Sterol auto-oxidation adversely affects human motor neuron viability and is a neuropathological feature of amyotrophic lateral sclerosis

Aberrant cholesterol homeostasis is implicated in the pathogenesis of amyotrophic lateral sclerosis (ALS), a fatal neuromuscular disease that is due to motor neuron (MN) death. Cellular toxicity from excess cholesterol is averted when it is enzymatically oxidized to oxysterols and bile acids (BAs) to promote its removal. In contrast, the auto oxidation of excess cholesterol is often detrimental to cellular survival. Although oxidized metabolites of cholesterol are altered in the blood and CSF of ALS patients, it is unknown if increased cholesterol oxidation occurs in the SC during ALS, and if exposure to oxidized cholesterol metabolites affects human MN viability. Here, we show that in the SOD1G93A mouse model of ALS that several oxysterols, BAs and auto oxidized sterols are increased in the lumbar SC, plasma, and feces during disease. Similar changes in cholesterol oxidation were found in the cervical SC of sporadic ALS patients. Notably, auto-oxidized sterols, but not oxysterols and BAs, were toxic to iPSC derived human MNs. Thus, increased cholesterol oxidation is a manifestation of ALS and non-regulated sterol oxidation likely contributes to MN death. Developing therapeutic approaches to restore cholesterol homeostasis in the SC may lead to a treatment for ALS.

www.nature.com/scientificreports/ mRNA levels were not altered in SOD1 G93A mice. However, significant progressive increases in CYP46A1 and CH25H mRNA levels were observed in SOD1 G93A mice vs. WT controls. Notably, CH25H levels were elevated more than 30-fold in end stage (ES) SOD1 G93A mice (Fig. 1b). Our results corroborate an earlier finding showing a similar increase in CH25H mRNA in microglia isolated from SOD1 G93A mice 27 . Next, we used LC-MS to measure oxysterol levels in the whole lumbar spinal cords of ES SOD1 G93A and WT mice (N = 5/group). We found that 25-hydroxycholesterol (25-OHC) was significantly elevated nearly tenfold in the spinal cords of ES SOD1 G93A mice (Fig. 1c). Levels of 24S-hydroxycholesterol (24S-OHC) and 27-hydroxycholesterol  were not altered in the spinal cords of SOD1 G93A mice (Fig. 1c). Interestingly, several oxysterols derived through the autooxidation of cholesterol (Fig. 1d) by reactive oxygen (e.g., OH . ) or nitrogen (ONOO -; peroxynitrite) 25 were also significantly elevated in the spinal cords (Fig. 1e,f) of SOD1 G93A mice compared to WT controls. These oxysterols include 7α-hydroxycholesterol (7α-OHC; generated by auto-oxidation or by CYP7A1), 7β-hydroxycholesterol (7β-OHC), 7-ketocholesterol (7-KC), 5α6α-epoxycholesterol (5a6a-EC), 5ß6ß-epoxycholesterol (5b6b-EC) and 5α,6ß-dihydroxycholestanol (THC). Emerging evidence suggests that BAs synthesis may also occur in the CNS 24,28 ; however, BA detection in CNS tissue remains to be demonstrated. We found that the expression levels of CYP7B1 and HSD3B7, enzymes that regulate acidic BA synthesis were significantly upregulated in the spinal cords of ES SOD1 G93A mice vs. WT controls (Fig. 1g). In contrast, mRNA levels of CYP8B1, a mediator of neutral BA synthesis was significantly downregulated in symptomatic (SYMP) and ES SOD1 G93A mice (Fig. 1g). These findings suggest that oxysterols are metabolized to BAs within the spinal cord of SOD1 G93A mice. In the liver, BA synthesis is under circadian regulation 29 ; therefore, in order to increase the likelihood of detecting disease-related changes in spinal cord BA levels, we measured primary, secondary and conjugated BAs at two different time points, 10 am and 6 pm, using spinal cords collected from ES SOD1 G93A and WT mice (N = 7/group/time point). Significant elevations (up to tenfold) in cholic acid (CA), chenodeoxycholic acid (CDCA), ursodeoxycholic acid (UDCA), β-Muricholic acid (β-MCA), taurocholic acid (TCA), taurochenodeoxycholic acid (TCDCA), taurodeoxycholic acid (TDCA), tauroursodeoxycholic acid TUDCA, Tauro-β-muricholic acid (TBMC), glycocholic acid (GCA), and glycodeoxycholic acid (GDCA) were found in the spinal cords of ES SOD1 G93A vs. their WT counterparts at the 10 am (Fig. 1h), but not the 6 pm time point ( Supplementary Fig. S1). Collectively, our findings indicate that increased FC oxidation is the spinal cord is indeed a manifestation of disease in ALS. We also demonstrate for the first time that BAs are detected in CNS tissue, and that their levels are likely circadian rhythm dependent.
