Low-Density Lipoprotein Receptor Deficiency Attenuates Neuroinflammation through the Induction of Apolipoprotein E

Objective We aimed to determine the role of the low-density lipoprotein receptor (LDLr) in neuroinflammation by inducing experimental autoimmune encephalomyelitis (EAE) in ldlr knock out mice. Methods MOG35–55 induced EAE in male and female ldlr−/− mice was assessed clinically and histopathologically. Expression of inflammatory mediators and apolipoprotein E (apoE) was investigated by qPCR. Changes in protein levels of apoE and tumor necrosis factor alpha (TNFα) were validated by western blot and ELISA, respectively. Results Ldlr−/−-attenuated EAE disease severity in female, but not in male, EAE mice marked by a reduced proinflammatory cytokine production in the central nervous system of female ldlr−/− mice. Macrophages from female ldlr−/− mice showed a similar decrease in proinflammatory mediators, an impaired capacity to phagocytose myelin and enhanced secretion of the anti-inflammatory apoE. Interestingly, apoE/ldlr double knock out abrogated the beneficial effect of ldlr depletion in EAE. Conclusion Collectively, we show that ldlr−/− reduces EAE disease severity in female but not in male EAE mice, and that this can be explained by increased levels of apoE in female ldlr−/− mice. Although the reason for the observed sexual dimorphism remains unclear, our findings show that LDLr and associated apoE levels are involved in neuroinflammatory processes.

inTrODUcTiOn Multiple sclerosis (MS) is an inflammatory autoimmune demyelinating disease of the central nervous system (CNS). Infiltrated macrophages and resident microglia are considered the dominant effector cells in MS and its animal model, experimental autoimmune encephalomyelitis (EAE) (1,2). Effector mechanisms include phagocytosis of myelin and the secretion of inflammatory and toxic mediators (3)(4)(5)(6)(7). However, increasing evidence indicates that phagocytes can also acquire a disease-resolving phenotype in the CNS. For instance, we recently defined that ingestion of myelin by macrophages alters the inflammatory phenotype by activating the cholesterol-and fatty acid-sensing nuclear Ldlr -/ -Attenuates Neuroinflammation through apoE Frontiers in Immunology | www.frontiersin.org November 2017 | Volume 8 | Article 1701 liver X receptors and peroxisome proliferator-activated receptors (8)(9)(10). These studies stress the pleiotropic role that phagocytes play in the pathophysiology of MS and indicate that lipid-signaling pathways drive the phenotype of phagocytes in MS lesions. The impact of cholesterol and its metabolites on the pathophysiology of MS is a topic of interest in MS-research (11)(12)(13)(14)(15)(16). In the MS brain, marked alterations in both myelin cholesterol and other lipid metabolites are found (13). Furthermore, plasma and cerebrospinal fluid (oxy)sterol levels are disturbed and closely correlate to neurodegenerative processes (14)(15)(16)(17). In addition, we showed that lipoprotein levels and function are altered in relapsing-remitting MS patients (RRMS), which may advance disease progression in these patients (18). More specifically, we found smaller low-density lipoprotein (LDL) particles in RRMS patients compared to healthy controls. Interestingly, the average LDL particle size was smaller in male RRMS patients compared to female RRMS patients, which suggests gender differences in lipid metabolism in MS. LDL size is described to be important for LDL function indicating the lipoprotein may be involved in the pathophysiology of disease. Smaller particles have an increased susceptibility to oxidation and a decreased LDL receptor (LDLr) affinity, which may promote their proinflammatory properties (19)(20)(21).
The LDLr is a 160 kDa cell surface glycoprotein that plays a key role in plasma cholesterol homeostasis (22). It binds two physiologically important ligands, apolipoprotein B-100 (apoB) and apolipoprotein E (apoE). The only known ligand for the LDLr in the CNS is apoE since apoB is not synthesized in the CNS and, in contrast to apoE, cannot cross the blood-brain barrier (22). Malfunctioning of the receptor in humans results in familial hypercholesterolemia (FH) and is an important risk factor for cardiovascular disease (23). FH is caused by genetic mutations that directly or indirectly affect the function of the LDLr. It is characterized by defective catabolism of LDL, which results in increased plasma cholesterol concentrations and lipid accumulation in macrophages and other immune cells. This promotes proinflammatory responses, including enhanced NF-κB signaling, inflammasome activation, and increased production of neutrophils and monocytes in the bone marrow and spleen (24)(25)(26)(27)(28). The ability of lipoproteins to modulate neuroinflammation has been suggested in recent studies (29,30). In addition, elevated LDL plasma levels were found to increase neurotoxicity in a mouse model for Alzheimer's disease (30). The above findings indicate that lipoproteins modulate neuroinflammation and neurodegeneration. However, the direct contribution of LDL and LDLr to these processes remains unclear.
Here, we determined the role of the LDLr in neuroinflammation by using the EAE model and ldlr knock out (ldlr −/− ) mice. We show that ldlr deficiency reduces disease severity in female, but not in male EAE mice. In line with this finding, the inflammatory burden was significantly lower in the CNS of female ldlr −/− mice compared to ldlr −/− male mice. Macrophages isolated from female ldlr −/− mice show an altered inflammatory profile and an impaired capacity to phagocytose myelin. Interestingly, apoE release was enhanced in macrophages derived from females and the beneficial effect of ldlr deficiency in EAE was absent in apoE/ ldlr double knock out mice. This shows that elevated apoE levels likely attenuate EAE severity in female ldlr −/− mice. Collectively, our findings indicate that changes in ldlr expression contribute to the progression of neuroinflammatory diseases in a genderspecific manner.

