JOURNAL
OF SURGICAL
RESEARCH
51.417-424
Fluid Percussion
(1991)
Barotrauma Chamber: A New in Vitro Model for Traumatic Brain Injury
SCOTT R. SHEPAFLD, M.D., JAMSHID SUSAN Aitken
Neurosurgery
Laboratory,
KUPFERMAN,
B. G. GHAJAR,
The New York Hospital-Cornell Submitted
M.D.,
B.A., AND ROBERT Medical
for publication
PH.D.,
J. HARIRI,
ROSEANNE
M.D.,
GIANNUZZI,
B.A.,
PH.D.
Center, 1300 York Avenue, New York, New York 10021
September
4, 1990
as well as to compare the various forms of injury employed in these different models. Although some authors have made efforts to quantitate the injury produced by various models in terms of the kinetic energy delivered to the intact organ [ 11, 121, the specific effect at the cellular level is difficult to interpret. Two new forms of experimental injury were developed, impulse loading and localized loading, to improve our capacity to simulate head injury in the laboratory. Impulse loading involved subjecting the head to acceleration/deceleration forces by setting it in motion without creating a direct impact [14]. The sleds developed by Ommaya and the Penn devices developed by Gennarelli created a system for isolating the effects of angular and/ or translational acceleration [8, 15,161. The Penn II device was able to create specific types of head injury by controlling exposure time to the accelerating force [ 1719]. Short acceleration times resulted in skull fracture and contusion, intermediate acceleration times caused primarily subdural hematomas and contusions, and longer acceleration times resulted in diffuse axonal injury [14, 20-241. Localized loading models, known as fluid percussion Academic I’rew. Inc. devices, involved mechanical loading of the brain with a fluid load [14]. The first such model was developed by Gurdijan and then modified by Lindgren and Rinder, INTRODUCIION who applied brief pulses of high amplitude forces directly to the vertex dura of rabbits [ 25-281. This method and control of the Historically, the study of traumatic brain injury has of injury enabled the quantification amount of force being applied to the brain, by altering had a wealth of animal models to select from. Various the amount of fluid injected into the cranial cavity. Sulmeans were employed to simulate injury, including blasting caps [ 11, air driven pistons [2, 31, hammers [4], livan et al. modified this model in the cat to deliver slightly longer pulses of amplitude equal to that of the pendula [5], sleds [6-81, rotary strikers [9], free falls Lindgren device [29]. In the Sulli,van device, the rate of [lo], spring mounted weights [4], arrows [ 11, and gun fluid flow into the cranial cavity is controlled by the shot missiles [ 11, 121. Although these methods were height and weight of a pendulum that initiates the fluid quite effective in producing various types of experimenflow. While this model does not allow the velocity, acceltal brain injury, determination of the exact forces causeration, or the magnitude of the deformation to be meaing the tissue damage was difficult or impossible. The sured, these variables are directly related to the velocity, different mechanical systems employed in these models acceleration, and volume of fluid injected and can be and various components placed between the origin of the extrapolated from this system. As a result of changing impact and the brain itself made it difficult to accurately and assess the force being transmitted to the brain [13, 141, t,he fluid dynamics in this model, pathophysiological
Advances in the understanding of the pathophysiology of traumatic brain injury have implicated a number of cellular events as fundamental to the evolution of neurologic dysfunction in this process. Following the primary biomechanical insult, a highly complex series of biochemical changes occur, some of which are reversible. The development of fluid percussion injury as an in viva mode1 for traumatic brain injury has greatly improved our ability to study this disease. However, a comparable in vitro model of biomechanical injury which would enable investigators to study the response to injury in isolated cell types has not been described. We have developed a model of transient barotrauma in cell culture to examine the effects of this form of injury on cell metabolism. This mode1 employs the same fluid percussion device commonly used in in vivo brain injury studies. The effect of this injury was evaluated in monolayers of human glial cells. Cell viability by trypan blue exclusion and the production of leukotrienes following increasing barotrauma was investigated. This model provided a reproducible method of subjecting cells in culture to forces similar to those currently used in animal experimental head injury. 6 leei
417
0022-4&304/91
$1.50
Copyright Q 1991 by Academic Press, Inc. All righta of reproduction in any form reserved.
418
JOURNAL
OF SURGICAL
RESEARCH:
morphological alterations proportional to the peak pressures of the fluid load are created. These changes were similar to those seen with human head injury; thus, the fluid percussion device has become an important model for brain injury [ 141. While the impulse and localized loading models provide very effective systems for studying the effects of supratentorial diffuse axonal injury and brainstem injury, these methods do not yield specific information about the metabolic and morphologic effects of injury to the specific constituents of the brain such as the neurons, supporting glia, or the vasculature. When morphologic techniques such as electron microscopy and metabolic studies such as PET scanning are used in conjunction with these models of head injury, some information about what is occurring at the cellular level in head injury may be implied. However, the most effective means of studying the pathophysiologic effects of head injury, specifically, those metabolic alterations responsible for cell death, requires isolation of the individual cellular components of the brain in a well controlled environment such as can be achieved in tissue culture. Unfortunately, there are few studies of the effects of mechanical forces comparable to head injury at the cellular level and as a result there are few techniques described for injuring cells. A common technique employed has been scraping cell monolayers with a sterile pipet, a very poor approximation of the deformation which occurs in head injury. We sought to apply the same forces of deformation created by the fluid percussion model of head injury to monolayers of human glia cells by adapting the fluid percussion device of Sullivan et ,aZ. for use in in vitro tissue culture systems.
MATERIALS
AND METHODS
Fluid Percussion Device We have used the device described by Sullivan et al. Briefly, a plexiglass cylinder 60 cm long with an internal diameter of 4.5 cm and a thickness of 0.5 cm is sealed at one end with a plexiglass stopper covered on its exposed end with a neoprene rubber pad. A second plexiglass cylinder, 7.2 cm long with an internal diameter of 1.1 cm is connected to the first cylinder on one end and to a hollow metal injury screw onthe other end. This system was obtained from the Department of Biomedical Engineering at the Medical College of Virginia. The cylinders are filled with saline at 37°C and an injury is produced when a 4.8-kg weight attached to a pendulum falls from a specific height. The arc of the pendulum is marked off in degrees from 0’ to 90” and impact is measured by a highresolution pressure transducer at the coupler of the fluid percussion device. Pressure within the piston is recorded and converted to atmospheres (1 atm = 760 mm Hg = 14.7 pounds per square inch).
VOL.
51, NO. 5, NOVEMBER
1991 Tygona tubing to connect injury chamber to fluid percussion
Pressure fitting
Rubber gasket
Latex membrane
Recessed
cylindrica Ee6rJ
I
mm petri dish
a I
FIG. 1. Schematic representation of cell injury chamber. As can be seen the chamber accommodates a 60-mm petri dish into a recessed cylindrical area. Nondistendable Tygon tubing is used to connect the cell injury chamber to the fluid percussion device piston.