Oxidized cholesterol metabolites are potential biomarkers of disease. Next we measured oxidized cholesterol metabolites in the feces, urine, serum and plasma of ES SOD1 G93A mice and WT controls (n = 5/group/sample matrix) to determine if disease changes in cholesterol oxidation found in the spinal cord are also present systemically as potential biomarkers of disease. Oxysterol levels were unaltered in the feces, urine, and serum of ES SOD1 G93A mice compared to WT controls. In plasma, however, enzymatically derived oxysterols and auto-oxidized metabolites of cholesterol were significantly elevated in ES SOD1 G93A mice compared to WT controls. Specifically, increases in 7αOHC, 24-OHC, 7β-OHC, 7-KC, 5α6α-EC and 5β6β-EC were detected (Fig. 2a, b). Increased fecal BA excretion plays a significant role in removing excess sterols from the body. Consistent with our observations of increased sterols in the spinal cord and plasma of SOD1 G93A mice, we also found that BA levels were also significantly elevated in feces during ALS. Specifically, levels of CDCA, LCA, TDCA and TLCA were significantly elevated in the feces of SOD1 G93A mice compared to their WT counterparts (Fig. 2c). Collectively, our findings suggest that measuring oxidized cholesterol metabolites in plasma and feces may potentially serve as potential biomarkers of ongoing cholesterol cacostasis that occurs in the spinal cord during disease course in ALS.
Human and mice show species differences in spinal cord oxidized sterol composition. It is unknown if humans and mice have comparative differences in spinal cord cholesterol oxidation. Thus, we compared the relative percentages of enzymatically derived oxysterols and BAs in human control spinal cord grey matter (GM), (ventral white matter) VWM and WT mouse spinal cord tissue homogenates (N = 6-7/group). For enzymatically derived oxysterols significant species differences were observed for 7α-OHC, 24-OHC and 27-OHC (Fig. 3a). Interestingly, levels of 27-OHC were several-fold higher in humans compared to mice. As expected, based on what has been reported for the liver, only α-muricholic acid (α-MCA) and β-MCA were detected in the spinal cords of mice. Species differences in the primary and secondary BAs CDCA and DCA were also found (Fig. 3b). Significant comparative differences in conjugated BAs were also found. Mice showed higher levels of taurine BA conjugates (i.e., TCA), whereas glycine BA conjugates (i.e., GCA, GCDCA, and GDCA) were relatively enriched in humans (Fig. 3b). Collectively, our findings suggest that human and mice oxidize cholesterol differently in the spinal cord.

