Microglial CX3CR1 promotes adult neurogenesis by inhibiting Sirt 1/p65 signaling independent of CX3CL1

Homo and heterozygote cx3cr1 mutant mice, which harbor a green fluorescent protein (EGFP) in their cx3cr1 loci, represent a widely used animal model to study microglia and peripheral myeloid cells. Here we report that microglia in the dentate gyrus (DG) of cx3cr1−/− mice displayed elevated microglial sirtuin 1 (SIRT1) expression levels and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) p65 activation, despite unaltered morphology when compared to cx3cr1+/− or cx3cr1+/+ controls. This phenotype was restricted to the DG and accompanied by reduced adult neurogenesis in cx3cr1−/− mice. Remarkably, adult neurogenesis was not affected by the lack of the CX3CR1-ligand, fractalkine (CX3CL1). Mechanistically, pharmacological activation of SIRT1 improved adult neurogenesis in the DG together with an enhanced performance of cx3cr1−/− mice in a hippocampus-dependent learning and memory task. The reverse condition was induced when SIRT1 was inhibited in cx3cr1−/− mice, causing reduced adult neurogenesis and lowered hippocampal cognitive abilities. In conclusion, our data indicate that deletion of CX3CR1 from microglia under resting conditions modifies brain areas with elevated cellular turnover independent of CX3CL1.

Introduction CX 3 CR1 is a seven transmembrane domain receptor coupled to G i and G z subtypes of G proteins, activation of which is linked to several intracellular second messengers like phosphoinositide 3-kinase (PI3K), protein kinase B (AKT) and nuclear factor kappa-light-chain-enhancer of activated B cells (NFkB) [1]. CX 3 CR1 is prominently expressed by monocytes, subsets of NK and dendritic cells, and the brain microglia [2]. Surprisingly little is known on which intracellular signaling pathways in microglia are affected by the lack of CX 3 CR1. Microglial cells are fundamentally distinct from other brain cells in that they are derived from primitive myeloid progenitors that arise during embryogenesis [3][4][5]. They represent the resident phagocytic cells in the brain, taking part in immune-mediated defense mechanisms and clearing cell debris [6]. Microglial cells are constantly surveying their surroundings and are implicated in synaptic pruning, during both development and throughout adulthood, and therefore believed to play a role in regulating homeostatic neuronal synaptic plasticity [7,8]. Neurons and microglia are thought to communicate with one another through neuronal expression of the CX 3 CR1 ligand CX 3 CL1 (or fractalkine). CX 3 CL1 is expressed at the cell membrane of selected neurons and binds to and activates CX 3 CR1 receptors on microglia [9,10]. CX 3 CL1 exists in two distinct forms: a full-length membrane-bound form and a shed form that contains the N-terminal chemokine domain. The shed chemokine domain of CX 3 CL1 acts, when cleaved, as a signaling molecule and can bind microglial-expressed CX 3 CR1 receptors [11], whereas its membrane-bound mucin stalk can serve as a cell adhesion molecule [12]. CX 3 CL1 is abundantly expressed in the granular cell layer of the rat dentate gyrus (DG), where addition of recombinant CX 3 CL1 reverses age-related decline in adult neurogenesis [13]. Both, cx3cr1 −/− and cx3cr1 +/− mice display reduced hippocampal neurogenesis compared with wild-type controls [13,14]. However, it is not clear to what extent CX 3 CL1 is mandatory for proliferation and neurogenesis. In addition to the already mentioned reduced hippocampal neurogenesis, cx3cr1 −/− mice were reported to exhibit excessive hippocampal IL-1β expression and either enhanced [15] or attenuated long term potentiation (LTP) [14] resulting in improved [15] or impaired cognitive functions [13,14]. It is important to mention that Maggi et al. were using exclusively female mice (3 months of age). Rogers et al. and Bachstetter et al. performed all experiments with male mice, 3 months and 4 months of age, respectively. In addition to these conflicting results, little is known about the intracellular signaling cascades activated by CX 3 CR1 deficiency, which might impact synaptic plasticity and cognition. One of these pathways could include the NF-kB signaling pathway, which may trigger microglial activation and induce the release of inflammatory factors including IL-1β, as seen following irradiation [16], during normal aging [17] or neurodegeneration [18]. Along these lines, sirtuin 1 (SIRT1), a member of the sirtuin family, might be modulated in microglia by the lack of CX 3 CR1 because it can suppress inflammatory responses by inhibiting NF-kB signaling [19,20].
Here we investigated the consequences of the microglial CX 3 CR1 deletion on cell morphology, activation of the NF-kB signaling pathway, the expression of SIRT1, interference with neuroblasts, immature neurons and cognition. Our findings indicate that, under brain homeostasis, hippocampal microglia from cx3cr1 −/− mice are very similar, if not identical, to microglia from adult wild-type animals in contrast to the situation in newborns or during development. However, hippocampal cx3cr1 −/− microglia show activation of SIRT1 and the NF-kB pathway in areas of adult neurogenesis. We found that manipulation of SIRT1 activation in cx3cr1 −/− mice directly impacts cognitive performance, while the same treatment had no detectable effect on cognition in wild-type littermates.

