral deficits; 3) abnormally high brain N-acetylaspartate levels that potentially reveal disrupted lipid synthesis and/or energy metabolism; and 4) dysfunction of pre-frontal cortex/auditory cortex signaling cascades that may reflect similar perturbations underlying secondary psychiatric disorders observed in warfighters after blast exposure.


Introduction
Neuropsychological symptoms in warfighters after exposures to blast have triggered considerable research interest in the pathophysiological manifestations of blast-induced traumatic brain injury (bTBI).Pre-clinical research models of blast are attractive tools to understand the pathogenesis of behavioral changes, identify relevant biomarkers, and characterize the neurobiological underpinnings of blast injury to potentially develop countermeasures to bTBI, including therapeutic interventions.2][3][4] Despite the valuable insights these models have provided, there is a growing interest in identifying models of bTBI in other species with neuroanatomical features that more closely resemble the human brain. 5,6Unlike mice and rats, ferrets are gyrencephalical and bear a number of anatomical and physiological similarities to humans that, along with their size and related practical advantages, point to their great potential as an experimental model for high-throughput translational bTBI research that serves to develop countermeasures to the pathophysiological perturbations observed in warfighters exposed to blast. 5,7,8euroimaging plays a critical role in the acute setting to guide appropriate management based upon the detection of brain injuries that require intervention or further monitoring.0][11][12][13] In an attempt to better diagnose concussions and more subtle brain injuries, many additional advanced neuroimaging techniques are actively being pursued such as diffusion tensor imaging (DTI), which is more sensitive in detecting microstructural tissue damage after mTBI than conventional imaging tools. 14However, DTI has produced mixed findings in mild TBI patients, 15 and a review showed an approximately equal number of studies reporting increased versus decreased fractional anisotropy (FA) in a wide range of white matter (WM) areas during the first 3 months after mTBI. 16[19] Hence, DKI has the potential to provide a distinct microstructural contrast in comparison with the DTI parameters.To date, few studies have investigated DKI in patients with mTBI [20][21][22] and have focused primarily on the most common DKI metrics such as mean kurtosis (MK), axial kurtosis (AK), and radial kurtosis (RK).In these studies, lower MK and increased AK have been reported in WM and thalamus in patients with mTBI compared with healthy controls both in the acute, semiacute (within 1 month), and chronic ( ‡6 months) phases post-injury, and differences in DKI were demonstrated in the absence of DTI abnormalities.In line with this, DKI, but not DTI, changes were associated with hippocampal and cortical astrogliosis in one of our previous rat blast TBI studies. 23These studies support the view that DKI can be more sensitive to the pathology underlying mTBI than DTI.
Magnetic resonance spectroscopy (MRS) allows for in vivo measurement of biochemicals that are undetectable by conventional MRI, thereby holding potential to identify mTBI patients who could benefit from specific neuropsychiatric and cognitive rehabilitation. 246][27][28] In our previous study using a rat model of direct cranial blast traumatic brain injury (TBI), 23 we found delayed neurofunctional and pathological abnormalities subsequent to the brain injury that were silent on conventional T 2 -weighted imaging, but microstructural and metabolic changes were observed with DKI and proton MRS, respectively.Increased mean kurtosis, which peaked at 21 days post-injury, was observed in the hippocampus and internal capsule.Concomitant increases in myo-inositol (Ins) and taurine (Tau) were also observed in the hippocampus, whereas early changes at 1 day in glutamine (Gln) were observed in the internal capsule, all indicating glial abnormality in these regions.
In this study, to assess the subtle changes that are associated with blast exposure, we performed in vivo longitudinal DTI/DKI, MRS, and a behavioral test to assess the biochemical and behavioral profile alterations after blast exposure in a ferret model.

