Strong Correlation of Genome-Wide Expression after Traumatic Brain Injury In Vitro and In Vivo Implicates a Role for SORLA

Abstract

The utility of in vitro models of traumatic brain injury (TBI) depends on their ability to recapitulate the in vivo TBI cascade. In this study, we used a genome-wide approach to compare changes in gene expression at several time points post-injury in both an in vitro model and an in vivo model of TBI. We found a total of 2073 differentially expressed genes in our in vitro model and 877 differentially expressed genes in our in vivo model when compared to noninjured controls. We found a strong correlation in gene expression changes between the two models (r = 0.69), providing confidence that the in vitro model represented at least part of the in vivo injury cascade. From these data, we searched for genes with significant changes in expression over time (analysis of covariance) and identified sorting protein-related receptor with A-type repeats (SORLA). SORLA directs amyloid precursor protein to the recycling pathway by direct binding and away from amyloid-beta producing enzymes. Mutations of SORLA have been linked to Alzheimer's disease (AD). We confirmed downregulation of SORLA expression in organotypic hippocampal slice cultures by immunohistochemistry and Western blotting and present preliminary data from human tissue that is consistent with these experimental results. Together, these data suggest that the in vitro model of TBI used in this study strongly recapitulates the in vivo TBI pathobiology and is well suited for future mechanistic or therapeutic studies. The data also suggest the possible involvement of SORLA in the post-traumatic cascade linking TBI to AD.

Introduction

Every year, in the United States alone, there are approximately 1.7 million traumatic brain injuries (TBIs) responsible for an estimated $76.5 billion in losses attributable to direct medical costs and lost productivity.^1,2 Unfortunately, despite the tremendous societal and economic cost and decades of research, current pharmacological treatments for TBI are lacking, and TBI remains a leading cause of long-term disability.^3

–6 It is estimated that 75% of TBIs are attributable to concussions or other forms of mild TBI, and a growing body of research on chronic traumatic encephalopathy has revealed serious and long-term detrimental effects of mild brain injuries.^7,8

The primary mechanical event sets in motion a myriad of secondary injury cascades that cause delayed cell death that may continue for days or longer.^9,10 Effective clinical treatments for TBI have been elusive, perhaps because of this complex pathophysiology. A better understanding of the secondary mechanisms leading to delayed cell death and dysfunction post-TBI may support the discovery and development of novel and clinically viable treatments. To this end, several in vivo and in vitro models have been developed to investigate TBI.^11,12 Both in vitro and in vivo models have respective strengths and weaknesses, but ultimately the utility of in vitro models depends on their ability to reproduce the in vivo TBI cascade.¹²

The primary motivation of this study was to determine the degree of transcriptomic similarity between an in vivo and an in vitro model of TBI. Global gene expression changes likely coincide with the activation of different molecular pathways as the secondary injury process unfolds. Therefore, a high degree of correlation between the models would support the utility of in vitro models as research tools to complement in vivo models.

Utilizing DNA/RNA microarrays, it is possible to quantify global changes in gene expression.¹³ This method eliminates the need to choose a specific gene of interest and produce the necessary tools (e.g., primers) for its study, but instead allows for thousands of genes to be analyzed simultaneously. From these data, entire physiological pathways can be studied, which may lead to further insight into injury mechanisms.

Previous studies of TBI have shed light on new genes and pathways of interest, but many of these studies focused on a single injury model or a single time point post-injury.^14
–16 Because of the complex nature of the post-traumatic pathology, temporal changes in gene expression may provide more insight than examination of a single time point. For example, temporal changes in gene expression have been studied for over a decade and have been integral in understanding physiological changes during circadian rhythm cycles.^17

–20 A temporal analysis could provide insight to the progression of a disease or pathology and could identify novel targets for therapies that would be overlooked with analysis at a single time point.

Our in vitro model, which has previously been characterized, utilized organotypic hippocampal slice cultures (OHSCs) that were subjected to an equibiaxial deformation, which was verified by high-speed video analysis.^12,21

–24 To serve as a comparison to the in vitro data, a group of animals were subjected to a closed-head injury using a pneumatically driven piston.^25
–27 Changes in gene expression between the two models were highly correlated at similar time points post-injury. From these data, we identified a transmembrane protein SORLA (sorting protein-related receptor with A-type repeats) that was significantly reduced post-TBI and confirmed its reduced expression with both Western blotting and immunohistochemistry (IHC) in our in vitro model. Interestingly, SORLA has been implicated in Alzheimer's disease (AD), and mutations that reduce SORLA's ability to function properly may lead to AD by decreasing amyloid-beta (Aβ) clearance, possibly linking TBI as a risk factor for AD.^28,29 In conclusion, our study confirms a strong correlation of the pathobiology between our in vitro model and that of an in vivo model, further supporting its utility as a research tool for the discovery of potential therapeutic targets post-TBI.

Methods

Organotypic hippocampal slice cultures

Animal procedures to generate OHSCs were approved by Columbia University's Institutional Animal Care and Use Committee (IACUC). OHSCs were cultured as previously described.^30
–32 Briefly, Sprague-Dawley rat pups (post-natal day [PND] 8–10) were rapidly decapitated, the hippocampus excised, and placed in ice-cold Gey's balanced salt solution (Life Technologies, Grand Island, NY) supplemented with d-glucose (4.5 mg/mL; Sigma-Aldrich, St. Louis, MO). Transverse sections (400-μm-thick) were cut using a McIlwain tissue chopper (Ted Pella, Redding, CA) and then plated on polydimethylsiloxane membranes (Specialty Manufacturing, Saginaw, MI) previously coated with laminin (80 μg/mL; Life Technologies, Carlsbad, CA) and poly-l-lysine (320 μg/mL; Sigma-Aldrich) in custom-made stainless steel wells. Cultures were initially fed Neurobasal (Life Technologies) medium supplemented with B27 (Life Technologies), GlutaMAX (2 μM; Life Technologies), and d-glucose (4.5 mg/mL; Sigma-Aldrich). Cultures were maintained under standard culture conditions at 37°C and 5% CO₂. After 3 days in vitro (DIV), the medium was changed to full serum medium containing 50% modified Eagle's medium (Sigma-Aldrich), 25% heat-inactivated horse serum (Life Technologies), 25% Hanks' balanced salt solution (Sigma-Aldrich), GlutaMAX (2 μM; Life Technologies), and d-glucose (4.5 mg/mL; Sigma-Aldrich). Medium was changed every 2–3 days.

In vitro model of traumatic brain injury

After 10 DIV, OHSCs cultured on silicone membranes were subjected to a moderate mechanical injury. The injury was induced by stretching the underlying silicone substrate to a predetermined strain at a pre-determined strain rate to produce the desired tissue injury under motion control. Our well-established model produces a highly accurate and reproducible injury to OHSCs.^12,23,32 Tissue deformation was verified by image analysis of high-speed video (MotionPro; Redlake, Pasadena, CA) at 1000 frames per second.²¹ Lagrangian strain of the tissue was determined by calculating the deformation gradient tensor by locating fiduciary markers on the tissue slice before and at maximal stretch using custom MATLAB software (The MathWorks, Inc., Natick, MA) scripts.²¹

In vivo model of traumatic brain injury

Animal procedures for closed-head TBI were approved by Drexel University's IACUC and were in compliance with the Guide for the Care and Use of Animals. Brain injuries were induced using a pneumatically driven controlled cortical impact device (AmScien Incorporated, Richmond, VA) as described by Raghupathi and colleagues.²⁷ Briefly, rat pups (PND11) were anesthetized with isofluorane (2%) using a nose cone, and once a loss of a tail-pinch reflex was observed, a midline incision was made to expose the skull. The periosteum was reflected, the animal was placed in a restrainer, and the head was supported by a soft foam pad to make it level with the body. The restrainer was positioned under the cortical impact device, the nose cone was removed, and the zero-point for the indenter was made on the skull over the left parietal cortex, midway between the lambda and bregma sutures. The exposed skull was subjected to a 5-mm-diameter piston impact of 3.0 mm in depth, 5 m/s velocity, and 100 ms in duration for moderate injury. Sham-injured animals underwent the same procedures, except that the piston was not activated. Total time from initiation of anesthesia to removal of the nose cone preceding zeroing the impactor tip was typically 5–6 min. After recovery from injury/anesthesia, animals were reanesthetized with isofluorane for the scalp to be sutured, and pups were returned to the dam. Animals were placed on a heating pad kept at 37°C to maintain body temperature throughout the procedures and recovery.

