Liquid Chromatography-Mass Spectrometry of Sterols
KeywordsBile Acid Multiple Reaction Monitoring Atmospheric Pressure Chemical Ionization Cholesterol Oxidase Multiple Reaction Monitoring Transition
According to the Lipid Maps classification system (Fahy et al. 2005) sterols are a class of compounds based on the cyclopentanoperhydrophenanthrene skeleton and include cholesterol and its cyclic precursors, steroid hormones, bile acids, and the ring opened secosterol vitamin D3.
As for other components of the “lipidome” the “sterolome” is extremely complex with a potentially huge number of distinct chemical entities. The complexity is magnified by variability of stereochemistry resulting in diastereoisomers with differing biological properties. This chapter will concentrate on sterol molecules found in the mammalian system, but similar LC-MS methods will be appropriate for sterols derived from plants and microbes.
Extraction Prior to LC-MS
There are many different extraction methods for sterols each tailored to a specific sterol-type and biological matrix. In terms of a global sterolomic analysis, extraction from tissue (or fluids) into absolute ethanol with homogenization and/or ultrasonication is recommended (Griffiths and Sjovall 2010). Cholesterol and the most nonpolar sterols can then be removed/isolated by extraction on a reversed-phase C18 solid phase extraction (SPE) column following dilution to and loading in 70% ethanol (Griffiths and Sjovall 2010), or alternatively on a straight-phase SPE column after solvent exchange into, e.g., toluene or hexane-dichloromethane (2:8; v/v). Cholesterol will elute from a C18 column in absolute ethanol (Griffiths and Sjovall 2010). Using straight-phase silica SPE sterols loaded in toluene can be fractionated into sterylester-rich, cholesterol-rich, and oxysterol-rich fractions by elution with hexane, 0.1% propan-2-ol in hexane, and 30% propan-2-ol in hexane, respectively (McDonald et al. 2007; Dzeletovic et al. 1995), or alternatively on a silicic acid column after loading in hexane-dichloromethane (2:8; v/v) with elution in hexane-dichloromethane (2:8; v/v) followed by ethyl acetate (Karu et al. 2007). Ethanol is preferred for general sterol extraction as it is a good solvent for sterols, steroids, and bile acids. The efficiency of extraction can be studied; however, determination of recoveries of added reference compounds, radioactive or not, is unsatisfactory as added compounds are not distributed in the sample as endogenous compounds. The only satisfactory method with minimal bias is repeated extractions of the pellet after ethanol extraction.
Sterols are present in biological systems both as the free molecules and as conjugates with fatty acids, sulfuric acid, sugars, including glucuronic acid, and the amino acids glycine and taurine. While sterols with a polarity-enhancing conjugating group will be extracted into ethanol, sterols esterified with fatty acids may be better extracted into a mixture of ethanol and propan-2-ol. Potassium hydroxide can then be added for saponification of esters. Following incubation and neutralization with, e.g., glacial acetic acid, sterols can be fractionated by SPE (Griffiths and Sjovall 2010). The saponification procedure can lead to loss of labile groups, e.g., dehydration of 7α-hydroxy-4-en-3-ones (Saeed et al. 2014) and also introduces the risk of autoxidation (Schroepfer 2000). It is recommended that the products of saponification are analyzed independently from a nonsaponified sterol sample.
Global LC-MS Methods for Sterol Analysis
An important consideration for the LC-MS analysis of cholesterol and more hydrophobic sterols is solubility. If not properly solubilized cholesterol will not be efficiently retained on a reversed-phase sorbent, the consequence of which is that it will elute earlier than expected and contaminate proceeding fractions. This is problematic as cholesterol is usually very much more abundant than its precursors or metabolites. Solvents recommended for cholesterol and similar sterols are methanol, ethanol, propan-2-ol, and combinations, thereof. Cholesterol is soluble in 70% ethanol, so caution should be taken with the use of more polar solvents.
