26-Hydroxycholesterol. Identification and quantitation in human serum.

Using isotope dilution mass spectrometry, 26-hydroxycholesterol was identified in the serum of normal adults. Total values ranged from 9.2 to 25.6 micrograms/100 ml of which 31-35% was free sterol. Density gradient ultracentrifugation indicates that the steroid is distributed among the low and high density lipoproteins.


Cholest-5-ene-3P, 26-diol (26-hydroxycholesterol) is a major sterol component of human meconium (1). Its presence in ~o,dgical fluids after neonatal life has not been established. A number of recent findings (2-6) imply that hydroxysterols have an important biologic role. Thus, using cell culture techniques, hydroxycholesterols have been shown to reduce cholesterol synthesis by inhibition of 3-hydroxy-3-methylglutaryl coenzyme A reductase (4), stimulate esterification of cholesterol in liver cells (2), and inhibit thymidine incorporation into DNA in lymphocytes
). In the latter study (6), a fraction isolated from normal human low density lipoprotein, referred to as LDL'-In, was also shown to have similar activity. For these reasons and because of our interest in the metabolism of 26-hydroxycholesterol (7, €9, we have developed an isotope dilution mass spectrometric method for the quantitation of free and esterified 26-hydroxycholesterol.

Cholest-5-ene-3fi,26-diol-
The compound and its tritiated and deuterated analogs were prepared from kryptogenin using minor modifications of the synthesis reported by Scheer et al. (9) and described previously in detail in this journal (7). Use of D,O and DCI during the Clemmensen reduction step introduced deuterium at the C-16 = 0 and C-22 = 0 groups of kryptogenin and adjacent carbons (10) to provide a series of deuterated analogs. Analysis of the mass spectra of the deuterated and natural compounds as the di-trimethylsilyl ethers using a Hewlett-Packard GLC-MS instrument Model 5992B (Fig. 1) indicated a range of deuterated 26-hydroxycholesteroIs varying from 4 to 11 mass units greater than the natural abundance ( m / z -546) with a preponderant analog at m/z-554. No m/z-546 could be detected in the recrystallized deuterated 26-hydroxycholestero1 which was used for the quantitative estimation of 26-hydroxycholestero1 in serum.
* This work was supported by Grant AM-16201 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
+ To whom reprint requests should be addressed. It was found that 26-hydroxycholesterol could be isolated from plasma or serum in sufficiently pure form for GLC-MS analysis by separation on Glycophase G controlled pore glass, 80-100 mesh (Pierce Biochemical). Use of this packing for the separation of bile acid methyl esters was recently described in this journal (11). For separation of sterols, 25 g of the packing are suspended in a mixture of hexane/ethyl acetate, 9:l (v/v), and slurried into a glass column (1.5 cm internal diameter X 17.5 cm). The flow rate is adjusted to 1.8 ml/min. Fractions containing 26-hydroxycholestero1 are combined and reduced to dryness.
For plasma or serum analysis, 5 pg of deuterated analog were added to 2.5 ml and the protein was precipitated by addition of 18.0 ml of 2,2-dimethoxypropane (Aldrich) and 0.12 ml of 12 N HCl (12). After centrifugation, the filtrate was decanted and evaporated to dryness with heating (37 "C) aided by a stream of nitrogen.
Serum was fractionated into VLDL, LDL, and HDL by density ultracentrifugation using KBr following precisely the methods utilized by Curtiss and Edgington ( 6 ) . Aliquots of 2.5 ml were taken for 26hydroxycholesterol analysis as described above.
For analysis of esterified 26-hydroxycholestero1, the sterol fraction was saponified using an established procedure for the saponification of cholesterol (13). To the residue, 3.0 ml of 0.38 M KOH in 94% ethanol (v/v) was added and heated at 50-55 "C for 60 min. The volume was then reduced to approximately 2.0 ml and 2.0 ml of water was added. After extraction with diethyl ether and backwashing to neutrality, the sterol fraction was taken to dryness and then dissolved in 1.6 ml of hexane/ethyl acetate, 91, and applied to the Glycophase G column. Utilizing [16-22-"H]26-hydroxycholesterol prepared from kryptogenin (7) in tracer amounts added to serum, it was found that recovery after protein precipitation was complete and after column chromatography ranged from 68 to 83%. In contrast, it was found that the disulfate of [ 16-22-3H]26-hydroxycholesterol precipitated with proteins and could be recovered as the free sterol by resuspension of the precipitate in tetrahydrofuran (97.54) containing 0.04 ml of 70% perchloric acid and allowing to remain at room temperature for 72 h. In applying this procedure to the precipitate from serum, we failed to detect additional 26-hydroxycholesterol.
The eluate from the Glycophase G column was silylated and injected on a 4-ft column (2 mm internal diameter) of SP2250 (Supelco Inc., Bellafonte, PA) at 260 "C. At this temperature, 25-hydroxycholesterol and 26-hydroxycholestero1 have retention times of 14.0 and 16.3 min, respectively. GLC-MS analysis of a serum to which no deuterated or tritiated 26-hydroxycholestero1 was added is depicted in Fig. 2. As shown in the upper panel, monitoring for m/z-546 shows no masses until a peak appears at 15.9 min. A complete spectrum taken at 16.3 min (lowerpanel) shows all the fragments characteristic of 26-hydroxycholesterol. Since Glycophase G column chromatography does not separate 25-from 26-hydroxycholestero1, the absence of a peak at 14 min indicates little or no compound in plasma. Also, the fragmentation pattern of 25-hydroxycholesterol has a prominent m/ 2-131 peak, representing cleavage between C-24 and C-25 and a diminutive m/z-129 peak which is the opposite of the fragmentation pattern of 26-hydroxycholestero1 (14) and readily distinguishes the two compounds. In preliminary studies, recrystallized cholesterol and cholesterol oleate in amounts 10-fold greater than present in 2.5 ml of serum were subjected to protein precipitation, saponification, column chromatography and GLC-MS analysis. Neither 25 nor 26-hydroxycholesterol could be detected indicating they were not derived from the starting compounds. Frc. 2. GLC-MS analysis of 26-hydroxycholestero1 in serum. A 2.5-ml aliquot of serum (from P. M., Table I) was fractionated on a previously standardized Glycophase G column without the addition of a deuterated analog after protein precipitation, saponification, and ether extraction. The eluate from the column was converted to the di-trimethylsilyl ether and injected onto a 4-foot column of 3% SP2250 a t 260 "C. The instrument was set to monitor the m/z-546 (molecular ion). As shown a t top, a single peak was obtained with a retention time of 16.3 min. A complete spectrum obtained at this time contained all the peaks characteristic of 26-hydroxycholestero1 (bottom). 0.313, and 0.156 p g of 26-hydroxycholesterol to 5 p g of the deuterated compound to give known m/z-554/546 ratios of 4, 8, 16, and 32, respectively. Analyeis of these standards by GLC-MS yielded observed m/z-554/546 ratios of 1.96 & 0.40 S.D., 3.49 f 0.98, 9.85 f 1.2, and 18.6 +-2.3 uncorrected for proportion of D, enrichment in the deuterated analog. A complete set of standards was run with each group of serum samples and a standard curve was constructed from the weighed and observed m/z-5541546 ratios as described by Biemann (15) from which the serum values could be calculated. Addition of 26-hydroxycholestero1 to 2.5 ml of serum ranging in amounts from 0.115 to 2.14 pg gave recovery of 84% * 10.9 S.D. Monitoring of m/z-554/553, -554/552, and -553/552 ratios established that no deuterium was lost during the processing of the serum. Table I, the total 26-hydroxycho~estero1 in serum in 8 normal individuals ranged from 9.2 to 25.6 pg/lOO ml. In 3 individuals, the free sterol ranged from 31 to 35%. No 26-hydroxychoIestero1 was found in the very low density lipoproteins. Both free and esterified 26-hydroxycholesterol are found in the low density and high density lipoproteins.

