Modeling both the mechanical and hypoxic features of traumatic brain injury in vitro in rats

Neuroscience Letters 328 (2002) 133–136 www.elsevier.com/locate/neulet

Modeling both the mechanical and hypoxic features of traumatic brain injury in vitro in rats Todd F. Glass a,*, Brandi Reeves a, Frank R. Sharp b a

Department of Pediatric Emergency Medicine (Location C 2008), Children’s Hospital Medical Center, 3333 Burnet Avenue, Cincinnati, OH 45229, USA b Department of Neurology and the Neuroscience Program, University of Cincinnati, Cincinnati, OH 45267, USA Received 18 January 2002; received in revised form 2 April 2002; accepted 3 May 2002

Abstract In traumatic brain injury, the brain is subjected to mechanical shear and varying degrees of hypoxia/ischemia. To compare effects of stretch injury, hypoxia, and the combination of both insults on neurons, mixed neuronal and astrocytic cultures were established from day 17 fetal rat brains. On days 17–19 in vitro, cultures were subjected to stretch injury or hypoxia of varying degrees, alone and in combination. Cultures were assayed for lactate dehydrogenase release and Trypan Blue uptake. Hypoxia or Stretch injury alone induced a graded response (P , 0:05) on both assays. Stretch 1 Hypoxia (4 or 6 h) resulted in significantly greater injury as compared with controls (P , 0:05), and as compared with either isolated Stretch or Hypoxia (1, 2, 4 or 6 h) alone (P , 0:05). q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Traumatic brain injury; Hypoxia; Ischemia; Model; Neuron; Axon; Death

Traumatic brain injury (TBI) is the most common cause of death in children [2]. Experimental and clinical evidence suggests that cerebral blood flow (CBF) decreases early following TBI [4,7]. In regions of marked parenchymal injury, CBF can approach ischemic levels (,20% normal) [13]. In adult and pediatric TBI patients, superimposition of hypoxic and ischemic insults significantly increases both morbidity and mortality [3,6]. Despite this, most in vivo and in vitro TBI models have not examined the role of hypoxia and ischemia. In addition, most in vitro models of brain injury have focused on isolated hypoxia/ischemia or mechanical deformation as causes of cell injury, and not combinations of both. The goal of this project was to modify an existing model of mechanical neuronal injury induced by directly stretching cultured cells adherent to a flexible membrane [8]. Following stretch, cultures can be subjected to various metabolic insults. Such a model will provide graduated mechanical and metabolic insults in combination, more closely reflecting the conditions seen in patients suffering TBI. * Corresponding author. Department of Pediatric Emergency Medicine (Location C 2008), Children’s Hospital Medical Center, 3333 Burnet Avenue, Cincinnati, OH 45229, USA. Tel.: 11-513636-7966; fax: 11-513-636-7967. E-mail address: [email protected] (T.F. Glass).

At gestation day 17, pregnant female rats were sacrificed with carbon dioxide. Fetuses were immediately removed from the uterus and placed in ice-cold phosphate-buffered saline (PBS). Fetal brains were harvested, the cortex dissected bilaterally, and placed in cold PBS. The tissues were dissociated with a scalpel followed by trituration through fire-polished glass pipettes in media containing trypsin. After filtering through a 40 mm mesh, the cells were plated at 1 £ 10 6 cells per well in 1 cc of plating medium (4.5 g/l Dulbecco’s modified Eagle’s medium (DMEM), 10% fetal bovine serum, 2 mM l-glutamine, 1% penicillin/streptomycin/amphotericinB) on specially designed six-well plates (FLEXCELL, Int., Hillsborough, NC). The culture surface of each well consists of a collagen coated flexible silastic membrane. The medium was not changed until day 7 in vitro, when growth medium (4.5 g/ l DMEM, 30% glucose, 5% horse serum, 1% penicillin/ streptomycin/amphotericinB) was introduced by halfvolume replacement on alternating days until used for experimentation on day 17 in vitro. Cultures were maintained in a normoxic, 37 8C, 5% CO2 incubator. Experiments were conducted at 17–19 days in vitro. There were three experimental groups. The first group, Stretch, was subjected to 5.5, 6.5 or 7.5 mm stretch using a controller (Biomedical Engineering, MCV, Richmond, VA) that deli-

