No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
An atlas of gas chromatographic profiles of neutral steroids in health and disease
... metabolome of what is now termed P450 oxidoreductase deficiency (ORD) was accurately described long ago by thin layer chromatography and GC/MS urine profile analysis and was named combined 17α- hydroxylase/21-hydroxylase deficiency111213. Patients with ORD present with a pathognomonic urine steroid profile showing increased metabolites typically seen in 21- hydroxylase deficiency, such as pregnanetriol, 17-OH-pregnanolone and pregnanetriolone (17OHP and 21-deoxycortisol metabolites) increased excretion of mineralocorticoid precursors (corticosterone metabolites) and decreased excretion of active androgen metabolites [13]. ...
Liquid chromatography tandem mass spectrometry (LC/MS/MS) is replacing classical methods for steroid hormone analysis. It requires small sample volumes and has given rise to improved specificity and short analysis times. Its growth has been fueled by criticism of the validity of steroid analysis by older techniques, testosterone measurements being a prime example. While this approach is the gold-standard for measurement of individual steroids, and panels of such compounds, LC/MS/MS is of limited use in defining novel metabolomes. GC/MS, in contrast, is unsuited to rapid high-sensitivity analysis of specific compounds, but remains the most powerful discovery tool for defining steroid disorder metabolomes. Since the 1930s almost all inborn errors in steroidogenesis have been first defined through their urinary steroid excretion. In the last 30 years, this has been exclusively carried out by GC/MS and has defined conditions such as AME syndrome, glucocorticoid remediable aldosteronism (GRA) and Smith–Lemli–Opitz syndrome. Our recent foci have been on P450 oxidoreductase deficiency (ORD) and apparent cortisone reductase deficiency (ACRD).
In contrast to LC/MS/MS methodology, a particular benefit of GC/MS is its non-selective nature; a scanned run will contain every steroid excreted, providing an integrated picture of an individual's metabolome. The “Achilles heel” of clinical GC/MS profiling may be data presentation. There is lack of familiarity with the multiple hormone metabolites excreted and diagnostic data are difficult for endocrinologists to comprehend. While several conditions are defined by the absolute concentration of steroid metabolites, many are readily diagnosed by ratios between steroid metabolites (precursor metabolite/product metabolite). Our work has led us to develop a simplified graphical representation of quantitative urinary steroid hormone profiles and diagnostic ratios.
The androgens are an important group of C19 steroids comprising, principally, testosterone and its ring A-saturated metabolite, 5α-dihydrotestosterone (5α-DHT), and the isomeric androstane-3,17-diols. In addition, 4-androstenedione (4-A) and dehydroepiandrosterone (DHA) are usually classed as androgens, although the androgenic activity of these compounds is only some 10–15% of that of testosterone itself in the case of 4-androstenedione, and even less in the case of DHA (for reviews see Shoppee, 1964a; Gower, 1984a).
Steroids consist of an essentially lipophilic (or hydrophobic, non-polar) cyclopentanoperhydrophenanthrene nucleus modified on the periphery of the nucleus or on the side chain by the addition of hydrophilic (or lipophobic, polar) groups. These additional groups are mainly hydroxyl and oxo or carboxylic acid groups. In addition, in the case of the estrogens, the presence of a phenolic A ring renders the normally neutral 3-hydroxyl acidic. During metabolism, steroids become more hydrophilic by esterification (conjugation) with glucuronic or sulphuric acid. Bile acids (containing a C-24 carboxylic acid group) may be linked through a peptide bond to glycine or taurine. Despite the addition of these polar groups, the essential non-polarity of the steroids means that they are all to varying degrees soluble in organic solvents and can thus be extracted from aqueous media by a solvent or solvent mixture of suitable polarity.
Bbackground:
Urine steroid profiles are used in clinical practice for the diagnosis and monitoring of disorders of steroidogenesis and adrenal pathologies. Machine learning (ML) algorithms are powerful computational tools used extensively for the recognition of patterns in large data sets. Here, we investigated the utility of various ML algorithms for the automated biochemical interpretation of urine steroid profiles to support current clinical practices.
Methods:
Data from 4619 urine steroid profiles processed between June 2012 and October 2016 were retrospectively collected. Of these, 1314 profiles were used to train and test various ML classifiers' abilities to differentiate between "No significant abnormality" and "?Abnormal" profiles. Further classifiers were trained and tested for their ability to predict the specific biochemical interpretation of the profiles.