Depending on the study, it may be beneficial to analyze specific classes of sterols individually, or alternatively take a more global lipidomic approach. If the latter is the aim neutral sterols can be fractionated according to polarity and separated from acidic metabolites by mixed mode anion exchange chromatography, then the most nonpolar neutral fraction separated by, e.g., isocratic reversed-phase LC with, e.g., a 75% methanol 25% propan-2-ol mobile phase. The more polar neutral fraction can then be separated by reversed-phase LC using a gradient system (e.g., A: 60% ethanol, 0.1% formic acid, B: propan-2-ol, 0.1% formic acid) (Theofilopoulos et al. 2013). Using the same column acidic sterols can be separated using a gradient (e.g., A: 10 mM ammonium acetate, pH 7.2, in 5% methanol, B: 10 mM ammonium acetate, pH 7.2, in 95% methanol).
Targeted LC-MS Methods for Sterol Analysis
Nonpolar Neutral Sterols
Nonpolar neutral sterols present a challenge for LC-MS analysis in that mammalian samples tend to be dominated by cholesterol making observation of similarly nonpolar minor sterols difficult. A further issue to consider is that nonpolar sterols tend to ionize poorly by atmospheric pressure ionization (API) methods, often suffer in-source dehydration, and fragment upon collision-induced dissociation (CID) into multiple low mass fragment ions. To overcome these issues, many groups now employ a derivatization approach; however, this can be laborious and in some cases introduce unwanted side products, but does offers benefits in terms of sensitivity.
LC-MS Without Derivatization
McDonald et al. from Dallas have published a method for comprehensive sterol analysis including the nonpolar sterols lanosterol, zymosterol, desmosterol, 7-dehydrocholesterol (7-DHC), cholesterol, vitamin D3; the plant sterols brassicasterol, campesterol, stigmasterol, sitosterol; and the fungal sterols ergosterol and vitamin D2 (McDonald et al. 2012). Their method is based on reversed-phase LC-MS/MS exploiting multiple reaction monitoring (MRM) on a triple quadrupole instrument. The sensitivity of their method allowed ng/mL detection of sterols in plasma. Sterols injected in 90% methanol were resolved on a C18 LC column using 96% methanol containing 0.1% acetic acid as mobile phase A and methanol containing 0.1% acetic acid as mobile phase B. The gradient started at 100% A and increased to 100% B. Sterols were ionized by atmospheric pressure chemical ionization (APCI) in the positive-ion mode optimized for either [M + H]+ or [M + H-H2O]+ ions. Others have optimized LC-MS/MS methods for specific sterols, e.g., 7-DHC and cholesterol for the specific identification of the autosomal recessive disorder Smith-Lemli-Opitz syndrome (SLOS) (Becker et al. 2015). Becker et al. injected sample in methanol:propan-2-ol (1:1, v/v) and separated the sterols on a C18 LC column using a gradient going from 75% methanol to 100% propan-2-ol and used positive-ion atmospheric pressure photoionization (APPI) to generate [M + H-H2O]+ ions ready for MRM analysis on a triple quadruple mass spectrometer (Becker et al. 2015).
LC-MS with Derivatization
An alternative derivatization strategy is to target the 3-hydroxy group of sterols and form esters with picolinic acid (Fig. 2) (Honda et al. 2008). The nitrogen of the pyridine ring is readily sodiated facilitating ionization of the derivative. The method pioneered by Honda et al. is applicable to cholesterol, its sterol precursors, and oxysterol metabolites (see below) (Honda et al. 2008; Honda et al. 2009). The derivatization reaction is performed by adding a reaction mixture of 2-methyl-6-nitrobenzoic anhydride, 4-(N,N-dimethylamino)pyridine, picolinic acid, and tetrahydrofuran and triethylamine to dried sample. The reaction is complete after 30 min at room temperature (Honda et al. 2008). After solvent exchange to acetonitrile the supernatant is suitable for injection onto an LC-MS/MS system. Honda et al. used a reversed-phase column, an acetonitrile, methanol, water gradient going from 40:40:20 (v/v/v) to 45:45:10 (v/v/v) with 0.1% acetic acid as a constant additive, electrospray ionization (ESI), and MRM on a triple quadrupole instrument (Honda et al. 2008). The MRM was rather nonspecific being the transition [M + Na + CH3CN]+ ➔ [M + Na]+. Limits of detection were 1 pg on-column allowing analysis of sterols from 1 μL of serum (Honda et al. 2008).
Polar Neutral Sterols: Oxysterols
As for nonpolar sterols, oxysterols have been analyzed by LC-MS with and without derivatization.
