Traumatic injury of cortical neurons causes changes in intracellular calcium stores and capacitative calcium influx.

Using an in vitro traumatic injury model, we examined the effects of mechanical (stretch) injury on intracellular Ca2+ store-mediated signaling in cultured cortical neurons using fura-2. We previously found that elevation of [Ca2+](i) by the endoplasmic reticulum Ca2+-ATPase inhibitor, thapsigargin, was abolished 15 min post-injury. In the current studies, pre-injury inhibition of phospholipase C with neomycin sulfate maintained Ca2+-replete stores 15 min post-injury, suggesting that the initial injury-induced store depletion may be due to increased inositol trisphosphate production. Thapsigargin-stimulated elevation of [Ca2+](i) returned with time after injury and was potentiated at 3 h. Stimulation with thapsigargin in Ca2+-free media revealed that the size of the Ca2+ stores was normal at 3 h post-injury. However, Ca2+ influx triggered by depletion of intracellular Ca2+ stores (capacitative Ca2+ influx) was enhanced 3 h after injury. Enhancement was blocked by inhibitors of cytosolic phospholipase A2 and cytochrome P450 epoxygenase. Since intracellular Ca2+ store-mediated signaling plays an important role in neuronal function, the observed changes may contribute to dysfunction produced by traumatic brain injury. Additionally, our results suggest that capacitative Ca2+ influx may be mediated by both conformational coupling and a diffusible messenger synthesized by the combined action of cytosolic PLA2 and P450.

Traumatic brain injury (TBI) 1 is a major cause of death and disability (1). The pathophysiology of TBI is believed to consist of two phases. A primary phase causes cellular damage at the moment of insult by direct mechanical disruption of the brain. A secondary, or delayed phase of damage, consists of biochem-ical and cellular events that are initiated at the time of insult but do not appear clinically for hours or days after injury. Clinically, this is often presented as patients who regain consciousness after head injury and demonstrate the ability to speak and obey commands, only to die hours or days later (2). At present, the biochemical and cellular mechanisms that comprise secondary injury are not entirely understood.
A fundamental pathological mechanism underlying secondary cell injury and death in the central nervous system after TBI is believed to be a disruption of Ca 2ϩ homeostasis (3). This "calcium hypothesis" for TBI is based on the finding that calcium in the cerebrospinal fluid decreases from 1 to 0.01 mM following traumatic injury (4,5). Concurrently, total tissue 45 Ca 2ϩ uptake is increased after in vivo injury (6,7). Therefore, substantial evidence supports the notion that total intracellular calcium is elevated in cells of the central nervous system after trauma. However, there are several pools of intracellular calcium that comprise total cell calcium. These include calcium bound to the plasma membrane, cytoskeletal elements, and calcium-binding proteins; calcium stored in the endoplasmic reticulum (ER), mitochondria, and nucleus; and intracellular free calcium ([Ca 2ϩ ] i ) (8). The relative distribution of calcium throughout the various cellular pools after traumatic injury is unknown.
Maintenance of [Ca 2ϩ ] i is of critical importance to the cell because of the vital role that Ca 2ϩ ions play in regulation of cellular functions, especially in neurons (3). In particular, [Ca 2ϩ ] i plays a critical role in the control of neuronal membrane excitability (8,9), maintenance of cytoskeletal integrity (10,11), and control of synaptic activity and neurotransmitter release (12)(13)(14)(15)(16). [Ca 2ϩ ] i is also an important signal transduction molecule in neurons and is necessary for the activation of calcium-dependent enzymes involved with normal cellular and physiological processes, including several protein kinases and phospholipases (3). Alterations of these important physiological mechanisms in neurons could have profound effects on overall brain function.
Currently, there is a strong need to apply biochemical studies to the pathology of TBI, particularly with respect to [Ca 2ϩ ] i dynamics. Due to technical limitations of in vivo injury models, the mechanisms of altered [Ca 2ϩ ] i signaling after TBI are not well documented. To investigate some of the possible mechanisms underlying traumatic injury, our laboratory has utilized an in vitro model that simulates a major component of TBI, that being tissue strain or stretch (17). To produce strain, cells are grown in tissue culture wells having a Silastic membrane bottom that is stretched by a 50-ms pulse of pressurized gas. We have validated our model by demonstrating that in vitro stretch injury produces many of the post-traumatic responses observed after in vivo TBI (17)(18)(19)(20)(21). In addition, we have previously identified novel neuronal alterations after in vitro injury, including a transient stretch-induced delayed depolarization (22,23), a stretch-induced reduction in the Mg 2ϩ blockade of the NMDA current (24), and reduced 3-hydroxy-5-methyl-4isoxazole receptor desensitization (25).
Using our in vitro model of injury, we have identified disruptions in store-controlled or capacitative calcium signaling in neurons (26). In the capacitative calcium influx pathway, G protein-coupled receptors linked to phospholipase C (PLC) catalyze the hydrolysis of phosphatidylinositol bisphosphate yielding diacylglycerol and inositol 1,4,5-trisphosphate (IP 3 ). IP 3 binds to receptors on IP 3 -sensitive intracellular calcium stores in the ER, resulting in the release of stored calcium (16,27). Depletion of calcium stores causes the production of a calcium influx factor (CIF), which stimulates influx of extracellular calcium through store-operated or second messenger-operated channels (SOCs/SMOCs) in the plasma membrane (capacitative calcium influx), thereby further increasing [Ca 2ϩ ] i (8,(27)(28)(29). Although the majority of data on capacitative calcium influx is derived from nonexcitable cells, there is now strong evidence for the existence of this pathway in neurons (30 -32).