0304-3940/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 2) 00 51 0- 4

134

T.F. Glass et al. / Neuroscience Letters 328 (2002) 133–136

vers a reproducible positive pressure pulse of compressed air. Deformation of the silastic membrane to these magnitudes correlates to approximately 30, 40 and 55% stretch of the cultured cell layer and results in a graded injury response [8]. The second group, Hypoxia, was exposed to 1, 2, 4, 6 or 24 h of hypoxia. Growth medium was exchanged for medium aerated with 5% CO2/95% nitrogen gas mixture for one hour. Plates were then sealed in an incubation chamber flushed with the same oxygen-free gas mixture. Media pO2 ranged from 115 mmHg following ‘aeration’ to 56 mmHg after 6 h of incubation under reduced oxygen tension (1.8%). Hypoxic medium was exchanged for normoxic medium following hypoxic incubation. All wells were then placed in a normoxic, 37 8C, 5% CO2 incubator for the remainder of the experiment. The third group, Stretch 1 Hypoxia, was exposed to a combined injury of 5.5 mm stretch with either 4 or 6 h of hypoxia. Control cells were not subjected to stretch or hypoxia. For control wells and for wells in the Stretch group, medium was exchanged for conditioned normoxic medium coincident with the first medium exchange in the Hypoxia and Stretch 1 Hypoxia groups. Twenty-four hours after completion of each injury protocol, lactate dehydrogenase (LDH) release and Trypan Blue uptake were assayed. Total protein, via the Bradford Assay, was used as an indexing parameter between wells and experimental series. For LDH, 50 ml of supernatant from each well, with 50 ml of LDH reagent (0.66 mM 2p-iodophenyl-3pnitrophenyl tetrazolium chloride (INT), 54 mM lactic acid, 1.3 mM NAD, 0.28 mM phenazine methosulfate in 0.2 M Tris Buffer), was transferred to a 96-well plate. The plate was incubated in the dark for 20 min, and then read on a spectrophotometer at 490 and 660 nm. LDH data are expressed as absorbance divided by protein concentration. The LDH released into the medium provides an index of cell death and membrane permeability to LDH [10]. For the Trypan Blue assay, 50 ml of 0.2% Trypan Blue was added directly to each culture well. After incubating for 10 min at room temperature, wells were washed three times with PBS and 500 ml 0.8% Triton X-100 was added. After incubating plates at 37 8C for 15 min, lysate was removed from each well and placed in Eppendorf tubes. These samples were sonicated and 250 ml from each was added to a 96-well plate. The plate was read on a spectrophotometer at 595 nm. Trypan Blue data are expressed as absorbance divided by protein concentration. The Trypan Blue measured provides an index of cell death based upon the inability of lethally injured cells to exclude Trypan Blue. Measuring the Trypan Blue spectrophotometrically makes it possible to quantify the results without manual cell counting [15]. Data were analyzed using the Kruskal–Wallis Test and Dunn’s multiple comparison post-hoc analysis. Commercially available software (GraphPad InStat and GraphPad Prizm) was used for the data analysis. Microscopic examination of the cultures was completed before and immediately after injury, and following Trypan Blue staining, before the protein, LDH, and Trypan Blue

assays. The culture procedures yielded mixed neuronal and glial cultures adherent to the collagen substrate, with clumps of neurons of different sizes distributed throughout each culture well (Fig. 1A). Multiple cell processes connect the neuronal clumps (Fig. 1A). Light microscopy revealed that there was minimal uptake of Trypan Blue in Control or the mild Hypoxia or Stretch groups (5.5 mm stretch; 1- or 2h hypoxia). In the intermediate (6.5 mm; 4- or 6-h) and severe (7.5 mm; 24-h) Hypoxia or Stretch, Trypan Blue uptake was observed in the neuronal clumps (Fig. 1B,C). There was essentially no Trypan Blue uptake seen in the cells of the base layer. Similar results were obtained in Stretch 1 Hypoxia. In addition, light microscopy showed

Fig. 1. Photomicrographs of typical cultures at 24 h following injury stained with Trypan Blue. (A) Control: note clumps of neurons with multiple interconnecting processes. This image demonstrates minimal Trypan Blue uptake typical in Control and mild Hypoxia or Stretch groups (5.5 mm stretch, 1- or 2-h hypoxia). Calibration bar, 78.7 mm. (B) Hypoxia (6-h): note Trypan Blue uptake in neurons, while none is seen in the surrounding astrocytic layer. Also, note preservation of neuronal processes. This image is typical of intermediate Stretch or Hypoxia (6.5 mm, 4-, 6-h). Calibration bar, 197 mm. (C) Hypoxia 1 Stretch (5.5 mm 1 6-h): this image demonstrates loss of structural integrity and dissociation of neurons typical of Hypoxia 1 Stretch and severe Hypoxia (24-h) or Stretch (7.5 mm). Again, note presence of Trypan Blue in neurons, but not in astrocytic layer. Calibration bar, 197 mm.

