- Email: [email protected]
Midgestational maternal urine steroid markers of fetal Smith–Lemli–Opitz (SLO) syndrome (7-dehydrocholesterol 7-reductase deficiency)
Steroids 64 (1999) 446 – 452
Rapid communication
Midgestational maternal urine steroid markers of fetal Smith–Lemli– Opitz (SLO) syndrome (7-dehydrocholesterol 7-reductase deficiency)夞 Cedric H.L. Shackletona,*, Esther Roitmana, Lisa E. Kratzb, Richard I. Kelleyb a
Children’s Hospital Oakland Research Institute, 747 52nd Street, Oakland, CA 94609, USA Division of Neurogenetics, Kennedy Krieger Institute, The Johns Hopkins University, Baltimore, MD, 21215 USA
b
Received 26 January 1999; received in revised form 17 February 1999; accepted 18 February 1999
Abstract Smith–Lemli–Opitz syndrome (SLOS) is a malformation syndrome associated with 7-dehydrocholesterol (7DHC) 7-reductase deficiency. Although SLOS can be detected in an affected fetus before midpregnancy by measurement of 7DHC levels in amniotic fluid or chorionic villus cells, a noninvasive, more routine method is needed. Accordingly, this study was instigated to search for specific steroids in maternal urine in an affected pregnancy that reflect the 7-reductase deficiency of the fetus, ie, steroids retaining 7,8-unsaturation. Steroids were characterized by gas chromatography/mass spectrometry after urinary extraction, conjugate separation, and derivatization. Most steroids in maternal urine from a patient carrying a SLOS fetus were identified as progesterone metabolites, and these were entirely conventional, showing no evidence of additional unsaturation. Unsaturated homologues of the cortisol metabolites were also not detected. However, unsaturated homologues of pregnane-3,16,20-triols and pregnane-3,17,20-triol were found. Most likely, these are 7,8-unsaturated homologues, but 8,9-unsaturation is also possible because of the known activity of ⌬7-⌬8-isomerase on 7DHC, which results in 8DHC being a prominent sterol in SLOS. Among these novel human steroids, the following were provisionally characterized: 5-pregn-7(or 8)-ene3␣,17␣,20␣-triol, 5-pregn-7(or 8)-ene-3␣,16␣,20␣-triol, and 5␣-pregn-7(or 8)-ene-3,16␣,20␣-triol. Confirmation of the position of unsaturation will require steroid synthesis. These novel steroids are not present in normal pregnancy urine and, therefore, are valuable for prenatal diagnosis of SLOS. In addition, separate studies have shown that 5-pregn-7(or 8)-ene-3␣,17␣,20␣-triol is present in urine of children and adults with SLOS, and so is a useful analyte for confirmation of the disorder throughout life. © 1999 Published by Elsevier Science Inc. All rights reserved. Keywords: Smith–Lemli–Opitz syndrome; Pregnancy urine steroids; ⌬7-⌬8-Pregnenetriols; Gas chromatography/mass spectrometry
1. Introduction Smith–Lemli–Opitz syndrome (SLOS) is an autosomal recessive malformation disorder associated with 7-dehydrocholesterol (7DHC) 7-reductase deficiency [1,2]. Because this enzyme catalyzes the final step in cholesterol synthesis, a primary feature of the disorder is cholesterol deficiency [2,3] with concomitant buildup of 7DHC. Presence of an active ⌬7,⌬8-isomerase also causes production of 8-dehydrocholesterol (8DHC) [4]. SLOS is a severe disorder, and 夞
C.H.L.S. was supported by Children’s Hospital Oakland Research Institute, and R.I.K. and L.E.K. are grateful for support from the National Organization of Rare Disorders. * Corresponding author. Tel.: ⫹1-510-428-3604; fax: ⫹1-510-4283608. E-mail address: [email protected] (C.H.L. Shackleton)