Results:
The best performing binary classifier could predict the interpretation of No significant abnormality and ?Abnormal profiles with a mean area under the ROC curve of 0.955 (95% CI, 0.949-0.961). In addition, the best performing multiclass classifier could predict the individual abnormal profile interpretation with a mean balanced accuracy of 0.873 (0.865-0.880).
Conclusions:
Here we have described the application of ML algorithms to the automated interpretation of urine steroid profiles. This provides a proof-of-concept application of ML algorithms to complex clinical laboratory data that has the potential to improve laboratory efficiency in a setting of limited staff resources.
Most of the methods currently employed for the determination of bile acids in biological samples require some form of initial extraction procedure. Techniques which are generally available include the following:
(1)
Liquid-liquid partition.
(2)
Protein precipitation.
(3)
Liquid-solid extraction — anion exchange resins, neutral polystyrene polymers of Amberlite XAD and reverse-phase octadecylsilane-bonded silica.
(4)
Liquid-gel extraction — Lipidex 1000.
The paper by Arlt et al. (1) on detecting malignancy in adrenocortical tumors illustrates well the benefits of cooperation between centers and of a systematic approach to classifying data, in this case on urinary steroid profiles. We have long used urine steroid profiling as a means of detecting steroid secretion by adrenocortical tumors (2), but definitive guidance on its accuracy in differentiating malignancy has previously been limited to small patient numbers or a limited set of steroid markers.
Histological classification systems have already served this field well. They continue to evolve (3) and are undoubtedly the best indicator currently available of malignant potential, but there is room to improve differentiation of adrenocortical adenoma from nodular hyperplasia (4).
We wish to offer some pointers on how these approaches could develop further. First, it is important to apply very clear criteria for identifying adrenal neoplasms: Arlt et al. describe their classification as being based on histology and presence of metastases, but only 36/147 cases were given a Weiss score, so it was not possible to judge the performance of their steroid profile findings against this gold standard. In the absence of histology, there are two potential confusing elements: (a) some adrenocortical adenomas might represent dominant nodules in nodular hyperplasia that can remain undetected because of the limited resolution of imaging for sub-centimeter lesions; and (b) not all non-metastatic lesions are necessarily benign, but they could not be diagnosed unless they were evaluated histologically. It would be useful to see the outcome of steroid metabolomic analysis by Arlt et al. applied only to those defined histologically. Second, the greatest need lies in determining the prognosis of those with adrenocortical carcinomas (ACC) on histology that do not have secondaries at presentation. Arlt et al. report no difference of steroid metabolome between those with and without secondaries.
Continuing surveillance of ACC patients initially without secondaries may enable histological and steroid pointers to become recognized. Third, although a large set (32) of steroid metabolites was considered, some reported markers of ACC (e.g., Pregnene-3α,16α,20α-triol, ref. 5) were not included. It has been our experience that close examination of profiles obtained unselectively by full scan gas chromatography-mass spectrometry (GC-MS) often reveals new and unexpected markers of ACC and these can prove crucial for detection of relapse. A system akin to the Weiss system of histological scoring, based on number of steroid markers abnormally increased, may also prove to be more accurate and easier for referral laboratories to apply. What is already clear is that if surgical removal of a mass in the adrenal cortex is contemplated, determining presence or absence of markers in a urine steroid profile should be part of the work up.
References
1. Arlt W, Biehl M, Taylor AE, Hahner S, Libe R, Hughes BA, Schneider P, Smith DJ, Stiekema H, Krone N, Porfiri E, Opocher G, Bertherat J, Mantero F, Allolio B, Terzolo M, Nightingale P, Shackleton CHL, Bertagna X, Fassnacht M, Stewart PM. 2011. Urine steroid metabolomics as a biomarker tool for detecting malignancy in adrenal tumors. J Clin Endocrinol Metab 96:3775-3784
2. Shackleton CHL, Taylor NF, Honour JW. 1980. An atlas of gas chromatographic profiles of urinary steroids in health and disease Monograph, Packard Becker Ltd.