Affymetrix gene chips

OHSC samples, including time-matched controls, were collected after the in vitro model of TBI at 1, 6, 12, and 24 h post-injury. Whole hippocampal samples, including time-matched controls, were collected after the in vivo model of TBI at 1, 6, and 24 h post-injury. Tissue was collected, flash frozen, and stored at −80°C until total RNA was isolated using an RNeasy Mini Kit (Qiagen, Valencia, CA), following the manufacturer's protocols. Quality of RNA was assessed using a NanoDrop 2000 (Thermo Fisher Scientific, Waltham, MA). Purified RNA was used to synthesize complementary (cDNA) using the HT One-Cycle cDNA Synthesis Kit (Affymetrix, Santa Clara, CA), following the manufacturer's protocols. Amplified complementary RNA cRNA was produced from the cDNA template using the 3′ IVT Labeling Kit (Affymetrix). The cRNA was then hybridized to the Affymetrix Rat 230 2.0 chips, washed, and scanned at the Columbia Genome Center according to the manufacturer's protocols.

Microarray data analysis

Using custom MATLAB scripts, each array was normalized for variations in background using the robust multi-array average procedure.³³ After performing a Student's t-test between experimental groups on an individual gene basis, p values were adjusted for false discovery rate (FDR) using Storey-Tibshirani's method to improve confidence in p values.³⁴ An FDR-corrected p value <0.05 and a 2-fold biological change in expression were chosen as criteria for identifying genes whose expression was altered both statistically and biologically.³⁵ Analysis of covariance (ANCOVA) was used to determine whether gene expression over time was significantly different between experimentally injured groups and noninjured controls. An ANCOVA p value <0.05 was considered significant. To enrich for the most significantly changed genes in our data set for further analysis, we focused on genes that were significant in both the Student's t-test and ANCOVA.

To determine the correlation between the two injury models, Pearson's correlation coefficient was determined for the 323 genes found to be significant in both the in vitro and in vivo models by the Student's t-test.

Using Database for Annotation, Visualization and Integrated Discovery (DAVID) Bioinformatics Resources provided by the National Institutes of Health (NIH), each time point was analyzed for the most enriched Gene Ontology (GO) terms.³⁶ We chose to focus on biological process terms and have excluded cellular component and molecular function GO terms from analysis. Specifically, GO FAT biological process terms, as described by DAVID, were used for analysis. The GO FAT subset excludes the broadest GO terms, so that they do not overshadow more-specific terms.

Western blots

Additional samples were collected 48 h post-injury to confirm the transcriptional results at the protein level. This time point was chosen to allow for changes in messenger RNA expression to be translated to protein expression. For in vitro injured and uninjured controls, nine slices were collected for protein extraction. Slices were rinsed twice with ice-cold phosphate-buffered saline and immediately placed in lysis buffer A (40 mM of HEPES, 120 mM of NaCl, 1 mM of ethylenediaminetetraacetic acid, 1% Triton X-100, 10 mM of sodium pyrophosphate, 50 mM of sodium fluoride, 0.5 mM of sodium orthovanadate, and 10 mM of β-glycerophosphate; Sigma-Aldrich). Samples were sonicated (Sonicator 3000; Misonix, Farmingdale, NY), incubated on ice, and then centrifuged to remove cell debris. Approximately 50 μg of protein per sample was loaded in a 4–12% Bis-Tris gel (Life Technologies), and proteins were separated by electrophoresis (150 V, 1 h). Proteins were then transferred to a nitrocellulose membrane (Life Technologies) using a semidry apparatus (Thermo Fisher Scientific). After transfer, the membrane was blocked in Tris-buffered saline (TBS; pH 7.4) with 5% bovine serum albumin (BSA) for 2 h. Membranes were incubated overnight at 4°C with primary antibodies (Anti-SORL1 [S9200; Sigma-Aldrich], anti-β-tubulin to serve as an internal control [32-2600; Life Technologies]) in TBS-T (0.1% Tween-20; pH 7.4) and 5% BSA. After primary antibody incubation, membranes were washed 3 × 10 min in TBS-T. For detection, membranes were labeled with a corresponding secondary antibody (goat/anti-rabbit Alexa Fluor 546 or goat/anti-mouse Alexa Fluor 647; Life Technologies). Fluorescence was detected using a CRi Maestro 2 Imaging System (PerkinElmer, Akron, OH). Grayscale images were analyzed in ImageJ (NIH, Bethesda, MD) using the gel analysis function.³⁷ SORLA band intensities were normalized by the intensity of the β-tubulin band.

Patient samples

Autopsy case material was obtained from the New York City Office of Chief Medical Examiner (OCME; New York, NY). According to the policies of the New York City OCME and New York University School of Medicine, this research did not require institutional review board (IRB) approval; however, it was submitted to and approved by the OCME legal office to insure that it complies with OCME policies and did not require IRB approval. Two types of subjects who died of TBI were sought: those that died immediately from the TBI and those that survived for at least 12 h post-TBI. The immediate-death cases served as controls for comparison of changes in post-traumatic protein expression in the delayed-death cases. Blocks of fixed brain tissue from subjects with TBI were embedded in paraffin.

Immunohistochemistry

All steps were performed with a Ventana Benchmark Ultra automated staining instrument (Ventana Medical Systems, Tucson, AZ). Formalin-fixed, paraffin-embedded tissue samples from either human (at indicated time points) or rat OHSCs (48 h post-injury) were cut at 8 or 6 um, respectively. Using the automated Ventana equipment, paraffin was removed at 72°C, and antigen retrieval was achieved by incubation with CC1 reagent (Ventana Medical Systems) at 95°C. After antigen retrieval, slides were incubated with anti-SORL1 (S9200, 1:1000; Sigma-Aldrich) for 32 min before being washed. Tissue nuclei were stained using hematoxylin. Slides were then dehydrated using gradient alcohols and cover-slipped. Adjacent sections were stained with Luxol fast blue and hematoxylin and eosin (LHE) for routine pathological analysis.

Quantification of staining in human tissue

Stained sections from autopsy cases were analyzed semiquantitatively by a blinded observer. For SORLA IHC, a 4-point rating scale (0, +, ++, and +++) was devised to correspond to the intensity of staining (none, weak, moderate, or strong, respectively). A similar scheme (0, +, ++, and +++) for the LHE-stained sections (none, rare, occasional, or frequent, respectively) was devised to quantify pathological findings, including shrunken neurons, vacuolization, neuronal loss, and dark neurons.

