7-Ketocholesterol

The cholesterol autoXidation products, 7-ketocholesterol and 7β-hydroXycholesterol are associated with serum neurofilaments in multiple sclerosis

Abstract

Background: Serum neurofilament light chain (sNfL) is an established marker of neuroaxonal injury in multiple sclerosis (MS).

Objectives: To investigate if oXysterols produced from non-enzymatic and enzymatic cholesterol oXidation are differentially associated with sNfL measurements in MS.

Methods: This longitudinal study included 62 relapsing-remitting (RR-MS) and 36 progressive MS (PMS) patients with baseline and 5-year follow-up measures of serum levels of 6 oXysterols, sNfL and lipids. The oXysterols, 24- hydroXycholesterol (24HC), 25HC, 27HC, 7αHC, 7βHC and 7-ketocholesterol (7KC), were measured using liquid chromatography-mass spectrometry. sNfL was measured using single molecular array assay. Serum high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C) levels were obtained from a lipid profile.

Results: The enzymatically produced oXysterols 24HC, 25HC, 27HC and 7αHC were not associated with sNfL. However, baseline levels of reactive oXygen species (ROS) produced oXysterols, 7KC (p = 0.032) and 7βHC (p = 0.0025), were positively associated with sNfL levels at follow-up. Follow-up 7KC (p = 0.038) levels were also associated with follow-up sNfL levels. The associations of 7KC or 7βHC with sNfL remained significant after adjusting for LDL-C or HDL-C.

Conclusions: 7KC and 7βHC, produced by ROS-mediated cholesterol oXidation are associated with neuroaxonal injury as assessed by sNfL in MS.

1. Introduction

Multiple sclerosis (MS) is a chronic disease of the central nervous system (CNS) characterized by blood brain barrier (BBB) breakdown, inflammation and lesion formation, demyelination and neuro- degeneration (Frohman et al., 2006).Cholesterol is a major structural component of myelin (Dietschy and Turley, 2004) and cholesterol dyshomeostasis is emerging as an important factor in MS pathophysiology (Hussain et al., 2020) and dis- ease progression (Fellows et al., 2015; Murali et al., 2020; Wein- stock-Guttman et al., 2011; Flauzino et al., 2019). However, the molecular and physiological mechanisms for the associations of cholesterol biomarkers with MS disease progression are not well un- derstood. OXysterols are biologically active cholesterol metabolites that regulate CNS cholesterol homeostasis (Jeitner et al., 2011). OXysterols are produced by both non-enzymatic and enzymatic cholesterol oXida- tion. Reactive oXidative species (ROS), which are present in excess under conditions of oXidative stress, are responsible for the non-enzymatic oXidation of cholesterol to 7-keto-cholesterol (7KC) and 7β-hydroX- ycholesterol (7βHC) (Duc et al., 2019). 7KC and 7βHC are considered biomarkers of oXidative stress and cause neuroinflammation (Testa et al., 2016) and cellular cytotoXicity in vitro (Nury et al., 2013). CNS cells are particularly susceptible to 7KC and 7βHC cytotoXicity (Testa et al., 2016; Nury et al., 2013). For comparison, we measured four additional enzymatically synthesized oXysterols: 24 hydroXycholesterol (24HC), 25 hydroXycholesterol (25HC), 27 hydroXycholesterol (27HC) and 7α hydroXycholesterol (7αHC). 24HC regulates cholesterol synthe- sis, and mediates cholesterol elimination from the brain (Lund et al., 2003; Bjorkhem et al., 1997; Lutjohann et al., 1996), 25HC is a mediator of inflammatory signaling, 7αHC is the product of the rate-limiting step in cholesterol elimination by CYP7A1; and 27HC is produced by CYP27A1 in the acidic pathway (Chawla et al., 2000).

Based on this background, we aimed to investigate the associations of non-enzymatic and enzymatically-derived oXysterols with the serum neurofilaments (sNfL) levels, an established marker of neuroaxonal injury in MS (Disanto et al., 2017; Preziosa et al., 2020). Our working hypothesis was that 7KC and 7βHC levels would be associated with greater neuroaxonal injury compared to 24HC, 25HC, 27HC and 7αHC
because they reflect ROS-mediated cholesterol oXidation via oXidative stress pathways and have cytotoXic effects.

