In the present study, we show that OGD induces both apoptotic and necrotic astrocytic cell death. We find that similar to what is observed for both acute and chronic neurodegenerative conditions, such as pain, amyotrophic lateral sclerosis, multiple sclerosis, and epilepsy [22–25, 31–33] OGD also results in an increase in mGluR5a expression. We find InsP formation in cortical astrocytes is increased following OGD and that the treatment of cultures with the mGluR selective antagonist MPEP, not only attenuates OGD-mediated InsP formation, but the genetic deletion of mGluR5 selectively protects against OGD-induced apoptosis. The effects of mGluR5 are independent of the activation of the ERK1/2 signaling pathway, but appear to be associated with increased signaling via the PLC-InsP3-calcium pathway and Homer protein interactions.
We show that the mGluR5a expressed in response to injury are functional and activated after the ischemic insult. Group I mGluRs are known for their high constitutive and agonist-independent activity when they are overexpressed in heterologous systems . However, in this case, because OGD-induced mGluR5 expression in astrocytes is not as high as those found in neuronal cultures or in transfected cells , increased mGluR5 activity may also be due to glutamate released by astrocytes in response to ischemia. Indeed, ischemic conditions could result in glutamate release by astrocytes through a variety of mechanisms . The observed increases in InsP formation following OGD are consistent with previous work demonstrating that OGD results in increased InsP formation, likely via a metabotropic glutamate receptor .
We observe that OGD treatment activates a time-dependent increase in ERK1/2 phosphorylation that appears to be independent of mGluR5 activity, as increases in ERK1/2 phosphorylation were not antagonized by treatment of cultures with the mGluR5 selective antagonist MPEP. Furthermore, the genetic deletion of the mGluR5 gene did not result in diminished ERK1/2 phosphorylation in response to OGD. Indeed, in contrast to the InsP3 pathway, the mGluR5-induced activation of MAPK pathway occurs via mechanisms that are largely independent of PLC activation and calcium release in both striatal neurons (Homer protein dependent) and astrocytes (EGF receptor transactivation) [19, 37, 38]. The OGD-induced increase in ERK activation in astrocytes, as previously reported , did not significantly change in presence of the mGluR5 antagonist MPEP. Thus, it is possible that glutamate released in response to OGD may activate alternative mGluR subtypes that may be directly linked to the activation pro-survival signal transduction pathways in astrocytes. Consistent with this, mGluR3 is implicated in protecting against astrocytes apoptosis following OGD .
Interestingly, the mGluR5-dependent activation of PLC-mediated signaling leads to excessive calcium release and excitotoxic striatal neuronal death, which could be prevented by the down regulation of Group I mGluR expression following estradiol treatment . Thus, a selective increase in PLC signaling likely correlates with a pro-apoptotic role for astrocytic mGluR5, as OGD-induced apoptosis is reduced in astrocytes derived from mGluR5 null embryos. Moreover, selective mGluR5 blockade has been reported to be neuroprotective . Thus, it appears that increased mGluR5 signaling in astrocytes might lead to cell death following OGD-increased mGluR5a protein expression levels. Interestingly, mGluR5 has been shown to be upregulated in the spinal cord of ALS patients [22, 24] and astrocytic cell death in an amyotrophic lateral sclerosis transgenic model has been shown to be attenuated by blocking the mGluR5 receptor in vivo. Moreover, the disease onset was also delayed by this treatment. Similarly, the brain permeable mGluR5 antagonist MTEP provides neuroprotection in an in vivo ischemic model .
Because the effect of the InsP3 receptor inhibitor Xestospongin C on apoptosis is not observed in astrocytes lacking the mGluR5, we have confirmed that the mGluR5-PLC-InsP3-calcium signaling pathway is responsible for induction of apoptosis by releasing calcium from intracellular stores. As reviewed by Hanson et al. , calcium released through InsP3 receptors is sequestered by mitochondria and mitochondrial calcium overload can induce the apoptotic cascade. In addition, calcium release can also stimulate calcium-sensitive enzymes such as the phosphatase calcineurin, which is involved in other apoptotic mechanisms . Furthermore, we show that the formation of a molecular complex between Homer proteins and mGluR5 contributes to the activation of a pro-apoptotic pathway, as the treatment of astrocytes with a Tat-Homer interference peptide reduced OGD-induced astrocytes apoptosis. This Tat-peptide was previously shown to block mGluR5/Homer interactions but also blocks Homer-dependent activation of ERK1/2 . Thus, although we did not observe reduced ERK1/2 phosphorylation in response to OGD following either MPEP treatment or the loss of mGluR5 expression, it is possible that mGluR5-dependent activation of the ERK1/2 pathway via a Homer-dependent mechanism may also function to antagonize mGluR5-mediated apoptosis of astrocytes that is induced as the consequence of PLC-mediated mGluR5 signaling.
In conclusion, because of the neuroprotective role of astrocytes, astrocytic cell death could have a huge impact on acute or chronic neurodegenerative conditions. Thus, targeting mGluR5-signaling pathway involved in the induction of astrocyte apoptosis may have therapeutic benefits following ischemic insults such as stroke. For example, selective blockade of mGluR5-dependent PLC signaling in astrocytes might help limit cellular loss in a variety of neurodegenerative diseases. In contrast, the selective activation of mGluR5 G protein-mediated PLC signaling might promote apoptotic cell death in brain tumors, such as astrocytomas. This is of particular importance considering that mutations in mGluR1a have been associated with a variety of tumors, and that many of the mutations bias receptor signaling via the activation of ERK1/2 at the expense of G protein signaling [44–46] suggesting it may be possible to design biased allosteric agonists to prevent both astrocytic and neuronal cell death.