Results

Comparison of in vitro and in vivo traumatic brain injury models

Across all time points, a total of 2073 unique genes were differentially expressed in at least one time point after in vitro TBI, and a total of 877 genes were differentially expressed in at least one time point after in vivo TBI compared to uninjured controls (Supplementary Data) (see online supplementary material at http://www.liebertpub.com). After in vitro TBI, 988, 866, 861, and 374 genes were significantly differentially expressed at 1, 6, 12, and 24 h post-injury, respectively, compared to uninjured controls. For the in vivo TBI model, 312, 445, and 308 genes were significantly differentially expressed at 1, 6, and 24 h post-injury, respectively, compared to uninjured controls. There were a large number of differentially expressed genes found at multiple time points for both models of injury, and this overlap is shown in Figure 1 A,B. Comparing the overlap between the two models, 323 genes were differentially expressed in both models of injury as shown in Figure 1C. In total, the in vitro model replicated around 37% of the changes in gene expression of the in vivo model of TBI, although almost 3 times as many genes were significantly regulated in the in vitro model.

FIG. 1.

Overview of significant changes in gene expression post-TBI. Venn diagrams show the number of overlapping genes whose expression was significantly altered compared to time-matched controls for (A) in vitro TBI, (B) in vivo TBI, and (C) in vitro versus in vivo TBI. TBI, traumatic brain injury.

We next investigated how correlated gene expression was between the two models of TBI for the 323 differentially expressed genes common to both the in vivo and in vitro models. The comparison of fold changes is graphed in Figure 2. There was a strong correlation in gene expression between the in vitro and in vivo models at matching time points (1 h, r = 0.54; 6 h, r = 0.64; 24 h, r = 0.59). Within the in vitro injury group, there were very strong correlations between the 6- and 12-h time points (r = 0.89) and strong correlations between the 6- and 24-h time points (r = 0.67) and the 12- and 24-h time points (r = 0.62). This trend was similar for the in vivo injury between the 6- and 24-h time points (r = 0.85). A moderately negative correlation between the 1-h time point and all later time points within a given model (Fig. 2) suggests that an abundance of genes were initially differentially expressed post-injury and returned to baseline over time.

FIG. 2.

Correlation of gene expression changes post-TBI. Scatter plots of log₂ fold changes of the 323 genes that were significantly and differentially expressed in both models. Pearson's correlation coefficients (r) are located in the top left corner. Axes represent log₂ fold change from −5 to 5. Dashed line encloses in vitro versus in vivo comparisons. Both models were strongly correlated, particularly at matching time points.

For an overall comparison of our in vitro to in vivo TBI models, we combined the gene expression fold changes for the 37% of genes that were significantly regulated in both models at identical time points post-injury (1 h in vitro vs. 1 h in vivo, 6 h in vitro vs. 6 h in vivo, and 24 h in vitro vs. 24 h in vivo). Gene expression fold changes for the time-point–matched analysis had a combined Pearson correlation of r = 0.69. This strong correlation gives us confidence that our in vitro model is at least partially representative of in vivo TBI.

Gene ontology analysis

After analysis for the most significantly enriched GO terms in our data set, TBI-relevant GO terms were chosen, and the number of genes within a GO term for in vitro and in vivo injuries were graphed (Fig. 3). Several TBI-relevant GO terms contained significantly differentially expressed genes after both injury models, and further, many of the genes in these relevant groups were identical as indicated in Figure 3.

FIG. 3.

GO analysis of significantly altered gene expression post-TBI. The number of significantly altered genes from each model within the specified GO groups is graphed. The vertical black bar represents the number of identical genes that were significantly regulated in both models within the specified GO group. TBI-relevant GO groups were remarkably similar between the in vitro and in vivo models, and many identical genes were altered in both models. GO, Gene Ontology; TBI, traumatic brain injury.

Analysis of covariance of in vitro traumatic brain injury

The comparison of significant genes between the ANCOVA analysis and Student's t-test yielded 32 genes representing changes of particular interest. These genes are listed in Table 1. For further analysis, we decided to focus on genes that were not involved in cell adhesion, were predominantly expressed in the central nervous system (CNS), and were previously linked to diseases of the CNS. Following these guidelines we further validated SORLA, which has been linked to AD, with Western blot and IHC.^28,29

Table 1.

Genes Significant by ANCOVA Analysis and Student's t-Test

Affymetrix ID	NCBI ID	Symbol	Name
1380783_at	81796	Slco5a1	Solute carrier organic anion transporter family, member 5A1
1368261_at	116508	Nrxn3	Neurexin 3
1393933_at	300652	Sorla	Sortilin-related receptor, LDLR class A repeats-containing
1368290_at	83476	Cyr61	Cysteine-rich, angiogenic inducer, 61
1394095_at	316687	Gin1	Gypsy retrotransposon integrase 1
1373179_at	690003	Btbd8	Acyltransferase-like 1B; similar to H43E16.1; BTB (POZ) domain-containing 8
1376574_at	84398	Cd93	CD93 molecule
1384144_at	360577	Nxn	Nucleoredoxin
1394286_at	311616	Gdap1l1	Ganglioside-induced differentiation-associated protein 1-like 1
1369217_at	58853	Nr4a3	Nuclear receptor subfamily 4, group A, member 3
1377920_at	294708	Sgtb	Small glutamine-rich tetratricopeptide repeat (TPR)-containing, beta
1385036_at	307309	Sncaip	Synuclein, alpha interacting protein
1391146_at	84407	Cdh11	Cadherin 11
1388332_at	363875	Rac1	Ras-related C3 botulinum toxin substrate 1
1383632_at	58976	Slc30a1	Solute carrier family 30 (zinc transporter), member 1
1388872_at	89784	Idi1	Isopentenyl-diphosphate delta isomerase 1
1389280_at	309900	Tulp1	Tubby-like protein 1
1396485_at	308890	Btbd10	BTB (POZ) domain-containing 10; similar to K⁺ channel tetramerization protein
1379101_at	310461	Dhx36	DEAH (Asp-Glu-Ala-His) box polypeptide 36
1371071_at	24400	Gnb1	Guanine nucleotide-binding protein (G protein), beta polypeptide 1; guanine nucleotide-binding protein (G protein), beta polypeptide 4
1390176_at	315019	Fam84b	Family with sequence similarity 84, member B
1391807_at	315309	Tcfcp2	Transcription factor CP2
1393725_at	291755	Dsg1b	desmoglein 1 beta
1378205_at	—	—	Unknown
1379531_at	—	—	Unknown
1386059_at	—	—	Unknown
1376458_at	—	—	Unknown
1374022_at	—	—	Unknown
1396050_at	—	—	Unknown
1390240_at	—	—	Unknown
1374713_at	—	—	Unknown
1390425_at	—	—	Unknown

ANCOVA, analysis of covariance; NCBI, National Center for Biotechnology Information.

Protein expression of sorting protein-related receptor with A-type repeats

Although SORLA is widely expressed in the brain, it is predominately expressed in neurons of the hippocampus.³⁸ Cytoplasmic puncta were observed in the cornu ammonis 1 (CA1) pyramidal neurons of OHSCs of uninjured controls (Fig. 4A,B). Interestingly, at 48 h after mechanical injury of the OHSCs, these puncta were reduced or completely missing from the cytoplasm of pyramidal neurons (Figure 4 C,D). Injured regions also exhibited a vacuolar appearance, suggesting cell death. These data further support our gene chip data, corroborating reduced SORLA expression post-injury. To further confirm this claim, we measured protein expression of SORLA by Western blotting after in vitro TBI. We found that SORLA expression was significantly reduced by 44% compared to uninjured control OHSCs (p < 0.05; Fig. 4E,F).

