Microglia depletion diminishes key elements of the leukotriene pathway in the brain of Alzheimer's Disease mice

Abstract

Leukotrienes (LTs) contribute to the neuropathology of chronic neurodegenerative disorders including Alzheimer's Disease (AD), where they mediate neuroinflammation and neuronal cell-death. In consequence, blocking the action of Leukotrienes (LTs) ameliorates pathologies and improves cognitive function in animal models of neurodegeneration. Surprisingly, the source of Leukotrienes (LTs) in the brain is largely unknown. Here, we identified the Leukotriene (LT) synthesis rate-limiting enzyme 5-Lipoxygenase (5-Lox) primarily in neurons and to a lesser extent in a subpopulation of microglia in human Alzheimer´s Disease (AD) hippocampus brain sections and in brains of APP Swedish PS1 dE9 (APP-PS1) mice, a transgenic model for Alzheimer´s Disease (AD) pathology. The 5-Lipoxygenase (5-Lox) activating protein (FLAP), which anchors 5-Lipoxygenase (5-Lox) to the membrane and mediates the contact to the substrate arachidonic acid, was confined exclusively to microglia with the entire microglia population expressing 5-Lipoxygenase activating protein (FLAP). To define the contribution of microglia in the Leukotriene (LT) biosynthesis pathway, we ablated microglia using the colony stimulating factor 1 receptor (CSF1R) inhibitor PLX5622 in wildtype (WT) and APP-PS1 mice. Microglia ablation not only diminished the expression of FLAP and of the Leukotriene (LT) receptor Cysteinylleukotriene receptor 1 (CysLTR1), as expected based on their microglia cell type-specific expression, but also drastically reduced 5-Lipoxygenase (5-Lox) mRNA expression in the brain and its protein expression in neurons, in particular in wildtype (WT) mice. In conclusion i) microglia are key in Leukotriene (LT) biosynthesis, and ii) they regulate neuronal 5-Lipoxygenase (5-Lox) expression implying a yet unknown signaling mechanism between neurons and microglia.

Keywords: Alzheimer’s disease; Cysteinyl-leukotrienes; Leukotrienes; Microglia; RNAseq.

Fig. 1

LT biosynthesis is part of metabolism of the omega 6 polyunsaturated arachidonic acid (AA), which is cleaved from plasmamembranes by phospholipase A₂ (PLA₂). AA is a substrate for several lipoxygenases. For the production of LTs it is first converted into 5-hydroxy-peroxy-eicosatetraenoic acid (5-HPETE) and afterwards into LTA4, both steps mediated by a complex of the enzyme 5-lipoxygenase (5-Lox) and 5-lipoxygenase activating protein (FLAP) [22]. From LTA4 arise either Leukotriene B4 (LTB4) formed by epoxide hydrolase or Leukotriene C4 (LTC4) formed by LTC4 synthase, which can be further metabolized to LTD4 and LTE4 [23]. LTC4, LTD4 and LTE4 are summarized under the term cysteinyl-leukotrienes (CysLTs). CysLTs display different affinities towards the receptors cysteinyl-leukotriene receptor 1 and 2 (CysLT1R [24]), CysLT2R [25]) and G-protein coupled receptor 17 (GPR17 [26, 27]). The enzyme 5-Lox is also involved in some steps of biosynthesis of specialized pro-resolving mediators from omega 3 poly-unsaturated fatty acids (DHA – docosahexaenoic acid; EPA – eicosapentaenoic acid)

Fig. 2

Immunohistochemical analysis for 5-Lox and FLAP expression in human AD brain. a 5-Lox immunoreactivity (using monoclonal 5-Lox antibody from BD Biosciences #610694) in human post-mortem AD (high, Braak 6) hippocampus specimen. 5-Lox staining was prominent in cells of the granular cell layer in the dentate gyrus. b FLAP was expressed in cells with glial cell morphology but not in neurons. c Double staining of 5-Lox with FLAP revealed only minor co-expression of 5-Lox in FLAP positive cells (arrows). Asterisks indicate double positive cells. d Double staining of 5-Lox with microglia marker Iba1 showed that non-neuronal 5-Lox expression is associated to some microglia cells (asterisk). Arrows indicate microglia without 5-Lox expression. e Percentage of Iba1⁺/5-Lox⁺ and Iba1⁺/5-Lox⁻ cells in Iba1⁺ cells from control and AD (f) hippocampus sections. Hematoxylin (a, b) was used as nucleus stain. Scale: 20 μm (a-c, d right image), 50 μm (d, left image)

Fig. 3

Immunohistochemical analysis for 5-Lox and FLAP expression in WT and APP-PS1 mouse brains. a In mouse brains the majority of 5-Lox (green, monoclonal 5-Lox antibody from Abcam #ab169755) immunoreactivity was observed in the granular layer of the dentate gyrus and co-localized with NeuN (white). b 5-Lox immunoreactivity (green, using polyclonal 5-Lox antibody from Abcam #ab39347) was also observed outside the granular layer in particular in APP-PS1 mice and co-localized with some Iba1 positive microglia (white) cells of WT and APP-PS1 transgenic mice (inserts). c Besides co-expression of 5-Lox (green, using polyclonal 5-Lox antibody from Abcam #ab39347) in some microglia cells, also GFAP (red) positive astrocytes showed immunoreactivity for 5-Lox in WT and APP-PS1 mice (inserts). d FLAP (green, polyclonal FLAP antibody from Novus Biologicals #NBP1–84666) was ubiquitously expressed in Iba1 positive microglia cells (white) in WT and APP-PS1 transgenic mice. Dapi (a-d) was used as nucleus stain. Scale: 20 μm (b-d right images), 50 μm (a-d left images)