2. Methods

2.1. Study population

Study Setting: Longitudinal, prospective observational study at the MS Center of the State University of New York at Buffalo
Clinical Study Design: The study utilized blood samples and data from a larger cardiovascular, environmental, and genetic risk factors study in MS (CEG-MS). The samples have been previously used to investigate the role of apolipoprotein AI and E and neuroaxonal injury (McComb et al., 2020). Additionally, our group has used the samples and data from these patients to investigate oXysterols and apolipopro- teins and their associations with MS disease progression (Fellows Maxwell et al., 2019). However, the inter-dependencies between oXy- sterols and sNfL have not yet been investigated.

At baseline and follow-up, MS patients underwent neurological ex- amination and EXpanded Disability Status Scale (EDSS) scores were obtained (Meyer-Moock et al., 2014). Subjects provided blood samples at baseline and at the 5-year follow-up visit. Patients were assessed during remissions and should not have experienced exacerbations or received intravenous corticosteroids within 30 days of clinical evalua- tions or blood sampling.

At baseline and 5-year follow-up, progressive MS (PMS) were re- ported as primary progressive (PP-MS) or secondary progressive (SP- MS) by an MS specialist according to classification of patients’ clinical characteristics based on published disease course criteria (Lublin et al., 2014). This sub-study included relapsing-remitting MS (RR-MS) and PMS patients with sNfL and oXysterol measurements.MS patients younger than 18 years old, healthy controls, clinically isolated syndrome patients, neuromyelitis optica spectrum disorder patients or patients with other neurological diseases were excluded. Informed Consent: The study protocol was approved by the Uni- versity at Buffalo Human Subjects Institutional Review Board and written informed consent was obtained from all participants.

2.2. Oxysterol assay

EDTA plasma samples were stored at 80 ◦C until use without prior freeze thaw cycles. Analysts were blinded to clinical status of samples. Total plasma oXysterols, inclusive of free and esterified oXysterols, were measured by room temperature saponification, isolated by solid phase extraction and analyzed by LC-MS by a modification of our method previously described in detail (Narayanaswamy et al., 2015). In brief, samples, matriX based calibrators and quality controls materials (200 µl) were vortex-miXed with 100 µl of deuterated internal standard miX (150 ng/ml each of vitamin D3 (d3), 22HC (d7), 7αHC (d7), 7KC (d7)) and 10 µl of 50 mg/dl ethanolic butylated hydroXytoluene. Sam- ples were saponified with 875 μl of 0.5 M ethanolic KOH for 3 h at room temperature, under argon. The pH was neutralized and the samples were loaded onto HyperSep C18 solid phase extraction cartridge (200 mg, 3 mL) that had been pre-conditioned with 1 ml of hexane: isopropanol (50:50 v/v), 1 ml of methanol and 2 ml of water, sequentially. Polar lipids were eluted off the SPE column using 4 ml of methanol: water (75:25 v/v). Non-polar sterols (including cholesterol and oXysterols) were eluted using 2 ml of hexane: isopropanol (50:50 v/v). The eluate was evaporated under nitrogen, reconstituted in 300 μl of methanol: water (90:10, v/v), and 75 μl injected for LC-MS analysis. OXysterols were analyzed on a Shimadzu Scientific (Columbia, MD) LCMS-2010A mass spectrometer system with APCI interface in positive ion mode. Mobile phase composition was 100% methanol to pump A and meth- anol: water (50:50 v/v) to pump B, both containing 0.1% formic acid. OXysterols were separated on a Supelcosil LC-18-S, 10 cm X 3.0 mm, 3 μm column (Sigma Aldrich, St. Louis, MO). Column oven temperature was set at 10 ◦C and flow rate was at 0.75 ml/min. The mobile phase gradient was 80% A for 10 min, linear increase to 100% A over 5 min and held for 13 min followed by re-equilibration at 80% A for 6 min; total run time was 34 min. Temperature settings for the MS were: interface at 400 ◦C, CDL at 230 ◦C and heat block at 200 ◦C. Nitrogen was used as a nebulizing gas for the ion source at a flow rate of 2.5 L/min. Data was acquired in time segmented- single ion monitoring (SIM) manner to achieve maximum sensitivity. In time segment 1, 22-HC(d7) at 374.30 m/z was used as internal standard for quantifying 24HC and 25HC at 367.30 m/z and 27HC at 385.30 m/z. In time segment 2, 7αHC(d7) at 374.30 m/z was used as the internal standard for quantifying 7αHC and 7βHC at 367.30 m/z. In time segment 3, 7-KC(d7) at 408.40 was used as internal standard to quantify 7KC at 401.40 m/z.