Cell Injury Device The hollow injury screw is connected by nondistensible Tygon tubing (o.d. f inch; thickness $ inch) to a stainless steel chamber specifically designed to hold the bottom plate of a 60-mm plastic petri dish (Fig. 1). The chamber is composed of two halves, a lower half (Fig. la) which houses the petri dish and an upper half (Fig. lb) that connects to the fluid percussion device (Fig. 2). The lower chamber is a cylinder with a diameter of 60 mm and a height of 65 mm with an extended top to accommodate the upper half of the injury chamber. Within the lower chamber is a recessed cylindrical area 54 mm in diameter and 10 mm in depth which tightly fits the bottom plate of a petri dish. The upper unit of the injury chamber has a diameter of 67 mm at the point at which it contacts the lower chamber, and then gradually tapers to a male adapter that fits the tubing. Within the male type adapter is a circular hollowing of 7 mm diameter which allows free flow of the saline from the fluid percussion device. As can be seen in Fig. 1, two rubber gaskets (0.9 mm thick, o.d. 66 mm, i.d. 62 mm) seal the connection between the upper and lower halves of the chamber. The upper gasket contains a waterproof membrane composed of latex, which seals the petri dish and prevents mixing of the saline from the fluid percussion device
SHEPARD
\
Cell injury chamber
ET AL.: IN VITRO FLUID
Penduium frame
FIG. 2. Depicted here is the fluid percussion device system developed for in uiuo simulation of traumatic brain injury. The cell injury chamber is coupled to the fluid percussion piston. Pressures within the system are monitored at the coupler between the chamber and the fluid percussion device.
with the media in the petri dish, yet transmits the fluid pulse between the upper and lower chambers. A clamp consisting of a circular sheet of stainless of 1 mm thickness with an inner C-shaped metal piece is outfitted with adapters for a screw and wing nut and placed around the connection between the upper and lower halves of the cell injury unit to keep these halves tightly opposed. Cell Cultures Primary human glial lines were established from normal brain samples obtained at surgery from patients requiring resection of normal tissue for access to underlying structures. Samples were placed in normal saline and immediately transported to the laboratory. The meninges and any visible vasculature were carefully dissected and removed under sterile conditions. The tissue was then finely minced with iridectomy scissors into l-mm cubes and washed three times with PBS (phosphatebuffered saline), and the samples were collected by centrifugation at 12Og for 5 min after each washing. After the final washing, the tissue samples were plated in 75 cm flasks (Falcon) with 10 ml of Waymouth’s MB 752/l supplemented with 20% fetal bovine serum, 1% penicillin (5000 u/ml), streptomycin (5000 pg/ml) , and 1% fungizone (250 pg/ml). (Media, serum, and all other tissue culture reagents were obtained from GIBCO, Grand Island, NY). Cells were incubated at 37°C in 5% CO,, grown to confluent monolayers, and subcultivated as described previously [ 301. Evaluation of Physical Forces Generated by Cell Culture Barotrauma Device: Calibration Experiments aimed at quantifying the physical forces produced by this model on cell monolayers were per-
PERCUSSION
BAROTRAUMA
419
MODEL
formed. As can be seen in Fig. 1, the upper half of the cell injury chamber is designed with a pressure fitting which allows the placement of a high-resolution, piezo-resistive pressure transducer to measure changes in pressure within this chamber. In addition, the fluid percussion device is outfitted with a high-speed pressure transducer (P23, Statham, Puerto Rey, PR) at the distal end of the fluid-filled piston (Fig. 2). Both transducers were connected to an analog to digital converter (Maclab, AD Instruments, Milford, MA.) and computer data acquisition system. Pressure transducers were calibrated against mercury-filled manometers prior to these experiments. The chamber was coupled to the fluid percussion device by 10 cm of nondistensible Tygon tubing and joints were sealed with locking nylon ties. Air was eliminated from the system and the piston percussed by the weighted pendulum at varying angles of incidence. Pressures within the cell barotrauma chamber and piston were recorded. Incident angles of pendulum swing from 10” to 45’ were studied in 5” increments. A total of five consecutive piston percussions were studied at each increment. The system was zeroed between each piston percussion and care was taken to prevent leakage of fluid from the system during this period. Peak pressures and impulse duration were examined. Effect of Fluid Percussion Injury
on Cell Morwlayers
Cell viability after fluid percussion was studied on human astroglial cell monolayers. Confluent cell monolayers in 60-mm petri dishes were employed. Prior to placement in the device, 200-~1 aliquots of supernatant media were aspirated and cell counts obtained on a standard hemocytometer to obtained baseline nonattached cell concentrations. Petri dishes were then placed within the chamber and subjected to 20- to 30-msec barotraumatic injuries between 0.5 and 4 atmospheres above ambient pressure. Supernatant cell counts were obtained to determine the percentage of detached cells. The uptake of the vital stain trypan blue by these detached cells was also determined and the percentage of stained cells versus unstained cells was calculated as an index of cell viability. Leukotriene Astroglial
C, Production Cells
by Barotraumatically
Injured
We have recently shown that human glial cells generate lipoxygenase products following injury. To investigate whether this model resulted in detectable levels of these compounds, leukotriene C, (LTC,) production by human astroglial cells was measured as a function of barotraumatic injury expressed as angle of incidence of the fluid percussion device and atmospheric injury. Culture were maintained at 37°C throughout this experiment by use of a heated water bath. Supernatants were obtained from cultures subjected to barotraumatic inju-
420
JOURNAL
OF SURGICAL
RESEARCH:
VOL.
51, NO. 5, NOVEMBER
1991
The maximum pressure exerted by this system ranged from 1.2 atmospheres at a pendulum angle of 10” to 5.7 atm at a pendulum angle of 45” (Table 1). Duration of each pressure impulse was also monitored and ranged from 20 to 30 msec. The duration of each pressure impulse varied inversely with the pendulum angle with the smallest angles resulting in the shorter impulse durations than the larger pendulum angles. Fluid Percussion Injury
0
5
10
Time
(Milliseconds1
is
20
FIG. 3.