T.F. Glass et al. / Neuroscience Letters 328 (2002) 133–136

that neuronal clumps with greater degrees of Trypan Blue uptake appeared to loose their structural integrity and begin to dissociate by 24 h after injury (Fig. 1C). Neuronal processes are not disturbed following mild Hypoxia or Stretch. Following intermediate Hypoxia or Stretch, the neuronal processes loosen from the adherent base layer of cells, but the processes themselves remain intact up to 24 h. Following severe Hypoxia or Stretch, the neuronal processes are initially preserved, but there is angulation and distortion of the structure of most. By 24 h, most neuronal processes are completely lost (Fig. 1C). This effect is more pronounced following 24-h hypoxia than 7.5 mm stretch. Following Stretch 1 Hypoxia, neuronal processes remained intact but appeared to loosen from the substrate, as seen in intermediate Stretch or Hypoxia. By 24 h following injury, Stretch 1 Hypoxia demonstrated significant loss of neuronal processes and dissociation of neuronal clumps. The ability to observe injury evolution in the structural integrity of neuronal processes, some of which must be axons, as distinct from the neuronal cell body, is an important opportunity for the study of TBI, where evidence of axonal injury or axonal shearing in vivo is particularly prominent [11]. In the Stretch and Hypoxia groups, progressive increases in the degree of injury produced progressively greater LDH

Fig. 2. (a) LDH release at 24 h for Hypoxia or Stretch compared with controls. Compared with controls, 5.5, 6.5 and 7.5 mm, and 4, 6 and 24 h are elevated (P , 0:05). (b) Trypan Blue uptake at 24-h Hypoxia or Stretch. Compared with controls, 5.5, 6.5 and 7.5 mm and 6 and 24 h are elevated (P , 0:05). (c) LDH release at 24 h for Hypoxia or Stretch vs. Hypoxia 1 Stretch. Compared with controls and Stretch or Hypoxia, 5.5 mm 1 4 h and 5.5 mm 1 6 h are elevated (P , 0:05). (d) Trypan Blue uptake at 24 h for Hypoxia or Stretch vs. Hypoxia 1 Stretch. Compared with controls and Stretch or Hypoxia, 5.5 mm 1 4 h and 5.5 mm 1 6 h are elevated (P , 0:05).

135

release compared with controls (Fig. 2a,b). All injury magnitudes in the Stretch group produced a statistically significant increase in LDH release compared with controls (Fig. 2a; P , 0:05). In Hypoxia, while 1- and 2-h hypoxia was not different from controls, there was a significant increase of LDH release following 4 and 6 h of hypoxia, with a much greater increase of LDH following 24 h of hypoxia (Fig. 2a; P , 0:05). Stretch 1 Hypoxia produced greater injury than either alone (Fig. 2c,d). Injury magnitudes of 5.5 mm stretch with 4- or 6-h hypoxia were chosen because each individually produces a small but significant increase of LDH release (Fig. 2a). Fig. 2c compares Stretch (5.5 mm), Hypoxia (4- or 6-h) and Stretch 1 Hypoxia. The data demonstrate that each injury leads to increased LDH release when compared with controls (P , 0:05). There was no statistically significant difference in the Stretch vs. Hypoxia comparison. Comparing Stretch 1 Hypoxia vs. the Hypoxia and Stretch groups, demonstrated increased LDH release (P , 0:05). The Trypan Blue analysis produced results identical to those obtained for LDH release (Fig. 2c,d). Stretch 1 Hypoxia, Stretch (5.5 mm) and Hypoxia (4- or 6-h) led to increased Trypan Blue uptake compared with controls (P , 0:05). Stretch vs. Hypoxia were not statistically different. Stretch 1 Hypoxia led to increased Trypan Blue uptake when compared with the Hypoxia and Stretch groups (Fig. 2d; P , 0:05). There is a linear correlation between LDH and cell death following in vitro neuronal injury [10]. However, hypoxic induction of LDH may change baseline LDH values over time [9,14]. The work of Ellis et al. has also shown that LDH release may correlate with a recoverable injury, not specifically death [8]. Trypan Blue is excluded by living cells, but taken up by non-viable cells. Thus, the Trypan Blue assay was selected to confirm the results of the LDH assay. The Trypan Blue results are identical to the LDH results, indicating that there is increased cell death following Stretch or Hypoxia, with an additional elevation in Stretch 1 Hypoxia. This confirms the effects of combined injury producing increased cell death. The impact of the superimposed insults appears to be additive in our model. Microscopic observations indicate that the Trypan Blue is taken up almost exclusively by neurons. Thus, it is concluded that the source of increased LDH is neuronal as well, as opposed to astrocytic elements of the culture. Direct visualization of the cultures also suggests that the injury sustained by the neurons in this culture system is one that evolves over time. Our model has several advantages over previously described in vitro models of TBI [1,5]. Unlike cellular laceration models, which are most commonly used to model trauma in vitro, intercellular connections between neurons in our model are not immediately disrupted. Like Cargill and Thibault’s model [5], ours can produce a precisely controlled injury, allowing for exploration of