2.5. Statistical analysis

SPSS (IBM Inc., Armonk, NY, version 24.0) statistical program was used. A p-value 0.05 was considered statistically significant.
SP-MS and primary progressive patients were categorized as PMS in the RR-MS vs. PMS disease course status variable. Cholesterol oXidation products (COP) were defined as the sum of 7KC and 7βHC levels. Disease-modifying therapies were categorized into no treatment, interferon-beta, glatirmaer acetate, natatlizumab, orals and others. The Interferon-beta category included AVONEX®, REBIF®, BETASERON®, EXTAVIA® and PLEGRIDY. Dimethyl fumarate, fingolimod and teri- flunomide were categorized as Orals. Body mass index (BMI) was calculated based on the subjects’ height and weight. Quartiles of oXy- sterols were calculated using the SPSS visual binning tool.
OXysterols (24HC, 25HC, 27HC, 7αHC, 7βHC and 7KC and COP) and sNfL were log-transformed (base 10) to reduce skew.

The differences in demographic and clinical between the RR-MS and variables (sex, statin use), the Mann-Whitney was used for baseline and follow-up EDSS and the χ2 test for baseline and follow-up disease modifying therapy (DMT). Multiple linear regression analyses were conducted with baseline or follow-up sNfL and treated as dependent variables. Regression analyses included the oXysterol of interest (either 24HC, 25HC, 27HC, 7αHC, 7βHC, 7KC or COP) as predictor variables of interest, and adjusted for age, sex, body mass index (BMI) and baseline progressive MS disease status (RR-MS or PMS) (McComb et al., 2020). Associations between dependent and predictor variables were identified using the partial correlation coefficient (rp), which assess the strength and direction of the association between the independent variable and a given predictor variable after adjusting for the other predictor variables. The partial correlation coefficient values were obtained from the multiple regres- sion using methods embedded in the SPSS software package. The partial correlation for a given predictor variable was considered significant if its regression slope was significant. In additional regression analyses, LDL-C or HDL-C were included as an additional predictor variable in the preceding framework.

Multiple linear regression analysis was used to determine the asso- ciations of baseline, follow-up or change in ApoE as dependent variables with of oXysterols. The regressions included the oXysterols of interest (baseline or follow-up of either 7KC, 7βHC, or COP) as a predictor var- iable and were adjusted for age, sex, body mass index, baseline pro- gressive MS disease status (RR-MS or PMS).
Two-way ANOVA adjusted for age, sex and body mass index was used to compare the levels of sNfL at follow-up with the lower and highest quartiles of baseline 7KC, follow-up 7KC, baseline 7βHC and baseline COP for baseline progressive MS disease status (RR-MS or PMS).

3. Results

Table 1 summarizes the clinical and demographic characteristics of the study sample which was comprised of 62 RR-MS and 36 PMS patients.
The RR-MS group had lower baseline age (mean: 44.5 SD 11.3 years in RR-MS vs. 56.4 6.11 years in PMS, p < 0.001, independent sample t-test) and baseline EXpanded Disability Status Scale scores (EDSS, median: 2.50, IQR: 1.50 in RR-MS vs. median: 5.00, IQR: 2.80 in PMS, p < 0.001, Mann-Whitney U test) compared to the PMS group that are representative of the respective MS disease course. At baseline, the PMS group was comprised of 29 secondary progressive MS patients and 7 primary progressive MS patients. The female to male ratio in the RR- MS and PMS groups were similar (p = 0.26, Fisher exact test). Dimethyl fumarate, fingolimod and teriflunomide were categorized as Orals. Interferon-beta includes AVONEX®, REBIF®, BETASERON®, EXTAVIA® and PLEGRIDY. Supplementary Table 1 summarizes the oXysterol profile (24HC, 25HC, 27HC, 7KC, 7αHC, 7βKC and COP), cholesterol profile (TC, LDL- C, HDL-C, ApoE) and sNfL measurements. The changes over time and differences between the RR-MS vs. PMS for these biomarkers have been reported in our publications (McComb et al., 2020; Fellows Maxwell et al., 2019). Additionally, previous work has characterized the associ- ations of sNfL levels with MRI measures of brain atrophy and lesion formation (Jakimovski et al., 2019). It is therefore necessary to deter- mine the associations of sNfL levels with oXysterols in RR-MS and PMS patients. Table 2 summarizes regression results adjusted for age, sex, BMI and baseline progressive MS disease status (RR-MS or PMS) for the associ- ation of each of the oXysterols with sNfL levels. The oXysterols, 24HC, 25HC, 27HC and 7αHC, which are produced by enzymatic cholesterol oXidation, were not associated with sNfL. Baseline 7KC (p = 0.032) and 7βHC (p = 0.0025) were associated with sNfL at follow-up. Baseline COP (p 0.0037), which is the sum of 7KC and 7βHC, was also associated with sNfL at follow-up. Follow-up 7KC (p 0.038) levels were associated with sNfL at follow-up. Baseline sNfL levels were not associated with baseline levels of any of the oXysterols (all p > 0.22, data not shown). Supplemental Table 2 summarizes the partial correlation results for the age, sex, BMI, and baseline progressive MS disease status (RR-MS or PMS).