LC-MS Without Derivatization
McDonald et al. have developed a LC-MS method applicable to a wide range of oxysterols extending from 4β-hydroxycholesterol at one end of the hydrophobicity range to 7α,25-dihydroxycholesterol and 7α(25R)26-dihydroxycholesterol at the other (McDonald et al. 2012). Note, we use here systematic nomenclature to describe side-chain hydroxylated sterols where according to IUPAC rules hydroxylation at the terminal carbon of cholesterol is at C-26 (Fakheri and Javitt 2012). The resulting stereochemistry is assumed to be 25R (Fig. 1). Oxysterols injected in 90% methanol were separated on a C18 LC column using a mobile phase of A: 70% acetonitrile with 5 mM NH4OAc, B: 1:1 acetonitrile, propan2-ol (v/v) with 5 mM NH4OAc, and a gradient going from 0% B to 100% B. Upon ESI the most prominent ions were [M + NH4]+ or [M + H]+, and MRM transitions to fragment ions having lost one or two molecules of water were exploited (McDonald et al. 2012). The sensitivity of the method allowed the detection of oxysterols in plasma at a level of 1 ng/mL with an instrument detection limit of <50 pg on column. One limitation of the chromatography system used was that 7α- and 7β-hydroxycholesterol were not resolved. The reproducibility of McDonald et al.’s method was good with relative standard errors mostly below 5%. Quantification was made with the use of isotope-labeled standards. Stiles et al., also in Dallas, have used this methodology in a study of 3230 serum samples measuring >60 sterols including oxysterols and vitamin D derivatives (Stiles et al. 2014). Oxysterols analyzed included 4β-, 7α-, 22R-, 24S-, 25-, and (25R)26-hydroxycholesterol; 7-oxocholesterol; and 7α,26-dihydroxycholesterol (Stiles et al. 2014). Other LC-MS methods developed for oxysterol analysis come from Rentsch and colleagues in Zurich and Bandaru and Haughey in Baltimore who both targeted on 24S- and (25R)26-hydroxycholesterol (Burkard et al. 2004; Karuna et al. 2009; Bandaru and Haughey 2014); DeBarber et al. in Portland who separated closely eluting 24S-, 25-, and (25R)26-hydroxycholesterol (DeBarber et al. 2008); and Björkhem’s group in Stockholm who focused on 7α-hydroxycholest-4-en-3-one (Lovgren-Sandblom et al. 2007).
LC-MS with Derivatization
A similar derivatization reagent to picolinic acid is its isomer nicotinic acid. Sidhu et al. have used derivatization to nicotinyl esters to enhance the analysis of oxysterols found in cerebrospinal fluid (CSF) and plasma (Sidhu et al. 2015). Oxysterols were extracted and dried and treated with a reagent mix of N,N-diisopropylcarbodiimide, nicotinic acid, 4-(N,N-dimethylamino)pyridine in chloroform at 50 °C for 1 h. Following removal of chloroform and reconstitution in methanol the oxysterols were ready for analysis. Derivatized oxysterols were separated on a reversed-phase C18 column using mobile phase of A: 0.1% formic acid in water, B: 0.1% formic acid in acetonitrile:methanol 1:4 (v/v) and a gradient going from 95% to 100% B. The oxysterols 4β-, 7α-, 7β-, 24S-, 25-, and (25R)26-hydroxycholesterol were found to give dinicotinyl esters, which with the exception of the 7α- and 7β-epimers separated on the LC column. The dinicotinyl esters give [M + H]+ and [M + 2H]2+ upon ESI and both generate [M + H-C6H5NO2]+ and [C6H5NO2 + H]+ as major product-ions suitable for MRM analysis. The lower limit of quantification for the method was 1 ng/mL from 50 μL of plasma.