Ferret blast-related traumatic brain injury procedure
The experimental protocol was approved at the Walter Reed Army Institute of Research, and research was conducted in an Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) International-accredited facility in compliance with the Animal Welfare Act and all other federal statutes and regulations relating to animals and experiments involving animals, adhering to principles stated in the Guide for Care and Use of Laboratory Animals, NRC Publication 2011 edition, and the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines.In all, 18 male Fitch ferrets (1193 -19 g, age 108.1 -0.96 days) were subjected to bTBI using an advanced blast simulator that recreates ''free-field''-like blast closely resembling that experienced by warfighters. 1,29errets were anesthetized with 4% isoflurane for 8 min.An advanced blast simulator (ABS) was used for this study, which creates flow conditions much more closely resembling the Friedlander waves produced in the free field by improvised explosive devices and other explosive detonations, and was used to characterize exposures 29,30 and eliminate artefactual blast wind, which has been the primary cause of substantial injury caused by exposures in constant diameter shock tubes.The ABS consisted of an 0.5-ft-long compression chamber separated from a 21-ft-long transition/expansion test section by rupturable VALMEX Ò membranes (Mehler Texnologies, Inc., Martinsville, VA) or acetate membranes (Grafix Inc, Maple Heights, OH).The anesthetized ferret was secured in a longitudinal (i.e., head-on) orientation in the test section.Ferrets were exposed to a blast overpressure of *138 KPa with a 4-to 5-msec positive pressure duration.The critical biomechanical loading to the experimental subject was determined from both the static (Ps) and dynamic pressure (Pd) of the blast wave, which were fully recorded by a combination of side-on and head-on piezoresistive pressure gauges (Endevco, Depew, NY) using an Astro-med tmx-18 acquisition system at an 800,000-Hz sampling rate.This generated a model of blast TBI.
All sham animals in these experiments were subjected to isoflurane anesthesia, loading in the shock tube, and recovery procedures as described above, but were not exposed to the blast overpressure.Nine blast ferrets and 9 sham ferrets were included in this study.

Imaging experiments
The imaging protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Maryland, with IACUC No. 0519003.Imaging experiments were performed at 3 days, 7 days, 1 month, 3 months, and 6 months postinjury.In vivo magnetic resonance (MR) experiments were performed on a Bruker Biospec 7T scanner (Bruker BioSpin MRI GmbH, Ettlingen, Germany) run by Paravision 6.0.Anesthesia was maintained with around 3% isoflurane during the imaging session.A Bruker 30-mm single-turn surface coil was used as a receiver.A corresponding Bruker 86-mm-volume coil was used as a transmitter.An MR compatible smallanimal monitoring and gating system (SA Instruments, Inc., Stony Brook, NY) was used to monitor animal respiration rate and body temperature, which was maintained at 37°C-38.5°C using a warm water bath circulation.
DTI and DKI data were collected using an echo planar imaging spin echo diffusion scheme with a total of 30 diffusion directions in the coronal view.In total, five b = 0 images were collected with: TE, 24.58 ms; TR, 2500 ms; repetitions, 2; and matrix size, 128 • 128.The rest of the parameters matched to the corresponding T 2 -weighed images.A scan with the b-value set to 1000 and 2000 s/mm 2 was collected.The diffusion gradient separation was 9.34 ms.The diffusion gradient duration was 3.0 ms. 1 H-MRS data were obtained from a voxel (3.5 • 3.5 • 3.5 mm 3 ) that covered the mPFC and left hippocampus (3.5 • 2 • 4 mm 3 ; Fig. 1).A short-TE point-resolved spectroscopy pulse sequence (TR/TE = 2500/10 ms, numerical aperture = 360) was used for MRS data acquisition. 31The unsuppressed water signal from the prescribed voxel was used as a reference for determining the specific metabolite concentrations.

Behavioral assessment
Behavioral assessments were conducted at 1, 3, and 6 months post-injury.The primary behavioral assessment performed in this study was a trap test for ferrets that was developed in-house.The test utilized a modified skunk trap with a panel to control the opening and closing of the trap at one end (Fig. 2A).The ferret was put inside the trap and allowed to acclimate for 2 min.After the acclimation, the panel was opened and the time taken to fully exit the trap (from nose at the opening to rear feet out) was recorded.The test was designed to measure level of anxiety and impulsivity.