frequently a choice is made to terminate pregnancies associated with affected fetuses. There remains a critical need for noninvasive midgestation diagnostic tests for affected pregnancies, because current methodology relies on measurement of 7DHC/cholesterol ratios in amniotic fluid or chorionic villus cells [5,6]. Fetuses affected by SLOS use 7DHC or 8DHC precursors to produce steroid metabolites with a C7 or C8 double bond in addition to the conventional one at C5. This we have deduced from analysis of neonatal SLOS urine and serum where such metabolites are still dominant [7]. We hypothesized that in an affected fetus, fetal ⌬5,7- or ⌬5,8-pregnadienes would reach the placenta where they would be converted to 3-oxo-4,7- or 3-oxo-4,8(9)-dienes before being transferred to the maternal compartment, subjected to further metabolism, glucuronidation, and excretion in maternal urine. In normal pregnancy, characterizing nonestrogenic
0039-128X/99/$ – see front matter © 1999 Published by Elsevier Science Inc. All rights reserved. PII: S 0 0 3 9 - 1 2 8 X ( 9 9 ) 0 0 0 2 6 - 4
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
metabolites of fetal precursors in maternal urine is a needlein-a-haystack problem, because metabolites of fetal origin are quantitatively dwarfed by identical or structurally similar metabolites of placental progesterone. However, in the SLOS condition the metabolites sought should retain the 7 or 8 double bond, even after losing the 5 double bond through placental metabolism. The object of this study was to attempt characterization of pregnene metabolites of fetal origin in maternal urine. Metabolites such as these could be used as diagnostic markers of SLOS-affected pregnancies.
447
2.2. Materials Reference steroids were obtained from Steraloids (Wilton, NH, USA) or Sigma (St. Louis, MO, USA). Sephadex LH-20 was a product of Pharmacia Biotech AB (Uppsala, Sweden) and Sep-pak cartridges a product of Waters Corp (Milford, MA, USA). -Glucuronidase/sulfatase was obtained from Sigma and Boehringer (Mannheim, Germany). Trimethylsilylimidazole was obtained from Pierce Chemicals (Rockford, IL, USA) and methoxyamine hydrochloride and sodium bismuthate were obtained from Sigma. Lipidex 5000 was obtained from the Packard Instrument Company (Meriden, CT, USA).
2. Materials and methods 2.3. Steroid extraction and analysis 2.1. Patient Urine was collected from a patient suspected of carrying a fetus with SLOS. The pregnancy, the second for the healthy 35-year-old woman, was studied at 17 weeks’ gestation because of the sonographic findings of growth retardation, bilateral polydactyly of the hands, and an unidentifiable aortic arch. The maternal serum unconjugated estriol level was undetectable (⬍0.15 ng/ml; multiples of median, MoM ⬍0.23). Based on the physical anomalies and the undetectable maternal serum estriol level, a fetal diagnosis of SLOS was entertained. Amniotic fluid sent for sterol analysis showed a markedly increased level of 7DHC of 19 974 ng/ml (normal, 6 ⫾ 3 ng/ml; n ⫽ 89), consistent with the diagnosis of SLOS. The pregnancy was terminated at 18 weeks’ gestation and the diagnosis confirmed by sterol analysis of the fetal tissues, which showed a 7DHC/cholesterol ratio of 9.64 (normal, ⬍0.03).