We have previously found that, immediately after injury, the [Ca 2ϩ ] i elevation elicited by the metabotropic glutamate receptor (mGluR) agonist, trans-(1S,3R)-1-amino-1,3-cyclopentanedicarboxylic acid (tACPD) was abolished, suggesting alterations to mGluRs, IP 3 receptors, IP 3 production, or capacitative calcium influx (26). To dissect the various components of the tACPD response, we utilized the pharmacological agent, thapsigargin, which inhibits the sarcoplasmic-endoplasmic reticulum Ca 2ϩ -ATPase (SERCA) of intracellular calcium stores (33). Thapsigargin-stimulated elevation of [Ca 2ϩ ] i was also inhibited 15 min after neuronal injury (26), suggesting that intracellular calcium stores are either empty or can no longer sequester calcium. However, 3 and 24 h after injury, intracellular calcium stores responded to tACPD and thapsigargin with significantly enhanced elevation of [Ca 2ϩ ] i . Therefore, we observe a primary phase in which intracellular calcium stores elicit a depressed response to stimuli after injury, followed by a secondary phase in which store-mediated signaling appears to be potentiated in neurons.
The initial phase, in which IP 3 -mediated signaling is diminished in injured neurons, is similar to other findings in which disruption of the cytoskeleton results in inhibition of agonistinduced [Ca 2ϩ ] i signaling (34). Therefore, the initial phase of decreased [Ca 2ϩ ] i signaling after neuronal injury may correlate with the primary phase of TBI and be due to direct mechanical disruption of the cytoskeleton. The delayed response in which we observe a potentiation in calcium store-mediated signaling may correlate with the secondary phase of TBI due to altered biochemical mechanisms. In the current report, we describe the apparent underlying mechanisms of the observed changes in intracellular calcium store-mediated signaling after mechanical injury. We report, for the first time, a delayed enhancement in capacitative calcium influx in mechanically injured cortical neurons. Despite several recent findings that are in strong support of a physical coupling and the possible involvement of a secretory mechanism in capacitative calcium influx (35)(36)(37)(38), the delayed enhancement in capacitative calcium influx that we observe in injured neurons appears to also involve activation of cytosolic phospholipase A 2 (cPLA 2 ) and a P450 monooxygenase system. These findings are consistent with the production of a diffusible signaling molecule in neurons, which is similar to our previous findings in cortical astrocytes (29). Therefore, alterations in capacitative calcium influx could involve both structural disruption in neurons as well as changes in CIF-mediated signaling. We hypothesize that these alterations in IP 3 -mediated calcium signaling pathways and intra-cellular calcium stores may play a role in the neuronal dysfunction that follows TBI.

EXPERIMENTAL PROCEDURES
Cell Culture-Mixed cultures of neuronal plus glial cells were prepared from 1-to 2-day-old neonatal Harlan Sprague-Dawley rat pups (Hilltop Lab Animals, Scottdale, PA) as described previously (18). In brief, brains were removed aseptically and placed in sterile dissecting saline. Cortices were dissected from the brain and cleaned of white matter and meninges, leaving the neocortex. The neocortex was placed in 5 ml of dissecting saline that contained 0.125% trypsin (Sigma) for 10 min and was transferred to a tube containing 5 ml of Dulbecco's modified Eagle's medium (DMEM, Mediatech, Herndon, VA) supplemented with 20 mM glucose, 10% fetal bovine serum (Summit Biotechnology, Fort Collins, CO), 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine. The tissue was washed, the supernatant removed, and 5 ml of culture medium added for a second wash. After washing, the supernatant was removed, and the tissue fragments were dissociated by trituration with a glass pipette in 5 ml of culture medium. The suspension was centrifuged for 10 min at 200 ϫ g. The supernatant was removed, and 5 ml of culture medium was added. The suspension was triturated, filtered with a 70-m nylon cell strainer, diluted with DMEM, and counted with a hemocytometer. One-ml aliquots of the cell suspension containing 1 ϫ 10 6 cells were seeded into each well of a collagen-coated Flex Plate (Flexcell International, McKeesport, PA). Cells were cultured in a 5% CO 2 incubator at 37°C. One day after plating, the media were removed and replaced with 1 ml of fresh medium. Three days later, 0.5 ml of medium was removed from each well and replaced with 0.5 ml of neuronal growth medium. Neuronal growth medium contained DMEM, 10 mM glucose, 100 units/ml penicillin, 100 g/ml streptomycin, and 5% horse serum (Life Technologies, Inc.). Cells were used for experiments within 10 -15 days after removal from the rat. Glial cells typically adhered to the membrane substrate, and the neurons adhered to the glial cells. Neuronal processes and interconnections between individual neuronal cells and neuronal cell clusters were readily apparent.