Histology
At 6 months post-blast, after behavior and imaging studies, animals were deeply anesthetized and perfusion-fixed using 4% paraformaldehyde at the University of Maryland.Fixed brains were sent to FDNeur-oTechnologies Inc. (Columbia, MD) for further processing for silver staining.Briefly, paraffinembedded sections were cut to 50 lm and stained with the FD NeuroSilverÔ Kit II in accordance with the manufacturer's protocol.Researchers were blinded to exposure groups and used light microscopy for imaging of brain regions of interest (ROIs).ImagePro 9 software (Media Cybernetics, Rockville, MD) was used to assess the silver-staining-positive regions.
Image analysis DKI and DTI model fitting was performed using customized diffusion kurtosis software, 32 and parametric maps were calculated for mean diffusivity (MD), FA, axial diffusivity (AD), radial diffusivity (RD), MK, AK, and RK.The ROIs, including medial pre-frontal cortex (mPFC), primary somatosensory cortex, striatum, hippocampus, thalamus, corpus callosum, internal capsule, and auditory cortex, were manually defined on the FA images while using the T 2 -weighted image for anatomical reference in FSLeyes (Fig. 3).Values of MD, FA, AD, RD, MK, AK, and RK were extracted respectively from each generated map using the manually defined ROIs.The detailed procedure has been published previously. 31 1 -MRS data were fitted using the LCModel package (version 6.3-0G; LCModel Inc., Oakville, ON, Canada). 33 threshold of the Cramer-Rao lower bounds £35 % was used to select metabolite concentrations that had an acceptable level of quantification reliability. 32

Statistical analysis
Statistical analysis was performed in R (version 4.0.2;R Core Team, 2013).Repeated-measures analysis of vari-ance (ANOVA) was performed on imaging and behavioral measures with group as between-subject effect and time as within-subject effect.Interaction effect between group and time was also considered.Post hoc pairwise comparison was performed with Tukey's method when significant main effects or interaction were detected.our experimental subjects (data not shown) at 3 days to 6 months post-exposure.However, we detected brain tumors in 2 sham ferrets.One ferret showed a benign tumor located in the cerebellum at the first MRI time point, which did not significantly grow or change shape during the total 6 months.Because the location of the tumor was not included in our ROIs and the experimental subject did not show any abnormal behavior or weight gain compared to its peers, we kept this ferret in our study.Another sham ferret showed a benign tumor located in left hippocampal region at the first time point and was excluded from this study because the region was included in our ROIs.Therefore, the ferrets used in the sham group decreased from 9 to 8 over the course of the study.

Imaging experiments T 2 -weighted MRI did not indicate the presence of hematoma, hemorrhage, or edema after bTBI in any of
No microstructural disruption of the TBI was statistically distinguished by the DTI and DKI data evaluated from the brain regions selected in this cohort.However, a widespread effect of time was shown on diffusion parameters across the brain, which reflected the ferret brain structural maturation (from approximately 3.5 to 9.5 months; Figure 3 and Tables 1 and 2).During the 6 months, globally increased FA, MK, RK, and AK values and decreased MD, AD, and RD values were observed in the ferrets.
MRS data showed a significant group main effect for glutamate (Glu; p = 0.047) and N-acetyl-aspartate (NAA; p = 0.048) in the mPFC.Both metabolites increased significantly in the blast ferrets.Glu increased transiently at 1 month post-blast ( p = 0.024), whereas NAA remained high at both 1 and 3 months post-blast ( p = 0.039 and p = 0.047, respectively) in the blast ferrets.

Behavior assessment
A significant main effect of group was detected for the time animals spent in the trap ( p = 0.01).The blast ferrets spent significantly less time to exit the trap than the sham ferrets at 1 and 3 months post-blast ( p = 0.012 and p = 0.005, respectively).The difference did not reach statistical significance at 6 months post-blast ( p = 0.256; Fig. 2B).

Histology
Silver staining showed a significant increase in argyrophilic inclusions in the anterior ectosylvian gyri area of the auditory cortex ( p < 0.05; Fig. 6).No statistically significant changes were observed in the thalamus and hippocampal regions examined at 6 months after blast exposure (data not shown).

