NLRP3 maintains healthy pericytes in the brain CURRENT STATUS: UNDER REVIEW

Background Growing evidence shows that pericytes are damaged in the brain of Alzheimer’s disease (AD), which potentially contributes to AD pathogenesis. NLRP3-contained inflammasome is activated in AD brain and considered as a promising target for therapy. However, how NLRP3 affects brain pericytes is unclear. In our study, we investigated physiological function of NLRP3 in pericytes. Methods Immunohistological methods and Western blot were used to investigate pericytes and vasculature in the brains of 9-month-old NLRP3-deficient and wild-type littermate mice. Pericytes were also cultured and treated with NLRP3 inhibitor, recombinant IL-1β and AKT inhibitor. Then, proliferation, apoptosis and expression of PDGFRβ and CD13 in pericytes were analysed with biochemical methods. To investigate underlying molecular mechanisms, phosphorylation of protein kinases such as AKT, ERK and NF- k B were quantified. Results We observed that NLRP3 deficiency reduced the coverage of PDGFRβ-positive pericytes and collagen type IV-immunereactive vasculature in the brain. NLRP3 deficiency was also shown to decrease PDGFRβ and CD13 proteins in isolated cerebral microvessels. In cultured pericytes, inhibition of NLRP3 with MCC950 attenuated cell proliferation but did not induce apoptosis. NLRP3 inhibition also decreased protein levels of PDGFRβ and CD13. On the contrary, treatments with IL-1β increased protein levels of PDGFRβ and CD13 in pericytes. The alteration of PDGFRβ and CD13 protein levels was correlated with phosphorylation of AKT. Inhibition of AKT reduced PDGFRβ and CD13 in cultured pericytes. Conclusions NLRP3 might be essential to maintain healthy pericytes in the brain

3 through activating AKT. Adverse effects on brain pericytes should be considered in the possible clinical therapies with NLRP3 inhibitors. Background Brain pericytes wrapping around endothelial cells regulate various functions in the brain, which include blood-brain barrier (BBB) permeability, angiogenesis, and capillary hemodynamic responses [1]. Pericytes express platelet-derived growth factor receptor β (PDGFRβ) and CD13. The binding of PDGFRβ with endothelial cellsreleased platelet-derived growth factor (PDGF)-B is essential for pericyte proliferation and integration into the blood vessel [2]. CD13 promotes angiogenesis in hypoxic tissues, and response to stimulation of angiogenic growth factors, such as vascular endothelial growth factor, basis fibroblast growth factor, and transforming growth factor [3]. Deficiency of PDGFRβ decreases pericyte number, accumulates blood-derived fibrin/fibrinogen, reduces vasculature and attenuates blood flow in the mouse brain, which finally leads to the white matter lesions characterized by loss of oligodendrocytes, demyelination, and axonal degeneration [4]. Growing evidence suggests that pericyte impairment mediates vascular dysfunction and contributes to pathogenesis of Alzheimer's disease (AD) [5]. In AD human brain, pericytes are lost in association with increased BBB permeability at very early disease stage [6,7]. In AD mouse models that overexpress Alzheimer's precursor protein (APP) in neurons, deletion of pericytes increases deposition of amyloid β peptide (Aβ) in both brain parenchyma and blood vessels, which potentially exaggerates cognitive deficits [8]. However, molecular mechanisms that regulate pericyte survival and activation in the brain are largely unknown.
NLRP3 is considered as a promising therapeutic target for AD patients [17].
However, effects of NLRP3 activation on pericytes and vascular dysfunction were not addressed. The animal models used in published studies have also limited ADassociated vascular pathology. Thus, whether NLRP3 inhibition protects or damages microvascular circulation in AD brain remains unclear. Moreover, AD pathology is mainly localized in temporal and parietal lobes, instead of covering the whole brain.
Between AD lesion sites as shown with Aβ deposits, neurofibrillary tangles and gliosis, the brain tissues are relatively or absolutely healthy [18]. It is, therefore, necessary to understand the physiological functions of NLRP3 in brain pericytes, which is helpful to predict potential off-target effects of NLRP3 inhibitors in the future anti-AD therapies.
In this study, we used NLRP3-knockout mice and treated cultured pericytes with NLRP3 inhibitor, MCC950, or IL-1β, a major product of NLRP3-contained inflammasome. We observed that NLRP3 might be essential for the maintenance of 5 healthy pericytes in the brain. We further observed that AKT (also known as protein kinase B) might mediate the physiological function of NLRP3 in pericytes.

