Metabolism of Non-Enzymatically Derived Oxysterols: Clues from sterol metabolic disorders.

Cholestane-3β,5α,6β-triol (3β,5α,6β-triol) is formed from cholestan-5,6-epoxide (5,6-EC) in a reaction catalysed by cholesterol epoxide hydrolase, following formation of 5,6-EC through free radical oxidation of cholesterol. 7-Oxocholesterol (7-OC) and 7β-hydroxycholesterol (7β-HC) can also be formed by free radical oxidation of cholesterol. Here we investigate how 3β,5α,6β-triol, 7-OC and 7β-HC are metabolised to bile acids. We show, by monitoring oxysterol metabolites in plasma samples rich in 3β,5α,6β-triol, 7-OC and 7β-HC, that these three oxysterols fall into novel branches of the acidic pathway of bile acid biosynthesis becoming (25R)26-hydroxylated then carboxylated, 24-hydroxylated and side-chain shortened to give the final products 3β,5α,6β-trihydroxycholanoic, 3β-hydroxy-7-oxochol-5-enoic and 3β,7β-dihydroxychol-5-enoic acids, respectively. The intermediates in these pathways may be causative of some phenotypical features of, and/or have diagnostic value for, the lysosomal storage diseases, Niemann Pick types C and B and lysosomal acid lipase deficiency. Free radical derived oxysterols are metabolised in human to unusual bile acids via novel branches of the acidic pathway, intermediates in these pathways are observed in plasma.

Although the end products of 7-OC and 5,6-EC/3β,5α,6β-triol metabolism have been defined, the in vivo biochemical pathways generating these products have yet to be fully elucidated. To study further the metabolism of 7-OC and 5,6-EC/3β,5α,6β-triol we took advantage of plasma samples from patients where levels of these substrates are particularly high. We find that 7-OC, 7β-HC and 3β,5α,6β-triol fall into new branches of the acidic pathway of bile acid biosynthesis and in patient plasma where the concentration of these metabolites is high, almost all the necessary intermediates to bile acids are observed. Buildup of these intermediates may be responsible for some of the clinical features of diseases where free radical oxidation of cholesterol is prevalent and their measurement may have diagnostic value.

Materials
Oxysterol standards were from Avanti Polar Lipids Inc (Alabaster, Al, USA), 3β,5α,6β-triHBA was prepared as in Ref. [29], other bile acid standards were kindly donated by Dr Jan Sjövall of Karolinska Institute, Stockholm, or as detailed in Griffiths et al. [30]. Materials for liquid chromatographymass spectrometry (LC-MS) analysis were as in Refs. [31,32].  The upper panels show RICs of NPB patient and lower panels NPB carrier plasma. Oxysterols with a the 3β,5α,6β-triol function are labelled in blue, those with a 3β,7β-dihydroxy function in red. Concentrations are given in the righthand corner of chromatograms. GP derivatised oxysterols give syn and anti conformers, which may or may not be chromatographically separated. For quantification of 3β,5α,6β,26-tetrol and 3β,5α,6β,24-tetraHCa extended chromatographic gradients were exploited as described in Griffiths et al. [30].

Patient samples
All participants or their parents/guardians provided informed consent and the study was performed with institutional review board approval (REC08/H1010/63) and adhered to the principles of the Declaration of Helsinki.

LC-MS methods
The LC-MS methods have been described in detail elsewhere [30,31]. In brief, a charge-tagging method was adopted [31,32], where sterols, including oxysterols and bile acids, with a 3β-hydroxy group were oxidised with bacterial cholesterol oxidase (ChOx) to 3-oxo analogues and derivatised with [ 2 H 5 ]-labelled Girard P (GP) reagent ( Fig. 2), then analysed by LC-MS at high mass-resolution (120,000 at m/ z 400, full-width at half-maximum height definition) with parallel multistage fragmentation (MS n ). Sterols with a natural oxo group were derivatised with [ 2 H 0 ]GP reagent in the absence of cholesterol oxidase and analysed together with the [ 2 H 5 ]GP-derivatised sterols in a single LC-MS(MS n ) run. Quantification was by the isotope dilution method.

Metabolism of 7-OC and 7β-HC
The metabolism of 7-OC has been of interest to many investigators. Lyons and Brown reported that it could be metabolised to 26H,7O-C by CYP27A1 in HepG2 cells, while Heo et al. showed that 26H,7O-C and the down-stream CYP27A1 metabolite 3β-hydroxy-7-oxocholest-5-en-(25R)26-oic acid (3βH,7O-CA) could be formed by retinal pigment epithelial cells [40,41]. An alternative route for metabolism of 7-OC is reduction to 7β-HC by HSD11B1 as shown by Hult et al., Larsson et al.,. 7-OC itself can be formed via radical reactions from cholesterol or enzymatically from 7-DHC by CYP7A1 or from 7β-HC by HSD11B2 oxidation [4,9,34] (Fig. 1). We have reported a metabolic pathway from 7-OC to 7-oxo-and 7β-hydroxy-Δ 5 -bile acids in SLOS patients where the 7-DHC concentrations are high in plasma and tissue [22][23][24][25]. We next sought to investigate if this pathway is active in other diseases where levels of 7-OC are elevated.