Materials and methods
This article does not contain any studies with human participants performed by any of the authors.

Mice
For all animal experiments "Principles of laboratory animal care" (NIH publication No. 86-23, revised 1985) were followed. All experiments were approved by the Federal Ministry for Nature, Environment and Consumers' Protection of the states of Baden-Württemberg (G12/71 and G11/50), and were carried out in accordance to the respective national, federal, and institutional regulations. Adult, 8-12 weeks old, male mice were used for the experiments. Mice were group housed up to five per cage with 12 h light/dark cycle with lights on at 6 a.m. Food and water were available ad libitum. cx3cr1 gfp/gfp on a C57BL/6 J background were obtained from the Jackson Laboratory. cx3cl1 −/− mice were described previously [21]. In all experiments, wild type littermates were used as controls.

Morris Water Maze (MWM) test
The MWM was used to measure spatial learning and memory [22]. The apparatus consisted of a black plastic pool with a diameter of 120 cm. A black escape-platform (square, 10 × 10 cm) was located 1.0 cm below (hidden) the water surface. The temperature of the water was kept constant throughout the experiment (20 ± 0.5°C), and a 10 min recovery period was allowed between the training trials. The training consisted of 6 consecutive days of testing, with four trials per day. If the mouse failed to find the escape platform within the maximum time (60 s), the animal was placed on the platform for 10 s by the experimenter. During the first 6 days of testing, the mice were trained with a hidden platform. The platform location was kept constant, and the starting position varied between four constant locations at the pool rim. Mice were placed in the water with their nose pointing toward the wall at one of the starting points in a random manner. On the 7 th day, the platform was removed, and the mice were allowed to swim for 60 s to determine their search bias. On testing day 8, mice were trained to find a visible platform, which had a 10 cm high pole with a white flag and was changed every trial to a new position. Timing of the latency to find the visible platform was started and ended by the experimenter. A computer running the BIOBSERVE software (BIOBSERVE) analyzed all variables of the MWM test. All behavioral experiments were carried out in a doubleblind fashion and mice were tested in random order.

Cell quantification
Cells were counted using the optical fractionator, a method for unbiased stereological analysis. This method was performed using a computer-assisted image analysis system, consisting of a Leica DMRB/DIC fluorescence microscope equipped with a computer-controlled motorized stage, a video camera, and the Stereo Investigator software (MicroBrightField, Williston, VT). The number of positive cells throughout the rostrocaudal extent of the dentate gyrus was counted with a coded one-in-16 series for frozen sections (40 μm). We used a modified version of the optical fractionator method as reported previously [35][36][37]. The total numbers of positive cells were multiplied by 16 and reported as total number of cells per dentate gyrus. For immunocytochemical analysis of paraffin sections (30 μm), serial coronal sections were collected spanning the rostrocaudal extent of the hippocampus. For quantification, every 12 th section was selected. Every positive cell was counted on these sections, and, to obtain the relative total number of cells in the dentate gyrus, these counts were multiplied by 12 to account for the sampling frequency [25]. In control experiments we could not detect significant differences in cell numbers, when samples were quantified by both quantitation methods.

Laser microdissection
Microdissection of microglia was performed using a Zeiss PALM MicroBeam as described previously, with modification [29]. Fast immunochemistry of serial sections was performed with CD11b antibodies (Serotec). Immunostained sections were counterstained with DAPI to facilitate the identification of individual cells. RNA was isolated with the Rneasy Micro Plus Kit (Qiagen), and reverse transcription (RT), preamplification, and real-time PCR were performed using Applied Biosystems reagents according to the manufacturer's recommendations. The primer pairs were used as described previously [29].