Fura-2 Loading-[Ca 2ϩ ] i was measured with the ratiometric dye, fura-2. Mixed cultures of neuronal plus glial cells were loaded with fura-2 AM as described previously (24,26). Briefly, cells were washed once and placed in 1 ml of Dulbecco's phosphate-buffered saline (DPBS) containing 1 mM Ca 2ϩ and Mg 2ϩ and supplemented with 1 mM glucose and 1% fatty acid-free bovine serum albumin. Fura-2 AM (5 M, Molecular Probes, Eugene, OR) was added, and the cells were incubated in the dark at 37°C for 1 h. Under these conditions, the astrocytes in the co-cultures retain fura-2 only weakly in the cytoplasm. In our studies, astrocytes required room temperature conditions for adequate fura-2 loading since the cells appeared to actively metabolize and compartmentalize fura-2, and lose fluorescence when maintained at 37°C (20,26,29). Thus, loading of mixed neuronal plus glial cultures at 37°C resulted in a more selective loading of neurons. The concentration of intracellular fura-2, the presence of unhydrolyzed fura-2 AM derivatives, and dye compartmentalization were routinely monitored (39 -41). We have found that these fura-2 loading techniques produce high concentrations of intracellular fura-2 in neurons, negligible dye compartmentalization, and minimal incomplete AM ester hydrolysis as determined by scanning the excitation spectra of loaded cells. After loading, mixed cultures were washed once with supplemented DPBS and were placed in a final volume of 2 ml of DPBS for measurement of [Ca 2ϩ ] i .
Cell Injury-Fura-2-loaded neuronal plus glial cultures were injured as described previously, using a commercially available model 94A Cell Injury Controller (Bioengineering Facility, Virginia Commonwealth University, Richmond, VA) (17). In brief, the Silastic membrane of the Flex Plate well is rapidly and transiently deformed by a 50-ms pulse of compressed gas, which deforms the Silastic membrane and adherent cells to varying degrees controlled by the pulse pressure. The extent of cell injury, as measured by uptake of propidium iodide and release of lactate dehydrogenase, is dependent on the degree of Silastic membrane deformation (17,18,26). Based on our previous work we have arbitrarily defined three levels of cell injury, mild, moderate, and severe, produced by deforming the silastic membrane on which the cells are grown by 5.5, 6.5, and 7.5 mm, respectively (17)(18)(19)(20). These degrees of membrane deformation result in a biaxial strain or "stretch" of 31, 38, and 54%, respectively (17). This range of cell stretch has been shown to be relevant to what would occur in humans after rotational acceleration/deceleration injury (42). The current experiments were all performed at a mild level of injury (5.5 mm deformation, 31% stretch). Correlation of degree of strain produced in models to strain-induced deficits in humans has shown that greater than 30% stretch produces persistent behavioral deficits. It is important to note that with the cell injury model, as currently used, there is no hypoxia superimposed on the injured cells.
Microspectrophotometry-[Ca 2ϩ ] i was measured at room temperature using a Ratiomaster (Photon Technologies International, South Brunswick, NJ) microspectrophotometry system. Excitation light was provided by a xenon arc lamp coupled to a RAM scanner that alternated excitation light between 350 and 380 nm. Band pass was set at Ϯ1 nm. Excitation light was delivered to the cells via fiber optics through the epifluorescence port of a Zeiss Standard 16 microscope coupled to a Zeiss Achroplan 40ϫ water immersion lens. Use of the water immersion lens allowed in situ measurements and easy manipulation of the Flex Plates during injury. However, the optical properties of the lens necessitated the ratioing of fura-2 at 350 and 380 nm. Emission was measured at 510 nm via a microphotometer. The entire system for data collection and analysis was computer-driven.
Flex plates containing neuronal plus glial cultures were placed on the microscope stage, and single neurons were selected for [Ca 2ϩ ] i measurement using slit width adjustments on the microphotometer. Basal [Ca 2ϩ ] i was recorded for several minutes at a sampling rate of one ratio every 0.2 s. In experiments utilizing neomycin sulfate (Sigma), an inhibitor of PLC, cells were pretreated with neomycin for 3 h prior to injury. Thapsigargin (Calbiochem) and tACPD (Research Biochemicals International, Natrick, MA) were added directly to culture wells at the indicated time after injury. In the indicated experiments, the SOC/ SMOC antagonist SKF 96365 (Calbiochem) was added to cultures for 10 min prior to application of thapsigargin. The cytochrome P450 inhibitor, econazole (Calbiochem), and the cPLA 2 inhibitor, arachidonyltrifluoromethyl ketone (AACOCF 3 , Calbiochem), were added to cultures for 2 min prior to thapsigargin stimulation. An additional inhibitor of cytochrome P450, N-methylsulfonyl-6-(2-propargyloxyphenyl) hexanimide (MS-PPOH, synthesized by Dr. J. R. Falck, University of Texas Southwestern Medical Center, Dallas) (29) was also used and was added to cultures 2 min prior to thapsigargin stimulation. MS-PPOH is a "suicide substrate" inhibitor of P450 arachidonic acid epoxygenase, designed to resemble the substrate arachidonic acid and inactivate the enzyme (43). All agents were added directly to culture wells in a volume of 200 l at the indicated time after injury. This volume ensured almost immediate mixing with the buffer, as determined by dye diffusion.
