Trauma-Induced Neurotoxicity in Rat Hippocampal Neurons
Background and Purpose We have previously shown that traumatic injury of hippocampal cells triggers release of a soluble neurotoxin that can be transferred to an uninjured culture. The mechanism of this trauma-induced neurotoxicity is independent of glutamate receptor activation. We extended this observation to study the mechanism of this neurotoxicity.
Methods Dissociated rat hippocampal neurons were traumatized by disrupting the culture by scratching the plate. The toxicity expressed by the injured culture was studied by transferring the medium to an uninjured culture and assessing the death rate by trypan blue exclusion.
Results This neurotoxin is stable in the medium at room temperature for several hours and withstands boiling. The molecular weight is between 100 and 500. The release and the effect of this toxin seem to be independent of glutamate receptor activation. The toxicity is unaffected by removal of extracellular calcium. However, dantrolene dose-dependently blocked the toxicity in the recipient culture, suggesting that the release of intracellular stores of calcium is involved in the toxic effect. This release of calcium is likely to be followed by an activation of nitric oxide synthase because competitive nitric oxide synthase inhibitors attenuated this toxicity. Consistent with this result, cholecystokinin octapeptide significantly reduced cell death when combined with this toxic medium.
Conclusions Traumatic injury of dissociated cells can propagate neurotoxicity in uninjured cells by a soluble toxin released into the extracellular space. This toxin causes a rise in cytosolic calcium that activates nitric oxide synthase that can be blocked by cholecystokinin.
There has been an intense effort to elucidate the mechanism of brain injury. The importance of this endeavor is obvious. Despite the prevalence of brain pathophysiology that encompasses acute and acquired as well as chronic and congenital processes, the therapeutic options available to counteract neuronal damage are still limited. From the clinical standpoint, elucidation of these processes that may lead to therapeutic manipulation is an urgent agenda in neurobiology.
Many neuropathological processes that lead to neuronal degeneration seem to converge at several steps that lead to eventual cellular demise. Many of these processes may not be universally recognized as prerequisites for neuronal degeneration. However, the mere existence of common pathways suggests that many pathophysiological processes share the same mechanism of cell death. At the same time, treatment directed against some of these processes could lead to effective therapies for multiple nervous system disorders.
Several of these common processes are well recognized. Excitotoxicity mediated by certain amino acid neurotransmitters has been recognized as a mediator of cell death.1 2 An important consequence of excitotoxicity is a rise in intracellular calcium concentration,2 another critical step, followed by an activation of NOS.3 4 Generation of nitric oxide has been demonstrated to be toxic to neurons by generation of free radicals.5
The aforementioned key events in cell death (ie, excitotoxicity, intracellular calcium, NOS, nitric oxide, and free radicals) are likely to be the points where multiple processes converge. At the same time, it is also likely that multiple processes diverge at each of these steps. Therefore, it is important to consider the fact that multiple and divergent processes could lead to (for example) excitotoxicity and/or free radical generation.
We recently reported that in vitro traumatic cell death in dissociated hippocampal cells6 is mediated, at least in part, by a substance that is released in the EC solution and that this toxicity can be transferred to an uninjured cell culture.7 This “neurotoxin” is released transiently after the traumatic insult because its effects are greatest when the medium is transferred within 5 minutes and decreases in a time-dependent fashion. This toxicity, induced by trauma and transferred to an uninjured culture, was independent of the NMDA receptor mechanism of neurotoxicity. Our findings suggested an alternative mechanism of neurotoxicity induced by traumatized hippocampal cell culture.
In this report we extend our previous findings to show that trauma-induced neurotoxicity is mediated by a heat-resistant toxin that can be blocked at various steps of the known cellular injury mechanisms.
Materials and Methods
The method of cell culture was identical to the procedure reported by Yamada et al8 and Sturm et al.9 Briefly, we used dissociated 1-day-postnatal rat hippocampal cell cultures kept 1 to 2 weeks in vitro for all experiments. Hippocampi from Sprague-Dawley rats were removed, minced, and triturated with fire-polished Pasteur pipettes after incubation in papain solution. The dissociated cells were centrifuged through Eagle’s minimum essential medium/NuSerum (1 mL/10 mL minimum essential medium) with bovine serum albumin and trypsin inhibitor. The cells were resuspended in minimum essential medium/NuSerum and plated at a cell density of 1 million cells per dish (35 mm) onto established cortical glial cultures. Some cells were plated onto 12-well plates at a density of 300 000 cells per well. Glial cells were harvested from cortices of 2- to 6-day-old rat pups. Passaged glial cells were plated onto poly-l-lysine-coated Petri dishes 4 days before neuronal plating. All experiments were performed at room temperature.
