25-Hydroxycholesterol is involved in the pathogenesis of amyotrophic lateral sclerosis

This study aimed to evaluate the levels of three major hydroxycholesterols (24-, 25-, and 27-hydroxycholesterols) in the serum and cerebrospinal fluid (CSF) of patients with amyotrophic lateral sclerosis (ALS), as well as to show their role in the pathogenesis of ALS experimental models. The level of 25-hydroxycholesterol were higher in untreated ALS patients (n = 30) than in controls without ALS (n = 33) and ALS patients treated with riluzole (n = 9) both in their serum and CSF. The level of 25-hydroxycholesterol in the serum of ALS patients were significantly associated with their disease severity and rate of progression. In the motor neuron-like cell line (NSC34) with the human mutant G93A superoxide dismutase 1 gene (mSOD1-G93A), 25-hydroxycholesterol induced motor neuronal death/ apoptosis via glycogen synthase kinase-3β and liver X receptor pathways; riluzole treatment attenuated these effects. The expressions of enzymes that synthesize 25-hydroxycholesterol were significantly increased in the brains of early symptomatic mSOD1G93A mice. Our data, obtained from patients with ALS, a cellular model of ALS, and an animal model of ALS, suggests that 25-hydroxycholesterol could be actively involved in the pathogenesis of ALS, mostly in the early symptomatic disease stage, by mediating neuronal apoptosis.


INTRODUCTION
Amyotrophic lateral sclerosis (ALS) is a progressive degenerative disease of the central nervous system (CNS) that lead to the progressive death of the motor neurons [1] and death about four years after disease onset [2]. Although various mechanisms, such as neuroinflammation, excitotoxicity, mitochondrial dysfunction, abnormal protein aggregation, and oxidative stress have been proposed to be associated with the pathogenesis of ALS, no single mechanism has been identified as responsible [3]. Cholesterol, a major constituent of the CNS, is synthesized de novo [4]. It is metabolized into diverse types of hydroxycholesterols (OHCs), thereby regulating cholesterol homeostasis [5], cellular apoptosis [6], and signaling [7].
We hypothesized that the synthesis of OHC in the CNS might contribute to the pathogenesis of ALS. The aim of this study was to measure the levels of three major OHCs  in serum (OHC serum ) and cerebrospinal fluid (OHC CSF ) of patients with ALS and to assess their role in the pathogenesis of ALS experimental models.
significantly in terms of demographic, body mass index, or levels of serum cholesterols ( Table 1).

The 25-OHC induced neuronal apoptosis, activated the GSK-3β pathway in ALS in vitro model, which was attenuated by riluzole
Among the three OHCs, the 25-OHC induced most severe motor neuronal death in a concentration of the 2.5uM to 40 uM ( Figure 5A) in a motor neuronal cell line stably expressing mutant G93A superoxide dismutase 1 gene (mSOD1-NSC34) cells. It also activated the GSK pathway ( Figure 5B) and induced cell apoptosis ( Figure  5C). Riluzole reduced 25-OHC-induced apoptosis ( Figure  5C), neuronal death ( Figure 5D), and activation of GSK-3 ( Figure 5E).
Western blot analysis showed that 25-OHC facilitated release of cytochrome c from injured mitochondria, activated caspase-3, and increased the cleavage of poly (ADP-ribose) polymerase (PARP) in mSOD1-NSC34 cells. However, this activation of cell death signaling cascades by 25-OHC was significantly reduced in a dose-dependent manner after treatment with riluzole ( Figure 6A-6C). Moreover, immunofluorescence staining revealed that 25-OHC increased the immunoreactivity of cleaved caspase-3 immunoreactivity and Poly ADP ribose polymerase (PARP), which were reduced by treatment with riluzole ( Figure 6D).
Administration of 25-OHC have activated of GSK-3β (Ser9) and phosphorylated tau (Ser396), which were prevented by treatment with riluzole ( Figures 7A  and 7B). Treatment with either GSK-3 inhibitor (10 μM SB415286) or riluzole reduced the GSK-3β activation induced by 25-OHC, although combined treatment with both GSK-3 inhibitor and riluzole did not produce synthetic effects above and beyond the effects of either on its own ( Figure 7C).
To determine a potential mechanism for the observed increases in 25-OHC in ALS, we examined the mRNA expression of CH25H, CYP3A4, and CYP7B1 in the brain and spinal cord of transgenic mice with the human mutant SOD1 gene containing a glycine 93 (Gly93) to alanine (Ala) substation (mSOD1-G93A mice) at different time points (postnatal day of 60, 90, and 120) ( Figure 9). In the asymptomatic stage (60 days), the mRNA expression of these enzymes in mSOD1 G93A mice did not differ between ALS and WT mice. However, in the early symptomatic stage (90 days), mRNA expression of the 25-OHC synthesizing enzymes, CH25H and CYP3A4, was increased significantly compared to agematched WT mice and asymptomatic mSOD1-G93A mice; these effects were primarily observed in brain tissue. The 25-OHC metabolizing enzyme, CYP7B1, was also moderately increased in the early symptomatic SOD1G93A mice, both in the brain and the spinal cord. Interestingly, these high levels of mRNA expressions of 25-OHC associated enzymes were only observed in the early symptomatic stage but not in the late symptomatic stage in mSOD1G93A mice.