Discussion
The observation here and elsewhere of elevated levels of 3β,5α,6βtriol and 7-OC in the lysosomal storage diseases NPC and NPB points to their endogenous formation [22]. While 7-OC can be formed enzymatically from 7-DHC in SLOS [43], there is little convincing evidence for the enzymatic formation of 5,6-EC, the precursor of 3β,5α,6βtriol, hence it is likely that it is formed through free radical reactions in vivo. The absence of high levels of 7-DHC in NPC and NPB also points to the formation of 7-OC via in vivo free radical reactions.

Formation of 3β,5α,6β-triHBA
Until recently little was known about the metabolism of 5,6-EC. It is established that 5,6-EC can be enzymatically converted to the 3β,5α,6β-triol by ChEH [35], but only in 2016 was it shown that the triol could be converted to the bile acid 3β,5α,6β-triHBA in man [28,29]. A similar product has also been proposed to be generated in rat [44,45], however, a description of the pathway from 3β,5α,6β-triol to 3β,5α,6β-triHBA has not previously been reported.
There are two major pathways of bile acid biosynthesis, the neutral and acidic pathways, and two more minor pathways, the sterol 24hydroxylase and 25-hydroxylase pathways [37]. The neutral pathway is initiated by 7α-hydroxylation of cholesterol by CYP7A1, but in the metabolism of 3β,5α,6β-triol the absence of a 7α-hydroxy group in the ultimate bile acid argues against operation of this pathway. The cholesterol 24-hydroxylase pathway is also an unlikely route for 3β,5α,6βtriol metabolism as this starts with a reaction catalysed by CYP46A1, predominantly expressed in brain [37]. This leaves the acidic and 25hydroxylase pathways for biosynthesis of 3β,5α,6β-triHBA from 3β,5α,6β-triol. As in the present study we have identified most of the intermediates in a novel branch of the acidic pathway from 3β,5α,6βtriol to 3β,5α,6β-triHBA in plasma, it is highly likely that this is the pathway followed (Fig. 4).

4.2.
Formation of 3βH,7O-Δ 5 -BA and 3β,7β-diH-Δ 5 -BA The metabolism of 7-OC has been the subject of greater interest than that of 3β,5α,6β-triol [40,41,46], as the former compound is found in atherosclerotic lesions [47]. 7-OC can be metabolised to 26H,7-OC and 3βH,7O-CA [40,41], hence it likely to also follow another branch of the acidic pathway to generate 3βH,7O-Δ 5 -BA. The data reported here confirms this hypothesis as judged by the preponderance of relevant pathway intermediates identified in plasma of NPC and NPB patients. 7-OC can alternatively be converted to 7β-HC [5][6][7][8] and this oxysterol may similarly fall into a related branch of the acidic pathway with the formation of 3β,7β-diH-Δ 5 -BA. These 7-OC and 7β-HC branches of the acidic may be interconvertible through HSD11B1 which can convert 7oxo to 7β-hydroxy groups and HSD11B2 which can catalyse the reverse reactions (Fig. 4). In fact, in an early study Alvelius et al. identified both 3βH,7O-Δ 5 -BA and 3β,7β-diH-Δ 5 -BA as their sulphate and glycine conjugates in urine of an NPC patient [26], although the plasma pattern of oxysterols was apparently normal. In the present study, the ratio of 7oxo-to 7β-hydroxy-sterols may provide an insight into the activity of HSD11B enzymes. HSD11B1 is the enzyme responsible for reducing 7-OC to 7β-HC in mouse and man [5,7], while HSD11B2 can also catalyse the reverse reaction [9]. The HSD11B1 enzyme also catalyses the reduction of cortisone to cortisol in man and of 11-dehydrocorticosterone to corticosterone in mouse, the reduced metabolites being ligands to the glucocorticoid receptor. On the other hand, HSD11B2 is the enzyme that oxidises cortisol to cortisone and has a similar activity towards 7β-HC and also 7β,25-dihydroxycholesterol [9,48,49]. Data from the present study indicates that as the acidic pathway proceeds the ratio of 7βhydroxy to 7-oxo metabolites increase, this is true for patients, controls and carriers. This may be explained by the HSD11B reductase having a dominant effect over the oxidase as the pathway descends.
In light of the low number of patient samples analysed the diagnostic value of bile acid precursors can only be speculated on. However, the high abundance of the C 27 acids 3β,5α,6β-triHCa, 3β,7β-diHCA ( Figure 3D) and 3βH,7O-CA (Fig. 5C) in NPC and NPB plasma suggests that these three acids in combination may diagnose these disorders. It may not be possible to distinguish these diseases from LALD as the three acids are elevated in this disorder also.

Conclusion
In man, most of the primary bile acids are synthesised in the liver through the neutral pathway of bile acid biosynthesis. The acidic pathway can proceed extrahepatically [37], while quantitatively less important in adult this pathway may be dominant in neonates [50]. We show here that novel branches of the acidic pathway, starting from oxysterols formed non-enzymatically, are important for the formation of unusual bile acids in patients with lysosomal storage disease. In healthy controls, the acidic pathway also proceeds to generate unusual 7β-hydroxy and 7-oxo bile acids but these are quantitatively minor.