All cultures were washed with EC solution containing (mmol/L) NaCl 140, KCl 3, Na-HEPES 10, CaCl2 1, MgCl2 2, and pH adjusted to 7.35. The EC fluids containing the desired compounds were freshly mixed on the day of the experiment. CCK and NS-CCK were dissolved in ethanol at 1 mmol/L before dilution with EC solution. The final concentration of CCK and NS-CCK used was 100 nmol/L (104 dilution factor). All chemicals were obtained from Sigma Chemical Co except CNQX, MK801, L-NAME, L-NMMA, and D-NMMA. CCK and NS-CCK were obtained from Research Biochemicals International.
The method of inducing cell injury was adapted from the published procedure of Tecoma et al.6 A 20-gauge needle was drawn across the Petri dish to produce a tear in the neuronal and glial layer in the 35-mm dish. A total of four tears were made in each culture dish. The EC fluid in the culture (total volume, 2 mL) was collected 5 minutes after the trauma and applied to the uninjured culture wells (volume per well, 1 mL). The cultures were placed back into the incubator overnight before the cell death assessment.
In some experiments, the fluid collected from the traumatized culture at 5 minutes was placed in a dialysis bag (100 or 500 D, Spectra/Por, Spectrum, Fischer). The fluid was equilibrated in plain EC fluid for 4 hours before the application to the naive culture.
The neuronal cell cultures were stained with 0.4% trypan blue in EC solution. Five minutes after they were stained with the dye solution, the cells were washed with EC solution. Dead neurons took up the dye and stained purple, while viable neurons were not stained. Both viable and nonviable cells were then counted with the use of a phase-contrast microscope within a grid.
In most instances there were at least two wells of control, with each batch of experiment in one or two 12-well plates. The control wells received the EC solution collected from an untraumatized dish. This resulted in the same number of fluid exchanges. In some experiments, two additional wells for control toxicity (5 minutes after injury) were also incorporated with each batch of experiments. In these wells, the traumatized fluid collected at 5 minutes was applied without any treatment. Those batches that had a high rate of control death (>0.40) and/or a low rate of control trauma-induced neurotoxicity (<0.55) were discarded.
As another control, plain EC solution was exchanged for the same number of times in six dishes that had an average death rate of 0.31±0.01 (Fig 1A⇓, “wash”). The average control death rate in all experiments was 0.31±0.01 (n=21, P>.5 compared with wash), and the traumatized fluid caused cell death of 0.63±0.01 (n=36).
The results are expressed as a ratio of dead cells to the total number of cells counted. The numerical data are expressed as mean±SE. The significance was tested by two-tailed t test, F test, and the Tukey test when appropriate. Results found to have a value of P<.05 were considered significant.
In an attempt to characterize the toxin, the traumatized EC fluid collected after 5 minutes was kept in the incubator for 2 hours before it was applied to the recipient, the uninjured wells (Fig 1A⇑). This delay had no significant effect on the death rate (0.69±0.05%, n=6, P>.5, compared with control injury). In six additional dishes, the traumatized fluid was boiled briefly (2 minutes) and applied to the recipient culture after it returned to room temperature. The treatment of the solution increased the toxicity to 0.82±0.02 (n=6, P<.001, compared with control death from the traumatized fluid collected after 5 minutes).
To develop an estimate of the size of the toxic mediator, dialysis experiments were performed (Fig 1A⇑). In these experiments the traumatized fluid was allowed to equilibrate with EC fluid in either 100- or 500-D dialysis bags for 4 hours at 37°C. After transfer of the dialyzed fluid into naive cultures of neurons, the cell cultures were allowed to incubate overnight. Those cells treated with traumatized fluid from 100-D dialysis bags retained the toxicity (0.60±.02, n=6), while those treated with fluid from 500-D dialysis bags had a death rate of only 0.17±0.08 (n=4). This result estimates the molecular weight of the toxic mediator to be between 100 and 500.