Oxidized cholesterol accumulation occurs in the spinal cord of ALS patients. To determine
if increased cholesterol oxidation also occurs in ALS patients, we analyzed GM and VWM spinal cord tissue homogenates for levels of enzymatically derived oxysterols, auto-oxidized sterols and BAs (N = 6-7/tissue type/ group). In contrast to SOD1 G93A mice, 25-OHC levels were not altered, but 24S-OHC levels were significantly increased in the VWM of ALS patients vs. controls (Fig. 3c). Levels of 7α-OHC, an oxysterol generated by both CYP7A and auto-oxidation, was significantly increased in the GM and VWM of ALS patients compared to controls (Fig. 3d). Similar to SOD1 G93A mice several auto-oxidized generated sterols including 7β-OHC, 7-KC, 5α6α-EC, 5β6β-EC and THC were significantly increased in the spinal cords of ALS patients compared to controls (Fig. 3e). Not surprisingly, since postmortem tissue samples are collected at variable time points, and our findings in mice suggest BA synthesis in the spinal cord has a circadian rhythm, we did not see significant For 27-OHC, CTG analysis showed that hMN viability was only adversely affected at 100 μM (Fig. 4c). Surprisingly, CTG and LDH analysis showed that BAs even at very high concentrations did not decrease hMN survival ( Supplementary Fig. S2). In contrast to enzymatically derived oxysterols, auto-oxidized sterols were relatively more detrimental to hMN survival. 7α-OHC promoted significant hMN demise at (10, 30 and 100 μM) (Fig. 5a). Notably, CTG analysis showed that 7β-OHC, 5α6α-EC, and THC adversely affected hMN viability starting at 1 μM (Fig. 5b,c,d). 7-KC and 5β6β-EC were also damaging to hMN survival starting at 3 and 10 μM respectively (Fig. 5e,f). Collectively, our findings indicate that enzymatically derived oxidized metabolites of cholesterol are relatively non-toxic to hMNs, whereas sterols generated through the auto-oxidation of cholesterol are detrimental to hMN viability.

Discussion
Similar to other neurometabolic diseases that feature neurodegeneration of the motor system 30 , the unique structural composition of the spinal cord may play a role in determining its relative vulnerability to pathogenic triggers of amyotrophic lateral sclerosis (ALS). Notably, the spinal cord is particularly enriched in cholesterol, Indeed, growing evidence suggests that disrupted cholesterol homeostasis (i.e., cholesterol cacostasis) contributes to disease pathogenesis in ALS. Cholesterol cacostasis occurs in the spinal cord during aging 32 , a risk factor for developing ALS, and a number of genes associated with familial ALS are also known to affect cholesterol homeostasis. For example, VCP, FUS and HNRNPA1 regulate levels of HMG-CoA reductase (HMGCR) 8,9,12 , the rate limiting enzyme in cholesterol biosynthesis. Moreover, ANG modulates the neutralization of free cholesterol (FC) to cholesterol esters (CEs) 11 , and VAPB and SIGMAR1 affect intra-organelle cholesterol transport 10,33 . And lastly, CEs are markedly increased in the spinal cords of ALS patients and in SOD1 G93A rodent models of disease [17][18][19] , and inducing cholesterol cacostasis in the SC of normal mice leads to ALS-like lipid neuropathology, astrogliosis, motor neuron (MN) death, and clinical features of ALS 18 . Sustained FC loading in mammalian cells leads to the activation of several pathological pathways implicated in the demise of MNs including ER stress, ROS generation, mitochondrial dysfunction, inflammation, and cell death 21,25,34 . FC-induced cellular toxicity is averted in part through the hydroxylation of FC to oxysterols, which is dependent upon ER (i.e., CYP7A1, CYP46A1 and CH25H) and mitochondrial (i.e., CYP27A1) resident enzymes. The relative solubility of enzymatically derived oxysterols facilitates their exit from cells 25 . Although oxysterols are elevated in the plasma 22 and CSF 20 of ALS patients, it is unknown if increased FC oxidation occurs in the spinal cord during disease. Here, we demonstrate that spinal cord oxysterol accumulation is a pathogenic feature of ALS. Levels of 25-OHC were increased over tenfold in the spinal cords of SOD1 G93A mice, but not in ALS patients. The latter finding was unexpected because similar to SOD1 G93A mice 27 CH25H expression is increased in the spinal cords of ALS patients 35 . Detecting disease related changes in patients may be challenging because some oxysterols are produced cyclically 29 