Materials and Methods
Mice NLRP3-encoding gene knockout (NLRP3 -/-) mice were kindly provided by N. Fasel (University of Lausanne, Lausanne, Switzerland) [19]. Mice with knockout of gene encoding myeloid differentiation primary response 88 (MyD88 -/-) were originally provided by S. Akira and K. Takeda (Osaka University, Osaka, Japan) [20]. Breeding between heterozygous mutants (+/-) on a C57BL/6 background were used to maintain mouse colonies. Mice were compared only between littermates. Animal experiments were performed in accordance with all relevant national rules and were authorized by the local research ethical committee.

Tissue collection and isolation of blood vessels
Animals were euthanized by inhalation of isofluorane and perfused with ice-cold phosphate-buffered saline. The brain was removed and divided. The left hemisphere was immediately fixed in 4% paraformaldehyde (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) in phosphate-buffered saline and embedded in paraffin for immunohistochemistry. The cortex and hippocampus were carefully dissected from the right hemisphere, snap-frozen in liquid nitrogen and stored at -80°C for biochemical analysis. Cortex and hippocampus were also used for isolation of brain vessel fragments according to the published protocol [21]. Briefly, brain tissues were homogenized in HEPES-contained Hanks' balanced salt solution (HBSS) and centrifuged at 4,400 g in HEPES-HBSS buffer supplemented with dextran from Leuconostoc spp. (molecular weight ~ 70,000; Sigma-Aldrich) to delete myelin. The vessel pellet was re-suspended in HEPES-HBSS buffer supplemented with 1% bovine serum albumin (Sigma-Aldrich) and filtered with 20 µm -mesh. The blood vessel fragments were collected on the top of filter and frozen at -80°C for further biochemical analysis. The grid size was set at 10 µm, which provided coefficient of error estimates of < 0.05.

Histological image acquisition and analysis
To quantify vasculature in the brain, our established protocol was used [22]. Briefly, 4 serial paraffin-embedded sections per mouse were deparaffinized, heated at 80°C in citrate buffer (10mM, pH = 6) for 1 hour and digested with Digest-All 3 (Pepsin) (Thermo Fisher Scientific) for 20 minutes. Thereafter, brain sections were stained with rabbit anti-collagen IV polyclonal antibody (Catalog: # ab6586; Abcam, Cambridge, UK) and Alexa488-conjugated goat anti-rabbit IgG (Thermo Fisher Scientific). After being mounted, the whole brain including hippocampus and cortex was imaged with Microlucida (MBF Bioscience). The length and branching points of collagen type IV staining-positive blood vessels were analyzed with a free software, 7 AngioTool (http://angiotool.nci.nih.gov) [23]. The parameters of analysis for all compared samples were kept constant. The length and branching points were adjusted with area of interest.
Densitometric analysis of band densities was performed with ImageJ software (https://imagej.nih.gov/ij/). For each sample, the protein level was calculated as a ratio of target protein/β-actin.

Culture of pericytes
Human primary brain vascular pericytes (HBPC) were immortalized by infecting cells with tsSV40T lentiviral particles [24]. The selected immortalized HBPC clone 37 (hereafter referred to as HBPC/ci37) was used for our study. HBPC/ci37 cells were

Treatments of pericytes for detection of PDGFRβ and CD13 and phosphorylated kinases
Pericytes were cultured in 12-well plate at 5.0 × 10 5 cells/well. Before experiments, we replaced culture medium with serum-free pericyte medium and cultured cells at 37°C for 3 days to facilitate cell differentiation [24]. Thereafter, pericytes were

Statistics
Data was presented as mean ± SEM for mice and mean ± SD for cells. For multiple comparisons, one-way or two-way ANOVA followed by Bonferroni or Tukey post hoc test. Two independent-samples Students t test was used to compare means for two groups of cases. All statistical analyses were performed with GraphPad Prism 5 version 5.01 for Windows (GraphPad Software, San Diego, CA, USA). Statistical significance was set at p < 0.05.

NLRP3 deficiency reduces pericyte cell coverage and decreases protein levels of PDGFRβ and CD13 in the brain.
To explore effects of NLRP3 on the maintenance of pericytes in the brain, we estimated the coverage of PDGFRβ-positive cells in brains from 9-month-old NLRP3-knockout (NLRP3 -/-) and wild-type (NLRP3 + /+ ) littermate mice. As shown in Fig. 1, A and B, deficiency of NLRP3 significantly decreases the coverage of pericytes (t test, p < 0.01; n = 10 per group). We then isolated blood vessels from brains of 9-monthold NLRP3 -/-, NLRP3 + /and NLRP3 + /+ littermate mice for the detection of pericyte protein markers, PDGFRβ and CD13. We observed that deletion of NLRP3 significantly reduces PDGFRβ and CD13 proteins in the cerebral blood vessels in a gene dose-dependent manner (Fig. 1, C and D; one-way ANOVA, p < 0.05; n ³ 7 per group). Unfortunately, we failed to detect cleaved caspase-3 and PCNA in the tissue lysate of blood vessels (data not shown), which are markers for cell apoptosis and cell proliferation, respectively.
In further experiments, we asked whether innate immune signaling serves a common effect on pericyte survival in the brain. We detected PDGFRβ and CD13 proteins in cerebral blood vessels isolated from 6-month-old MyD88 -/-, MyD88 + /and MyD88 + /+ littermate mice. MyD88 is a common adaptor down-stream to most TLRs and also mediates the inflammatory activation of IL-1β [25]. We observed that protein levels of CD13 and PDGFRβ were significantly lower in MyD88-deficient mice than in MyD88-wildtype controls (Fig. 1, E and F; one-way ANOVA, p < 0.05; n ³ 6 per group).