MaTerials anD MeThODs animals
Ldlr −/− and apoE −/− mice on a C57BL/6 background and C57BL/6 wild-type (WT) mice were purchased from Harlan Laboratories. Animals were fed a regular diet and housed in the animal facility of the Biomedical Research Institute of Hasselt University. All experiments were performed according to institutional guidelines and were approved by the ethical committee for animal experiments of Hasselt University.

induction of eae
Eleven-week-old mice were immunized subcutaneously at the base of the tail with 200 µg of recombinant human myelin oligodendrocyte glycoprotein MOG35-55 emulsified in 100 µl complete Freund's adjuvant supplemented with 4 mg/ml of Mycobacterium tuberculosis (H37RA strain) according to manufacturer's guidelines (Hooke Laboratories, Lawrence, USA). Within 2 h and after 22-26 h, mice were intraperitoneally injected with 0.1 ml pertussis toxin. Immunized mice were weighed and scored daily by following a five-point standardized rating of clinical symptoms: 0, no signs; 1, loss of tail tonus; 2, flaccid tail; 3, hind limb paresis; 4, hind limb paralysis; 5, death. Mice were sacrificed at day 18 (peak) and at day 27 post immunization.

T cell Proliferation
Spleen and inguinal lymph nodes (LN) were isolated from mice 9 days after EAE induction. T cells were isolated from LN and spleen and cultured in RPMI medium containing 20 µM β-mercaptoethanol, 2% mouse serum, 1% non-essential amino acids, 1% sodium pyruvate, 0.5% P/S. Cells were plated in a 96-well plate at a density of 3 × 10 5 cells/well and subsequently stimulated with 10 µg/ml MOG for 48 h. Next, 1μCi [ 3 H] thymidine (Amersham biosciences, UK) was added for 18 h after which cells were collected using an automatic cell harvester (Pharmacia, Uppsala, Sweden). A β plate liquid scintillation counter (Perkin Elmer, lifesciences, Wellesly, MA, USA) was used to quantify radioactivity and results are expressed as stimulation index (SI). The SI shows the relative proliferation of MOG-stimulated T cells compared to non-stimulated T cells (SI = 1).

Quantitative Pcr (qPcr)
RNA was extracted from tissue or cells using the RNeasy mini kit (Qiagen, Venlo, The Netherlands). In short, lysis was performed with Qiazol Lysis reagent (Qiagen) supplemented with 1% β-mercaptoethanol (Sigma-Aldrich). RNA concentration and quality were determined with a Nanodrop spectrophotometer (Isogen Life Science, The Netherlands). cDNA synthesis was conducted using the Quanta qScript cDNA synthesis kit (Quanta Biosciences, Boston, MA, USA) per manufacturer's instructions. qPCR was performed on a StepOnePlus™ Real-Time PCR system (Applied biosystems, Ghent, Belgium) using the SYBR green method (Applied Biosystems). The master mix contained 1× SYBR green, 10 µM primers, 12.5 ng cDNA, and nuclease free water. Data were normalized to the most stable reference genes and the ΔΔCt method was used to determine relative quantification of gene expression. Primers were designed using NCBI's Primer-blast (details are shown in Table 1).

nO Production
Peritoneal macrophages in culture were stimulated with 100 ng/ml LPS (Sigma-Aldrich) prior to the assay. NO production was determined in the supernatant of peritoneal macrophages using a Griess-reagent assay (Promega, Leuven, Belgium) according to the manufacturer's instructions.