Another derivatization method targeting hydroxy groups and using carbodiimide catalysis is that developed by Jiang et al. for the oxysterols 7α-, 7β-, 24S-, 25-, and (25R)26-hydroxycholesterol (Jiang et al. 2007), in this case derivatization is to N,N-dimethylglycine esters. A reagent mixture of dimethylglycine, 4-(N,N-dimethylamino)pyridine and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide in chloroform is added to the dried sample and heated at 50 °C overnight. The reaction products are then extracted into diethylether against 0.1 N ammonia solution. Bisdimethylglycine esters are formed with hydroxycholesterols while 7-oxocholesterol gives the monodimethylglycine ester (Fig. 3). Subsequently Jiang et al. developed the method further by incorporating LC separation (Jiang et al. 2011). They exploited the N,N-dimethylglycine ester derivatization for the diagnosis of the rare neurodegenerative lysosomal storage disease Niemann-Pick type C1 (NPC1) which shows elevated levels of 7-oxocholesterol and cholestane-3β,5α,6β-triol in plasma. The derivatization reaction was performed as described above, but the temperature was 45 °C and the reaction time reduced to 1 h. The reaction was quenched with methanol, dried under nitrogen, and the derivatives reconstituted in 80% methanol for injection. LC separation was on a reversed-phase column using a mobile phase of A: 0.015% trichloroacetic acid, 0.5% acetic acid in water and B: 0.015% trichloroacetic acid, 0.5% acetic acid in acetonitrile. The gradient was from 35% B to 100% B. APCI was used in combination with MRM. Cholestane-3β,5α,6β-triol was found to give a more abundant [M + H]+ than [M + 2H]2+ ion for the double derivative while 7-oxocholesterol gave the [M + H]+ for the monoderivative. [M + H]+ ions upon CID gave a prominent fragment corresponding to the loss of dimethylglycine and at m/z 104 corresponding to protonated dimethylglycine, both appropriate for MRM transitions. This derivatization is being used by others for the identification of NPC1 (Pataj et al. 2016; Klinke et al. 2015; Romanello et al. 2016). Interestingly, Klinke et al. found that Niemann-Pick type A and B patients also showed elevated levels of cholestane-3β,5α,6β-triol in plasma (Klinke et al. 2015). Pataj et al. found that the dimethylglycine derivatization method was also applicable to the diagnosis of cerebrotendinous xanthomatosis (CTX) an inborn error of cholesterol metabolism where (25R)26-hydroxycholesterol is essentially absent from plasma and tissues (Pataj et al. 2016). Rather than using APCI on a triple quadrupole instrument Pataj et al. used ESI on an Orbitrap high resolution instrument (Pataj et al. 2016). Johnson and colleagues in Australia have used derivatization to dimethylaminobutyric esters to enhance the analysis of cholestane-3β,5α,6β-triol and 7-oxocholesterol in the diagnosis of NPC (Boenzi et al. 2014). Following extraction of oxysterols from 50 μL of plasma into ethylacetate and lyophilization, dimethylaminobutyric acid imidazolide (prepared from dimethylaminobutyric acid and carbonyldiimidazole hydrochloride in methylenechloride and stable for up to 1 week in a desiccator) was added and the mixture allowed to incubate at 65 °C for 15 min (Fig. 3). The mixture was lyophilized ready for LC-MS/MS analysis on a reversed-phase C18 column using a mobile phase of A: 5 mM ammonium formate, pH 3, B: acetonitrile, and a gradient of 40–100% B. The sample was injected in a 60/40 (v/v) solution of A:B. Monoderivatives were formed with cholestane-3β,5α,6β-triol and 7-oxocholesterol which both gave [M + H]+ ions upon ESI and fragmented upon CID to give a major product-ion at m/z 132 corresponding to protonated dimethylaminobutyric acid. The assay was linear over a concentration range of 1–200 ng/mL and showed both intraday and interday coefficients of variation (CV) of below 15%. For control plasma median concentrations of cholestane-3β,5α,6β-triol and 7-oxocholesterol were determined to be 6.4 ng/mL and 26.6 ng/mL and in NPC patients 55.5 ng/mL and 120 ng/mL, respectively. All samples from NPC patients showed abnormal concentration of cholestane-3β,5α,6β-triol, and it was concluded that this was the “gold standard” disease biomarker. A danger of basing a diagnosis on cholestane-3β,5α,6β-triol and 7-oxocholesterol concentrations is that the latter is formed from cholesterol, the major sterol in plasma, by reaction with air, as is the former following hydrolysis of the cholesterol autoxidation product 5,6-epoxycholesterol (Schroepfer 2000).