NLRP3 deficiency reduces vasculature in the brain.
Pericytes are essential for the development of cerebral circulation. Dysfunction of pericytes reduces vasculature [1,4]. We asked whether NLRP3 deficiency affects the structure of cerebral blood vessels. We observed that, in 9-month-old mouse brains, deficiency of NLRP3 significantly reduced the total length and branching pints of collagen type IV-positive blood vessels (Fig. 2, A -C; one-way ANOVA, p < 0.05; n ³ 6 per group). The reduction of brain vasculature was dependent on the copies of NLRP3-encoding gene.

NLRP3 inhibition attenuates cell proliferation in cultured pericytes.
After we observed that NLRP3 deficiency decreased the number of pericytes in the brain, we continued to investigate underlying mechanisms mediating the effects of NLRP3 on pericytes. After treating cultured pericytes with NLRP3 inhibitor, MCC950, at different concentrations, we observed that NLRP3 inhibition significantly reduced the conversion of MTT into its colorful product in a dose-dependent manner (Fig. 3, A; one-way ANOVA, p < 0.05; n = 4 per group). In further experiments, we detected no cleavage of caspase-3 in MCC950-treated cells (Fig. 3, B), while MCC950 treatments significantly decreased protein levels of both PCNA (Fig. 3, C and D; oneway ANOVA, p < 0.05; n = 3 per group) and Ki-67 (Fig. 3, E and F; one-way ANOVA, p < 0.05; n = 4 per group), which are two typical protein markers for cell proliferation. Thus, inhibition of NLRP3 potentially suppressed proliferation of pericytes.

NLRP3 inhibition attenuates expression of PDGFRβ and CD13 in cultured pericytes.
PDGFRβ and CD13 are two protein markers of pericytes in the brain, which mediate physiological and pathophysiological functions of pericytes [2,3]. We observed that treatments with MCC950 inhibited expression of PDGFRβ and CD13 in pericytes in a dose-dependent manner (Fig. 4, A; One-way ANOVA, p < 0.05; n = 4 per group). In order to analyze underlying mechanisms, through which NLRP3 drives pericyte differentiation, we detected phosphorylation of AKT, ERK and NF-kB in MCC950treated cells. Activation of AKT and ERK is involved in pericyte proliferation and migration [26,27]. We observed that inhibition of NLRP3 reduced the protein levels of both phosphorylated AKT and ERK in a dose-dependent manner (Fig. 4, A; Oneway ANOVA, p < 0.05; n = 4 per group). However, phosphorylation of NF-kB in pericytes was not significantly altered by treatments with MCC950 (Fig. 4, A; Oneway ANOVA, p = 0.094; n = 3 per group).

IL-1β increases expression of PDGFRβ and CD13 in cultured pericytes.
NLRP3-contained inflammasome activates caspase 1 and subsequently cleaves pro-IL-1β into active IL-1β [28]. Due to the low level of IL-1β released from nonactivated pericytes, we could not detect reduction of IL-1β secreted from NLRP3deficient pericytes compared to NLRP3-wild-type pericytes (data not shown).
However, we hypothesized that IL-1β affects differentiation of pericytes. We treated cultured pericytes with IL-1β at different concentrations. Very interestingly, IL-1β increases the protein expression of PDGFRβ and CD13 also with a concentrationsdependent pattern (Fig. 5, A-C; One-way ANOVA, p < 0.05; n = 3 per group). As potential mechanisms mediating effects of IL-1β activation, we observed that IL-1β treatments significantly increased phosphorylation of AKT but of ERK (Fig. 5, D-F; One-way ANOVA, p < 0.05; n = 3 per group).