In experiments in which [Ca 2ϩ ] i was measured at 3 h post-injury, the cells were injured in DMEM without serum or phenol red. After injury, serum was added back to the media, and cells were incubated for the 3-h interval. Uninjured controls were incubated under the same conditions. Cells were loaded with fura-2 AM 1 h prior to the end of the incubation period. The concentration of intracellular fura-2 was routinely monitored (39) at 3 h post-injury and did not significantly differ from controls.
In Ca 2ϩ -free experiments, supplemented DPBS was replaced with Ca 2ϩ -free DPBS. Nominally Ca 2ϩ -free conditions caused a small but nonsignificant drop in basal [Ca 2ϩ ] i as previously reported by others (30,44,45). After basal [Ca 2ϩ ] i had stabilized, thapsigargin was applied to culture wells. To preserve the original status of intracellular calcium stores, neurons were maintained in nominally Ca 2ϩ -free conditions for a maximum of 2 min before stimulation with thapsigargin.
[Ca 2ϩ ] i was calculated as described previously (46) using a correction for intracellular viscosity (39). Autofluorescence at 350 and 380 nm was recorded from neurons of neuronal plus glial cultures that were not loaded with fura-2 and was subtracted from all measurements. Leakage of fura-2 into the medium after injury was monitored by measuring intracellular fura-2 concentration at the isosbestic wavelength, 362 nm (46). Additional measurements were made to assess fura-2 fluorescence directly in the medium. Consistent with the lack of propidium iodide uptake that we have previously reported immediately after injury in neurons (26), no significant leakage of fura-2 was observed following injury.
Statistical Analysis-Data are mean Ϯ S.E. values. Statistical significance was established by analysis of variance followed by Fisher's protected least significant difference test.

RESULTS
We have previously reported that tACPD and thapsigarginstimulated elevation of [Ca 2ϩ ] i is almost completely abolished 15 min after injury, suggesting that intracellular calcium stores are depleted (26). However, intracellular calcium stores elicit a delayed enhancement in response to stimuli at 3 h post-injury. Therefore, we observe two phases of altered cal-cium store-mediated signaling after injury, an early or immediate phase in which the stores do not respond to stimuli and a late or delayed phase in which store-mediated signaling is potentiated. The current experiments were designed to decipher the mechanisms of the two different phases of calcium store-mediated signaling that are observed after traumatic injury of cortical neurons.
Immediately after Injury IP 3 -sensitive Calcium Stores Are Depleted-One possible explanation for the diminished response to tACPD and thapsigargin that we observe initially after injury is increased activation of PLC, resulting in elevated IP 3 . High levels of IP 3 may stimulate IP 3 receptors on intracellular calcium stores, possibly maintaining them in an empty state (47). Therefore, we used the PLC inhibitor, neomycin sulfate (48,49), to evaluate the effect of PLC activation on store-mediated calcium signaling in stretch-injured neurons.
Pretreatment of uninjured neurons with 100 M neomycin sulfate completely blocked the elevation of [Ca 2ϩ ] i in response to 100 M tACPD, suggesting that neomycin inhibits PLC activation in cortical neurons (data not shown). post-injury. However, when neurons are pretreated with 100 M neomycin sulfate 3 h prior to injury, the thapsigarginstimulated elevation of [Ca 2ϩ ] i was 50 Ϯ 8 nM, suggesting that blockade of PLC maintains calcium-replete intracellular calcium stores after injury. These results suggest that activation of PLC may cause intracellular Ca 2ϩ stores to remain in an empty state after stretch injury.
Thapsigargin-stimulated Elevation of [Ca 2ϩ ] i Is Restored with Time after Injury-We have previously found that the [Ca 2ϩ ] i elevation elicited by tACPD or thapsigargin is potentiated at 3 h post-injury (26) suggesting that intracellular calcium stores undergo a delayed enhancement in their response to stimuli after injury. However, the status of intracellular calcium stores between 15 min and 3 h post-injury is unknown. Therefore, we investigated the time course of thapsigarginstimulated elevation of [Ca 2ϩ ] i after injury (Fig. 2). Application of 1 M thapsigargin increased [Ca 2ϩ ] i 43 Ϯ 6 nM above basal in uninjured neurons. Fifteen minutes after injury thapsigarginstimulated elevation of [Ca 2ϩ ] i is almost completely abolished, being only 8 Ϯ 4 nM above basal. The elevation of [Ca 2ϩ ] i elicited by 1 M thapsigargin increased with time after injury, being 34 Ϯ 18 nM 1 h after injury, 61 Ϯ 20 nM 2 h after injury, and 87 Ϯ 19 nM 3 h after injury, at which time the response was potentiated as compared with uninjured neurons. Therefore, after injury we observe a primary phase in which calcium signaling mediated through intracellular stores is depressed, followed by a secondary phase in which store-mediated signaling appears to be potentiated.