Microglia isolation and flow cytometry
Adult microglia cells were harvested from dissected hippocampal tissue using density gradient separation and were prepared as described before [27]. In short, samples were stained for CD11b and CD45 (eBioscience, BD Pharmingen) [4]. Cell suspensions were acquired on a FACS Canto II (Becton Dickinson) or cell populations were sorted with a MoFlo Astrios (Beckman Coulter) and further processed. Data were analyzed using FlowJo software (Tree Star).
Mice underwent diffusion tensor imaging (DTI) examination at 156 2 × 250 μm 3 spatial resolution with a cryogenic cooled resonator (CCR) at ultrahigh field (7 T) as described previously [39,40]. Diffusion images were acquired along 30 gradient directions plus 5 references without diffusion encoding with a total acquisition time of 35 min. Fractional anisotropy (FA) maps were statistically compared by whole brain-based spatial statistics (WBSS) at the group level vs. wt controls.

Sirtuin 1 activity assay
To quantify sirtuin 1 (Sirt1) activity, nuclear extracts from sorted microglia (pooled from hippocampi of three mice per group) were prepared. Nuclear extracts were used to measure deacetylase activity of an acetylated histone using Epigenase Universal SIRT Activity/Inhibition Assay Kit (Epigentek, Farmingdale, NY) [41].

Statistical analysis
Statistical analysis was performed using GraphPad Prism (GraphPad Software, Version 6.0, La Jolla, USA). In general, chosen sample sizes are similar to those reported in previous publications [38]. All data were tested for normality applying the Kolmogorov-Smirnov test. If normality was given, an unpaired t test was applied. If the data did not meet the criteria of normality, the Mann-Whitney U test was applied. To test for effects of treatment or genotype a two-factorial analysis of variance (ANOVA) with post hoc Bonferroni test or Tukey-Kramer HSD test was applied. Differences were considered significant when p < 0.05. Number of animals per group is given as "n". To obtain unbiased data, experimental mice were all processed together by technicians and cell quantifications were performed blinded by two scientists independently and separately.