As shown in
The findings establish that 26-hydroxycholesterol is present throughout life. Although it is present as the disulfate in meconium (l), it can also be metabolized to both chenodeoxycholic and cholic acids (8) and is the likely source of 3phydroxy-5-cholenoic acid, also present in meconium (16) and urine as a sulfate (17).
Frederickson and Ono initially identified both 25-and 26hydroxycholesterol as metabolites of cholesterol after incubation with mouse liver mitochondria (18). Danielsson (19) confiied the enzymic origin of 26-hydroxycholestero1 but considered 25-hydroxycholestero1 to be an autooxidation product.
Extensive studies by Van Lier and Smith (20, 21) established that 25-hydroxycholesterol is an autooxidation product and that 26-hydroxycholestero1 in human atheromata is of enzyme origin. A recent report (22) that air oxidation of thin films of cholesterol on glass plates at 70 "C for 6 months yielded 0.1% 25-hydroxycholestero1 and no detectable 26-hydroxycholesterol further supports the view that 26-hydroxycholesterol is an unlikely autooxidation product. In the present studies as part of the development of the serum method, 12.5 mg of cholesterol were carried through the entire procedure without any detectable 26-hydroxycholestero1 being found by mass spectrometry. Also, the VLDL separated by ultracentrifugation and carried through the entire procedure had no detectable 25-or 26-hydroxycholesterol. The absence of 25-hydroxycholesterol in our processed samples, which would have been detected a t 14.0 min as an m/z-546 peak, implies that the 26-hydroxycholesterol found in serum was not generated by the methods used. We concur therefore, with the view of Van Lier and Smith that the 26-hydroxycholesterol present in vivo is of enzymic origin and that our more sensitive isotope dilution mass spectrometric method establishes its presence in serum. This interpretation is fully in keeping with the recent report that the C 2 7 steroid 26-hydroxylase in human liver mitochondria is the major determinant of normal side chain oxidation of cholesterol to CZ4 bile acids (23).
In vitro studies thus far have utilized mostly 25-hydroxycholesterol because it is more readily available. However, Dr. Donald McNamara has found that 26-hydroxycholesterol prepared as described (7) is at least as active as 25-hydroxycholesterol in the inhibition of cholesterol synthesis in vitro.' In addition, the presence of 26-hydroxycholesterol in LDL raises the possibility that it may be the mediator of its inhibitory effects in cell culture.
Brown et al. (24) developed the concept that the normal regulation of cholesterol synthesis via the activity of hydroxymethylglutaryl-Co A reductase is dependent on the internalization of cholesterol obtained from the binding of LDL to a receptor. on the cell surface. Kandutsch et al. (25) propose that the actual mediator is an oxygenated sterol rather than cholesterol. Our studies establish that 26-hydroxycholesterol is present in biologic fluids and therefore could play a role in the regulation of cholesterol metabolism. With the methods that have been developed, this possibility can be explored rigorously.