Phagocytosis assay
Myelin was isolated from brain tissue of healthy adult WT C57BL/6 OlaHSD mice (Harlan) by means of sucrose density gradient centrifugation, as previously described (31

statistical analysis
Data were statistically analyzed using GraphPad Prism for windows (version 5.0) and are reported as mean ± SEM. D' Agostino and Pearson omnibus normality test was used to test Gaussian distribution. A two-tailed unpaired student t-test was used for normally distributed data sets [t(df); P]. The Mann-Whitney analysis (MWU, n1, n2; P) was used for data sets that did not pass normality. EAE scores were analyzed using two-way ANOVA (Bonferroni's post hoc multiple comparison test). *p < 0.05, **p < 0.01, and ***p < 0.001.

resUlTs
Ldlr Deficiency reduces eae severity in Female but not in Male Mice To elucidate whether the LDLr contributes to neuroinflammation, we induced EAE in male and female ldlr −/− mice. We show that ldlr deficiency has a sex-specific effect on the EAE course. Female ldlr −/− mice exhibited a decreased EAE disease severity compared to WT mice, whereas in male mice ldlr deficiency did not affect EAE disease severity (Figures 1A,B). In female animals, the mean peak of disease symptoms was reached around day 17. The mean disease score was attenuated in ldlr −/− (1.504 ± 0.1995) compared to WT (2.318 ± 0.1782) mice ( Figure 1C). Accordingly, the maximum disease score was lower in LDLr −/− (1.70 ± 0.30) compared to WT (2.68 ± 0.29) mice ( Figure 1D). The ldlr −/− and WT groups displayed a similar day of onset (14.29 ± 1.04 and 12.41 ± 0.34, respectively Figure 1E).
Ldlr Deficiency has no significant influence on immune cell infiltration into the cns Experimental autoimmune encephalomyelitis is characterized by the infiltration of peripheral immune cells into the CNS leading to a local inflammatory response. To determine whether this process is altered by ldlr deficiency in female mice, the accumulation of T cells (CD3) and macrophages (F4/80) in the CNS of WT and ldlr deficient female EAE mice was assessed by immunohistochemistry at day 18 and day 33 post immunization (Figure 2; Figures S1 and S2 in Supplementary Material). Despite a reduced disease severity, no significant differences in the number of infiltrated macrophages and T cells into the spinal cord tissue were observed comparing female WT EAE mice and ldlr-deficient EAE mice, although a trend on day 18 is visible.

Ldlr Deficiency in Female Mice has no influence on T cell Proliferation
T cell proliferation is an important hallmark of EAE and is essential for the initiation of EAE pathogenesis. Since T cells are dependent on cholesterol in order to proliferate (32), we investigated the influence of ldlr on T cell proliferation during EAE. T cells from both ldlr −/− and WT mice were isolated on day 9 post immunization and a thymidine assay was used to determine their proliferation capacity. No significant differences in T cell proliferation between ldlr −/− and WT mice induced with EAE ( Figure 3A). In addition, we determined mRNA expression of transcription factors such as T-bet, GATA3, RORγt, and Foxp3, which are involved in T cell differentiation into Th1, Th2, Th17, and Treg subsets, in LN and spleen but found no difference between ldlr −/− and WT female mice (Figures 3B,C). Lastly, we analyzed the mRNA expression levels of cytokines important in T cell differentiation (IFNγ, IL (Figures 4A,B), whereas the expression of these inflammatory genes is not changed in male ldlr −/− mice ( Figure 4C). These data indicate that female ldlr −/− mice have decreased inflammation in the spinal cord compared to male ldlr −/− and female WT mice.