DISCUSSION
This study showed that the 25-OHC is involved in the pathogenesis of ALS. The levels of 25-OHC CSF and 25-OHC serum were higher in riluzole-naive ALS patients than in controls and/or riluzole-treated ALS patients. Moreover the level of 25-OHC serum in ALS patients were significantly associated with their disease severity and rate of progressions. The 25-OHC also induced cellular apoptosis and activated the GSK-3β/ LXR pathways in a cellular model of ALS. Lastly the expression of enzymes important for synthesis and metabolism of 25-OHC was increased in an animal model of ALS, in particular, at the early symptomatic stage of the disease.
Our observations are in line with several evidences from the existing literature. Specifically, the expression of the 25-hydroxylase gene was shown to be increased in autopsied tissues of patients with ALS [12]. The inactivation of liver X receptor beta (LXRß), which promotes cholesterol transport in the CNS, has induced the pathogenesis of ALS pathogenesis in a mouse model [13]. The mutation in the CYP7B1 gene, responsible for 25-OHC metabolism, caused the hereditary degenerative disease involving motor neurons [14]. The metabolism for lipid were dysregulated in both mouse models of ALS and patients with ALS [15,16]. Finally, 25-OHC can induce mitochondria-dependent cell apoptosis via activation of glycogen synthesis kinase-3β (GSK-3β) [17], an enzyme analysis, performed with antibodies against p-GSK3β (Ser9) and its substrate p-TAU (Ser396) revealed that 25-OHC activated A. GSK3β and B. subsequent tau pathway. C. Both the GSK-3 inhibitor (10 μM SB415286) and riluzole treatments reduced 25-OHC-induced GSK-3β activation. Compounded effects of combined treatment of t GSK-3 inhibitor and riluzole did not differ from the either of these treatments alone (C). Data are expressed as a ratio of the simultaneously assayed non-treated group's value and were compared using Tukey's test after a one-way ANOVA (n=5). *p<0.05 when compared with non-treated group. #p<0.05 when compared with 5 μM 25-OHC treated group. NS, not significant. considered to be involved in the disease progression of ALS [18].
In a recent study of OHCs in ALS patients, levels of 25-OHC serum tended to be higher in ALS patients than controls, but this difference did not reach statistical significance [19], in contrast to our observations. The reason for this discrepancy may be that our patient inclusion criteria were stringent, such that possible or probable cases of ALS according to the El Escorial criteria were excluded [20]. Moreover, we have controlled the compounding effects of riluzole treatment on OHC level, measured the severity of patients by ALSFRSr, and analyzed the OHC CSF , as well as the OHC serum , which could have contributed the more robust and readily observable results in our study.
Among the three major OHCs, the 25-OHC induced most severe motor neuronal death in vitro model of ALS compared to the 24-and 27-OHC ( Figure 5A). The 25-OHC is produced in the macrophage and/or glia by the enzyme CH25H [21,22], in response to Toll-like receptor-4 (TLR-4) signaling [23] Moreover patients with ALS had increased expression of the TLR-4 in their glia of spinal cord. [24]. We speculate that this 25-OHC could mediate the motor neuronal death of ALS, via GSK3-ß and LXR pathway, in response to the upregulated the inflammatory signals of the glia (Figure 10).
The OHC engaged in diverse biological activities, including cell signaling, cellular cholesterol synthesis, and cytotoxicity [25,26]. Among these metabolites, the level of 24-OHC has been shown to increase in patients with Alzheimer's disease and is associated with disease severity [27]. The metabolite 27-OHC is proposed to reflect the effect of the peripheral hypercholesterolemia to the CNS via its net flux from peripheral circulation to the CNS [28]. Recently, 25-OHC has drawn attention for its role as both an anti-inflammatory and pro-inflammatory mediator [11]. In addition to these important roles of OHCs, we speculated that 25-OHC might play a pathogenic role in ALS by mediating GSK-3ß activation, LXR signaling, and neuronal apoptosis.
Although disease severity in patients with ALS patients was correlated with levels of 25-OHC serum , it was not correlated with levels of 25-OHC CSF . This may explained by our finding that the expression of 25-OHC synthesizing enzyme was only transiently increased in the CNS of ALS mice during the early symptomatic stage of  The qRT-PCR reactions were performed with mRNA with primers for the 25-OHC synthesizing enzyme (CH25H and CYP3A4) and for the 25-OHC metabolizing enzyme (CYP7B1) using A-C. brain and D-F. spinal cord at different disease stages (60, 90, and 120 days) of SOD1 G93A mice and age-matched WT controls. (A). The expression of CH25H and (B) CYP3A4 mRNA were significantly increased in the brain tissues of early symptomatic stage (90 days) SOD1 G93A mice compared to age-matched WT controls and asymptomatic SOD1 G93A mice. (C and F). The expression of CYP7B1 was also modestly increased in the brain and spinal cord of SOD1 G93A mice in early symptomatic stages. These high expression levels of 25-OHC-associated enzymes were not observed in the late symptomatic stage of SOD1 G93A mice. Group means ± SEM are plotted; #p <0.05; SOD1 G93A mice compared to WT mice on the same day. *p <0.05; SOD1 G93A mice compared at different time points. Differences between groups were analyzed with two-tailed Student's t-tests. Abbreviations: CH25H = cholesterol 25-hydroxylase, CYP3A4= Cytochrome P450 3A4, CYP7B1 = 25-hydroxycholesterol 7-alpha-hydroxylase, SOD1 = superoxide dismutase 1, WT = wild type. the disease. It may be that this temporarily high expression of 25-OHC-synthesizing enzymes may have interfered with the production of 25-OHC in the CNS of patients who are in the advanced stages of ALS.
The precise cause of decreased 25-OHC CSF in ALSriluzole patients remains unclear. However, in light of a previous study showing that glutamate excitotoxicity increased both cholesterol and cholesterol oxidation products in the brain [14], these results may be attributable to anti-glutamatergic effects of riluzole [29].
Our study has several limitations. First, the number of patients were relatively small. Second, the 25-OHC may not be solely responsible for the whole pathogenesis of ALS. Because the ALS-riluzole group, who had relatively low levels of 25-OHC CSF , continued to experience the disease progression [30]. Moreover, the 25-OHC might also mediate diverse disease conditions, including viral infection [31], X-linked adrenoleukodystrophy [8], multiple sclerosis [32], and neuromyelitis optica [33], other than the ALS [23]. Therefore, it seems that the 25-OHC might be one of the diverse inflammatory mediator rather than a specific marker for ALS. Second, we have used the serum and CSF from patients as controls, rather than those from healthy volunteers. This is for the ethical issues of performing the lumbar punctures on the healthy subject. However, as we have excluded both the degenerative or inflammatory diseases of the CNS from the controls, we speculate that this issue would not significantly affect the result of our study. Lastly, the level of OHCs in the CSF and serum of ALS animal could also have been another interesting point, which were not performed in our study because the concentrations of the OHC were expected to be extremely low in mice.
Our data showed that the 25-OHC could be an important mediator of the ALS pathogenesis involving in the GSK3-ß activation and neuronal apoptosis, which could eventually trigger the onset or accelerating the progression of ALS symptoms. Our results may reveal a novel ALS treatment target and provide insight into the involvement in of CNS cholesterol metabolism in ALS pathogenesis.

