The objective of this study was to investigate the cellular mechanisms involved in glutamate-induced apoptotic cell death in neuronal cells. We used the mouse hippocampal cell line HT22 and primary fetal rat cortical cells to demonstrate that glutamate can induce neuronal apoptotic cell death by apoptotic mechanisms and that the process can be reversed or inhibited by equine estrogens such as 17β-E2 or Δ8,17β-E2. The results further indicate that glutamate-induced apoptotic cell death involves the activation of apoptotic proteases calpain and caspase-3 as well as upregulation and/or translocation of apoptosis-inducing factor (AIF) from mitochondria into cytosol and nuclei, where it causes nuclear condensation and apoptosis by a caspase-independent pathway [44, 45]. Our results reveal that in these two cell types, primary cortical cells and the HT22 cell line, cells display different patterns in regulating specific proteins involved in apoptosis. Thus, in primary cortical cells, glutamate-induced apoptotic cell death is associated with activation of calpain and caspase-3, which is in agreement with studies using different stimuli such as staurosporine , and release of apoptosis-inducing factor (AIF) from mitochondria into cytosol and nuclei. However in the mouse hippocampal cell line HT22, glutamate-induced apoptotic cell death involved calpain protease activity and upregulation of AIF protein, but not caspase-3 activity. Taken together, our data demonstrate that glutamate-induced apoptotic cell death in neuronal cells could be mediated by the apoptotic proteases calpain and caspase-3 as well as mitochondrial apoptosis effector AIF via a caspase-dependent or caspase-independent apoptotic pathway that depends on the cell type. In HT22 cells, caspase-3 protein was not detectable and glutamate-induced cell death occurred without elevation of caspase-3 activity (data not shown). Caspase inhibitors, either the specific caspase-3 inhibitor DEVD or the pancaspase inhibitor z-VAD, are not capable of protecting against glutamate-induced cell death in HT22 cells, while they do inhibit glutamate-induced cell death in primary cortical cells. In addition, glutamate-induced DNA fragmentation (ladder) can be reduced or blocked by calpain inhibitors but not by caspase inhibitors in this cell line. These results strongly indicate that glutamate-induced apoptotic cell death in HT22 cells is executed through a caspase-independent pathway as has been observed in platelets .
Caspase-3 is considered a key executioner in the apoptotic cell death cascade and shares numerous substrates with Ca2+-dependent protease calpain such as cytoskeletal protein α-fodrin in which they cleave different epitope sites once activated . Though they both are capable of executing apoptotic cell death, they appear to play different roles in this process. Our results show that immunoblots probed with an anti-α-fodrin monoclonal antibody could detect intact α-fodrin (MWr = 240 KDa) and 145 KDa and 120 KDa BDPs . Treatment with glutamate induced accumulation of both 145 KDa and 120 KDa BDPs in primary cortical cells, which could be reduced with both calpain inhibitor (ALLN) and caspase-3 inhibitor (z-DEVD-fmk). Thus, calpain inhibition specifically protects against glutamate-induced production of 145 BDP and caspase-3 inhibition inhibits production of 120 BDP. There were no synergetic effects on glutamate-induced apoptotic cell death when these protease inhibitors were combined. These studies indicate that both calpain and caspase-3 are indeed activated, and their activation occurs sequentially and independently after cells undergo apoptosis, suggesting they may execute cell death to some extent by distinct mechanisms in primary cortical cells. There is no evidence showing that they have overlapping roles in cell death as described previously in neuronal or non-neuronal cells . Moreover, in HT22 cells, only calpain-mediated α-fodrin 145 KDa BDP appeared after glutamate treatment, but 120 KDa BDP generated specifically by caspase-3 [47, 48] could not be detected. Furthermore, glutamate-induced cleavage of α-fodrin to 145 BDPs could be reduced or blocked by calpain inhibitors, further indicating that glutamate-induced apoptosis is mediated by a calpain cascade via a caspase-independent pathway in HT22 cells. Additional evidence for calpain's involvement in apoptotic cell death in central nervous system is supported by recent studies examining in
vivotraumatic brain injury TBI  and ischemia in the rat heart , in which calpain-mediated cleavage of α-fodrin and DNA fragmentation have been detected following these insults.