A representative pressure tracing derived from data recorded during these experiments. Note the impulse duration of approximately 20 msec. Secondary, smaller peak probably represents some residual elasticity within the system.
ries ranging from approximately 20 to 50” (approximately 2 to 6 atmospheres) 15 min following injury. LTC, levels were measured by means of a competitive displacement, enzyme-linked immunoassay (EIA) (Cayman Chemical, Ann Arbor, MI) and quantitated against a standard curve generated from known concentrations of immunoreactive LTC, and expressed as pg/ml and pg/lOO,OOO cells.
of Cell Monolayers
The effect of barotraumatic injury on glial cell monolayers was also evaluated. Detachment of cells from the petri dish occurred as a function of the magnitude of the barotrauma these monolayers were subjected to. As can be seen in Fig. 6A, confluent glial cell monolayers adhere tightly to the surface of the petri dish. Following barotrauma, the injured monolayers (Fig. 6B) lift off the surface of the dish and detached cells are detected in the supernatants. The percentage of cells detached from the monolayers adherent to the bottom of the petri dish increased as a function of the magnitude of the barotraumatic injury. Cell counts were obtained from supernatants following increasing barotraumatic injury. Detached cells were quantitated by counting total cells by light microscopic cytometry in 100~~1 aliquots of supernatant obtained 5 min following injury. Total detached cells were expressed as a function of total supernatant volume. Adherent cells were then counted using a similar technique after trypsin dissociation. The percentage
RESULTS
Calibration and Evaluation of Physical Forces Generated by Cell Culture Barotrauma Device Pressure dynamics characteristic of this model were evaluated by recording pressure within the chamber as well as within the fluid percussion piston. Our experience indicated that chamber and piston pressures were virtually identical and since all pressure tracings presented herein are derived from the piston transducer, all pressure data used in this study are based on piston recordings. A representative pressure tracing is seen in Fig. 3. All impulses lasted between 20 and 30 msec. Calibration of the system against mercury manometry yielded an equation from which we calculated chamber pressure from voltage differentials provided by the piston transducer (Fig. 4). This equation is Y = -83.8784
600
500
400
?I 300
200
100
0
+ 0.13X (r = 0.95).
Chamber pressures resulting from increasing angle of incidence are shown represented as impulse magnitude in millivolts recorded by the piston transducer as well as converted to atmospheres (Fig. 5 and Table 1).
FIG. 4. Calibration of fluid percussion system. Data recorded from the high-speed piezo resistive pressure transducer measured in millivolts is calibrated against known mercury manometry. This relationship is then used to calculate peak pressures within the chamber during each fluid percussion.
SHEPARD
ET AL.: IN
VI7’RO
FLUID
of detached cells was calculated from the ratio of total detached cells divided by the sum of detached and adherent cells. Cell viability was calculated in the supernatant cell suspensions by counting the number of cells taking up the vital stain trypan blue as a function of the total number of detached cells and expressed as a ratio. As can be seen in Fig. 7, cell detachment increased with increasing magnitude of barotraumatic injury. The percentage of detached cells increased dramatically when injury was in excess of 2.5 atmospheres. In addition, cell viability appears to be relatively well maintained below 3 atmospheres of barotrauma (greater than 90% viability), yet nonviability rises rapidly to greater than 80% nonviable at barotrauma approaching 6 atmospheres. Leukotriene Production Monolayers
by Barotraumutized
PERCUSSION
BAROTRAUMA
TABLE Angle of incidence 10 15 20 25 30 35 40 45
1
Impulse (mV)
Pressure (mm I-k)
37.26 63.54 95.36 122.54 157.28 202.23 256.26 476.4
931.8 1133.69 1378.7 1587.8 1852.9 2201 2616.4 4309.7
+ f k -t + + + +
421
MODEL
5.87 3.14 11.26 6.31 15.08 15.81 28.17 62.27
Atmospheres 1.2 1.49 1.81 2.09 2.44 2.9 3.44 5.67
production occurred rapidly after injury, with peak levels of LTC, appearing within 15 min after injury and rapidly declining 30 to 60 min post injury.
Glial Cell
Evaluating leukotriene C4 production as an independent index of glial cell injury yielded the following results. Fifteen minutes after injury the astroglia produced LTC, in quantities ranging from 20 pg/ml (500 pg/106 cells) at a pendulum angle of 20” to 180 pg/ml (4.5 ng/ lo6 cells) at a pendulum angle of 30” (Fig. 8). Time course analysis of LTC, production after injury demonstrated that LTC, levels peaked at 15 min following barotrauma and this time point was selected for study. LTC,
FIG. 5. Chamber pressure and impulse magnitude vs pendulum angle. As is apparent in this graph, pressure within the chamber is logarithmically related to the angle of incidence of the pendulum. Both impulse magnitude recorded in millivolts and chamber pressure recorded in atmospheres are shown.
DISCUSSION
This adaptation of the fluid percussion system represents a powerful method for creating a reliable, highly reproducible injury to cell monolayers. Rigorous testing of this system has demonstrated a close correlation between peak chamber pressures and the pendulum angle from which the 4.8-kg weight is swung to impact the fluid filled piston. Thus, by setting the pendulum angle, a reproducible specified level of injury is obtained. Previous studies of fluid percussion systems have demonstrated that the level of injury produced correlates with the peak pressure exerted by the system rather than with the mean pressure exerted by the system. Two means were employed to demonstrate the biological effects of injury by this system. First, the percentage of cells detached from the cell monolayer as a function of the degree of injury was examined and the resulting percentage of nonviable cells as a function of the degree of injury were examined. Both of these parameters demonstrated direct, almost linear relationships with the degree of injury, demonstrating that increased peak pressures experienced by the cells results in increased levels of injury to the cells. Second, LTC, production as a function of cell injury was examined. As the degree of injury increased, the cells produced greater amounts of LTC,, suggesting a relationship between the degree of injury and the level of LTC, produced. Interestingly, the production of LTC, sharply declines at pendulum angles greater than 30”. It is possible that excessive injury produced at higher angles results in irreversible damage to the metabolic machinery necessary to generate lipoxygenase products from liberated arachidonic acid. LTC, production was employed only as an independent index of cell injury and more in depth analysis will not be presented herein. The exact mechanism by which barotrauma causes cell injury is not yet defined. Plausible considerations
FIG. 6. (A) A light micrograph of glial cell monolayers after barotrauma. (Magnification 400X).
in culture.
(B) The detachment
422
of the monolayer
from the surface of the petri dish
SHEPARD
ET AL.: IN
VITRO
FLUID
PERCUSSION
include mechanical disruption of cell membranes, particularly a direct effect on phospholipid bilayers. Perhaps more likely would be a direct effect on integral membrane proteins, especially those membrane-bound ATPases responsible for ionic homeostasis. Since activity of these proteins is dependent upon allosteric relationships within the phospholipid bilayer, mechanical forces may alter electrostatic relationships and render these molecules temporarily inactive. Another possible injurious effect would be heat generation within the cell chamber as a result of rapid compression of the chamber contents. Although we took every step to insure the chamber temperature was maintained at 37°C during the experiment, we were technically limited in our ability to follow temperatures at the level of the cell monolayers. This simple system for the in vitro study of the effects of mechanical injury on cell culture systems using forces comparable to those employed in our most widely used animal model may enhance our capacity to investigate the cellular effects of mechanical injury on isolated cell populations. By isolating the effects of barotraumatic injury upon individual components of the nervous system, the effects of injury on specific cell types may be investigated. Also, this system allows for the study of certain aspects of barotraumatic injury without the need to sacrifice animals and is more economical than in uivo systems. Furthermore, the effects of various pharmacologic and metabolic interventions may be investigated.
1OC
BAROTRAUMA
423
MODEL 5000
2co
4000
3000
1 j$
8. 100 I
5 s
2000
e r;l
1000
0
I
10
20
-
I
30
.