FIG. 4.

SORLA expression after in vitro injury. Micrographs of SORLA staining in uninjured (A: 4×; B: 40×) and moderately injured (C: 4×; D: 40×) OHSCs. An overall decrease in SORLA staining was observed post-injury, and distinct SORLA staining puncta were missing from the apical dendrites. (E) Normalized SORLA protein expression after in vitro injury of OHSCs was significantly decreased post-injury. (F) Representative Western blot of SORLA and β-tubulin. Scale bar (A and C) = 1 mm; (B and D) = 100 um. OHSCs, organotypic hippocampal slice cultures; SORLA, sorting protein-related receptor with A-type repeats. Color image is available at www.liebertpub.com/neu

Next, we assessed changes in SORLA expression in the brains of human subjects after severe TBI. Sections of brain tissue from 8 TBI subjects were studied with details given in Table 2. Of these subjects, 5 died shortly post-injury, and 3 had prolonged survival (≥12 h). Semiquantitative results are summarized in Table 3. In the immediate deaths, a granular label was detected prominently within neurons, but the number and sizes of these granules varied considerably across brain regions. Neurons in CA1 often had single or scattered medium-sized cytoplasmic granules (∼1 μm in diameter; Fig. 5A). Within the dentate gyrus and entorhinal cortex, numerous small, finely granular staining was present that filled the cytoplasm (Fig. 5B,C). Neurons in the entorhinal cortex often had an additional single large, perinuclear granule (∼8 μm). A diffuse granular label of the cytoplasm was also detected occasionally within microglia. Staining in astroglia and oligodendroglia was not prominent. In the cases that died ≥12 h post-injury, the signal intensity of SORLA staining was less prominent (Fig. 5D); however, the significance of this is uncertain given the small sample size and the heterogeneity that is associated with human tissue. In the frontal white matter from 1 case, we found numerous SORLA-positive cellular processes (Fig. 5E,F). The significance of this finding is also unclear, but this is a brain region known to be vulnerable to diffuse axonal injury and may reflect secondary changes in axons. Together, these results open up the possibility that changes in SORLA expression may occur in human brain post-TBI.

FIG. 5.

SORLA expression after human TBI. Localization of SORLA in human post-mortem brain post-TBI. (A–C) Immunohistochemical staining with antisera targeting SORLA revealed strong labeling within the cytoplasm of neurons of granules that differ in size and number across brain region. CA1 neurons had single or small numbers of medium-sized perinuclear granules. Dentate gyrus granule neurons had abundant diffuse granular labeling. Neurons in the entorhinal cortex displayed diffuse granular labeling, but also often contained a single large perinuclear granule as well. (D) Overall, the abundance of diffuse granular label was greatly diminished in the 3 TBI cases with extended survival. (E and F) In addition, antisera to SORLA also labeled scattered abnormal and swollen cellular processes within the white matter (highlighted by arrows in E). Scale bar (A–E) = 50 um; (F) = 10 um. SORLA, sorting protein-related receptor with A-type repeats; TBI, traumatic brain injury. Color image is available at www.liebertpub.com/neu

Table 2.

Patient Information on Human Control and TBI Cases for Immunohistochemistry

Case	Age	Sex	TBI mechanism	Other neuropathological findings	TBI-death interval
1	22	M	Motor vehicle accident	Cortical contusions, subdural hemorrhage	0 h
2	20	F	Motor vehicle accident	Subarachnoid hemorrhage, cortical contusions, diffuse axonal injury	0 h
3	21	M	Motor vehicle accident	Subdural hemorrhage, neck injuries	0 h
4	21	M	Motorcycle accident	Subarachnoid hemorrhage	0 h
5	49	M	Pedestrian hit by car	Subural and subaracnhoid hemorrhage, cortical contusions, diffuse vascular injury	0 h
6	37	M	Fell from building	Subdural and subarachnoid hemorrhage	12–24 h
7	32	M	Fell from building	Subarachnoid hemorrhage	12–24 h
8	44	M	Motor vehicle accident	Subarachnoid and subdural hemorrhage	1 month

TBI, traumatic brain injury.

Table 3.

Qualitative Assessment of Human SORLA and H + E Staining

	SORLA IHC					Luxol fast blue/hematoxylin and eosin
Case	DG	CA1	Sub	EC	PG	Shrunken neurons	Vacuolization	Neuronal loss	Dark neurons
1	0	+	0	++	++	0	+	0	++
2	+++	+	++	+++	+++	0	0	0	+
3	++	+	++	+++	+++	+	0	0	++
4	+	+	+	+++	+++	++	++	0	++
5	+	+	+	++	+	+	0	0	++
6	+	+	0	+	+	++	++	0	+
7	0	+	++	++	++	0	0	0	0
8	+	+	+	+	+	0	0	0	+

0 = none; + = weak/rare; ++ = moderate/occasional; +++ = strong/frequent.

SORLA, sorting protein-related receptor with A-type repeats; H+E, hematoxylin and eosin; IHC, immunohistochemistry; DG, dentate gyrus; CA1, cornu ammonis 1; Sub, subiculum; EC, entorhinal cortex; PG, parahippocampal gyrus.

Discussion

In vitro models of TBI have the advantage of higher throughput than in vivo methods, but their utility depends on their ability to replicate the in vivo injury cascade(s). To that end, this study compared the global changes in gene expression post-TBI in both an in vitro and an in vivo model. Time points were chosen to investigate the near-immediate cellular response (1 h), which may lead to downstream cascades at later time points (6 and 12 h) and potential stabilization and recovery (24 h). Although the effects from TBI, such as cell death, can persist for several hours or days after the initial injury, this study focused on the time frame post-TBI during which therapeutic interventions may be successful, and our data may point to novel targets for future clinical treatments.^9,10,39,40 The authors acknowledge that additional time points would be relevant and could further elucidate long-term pathological effects of TBI. Nonetheless, the current study revealed interesting and novel effects of TBI that were identified with an in vitro study and confirmed in vivo.

Some of the significant gene expression changes in our current study have been identified previously either in genome-wide or single-gene studies.^{14
–16,41

–45} Di Pietro and colleagues, using the same in vitro model as ours, investigated mild (10%) and severe (50%) injuries in OHSCs at 24 h post-injury and found significant changes in expression of inflammatory and apoptotic genes, such as Cox-2, Cxcl2, and Mmp2.^16,46,47 However, only a single time point of 24 h was used for their analysis. Our data show many changes in gene regulation at 1 h post-injury that are not evident at 24 h. It is possible that, in their model, the severe injury may induce significant gene expression changes at an earlier time point than mild TBI, which would be overlooked by focusing on a single time point. Both Crack and colleagues and Poulsen and colleagues used an in vivo TBI model and analyzed gene expression changes at multiple time points post-injury.^15,48 However, no temporal analysis method was employed; instead analysis focused on gene expression changes at individual time points. In both studies, similar inflammatory genes were found to be differentially expressed. A reanalysis of their data sets, taking into account that temporal aspects could reveal interesting and statistically significant changes that would otherwise be missed. White and colleagues analyzed gene expression changes 24 h post-TBI in rats.⁴¹ They compared ipsilateral to contralateral tissue and found significant changes in inflammatory genes. The researchers acknowledged the temporal limitation of their study and suggested that a larger study investigating many time points would be necessary to track changes in gene expression. To our knowledge, ours is the first study of global gene expression post-TBI that takes advantage of multiple time points (ANCOVA) to further enrich our conclusions and provide novel targets for future study.

