PGE₂ EP1 receptor exacerbated neurotoxicity in mouse model of cerebral ischemia and Alzheimer’s disease

1. Introduction

Accumulating evidence linking stroke and Alzheimer’s disease (AD) indicates that each exacerbates the severity of the other (Jendroska et al., 1995; Koistinaho and Koistinaho, 2005). Because there are few effective treatments for AD, especially when stroke supervenes in these patients, finding a common pathway that exacerbates neuronal loss and brain damage might provide a potential therapeutic target. Over the years, prostaglandin E₂ (PGE₂) and cyclooxygenase-2 (COX-2), the rate-limiting enzyme responsible for prostaglandin production, have become largely discussed targets for exploration in therapy for AD and stroke patients (Ahmad et al., 2008; Bazan et al., 2002; Santovito et al., 2009). Clinical use of non-steroidal anti-inflammatory drugs or selective COX inhibitors is associated with various concerning side effects (Donnelly and Hawkey, 1997; Ray et al., 2004); thus, studying the downstream pathway may provide a unique therapeutic target which would allowed to be more selective and at the same time not simply blocking synthesis of all prostaglandins. Previously, the EP1 receptor expression in CNS neurons has been documented in the literature; for example, signals for EP1 receptor mRNA have been exposed in C57BL/6 mouse brain e.g. www.brain-map.org, and the EP1 receptor has also be documented by various independent teams at the protein level in the CNS and on neuronal cells (Kawano et al., 2006; Carlson et al., 2009). Additionally, previous studies by others and by us have provided physiological evidence that deletion of EP1 receptor significantly attenuates focal ischemic and excitotoxic brain damages and that treatment with selective receptor ligands provided significant anatomical and functional biological outcomes (Ahmad et al., 2006; Kawano et al., 2006). Thus, our study focuses on delineating the role of PGE₂ EP1 receptor in AD model in the setting of cerebral ischemia. Based on this team effort, we propose that the EP1 receptor may be critical in the inflammatory and cell death pathways that are both hallmarks in the neurodegenerative conditions of AD and stroke.

We used the APPswe/PS1ΔE9 double gene mutation (APP/PS1) AD mouse model and crossed them with EP1 gene knockout (EP1^−/−) mice, and explored the potential EP1 receptor role in stroke and AD-like conditions. We compared β-amyloid (Aβ) plaques, Aβ40 and Aβ42 levels, and behavioral outcomes of ischemia in APP/PS1xEP1^−/− and APP/PS1 mice in together with their none-AD control mice. Furthermore, to explore the potential mechanism of the neuroprotective effect of genetic knocking out the EP1 receptor, we investigated intracellular Ca²⁺ response in an Aβ-enriched environment in cultured neurons.

2. Materials and Methods

3. Results

First, we compared Aβ plaque number, Aβ concentration and neurobehavior functions between naïve APP/PS1 and APP/PS1xEP1^−/− mice. Significantly lower Aβ burden in APP/PS1xEP1^−/− mice was evidenced by fewer plaques in the hippocampal area and a lower concentration of soluble Aβ42 and Aβ40 (Fig 1A). The reduced plaque load and Aβ levels may explain the benefits seen in memory function in APP/PS1xEP1^−/− mice (Fig. 1A). To ascertain whether the Aβ-mediated cell death mechanism is different in the absence of EP1, we exposed hippocampal slice cultures and neuronal cell cultures to Aβ42. The CA1 regions in slice cultures and the cultured neurons from EP1^−/− mice each were significantly more resistant to Aβ42 toxicity than were the corresponding WT controls (Fig. 1B). We then performed Ca²⁺ imaging in neuronal cell cultures to explore the potential mechanism by which EP1 receptor inhibition improves neuronal survival during Aβ-induced toxicity. The [Ca²⁺]_i and Ca²⁺ transients were increased after Aβ42 treatment in WT controls; this increase was prevented by pretreatment with an EP1 receptor antagonist (SC51089) and blocked in EP1^−/− neurons (Fig. 1C).