The mean values of sNfL for the lower quartiles vs. highest quartile of associated with neuroaxonal injury as assessed by sNfL measurements in MS patients. Higher levels of baseline 7KC and 7βHC, which are pro- duced from ROS-mediated oXidation of cholesterol, were associated with increased levels of sNfL at follow-up. The four enzymatically- were not abrogated upon adjusting for LDL-C and HDL-C.

Studies with human endothelial cells and fibroblasts (Lizard et al., 1999), lymphocytes (Miguet et al., 2001), tumoral and human brain cells (Nury et al., 2013) and mouse models (Indaram et al., 2015; VejuX et al., 2020) have demonstrated that 7KC and 7βHC, are the primary oXysterols responsible for apoptosis and cellular toXicity. Additionally, 24HC exhibits a protective activity in preventing cell death from 7KC toXicity (Okabe et al., 2013). Interconversion between 7KC and 7βHC has also been characterized in rodent liver microsomes (Larsson et al., 2007). Thus, the associations we found are consistent with the possi- bility that plasma 7KC and 7βHC levels are surrogate biomarkers that reflect cholesterol oXidation during the ROS-mediated injury causing neuroaxonal injury.
ROS are short-lived but powerful oXidizing agents that react rapidly with cellular lipids, proteins and DNA and compromise the integrity of physiological processes. OXidative stress results from an imbalance be- tween the damaging effects of endogenous ROS and the detoXifying ef- fects of antioXidant defense molecules that eliminate or scavenge ROS and repair biomolecules damaged by ROS. Increased oXidative stress is an effector mechanism of many pathological processes including mito- chondrial dysfunction, apoptosis, neurodegeneration and tissue aging follow-up 7KC (p 0.022) remained significant upon inclusion of LDL-C as a predictor. Likewise, the associations of follow-up sNfL with baseline levels of 7KC (p 0.018), 7βHC (p 0.0023) and COP (p 0.0020) and follow-up 7KC (p 0.041), remained significant upon inclusion of HDL-C as a predictor.

In our previous work, we have found that sNfL levels at follow-up were positively associated with serum ApoE (McComb et al., 2020). We therefore explored the association of 7KC, 7βHC and COP with ApoE. (Table 4). We did not obtain evidence for associations between any of baseline oXysterol levels assessed with follow-up ApoE. Baseline 7KC (p 0.0083), baseline 7βHC (p 0.0050) and baseline COP (p 0.0043) were associated with baseline ApoE levels. Additionally, change in 7KC (p 0.0031) and change in COP (p 0.030) were associated with change in ApoE from baseline to follow-up.

These results demonstrate that 7KC and 7βHC, which are produced by oXidative stress-mediated cholesterol autoXidation, are associated with neuroaxonal injury. These associations remained significant after adjusting for LDL-C or HDL-C. The enzymatically-produced oXysterols 7αHC, 24HC, 25HC and 27HC were not associated with neuroaxonal injury.

4. Discussion

In this work, we investigated whether oXysterols produced from non- biomarkers such as oXidized LDL-C, malondialdehyde and 4-hydroXyno- nenal are prominent in early and actively demyelinating MS lesions (Newcombe et al., 1994). Oligodendrocytes and neurons are also oXidatively injured during MS demyelination and axonal injury (Haider et al., 2011). sNfL levels can distinguish MS patients from healthy con- trols and MS patients with contrast-enhancing lesions from those without contrast-enhancing lesions (Disanto et al., 2017). sNfL is a predictor of brain atrophy and disease worsening on EDSS (Barro et al., 2018; Siller et al., 2018). sNfL are increased in patients with clinically isolated syndrome (Disanto et al., 2016) and additionally, have been shown to be prognostic markers predict the conversion of individuals with radiologically isolated syndrome to clinically definite MS (Matu- te-Blanch et al., 2018). In our work, baseline sNfL was associated with T1, T2 and gadolinium lesion volumes. Baseline sNfL was also associated with longitudinal changes in whole brain, gray matter, deep gray matter and thalamic volume changes over 5-year follow-up (Jakimovski et al., 2019). We did not find associations between oXysterols and any of the MRI measures.

