Scientific literature on natural compounds has been recently accumulating an enormous amount of reports describing the neuroprotective ability of natural polyphenolic substances in different models of brain injury both in vitro and in vivo. Resveratrol is one of the most promising compounds, considering recent evidence indicating its potent ability to reduce cerebral damage after ischemia/hypoxia [7, 13, 17], trauma , excitotoxicity , and other conditions leading to neuronal demise [19, 21].
It is widely accepted that neuroinflammation is a key player in various pathological events associated with brain injury. More specifically, microglial activation and the subsequent release of pro-inflammatory cytokines, ROS, and prostaglandins play a role of paramount importance in cerebral damage. Inhibition of COX-2 induction and/or activity has been proven to reduce brain injury after ischemia [46–50], excitotoxicity [47, 51], and MPTP-induced neurodegeneration [52–54]. The most important mechanisms associated with the toxic effects of enhanced COX activity during neuroinflammation include production of PGE2 [55–59] and formation of free radicals leading to oxidative stress [34, 37, 60, 61].
In the present study, we have found that resveratrol is a potent inhibitor of PGE2 and free radical formation by activated microglial cells. These findings add significant information on the molecular mechanisms involved in the neuroprotective effect of this compound. The ability of resveratrol to reduce PGE2 production comes from the modulation of multiple events in the COX/PGE2 pathway: 1) resveratrol is a potent inactivator of the peroxidase reaction of COX-1 , and thus is considered a relatively selective inhibitor of this isozyme; 2) resveratrol significantly diminished LPS-induced expression of mPGES-1 (Figs. 3 and 4), the most important terminal synthase responsible for PGE2 synthesis in activated microglia ; and 3) production of 8-iso-PGF2α, a reliable indicator of free radical generation, is dramatically reduced by low concentrations of resveratrol (Fig. 5).
The relative contribution of each of these mechanisms to the overall reduction in PGE2 by resveratrol is difficult to address based on present data. However, some issues deserve further discussion. The ability of resveratrol to inhibit the peroxidase activity of COX-1 is a well-known pharmacological effect of this compound [43, 44]. COX-1 is constitutively expressed in microglia under resting conditions, and its expression is not induced by LPS as shown in Figs. 3A and 4A, and reported by us before . However, according to present results, there is a significant contribution of COX-1 to PGE2 formation by microglia upon LPS challenge. This is supported by a previous study performed in LPS-stimulated human adult microglial cells, in which selective inhibition of COX-1 was also very effective in reducing PGE2 production . Results from control experiments using other highly selective COX-1 inhibitors (SC-560 and VAS), in addition to resveratrol, indicate that COX-1 isoform is not only important in LPS-induced PGE2 synthesis, but it is also a key source of free radicals in microglia. This is an unexpected observation, and represents the first evidence that microglial COX-1 activity is a significant source of free radicals during neuroinflammation.
It has been previously shown that increased COX activity is associated with oxidative damage following different types of brain injury, including excitotoxicity [61, 63, 64], ischemia [60, 65, 66], and traumatic brain injury . Furthermore, treatment with COX inhibitors has been proven to significantly reduce IL-1β- and LPS-induced oxidative damage in neuronal and microglial cells, respectively [34, 37].
The interesting pharmacological properties of resveratrol in terms of inhibition of COX-1 and direct antioxidant ability, may underlie the dramatic attenuation of PGE2 and free radical production by LPS-activated microglia. It has long been known that free radicals directly increase COX activity and conversely antioxidants reduce COX catalytic activity [68–70]. Based on these reports, one may speculate that maintenance of microglial redox status by resveratrol contributes to the reduction in COX activity and PGE2 production in these cells. This is further supported by our finding that other antioxidants (Trolox C and α-tocopherol) are also able to reduce PGE2 formation (Fig. 1B). Antioxidants have been shown to modulate microglial activation [71–74]. In an earlier study, vitamin E was found to attenuate nitric oxide production and the induction of IL-1α and TNFα expression through suppression of signaling events necessary for microglial activation . Furthermore, inhibition of ROS generation in LPS-activated microglia can reduce PGE2 production as reported previously by Wang et al .
An important new observation of the present study is the dramatic reduction in LPS-mediated expression of mPGES-1 in cells treated with resveratrol (Figs. 3 and 4). Based on very recent and convincing data, production of PGE2 in microglia following LPS treatment is almost entirely dependent on the activity of mPGES-1 . Thus, blockade of mPGES-1 expression by resveratrol is an additional effect of this compound that contributes to its potent ability to block PGE2 synthesis in microglia. Resveratrol effects on mPGES-1 expression seem to be independent on its ability to reduce COX activity/PGE2 formation since other COX inhibitors were unable to modify LPS-induced mPGES-1 upregulation in microglial cells despite their potent inhibitory effects on PGE2 production (Figs. 6 and 7).
To the best of our knowledge, the present study is the first to document the ability of resveratrol to reduce mPGES-1 expression, as shown here in activated microglia. It is of great relevance that resveratrol reduced mPGES-1, but not COX-2 expression (Fig. 4). This suggests that LPS-induced microglial expression of mPGES-1 proceeds through molecular mechanisms which are different from the ones involved in COX-2 induction, providing for the first time evidence that the expression of mPGES-1 and COX-2 are not always coupled as suggested by other authors [75, 76].
It has been recently reported that resveratrol reduces COX-2 expression in mouse BV-2 microglial cells through a mechanism that possibly involves inhibition of NF-κB activation . Our present findings are not in line with these previous observations. The reasons for these discrepancies are not known, but might be related to different cell types and species (mouse BV-2 versus primary rat microglia).
At this point we have not yet identified the exact signal transduction pathway involved in resveratrol's effects on mPGES-1 expression. However, present findings will inspire new investigations in order to elucidate the differential signal transduction pathways responsible for the expression of mPGES-1 and COX-2 in microglia. This is of importance since this could lead to the discovery of new targets for attenuating microglial activation and PGE2 synthesis.
Because of the downstream position of mPGES-1 in the PGE2-synthesizing cascade, selective pharmacological blockade of its expression, as shown here for resveratrol, would affect only the pro-inflammatory PGE2, and would not decrease the production of other physiologically important prostanoids. Although COX-2 inhibitors have been highly marketed in the last five years including clinical trials in Alzheimer's disease, there are also important side effects associated with this new group of drugs [78, 79]. Thus, it is of paramount importance our finding that resveratrol specifically reduced mPGES-1 without affecting COX-2 levels. Recent evidence indicates that enhanced expression and activity of mPGES-1 is a critical pathological event during inflammation, both in the CNS and in the periphery. Mice lacking the mPGES-1 gene are protected against stroke-induced injury  and display reduced pain hypersensitivity and inflammation [81, 82].