Calpain is a calcium-dependent neutral protease with two isoenzyme forms, μ-calpain and m-calpain, distinguished by their in
vitrocalcium requirements. Calpain exists as a proenzyme in an inactive form in cytosol where the normal range of Ca2+ concentration is 50–100 nM in resting cells [25, 51]. An elevation of cytoplasmic free Ca2+ concentration triggers calpain activation, which also depends upon the amount of calcium and site of activation. μ-Calpain is activated in the presence of low micromolar levels of Ca2+, and m-calpain requires millimolar Ca2+ levels for its activation. After calpain is activated, it cleaves many cytoskeletal proteins and signaling molecules such as fodrin, filamine, neurofilament protein tau and tubulin [24, 51]. Fodrin (α II spectrin) was found to be a specific calpain substrate [52, 53]. In our studies, a calpain-mediated cleavage fragment of α-fodrin was observed in two cell types exposed to glutamate, as seen in the dystrophic neuritis and senile plaques in Alzheimer's diseased brains , suggesting that calpain could be a major apoptotic protease involved in the pathogenesis of neurodegenerative diseases.
It has been reported that calpain could be activated in both necrotic and apoptotic conditions as observed in various neurological and neurodegenerative disorders, such as cerebral ischemia and glutamate toxicity [55, 56], while caspase-3 is only activated in neuronal apoptosis [44, 57–59]. Nevertheless, in our model systems, glutamate-induced cell death in HT22 cells was found to involve the mechanism of apoptosis (programmed cell death) characterized by the presence of DNA fragmentation (DNA ladder) and morphological cell changes, and this process could be attenuated by calpain inhibitors and not caspase inhibitors. Thus, in cortical cells, both 145 KDa and 120 KDa BDPs were formed in untreated cells and cells treated with glutamate, which was also associated with DNA fragmentation (data not shown), and calpain-mediated 145 KDa BDP accumulated to a greater extent than 120 BDP. Furthermore, activation of calpain precedes that of caspase-3, glutamate-induced LDH release (cell death) and DNA fragmentation in these cells. Taken together, our results strongly suggest that calpain, rather than caspase-3, plays a critical role in mediating apoptosis characterized by oligonucleosomal DNA fragmentation (DNA ladder) and is a key apoptotic executioner in neuronal apoptosis, especially in HT22 cells. Whether calpain activity is associated with necrosis needs to be investigated.
We have previously reported that glutamate induces cytochrome c release from the mitochondria into the cytoplasm where it activates caspase-3 and causes apoptotic cell death in primary cortical cells via a caspase-dependent pathway . In the present paper, we show that apoptosis inducing factor (AIF), normally confined in the mitochondria, is also released into the cytosol and nuclei in primary cortical cells undergoing apoptosis and upregulation of AIF also occurs in HT22 cells after glutamate treatment. In contrast to cytochrome c, AIF acts in a caspase-independent fashion [19, 60]. AIF is a ubiquitously expressed flavoprotein with significant homology to bacterial oxidoreductases and has NADH oxidase activity . Under normal circumstances, transcription and translation of nuclear AIF gene give rise to a precursor molecule (67 KDa) that carries a putative mitochondrial localization sequence in its NH2 terminus . Upon import into the mitochondrial intermembrane space, this 100 amino acid precursor is cleaved by a local peptidase leading to generation of the mature AIF molecule (57 KDa). When apoptotic cell death is induced, AIF translocates through the outer mitochondrial membrane to the cytosol and nucleus, where it leads to nuclear chromatin condensation and a large-scale DNA fragmentation (high molecular weight DNA fragments) and apoptosis in a caspase-independent manner [10, 11]. Recent studies show that microinjection of mature AIF and precursor AIF can both induce chromatin condensation and cell death [61, 62]. It was further found that transfection enforced overexpression of the wild type precursor AIF (precursor AIF protein) also diminishes the permeability of the mitochondrial membrane (ΔΨμ), induces AIF release from the mitochondria and causes apoptosis. Based on these studies, glutamate-induced overexpression of AIF in HT22 cells might also trigger AIF release or translocation from the mitochondria into the cytosol and nuclei. This process needs to be further elucidated. Taken together with previous studies, our data indicate that glutamate exerts its toxic action through two parallel pathways: a caspase-dependent pathway mediated by caspase-3, the final apoptotic effector, and a caspase-3 independent apoptotic pathway involving the upregulation of AIF and/or mitochondrial AIF release as well as calpain-modulated effects (cell death).