The authors of this chapter have also extensively exploited the GP derivatization methodology for the diagnosis of the rare disease SLOS, CTX, NPC, and spastic paraplegia type 5 (Griffiths et al. 2013; Theofilopoulos et al. 2014; Griffiths et al. 2016b; Griffiths et al. 2016a; Dai et al. 2014). We have also applied the technology to the discovery of potential biomarkers for amyotrophic lateral sclerosis and multiple sclerosis (Crick et al. 2016; Abdel-Khalik et al. 2017). We developed the Girard derivatization method further by incorporating an enzyme catalyzed oxidation step so that sterols with a 3β-hydroxy-5-ene or 3β-hydroxy-5α-hydrogen structure could be converted to 3-oxo-4-enes or 3-oxo sterols, respectively (Karu et al. 2007; Griffiths et al. 2003). The method was extended to allow sterols with a 3α-hydroxy-5β-hydrogen structure to be converted to 3-oxo equivalents. Cholesterol oxidase from Streptomyces sp. was used to oxidize 3β-hydroxy groups, while 3α-hydroxysteroid dehydrogenase from Pseudomonas testosterone was used to oxidize 3α-hydroxy groups (Fig. 4). To allow for the differentiation of sterols with a natural 3-oxo group from those oxidized to contain one, we utilize isotope-labeled versions of the GP reagent, where the [2H5]GP reagent is used to derivatize an aliquot of sample treated with enzyme while [2H0]GP is used to derivatize a separate aliquot of sample untreated by enzyme (Crick et al. 2015a) (Fig. 4). In brief, our methodology is as follows. Sterols are extracted from fluid or tissue into ethanol. After dilution to 70% ethanol, nonpolar sterols, including cholesterol, desmosterol, and 7-DHC are extracted on a C18 column, while more polar sterols including oxysterols, bile acids, and C21–18 steroids flow through in 70% ethanol. This provides an “oxysterol fraction” devoid of cholesterol and eliminates the possibility of introducing oxysterol artifacts from cholesterol oxidation during subsequent sample preparation. In most studies we target nonesterified oxysterols; if we are interested in oxysterols esterified with fatty acids, we perform a saponification step during the initial ethanolic extraction, and after neutralization and dilution, sterols can be fractionated as above. The “oxysterol fraction” from the C18 column is then divided into two subfractions A and B then lyophilized. The subfractions are reconstituted in propan-2-ol. To fraction A enzyme is added in buffer to oxidize 3-hydroxy to 3-oxo groups. In most studies we have used cholesterol oxidase in 50 mM phosphate buffer pH 7 targeting upon 3β-hydroxy-containing sterols. After an incubation period of 1 h at 37 °C the reaction is quenched by addition of methanol. Fraction B is treated in an identical manner but in the absence of cholesterol oxidase. Derivatization is completed by the addition of acetic acid catalyst and [2H5]GP to subfraction A and [2H0]GP to subfraction B and the mixture left overnight in the dark. The next day we use a recycling SPE protocol using Oasis HLB columns to remove excess GP reagent and derivatized oxysterols are eluted in methanol (Crick et al. 2015b). Following dilution to 60% methanol oxysterol subfractions A and B can be combined and injected on to the LC column. We use a reversed-phase C18 column with mobile phase A: 33.3% methanol, 16.7% acetonitrile, 0.1% formic acid, B: 63.3% methanol, 31.7% acetonitrile, 0.1% formic acid, and a gradient of 20% to 80% B. The cholesterol-rich fraction is treated in an identical manner. We perform analysis on an LTQ-Orbitrap, exploiting the high-resolution capability of the instrument to generate RICs for the ions of interest. ESI gives intense [M]+ ions suitable for quantification by isotope-dilution methods. [M]+ ions from fraction B are 5 Da lighter than equivalents from fraction A (Fig. 4), but elute with the same retention time. The method is suitable for quantification of oxysterols from plasma at sub-ng/mL levels and from CSF at concentrations of 30 pg/mL and above (Crick et al. 2016).
A major advantage of the Girard derivatives is that they give informative fragmentation patterns upon CID. On triple quadrupole and Q-TOF type instruments the major fragment-ions correspond to the loss of the pyridine ring from GP or trimethylamine from Girard T (GT) derivatives providing suitable transitions for high-sensitivity MRM analysis (Fig. 4) (Griffiths et al. 2003; Griffiths et al. 2006). The resulting [M-Py]+ and [M-TMA]+ ions fragment further to give structurally informative MS/MS spectra (Griffiths et al. 2003; Griffiths et al. 2006). This can be exploited on ion-trap instruments by isolating [M-Py]+ or [M-TMA]+ fragments in the ion-trap and fragmenting them further in a multistage fragmentation (MSn) event (Karu et al. 2007). We have found that even an apparently minor change in structure, e.g., epimerization of the 7-hydroxy group, results in a characteristically different spectrum. The major objection to the methodology described above is its laborious nature.