I
40
.
I
50
0
.
60
Pendulum Angle FIG. 8. Production of leukotrienes C, as measured by enzyme linked amino assay and recorded as a function of the pendulum angle used to induce barotraumatic injury is shown. Leukotriene production recorded in picograms per milliliter as well as picograms per one hundred thousand cells is represented. Leukotriene production is recorded solely as an additional index of cell injury.
8C
REFERENCES 6C
1. 2. 3.
4.
5. 6.
7.
FIG. 7. The effect of harotrauma on glial cell monolayers. Both the percentage of detached cells and the percentage of nonviable cells as a function of atmospheres of barotraumatic injury are represented here.
8.
Govons, S. R., Govons, R. B., Van Huss, W. D., and Heusaer, W. W. Brain concussion in the rat. Exp. Neural. 34: 121,1972. Bergren, D. R., and Beckman, D. L. Pulmonary surface tension and head injury. J. Trauma 15: 336,1975. Ommaya, A. K., Flamm, E. S., and Mahone, R. H. Cerebral concussion in the monkey: An experimental model. Science 153: 211, 1966. Gurdjian, E. S., and Webster, J. F. Experimental head injury with reference to the mechanical factors in acute trauma. Surg. Gynecol. Obstet. 76: 623, 1943. Denny-Brown, D., and Russell, W. R. Experimental cerebral concussion. Brain 64: 93, 1941. Tsubolawa, T., Nakamura, S., Hayashi, N., et al. Experimental primary fatal head injury caused by linear acceleration-biomechanics and pathogenesis. Neural. Med. Chir. 15: 57, 1975. Kobrine, A., and Kempe, L. G. Studies in head injury. Part 1: An experimental model of closed head injury. Surg. Neurol. 1: 34, 1973. Ommaya, A. K. Experimental head injury in the monkey. In W. F. Cavenes and A. E. Walker (Eds.), Head Injury Conference Proceedings, 1966. Pp. 260-275.
424 9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
JOURNAL
OF SURGICAL
RESEARCH
Chason, J. L., Fernand, 0. V., and Hodgson, V. R. Experimental brain concussion morphogenic findings and a new cytologic hypothesis. J Trauma 6: 767, 1966. Huger, F., and Patrick, G. Effect of concussive head injury on central catecholamine levels and synthesis rates in rat brain regions. J. Neurochem. 33: 89, 1979. Crockard, H. A., Brown, F. D., Calica, A. B., Johns, L. M., and Mullan, S. Physiological consequences of experimental cerebral missile injury and use of data analysis to predict survival. J. Neurosurg. 46: 784, 1977. Crockard, H. A., Brown, F. D., Johns, L. M., and Mullan, S. An experimental cerebral missile injury model in primates. J. Neurosurg. 46: 776,1977. Hayes, R. L., and Ellison, M. D. Animal models of concussive head injury. In Becker and Gudeman (Eds.), Head Injury, Philadelphia: W.B. Saunders, 1989. Gennarelli, T. A., and Thibault, L. E. Biological models of head injury. In D. P. Becker and J. T. Povlishock (Eds.), NZH Central Nervous System Trauma Status Report, 1985. Pp. 103-106. Ommaya, A. K., Corrao, P., and Letcher, F. S. Head injury in the chimpanzee. Part 1: Biodynamics of traumatic unconsciousness. J. Neurosurg. 39: 152, 1973. Ommaya, A. K., Grubb, R. L., and Naumann, R. A. Coup and contrecoup injury: Observations on the mechanics of visible brain injuries in the rhesus monkey. J. Neurosurg. 35: 503,197l. Gennarelli, T. A., Adams, J. H., and Graham, D. I. Acceleration induced head injury in the monkey: Neuropathology. Appl. Neurobiol. 6: 234, 1980. Gennarelli, T. A., Adams, J. H., Graham, D. I., and Jellinger, K. Acceleration induced head injury in the monkey: The model its mechanical and physiological correlates. Acta Neuropath. Suppl. VII, 23, 1981. Gennarelli, T. A., Seqawa, H., Wald, V., et al. Physiological response to angular acceleration of the head. In R. G. Grossman and P. L. Gildenbere (Eds.). Seminars in Head Zniurv: - “, Basic and Clinical Aspects. 1982: Pp. 129-140. Gennarelli, T. A., Thibault, L. E., Adams, J. H., et al. Diffuse axonal injury and traumatic coma in the primate. Ann. Neurol. 12: 564, 1982.
VOL.
51, NO. 5, NOVEMBER
1991
21.
Gennarelli, son of linear concussion. SAE. 1971.
22.
Gennarelli, T. A., Thibault, L. E., and Ommaya, A. K. Pathophysiologic responses to rotational and translational acceleration of the head. In 16th Stapp Car Crash Conference Proceedings SAE. 1972. Pp. 296-308.
23.
Gennarelli, T. A., and Thibault, L. E. Experimental production of prolonged traumatic coma in the primate. In R. Villiani et al. (Eds.), Advances in Neurotraumatology. 1982. Pp. 31-33.
24.
Gurdijian, E. S., Hodgson, V. R., Thomas, L. M., et al. Highspeed techniques in head injury research. Demonstration of relative movements of scalp, skull and intracranial contents during impact. Med. Sci. 18: 45, 1967.
25.
Chason, J. L., Hardy, W. G., Webster, J. E., and Gurdjian, E. S. Alterations in cell structure of the brain associated with experimental concussion. J. Neurosurg. 16: 135, 1958.
26.
Lewett, W., Jenkins, L. W., and Miller, J. D. Effects of experimental fluid percussion injury of the brain on cerebrovascular reactivity to hypoxia and to hypercapnia. J. Neurosurg. 66: 332, 1982.
27.
Lindgren, S., and Rinder, L. Experimental studies in head injury. I: Some factors influencing results of model experiments. Biophysik 3: 320,1965.
28.
Rinder, L., and Olsson, Y. Studies on vascular permeability changes in experimental brain concussion. I: Distribution of circulating fluorescent indicators in brain and cervical cord after sudden mechanical loading of the brain. Acta Neuropathol. 11: 183, 1968.
29.
Sullivan, H. G., Martinez, sion model of mechanical 45: 520, 1976.
30.
Hariri, R. J., Ghajar, J. B. G., Pomerantz, K. B., Haiar, D. P., Giannuzzi, R. F., Tomich, E., Andrews, D. W., Patterson, R. H. Human glial cell production of lipoxygenase generated eicosanoids: A potential role in the pathophysiology of vascular changes following traumatic brain injury. J. Trauma 29(g): 1203, 1989.
T. A., Ommaya, A. K., and Thibault, L. E. Compariand rotational acceleration in experimental cerebral In 15th Stapp Car Crash Conference Proceedings Pp. 797-803.
J., Becker, D. P., et al. Fluid-percusbrain injury in the cat. J. Neurosurg.