The Size of Intracellular Calcium Stores Is Unaltered after Injury-There are several possible explanations for the mechanism underlying enhanced thapsigargin-stimulated elevation of [Ca 2ϩ ] i 3 h after injury. One possibility is that calcium is released from the stores more rapidly in injured neurons. However, we have previously found that the release of calcium from intracellular calcium stores is actually slower in injured neurons when compared with uninjured cells (26). Since we have also found that [Ca 2ϩ ] i is elevated immediately after mild injury of neurons but returns to basal levels by 3 h post-injury (26) intracellular calcium stores may play a role in restoring basal [Ca 2ϩ ] i by sequestering additional calcium after injury. An increased uptake of calcium after injury could result in an enlargement of the intracellular calcium store pool and may explain the potentiated store-mediated signaling 3 h post-injury.
To investigate the possibility that the size of intracellular calcium stores had increased at 3 h after injury, neurons were stimulated with 1 M thapsigargin in Ca 2ϩ -free DPBS (Fig. 3). Basal [Ca 2ϩ ] i in uninjured neurons placed in nominally Ca 2ϩfree conditions was 81 Ϯ 8 nM. Stimulation of uninjured neurons with 1 M thapsigargin in Ca 2ϩ -free DPBS resulted in an increase in [Ca 2ϩ ] i of 24 Ϯ 6 nM, suggesting that approximately half of the thapsigargin-stimulated elevation of [Ca 2ϩ ] i is due to release of calcium from intracellular stores. The thapsigargin-stimulated elevation of [Ca 2ϩ ] i in Ca 2ϩ -free DBPS was 32 Ϯ 17 nM at 3 h post-injury, similar to uninjured neurons. Thus, the size of the intracellular calcium stores is unchanged 3 h after injury. This implies that the increased [Ca 2ϩ ] i response at 3 h post-injury is dependent on influx of extracellular calcium.
Injury Causes a Delayed Enhancement in Capacitative Calcium Influx-Since the size of the intracellular calcium stores was apparently unaltered 3 h after injury, we tested whether the enhancement in thapsigargin-stimulated elevation of [Ca 2ϩ ] i was due to an increase in capacitative calcium influx. To do this, we utilized the SOC/SMOC antagonist, SKF 96365 (50). In Fig. 4 increase in capacitative calcium influx and not to a change in the size of the intracellular calcium store pool. Application of 1 M SKF 96365 had no effect on the [Ca 2ϩ ] i elevation elicited by 50 mM KCl, indicating that SKF 96365 had no effect on voltagegated calcium channels at the concentration used in these experiments.
Capacitative Calcium Influx May Be Enhanced Due to Alterations in the Production of Calcium Influx Factor-The effects of cPLA 2 on capacitative calcium influx in neurons was assessed with AACOCF 3 , an inhibitor of cPLA 2 . As shown in Fig.  5, 5 M AACOCF 3 had no effect on basal [Ca 2ϩ ] i . Application of 5 M AACOCF 3 did not alter the thapsigargin-stimulated elevation of [Ca 2ϩ ] i in calcium-free medium, suggesting that AA-COCF 3 had no effect on intracellular calcium stores (data not shown). AACOCF 3 also had no effect on the [Ca 2ϩ ] i elevation induced by 50 mM KCl, suggesting that voltage-gated calcium signaling was unaffected (data not shown). However, consistent with our published results in astrocytes (29), AACOCF 3 significantly decreased thapsigargin-stimulated elevation of [Ca 2ϩ ] i in uninjured neurons (Fig. 5, light gray bars), to a level consistent with blockade of capacitative calcium influx. In injured neurons, AACOCF 3 effectively blocked the enhancement in thapsigargin-stimulated calcium influx observed at 3 h postinjury (Fig. 5, dark bars). These results suggest that activation of cPLA 2 may be involved in the delayed enhancement of capacitative calcium influx in stretch-injured neurons.
To dissect the effects of cytochrome P450 on the enhancement of capacitative calcium influx in injured neurons, we utilized two P450 inhibitors, MS-PPOH and econazole. The cytochrome P450 enzymes comprise a large family of isoforms (51) of which econazole is a generalized inhibitor. MS-PPOH is a specific suicide substrate inhibitor of P450 AA epoxygenase, designed to resemble the substrate AA and inactivate the enzyme (43). Neither 1 M MS-PPOH nor 10 M econazole had any effect on the thapsigargin-stimulated elevation of [Ca 2ϩ ] i in calcium-free medium, suggesting that the intracellular calcium stores were unaffected by these agents. Additionally, MS-PPOH and econazole did not alter the [Ca 2ϩ ] i elevation induced by 50 mM KCl, suggesting that voltage-gated calcium channels were not affected at the concentrations utilized in these experiments (data not shown).