Roberg-Larsen and Wilson and colleagues have enhanced the sensitivity for oxysterol analysis using the GT derivative and miniaturized columns with low-flow rate ESI (Roberg-Larsen et al. 2014; Roberg-Larsen et al. 2016). Similar studies were performed earlier by Karu et al. but with GP derivatization (Karu et al. 2011). Roberg-Larsen et al. exploited on-line C18 sample clean-up using an automatic filtration and back-flush setup prior to the analytical column (Roberg-Larsen et al. 2014). Using a nano-LC setup they were able to achieve a lower limit of quantification of about 50 fg on-column and were able to determine fg quantities of oxysterols from 10,000 cells (Roberg-Larsen et al. 2014). Following treatment with cholesterol oxidase oxysterols were derivatized in a similar manner as described above but with the GT reagent replacing the GP reagent. The derivatized sample solution was then injected onto the LC system in 0.1% formic acid. Cell debris and precipitated matter was trapped on an on-line filter while derivatized oxysterols were trapped on a C18 trap-column. Unreacted derivatization reagent was flushed to waste. By valve-switching the on-line filter was back-flushed to waste using the loading solvent, while the trap column was switched in-series with an analytical C18 column (0.1 mm i.d.). Oxysterols were separated with an iscocratic mobile phase of methanol 95% containing 0.1% formic acid at a flow-rate of 0.5 μL/min. Roberg-Larsen et al. used an Q-Exactive quadrupole–Orbitrap type instrument and exploited the [M]+➔[M-TMA]+ transitions for quantification (Roberg-Larsen et al. 2014). A disadvantage of the nanoscale format, despite the improvement in sensitivity, is the long analysis times as a consequence of operating at sub-μL/min flow-rates. Roberg-Larsen et al. subsequently scaled-up to capillary column LC (0.3 mm i.d.) with a flow-rate of 5 μL/min; this allowed both shorter run-times and faster system equilibration while maintaining high sensitivity and allowing increased injection volumes compared to their nano-LC system (Roberg-Larsen et al. 2016). Using this capillary LC system they were able to determine increased levels of (25R)26-hydroxycholestrerol in exosomes from an ER+ breast cancer cell line compared to an ER- cell line (Roberg-Larsen et al. 2016).
There are well-described methods for the LC-MS analysis of bile acids from urine and plasma/serum (Makin and Gower 2010). Many exploit reversed-phase SPE followed by reversed-phase LC (Zhang et al. 2015), but others exploit protein precipitation (Krautbauer et al. 2016) or just dilute the sample in methanol prior to injection on the LC column (Mazzacuva et al. 2016). There are fewer methods described for the LC-MS analysis of bile acids from bile or feces (Mi et al. 2016; Kakiyama et al. 2014).