Ten milliliters of the urine were extracted by Sep-pak C18 cartridge [8] and the extract was fractionated into free ⫹ glucuronide, monosulfate, and disulfate fractions by using the Sephadex LH-20 methodology of Ja¨nne et al. [9]. We reasoned that potential steroids of interest would be in the free ⫹ glucuronide fraction, so this was studied in depth. Although the fetal precursors reach the placenta principally as steroid sulfates, placental desulfation, conversion to 3-oxo-4-ene steroids, and maternal hepatic metabolism would ensure that the pregnadienes would be excreted principally as urinary steroid glucuronides. The separated steroid conjugate fractions were hydrolyzed with Helix pomatia -glucuronidase and sulfatase [10], and after extraction the freed steroids were fractionated by Sephadex LH-20 chromatography by using the method of Setchell and Shackleton [11]. Six-gram columns were used with cyclohexane/ethanol (4:1) as eluant. The
Table 1 Major components of Sephadex LH-20 chromatographic fractions of free ⫹ glucuronidated steroids: Presence of ring B unsaturated homologues indicated Fraction
Steroid
Fraction 2 12–22 ml
Androsterone Etiocholanolone Pregnanolones Pregnanediols
Fraction 3 22–37 ml
11-Oxo-etiocholanolone 11-Oxo-androsterone 11-Hydroxyetiocholanolone 16␣-Hydroxypregnanolone 17␣-Hydroxypregnanolone 6␣-Hydroxypregnanolones Pregnane-3,6,20-triols Pregnane-3,16,20-triols 5-Pregnane-3␣,17␣,20␣-triol 11-Oxo-Pregnanediol
⌬7 or ⌬8 Homologue
Fraction
Steroid
Fraction 4 37–55 ml
Hydroxyandrosterone Androstane-3,16,17-triols Pregnane-3,6,20-triols 11-Oxo-pregnanetriol Pregnane-3,20,21-triol Pregnanetetrol
Fraction 5 55–80 ml
Tetrahydrocortisone ␣-Cortolone -Cortolone
Fraction 6 80–160 ml
Tetrahydrocortisol 5␣-Tetrahydrocortisol ␣-Cortol -Cortol 1-Hydroxycortolone 6␣-Hydroxycortolone Estriol
⌬7 or ⌬8 Homologue
⫹
(⫹)
⫹ ⫹
⫹
448
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
following six fractions were collected: 0 –12, 12–22, 22–37, 37–55, 55– 80, and 80 –160 ml. A portion of each fraction was derivatized to form methyloxime trimethylsilyl (MOTMS) ethers, which were analyzed by gas chromatography/ mass spectrometry [10]. Certain fractions suspected of containing ⌬7-or ⌬8-unsaturated pregnenes were further subjected to bismuthate oxidation before derivatization as TMS and MO-TMS derivatives. 2.3.1. Bismuthate oxidation This was undertaken to oxidatively remove the side chains of 17-hydroxy pregnenes to provide 17-oxosteroids for use in confirmatory identification. To a 20% aliquot of the free ⫹ glucuronide fraction was added 0.1 ml of acetic acid, 0.1 ml of water, and 5 mg of sodium bismuthate. After sonication, the reaction tube was left in the dark for 2 h, diluted with 2 ml of 0.1 M acetate buffer, and extracted twice with 3 ml of methylene chloride. The methylene chloride extracts were dried under nitrogen and derivatized. 2.3.2. Gas chromatography/mass spectrometry conditions A Hewlett-Packard 5971 instrument was used with 15-m DB-1 capillary columns. Samples were injected splitless at 50°C and, after a 3-min delay, the oven was taken to a starting temperature of 210°C at a rate of 30°C/min. The temperature was programmed to 300°C at 3°C/min. After a 12-min delay, repetitive scanning was performed every 1.3 s over the mass range 90 to 650 amu.