Ldlr Deficiency suppresses the inflammatory Phenotype of Macrophages in Female Mice
As the ldlr is highly expressed on macrophages and we observed a reduced expression of cytokines typically released by macrophages in female ldlr −/− EAE mice, we investigated the influence of ldlr deficiency on the inflammatory phenotype of macrophages in vitro. We found that female LPS-stimulated ldlr −/− macrophages produce less TNFα and NO compared to macrophages isolated from WT females (Figures 5A,B). In addition, we show that macrophages isolated from ldlr −/− female mice less efficiently internalize myelin compared to macrophages isolated from WT females ( Figure 5C). Interestingly, expression of lrp1, a major receptor involved in the phagocytosis of myelin (33), was reduced as well ( Figure 5D). The expression of the proinflammatory genes Tnfα, IL-6 and Il-12 is reduced in ldlr −/− female mice, whereas the anti-inflammatory gene Arg1 is increased in macrophages isolated from ldlr −/− female mice compared to WT mice (Figures 5E-I). These data indicate that macrophages isolated from female ldlr −/− mice have an attenuated

apoe expression is increased in activated Peritoneal Macrophages isolated from ldlr −/− Female Mice
Apolipoprotein E is known to induce an anti-inflammatory phenotype in macrophages (34). Interestingly, we show that ApoE mRNA expression is significantly increased in activated peritoneal macrophages from ldlr −/− female mice compared to both male ldlr −/− mice and female WT mice (Figures 6A-C). Next, we used Western blot to determine the amount of apoE secreted into the culture medium. We found that only LPS-stimulated macrophages isolated from female ldlr −/− mice, and not those from WT and male ldlr −/− mice, secreted significant levels of apoE ( Figure 6D). These data suggest that ldlr deficiency in female mice is responsible for the increase and subsequent secretion of apoE by macrophages.
lack of apoe abrogates the reduced neuroinflammatory response in Female To define to what extend an increase in apoE accounts for the observed reduction in EAE disease severity in ldlr −/− female mice and the observed less-inflammatory phenotype of macrophages isolated from these mice, we crossbred apoE −/− and ldlr −/− mice. In contrast to female ldlr −/− mice, female apoE −/− ldlr −/− mice did not exhibit an attenuated EAE disease course compared to WT mice ( Figure 7A). In addition, the observed decreased capacity to phagocytose myelin and produce TNFα by macrophages from ldlr −/− female mice was counteracted by depletion of apoE (Figures 7B,C). Interestingly, the impairment in NO production observed in ldlr −/− macrophages was also present in apoE −/− ldlr −/− macrophages indicating apoE is not accountable for this effect (Figure 7D). ApoE −/− alone did not alter EAE severity in female mice compared to WT mice ( Figure 7E).
These data indicate that the lack of apoE largely restores the impaired inflammatory response induced by ldlr deficiency in female mice and that the effect of ldlr −/− on NO production is not responsible for the ameliorated disease course in these mice.