Patients
Patients who conformed to the diagnosis of definite or probable ALS according to the revised El Escorial criteria [20] were recruited from the ALS Clinic at the Seoul National University Hospital between December 2009 and May 2012. The following patients were excluded: those diagnosed more than 3 years before study commencement, those aged < 20 years or >80 years, those with respiratory support (tracheostomy, noninvasive ventilation, or oxygen inhalation), those fed via levin tube or gastrostomy tube (percutaneous endoscopic gastrostomy, PEG), those with presence of an acute systemic illness (per se, aspiration pneumonia), previous enrollment in the clinical drug or stem cell trial for ALS, or absence of written consent. ALS patients were also dichotomized according to the use of riluzole, as ALSnaive group and ALS-riluzole group [30].
The control group was comprised of patients who visited the Neurology Clinic at the Seoul National University Hospital. Patients were excluded from the control group if they had been diagnosed with a either neurodegenerative [20] or neuroinflammatory disease [34] of the central nervous system, any clinical or electrophysiological evidence of motor neuron disease [20], or if they had used an oral or intravenous steroid that might have affected cholesterol metabolism.

Clinical parameter and sampling protocols
Age, gender, level of serum total cholesterol, body weight, and height were assessed in patients with ALS and non-ALS controls. The ALSFRSr [35] and duration of disease were also assessed in patients with ALS. The progression rate (∆FS) was calculated with minor modification of the previous study as follows; ∆ FS = (48 -ALSFRSr at "time of sampling")/duration from onset to diagnosis (month) [36]. All samples of CSF and serum were immediately centrifuged and stored at -80 o C after collection, according to standard protocols [37].