In this study, we investigated the associations of 7KC and 7βHC with ApoE based on the finding in our prior work that ApoE was associated with sNfL (McComb et al., 2020). We found that changes in 7KC and 7βHC were also associated with changes in ApoE. Lipoproteins do not enter the CNS and ApoE, the predominant apolipoprotein of the CNS, contributes to cholesterol transport (Rebeck, 2017). The associations of ApoE with sNfL are intriguing because it is possible that ApoE and NfL are released concomitantly from the CNS when neuroaxonal injury oc- curs. However, serum ApoE is also modulated by changes in peripheral lipid homeostasis, oXysterols (Fellows Maxwell et al., 2019) and cyto- kines such as transforming growth factor-beta (McComb et al., 2020). In human adipocytes, ApoE is induced by oXidative stress and protects against ROS-mediated injury (Tarnus et al., 2009). ApoE deficient mice exhibit increased oXidative stress, inflammation and neuroaxonal injury after spinal cord injury (Yang et al., 2018). Thus, the associations of additional mechanistic dots in the pathophysiology of neuroaxonal injury in MS.

24HC and 27HC, which are produced from cholesterol oXidation by the cytochrome P450 (CYP) enzymes, CYP46A1 and CYP27A1 (Testa et al., 2016), contribute to the maintenance and cholesterol homeostasis in the brain by regulating its effluX and uptake, respectively (Bjorkhem et al., 2019). Interestingly, we did not find evidence for associations between 24HC or 27HC with sNfL.

Cholestane-3β,5α,6β-triol (CT) is “tri-hydroXylated” metabolic product of cholesterol oXidation that is increased with 7KC in Niemann- Pick type C, cerebrotendinous Xanthomatosis and lysosomal acid lipase deficiency (Boenzi et al., 2016; Pajares et al., 2015). We did not obtain CT levels because our analytical methods, which were optimized to resolve “mono-oXygenated” oXysterols, particularly 7KC vs. 7βHC and 24HC vs. 25HC (Supplementary Figure), were incapable of assessing the CT due to its relative polarity compared to the other oXysterols. We usage as a covariate (results not shown). This suggests that statin usage was not associated with baseline or follow-up sNfL. However, it is plausible that we did not obtain evidence for statin effects because these were obscured. For example, if statin-treated group was hypercholes- terolemic and became normocholesterolemic, this would make it diffi- cult to distinguish from the group that did not receive statin treatment. There were minor differences in the association patterns of 7KC, 7βHC and COP with follow-up sNfL levels (Table 2), e.g., follow-up sNfL was associated with both baseline and follow-up 7KC; however, corre- sponding associations were found only for baseline 7βHC and COP. The exact reasons for these differences are not known. However, it is possible that underlying differences the underlying chemistry at the 7-position of cholesterol and the pathophysiological milieu could be potential contributing factors. The levels of 7βHC are lower than 7KC possibly because 7KC can be produced as a result of oXidation of 7βHC. Thus, 7KC may be preferentially produced under conditions where the oXidative stress is more severe. We did not find associations between EDSS and 7KC, 7βHC and COP.

We did not obtain evidence for associations of 7KC, 7βHC and COP with EDSS. EDSS reflects the accumulation of disability due to the MS disease process, whereas sNfL is a surrogate marker for ongoing neu- roaxonal injury occurring in central nervous system. 7KC and 7βHC in contrast are biomarkers associated with oXidative stress that are likely secondary to the inflammatory processes in MS, and relatively more transient and distal from the irreversible accumulation of neuro- degeneration responsible for disability. Also, our ordinal regression analyses of EDSS (results not shown), included the progressive MS disease status (RR-MS or PMS) as a predictor, which may have reduced the power to detect the associations of oXysterols with EDSS.

In conclusion, our results provide evidence for a possible role for 27‑dione emerged between 3.5 and 4 min (Supplementary Figure 1) and the remaining “mono-oXygenated” cholesterol metabolites eluted much later. Nevertheless, further investigation of tri-hydroXylated cholesterol
metabolites may provide additional insight.