and human tissue samples are often collected at variable circadian time www.nature.com/scientificreports/ points. Accumulation of 24S-OHC was found in the spinal cords of ALS patients, but not in SOD1 G93A mice. Interestingly, however, 24-OHC plasma levels were elevated in SOD1 G93A mice, which may be reflective of an earlier change in spinal cord 24-OHC levels. In both ALS patients and SOD1 G93A mice 7α-OHC spinal cord levels were elevated. For 27-OHC marked species differences were found with humans being relatively more dependent on this pathway of sterol oxidation. Consistent with these findings mutations in CYP27A1 are associated with cerebrotendinous xanthomatosis, a rare disease which sometimes presents as a clinical mimic of ALS due to upper motor neuron loss in humans 24,36 , but not in mice 37 . Oxysterols also act as potent signaling molecules to regulate cholesterol synthesis, metabolism, transport, and efflux through their modulation of the transcription factors sterol regulatory element binding protein 2 (SREBP2) and liver X receptors (LXRs). SREBP2 increases cellular cholesterol levels, whereas the activation of LXRs by oxysterols initiates the opposite effect 21,25,31 . Our finding of a marked increase in 25-OHC levels may explain why SREBP2 activation progressively decreases with disease in SOD1 G93A mice 18 , as 25-OHC precludes the proteolytic processing of SREBP2 into its activate form 38 . Moreover, elevated 24S-OHC levels may account for reductions in HMGRC that are observed in some ALS patients 18 ; 24S-OHC directly triggers the proteolytic   21 . Sustained oxysterol activation of LXRs also promotes the expression of members of the ATP-binding cassette (ABC) transporter super-family, which expel cholesterol from cells. Thus, increased oxysterol signaling may also potentially explain reported increases in ABCG1 and ABCA1 found in SOD1 G93A mice 18,39 . Notably, disruption of LXR signaling in mice leads to an ALS phenotype 40 and gene variants in LXR are associated with altered disease progression in ALS patients 5 . Oxysterols undergo further hydroxylation and oxidation to generate bile acids (BAs), a metabolic route that is under circadian control and reportedly circumscribed to the liver 41 . Several lines of evidence, however, suggest that BAs are synthesized locally within the CNS and that they have important roles in modulating MN viability. For example, cholestenoic acids activate LXRs to modulate MN survival in vitro and mutations in enzymes that regulate the synthesis of BA intermediates are associated with neuromuscular diseases including hereditary spastic paresis type 5, cerebrotendinous xanthomatosis, and adult-onset motor neuropathy 24 . Indeed, our current findings suggest that BA synthesis does occur in the spinal cord, and that it is also upregulated during disease course in ALS. Interestingly, in SOD1 G93A mice expression of CYP7B1, a regulator of acidic BA synthesis was increased in the spinal cord. This may be part of a compensatory response to avert disease because it is thought that reduced levels of acid BA precursors in the CSF of ALS patients contributes to disease progression 20 . Our measurement of BA levels showed that several primary, secondary and conjugated BAs were increased in the spinal cords of SOD1 G93A mice in a circadian dependent manner. This observation highlights the importance of taking circadian rhythms into consideration when trying to understand pathogenic mechanisms of disease. Not surprisingly, since human tissue collected at autopsy often occurs at varying times, BA levels in the spinal cords of ALS patients and controls were highly variable. Non-enzymatic or unregulated oxidation also occurs when sterols are inadvertently exposed to reactive oxygen or nitrogen species (e.g., OH . and ONOO -) or to immune cell oxidizers such as H 2 0 2 and HOCl 25 . This potential route of sterol oxidation seems likely during ALS because ONOOaccumulation 42 and neuroinflammation are robust features of disease 43 . Indeed, we found that several auto-oxidized sterols were elevated in the spinal cords of ALS patients and SOD1 G93A mice. Moreover, we also found that several auto-oxidized sterols, as well as enzymatically derived oxysterols and BAs were also elevated in the plasma and feces of SOD1 G93A mice. Thus, oxidized sterols may potentially serve as biomarkers of ongoing cholesterol cacostasis during ALS.