Measurement for the CSF and serum level of 24-, 25-, and 27-hydroxycholesterols
Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analyses were performed at Korea Institute of Science and Technology (KIST), according to our recent methods [33]. Briefly, an API 4000 triple-quadrupole mass spectrometer (AB Sciex, Toronto, Canada) with an electrospray ionization source, in the positive ionization mode was used. The electrospray source was coupled online with a Shimadzu ultra-fast LC system (Shimadzu Corporation, Kyoto, Japan).

Standard protocol approval, registration, and patient consent
This study was approved by the Seoul National University Hospital Institutional Review Board (IRB number: H-0904-057-279). All patients provided written informed consent prior to participation. This study was approved by the IACUC (Institutional Animal Care and Use Committee) of Hanyang University.

Evaluation of apoptotic cells
To evaluate apoptosis, cells were stained by annexin V-FITC apoptosis detection kit (Sigma) following the manufacturer's instructions. Cells were washed and briefly trypsinized, and then washed twice with cold PBS. Cells were pelleted by centrifugation, resuspended in 1X binding buffer, and incubated with a staining solution (annexin V-FITC and PI) for 15 min. in the dark at 4°C. The cells were then resuspended in 1X binding buffer. Samples were kept on ice during the entire procedure and analyzed immediately by flow cytometry. Ten thousand cells from each sample were scanned and analyzed by FACS Canto (Becton Dickinson) using the standard configuration and parameters. Data acquisition and analysis was performed using the FACS DIVA software (BD Biosciences, San Jose, California, USA). Necrosis and apoptosis were determined by PI (FL2) and annexin V-FITC (FL1) fluorescence, respectively.

Transgenic SOD1G93A mice
B6SJL-Tg (SOD1-G93A) 1Gur/J mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). The mice were housed under a 12-h light/dark cycle and bred following the supplier's protocol. The presence of the human G93A transgene was confirmed by polymerase chain reaction (PCR) and evaluated for transgene copy numbers [45]. In this strain, the first symptoms and endstage symptoms of ALS appear at approximately 77 and 136 days of age [46]. Thus, we divided the SOD1 G93A mice into three groups that were evaluated at the following time points: 60 (asymptomatic), 90 (early symptomatic), and 120 days of age (late symptomatic) [47]. Wild-type (WT) mice (B6SJLF1/J) obtained from the Jackson Laboratory were used as controls. All animal experiments were approved by the Institutional Animal Care and Use Committee of Hanyang University.

Real-time polymerase chain reaction (RT-PCR)
We quantified mRNA expression of CH25H and CYP3A4, which synthesize 25-OHC, and the CYP7B1, which metabolizes 25-OHC, using the mSOD1G93A mice (n = 30) and age-matched WT mice (n = 30). Mice were randomly assigned into three groups that were evaluated at each time points: n = 10. RNA was isolated from the lumbar spinal cord and cortical brain regions of mice and extracted using RNeasy Mini Prep (Qiagen, Germany). The quantity and integrity of RNA was evaluated by measuring the absorbance at 260 nm via Nanodrop (Thermo Scientific, ND-2000). The cDNA was synthesized from 2.5 μg of RNA using Vilo cDNA kit (Invitrogen, USA) per the manufacturer's protocol. Primers for quantitative polymerase chain reaction (PCR) were purchased from Qiagen. We analyzed CYP3A4 (Qiagen, Germany, PPM03943B), CYP7B1 (Qiagen, Germany, PPM03979F), CH25H (Qiagen, Germany, PPM27829A), and GAPDH (Qiagen, Germany, PPM02946E). Firststrand cDNA was amplified using Power SYBR Green PCR master mix (Applied Biosystems, USA) with primers. Real-time RT-PCR was conducted with Applied Biosystems StepOnePlus TM (Carlsbad, USA) at 95°C for 10 min, followed by 40 cycles of 15 sec at 95°C and 1 min at 60°C. A melting curve was performed to check the specificity of amplification. The relative quantity (RQ) was calculated by 2-ΔΔCt, taking GAPDH as the interval standard control. Each sample was analyzed in triplicate.

Statistical analyses
Clinical data are presented as means ± standard deviation. Experimental data are presented as means ± SEM of five or more independent experiments. Differences between groups were analyzed by Student's t-test or one-way ANOVA followed by Tukey's post hoc comparisons. Two-tailed p-values < 0.05 were considered statistically significant. Linear regression analyses were used to test for associations between values. All statistical analyses were performed using the SPSS 17.0 software package for Windows (SPSS, Seoul, Korea).