OF SURGICAL
RESEARCH
51.417-424
Fluid Percussion
(1991)
Barotrauma Chamber: A New in Vitro Model for Traumatic Brain Injury
SCOTT R. SHEPAFLD, M.D., JAMSHID SUSAN Aitken
Neurosurgery
Laboratory,
KUPFERMAN,
B. G. GHAJAR,
The New York Hospital-Cornell Submitted
M.D.,
B.A., AND ROBERT Medical
for publication
PH.D.,
J. HARIRI,
ROSEANNE
M.D.,
GIANNUZZI,
B.A.,
PH.D.
Center, 1300 York Avenue, New York, New York 10021
September
4, 1990
as well as to compare the various forms of injury employed in these different models. Although some authors have made efforts to quantitate the injury produced by various models in terms of the kinetic energy delivered to the intact organ [ 11, 121, the specific effect at the cellular level is difficult to interpret. Two new forms of experimental injury were developed, impulse loading and localized loading, to improve our capacity to simulate head injury in the laboratory. Impulse loading involved subjecting the head to acceleration/deceleration forces by setting it in motion without creating a direct impact [14]. The sleds developed by Ommaya and the Penn devices developed by Gennarelli created a system for isolating the effects of angular and/ or translational acceleration [8, 15,161. The Penn II device was able to create specific types of head injury by controlling exposure time to the accelerating force [ 1719]. Short acceleration times resulted in skull fracture and contusion, intermediate acceleration times caused primarily subdural hematomas and contusions, and longer acceleration times resulted in diffuse axonal injury [14, 20-241. Localized loading models, known as fluid percussion Academic I’rew. Inc. devices, involved mechanical loading of the brain with a fluid load [14]. The first such model was developed by Gurdijan and then modified by Lindgren and Rinder, INTRODUCIION who applied brief pulses of high amplitude forces directly to the vertex dura of rabbits [ 25-281. This method and control of the Historically, the study of traumatic brain injury has of injury enabled the quantification amount of force being applied to the brain, by altering had a wealth of animal models to select from. Various the amount of fluid injected into the cranial cavity. Sulmeans were employed to simulate injury, including blasting caps [ 11, air driven pistons [2, 31, hammers [4], livan et al. modified this model in the cat to deliver slightly longer pulses of amplitude equal to that of the pendula [5], sleds [6-81, rotary strikers [9], free falls Lindgren device [29]. In the Sulli,van device, the rate of [lo], spring mounted weights [4], arrows [ 11, and gun fluid flow into the cranial cavity is controlled by the shot missiles [ 11, 121. Although these methods were height and weight of a pendulum that initiates the fluid quite effective in producing various types of experimenflow. While this model does not allow the velocity, acceltal brain injury, determination of the exact forces causeration, or the magnitude of the deformation to be meaing the tissue damage was difficult or impossible. The sured, these variables are directly related to the velocity, different mechanical systems employed in these models acceleration, and volume of fluid injected and can be and various components placed between the origin of the extrapolated from this system. As a result of changing impact and the brain itself made it difficult to accurately and assess the force being transmitted to the brain [13, 141, t,he fluid dynamics in this model, pathophysiological
Advances in the understanding of the pathophysiology of traumatic brain injury have implicated a number of cellular events as fundamental to the evolution of neurologic dysfunction in this process. Following the primary biomechanical insult, a highly complex series of biochemical changes occur, some of which are reversible. The development of fluid percussion injury as an in viva mode1 for traumatic brain injury has greatly improved our ability to study this disease. However, a comparable in vitro model of biomechanical injury which would enable investigators to study the response to injury in isolated cell types has not been described. We have developed a model of transient barotrauma in cell culture to examine the effects of this form of injury on cell metabolism. This mode1 employs the same fluid percussion device commonly used in in vivo brain injury studies. The effect of this injury was evaluated in monolayers of human glial cells. Cell viability by trypan blue exclusion and the production of leukotrienes following increasing barotrauma was investigated. This model provided a reproducible method of subjecting cells in culture to forces similar to those currently used in animal experimental head injury. 6 leei
417
0022-4&304/91
$1.50
Copyright Q 1991 by Academic Press, Inc. All righta of reproduction in any form reserved.
418
JOURNAL
OF SURGICAL
RESEARCH:
morphological alterations proportional to the peak pressures of the fluid load are created. These changes were similar to those seen with human head injury; thus, the fluid percussion device has become an important model for brain injury [ 141. While the impulse and localized loading models provide very effective systems for studying the effects of supratentorial diffuse axonal injury and brainstem injury, these methods do not yield specific information about the metabolic and morphologic effects of injury to the specific constituents of the brain such as the neurons, supporting glia, or the vasculature. When morphologic techniques such as electron microscopy and metabolic studies such as PET scanning are used in conjunction with these models of head injury, some information about what is occurring at the cellular level in head injury may be implied. However, the most effective means of studying the pathophysiologic effects of head injury, specifically, those metabolic alterations responsible for cell death, requires isolation of the individual cellular components of the brain in a well controlled environment such as can be achieved in tissue culture. Unfortunately, there are few studies of the effects of mechanical forces comparable to head injury at the cellular level and as a result there are few techniques described for injuring cells. A common technique employed has been scraping cell monolayers with a sterile pipet, a very poor approximation of the deformation which occurs in head injury. We sought to apply the same forces of deformation created by the fluid percussion model of head injury to monolayers of human glia cells by adapting the fluid percussion device of Sullivan et ,aZ. for use in in vitro tissue culture systems.
MATERIALS
AND METHODS
Fluid Percussion Device We have used the device described by Sullivan et al. Briefly, a plexiglass cylinder 60 cm long with an internal diameter of 4.5 cm and a thickness of 0.5 cm is sealed at one end with a plexiglass stopper covered on its exposed end with a neoprene rubber pad. A second plexiglass cylinder, 7.2 cm long with an internal diameter of 1.1 cm is connected to the first cylinder on one end and to a hollow metal injury screw onthe other end. This system was obtained from the Department of Biomedical Engineering at the Medical College of Virginia. The cylinders are filled with saline at 37°C and an injury is produced when a 4.8-kg weight attached to a pendulum falls from a specific height. The arc of the pendulum is marked off in degrees from 0’ to 90” and impact is measured by a highresolution pressure transducer at the coupler of the fluid percussion device. Pressure within the piston is recorded and converted to atmospheres (1 atm = 760 mm Hg = 14.7 pounds per square inch).
VOL.
51, NO. 5, NOVEMBER
1991 Tygona tubing to connect injury chamber to fluid percussion
Pressure fitting
Rubber gasket
Latex membrane
Recessed
cylindrica Ee6rJ
I
mm petri dish
a I
FIG. 1. Schematic representation of cell injury chamber. As can be seen the chamber accommodates a 60-mm petri dish into a recessed cylindrical area. Nondistendable Tygon tubing is used to connect the cell injury chamber to the fluid percussion device piston.