Clayton and colleagues in London have developed an LC-MS/MS method suitable for bile acids extracted from urine by dilution in methanol and from blood on screening cards by elution with methanol (Mazzacuva et al. 2016). They separate bile acids on a reversed-phase C18 column using mobile phase of A: 0.01% formic acid, B: methanol, and a gradient going from 40% B to 99% B. Negative-ion ESI is used and quantification by MRM. Clayton et al. screen for cholic, chenodeoxycholic, deoxycholic, lithocholic, hyocholic, and hyodeoxycholic acids in their unconjugated forms and conjugated with glycine or taurine (Mazzacuva et al. 2016). They exploit the MRM transitions [M-H]−➔m/z 74 and [M-H]−➔m/z 80 for glycine and taurine conjugates while selected ion recording of [M-H]− ions is used to detect and quantify unconjugated bile acids. Clayton and colleagues have an interest in inborn errors of bile acid biosynthesis and have described unusual bile acids characteristic for different disorders, e.g., 3β-hydroxy-Δ5-C27-steroid oxidoreductase (HSD3B7) deficiency, Δ4–3-oxo-5β-reductase (AKR1D1) deficiency, and NPC. The bile acids characteristic of 3β-hydroxy-Δ5-C27-steroid oxidoreductase deficiency include 3β,7α-dihydroxychol-5-enoic and 3β,7α,12α-trihydroxychol-5-enoic acids conjugated with glycine and/or sulfuric acid. Δ4–3-oxo-5β-reductase deficiency is characterized by 7α-hydroxy-3-oxochol-4-enoic and 7α,12α-dihydroxy-3-oxochol-4-enoic acids conjugated with glycine or taurine, while NPC is characterized by 3β,7β-dihydroxychol-5-enoic acid conjugated with glycine alone or with N-acetylglucosamine (GlcNAc) or as the double conjugates with glycine and GlcNAC, taurine and GlcNAc, or sulfuric acid and GlcNAc (Mazzacuva et al. 2016). The transition [M-H]−➔ m/z 97 is characteristic for sulfuric acid conjugates. The neutral-loss of 203 Da is characteristic for [M-H]− ions of GlcNAc conjugates and was used in the MRM transition [M-H]−➔[M-H-203]− to identify and quantify GlcNAc conjugates. However, a common homozygous polymorphism in the UGT3A1 gene leads to an absence in activity of UDP N-acetylglucosaminyl transferase and an inability to form GlcNAc conjugates; thus not all NPC patients show elevated levels of GlcNAc conjugates. Instead Clayton and colleagues identified a trihydroxy bile acid with a probable structure of 3β,5α,6β-trihydroxycholanoyl-glycine as a better diagnostic marker of NPC in plasma or blood (Mazzacuva et al. 2016).
There is a growing interest in fecal bile acid profiles. This is driven by the realization of the importance of the gut microbiota which can metabolize bile acids secreted with bile into the intestine. Bioconversions catalyzed by microbiota alter the signaling properties of bile acids, but bile acids can also alter the composition of the microbiota (Wahlstrom et al. 2016). Kakiyama et al. have developed a simple LC-MS/MS method for fecal analysis of bile acids (Kakiyama et al. 2014). Powdered lyophilized stool is suspended in cold acetate buffer pH 5.6 then refluxed with ethanol. Following centrifugation the supernatant is diluted with 15% ethanol and bile acids extracted on a C18 SPE column. The column is washed with 25% ethanol and bile acids eluted in absolute ethanol. The eluate is concentrated and redissolve in 50% ethanol ready for LC-MS/MS analysis (Kakiyama et al. 2014). Kakiyama et al. separated the bile acids on a C18 LC column using a mixture of ethanol, methanol, and 5 mM ammonium acetate as mobile phase, bile acids were ionized by negative-ion ESI and identified and quantified by MRM (Kakiyama et al. 2014; Muto et al. 2012). Thirty six C24 bile acids were analyzed including unconjugated, glycine or taurine conjugated, as well as C-3 sulfated bile acids.
The bile acid concentration in bile is about 10 μg/μL, in which case dilution in 50% methanol with heating to 60–80 °C followed by centrifugation is sufficient sample preparation prior to LC-MS/MS analysis. The LC-MS/MS methods discussed above are then suitable for bile acid analysis.
It is beyond the scope of this chapter to discuss the LC-MS/MS of C21–18 steroids in detail. The interested reader is directed to comprehensive chapters in the book Steroid Analysis edited by Makin and Gower (Makin and Gower 2010).
The gold standard method for quantification by LC-MS and LC-MS/MS is stable isotope dilution MS where a known amount of an isotope-labeled version of the analyte of interest is added as early as possible during the sample preparation procedure. As the target analyte and the isotope-labeled standard will have (almost) identical physical and chemical properties any losses of analyte during sample preparation will be corrected for, elution times from the LC column will be (almost) the same, and response factors will be equivalent. An isotope-labeled version of the target analyte is not always available in which case a close structural analogue, not present in the sample to be analyzed, may be used.
To maximize throughput there is a drive towards automation. The use of on-line multidimensional chromatography is an attractive option where sterols can be fractionated according to pKa and polarity on a mixed mode ion exchange column and then separated according to hydrophobicity on a reversed-phase analytical column prior to MS/MS analysis. In an effort to gain maximum sensitivity one option is derivatization. This may become more popular when coupled with on-line sample clean up, e.g., using the filtration back flush SPE method. Alternatively, sensitivity can be enhanced by exploiting nano-LC-MS/MS or capillary-LC-MS/MS with or without derivatization. One note of caution, in omics-type experiments there is a growing tendency to use the term “identified” when a measured m/z is matched to an entry in a database. This may be supported by retention time information. The analyst should be reminded of the extreme diversity of sterol isomers with the exactly same m/z and possibly similar retention time. In the absence of a comparison to an authentic standard the term “annotated” may be more appropriate than “identified.”