Comparison of in vitro and in vivo traumatic brain injury models

Gene expression between the two models was strongly and positively correlated 1 h post-injury. Interestingly, these initial changes in gene expression were moderately negatively correlated with later changes in gene expression suggesting a transient burst of expression followed by a return to baseline similar to post-traumatic immediate early gene expression observed by others.^49

–52 This relationship was apparent in both the in vitro and in vivo models of our study. The strongest correlations within a model were between the in vitro 6- and 12-h time points and the in vivo 6- and 24-h time points. This trend extended to the cross-model comparison at the later time points (6, 12, and 24 h) as well. Biological fold changes between the in vitro and in vivo model at matched time points (1 h in vitro vs. 1 h in vivo, 6 h in vitro vs. 6 h in vivo, and 24 h in vitro vs. 24 h in vivo) had a strong correlation (r = 0.69). These data suggest that the post-injury changes in gene expression follow a similar trend and therefore the in vitro model is well representative of the in vivo model.

Gene oncology analysis

Several TBI-related GO terms were enriched post-injury in both models. Not surprisingly, several GO terms related to inflammation were enriched including response to wounding (GO:0009611), response to extracellular stimulus (GO:0009991), defense response (GO:0006952), cellular response to stress (GO:0033554), and inflammatory response (GO:0006954). These GO terms contain genes activated by damage to the organism. Response to wounding specifically encompasses genes that are activated directly by an external stimulus, whereas defense response is the reaction of a foreign body introduced to a cell from an attack. Genes from both models that fall within these GO categories included Cd14, Cox-2, Ptges, Cxcl2, and Ccl2. Ccl2 has been implicated in recruitment of macrophages post-TBI, and elevated levels can be measured for 10 days post-trauma in severe TBI patients.⁵³ Cox-2 is one of the most widely studied enzymes responsible for inflammation, and pain and is the target of most nonsteroidal anti-inflammatory drugs, such as ibuprofen.^54,55 Cox-2 has previously been targeted as a potential therapy for TBI and has shown promise in pre-clinical studies.⁴⁷

Secondary sequelae post-TBI can lead to delayed cell death for hours, days, or even years.^56,57 Genes within the GO category positive regulation of apoptosis (GO:0043065) were significantly altered in both injury models and could prove to be targets for therapeutic intervention. For instance, Wwox is an essential mediator of tumor necrosis factor alpha–induced apoptosis and is upregulated post-injury in our models of TBI.⁵⁸ Mmp2, which was significantly increased after closed-head TBI and is implicated in synapse loss, was also found to be upregulated in our study immediately post-injury.⁵⁹ Jnk2 plays an important role in neuronal cell death as well as astrocyte activation, and inhibitors to Jnk2 have shown promising pre-clinical results.^60

–63

The GO term transcription (GO:0006350) included several differentially expressed genes in our study that could have wide-ranging effects on cellular response post-injury. Some of these genes are well-known markers of injury, such as Atf3, and others are part of the transcription machinery, such as Polr2a, which forms the largest subunit of RNA polymerase II.⁶⁴ Perhaps more interestingly, some of the identified genes are directly associated with development and neurogenesis, including Notch-2 and Adnp, respectively. It has previously been reported that peptides based on Adnp can have direct, neuroprotective effects after closed-head injury in mice.⁶⁵ Adnp has also been shown to be vital to embryogenesis and neurogenesis and could have direct implications in neural repair post-injury.⁶⁶

Sorting protein-related receptor with A-type repeats downregulation after traumatic brain injury

By combining different statistical approaches (single-time-point t-tests and ANCOVA), 32 genes were identified as being of particular interest. From this enriched subset of genes, Sorla, which plays a key role in recycling amyloid precursor protein (APP) from the cell surface through the endocytic pathway, was identified. Accumulation of APP in the axon is a hallmark pathology of TBI observed in multiple animal models as well as clinically.^67

–72 In the pig model of inertial brain injury, accumulation of APP with BACE, PS-1, and activated caspases has been suggested to lead to increased Aβ processing, leading to formation of amyloid plaques.⁷³ Our results suggest another molecular contributor in the accumulation of Aβ post-TBI. Specifically, SORLA directs trafficking of APP into recycling pathways by direct binding. Common mutations of SORLA, which are predictive of AD, reduce the binding affinity of SORLA for APP, leaving APP to enter Aβ-generating compartments and accumulate.^28,29 It is speculated that SORLA, in this manner, is mechanistically involved in AD. Accumulating evidence from epidemiological studies implicates TBI as a possible risk factor in AD development.⁷⁴ Although the link between TBI and AD is controversial, extensive research continues to explore the relationship, if any, between injury and disease. At the forefront of this research is an apolipoprotein E (APOE) genotype (ɛ4) that results in higher Aβ deposition post-TBI. However, the link between the APOE ɛ4 genotype and TBI is inconclusive, and no other biological explanation has been well substantiated. In our model, Sorla was downregulated post-TBI, which may lead to a disruption in APP recycling leading to the accumulation of Aβ, suggesting a possible link to AD. Therefore, the identification of decreased Sorla expression post-injury was an interesting and potentially clinically relevant discovery from our in vitro model of TBI.

IHC staining of SORLA in uninjured OHSCs was similar to previous reports in both rat and human tissue.³⁸ Punctate staining was specific to neurons located at the base of the apical dendrites within structures hypothesized to be butrysomes (Fig. 4A,B), suggesting that SORLA expression was not altered during the process of culturing the OHSCs. Further confirming our microarray analysis, SORLA expression was considerably lower, and puncta were almost nonexistent, after in vitro TBI (Fig. 4C,D). Western blotting analysis of SORLA expression confirmed that SORLA was significantly downregulated after mechanical injury in our in vitro model. Preliminary results of human tissue show a similar response, although a larger sample size is necessary to confirm these changes. Further, human samples suitable for analysis by Western blot would strengthen our confidence that these changes are truly biologically significant. These results suggest a potential mechanism for Aβ accumulation post-injury and that SORLA may be a potential therapeutic target for post-TBI treatment, given that an abundance of Aβ can lead to neuronal cell death.⁷⁵ Further research to determine the injury threshold for SORLA reduction and temporal response are necessary to better understand the connection between TBI and SORLA changes. These data suggest the potential therapeutic benefit of targeting SORLA post-injury and may expose links between TBI and neurodegenerative diseases.

Although several genes were highly correlated in both models of TBI, there were many genes that did not overlap between the two models. It is possible that the severity of injury was not equal in both models, although moderate levels were specifically chosen. Further work will be required to determine the importance of these nonoverlapping genes. Although many inflammatory genes were upregulated in our in vitro model, the OHSCs are missing the systemic component of the inflammatory response, which could have major effects on the efficacy of any therapy tested in vitro. Other limitations of the in vitro model include a lack of a blood–brain barrier and circulatory system. We also opted to include more time points in our study at the expense of more arrays per experimental condition. Although the high correlation of genes is a promising result, a more extensive study with additional arrays would provide higher statistical confidence.