As shown in Fig. 5, MS-PPOH significantly decreased the thapsigargin-stimulated [Ca 2ϩ ] i elevation in uninjured neurons, consistent with blockade of capacitative calcium influx and similar to our previous observations in astrocytes (29). Econazole did not significantly attenuate thapsigargin-stimulated elevation of [Ca 2ϩ ] i in uninjured cells; however, it did block the enhanced thapsigargin-stimulated elevation of [Ca 2ϩ ] i at 3 h post-injury. MS-PPOH attenuated thapsigarginstimulated elevation of [Ca 2ϩ ] i at 3 h post-injury in two of five neurons examined but was ineffective in blocking the potentiated response in three neurons, hence the large error bar in Fig. 5. Taken together, the results suggest the involvement of cytochrome P450 in the delayed enhancement of capacitative calcium influx observed in injured neurons. DISCUSSION A fundamental pathological mechanism underlying secondary cell injury in the central nervous system following TBI is proposed to be a disruption of Ca 2ϩ homeostasis (3). Although total intracellular calcium is elevated in the brain after traumatic injury, the relative distribution of calcium throughout the various cellular pools is unknown. For example, much of the calcium influx after injury may be buffered by the plasma membrane, cytoskeletal elements, and by cytoplasmic proteins such as calmodulin, calbindin, and parvalbumin (52). These processes may protect against dramatic changes in [Ca 2ϩ ] i levels. Concurrently, additional uptake of calcium by the mitochondria could be detrimental, based on reports that calcium overload may disrupt energy metabolism (53)(54)(55). Also, greater uptake of calcium by the nucleus could adversely affect genetic mechanisms by activating calcium-dependent endonucleases (56,57). In addition, [Ca 2ϩ ] i plays a critical role in regulating several important neuronal functions (8 -16). Therefore, alterations in [Ca 2ϩ ] i homeostasis could trigger neuronal degeneration and cell death which could ultimately result in clinical disability. We have previously described alterations in IP 3mediated signaling and intracellular calcium stores in cortical neurons after stretch-induced traumatic injury (26). The current report seeks to elucidate some of the underlying biochemical mechanisms of the injury-induced changes in [Ca 2ϩ ] i -mediated signaling in neurons.
We previously found that NMDA-stimulated elevation of [Ca 2ϩ ] i was potentiated 15 min after injury but was normal by 3 h post-injury in neurons of neuronal plus glial cultures (26). Glutamate-stimulated elevation of [Ca 2ϩ ] i was also enhanced 15 min after injury and remained enhanced at 3 h post-injury. The discrepancy in the responses to NMDA and glutamate 3 h after injury suggested a possible alteration in mGluRs coupled to intracellular calcium stores. We further found that elevation of [Ca 2ϩ ] i elicited by the mGluR agonist, tACPD, or by thapsigargin was nearly abolished 15 min after injury (26). This observation may be due to the elevated [Ca 2ϩ ] i at 15 min after injury, which could desensitize the IP 3 receptor (58). However, thapsigargin acts independently of the IP 3 receptor, suggesting that the stores are either empty or can no longer sequester calcium.
Using the PLC inhibitor, neomycin sulfate (48, 49), we were able to investigate the effect of PLC blockade on the status of intracellular calcium stores after injury (see Fig. 6). Neomycin pretreatment restored thapsigargin-stimulated elevation of [Ca 2ϩ ] i in injured neurons. In neomycin-pretreated cells, the [Ca 2ϩ ] i elevation elicited by thapsigargin was restored despite a significant injury-induced elevation of [Ca 2ϩ ] i at 15 min postinjury. These results show that pre-injury blockade of PLC maintains calcium-replete intracellular stores after injury and suggest that G protein-coupled receptors linked to PLC (i.e. Group I mGluRs) may be aberrantly activated after mechanical injury. Activation of PLC and subsequent generation of IP 3 appear to be responsible for the decreased calcium store-mediated signaling 15 min after injury. These findings are consistent with reports following in vivo TBI in which IP 3 levels are elevated (59,60) and with our previous findings in cortical astrocytes demonstrating that inositol polyphosphates are elevated 10-fold after stretch injury (61).
Calcium stores appear to refill with time after injury as evidenced by the ability of thapsigargin to elicit an increase in [Ca 2ϩ ] i at 1 and 2 h after injury. Three hours after injury, we observed an enhancement of thapsigargin-stimulated elevation of [Ca 2ϩ ] i , consistent with our previous findings (26). We have previously found that the enhancement in thapsigargin-stimulated elevation of [Ca 2ϩ ] i is not due to a more rapid release of Ca 2ϩ from intracellular stores (26). An alternative mechanism for the enhanced response to thapsigargin is that calcium stores sequester additional calcium following injury. Further uptake of calcium could result in an increase in the size of the intracellular calcium store pool and may explain the enhanced thapsigargin-stimulated elevation of [Ca 2ϩ ] i observed 3 h postinjury. However, neurons that were stimulated with thapsigargin under Ca 2ϩ -free conditions 3 h post-injury elicited an elevation of [Ca 2ϩ ] i that was similar to uninjured controls. These findings suggest that the size of calcium stores does not increase after injury and implies that extracellular Ca 2ϩ is necessary for the enhanced thapsigargin-stimulated elevation of [Ca 2ϩ ] i observed 3 h post-injury.
Since an increase in the size of intracellular calcium stores did not seem to be the cause of the delayed enhancement in calcium store-mediated signaling, this suggests that capacitative calcium influx may be enhanced 3 h after injury. We investigated the possibility that capacitative calcium influx was enhanced by stimulating neurons with thapsigargin in the presence of the SOC/SMOC antagonist, SKF 96365, at 3 h post-injury. Application of SKF 96365 blocked the enhanced thapsigargin-stimulated elevation of [Ca 2ϩ ] i at 3 h post-injury. Taken together with the results observed under Ca 2ϩ -free conditions, our findings strongly suggest that an increase in capacitative calcium influx is responsible for the potentiated thapsigargin-stimulated elevation of [Ca 2ϩ ] i observed 3 h after injury. These results imply that the biochemical mechanisms that link the status of calcium stores with influx of calcium through the plasma membrane can be disrupted by mechanical trauma.