Conflict of Intrest
WJG and YW have made the patent application number WO 2014037725 A1 “Kit and method for the quantification of steroids.”
- Boenzi S, Deodato F, Taurisano R, Martinelli D, Verrigni D, Carrozzo R, et al. A new simple and rapid LC-ESI-MS/MS method for quantification of plasma oxysterols as dimethylaminobutyrate esters. Its successful use for the diagnosis of Niemann-Pick type C disease. Clin Chim Acta. 2014;437:93–100.CrossRefPubMedGoogle Scholar
- Crick PJ, Griffiths WJ, Zhang J, Beibel M, Abdel-Khalik J, Kuhle J, et al. Reduced plasma levels of 25-hydroxycholesterol and increased cerebrospinal fluid levels of bile acid precursors in multiple sclerosis patients. Mol Neurobiol. 2016. Epub ahead of print.Google Scholar
- DeBarber AE, Sandlers Y, Pappu AS, Merkens LS, Duell PB, Lear SR, et al. Profiling sterols in cerebrotendinous xanthomatosis: utility of Girard derivatization and high resolution exact mass LC-ESI-MS(n) analysis. J Chromatogr B Analyt Technol Biomed Life Sci. 2011;879(17–18):1384–92.CrossRefPubMedGoogle Scholar
- Griffiths WJ, Abdel-Khalik J, Crick PJ, Ogundare M, Shackleton CH, Tuschl K, et al. Sterols and oxysterols in plasma from Smith-Lemli-Opitz syndrome patients. J Steroid Biochem Mol Biol. 2016a. doi:10.1016/j.jsbmb.2016.03.018.Google Scholar
- Griffiths WJ, Abdel-Khalik J, Crick PT, Wang Y. Unravelling new pathways of sterol metabolism. LE STUDIUM Conference, Lipids, nanotechnology and cancer; Hôtel de Ville de Tours, France. 2016b.Google Scholar
- Lovgren-Sandblom A, Heverin M, Larsson H, Lundstrom E, Wahren J, Diczfalusy U, et al. Novel LC-MS/MS method for assay of 7alpha-hydroxy-4-cholesten-3-one in human plasma. Evidence for a significant extrahepatic metabolism. J Chromatogr B Analyt Technol Biomed Life Sci. 2007;856(1–2):15–9.CrossRefPubMedGoogle Scholar
- Luchetti G, Sircar R, Kong JH, Nachtergaele S, Sagner A, Byrne EF, et al. Cholesterol activates the G-protein coupled receptor Smoothened to promote Hedgehog signaling. elife. 2016;5.Google Scholar
- Makin HL, Gower DB (editors) Steroid Analysis. Springer Science+Business Media B.V.; 2010.Google Scholar
- Roberg-Larsen H, Lund K, Seterdal KE, Solheim S, Vehus T, Solberg N, et al. Mass spectrometric detection of 27-hydroxycholesterol in breast cancer exosomes. J Steroid Biochem Mol Biol. 2016. Epub ahead of print.Google Scholar
- Romanello M, Zampieri S, Bortolotti N, Deroma L, Sechi A, Fiumara A, et al. Comprehensive evaluation of plasma 7-ketocholesterol and cholestan-3beta,5alpha,6beta-Triol in an Italian cohort of patients affected by Niemann-Pick disease due to NPC1 and SMPD1 mutations. Clin Chim Acta. 2016;455:39–45.CrossRefPubMedGoogle Scholar
- Zhang W, Jha P, Wolfe B, Gioiello A, Pellicciari R, Wang J, et al. Tandem mass spectrometric determination of atypical 3beta-hydroxy-Delta5-bile acids in patients with 3beta-hydroxy-Delta5-C27-steroid oxidoreductase deficiency: application to diagnosis and monitoring of bile acid therapeutic response. Clin Chem. 2015;61(7):955–63.CrossRefPubMedGoogle Scholar