Utilizing microarrays, we compared our in vitro model of TBI to an in vivo model. We found many overlapping genes that were categorized in similar GO groups. The correlation of gene expression between models was high; of the 12 Pearson correlations made between the two models, seven had a strong positive correlation (r > 0.50). A remaining two had moderately negative correlations (r < −0.40) and were comparisons of the in vivo 1 h to the in vitro 6- and 12-h time points. These negative correlations could potentially represent immediate early genes responding to the initial mechanical injury, which then normalize at later time points, but this speculation would require further study to confirm. These data, taken together and with the previously mentioned limitations, provide confidence in using the in vitro model to represent some of the in vivo TBI cascade. This important conclusion supports the use of in vitro models such as ours for drug discovery and mechanistic studies, which may reduce the number of animals necessary for TBI research. Further, our study identified significant changes in Sorla expression, which previously has not been implicated in TBI research, and may provide a new connection between TBI and AD, as well as an avenue for therapeutic intervention.

Footnotes

Acknowledgments

This work was supported, in part, by NIH NINDS R03 NS054256 (BMIII), National Institute of Child Health and Development R01 HD 069163 (RR), a National Science Foundation Graduate Research Fellowship (to M.R.L.), and a Natural Sciences and Engineering Research Council of Canada postgraduate scholarship (to B.S.E.).

Author Disclosure Statement

No competing financial interests exist.

References

Coronado

V.G.

, Xu

, Basavaraju

S.V.

, McGuire

L.C.

, Wald

M.M.

, Faul

M.D.

, Guzman

B.R.

, and Hemphill

J.D.

(2011). Surveillance for traumatic brain injury-related deaths—United States, 1997–2007. MMWR Surveill. Summ., 60, 1–32.

Roger

V.L.

, Go

A.S.

, Lloyd-Jones

D.M.

, Adams

R.J.

, Berry

J.D.

, Brown

T.M.

, Carnethon

M.R.

, Dai

, de Simone

, Ford

E.S.

, Fox

C.S.

, Fullerton

H.J.

, Gillespie

, Greenlund

K.J.

, Hailpern

S.M.

, Heit

J.A.

, Ho

P.M.

, Howard

V.J.

, Kissela

B.M.

, Kittner

S.J.

, Lackland

D.T.

, Lichtman

J.H.

, Lisabeth

L.D.

, Makuc

D.M.

, Marcus

G.M.

, Marelli

, Matchar

D.B.

, McDermott

M.M.

, Meigs

J.B.

, Moy

C.S.

, Mozaffarian

, Mussolino

M.E.

, Nichol

, Paynter

N.P.

, Rosamond

W.D.

, Sorlie

P.D.

, Stafford

R.S.

, Turan

T.N.

, Turner

M.B.

, Wong

N.D.

, and Wylie-Rosett

(2011). Heart disease and stroke statistics—2011 update: a report from the American Heart Association. Circulation, 123, e18–e209.

Narayan

R.K.

, Michel

M.E.

, Ansell

, Baethmann

, Biegon

, Bracken

M.B.

, Bullock

M.R.

, Choi

S.C.

, Clifton

G.L.

, Contant

C.F.

, Coplin

W.M.

, Dietrich

W.D.

, Ghajar

, Grady

S.M.

, Grossman

R.G.

, Hall

E.D.

, Heetderks

, Hovda

D.A.

, Jallo

, Katz

R.L.

, Knoller

, Kochanek

P.M.

, Maas

A.I.

, Majde

, Marion

D.W.

, Marmarou

, Marshall

L.F.

, McIntosh

T.K.

, Miller

, Mohberg

, Muizelaar

J.P.

, Pitts

L.H.

, Quinn

, Riesenfeld

, Robertson

C.S.

, Strauss

K.I.

, Teasdale

, Temkin

, Tuma

, Wade

, Walker

M.D.

, Weinrich

, Whyte

, Wilberger

, Young

A.B.

and Yurkewicz

(2002). Clinical trials in head injury. J. Neurotrauma, 19, 503–557.

Willis

, Lybrand

, and Bellamy

(2004). Excitatory amino acid inhibitors for traumatic brain injury. Cochrane Database Syst. Rev., CD003986.

Taussky

, Tawk

R.G.

, Daugherty

W.P.

, and Hanel

R.A.

(2011). Medical therapy for ischemic stroke: review of intravenous and intra-arterial treatment options. World. Neurosurg., 76, 6 Suppl., S9–S15.

Kelley

R.E.

, and Martin-Schild

(2012). Ischemic stroke: emergencies and management. Neurol. Clin., 30, 187–210, ix.

Baugh

C.M.

, Robbins

C.A.

, Stern

R.A.

, and McKee

A.C.

(2014). Current understanding of chronic traumatic encephalopathy. Curr. Treat. Options Neurol., 16, 306.

Gavett

B.E.

, Stern

R.A.

, and McKee

A.C.

(2011). Chronic traumatic encephalopathy: a potential late effect of sport-related concussive and subconcussive head trauma. Clin. Sports Med., 30, 179–188, xi.

Moretti

, Cristofori

, Weaver

S.M.

, Chau

, Portelli

J.N.

, and Grafman

(2012). Cognitive decline in older adults with a history of traumatic brain injury. Lancet Neurol. 11, 1103–1112.

10.

Stoica

B.A.

, and Faden

A.I.

(2010). Cell death mechanisms and modulation in traumatic brain injury. Neurotherapeutics, 7, 3–12.

11.

Gold

E.M.

, Su

, Lopez-Velazquez

, Haus

D.L.

, Perez

, Lacuesta

G.A.

, Anderson

A.J.

, and Cummings

B.J.

(2013). Functional assessment of long-term deficits in rodent models of traumatic brain injury. Regen. Med., 8, 483–516.

12.

Morrison

III , Elkin

B.S.

, Dolle

J.P.

, and Yarmush

M.L.

(2011). In vitro models of traumatic brain injury. Annu. Rev. Biomed. Eng., 13, 91–126.

13.

Fodor

S.P.

, Rava

R.P.

, Huang

X.C.

, Pease

A.C.

, Holmes

C.P.

, and Adams

C.L.

(1993). Multiplexed biochemical assays with biological chips. Nature, 364, 555–556.

14.

Morrison

III , Eberwine

J.H.

, Meaney

D.F.

, and McIntosh

T.K.

(2000). Traumatic injury induces differential expression of cell death genes in organotypic brain slice cultures determined by complementary DNA array hybridization. Neuroscience, 96, 131–139.

15.

Poulsen

C.B.

, Penkowa

, Borup

, Nielsen

F.C.

, Caceres

, Quintana

, Molinero

, Carrasco

, Giralt

, and Hidalgo

(2005). Brain response to traumatic brain injury in wild-type and interleukin-6 knockout mice: a microarray analysis. J. Neurochem., 92, 417–432.

16.

Di Pietro

, Amin

, Pernagallo

, Lazzarino

, Tavazzi

, Vagnozzi

, Pringle

, and Belli

(2010). Transcriptomics of traumatic brain injury: gene expression and molecular pathways of different grades of insult in a rat organotypic hippocampal culture model. J. Neurotrauma, 27, 349–359.

17.

Hazen

S.P.

, Borevitz

J.O.

, Harmon

F.G.

, Pruneda-Paz

J.L.

, Schultz

T.F.

, Yanovsky

M.J.

, Liljegren

S.J.

, Ecker

J.R.

, and Kay

S.A.

(2005). Rapid array mapping of circadian clock and developmental mutations in Arabidopsis. Plant Physiol. 138, 990–997.

18.

Menger

G.J.

, Lu

, Thomas

, Cassone

V.M.

, and Earnest

D.J.

(2005). Circadian profiling of the transcriptome in immortalized rat SCN cells. Physiol. Genomics, 21, 370–381.

19.

Aryee

M.J.

, Gutierrez-Pabello

J.A.