136

T.F. Glass et al. / Neuroscience Letters 328 (2002) 133–136

degrees of mechanical trauma. In addition, it provides a means of delivering graded hypoxic insults as well. Reduction of oxygen tension in the culture medium, instead of chemical blockade of the respiratory chain, should induce injury pathways that are more similar to those seen in clinical hypoxic or ischemic insults. Our model exhibits an exacerbation of mechanical injury when hypoxia is superimposed. This is similar to TBI in the patient, which is characterized by mechanical stresses that distort but do not necessarily disrupt tissues and metabolic stresses that exacerbate the evolution of injury [6,12,13]. This system replicates the combination of mechanical and metabolic insults seen in vivo, and allows investigation of thresholds of activation for various pathways of injury evolution.

[7]

[8]

[9]

[10] [1] Allen, J.W., Knoblach, S.M. and Faden, A.I., Combined mechanical trauma and metabolic impairment in vitro induces NMDA receptor-dependent neuronal cell death and caspase-3-dependent apoptosis, FASEB J., 13(13) (1999) 1875–1882. [2] Boswell, W.C., Boyd, C.R., Schaffner, D., Williams, J.S. and Frantz, E., Prevention of pediatric mortality from trauma: are current measures adequate? South. Med. J., 89(2) (1996) 218–220. [3] Boswell, W.C., Boyd, C.R., Schaffner, D., Williams, J.S. and Frantz, E., Prevention of pediatric mortality from trauma: are current measures adequate? South. Med. J., 89(2) (1996) 218–220. [4] Bouma, G.J., Muizelaar, J.P., Choi, S.C., Newlon, P.G. and Young, H.F., Cerebral circulation and metabolism after severe traumatic brain injury: the elusive role of ischemia, J. Neurosurg., 75(5) (1991) 685–693. [5] Cargill 2nd, R.S. and Thibault, L.E., Acute alterations in [Ca2 1 ]i in NG108-15 cells subjected to high strain rate deformation and chemical hypoxia: an in vitro model for neural trauma, J. Neurotrauma, 13(7) (1996) 395–407. [6] Chesnut, R.M., Marshall, S.B., Piek, J., Blunt, B.A., Klauber, M.R. and Marshall, L.F., Early and late systemic hypoten-

[11]

[12]

[13]

[14]

[15]

sion as a frequent and fundamental source of cerebral ischemia following severe brain injury in the Traumatic Coma Data Bank, Acta Neurochir. Suppl. (Wien), 59 (1993) 121–125. Dietrich, W.D., Alonso, O., Busto, R., Prado, R., Zhao, W., Dewanjee, M.K. and Ginsberg, M.D., Posttraumatic cerebral ischemia after fluid percussion brain injury: an autoradiographic and histopathological study in rats, Neurosurgery, 3 (1998) 585–593 (Discussion pp. 593-594). Ellis, E.F., McKinney, J.S., Willoughby, K.A., Liang, S. and Povlishock, J.T., A new model for rapid stretch-induced injury of cells in culture: characterization of the model using astrocytes, J. Neurotrama, 12(3) (1995) 325–339. Firth, J.D., Ebert, B.L. and Ratcliffe, P.J., Hypoxic regulation of lactate dehydrogenase A: interaction between hypoxiainducible factor 1 and cAMP response elements, J. Biol. Chem., 270 (1995) 21021–21027. Koh, J.Y. and Choi, D.W., Quantitative determination of glutamate mediated cortical neuronal injury in cell culture by LDH efflux assay, J. Neurosci. Methods, 20(1) (1987) 83– 90. Povlishock, J.T. and Jenkins, L.W., Are the pathobiological changes evoked by traumatic brain injury immediate and irreversible? Brain Pathol., 5 (4) (1995) 415–426. Reinert, M., Hoelper, B., Doppenberg, E., Zauner, A. and Bullock, R., Substrate delivery and ionic balance disturbance after severe human head injury, Acta Neurochir. Suppl., 76 (2000) 439–444. Schroder, M.L., Muizelaar, J.P., Bullock, M.R., Salvant, J.B. and Povlishock, J.T., Focal ischemia due to traumatic contusions documented by stable xenon-CT and ultrastructural studies, J. Neurosurg., 6 (1995) 966–971. Semenza, G.L., Roth, P.H., Fang, H.-M. and Wang, G.L., Transcriptional regulation of genes encoding glycolytic enzymes by hypoxia-inducible factor 1, J. Biol. Chem., 269 (1994) 23757–23763. Uliasz, T.F. and Hewett, S.J., A microtiter trypan blue absorbance assay for the quantitative determination of excitotoxic neuronal injury in cell culture, J. Neurosci. Methods, 100(1–2) (2000) 157–163.