Cell Injury Device The hollow injury screw is connected by nondistensible Tygon tubing (o.d. f inch; thickness $ inch) to a stainless steel chamber specifically designed to hold the bottom plate of a 60-mm plastic petri dish (Fig. 1). The chamber is composed of two halves, a lower half (Fig. la) which houses the petri dish and an upper half (Fig. lb) that connects to the fluid percussion device (Fig. 2). The lower chamber is a cylinder with a diameter of 60 mm and a height of 65 mm with an extended top to accommodate the upper half of the injury chamber. Within the lower chamber is a recessed cylindrical area 54 mm in diameter and 10 mm in depth which tightly fits the bottom plate of a petri dish. The upper unit of the injury chamber has a diameter of 67 mm at the point at which it contacts the lower chamber, and then gradually tapers to a male adapter that fits the tubing. Within the male type adapter is a circular hollowing of 7 mm diameter which allows free flow of the saline from the fluid percussion device. As can be seen in Fig. 1, two rubber gaskets (0.9 mm thick, o.d. 66 mm, i.d. 62 mm) seal the connection between the upper and lower halves of the chamber. The upper gasket contains a waterproof membrane composed of latex, which seals the petri dish and prevents mixing of the saline from the fluid percussion device
SHEPARD
\
Cell injury chamber
ET AL.: IN VITRO FLUID
Penduium frame
FIG. 2. Depicted here is the fluid percussion device system developed for in uiuo simulation of traumatic brain injury. The cell injury chamber is coupled to the fluid percussion piston. Pressures within the system are monitored at the coupler between the chamber and the fluid percussion device.
with the media in the petri dish, yet transmits the fluid pulse between the upper and lower chambers. A clamp consisting of a circular sheet of stainless of 1 mm thickness with an inner C-shaped metal piece is outfitted with adapters for a screw and wing nut and placed around the connection between the upper and lower halves of the cell injury unit to keep these halves tightly opposed. Cell Cultures Primary human glial lines were established from normal brain samples obtained at surgery from patients requiring resection of normal tissue for access to underlying structures. Samples were placed in normal saline and immediately transported to the laboratory. The meninges and any visible vasculature were carefully dissected and removed under sterile conditions. The tissue was then finely minced with iridectomy scissors into l-mm cubes and washed three times with PBS (phosphatebuffered saline), and the samples were collected by centrifugation at 12Og for 5 min after each washing. After the final washing, the tissue samples were plated in 75 cm flasks (Falcon) with 10 ml of Waymouth’s MB 752/l supplemented with 20% fetal bovine serum, 1% penicillin (5000 u/ml), streptomycin (5000 pg/ml) , and 1% fungizone (250 pg/ml). (Media, serum, and all other tissue culture reagents were obtained from GIBCO, Grand Island, NY). Cells were incubated at 37°C in 5% CO,, grown to confluent monolayers, and subcultivated as described previously [ 301. Evaluation of Physical Forces Generated by Cell Culture Barotrauma Device: Calibration Experiments aimed at quantifying the physical forces produced by this model on cell monolayers were per-
PERCUSSION
BAROTRAUMA
419
MODEL
formed. As can be seen in Fig. 1, the upper half of the cell injury chamber is designed with a pressure fitting which allows the placement of a high-resolution, piezo-resistive pressure transducer to measure changes in pressure within this chamber. In addition, the fluid percussion device is outfitted with a high-speed pressure transducer (P23, Statham, Puerto Rey, PR) at the distal end of the fluid-filled piston (Fig. 2). Both transducers were connected to an analog to digital converter (Maclab, AD Instruments, Milford, MA.) and computer data acquisition system. Pressure transducers were calibrated against mercury-filled manometers prior to these experiments. The chamber was coupled to the fluid percussion device by 10 cm of nondistensible Tygon tubing and joints were sealed with locking nylon ties. Air was eliminated from the system and the piston percussed by the weighted pendulum at varying angles of incidence. Pressures within the cell barotrauma chamber and piston were recorded. Incident angles of pendulum swing from 10” to 45’ were studied in 5” increments. A total of five consecutive piston percussions were studied at each increment. The system was zeroed between each piston percussion and care was taken to prevent leakage of fluid from the system during this period. Peak pressures and impulse duration were examined. Effect of Fluid Percussion Injury
on Cell Morwlayers
Cell viability after fluid percussion was studied on human astroglial cell monolayers. Confluent cell monolayers in 60-mm petri dishes were employed. Prior to placement in the device, 200-~1 aliquots of supernatant media were aspirated and cell counts obtained on a standard hemocytometer to obtained baseline nonattached cell concentrations. Petri dishes were then placed within the chamber and subjected to 20- to 30-msec barotraumatic injuries between 0.5 and 4 atmospheres above ambient pressure. Supernatant cell counts were obtained to determine the percentage of detached cells. The uptake of the vital stain trypan blue by these detached cells was also determined and the percentage of stained cells versus unstained cells was calculated as an index of cell viability. Leukotriene Astroglial
C, Production Cells
by Barotraumatically
Injured
We have recently shown that human glial cells generate lipoxygenase products following injury. To investigate whether this model resulted in detectable levels of these compounds, leukotriene C, (LTC,) production by human astroglial cells was measured as a function of barotraumatic injury expressed as angle of incidence of the fluid percussion device and atmospheric injury. Culture were maintained at 37°C throughout this experiment by use of a heated water bath. Supernatants were obtained from cultures subjected to barotraumatic inju-
420
JOURNAL
OF SURGICAL
RESEARCH:
VOL.
51, NO. 5, NOVEMBER
1991
The maximum pressure exerted by this system ranged from 1.2 atmospheres at a pendulum angle of 10” to 5.7 atm at a pendulum angle of 45” (Table 1). Duration of each pressure impulse was also monitored and ranged from 20 to 30 msec. The duration of each pressure impulse varied inversely with the pendulum angle with the smallest angles resulting in the shorter impulse durations than the larger pendulum angles. Fluid Percussion Injury
0
5
10
Time
(Milliseconds1
is
20
FIG. 3.