We have previously reported that a number of equine estrogens, which are components of the drug CEE used extensively for management of vasomotor symptoms and osteoporosis in postmenopausal women, are potent antioxidants and protect neuronal cells against cell death induced by oxidized LDL or glutamate [32, 33, 35, 63–65]. Our current findings indicate that estrogens prevent glutamate-induced cell death by reducing the upregulation and/or translocation of AIF from the mitochondria into the cytosol as well as by attenuating activity of calpain and caspase-3 in primary cortical cells and HT22 cells, with Δ8, 17β-estradiol being more potent than 17β-estradiol. Although neurotoxic effects of oxidized LDL and glutamate can be inhibited differentially by various estrogens, [32, 33, 35], the data presented in this paper, to our knowledge, represents the first comprehensive analysis of the involvement of multiple apoptotic proteins in glutamate-induced apoptosis in different neuronal cells that can be differentially inhibited by equine estrogens. Recent studies have shown that estrogen can decrease calpain activity in rat C6 glial cells treated with H2O2 and in ischemic muscle [66, 67]. The regulation by estrogens on the translocation of apoptotic effectors from the mitochondria into the cytosol such as cytochrome c and the activation of caspase-3 has been discussed in previous reports in which estrogens might directly modulate the mitochondrial Ca2+ content and mitochondrial trans-membrane potential . As the estrogen receptor has been shown to be a substrate of calpain, estrogen-mediated reduction of calpain activity may prevent degradation of the ERs, and enable receptor-mediated events to proceed by genomic mechanism . Whereas in HT22 cells lacking estrogen receptors, estrogen effects are most-likely mediated to some extent by non-genomic mechanism [4, 69]. It may also be related to an effect on voltage-gated Ca2+ channels and reduced post-injury influx of Ca2+ . It has been reported that estrogen treatment can attenuate Ca2+ influx in cultured neurons . However, our earlier findings indicate that estrogens protected primary cortical cells against glutamate induced excitotoxicity by a mechanism that appears to be independent of Ca2+ influx . In this study in primary cultures of cortical neurons, we investigated the glutamate-induced excitotoxicity by exposing the neurons to μM concentration of glutamate for only 20 minutes. The results from this study indicated that glutamate-induced cell death occurs through NMDA receptors and NOS-linked mechanism independent of Ca2+ influx and this form of cell death was also prevented by estrogens. A recent study  has confirmed our latter observations in neuronal cells derived from fetal rat brains, however, these investigators also used middle aged and old rats neurons where there was a profound loss of calcium homeostasis more so in the older age group. These observations indicate that neurons from older rats are more sensitive to glutamate toxicity, however, as in neurons from fetal brains, estrogens were also able to protect these older neurons. Whether the role of calpain, caspase-3 and AIF is altered in cortical neurons derived from older animals remains to be investigated.
Pre-treatment of these neurons with 17β-estradiol prior to exposure to glutamate attenuate and delay the adverse effects on intracellular Ca2+ homeostasis. Whether the role of calpain, caspase-3 and AIF is altered in cortical neurons derived from older rat brains remains to be investigated. Our observations further suggest that the neuroprotective effects of estrogens can involve both genomic and non-genomic mechanisms that depend on the neuronal cell type and the cytotoxin used to induce apoptosis.