Inhibition of AKT suppresses expression of PDGFRβ and CD13 in cultured pericytes.
As activation of AKT in pericytes was suppressed by NLRP3 inhibition but enhanced upon IL-1β activation, we supposed that AKT signaling plays an important role in the 13 differentiation of pericytes. We treated pericytes with AKT inhibitors at 0, 0.5, 1 and 5 µM. Phosphorylation of AKT and phosphorylation of GSK3β, a kinase down-stream to AKT, were both reduced (Fig. 6, A-C; One-way ANOVA, p < 0.05; n = 4 per group), which verified the successful inhibition of AKT signaling. With such an inhibition, expression of both PDGFRβ and CD13 was significantly down-regulated in a dosedependent manner (Fig. 6, D-F; One-way ANOVA, p < 0.05; n = 3 per group). discussion Pericytes play a central role in regulating microvascular circulation and BBB function in the brain [1]. Our study demonstrated that deletion of NLRP3 under physiological conditions decreases the coverage of pericytes and protein levels of PDGFRβ and CD13 in cerebral blood vessels. PDGFRβ and CD13, together with neural/glial antigen-2 and CD146 are expressed in capillary-associated pericytes, and often used as protein markers of pericytes [29]. PDGFRβ and CD13 also trigger proliferation and migration of pericytes after stimulation with angiogenesisassociated growth factors [2,3]. Thus, the reduction of CD13 and PDGFRβ represents not only loss of pericytes but also dysfunction of pericytes in the brain.
Indeed, we observed that deletion of NLRP3 reduces the vasculature in the brain, which corroborates a recent observation that dysfunction of pericytes decreases the length of cerebral blood vessels in PDGFRβ-mutated mouse brain [4]. Our study suggests that NLRP3 is essential in the maintenance of functional pericytes in healthy brains.
We observed that deficiency of either NLRP3 or MyD88 decreases protein levels of PDGFRβ and CD13 in the mouse brain. We supposed that NLRP3 drives a basal inflammatory activation in pericytes and promotes pericyte survival, although it was difficult to detect the secretion of IL-1β from pericytes in the brain. In cultured pericytes, we did find that inhibition of NLRP3 attenuated phosphorylation of multiple inflammation-related kinases, such as AKT and ERK, and perhaps also NFkB (p = 0.094), which is correlated with decreased cell proliferation and PDGFRβ and CD13 expression. Moreover, treatments with IL-1β increase PDGFRβ and CD13 expression in our cultured pericytes. It is consistent with a report that TNF-α at 10ng/ml promotes cultured pericytes to proliferate and migrate [30]. Thus, it is not surprising that angiogenesis is activated with pericyte proliferation in inflammatory lesion sites of multiple sclerosis [31]. During early wound healing, NLRP3 facilitates angiogenesis; however, production of IL-1β appeared not to be always necessary for angiogenesis [32].
It should be noted that NLRP3 might not provide protective effects on pericyte when the inflammatory activation surrounding pericytes is severe and lasts for a long term. In AD brain, pericytes are damaged and lost [6,7], while NLRP3contained inflammasome is activated [13]. However, there are potentially other mechanisms than sustained NLRP3 activation, which damage pericytes. For example, brain-delivered neurotrophic factor (BDNF) drops down and the activation of BDNF receptor, TrkB, is impaired in AD brain [33]. TrkB signaling regulates pericyte migration. Deletion of TrkB in pericytes reduces pericyte density and causes abnormal vasculogenesis in the heart [34]. In our unpublished experiments, we observed that deletion of BDNF in astrocytes or neurons decreases protein levels of PDGFRβ and CD13 in cerebral blood vessels (Wenqiang Quan, Qinghua Luo, Zhengyu Tang, Michael Menger, Klaus Fassbender and Yang Liu). Deletion of NLRP3 in APP or Tau-transgenic mice was reported to rescue neuronal functions and shift microglial activation from pro-inflammatory to anti-inflammatory profile [13,16].
Unfortunately, the microvascular circulation in NLRP3-deficient AD mice has not been investigated.
AKT is a known kinase to regulate cell survival, proliferation, and angiogenesis in response to extracellular signals [35]. AKT activation prevents pericyte loss in diabetic retinopathy [36]. In our experiments, AKT activation might mediate protective effects of NLRP3 on pericytes, as AKT phosphorylation is reduced by NLRP3 inhibition but enhanced by IL-1β activation. Moreover, inhibition of AKT down-regulates expression of PDGFRβ and CD13 in pericytes. Interestingly, PDGFRβ activation induces phosphorylation of AKT [37]. Thus, PDGFRβ and AKT activate each other and form a potential positive feed-back to maintain healthy pericytes in the brain. conclusions Our study suggested that NLRP3 activation maintains healthy pericytes in the brain through activating AKT signaling pathway. In the potential anti-AD or other therapies, in which NLRP3 inhibitors are administered, the potential adverse effects of NLRP3 inhibition on pericyte function and microcirculation should be considered.

Availability of data and materials
The data and materials published in this study are available from the corresponding 16 author on reasonable request.   Inhibition of AKT signaling pathway suppresses protein expression of PDGFRβ and CD13 in cu