A representative pressure tracing derived from data recorded during these experiments. Note the impulse duration of approximately 20 msec. Secondary, smaller peak probably represents some residual elasticity within the system.
ries ranging from approximately 20 to 50” (approximately 2 to 6 atmospheres) 15 min following injury. LTC, levels were measured by means of a competitive displacement, enzyme-linked immunoassay (EIA) (Cayman Chemical, Ann Arbor, MI) and quantitated against a standard curve generated from known concentrations of immunoreactive LTC, and expressed as pg/ml and pg/lOO,OOO cells.
of Cell Monolayers
The effect of barotraumatic injury on glial cell monolayers was also evaluated. Detachment of cells from the petri dish occurred as a function of the magnitude of the barotrauma these monolayers were subjected to. As can be seen in Fig. 6A, confluent glial cell monolayers adhere tightly to the surface of the petri dish. Following barotrauma, the injured monolayers (Fig. 6B) lift off the surface of the dish and detached cells are detected in the supernatants. The percentage of cells detached from the monolayers adherent to the bottom of the petri dish increased as a function of the magnitude of the barotraumatic injury. Cell counts were obtained from supernatants following increasing barotraumatic injury. Detached cells were quantitated by counting total cells by light microscopic cytometry in 100~~1 aliquots of supernatant obtained 5 min following injury. Total detached cells were expressed as a function of total supernatant volume. Adherent cells were then counted using a similar technique after trypsin dissociation. The percentage
RESULTS
Calibration and Evaluation of Physical Forces Generated by Cell Culture Barotrauma Device Pressure dynamics characteristic of this model were evaluated by recording pressure within the chamber as well as within the fluid percussion piston. Our experience indicated that chamber and piston pressures were virtually identical and since all pressure tracings presented herein are derived from the piston transducer, all pressure data used in this study are based on piston recordings. A representative pressure tracing is seen in Fig. 3. All impulses lasted between 20 and 30 msec. Calibration of the system against mercury manometry yielded an equation from which we calculated chamber pressure from voltage differentials provided by the piston transducer (Fig. 4). This equation is Y = -83.8784
600
500
400
?I 300
200
100
0
+ 0.13X (r = 0.95).
Chamber pressures resulting from increasing angle of incidence are shown represented as impulse magnitude in millivolts recorded by the piston transducer as well as converted to atmospheres (Fig. 5 and Table 1).
FIG. 4. Calibration of fluid percussion system. Data recorded from the high-speed piezo resistive pressure transducer measured in millivolts is calibrated against known mercury manometry. This relationship is then used to calculate peak pressures within the chamber during each fluid percussion.
SHEPARD
ET AL.: IN
VI7’RO
FLUID
of detached cells was calculated from the ratio of total detached cells divided by the sum of detached and adherent cells. Cell viability was calculated in the supernatant cell suspensions by counting the number of cells taking up the vital stain trypan blue as a function of the total number of detached cells and expressed as a ratio. As can be seen in Fig. 7, cell detachment increased with increasing magnitude of barotraumatic injury. The percentage of detached cells increased dramatically when injury was in excess of 2.5 atmospheres. In addition, cell viability appears to be relatively well maintained below 3 atmospheres of barotrauma (greater than 90% viability), yet nonviability rises rapidly to greater than 80% nonviable at barotrauma approaching 6 atmospheres. Leukotriene Production Monolayers
by Barotraumutized
PERCUSSION
BAROTRAUMA
TABLE Angle of incidence 10 15 20 25 30 35 40 45
1
Impulse (mV)
Pressure (mm I-k)
37.26 63.54 95.36 122.54 157.28 202.23 256.26 476.4
931.8 1133.69 1378.7 1587.8 1852.9 2201 2616.4 4309.7
+ f k -t + + + +
421
MODEL
5.87 3.14 11.26 6.31 15.08 15.81 28.17 62.27
Atmospheres 1.2 1.49 1.81 2.09 2.44 2.9 3.44 5.67
production occurred rapidly after injury, with peak levels of LTC, appearing within 15 min after injury and rapidly declining 30 to 60 min post injury.
Glial Cell
Evaluating leukotriene C4 production as an independent index of glial cell injury yielded the following results. Fifteen minutes after injury the astroglia produced LTC, in quantities ranging from 20 pg/ml (500 pg/106 cells) at a pendulum angle of 20” to 180 pg/ml (4.5 ng/ lo6 cells) at a pendulum angle of 30” (Fig. 8). Time course analysis of LTC, production after injury demonstrated that LTC, levels peaked at 15 min following barotrauma and this time point was selected for study. LTC,
FIG. 5. Chamber pressure and impulse magnitude vs pendulum angle. As is apparent in this graph, pressure within the chamber is logarithmically related to the angle of incidence of the pendulum. Both impulse magnitude recorded in millivolts and chamber pressure recorded in atmospheres are shown.
DISCUSSION
This adaptation of the fluid percussion system represents a powerful method for creating a reliable, highly reproducible injury to cell monolayers. Rigorous testing of this system has demonstrated a close correlation between peak chamber pressures and the pendulum angle from which the 4.8-kg weight is swung to impact the fluid filled piston. Thus, by setting the pendulum angle, a reproducible specified level of injury is obtained. Previous studies of fluid percussion systems have demonstrated that the level of injury produced correlates with the peak pressure exerted by the system rather than with the mean pressure exerted by the system. Two means were employed to demonstrate the biological effects of injury by this system. First, the percentage of cells detached from the cell monolayer as a function of the degree of injury was examined and the resulting percentage of nonviable cells as a function of the degree of injury were examined. Both of these parameters demonstrated direct, almost linear relationships with the degree of injury, demonstrating that increased peak pressures experienced by the cells results in increased levels of injury to the cells. Second, LTC, production as a function of cell injury was examined. As the degree of injury increased, the cells produced greater amounts of LTC,, suggesting a relationship between the degree of injury and the level of LTC, produced. Interestingly, the production of LTC, sharply declines at pendulum angles greater than 30”. It is possible that excessive injury produced at higher angles results in irreversible damage to the metabolic machinery necessary to generate lipoxygenase products from liberated arachidonic acid. LTC, production was employed only as an independent index of cell injury and more in depth analysis will not be presented herein. The exact mechanism by which barotrauma causes cell injury is not yet defined. Plausible considerations
FIG. 6. (A) A light micrograph of glial cell monolayers after barotrauma. (Magnification 400X).
in culture.
(B) The detachment
422
of the monolayer
from the surface of the petri dish
SHEPARD
ET AL.: IN
VITRO
FLUID
PERCUSSION
include mechanical disruption of cell membranes, particularly a direct effect on phospholipid bilayers. Perhaps more likely would be a direct effect on integral membrane proteins, especially those membrane-bound ATPases responsible for ionic homeostasis. Since activity of these proteins is dependent upon allosteric relationships within the phospholipid bilayer, mechanical forces may alter electrostatic relationships and render these molecules temporarily inactive. Another possible injurious effect would be heat generation within the cell chamber as a result of rapid compression of the chamber contents. Although we took every step to insure the chamber temperature was maintained at 37°C during the experiment, we were technically limited in our ability to follow temperatures at the level of the cell monolayers. This simple system for the in vitro study of the effects of mechanical injury on cell culture systems using forces comparable to those employed in our most widely used animal model may enhance our capacity to investigate the cellular effects of mechanical injury on isolated cell populations. By isolating the effects of barotraumatic injury upon individual components of the nervous system, the effects of injury on specific cell types may be investigated. Also, this system allows for the study of certain aspects of barotraumatic injury without the need to sacrifice animals and is more economical than in uivo systems. Furthermore, the effects of various pharmacologic and metabolic interventions may be investigated.
1OC
BAROTRAUMA
423
MODEL 5000
2co
4000
3000
1 j$
8. 100 I
5 s
2000
e r;l
1000
0
I
10
20
-
I
30
.