Discussion
With the present study we show that, under resting conditions, SIRT1 and the NF-kB pathway are activated in cx3cr1 −/− microglia residing within the murine DG area. This activation seems to be restricted to the DG and was largely diminished in the hippocampal CA1 region. As a result of pharmacological SIRT1 activation, impaired adult neurogenesis and lowered hippocampal cognitive performance was restored in cx3cr1 −/− mice.
We hypothesize that the NF-kB signaling pathway and SIRT1 enzyme in microglia interact to maintain cellular homeostasis in vivo. Since the fractalkine receptor CX 3 CR1 inhibits cAMP signaling including the cAMPdependent protein kinase A (PKA) via coupling to a G iprotein coupled receptor [43], deletion of CX 3 CR1 from microglia might facilitate activation of PKA and subsequently NF-kB activation [49]. Stimuli causing PKA activation appear to be restricted to certain brain regions with e.g. enhanced cellular turnover like the DG because only marginal acp65 signals were detected outside the DG as seen in the CA1 area. In response to increased acetylation of p65 in cx3cr1 −/− microglia, SIRT1 activity is amplified, most likely to counteract excessive NF-kB signaling. Activation of SIRT1 can induce deacetylation of the RelA/p65 component of the NF-kB complex. The deacetylation of Lys310 inhibits the transactivation capacity of RelA/p65 subunit and consequently suppresses the transcription of the NF-kB-dependent gene expression [20]. However, in cx3cr1 −/− microglia SIRT1 activity, although elevated, appears to be insufficient to prevent NF-kB-dependent gene expression [50] as indicated by elevated protein levels of IL-1β in the hippocampus [13] or by increased CXCL10, TNF-α and IL-1β mRNA expression in microglia micro-dissected from DG of cx3cr1 −/− mice (Fig. 4). Only additional SIRT1 activation can effectively counteract activation of the NF-kB pathway. Interestingly, in wild-type mice where no microglial NF-kB activation was detectable, activation of SIRT1 had no effect on adult neurogenesis or performance in the water maze test (data not shown). Previous studies have shown that especially IL-1β can impair adult neurogenesis by decreasing proliferation in the DG of cx3cr1 −/− mice [13]. A similar mechanism was observed in wild-type mice under stressed conditions when elevated IL-1β reduced hippocampal neurogenesis and administration of an IL-1 receptor antagonist restored the neurogenesis rate following stress exposure [51]. Notably, despite the pro-inflammatory activation status of cx3cr1 −/− microglia, no morphological changes could be detected in comparison to cx3cr1 +/+ microglia, confirming previous findings within an independent set of experiments [52]. During development in the hippocampus, microglial morphology seems to depend, at least partially, on the fractalkine receptor considering that at P8 a small population of cx3cr1 −/+ cells was typified by a very large surface area. This group of cells was absent in brains of cx3cr1 −/− mice [53]. However, one report based on immunohistochemistry indicates that under normal physiological conditions genetic cx3cr1 deficiency is associated with microglial alterations, including increased cell number and enlargement of the soma [54]. While the receptor for CX 3 CL1, CX 3 CR1, is highly expressed on microglia, CX 3 CL1 is constitutively expressed at high levels on healthy neurons. CX 3 CL1 is expressed as a transmembrane protein that can be cleaved in a soluble form, consisting of the extracellular N-terminal chemokine domain, by the activity of the lysosomal cysteine protease cathepsin S (CatS) or by members of the disintegrin and metalloproteinase (ADAM) family, ADAM-10 and ADAM-17 [55]. The anatomical expression of CX 3 CL1 on neurons and CX 3 CR1 on microglia suggests that neurons may maintain microglia in a surveilling/ ramified state through a repressive CX 3 CL1 signal [56]. From this point of view it is unexpected to find intact proliferation of neural stem/progenitor cells and adult neurogenesis in CX 3 CL1-deficient mice while the same process is impaired in mice lacking CX 3 CR1. If binding of CX 3 CL1 elicits a tonic inhibitory signal, which maintains microglia in a quiescent state, its absence should result in activated or pro-inflammatory microglia with negative impact on neurogenesis and neurodevelopment. While there is a consensus among studies for a neuroprotective role of CX 3 CL1 signaling in vitro, some in vivo studies even suggest a neurotoxic role of CX 3 CL1 as seen in animal models for Alzheimer's-and Parkinson's disease. Here, CX 3 CL1 can act as a repressor of microglial phagocytic activity and cause overall microglia activation [57,58]. A further remarkable mode of action was observed in lung endothelial cells, which respond to stimuli and produce CX 3 CL1. This leads to the endothelial attachment of the subset mononuclear leukocytes that express the sole CX 3 CL1 receptor, CX 3 CR1 [59]. Following a challenge with lipopolysaccharides (LPS), cx3cl1 −/− mice exhibit reduced expression of CX 3 CR1 and impaired NF-kB signaling in lung tissue when compared to wt controls [60]. CX 3 CL1 is a molecule that may have various activities, with either no, beneficial or destructive potential, likely depending on the activation state of its main target cells. In support of our findings cx3cl1 −/− mice do not have histologic abnormalities in any major organs (including the brain), hematopoietic lineages in blood and lymphoid tissue are essentially normal, and they do not exhibit any overt behavioral abnormalities [21]. There is also the possibility of an alternative CX 3 CR1 ligand, which might act similar to CX 3 CL1. In humans eotaxin-3/CC chemokine ligand 26 was recently reported to be a functional ligand for CX 3 CR1 [61]; in mice, the CCL26 gene, however, may be a pseudogene since no cDNA or expressed sequence tag (EST) has been reported [62]. Similar to previous reports we found that hippocampal neurogenesis was decreased in mice that lack CX 3 CR1 [15]. These mice display significant deficits in cognitive functions and LTP induction due to increased action of IL-1β [14]. There are two reports indicating that cx3cr1 −/− mice have improved hippocampal cognitive abilities compared to wild-type controls [15,54] with either enhanced [54] or impaired [15] generation of neuronal precursors. The reason for this inconsistency is presently unclear.

Conclusions
Our findings indicate that in the SGZ of the DG area in cx3cr1 −/− mice the number of DCX + cells is reduced, independent of the CX 3 CR1 ligand CX 3 CL1. Enhanced microglial NF-kB-dependent gene expression in the DG results in elevated levels of chemokines such as IL-1β and consequently in the inhibition of neurogenesis and spatial cognitive function. Manipulation of SIRT1 activity interferes with NF-kB signaling, adult neurogenesis and ultimately hippocampal learning and memory.