3. Results Results will be presented according to the steroids found in the six fractions from Sephadex LH-20 chromatography. A nearly complete list of steroids characterized in the individual fractions is shown in Table 1. Fraction 1 (0 –12 ml) contained minor amounts of monosubstituted sterols and contaminants, and was not further studied. Fraction 2 (12–22 ml) contained di-substituted C19 and C21 steroids of which the principal constituents were pregnanolone and pregnanediol isomers. The steroids were entirely conventional and there was no evidence for the presence of unsaturated ⌬7 or ⌬8 homologues. Fraction 3 (22–37 ml) contained hydroxypregnanolones, certain pregnanetriols, and tri-substituted C19 steroids. Components included 6␣-hydroxypregnanolones, 16␣-hydroxypregnanolones, 17-hydroxypregnanolone, and pregnane-3,16,20-triols. A 16-hydroxydehydropregnanolone was tentatively identified, and because this was not 3,16␣dihydroxy-5-pregnen-20-one from retention time and mass spectrum, a ⌬7 or ⌬8 steroid must be considered. Pregnanetriol (5-pregnane-3␣,17␣,20␣-triol) was identified at the correct retention time but superimposed was a compound whose mass spectrum had pregnanetriol-like
Fig. 1. The identification of 5-pregn-7-ene-3␣,17␣,20␣-triol. The upper panel (A) is the spectrum obtained at the retention time of 5-pregnane3␣,17␣,20␣-triol tris-trimethylsilyl (TMS) ether in the SLOS pregnancy. The ion components m/z 253, 243 should not be present as illustrated in the spectrum of pregnanetriol in normal pregnancy urine (B, mirror image spectrum). Spectrum C was obtained by subtraction of (B) from (A) and represents the spectrum of steroid identified as the dehydropregnanetriol, probably ⌬7.
fragment ions of m/z values 2 amu less (Fig. 1). This is illustrated in the “mirror-image” spectra shown in Fig. 1A and B. The upper panel (A) is a spectrum of the pregnanetriol chromatographic peak from the SLOS pregnancy. The presence of the second component with fragment ions of 2 mass units less is clearly illustrated. The lower (reverse) spectrum (B) shows a pregnanetriol spectrum from the equivalent steroid fraction of a normal pregnancy, and the absence of the second component is clearly evident. As superimposition of pregnanetriol and an unsaturated pregnanetriol had never been seen in studies of normal pregnancy steroids, there was a strong likelihood that the new component was a ⌬7 or ⌬8 homologue of pregnanetriol. From the identity of retention time it was also likely that the unsaturated compound shared 3␣- and 20␣-hydroxyl
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
449
Fig. 2. 17-Ketosteroids produced by sodium bismuthate oxidation of the pregnanetriol fraction. (A) The spectrum of the methyloxime trimethylsilyl (MO-TMS) derivative of the steroid identified as etiocholanolone (M⫹391). (B) The MO-TMS derivative of 7-dehydroetiocholanolone (M⫹389). (C) The TMS derivative of etiocholanolone (M⫹362). (D) The TMS derivative of 7-dehydroetiocholanolone (M⫹360).
groups, and a 5-hydrogen. The spectra of 3,17,20-pregnanetriols are quite distinctive. They have important ions at m/z 117, representing C20 and C21, but the major fragments are those formed by the loss of 117 mass units. The small parent ion of pregnanetriol is at m/z 552, but a major ion (base peak) is at m/z 435 (M-117), and other important fragments are m/z 345 (M-117–90) and m/z 255 (M-117– 90-90). By subtracting the normal pregnancy ‘pregnanetriol,’ a spectrum of a purported ⌬7 or 8 pregnanetriol is obtained (Fig. 1C). This pregnenetriol spectrum has a lowintensity ion at m/z 433 (M-117) but strong ions at m/z 343 (M-117–90) and m/z 253 (M-117–90-90). Although this pair of compounds comprised the major pregnanetriol peak, an earlier eluting pair was present in a lesser amount. These probably represented the 3␣-hydroxy-5␣ epimers. To obtain confirmatory information on these structures, Fraction 3 was subjected to sodium bismuthate oxidation to remove the side chain of these pregnanetriols. Any di-functional C19 steroids that would subsequently be found in this fraction had to come from pregnanetriols, as none were present in Fraction 3 before oxidation. The bismuthate oxidation products were analyzed as MO-TMS and TMS ethers. Two pairs of steroids were found equivalent to androsterone and etiocholanolone and their unsaturated homologues. Representative spectra of etiocholanolone as MO-TMS ether and its ⌬7 or 8 homologue are shown in Fig. 2A and B, and spectra of the TMS derivatives are shown in Fig. 2C and D. The earlier