, Kramnik

, Maiti

, and Quackenbush

(2009). An improved empirical bayes approach to estimating differential gene expression in microarray time-course data: BETR (Bayesian Estimation of Temporal Regulation). BMC Bioinformatics, 10, 409.

20.

Leek

J.T.

, Monsen

, Dabney

A.R.

, and Storey

J.D.

(2006). EDGE: extraction and analysis of differential gene expression. Bioinformatics, 22, 507–508.

21.

Elkin

B.S.

, and Morrison

III. (2007). Region-specific tolerance criteria for the living brain. Stapp Car Crash J. 51, 127–138.

22.

, Tsay

, Lacour

S.P.

, Wagner

, and Morrison

(2006). Stretchable microelectrode arrays a tool for discovering mechanisms of functional deficits underlying traumatic brain injury and interfacing neurons with neuroprosthetics. Conf. Proc. IEEE Eng. Med. Biol. Soc., 1 Suppl., 6732–6735.

23.

Cater

H.L.

, Gitterman

, Davis

S.M.

, Benham

C.D.

, Morrison

III , and Sundstrom

L.E.

(2007). Stretch-induced injury in organotypic hippocampal slice cultures reproduces in vivo post-traumatic neurodegeneration: role of glutamate receptors and voltage-dependent calcium channels. J. Neurochem., 101, 434–447.

24.

Morrison

III , Meaney

D.F.

, Margulies

S.S.

and McIntosh

T.K.

(2000). Dynamic mechanical stretch of organotypic brain slice cultures induces differential genomic expression: relationship to mechanical parameters. J. Biomech. Eng., 122, 224–230.

25.

DiLeonardi

A.M.

, Huh

J.W.

, and Raghupathi

(2009). Impaired axonal transport and neurofilament compaction occur in separate populations of injured axons following diffuse brain injury in the immature rat. Brain Res. 1263, 174–182.

26.

Huh

J.W.

, Widing

A.G.

, and Raghupathi

(2008). Midline brain injury in the immature rat induces sustained cognitive deficits, bihemispheric axonal injury and neurodegeneration. Exp. Neurol., 213, 84–92.

27.

Raghupathi

, and Huh

J.W.

(2007). Diffuse brain injury in the immature rat: evidence for an age-at-injury effect on cognitive function and histopathologic damage. J. Neurotrauma, 24, 1596–1608.

28.

Buggia-Prévot

, and Thinakaran

(2014). Sorting the role of SORLA in Alzheimer's disease. Sci Transl Med, 6, 223fs228.

29.

Caglayan

, Takagi-Niidome

, Liao

, Carlo

A.S.

, Schmidt

, Burgert

, Kitago

, Fuchtbauer

E.M.

, Fuchtbauer

, Holtzman

D.M.

, Takagi

, and Willnow

T.E.

(2014). Lysosomal sorting of amyloid-beta by the SORLA receptor is impaired by a familial Alzheimer's disease mutation. Sci. Transl. Med., 6, 223ra220.

30.

Stoppini

, Buchs

P.A.

and Muller

(1991). A simple method for organotypic cultures of nervous tissue. J. Neurosci. Methods, 37, 173–182.

31.

Morrison

, Pringle

A.K.

, McManus

, Ellard

, Bradley

, Signorelli

, Iannotti

, and Sundstrom

L.E.

(2002). L-Arginyl-3,4-Spermidine is neuroprotective in several in vitro models of neurodegeneration and in vivo ischaemia without suppressing synaptic transmission. Brit. J. Pharmacol., 137, 1255–1268.

32.

Sundstrom

, Morrison

, Bradley

and Pringle

(2005). Organotypic cultures as tools for functional screening in the CNS. Drug Discov Today, 10, 993–1000.

33.

Irizarry

R.A.

, Hobbs

, Collin

, Beazer-Barclay

Y.D.

, Antonellis

K.J.

, Scherf

and Speed

T.P.

(2003). Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics, 4, 249–264.

34.

Storey

J.D.

and Tibshirani

(2003). Statistical significance for genomewide studies. Proc Natl Acad Sci U S A, 100, 9440–9445.

35.

Brettschneider

, Collin

, Bolstad

B.M.

and Speed

T.P.

(2008). Quality assessment for short oligonucleotide microarray data. Technometrics, 50, 241–264.

36.

Huang da

, Sherman

B.T.

and Lempicki

R.A.

(2009). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc, 4, 44–57.

37.

Schneider

C.A.

, Rasband

W.S.

and Eliceiri

K.W.

(2012). NIH Image to ImageJ: 25 years of image analysis. Nat Methods, 9, 671–675.

38.

Motoi

, Aizawa

, Haga

, Nakamura

, Namba

, and Ikeda

(1999). Neuronal localization of a novel mosaic apolipoprotein E receptor, LR11, in rat and human brain. Brain Res. 833, 209–215.

39.

Margulies

, and Hicks

(2009). Combination therapies for traumatic brain injury: prospective considerations. J. Neurotrauma, 26, 925–939.

40.

Margulies

, Anderson

, Atif

, Badaut

, Clark

, Empey

, Guseva

, Hoane

, Huh

, Pauly

, Raghupathi

, Scheff

, Stein

, Tang

, and Hicks

(2016). Combination therapies for traumatic brain injury: retrospective considerations. J. Neurotrauma, 33, 101–112.

41.

White

T.E.

, Ford

G.D.

, Surles-Zeigler

M.C.

, Gates

A.S.

, Laplaca

M.C.

, and Ford

B.D.

(2013). Gene expression patterns following unilateral traumatic brain injury reveals a local pro-inflammatory and remote anti-inflammatory response. BMC Genomics, 14, 282.

42.

Ishikawa

, Uchino

, Morota

, Li

, Takahashi

, Ikeda

, Ishii

, and Shibasaki

(2006). Search for novel gene markers of traumatic brain injury by time differential microarray analysis. Acta Neurochir. Suppl., 96, 163–167.

43.

Di Pietro

, Amorini

A.M.

, Tavazzi

, Hovda

D.A.

, Signoretti

, Giza

C.C.

, Lazzarino

, Vagnozzi

, and Belli

(2013). Potentially neuroprotective gene modulation in an in vitro model of mild traumatic brain injury. Mol. Cell. Biochem., 375, 185–198.

44.

Barr

T.L.

, Alexander

, and Conley

(2011). Gene expression profiling for discovery of novel targets in human traumatic brain injury. Biol. Res. Nurs., 13, 140–153.

45.

Matzilevich

D.A.

, Rall

J.M.

, Moore

A.N.

, Grill

R.J.

, and Dash

P.K.

(2002). High-density microarray analysis of hippocampal gene expression following experimental brain injury. J. Neurosci. Res., 67, 646–663.

46.

Strauss

K.I.

, Barbe

M.F.

, Marshall

R.M.

, Raghupathi

, Mehta

, and Narayan

R.K.

(2000). Prolonged cyclooxygenase-2 induction in neurons and glia following traumatic brain injury in the rat. J. Neurotrauma, 17, 695–711.

47.

Strauss

K.I.

(2008). Antiinflammatory and neuroprotective actions of COX2 inhibitors in the injured brain. Brain Behav. Immun., 22, 285–298.

48.

Crack

P.J.

, Gould

, Bye

, Ross

, Ali

, Habgood

M.D.

, Morganti-Kossman

, Saunders

N.R.

, and Hertzog

P.J.

(2009). The genomic profile of the cerebral cortex after closed head injury in mice: effects of minocycline. J. Neural Transm., 116, 1–12.

49.

Robertson

H.A.