3. Lau SK, Weiss LM. 2009, The Weiss system for evaluating adrenocortical neoplasms: 25 years later. Hum Pathol 40:757-768
4. Blanes A, Diaz-Cano SJ. 2007. Histologic criteria for adrenocortical proliferative lesions. Anatomic Pathol 127:398-408
5. Tiu SC, Chan AOK, Taylor NF, Leung PY, Choi CH, Shek CC. 2009. Use of urinary steroid profiling for diagnosing and monitoring adrenocortical tumours. Hong Kong Med J 15:463-470
Three patients with advanced adrenocortical carcinoma were treated with a combination of intermittent streptozocin and continuous o,p'DDD. Two patients were treated preoperatively and the primary tumors, initially considered as inoperable, could be resected after 19 and 5.5 months, respectively. In the patient with the longer treatment (35 months), lung and lymph node metastases disappeared and she has no evidence of recurrent disease 6.5 years after start of therapy. One patient was followed by magnetic resonance imaging (MRI) and urinary steroid secretion. The MRI gave a good visualization of the tumor. Measurements of relaxation times showed a significant decrease in T1 values. The urinary steroid profile showed an increased secretion of 3 beta-hydroxy-5-ene steroids and tetrahydro-11-deoxy-cortisol. Treatment with streptozocin and o,p'DDD initially increased 16-oxygenation of dehydroepiandrosterone and androst-5-en-3 beta,17 beta-diol, followed by a decrease in the secretion of all urinary steroids. The third patient received postoperative treatment with no effect on metastatic disease in the lungs, she died 9 months after start of treatment. The therapeutic approach with the combination regimen of streptozocin and o,p'DDD pretreatment plus aggressive surgery has to be further evaluated, as well as MRI and urinary steroid profile as methods to monitor the effect of therapy.
In a study using gas chromatography‐mass spectrometry (GC‐MS) on urine specimens from 16 normal infants and 16 infants with congenital adrenal hyperplasia (CAH) due to 21‐hydroxylase deficiency (aged 1 day to 4 weeks), the major steroids recognized in all infants were: 16α‐hydroxy‐dehydroepiandrosterone, 16β‐hydroxy‐dehydroepiandrosterone, 16‐oxo‐androstenediol, androstenetriol, 15β, 17α‐dihydroxy‐pregnenolone and 16α‐hydroxy‐pregnenolone.
Pregnanetriol was detectable in three normal infants (aged 3, 6 and 15 days) but the levels seen in 15 CAH patients were in a higher range. Pregnanetriolone, 5β‐17‐hydroxy‐pregnanolone and 15β, 17α‐dihydroxy‐pregnanolone were present in the urine of 15 CAH patients, but were not detectable in any of the normal infants. The older the patient, the higher the level was of each of these four steroids.
The results indicate that, even on day 1, patients with CAH due to 21‐hydroxylase deficiency may be positively identified using GC‐MS of urine specimens. This does not preclude the possibility that a minority of patients with CAH, most likely those with mild 21‐hydroxylase deficiency, may not exhibit the characteristic GC‐MS findings on day 1, as seen in one of the 18 CAH patients.
The development of tandem mass spectrometry (m.s./m.s.) techniques for the rapid structural elucidation and quantification of metabolites is described. With these techniques the mass spectrometer serves as both a separation device as well as a tool for structure elucidation. Ion formation by desorption ionization techniques and mass analysis with an appropriate combination of neutral loss and precursor ion scans can then be used to classify the number and type of primary metabolites and polar drug conjugates that are present in a complex sample matrix. The resulting information can be used to assess the overall biotransformation routes that are available to a drug or other xenobiotic substance.
A new mass-spectrometric technique relies on ionization during bombardment of the analyte (dissolved in a liquid matrix, usually glycerol) by an atom beam (e.g., Ar0, Xe0). This technique, termed "fast atom bombardment," is particularly useful in the characterization of polar charged molecules. A neutral beam is not essential, and a primary beam of cesium ions has been successfully used to produce spectra equivalent to those obtained by fast atom bombardment. In this communication I report data on the use of both ion and atom primary beams for producing secondary-ion mass spectra of conjugated steroids. In negative-ion spectra produced for steroid glucuronides and sulfates, the ion [M - H]- is invariably the major high-mass peak, and the lack of substantial fragmentation allows assay of relatively complex mixtures if the analytes differ in mass. I describe here the use of secondary-ion mass spectrometry for distinguishing, by urinary steroid analysis, patients with the four enzyme defects that can affect cortisol synthesis: defects in 17 alpha-hydroxylase, 3 beta-hydroxysteroid dehydrogenase/isomerase, 21-hydroxylase, and 11 beta-hydroxylase.