I
40
.
I
50
0
.
60
Pendulum Angle FIG. 8. Production of leukotrienes C, as measured by enzyme linked amino assay and recorded as a function of the pendulum angle used to induce barotraumatic injury is shown. Leukotriene production recorded in picograms per milliliter as well as picograms per one hundred thousand cells is represented. Leukotriene production is recorded solely as an additional index of cell injury.
8C
REFERENCES 6C
1. 2. 3.
4.
5. 6.
7.
FIG. 7. The effect of harotrauma on glial cell monolayers. Both the percentage of detached cells and the percentage of nonviable cells as a function of atmospheres of barotraumatic injury are represented here.
8.
Govons, S. R., Govons, R. B., Van Huss, W. D., and Heusaer, W. W. Brain concussion in the rat. Exp. Neural. 34: 121,1972. Bergren, D. R., and Beckman, D. L. Pulmonary surface tension and head injury. J. Trauma 15: 336,1975. Ommaya, A. K., Flamm, E. S., and Mahone, R. H. Cerebral concussion in the monkey: An experimental model. Science 153: 211, 1966. Gurdjian, E. S., and Webster, J. F. Experimental head injury with reference to the mechanical factors in acute trauma. Surg. Gynecol. Obstet. 76: 623, 1943. Denny-Brown, D., and Russell, W. R. Experimental cerebral concussion. Brain 64: 93, 1941. Tsubolawa, T., Nakamura, S., Hayashi, N., et al. Experimental primary fatal head injury caused by linear acceleration-biomechanics and pathogenesis. Neural. Med. Chir. 15: 57, 1975. Kobrine, A., and Kempe, L. G. Studies in head injury. Part 1: An experimental model of closed head injury. Surg. Neurol. 1: 34, 1973. Ommaya, A. K. Experimental head injury in the monkey. In W. F. Cavenes and A. E. Walker (Eds.), Head Injury Conference Proceedings, 1966. Pp. 260-275.
424 9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
JOURNAL
OF SURGICAL
RESEARCH
Chason, J. L., Fernand, 0. V., and Hodgson, V. R. Experimental brain concussion morphogenic findings and a new cytologic hypothesis. J Trauma 6: 767, 1966. Huger, F., and Patrick, G. Effect of concussive head injury on central catecholamine levels and synthesis rates in rat brain regions. J. Neurochem. 33: 89, 1979. Crockard, H. A., Brown, F. D., Calica, A. B., Johns, L. M., and Mullan, S. Physiological consequences of experimental cerebral missile injury and use of data analysis to predict survival. J. Neurosurg. 46: 784, 1977. Crockard, H. A., Brown, F. D., Johns, L. M., and Mullan, S. An experimental cerebral missile injury model in primates. J. Neurosurg. 46: 776,1977. Hayes, R. L., and Ellison, M. D. Animal models of concussive head injury. In Becker and Gudeman (Eds.), Head Injury, Philadelphia: W.B. Saunders, 1989. Gennarelli, T. A., and Thibault, L. E. Biological models of head injury. In D. P. Becker and J. T. Povlishock (Eds.), NZH Central Nervous System Trauma Status Report, 1985. Pp. 103-106. Ommaya, A. K., Corrao, P., and Letcher, F. S. Head injury in the chimpanzee. Part 1: Biodynamics of traumatic unconsciousness. J. Neurosurg. 39: 152, 1973. Ommaya, A. K., Grubb, R. L., and Naumann, R. A. Coup and contrecoup injury: Observations on the mechanics of visible brain injuries in the rhesus monkey. J. Neurosurg. 35: 503,197l. Gennarelli, T. A., Adams, J. H., and Graham, D. I. Acceleration induced head injury in the monkey: Neuropathology. Appl. Neurobiol. 6: 234, 1980. Gennarelli, T. A., Adams, J. H., Graham, D. I., and Jellinger, K. Acceleration induced head injury in the monkey: The model its mechanical and physiological correlates. Acta Neuropath. Suppl. VII, 23, 1981. Gennarelli, T. A., Seqawa, H., Wald, V., et al. Physiological response to angular acceleration of the head. In R. G. Grossman and P. L. Gildenbere (Eds.). Seminars in Head Zniurv: - “, Basic and Clinical Aspects. 1982: Pp. 129-140. Gennarelli, T. A., Thibault, L. E., Adams, J. H., et al. Diffuse axonal injury and traumatic coma in the primate. Ann. Neurol. 12: 564, 1982.
VOL.
51, NO. 5, NOVEMBER
1991
21.
Gennarelli, son of linear concussion. SAE. 1971.
22.
Gennarelli, T. A., Thibault, L. E., and Ommaya, A. K. Pathophysiologic responses to rotational and translational acceleration of the head. In 16th Stapp Car Crash Conference Proceedings SAE. 1972. Pp. 296-308.
23.
Gennarelli, T. A., and Thibault, L. E. Experimental production of prolonged traumatic coma in the primate. In R. Villiani et al. (Eds.), Advances in Neurotraumatology. 1982. Pp. 31-33.
24.
Gurdijian, E. S., Hodgson, V. R., Thomas, L. M., et al. Highspeed techniques in head injury research. Demonstration of relative movements of scalp, skull and intracranial contents during impact. Med. Sci. 18: 45, 1967.
25.
Chason, J. L., Hardy, W. G., Webster, J. E., and Gurdjian, E. S. Alterations in cell structure of the brain associated with experimental concussion. J. Neurosurg. 16: 135, 1958.
26.
Lewett, W., Jenkins, L. W., and Miller, J. D. Effects of experimental fluid percussion injury of the brain on cerebrovascular reactivity to hypoxia and to hypercapnia. J. Neurosurg. 66: 332, 1982.
27.
Lindgren, S., and Rinder, L. Experimental studies in head injury. I: Some factors influencing results of model experiments. Biophysik 3: 320,1965.
28.
Rinder, L., and Olsson, Y. Studies on vascular permeability changes in experimental brain concussion. I: Distribution of circulating fluorescent indicators in brain and cervical cord after sudden mechanical loading of the brain. Acta Neuropathol. 11: 183, 1968.
29.
Sullivan, H. G., Martinez, sion model of mechanical 45: 520, 1976.
30.
Hariri, R. J., Ghajar, J. B. G., Pomerantz, K. B., Haiar, D. P., Giannuzzi, R. F., Tomich, E., Andrews, D. W., Patterson, R. H. Human glial cell production of lipoxygenase generated eicosanoids: A potential role in the pathophysiology of vascular changes following traumatic brain injury. J. Trauma 29(g): 1203, 1989.
T. A., Ommaya, A. K., and Thibault, L. E. Compariand rotational acceleration in experimental cerebral In 15th Stapp Car Crash Conference Proceedings Pp. 797-803.
J., Becker, D. P., et al. Fluid-percusbrain injury in the cat. J. Neurosurg.