Fig. 4

Experimental design of microglia ablation via CSFR1 inhibition. a 12 month old mice were treated with the CSF1R inhibitor PLX5622 for a total of 28 days. Treatment was applied in form of control chow or PLX5622 containing chow. b WT animals as well as APP-PS1 animals treated with PLX5622 showed a reduction of microglia cells (Iba1⁺ and TMEM119⁺) in the hippocampus compared to the respective control groups. Dapi was used as nucleus stain. Scale: 50 μm (b)

Fig. 5

Hippocampal transcriptome analysis revealed significantly downregulated microglia genes and downregulated LT signaling related genes in PLX5622 treated mice. a Volcano blots of WT + PLX5622 vs. WT Control and APP-PS1 + PLX5622 vs. APP-PS1 Control (b) comparisons illustrating representative microglia genes (Mrc1, TMEM119, Cd33, Aif1) that were significantly downregulated (adjusted p-value < 0.05) upon microglia cell ablation. Surprisingly, genes involved in the LT signaling pathway (Alox5ap, Alox5, Cysltr1, Syk, Fcgr1, Tlr4, Ncf1, Rgs10) were significantly downregulated in the hippocampus of WT and APP-PS1 mice upon PLX5622 treatment. Blue spots indicate microglia genes; red spots are LTs related genes and grey spots represent all other genes

Fig. 6

qPCR validation of hippocampal mRNA expression for LT synthesis related genes: a Microglia ablation in WT and APP-PS1 mice resulted in significantly lower mRNA expression of Alox5ap and Alox5. b Expression of cysteinyl-LT receptor CysLTR1 was significantly decreased upon microglia ablation in WT and APP-PS1 mice. One-way analysis of variance with Bonferroni’s multiple comparison test was used. P-values < 0.05 were considered significant. Data are shown as mean with SEM

Fig. 7

Immunohistochemical analysis of FLAP expression in the hippocampus. a Overall FLAP (green, polyclonal FLAP antibody from Novus Biologicals NBP1–84666) expression was more prominent in APP-PS1 compared to WT control animals and highly reduced with PLX5622 treatment. b FLAP co-localized with Iba1 (white) positive cells in all groups (arrows). Most interestingly, the intensity of FLAP staining in microglia at sites of amyloid plaques was reduced (red asterisk). Dapi was used as nucleus stain. Scale: 50 μm (a), 20 μm (b)

Fig. 8

Immunohistochemical analysis of 5-Lox expression in the hippocampus. a 5-Lox (red, monoclonal 5-Lox antibody from Abcam #ab169755) was primarily expressed in neurons of the granular cell layer but did also co-localize with some Iba1 (white) positive cells (arrow or inserts). b We analyzed the percentage (%) of 5-Lox⁺ cells in the Iba1⁺ cell population revealing increased numbers of 5-Lox + microglia in APP-PS1 mice compared to WT animals. 5-Lox⁺ microglia numbers were reduced in APP-PS1 + PLX5622 animals. c The overall 5-Lox staining (% area stained by 5-Lox) was increased in APP-PS1 mice compared to WT animals and reduced upon microglia depletion in WT + PLX5622 mice. Remaining microglia in PLX5622 treated mice displayed altered cell morphology. One-way analysis of variance with Tukey’s multiple comparison test (b, c) and unpaired t test comparing WT vs. WT + PLX5622 (c) was used. P-values < 0.05 were considered significant. Data are shown as mean with SD. Dapi was used as nucleus stain. Scale: 100 μm and 20 μm inserts (a)

Fig. 9

3D Render of images with Imaris Software. 5-Lox immunoreactivity (red, using polyclonal 5-Lox antibody from Abcam #ab39347) is shown in NeuN positive neurons (green) of the granular cell layer in the dentate gyrus and in some Iba1 positive microglia cells (white). In microglia cells 5-Lox staining clustered at the nuclear membrane indicating an activated form of the 5-Lox enzyme (arrow). Scale: 20 μm and 5 μm inserts

Fig. 10

3D Render of images with Imaris Software. 5-Lox (red, using polyclonal 5-Lox antibody from Abcam #ab39347) is expressed in FLAP (green, polyclonal FLAP antibody from Novus Biologicals NB300–891) positive cells associated with the amyloid plaque. 5-Lox and FLAP are in close contact at sites of the nuclear membrane. Dapi was used as nucleus stain. Scale: 5 μm, 1 μm (insert left), 2 μm (insert right)

Fig. 11

Hypothetical model for microglia as source of LTs in the brain. Key elements of the LTs synthesis pathway are FLAP and 5-Lox, which are co-expressed in some microglia cells (green). Additionally, 5-Lox staining was observed in neurons (blue). It is known that 5-Lox and FLAP protein need to form an intracellular complex to start the synthesis pathway resulting in the production of LTs (left panel). With the herein shown microglia ablation experiment (+PLX5622), we could confirm microglia cells as potential source for LTs. In microglia ablated brains (right panel) FLAP and 5-Lox expression as well as CysLTR1 was highly decreased. Additionally, our data supports the idea of a putative microglia neuron interaction, as 5-Lox immunoreactivity was reduced in neurons of microglia ablated WT brains. LTs thereby possibly affect microglia in an autocrine manner via the CysLTR1, and neurons via GPR17. We hypothesize a putative feedback loop between neurons and microglia involving LTs signaling in the context of neuroinflammation. Image was created with BioRender.com