Data are presented on the mass spectrometry of intact steroid conjugates. The principal technique used was secondary ion mass spectrometry (SIMS) using a Cs+ ion beam for ionization, although comparable data were obtained by fast atom bombardment (FAB) using a Xeo beam. In both techniques the samples were analyzed in a liquid matrix (glycerol). Positive and negative ion spectra have been obtained, the latter being most useful for steroid sulfate and glucuronide analysis. The negative ion spectra are dominated by a pseudomolecular ion at m/z [M-H]- (M of free acid) and the lack of marked fragmentation permits mixtures of steroids to be resolved in a single spectrum, providing they differ in mass. Preliminary data on the separate analysis of individual components from urine and plasma of patients with assorted disorders of steroid synthesis and metabolism are presented. This technique shows great promise for the clinical analysis of steroid conjugates without the need for enzymic hydrolysis or chromatographic separation of individual steroids.
The synthesis and identification of 12 A-ring reduced 6 alpha-(and 6 beta-)hydroxylated compounds derived from 11-deoxycortisol (S), corticosterone (B) and 11-dehydrocorticosterone (A) are reported here. These steroids were prepared in two steps from the corresponding 6 6 alpha-(and 6 beta-)hydroxy-4-pregnene-3-ones. Selective reduction of the 4,5 double bond yielded 12 6 alpha-(and 6 beta)hydroxy-5 alpha-(and 5 beta)pregnane-3,20-diones. Enzymatic reduction of these compounds with NADH and 3 alpha-hydroxysteroid dehydrogenase yielded the corresponding tetrahydro steroids. The steroids were characterized by high performance liquid chromatography (HPLC), gas chromatography mass spectrometry (GC and GC/MS) and in part by 1H-NMR. 6 beta OH-THS and 6 beta OH-5 alpha THS were identified by 1H-NMR. The structures of the two precursors, i.e. 6 beta OH-5 beta DHS and 6 beta OH-5 alpha DHS were confirmed by 1H-NMR using two-dimensional spectra. 6 alpha OH-THS was identified by comparing its HPLC, GC and MS data with those of the steroid obtained by enzymatic oxidation of the standard reference steroid 6 alpha OH-20 beta HHS to the corresponding 20-ketosteroid. The other steroids, e.g. 6 alpha OH-THB and 6 alpha OH-5 alpha THB were identified by using the proved sequence of elution of each of the epimer pairs on the normal phase HPLC column (5 alpha < 5 beta), and by the reversed order of elution of the same epimer pair as the methoxime-trimethylsilyl ethers on the GC column (5 alpha > 5 beta) and by the mass spectra, with the exception of 6 beta OH-THA.
The identification of 3 new 15 beta-hydroxylated 21-deoxy-pregnanes in the urinary steroid profile of a 4-month-old girl with congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency (21OHD) is reported here. These steroids were identified by gas chromatography and gas chromatography-mass spectrometry as 3 alpha,15 beta,17-trihydroxy-5 alpha-pregnan-20-one (5 alpha II), 3 alpha,15 beta,17,20 alpha-tetrahydroxy-5 alpha-pregnane, and 3 alpha,15 beta,17,20 alpha-tetrahydroxy-5 beta-pregnane (20 alpha DH-II). Two other compounds in the urine, 3 beta,15 beta,17- trihydroxy-5 alpha-pregnan-20-one and 3 beta,15 beta,17-trihydroxy-5 beta-pregnan-20-one were also characterized. The identification of the former 3 steroids was obtained by comparing their methylene unit values and mass spectral data with the corresponding data of the standard steroids synthesized from 15 beta,17-dihydroxy-4-pregnene-3,20-dione. Seven other synthesized and identified 15 beta-hydroxylated steroids were 3 alpha,15 beta,17-trihydroxy-5 beta-pregnan- 20-one (II), 3 alpha,15 beta,17,20 beta-tetrahydroxy-5 beta-pregnane, 15 beta,17-dihydroxy-5 alpha-pregnane-3,20-dione, 15 beta,17-dihydroxy-5 beta-pregnane-3,20-dione, 3 alpha,15 beta-dihydroxy-5 alpha-androstan-17-one (15 beta OH-An), 3 alpha,15 beta-dihydroxy-5 beta-androstan-17-one (15 beta OH-Et) and 3 alpha,15 beta,17,20 beta- tetrahydroxy-5 alpha-pregnane. Of these the latter two have not been reported previously. This study supports the findings that 15 beta-hydroxylated steroids are common in the neonate and could play an important role in the diagnosis of CAH due to 21OHD, where II and the newly identified steroids from this investigation viz., 5 alpha II and 20 alpha DH-II appear the most important 15 beta-hydroxysteroid markers for this disease.
ResearchGate has not been able to resolve any references for this publication.