(1992). Immediate-early genes, neuronal plasticity, and memory. Biochem. Cell Biol., 70, 729–737.

50.

Dutcher

S.A.

, Underwood

B.D.

, Walker

P.D.

, Diaz

F.G.

, and Michael

D.B.

(1999). Patterns of immediate early gene mRNA expression following rodent and human traumatic brain injury. Neurol. Res., 21, 234–242.

51.

Hayes

R.L.

, Yang

, Raghupathi

, and McIntosh

T.K.

(1995). Changes in gene expression following traumatic brain injury in the rat. J. Neurotrauma, 12, 779–790.

52.

Raghupathi

, and McIntosh

T.K.

(1996). Regionally and temporally distinct patterns of induction of c-fos, c-jun and junB mRNAs following experimental brain injury in the rat. Brain Res. Mol. brain Res., 37, 134–144.

53.

Semple

B.D.

, Bye

, Rancan

, Ziebell

J.M.

, and Morganti-Kossmann

M.C.

(2010). Role of CCL2 (MCP-1) in traumatic brain injury (TBI): evidence from severe TBI patients and CCL2-/- mice. J. Cereb. Blood Flow Metab., 30, 769–782.

54.

Mitchell

J.A.

, Akarasereenont

, Thiemermann

, Flower

R.J.

, and Vane

J.R.

(1993). Selectivity of nonsteroidal antiinflammatory drugs as inhibitors of constitutive and inducible cyclooxygenase. Proc. Natl. Acad. Sci. U. S. A., 90, 11693–11697.

55.

Engelhardt

(1996). Pharmacology of meloxicam, a new non-steroidal anti-inflammatory drug with an improved safety profile through preferential inhibition of COX-2. Br. J. Rheumatol., 35, Suppl. 1, 4–12.

56.

Chen

X.K.

, Xiong

Y.F.

, and Zhou

(2006). “Kiss-and-run” exocytosis in astrocytes. Neuroscientist, 12, 375–378.

57.

Sauaia

, Moore

F.A.

, Moore

E.E.

, Moser

K.S.

, Brennan

, Read

R.A.

, and Pons

P.T.

(1995). Epidemiology of trauma deaths: a reassessment. J. Trauma, 38, 185–193.

58.

Yang

, and Zhang

(2008). WWOX tumor suppressor gene. Histol. Histopathol., 23, 877–882.

59.

Ding

J.Y.

, Kreipke

C.W.

, Schafer

, Speirs

S.L.

, and Rafols

J.A.

(2009). Synapse loss regulated by matrix metalloproteinases in traumatic brain injury is associated with hypoxia inducible factor-1alpha expression. Brain Res. 1268, 125–134.

60.

Fernandes

K.A.

, Harder

J.M.

, Fornarola

L.B.

, Freeman

R.S.

, Clark

A.F.

, Pang

I.H.

, John

S.W.

, and Libby

R.T.

(2012). JNK2 and JNK3 are major regulators of axonal injury-induced retinal ganglion cell death. Neurobiol. Dis., 46, 393–401.

61.

Kang

W.H.

, Simon

M.J.

, Gao

, Banta

, and Morrison

III. (2011). Attenuation of astrocyte activation by TAT-mediated delivery of a peptide JNK inhibitor. J. Neurotrauma, 28, 1219–1228.

62.

Ries

, Silva

R.M.

, Oo

T.F.

, Cheng

H.C.

, Rzhetskaya

, Kholodilov

, Flavell

R.A.

, Kuan

C.Y.

, Rakic

, and Burke

R.E.

(2008). JNK2 and JNK3 combined are essential for apoptosis in dopamine neurons of the substantia nigra, but are not required for axon degeneration. J. Neurochem., 107, 1578–1588.

63.

Tran

H.T.

, Sanchez

, and Brody

D.L.

(2012). Inhibition of JNK by a peptide inhibitor reduces traumatic brain injury-induced tauopathy in transgenic mice. J. Neuropathol. Exp. Neurol., 71, 116–129.

64.

Greer

J.E.

, McGinn

M.J.

, and Povlishock

J.T.

(2011). Diffuse traumatic axonal injury in the mouse induces atrophy, c-Jun activation, and axonal outgrowth in the axotomized neuronal population. J. Neurosci., 31, 5089–5105.

65.

Beni-Adani

, Gozes

, Cohen

, Assaf

, Steingart

R.A.

, Brenneman

D.E.

, Eizenberg

, Trembolver

, and Shohami

(2001). A peptide derived from activity-dependent neuroprotective protein (ADNP) ameliorates injury response in closed head injury in mice. J. Pharmacol. Exp. Ther., 296, 57–63.

66.

Mandel

, Rechavi

, and Gozes

(2007). Activity-dependent neuroprotective protein (ADNP) differentially interacts with chromatin to regulate genes essential for embryogenesis. Dev. Biol., 303, 814–824.

67.

Uryu

, Chen

X.H.

, Martinez

, Browne

K.D.

, Johnson

V.E.

, Graham

D.I.

, Lee

V.M.

, Trojanowski

J.Q.

, and Smith

D.H.

(2007). Multiple proteins implicated in neurodegenerative diseases accumulate in axons after brain trauma in humans. Exp. Neurol., 208, 185–192.

68.

Koizumi

, and Povlishock

J.T.

(1998). Posttraumatic hypothermia in the treatment of axonal damage in an animal model of traumatic axonal injury. J. Neurosurg., 89, 303–309.

69.

Stone

J.R.

, Okonkwo

D.O.

, Singleton

R.H.

, Mutlu

L.K.

, Helm

G.A.

, and Povlishock

J.T.

(2002). Caspase-3-mediated cleavage of amyloid precursor protein and formation of amyloid beta peptide in traumatic axonal injury. J. Neurotrauma, 19, 601–614.

70.

Stone

J.R.

, Okonkwo

D.O.

, Dialo

A.O.

, Rubin

D.G.

, Mutlu

L.K.

, Povlishock

J.T.

, and Helm

G.A.

(2004). Impaired axonal transport and altered axolemmal permeability occur in distinct populations of damaged axons following traumatic brain injury. Exp. Neurol., 190, 59–69.

71.

Washington

P.M.

, Morffy

, Parsadanian

, Zapple

D.N.

, and Burns

M.P.

(2014). Experimental traumatic brain injury induces rapid aggregation and oligomerization of amyloid-beta in an Alzheimer's disease mouse model. J. Neurotrauma, 31, 125–134.

72.

Marklund

, Farrokhnia

, Hanell

, Vanmechelen

, Enblad

, Zetterberg

, Blennow

, and Hillered

(2014). Monitoring of beta-amyloid dynamics after human traumatic brain injury. J. Neurotrauma, 31, 42–55.

73.

Chen

X.H.

, Siman

, Iwata

, Meaney

D.F.

, Trojanowski

J.Q.

, and Smith

D.H.

(2004). Long-term accumulation of amyloid-beta, beta-secretase, presenilin-1, and caspase-3 in damaged axons following brain trauma. Am. J. Pathol., 165, 357–371.

74.

Van Den Heuvel

, Thornton

, and Vink

(2007). Traumatic brain injury and Alzheimer's disease: a review. Prog. Brain Res., 161, 303–316.

75.

Reifert

, Hartung-Cranston

, and Feinstein

S.C.

(2011). Amyloid beta-mediated cell death of cultured hippocampal neurons reveals extensive Tau fragmentation without increased full-length tau phosphorylation. J. Biol. Chem., 286, 20797–20811.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.19 MB