Recently, Bano et al  reported that the Na+/Ca2+ exchanger (NCX) which under normal physiological conditions, transports Na+ ions into the cell and Ca2+ out, is destroyed by Ca2+-activated calpain under ischaemic conditions. This proteolytic inactivation of NCX could allow further accumulation of Ca2+ or delay its elimination, that ultimately results in the destruction of the neurons. Whether glutamate induced Ca2+ influx and the activation of calpain observed in our neuronal cells also results in the inactivation of NCX that can be prevented by estrogen remains to be investigated.
Although our present data and those of a number of previous studies [20, 21, 23, 24, 26–29] indicate that calpains play a critical role in mediating apoptotic cell death, others indicate that in cerebellar granule cells and rat fetal hippocampal neurons activation of NMDA receptors is necessary for calpain-1 activation but no association between calpain activation and glutamate-induced cell death was observed [74–76]. Differences between these observations could be due to several factors such as cell type, primary cell cultures or stable cell line cultures, tissue culture conditions, source of cells (fetal, middle age or adult brains), age of cultures, purity of cultures concentration and duration (minutes to hours) of glutamate exposure, nature of the neurotoxin used, and extent of Ca2+ influx [13, 74–76]. Along with these, a number of additional differences have been indicated . Further detailed studies are needed to delineate the role of calpains in glutamate-induced apoptotic neuronal cell death in clearly defined model systems. It is also important to recognize that at least two distinct pathways for glutamate-induced cell death have been described: the excitotoxic pathway and the oxidative pathway. The excitotoxic pathway involves the overactivation of glutamate receptors that leads to both rapid and slowly triggered cytotoxic events. The rapid effects involve the activation of the NMDA receptor and leads to a large Ca2+ influx that is detrimental to cell viability. The oxidative pathway involves the breakdown of the glutamate-cystine antiporter and a drop in glutathione levels that allows for aberrant formation of reactive oxygen species (ROS) that are neurotoxic [77–79]. In most of our studies, we and a number of other investigators have used higher concentration (mM) and exposed the cultures to longer exposure times (hours vs minutes) to glutamate. Under these conditions, we usually get 30 to 50% cell death. This cell death occurs most likely by mechanism distinct from excitotoxicity induced by acute exposure of neurons to μM concentrations of glutamate. We think as indicated in the first sentence in the background it is the high concentrations (mM) of the excitatory neurotransmitter glutamate that accumulates in the aging brain that may play a role in neurodegeneration associated with normal aging. In contrast, in acute trauma to the brain, excitotoxicity may be the key factor. In our recent study  in primary cultures of cortical neurons, we investigated the glutamate-induced excitotoxicity by exposing the neurons to μM concentration of glutamate for only 20 minutes. The results from this study indicated that glutamate-induced cell death occurs through NMDA receptors and NOS-linked mechanism independent of Ca2+ influx and this form of cell death was also prevented by estrogens. Thus, depending on the objectives of the study, we have used different concentrations of glutamate. Important to note that in HT22 cells, NMDA receptors are absent and therefore this form of excitotoxicity may not occur in these types of neuronal cells.
Since in our study both NMDA receptor and estrogen receptor positive (primary cortical cells) and negative (HT22) cells were used, the results therefore indicate that at least two different mechanisms are involved in the activation of calpains. Which of these mechanisms is involved in neurodegenerative diseases such as Alzheimer's remains to be investigated. It is likely that in the aging brain, neurons are more sensitive to chronically high levels (mM) of glutamate that can result in the activation of calpains and the initiation of neurodegenerative process. Over time, the caspase-3 system is also activated leading to a higher degree of apoptotic neuronal cell death and progression of the disease process. The demonstration that equine estrogens can prevent the activation of both types of proteases, provides an avenue of developing new therapeutic interventions for the prevention of neurodegenerative diseases.