We then asked whether the reduced Aβ burden in APP/PS1xEP1^−/− mice led to reduced injury development after cerebral ischemia. We approached this question in vitro by exposing cultured hippocampal slices from each of the mouse genotypes to OGD. Cell death was significantly decreased in EP1^−/− tissue (P < 0.001) and increased in the APP/PS1 tissue compared to WT controls (P < 0.01) but was significantly attenuated in the APP/PS1xEP1^−/− tissue compared to APP/PS1 tissue (P < 0.001; Fig. 2A).

To correlate our results with in vivo experiments, we assessed delayed hippocampal neuronal damage by subjecting mice to global forebrain ischemia. Similar to the in vitro findings, CA1 neuronal loss was much greater in the APP/PS1 than in the WT mice and was significantly attenuated in the APP/PS1xEP1^−/− mice compared to the APP/PS1 mice (Fig. 2B). Hippocampal damage caused by ischemic insult can lead to impaired memory retention, reflected in shortened response latency in the passive avoidance task (Sakurai et al., 2008). Correlating with the anatomical differences, APP/PS1 mice had impaired memory retention after ischemic insult compared to WT mice subjected to ischemia and sham-operated APP/PS1 mice. There was no statistical difference between the APP/PS1xEP1^−/− and APP/PS1 ischemia groups, but the APP/PS1xEP1^−/− mice tended to show greater memory retention (Fig. 2C). We also questioned whether Aβ burden increases after ischemic insult in APP/PS1 mice and if so, whether this increase is attenuated in the absence of the EP1 receptor. We observed less intense Aβ immunostaining in the CA1 neurons of APP/PS1xEP1^−/− mice after ischemia compared to that of APP/PS1 mice (Fig. 2D). the ELISA assay revealed that the insoluble Aβ42 concentration rose significantly after ischemia in APP/PS1 mice but remained unchanged in APP/PS1xEP1^−/− mice. Both the soluble and insoluble Aβ concentrations were higher in APP/PS1 mice than in APP/PS1xEP1^−/− mice post-ischemia (Fig. 2E). The increase of insoluble Aβ in APP/PS1 mice suggests a tendency for soluble Aβ to accumulate and potentially fibrillate after ischemia. Deletion of EP1 appeared to attenuate the increase of Aβ generation and accumulation.

4. Discussion

We observed that following transient global forebrain ischemia, APP/PS1 double mutant mice sustain more severe neuronal and functional damage than WT controls. This suggests that the Aβ burden is at least partially responsible for such outcomes. Knowing that brain levels of both COX-2 and PGE₂ are increased in AD and in stroke, we questioned whether the deletion of the EP1 receptor—which has been proposed to be a key modulator of COX-2/PGE₂-related neurotoxicity—would be sufficient to alleviate most of the negative outcomes. In our study, we provide the first evidence that genetic deletion of EP1 is sufficient to decrease basal and post-ischemic levels of Aβ in APP/PS1 mice. We discovered that hippocampal neuron cell death and neurobehavioral deficits after ischemic insult are attenuated in the APP/PS1^−/−xEP1 as compared to APP/PS1 mice. Isolated neurons and brain slice cultures revealed that neurons derived from EP1^−/− mice were more resistant to Aβ toxicity than those from WT mice. Neuronal Ca²⁺ signaling and Ca²⁺ homeostasis regulate multiple neuronal functions, including synaptic transmission, plasticity, and cell survival (Wojda et al., 2008). EP1 receptor inhibition prevented induction of Ca²⁺ signaling initiated by Aβ exposure by decreasing the frequency of Ca²⁺ transients and mitigating the increase in baseline Ca²⁺ levels. Prevention of Aβ-associated Ca²⁺ signaling in neurons by EP1 receptor inhibition is likely an additional pathway for the neuroprotective effects seen. Our study provides the rational to further explore potential therapeutic agents targeting EP1 receptor in stroke and AD-like neuropathologies.

Supplementary Material