In our previous report,7 we demonstrated that the glutamate concentration in the traumatized culture never exceeded 20 nmol/L and that MK801, an NMDA receptor antagonist added to EC fluid in the donor (ie, traumatized) culture before injury, had no effect on the neurotoxicity seen in the recipient culture. It is still possible that glutamate receptor activation or glutamate release in the recipient culture could be involved in the mechanism of this neurotoxicity. To test this possibility, CNQX (30 μmol/L) and MK801 (10 μmol/L) alone and in combination were added to the traumatized fluid after removal from the traumatized donor culture (Fig 1B⇑). None of these glutamate receptor antagonists had any effect on the rate of cell death (P>.2, F test comparing all injury groups including control injury).
A rise in the intracellular calcium concentration is one of the critical steps in neurotoxic mechanisms.2 10 To determine whether this toxicity is dependent on entry of calcium into the cell from the extracellular space, all extracellular calcium was removed and replaced with magnesium (Fig 2A⇓). The control death rate in low-calcium EC fluid was 0.34±0.02 (n=4). The removal of calcium had no effect on the neurotoxicity (0.62±0.03, n=10, P>.2, compared with toxicity in standard EC fluid).
Another possible mechanism of rise in intracellular calcium is by release from the intracellular store. This release of calcium can be blocked by dantrolene.11 12 Consistent with this notion, dantrolene dose-dependently protected the neurons from the trauma-induced neurotoxin (Fig 2B⇑). In these experiments, dantrolene was added to the toxic fluid after collection from the traumatized culture and before the medium exchange in the naive culture.
The stability of this toxin in heat suggests that nitric oxide is not the toxin that is transferred with the medium to the recipient culture. However, it is possible that the effect of this toxin could be mediated by the generation of nitric oxide after the transfer, particularly since the rise in intracellular calcium could activate NOS.3 To test this hypothesis, a competitive NOS inhibitor, L-NAME (10 and 100 μmol/L), was added to the fluid collected after trauma (Fig 3A⇓). The protective effect of L-NAME was dose dependent (0.47±0.02, n=4; 0.40±0.03, n=4, respectively). Another competitive NOS inhibitor, L-NMMA (100 μmol/L), also had a protective effect (0.39±0.02, n=6), but its less-active enantiomer, D-NMMA, had no protective effect (0.68±.03, n=6, P<.005, paired t test compared with L-NMMA).
CCK, a gastric octapeptide, has been demonstrated to protect dissociated cortical neurons from glutamate-induced excitotoxicity.13 14 Studies by Tamura et al14 demonstrated that CCK-mediated protection does not involve antagonism of glutamate receptor activation. Instead, CCK directly inhibits the activation of NOS after NMDA receptor activation, with a resultant rise in cytosolic calcium. If the neurotoxicity induced by trauma in the recipient culture is mediated by a rise in cytosolic calcium and activation of NOS, CCK may also attenuate this toxicity. In these experiments CCK was added to the traumatized fluid after the collection from the donor culture. At 100 nmol/L, CCK significantly protected the neurons from trauma-induced neurotoxicity (0.29±0.05, n=7, P<.001). The protective effect of CCK was reproduced by NS-CCK (0.34±0.02, n=5, P<.001), a specific agonist for the central CCKB receptor.15
Our observations suggest that there is a mechanism of propagation of neuronal destruction triggered by physical injury of the neuronal tissue. At the present time the identity of the toxin is unknown. However, this toxin seems to be relatively stable at 37°C for at least 2 hours and can withstand 2 minutes of exposure to boiling temperature. This stability of the compound essentially rules out any unstable compounds, such as free radicals and gases, as the toxin. It is also unlikely that it is a large peptide.
The mechanism of this trauma-induced toxicity is dependent on the activation of NOS, which can be activated by calcium/calmodulin complex.3 4 Consistent with this observation, this toxicity is also dependent on the rise in cytosolic calcium, which seems to be released from the intracellular store instead of from entry from the extracellular space.
Excitotoxicity is an important component of trauma-induced neuronal destruction.6 However, other possible neurotoxic mechanisms have also been implicated.16 For example, a similar observation of trauma-induced release of an unknown toxin has been reported.17 The toxin released in this case was also heat resistant and was released by reactive mononuclear phagocytes. An important variance of the neurotoxicity described in the present report is that glutamate antagonists had no protective effect against this unknown toxin.
