Knockout of the FLAP gene was associated with ceased LT production and amelioration of stroke damage in terms of mortality-adjusted infarct size. Furthermore, there was a clear trend of improved mortality-adjusted functional test performance in the knockout group.
Activation of LT synthesis involves translocation to the nuclear envelope and endoplasmic reticulum of cytosolic phospholipase A2 (cPLA2) and of 5-LO from the cytosol and nuclear matrix. cPLA2 subsequently releases AA from membrane phospholipids [3, 5, 35]. FLAP resides at these locations as an integral membrane protein [3–6, 8] and facilitates the transfer of AA from cPLA2 to 5-LO . The labile intermediate LTA4, formed by 5-LO, is converted to LTB4 by LTA4 hydrolase or to LTC4 by LTC4 synthase. LTC4 is exported from the cell and metabolized to two other cysteinyl (Cys) LTs, LTD4 and LTE4. Both 5-LO and FLAP are required for LTA4 synthesis from endogenous AA , and both FLAP knockout mice and 5-LO knockout mice lack detectable LT production . This was confirmed for the FLAP knockout mice in the current study by stimulating WT and KO bone marrow cells under conditions which allow LTB4 formation, but not LTC4 formation to occur due to lack of cysteine in the incubation mixture [33, 34]. Thus LTB4 represents the total production of LTs from LTA4. LTB4 is a potent chemoattractant, which recruits inflammatory cells to sites of inflammation [38–40]. It also contributes to leukocyte accumulation by attenuation of leukocyte apoptosis [41, 42]. CysLTs cause wide-spread plasma leakage by increasing vascular permeability and attract subsets of T-cells [43–45]. They also activate dendritic cells and their cytokine release [46, 47] as well as mast cell cytokine production , which may also influence the inflammatory state. Hence, alterations in the LT pathway affect the inflammatory response, and such alterations could in turn have an impact on cerebral ischemia.
It should be noted, however, that inflammation is not only an important feature in the specific infarct process, but also in the pathology of atherosclerosis. As mentioned above, several genetic studies on human populations have linked LT-related genes to altered stroke incidence [15–18]. Such studies, however, do not confer firm evidence regarding the mechanisms of the effects. This emphasizes the importance of animal studies, since they are sin qua non for investigating biological mechanisms. As already mentioned, several reports, using models similar to those in the current study, have demonstrated protection against stroke by drugs blocking LT effects, such as montelukast [23, 24] and pranlukast [27, 28], strongly suggesting LT-related effects on the specific infarct pathophysiology. Experiments using drugs and those employing genetically modified animals are important complements to each other. The study using a 5-LO knockout mouse strain, mentioned above, showed no effects on infarct size after transient MCAo. That study, however, only included 6 mice per transient MCAo group, and with infarct size coefficient of variation (standard deviation divided by mean value) of around 42% and α = 0.05, the chance (statistical power) of detecting for example a 30% difference was 45.1%. In other words, the study was underpowered in this specific respect, thus not substantiating negative conclusions . To the best of our knowledge, the current study is the first to show that a genetic distortion of the LT system ameliorates the detrimental effects of cerebral ischemia.
It should be noted that altered inflammatory response is not the only possible mechanism for the decreased infarct volumes in the knockout group in the current experiment. Even though the cerebrovascular anatomy was similar between the groups, the lack of leukotriene production may theoretically have affected the blood flow to the brain, which in turn could decrease the infarct size. For example, exogenous LTD4 increases the blood pressure in rats  and the FLAP inhibitor MK886 ameliorates hypertension in L-NAME treated rats .
It is a well-known problem that MCAo studies often suffer from high random variability regarding infarct size, and numerous attempts have been made to address this [51–54]. The source of this variability can be a consequence of for example inconsistency in the filament insertion procedure and to subtle, individual variations in cerebral vasculature as well as in peroperative hydration status and body temperature. In the current study, efforts to minimize random variability included a strictly standardized operation procedure performed by one single surgeon, peroperative surveillance of physiological parameters and the use of an inbred mouse strain, minimizing inter-individual differences.
Strengths and weaknesses of the current study
A frequent problem in animal stroke studies is that mortality is neglected, and not included in the final analysis. A strength of the current study is that this was addressed by combining mortality with infarct size and functional score, respectively, in two mortality-adjusted non-parametrical models. The advantage of this approach is that the importance of the extreme outcome of death is acknowledged. A theoretical drawback is that if mortality was very high in one of the groups, that factor itself could contribute with so much group difference that any other variable combined with the mortality rate would seem significant. In the current study, with only 2 included cases of death, this was not a concern.
Even though 2,3,5-triphenyltetrazolium chloride (TTC) staining is a well-used and validated method for infarct size assessment, other staining procedures could have provided differentiated information regarding the mode of cell death. It should be noted that any eventual differences between the groups regarding mode of cell death remain undisclosed in the current experimental setup. Such differences may potentially contribute to explaining the mechanism of the reduced infarct sizes in the knockout group, and merit attention in future studies.
In a preceding pilot study, different MCA occlusion times were tested with the result that for this specific mouse type, 120 minutes was needed to ensure a relatively consistent infarction. This however caused quite high mortality if longer convalescence periods were adopted, which was why we settled for 24 h even though longer survival times could be beneficial. Theoretically, the infarct evolution may merely have been delayed in the knockout group.