It is unclear if increased sterol oxidation is a compensatory response to disease (to reduce FC levels) or if it contributes to disease pathogenesis. Several studies suggest that oxidized sterols may adversely affect the viability of human MNs. For example, 24S-OHC triggers necroptosis in mouse neurons 44 , a cell death pathway implicated in ALS 45 . In addition, in rat organotypic brain slice cultures 24-OHC and 25-OHC markedly reduced ChATpositive neuronal survival 46 , and 25-OHC reduces the survival of SOD1 G93A NSC34 motor neuron like cells 22 . www.nature.com/scientificreports/ Furthermore, at high concentrations BAs function as membrane solubilizers, induce mitochondrial dysfunction and cause death in several cell types 34 . Moreover, sterols generated through auto-oxidation induce cell lysosomal and mitochondrial dysfunction 47 , and oxiapoptophagy 48 -a form of cell death involving simultaneous oxidative stress, apoptosis, and autophagy. Interestingly, in our experiments enzymatically derived oxysterols and BAs were relatively harmless to human MNs; and in some instances (e.g. 24-OHC) even improved human MN viability. In contrast to the enzymatically derived oxysterols; however, we found that several auto-oxidized generated sterols did adversely affected human MN survival. Collectively, these findings suggest that enzymatically regulated cholesterol oxidation during ALS may be a compensatory response to avert cholesterol cacostasis; whereas, exposure to auto-oxidized derivatives of cholesterol likely contributes to MN death. These findings should be interpreted with caution; however, because oxidized sterols in our experiments were not delivered to MNs using intracellular carrier proteins (e.g., oxysterol or fatty acid binding proteins) which would enable sterols to more efficiently affect MN physiology. Moreover, the indirect effect that these oxidized sterols have on MN survival through their effects on additional cell types that contribute to disease pathogenesis remains largely unknown. Nevertheless, our findings suggest that therapeutic approaches that restore CNS cholesterol metabolism to its optimal homeostatic state (i.e., eustasis) may lead to a disease-modifying treatment for ALS.

Materials and methods
Human tissue samples. Cervical spinal cord segments from six male patients who died of respiratory failure caused by advanced ALS were used in our analyses. The control cervical spinal cord segments used in our experiments came from seven age-matched male individuals without evidence of neurological disease. All human tissue samples were obtained from the NICHD Brain and Tissue Bank for Developmental Disorders at the University of Maryland, Baltimore, MD, contact HHSN275200900011C, Ref.

Animals.
Transgenic male ALS mice that express the mutant SOD1 G93A transgene at high levels (B6SJL background strain) were divided equally among groups. Mutant SOD1 G93A gene copy number and protein expression were confirmed by PCR and western blot analysis, respectively. Animals were housed under light:dark (12:12 h) cycles and provided with food and water ad libitum. At 75 days of age, food pellets were place on the cage floor to facilitate access and consumption. Mice were scored daily into the following phases: symptomatic (SYMP; median age of 82 days) = abnormal hind limb splay, end stage (ES; median age of 103 days) = onset of limb paralysis (typically hind limb), and moribund (MB; median age of 122 days) = unable to right themselves within 30 s after being placed on their backs. Wild type mice used were unaffected siblings from the same breed pairs. All samples taken from mice were collected between 9:45 and 10:15 AM unless stated otherwise to avoid variation due to circadian fluctuations in metabolism. This time point was selected because previously we showed that increased corticosterone signaling and glycogen accumulation-metabolic consequences of increased BA signaling-occur in SOD1 G93A mice during this sampling period 49,50 . At sacrifice all mice were perfused with cold 1X PBS (pH 7.4); the lumbar spinal cords were removed, flash frozen in liquid nitrogen and then stored at -80C. The N values and tissue samples used from each mouse is described in detail below for each assay. All experimental procedures and methods were performed in accordance with protocols and regulations approved by Sanofi Genzyme's Institutional Animal Care and Use Committee (IACUC).