Discussion
We report here that blast injury in the gyrencephalic ferret resulted in increased behavioral impulsivity, a exited the opening faster than the sham ferrets, which likely represents a possible increased impulsive decision making, which has been earlier attributed to a dysfunction in the pre-frontal cortical/amygdalar axis post-blast. 345][36] Impulsive behavior can either increase the chance of experiencing TBI, or TBI can lead to a decline in impulse control.It is therefore likely that the present observation of increased impulsivity may be attributable to the effect of blast. 37ur findings of increased Glu are consistent with earlier reports that showed excitotoxicity and associated cognitive dysfunctions during acute and subacute stages post-TBI, 40 which were also observed in a rat model of blast exposure.][40] Considering that TBI is a risk factor for schizophrenia 41 and psychotic disorders, 42 our MRS results, together with the observed increased impulsivity, suggest a link between dysfunction of the pre-frontal cortex and potential secondary psychiatric disorders after mild blast TBI.
NAA plays an important role in lipid synthesis and energy metabolism.Abnormally high brain NAA levels correlated with deficient axonal myelin sheath development. 43The NAA in TBI ferrets showed significantly FIG. 4. Effect of blast on cerebral metabolomics profiles in pre-frontal cortex.*p < 0.05, **p < 0.01, between group or time differences.
higher levels compared to sham animals at both 1 and 3 months, which may indicate a disruption of the utilization of NAA in lipid synthesis and/or energy metabolism at the time.5][46][47] The continuation of trends in changes of Tau and tCr, which start in the fetal stage of development, has been postulated to reflect the increase in neural activity associated with maturation. 52ater diffusion patterns, which are measured by DTI and DKI, have been considered useful reflections of the underlying axonal organization of the brain. 531][22] This discrepancy may likely be attributable to the mild nature of the bTBI model evaluated in this study.This study instead revealed maturation changes of DKI parameters in a group of brain regions.Studies in humans have shown that the maturation of WM involves a phase of axonal myelination, which corresponds to the ensheathment of oligodendroglial processes around the axons. 19,54,55imilar to our findings, clinical studies have shown an increase FA values in blast-exposed veterans when compared to non-blast cohorts, suggesting a possible clinical relevance of this ferret injury model.As has been shown in both pre-clinical and clinical studies, the increase in the FA values corresponds with astrogliosis after blast exposure. 7,8,56he argyrophilic inclusions in the auditory cortex of the blast-exposed ferret likely reflects ongoing neurodegeneration and axonal degeneration, even at 6 months postblast, which is noteworthy given that the auditory system is highly sensitive to blast injury.The absence of argyophilic inclusions in the other regions examined, including the hippocampus and thalamus, are possibly reflective of earlier degeneration that did not persist through 6 months post-blast.However, it is possible that brain regions other than those examined in the present study (auditory cortex, hippocampus, and thalamus) might display increased intensity of argyrophilic inclusions.Collectively, these findings show potential maturation changes at a microscopic level, which are apparent with advanced imaging techniques such as DKI and MRS, while also demonstrating the clinical relevance of the readouts that a ferret model can present.Using the ferret as a gyrencephalic animal model to evaluate the effects of blast injuries shows that they display several features resembling those observed in human victims of blast exposures.In particular, the ferret blast TBI model showed a disrupted functioning of the pre-frontal cortex, which could possibly link to the psychiatric disorders associated with blast exposure.As these and other corroborations with clinical findings emerge, we anticipate that they will show ferrets to be a highly translatable model to understand blast-related TBI and potentially point to their utility for future therapeutic evaluations, which, stemming from greater neuroanatomical similarities, may have improved translatability than has been achieved with rodent injury models.

FIG. 2 .
FIG. 2. (A) In-house trap test for the ferret.(B) Time required to walk out from the trap of the sham and blast-exposed ferrets at 1, 3, and 6 months post-blast.*p < 0.05, **p < 0.01, differences between sham and blast animals.

FIG. 6 .
FIG. 6.Effect of blast on argyrophilic inclusions in the auditory cortex.Note a significant increase in the density of silver-stained inclusions at 6 months post-blast.*p < 0.05.Scale bar = 100 lm.

Table 1 .
Effect of Time on DTI/DKI Parameters