The exact mechanism for the activation of capacitative calcium influx is still a topic of considerable debate. The "conformational coupling" hypothesis as first proposed by Irvine (62) suggests a direct physical connection between the IP 3 receptor and plasma membrane channels that control capacitative calcium influx. When the IP 3 receptor is activated, it is proposed that an alteration in the physical coupling causes influx of calcium into the cell via plasma membrane channels. The "diffusible messenger" hypothesis suggests that when IP 3 -sensitive stores are activated and depleted of calcium, a diffusible messenger (CIF) is released from the stores that binds to SOCs/ SMOCs on the plasma membrane and activates capacitative calcium influx (see Fig. 6). There have been several recent findings that are in strong support of a physical coupling and/or FIG. 6. Status of intracellular calcium stores and capacitative calcium influx in neurons after injury. To investigate the status of intracellular calcium stores after injury, we utilized the SERCA inhibitor, thapsigargin. Fifteen min after injury thapsigargin-stimulated elevation of [Ca 2ϩ ] i is abolished suggesting intracellular calcium stores are depleted. However, pretreatment of neurons with the PLC inhibitor, neomycin sulfate, maintained calcium-replete stores after injury, suggesting elevated levels of IP 3 may be responsible for empty calcium stores. Three hours after injury, thapsigargin-stimulated elevation of [Ca 2ϩ ] i is potentiated. This enhanced response is not evident in Ca 2ϩfree media and is blocked by the SOC/SMOC antagonist, SKF 96365, suggesting capacitative calcium influx is enhanced. Enhanced thapsigargin-stimulated elevation of [Ca 2ϩ ] i is also blocked by econazole and MS-PPOH, inhibitors of P450 metabolism of AA, and by the cPLA 2 inhibitor, AACOCF 3 , suggesting alterations in the production of CIF or changes in the sensitivity of SOCs/SMOCs. A disrupted physical coupling between IP 3 receptors and SOCs/SMOCs may also contribute to the observed effects. In addition, P450 and cPLA 2 may be part of a physical connection. PM, plasma membrane; IP 3 R, IP 3 receptor. a secretory mechanism in capacitative calcium influx (35)(36)(37)(38). However, our previous work also suggests that CIF, possibly an epoxide metabolite of AA, may be a component of conformational coupling and/or secretory models.
We have previously reported that in astrocytes, cPLA 2 and cytochrome P450 are coupled to activation of capacitative calcium influx through the epoxide metabolite of AA, 5,6-EET (29). Production of CIF has not been evaluated in cortical neurons. To investigate the possibility that the enhancement in thapsigargin-stimulated elevation of [Ca 2ϩ ] i at 3 h post-injury is due to an alteration in CIF production, we used inhibitors of two critical steps in 5,6-EET synthesis, release of AA from phospholipid stores by cPLA 2 and cytochrome P450 metabolism of AA. The enhancement in thapsigargin-stimulated elevation of [Ca 2ϩ ] i at 3 h post-injury was blocked by AACOCF 3 and econazole, suggesting that both cPLA 2 activation and cytochrome P450 are involved in this enhanced response. The fact that MS-PPOH was not able to more effectively block the enhanced elevation of [Ca 2ϩ ] i elicited by thapsigargin 3 h after injury in all neurons is difficult to interpret, especially since this compound is reported to be a potent and specific inhibitor of the AA epoxygenase that produces 5,6-EET (29,43). However, there are several isoforms of cytochrome P450 enzymes that differ among cell types (51); therefore, it is possible that MS-PPOH is not a specific inhibitor of CIF production in all cortical neurons. MS-PPOH did significantly decrease thapsigargin-stimulated elevation of [Ca 2ϩ ] i in uninjured neurons to a level that was consistent with thapsigargin stimulation under Ca 2ϩ -free conditions, suggesting that it does inhibit capacitative calcium influx in uninjured neurons. It is possible that injury induces expression of additional P450 enzymes that are differentially selective to pharmacological inhibition. Nonetheless, the overall results suggest that cPLA 2 and cytochrome P450 may be components of CIF in cortical neurons and that these components are involved in the enhanced thapsigarginstimulated elevation of [Ca 2ϩ ] i observed 3 h after injury (Fig.  6).