Experimental design and statistics.
All experiments were carried out in compliance with the ARRIVE guidelines. The experimental design for each assay is described in its subsection below. Normality was determined using the Shapiro-Wilk normality test. Data sets that failed the normality test were analyzed with a Mann-Whitney test. A two-tailed unpaired t-test was used to compare data sets that passed the normality test and that had equal variances. If data set variances were significantly different, then an unpaired t-test with Welch's correction was used to compare groups that passed the normality test. Statistical tests comparing multiple groups were performed using a 1-way analysis of variance (ANOVA) followed by a Dunnett's multiple comparison post hoc test to find differences between group means. A value of p < 0.05 was considered statistically significant. All statistical tests were performed using GraphPad Prism Software 8.0.
Homogenates were placed on ice for 10 min and then centrifuged at 6300 X g at 4 °C for 10 min. Oxysterols were extracted from the tissue homogenate supernatants, mouse plasma (100ul; n = 5/group) and mouse serum (100 ul; n = 5/group) with methanol using the BIOCRATES Kit filter plate, which was preloaded with an internal standard mixture. The extract was then subjected to alkaline hydrolysis to release oxysterols from their respective esters. After neutralization the metabolites were determined by UHPLC-tandem mass spectrometry (LC-MS/MS) with Multiple Reaction Monitoring (MRM) in positive mode using a SCIEX API QTRAP5500 mass spectrometer with electrospray ionization. Bile acid (BA) analysis was carried out by Metabolon, Inc on spinal cord tissue samples (175-200 mg/sample). For human samples, grey and white matter tissue segments (n = 6-7/tissue type/group) from the cervical spinal cord were analyzed separately. For mouse experiments entire spinal cords were used (n = 14/group/time point). Spinal cord tissue from 2 different mice (80-100 mg tissue per mouse) were pooled together to generate 1 sample; thus, reported BA levels were generated from n = 7/group/time point. Deuterium-labeled internal standards were added to tissue samples and the mixture was solubilized in methanol followed by a crash extraction. Samples  ) at a concentration of 10,000 neurons/well. After seeding, plates were placed in a humidified chamber inside a standard cell culture incubator with 5% CO2 at 37 °C. On Day 7, 10 µL of the corresponding treatment solution was added to each well. For the negative control condition ("vehicle"), medium alone was added. An additional 90 µL of fresh culture medium was then added to all wells to reach a final volume of 200 µL/well. Each treatment was tested in triplicate in an 8-point dose-dependent manner. On Day 10, the plates in each respective group were processed to determine ATP concentration using CellTiter-Glo 2.0 assay (Promega, G9242) according to the manufacture's protocol. All plates and reagents were brought to room temperature before beginning the assay. First, 100 μL of medium was gently removed from each well. Next, 100 μL of the CellTiter-Glo buffer/luciferase solution was added to the wells. After 2 min of shaking at 300 rpm and 10 min of stabilization, the relative luminescence signal was measured using a plate reader (Tecan). Plates were processed in a staggered manner to eliminate any differences in assay timing. The release of lactate dehydrogenase (LDH) into the culture medium was also measured on day 10. LDH is normally present only in small amounts in the culture medium but increases in concentration when released from dead or dying cells. The LDH concentration was measured according to the manufacture's protocol (Lactate Dehydrogenase Activity Assay Kit, Sigma #MAK066-1KT). Briefly, 50 μL of culture medium (supernatant) was gently removed from each well and then transferred to a new clear-bottom plate. Assay reagents were then added to reach a final volume of 100 μL. The plates were incubated at 37 °C in an incubator for 30 min. The absorbance signal at 485 nm, which was proportional to the LDH concentration, was then measured using a plate reader (Tecan).

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.