CCK has been shown to be widely distributed throughout the brain15 by radioimmunoassay. High levels of CCK precursor-specific mRNA have been detected in the rat hippocampus.18 However, the functional role of this neurotransmitter and its receptors remains to be clearly elucidated.
There are two types of CCK receptors. CCKA is known as the peripheral type of receptor and is found primarily in autonomic ganglia and the gastrointestinal system, although these receptors occur in discrete brain regions as well. CCKB receptors, through the use of radioligand binding19 and selective nonpeptide CCK receptor antagonists,20 have been shown to be abundantly distributed throughout the mammalian brain and have been implicated in most of the effects of CCK in the brain.
The purpose of using CCK as a protective agent in the present study was to indicate that trauma-induced neurotoxicity is mediated by an activation of NOS without activation of glutamate receptors. CCK has been shown to protect cortical neuron cultures from excitotoxicity by activation of the CCKB receptor21 and by attenuation of the nitric oxide mechanism. Consistent with this observation, the neuroprotective effect of CCK against trauma-induced neurotoxicity is reproduced by NS-CCK, which is a specific agonist for the CCKB receptor. Our results indicate that the site of CCK action is a step after the rise in cytosolic calcium.
An important consideration in the mechanism of cellular injury is the role of endogenous protective mechanisms.22 23 Some of these endogenous protective factors seem to be peptides already present within the neuronal milieu. Destruction of such protective factors could account for the increase in toxicity with the heat-treated traumatic fluid (Fig 1A⇑). The fact that the medium dialyzed in 500-D membrane had even less toxicity than the control also could be consistent with the presence of endogenous protective factors.23
Even in this limited model of in vitro physical cell injury, the outcome of cell survival seems to be a consequence of multiple toxic and protective effects. The observations described above indicate that cell death could be mediated by a mechanism independent of glutamate excitotoxicity. On the other hand, this toxic effect seems to converge with excitotoxicity at the point where the rise in cytosolic calcium activates the final steps that lead to cell destruction.
Selected Abbreviations and Acronyms
|L-NAME||=||NG-nitro-l-arginine methyl ester|
|NOS||=||nitric oxide synthase|
This study was supported by National Institutes of Health grant K0801547 (Dr Yoon). The cell cultures were prepared by Sharon Werner.
Reprint requests to Kong-Woo Yoon, MD, Division of Neurosurgery, St Louis University Health Science Center, 3635 Vista Ave at Grand Blvd, St Louis, MO 63110-0250. E-mail firstname.lastname@example.org.
- Received April 17, 1995.
- Revision received August 25, 1995.
- Accepted September 27, 1995.
- Copyright © 1996 by American Heart Association
Choi DW. Ionic dependence of glutamate neurotoxicity. J Neurosci. 1987;7:369-379.
Bredt DS, Ferris CD, Snyder SH. Nitric oxide synthase regulatory sites. J Biol Chem. 1992;267:10976-10981.
Dawson VL, Dawson TM, Bartley DA, Uhl GR, Snyder SH. Mechanisms of nitric oxide-mediated neurotoxicity in primary brain cultures. J Neurosci. 1993;13:2651-2661.
Yamada KA, Dubinsky JM, Rothman SM. Quantitative physiological characterization of a quinoxalinedione nonNMDA receptor antagonist. J Neurosci. 1989;9:3230-3236.
Van Winkle WB. Calcium release from skeletal muscle sarcoplasmic reticulum: site of action of dantrolene sodium? Science. 1976;193:1130-1131.
Beinfeld MC, Meyer DK, Eskay RL, Jensen RT, Brownstein MJ. The distribution of cholecystokinin in the central nervous system of the rat as determined by radioimmunoassay. Brain Res. 1981;212:13-16.
Takahashi Y, Kuwano R, Usui H, Araki K, Hasegawa M, Suz Y. Expression of the cholecystokinin gene in the neuron and paraneuron. Arch Histol Cytol. 1989;52:63-68.
Woodruff GN, Hill DR, Boden P, Pinnock R, Singh L, Hughes J. Functional role of brain CCK receptors. Neuropeptides. 1991;19:45-56.
Giulian D, Vaca K, Corpuz M. Brain glia release factors with opposing action upon neuronal survival. J Neurosci. 1993;13:29-37.