The injury we are administering to neurons in the current study is mechanical stretch; therefore it is likely that there is disruption of cytoskeletal elements after injury. Previous examination of light microscopy and transmission electron micrographs of neurons in neuronal plus glial cultures has revealed only minor ultrastructural changes in neurons after mild stretch injury. 2 However, by using rhodamine-phalloidin staining for actin, we have recently observed substantial disorganization of actin filaments in injured astrocytes and neurons. 2 It is possible that even minor disruptions of cytoskeletal elements may contribute to some of the effects we report. For example, Pedrosa Ribeiro et al. (34) have shown that cytoskeletal disruption inhibits agonist-elicited mobilization of calcium from IP 3sensitive stores but does not interfere with thapsigargin-stimulated calcium influx. Although our data suggest that elevated levels of IP 3 are most likely responsible for the attenuated calcium-mediated signaling in the immediate phase after injury, a disrupted cytoskeleton caused by the mechanical insult may also contribute to this observation. Also, it is possible that capacitative calcium influx involves both CIF and a physical interaction between the IP 3 -sensitive stores and plasma membrane channels. A strong physical connection with membrane channels as well as CIF may be necessary to activate capacitative calcium influx. For example, cPLA 2 may alter the membrane in the vicinity of SOCs/SMOCs. Likewise, the presence of 5,6-EET, with its epoxide group, may also alter membrane characteristics and impact the tertiary conformation of mem-brane proteins. Therefore, the enhancement in store-mediated signaling 3 h post-injury may be due to alterations in CIF production, the sensitivity of SOCs/SMOCs to CIF, a damaged physical connection of the stores with the plasma membrane, or to a combination of these factors. Since structural alterations are likely to occur after mechanical injury, a more thorough examination of cytoskeletal abnormalities associated with stretch injury in neurons may shed further light on this issue.
A disruption in [Ca 2ϩ ] i homeostasis mediated by intracellular calcium stores could trigger cellular degeneration and cell death, resulting ultimately in behavioral dysfunction. For example, a depletion of ER calcium stores is reported to suppress protein synthesis (63,64), which may contribute to neuronal injury and death. Although we have not directly evaluated cell death in the current studies, we have found that uptake of propidium iodide is minimal immediately after stretch injury of neurons (26) but is increased with time post-injury, suggesting delayed neuronal injury. Previous reports have shown that blocking the release of calcium from intracellular stores with dantrolene protects against NMDA and glutamate-induced neurotoxicity even when [Ca 2ϩ ] i levels are elevated (65,66), suggesting a role for calcium stores in neuronal death. It is also possible that the potentiation of intracellular store signaling observed at 3 h post-injury may be an eventual signal for delayed neuronal death. Although the present study demonstrates enhancement of thapsigargin-stimulated capacitative calcium signaling at 3 h post-injury, we have previously demonstrated that glutamate and tACPD signaling are also enhanced, suggesting that IP 3 -mediated signaling as a whole is potentiated. Initial depletion of intracellular calcium stores may trigger an apoptotic pathway, and enhanced IP 3 -mediated signaling 3 h post-injury could lead to the eventual increase in the synthesis of apoptotic enzymes such as proteases and endonucleases. Recent evidence from non-neuronal cells suggests the involvement of the IP 3 pathway in delayed cell death in that cells undergoing apoptosis have an increased amount of IP 3 receptors (67) and cells that are deficient in IP 3 receptors are resistant to apoptosis (68).
Our findings suggest that IP 3 -mediated signaling, intracellular calcium stores, and SOC/SMOC may be potential targets for therapeutic intervention in TBI. A recent clinical study has indicated that treatment of patients suffering from ischemic stroke within 48 h with conventional heparin, an anticoagulant that cannot cross cell membranes, had no beneficial effects on outcome at 6 months (69). However, treatment within 48 h with low molecular weight heparin, which can cross the cell membrane and inhibit the release of stored calcium, did improve outcome at 6 months. These findings were attributed to the intracellular effects of low molecular weight heparin, most likely antagonism of IP 3 receptors. IP 3 receptor antagonists may offer a considerable advantage over drugs such as mGluR antagonists for the treatment of TBI because several neurotransmitters signal through IP 3 -linked pathways, including glutamate and acetylcholine. Rather than using a mixture of pharmaceutical agents to block several types of mGluRs and acetylcholine receptors, IP 3 receptor antagonists would elicit their action downstream of receptor activation, regardless of the type of receptors being activated. Therefore, IP 3 receptor antagonists may be more efficient than a mixture designed to block IP 3 -linked signaling at the receptor level. In addition, carboxyamidotriazole, an agent which blocks calcium influx primarily by action at SOCs/SMOCs, has shown beneficial effects in clinical trials for the treatment of refractory cancers (70). Therefore, agents that can block the release of calcium from intracellular stores or capacitative calcium influx have shown promise in certain disease conditions. To our knowledge, this is the first report on the mechanisms of traumatically induced changes in calcium store-mediated signaling in neurons. Although in vitro results do not always correlate with the in vivo situation, we have shown that many of the post-traumatic occurrences of in vivo TBI also occur in our model of in vitro stretch injury, suggesting that similar pathologies do occur in vivo and in vitro. Our results suggest that blockade of SOCs/SMOCs with agents such as SKF 96365 may have potential benefits in restoring normal Ca 2ϩ -mediated signaling processes in neurons after a traumatic insult. Also, inhibition of cytochrome P450 with drugs similar to the antifungal agent, econazole, may aid in maintaining Ca 2ϩ homeostasis following trauma. Although further experimentation in in vitro and in vivo models of trauma is necessary, capacitative calcium influx and IP 3 receptors may serve as useful targets for pharmaceutical intervention in the future treatment of TBI. Finally, the present work suggests that both the conformational coupling and the diffusible messenger models for capacitative calcium influx may be intertwined in the mechanism underlying this important signal transduction pathway.