DiscUssiOn
In this study, we show that ldlr −/− attenuates neuroinflammation in female but not in male EAE mice, and that macrophages isolated from female mice have an attenuated inflammatory response and reduced phagocytic capacity. Interestingly, apoE deficiency counteracts the impact of ldlr deficiency in females on the course and pathology of EAE. These findings indicate that an increased apoE expression in macrophages apoE in female ldlr −/− mice is at least partially responsible for the observed reduction in EAE severity. We show that ldlr deficiency in female mice decreases the inflammatory burden in the CNS but does not affect T cell proliferation and differentiation, and immune cell infiltration into the CNS. However, the levels of proinflammatory mediators in the CNS were reduced in these mice and isolated macrophages showed an impaired inflammatory response. These findings suggest that ldlr deficiency affects the inflammatory and phagocytic properties of macrophages, thereby altering EAE disease severity.
In concordance, levels of the proinflammatory markers Tnfα, IL-6, and Il-12 were reduced while an increased gene expression of the anti-inflammatory marker Arg1 was observed in macrophages isolated from ldlr −/− female mice compared to WT macrophages. Moreover, ldlr deficiency decreases macrophage mediated myelin phagocytosis and expression of lrp1. LRP1 is a member of the LDLr family that also functions as a scavenger receptor and has been shown to be directly involved in myelin phagocytosis (33). Myelin degradation and phagocytosis by macrophages enhances demyelination which leads to neurological symptoms in MS (3). Although speculative, it is possible that lack of ldlr results in decreased Lrp1 expression, thereby indirectly impairing myelin phagocytosis. Together, an altered macrophage response at least partially accounts for the beneficial effects of ldlr deficiency on EAE severity. In contrast to the reduction in EAE severity in ldlr −/− mice, the clinical outcome was not altered in apoE −/− ldlr −/− mice. Activated peritoneal macrophages isolated from ldlr −/− female mice significantly increased apoE expression compared to WT or their male counterparts and, the beneficial effect of ldlr deficiency on myelin phagocytosis and TNFα production was absent in apoE −/− ldlr −/− macrophages. Although the LDLr is described to directly regulate apoE levels in the CNS (35), we found no elevated apoE gene expression in the CNS of ldlr −/− EAE mice (data not shown). Studies have reported that increased expression of apoE leads to a less-inflammatory phenotype in macrophages by downregulating M1-like and upregulating M2-like markers. In addition, apoE suppresses microglial activation and the release of TNFα (34,36,37). Furthermore, treatment of EAE animals with apoE-derived peptides ameliorates the clinical course of EAE and apoE deficiency exacerbates EAE by increasing immune reactivity and hampering CNS repair mechanisms (38,39). These findings suggest that the LDLr suppresses ApoE expression thereby exacerbating neuroinflammatory responses in females.
The reason for the sexual dimorphism in apoE expression remains elusive. It is reported that apoE exerts moderate sexspecific effects on neuroinflammation. Schrewe and coworkers found that the EAE disease course was slightly attenuated in male apoE −/− mice, whereas EAE was more severe in female apoE −/− mice compared to WT mice. In our hands apoE −/− did not alter EAE disease outcome in female mice compared to WT mice. Methodological differences (e.g., immunization protocol) may explain these discrepancies. Moreover, in Schrewe's study differences between WT and apoE −/− female mice were only visible after 30-36 days post immunization, whereas our experiment lasted for up to 22 days post immunization. Nevertheless, the question remains why only macrophages from ldlr −/− female mice abundantly secrete apoE. Macrophages express receptors for sex hormones such as estrogen and testosterone and these hormones modulate macrophage activity (40,41). Interestingly, administration of estrogens has been shown to directly regulate hepatic LDLr activity in human cells in vitro and 17β-estradiol increases the expression of apoE during the maturation stage of human monocytes to macrophages (42)(43)(44). As lack of LDLr has been shown to increase apoE levels also (35), both mechanisms may contribute to the increased apoE levels observed in female ldlr −/− mice.
However, it remains to be elucidated whether sex hormones are exclusively responsible for the observed effects. Studies investigating plasma lipid and lipoprotein concentrations across the menstrual cycle in women report virtually no variation (45)(46)(47), and those that do, report only small reductions in LDL during the luteal phase and no changes in plasma triglyceride or HDL concentrations (45,46). Therefore, the sexual dimorphism in lipid metabolism and its impact on neuroinflammation appears to be the result of an intricate network of sex hormones in combination with other modulators of lipid metabolism. The key pathways and mediators of this network remain to be elucidated, but there are many other factors involved such as insulin and adipokines (48), and gene expression and imprinting (49), that need to be explored in the future.
In addition to apoE, differences in female and male mice may be the result of differences in LDLr levels and function. However, there are few studies about sex-related differences in LDLr behavior. Segatto and coworkers determined age-and sexrelated differences in extra-hepatic LDLr expression and found that LDLr protein expression was decreased in 12-month-old (onset of menopause) female rat brains, which was completely restored by estrogen treatment. In skeletal muscle, LDLr levels were increased in both male and female aged rats, whereas in the heart no modifications were observed either in aged rats or rats of a specific gender (50). These data illustrate the complex age-and sex-related and tissue-specific regulation of LDLr expression, and warrant the need of more research to clarify the implications of these differences for neuroinflammatory diseases and other conditions involving the LDLr.
Collectively, we show that ldlr −/− attenuates EAE disease severity in female but not in male EAE mice. Although the nature of the observed sexual dimorphism remains unclear, our findings show that LDLr and associated apoE levels are involved in neuroinflammatory diseases such as multiple sclerosis, which may have implications for future treatment strategies.