eluting androsterone and dehydroandrosterone pair showed essentially identical spectra to the etiocholanolones. In the mass spectrum of the MO-TMS derivative of etiocholanolone, the molecular ion is almost unseen at m/z 391, but major fragment ions are present at m/z 360 (M-31) and m/z 270 (M-31–90). This steroid had identical mass spectrum and retention time to authentic etiocholanolone. For the mass spectrum of the MO-TMS derivative of the steroid identified as dehydroetiocholanolone (Fig. 2B), the molecular ion is at m/z 389 and loss of the oxime is represented by the ion at m/z 358 (M-31). All these ions are small in comparison with the series representing loss of trimethylsilanol, i.e. m/z 299 (M-90), m/z 284 (M-90 –15), and m/z 268 (M-31–90). An ion of high abundance is seen at m/z 253 (M-31–90-15). The mass spectrum and retention time of etiocholanolone TMS ether were identical to the authentic compound. The molecular ion was at m/z 362 and important fragments at m/z 347 (M-15), m/z 272 (M-90), m/z 257 (M-15–90), and m/z 244 (M-118). The 118 mass unit fragment probably represents -90 –28; 28 mass units being CO or C2H4. The mass spectrum of the TMS derivative of ⌬7 or 8 etiocholanolone is remarkably similar to that of the fully saturated analogue, the intensity of the homologous ions (m/z and m/z ⫺2) being almost identical. The molecular ion of dehydroetiocholanolone is at m/z 360, base peak at m/z 270, and a prominent ion at m/z 242 (M-118). From the spectra of the native pregnane- and pregnene-
450
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
Fig. 3. Mass spectrum of steroid identified as 5-pregn-7-ene-3␣,16␣,20␣-triol (M⫹550). Two other steroids with essentially identical mass spectrum were found with different retention times.
triols and their C19 steroid oxidation products, we characterize the major compound as 5-pregn-7(or 8)-ene3␣,17␣,20␣-triol and the minor as 5␣-pregn-7(or 8)-ene3␣,17␣,20␣-triol. Confirmation of the location of the double bond would require isolation of a substantial amount of the materials to allow use of other physico-chemical techniques or synthesis of authentic compounds. In addition to the pregnene-3,17,20-triols characterized, three steroids with nearly identical spectra consistent with pregnene-3,16,20-triols were detected. In contrast to the pregnane-3,17,20-triol homologue, these steroids were separable from the conventional pregnancy steroids 5-pregnane-3␣,16␣,20␣-triol, 5␣-pregnane3␣,16␣,20␣-triol, and 5␣-pregnane-3,16␣,20␣-triol. A representative spectrum of the TMS ether is illustrated in Fig. 3. The molecular ion (unseen) is at m/z 550 and prominent ions are formed by loss of trimethylsilanol (-90), ie, m/z 460 (M-90), m/z 370 (M-90 –90), and m/z 280 (M-90 –90-90). Loss of 117 mass units is a minor fragmentation in pregnane-3,16,20-triols compared with pregnane-3,17,20-triols, although such an ion is seen at m/z 343 (M-90 –117). The m/z 117 ion representing the C-20,21 side chain is significant, proving the 21-deoxy nature of the steroid, but the most distinctive ion is to the base peak at m/z 143. This ion probably represents C16, C17, C20, C21 with a single remaining silylated hydroxyl. Further evidence that these steroids had a 16- rather than 17-hydroxyl comes from their stability to bismuthate oxidation, i.e. C19 steroids were not formed. By analogy to their saturated counterparts, we believe these steroids (in order of elution) have the structures 5-pregn-7(or 8)-ene3␣,16␣,20␣-triol, 5␣-pregn-7(or 8)-ene-3␣,16␣,20␣-triol, and 5␣-pregn-7(or 8)-ene-3,16␣,20␣-triol. 16␣-Hydroxylation appears to be exclusive for pregnanes in humans and 20␣reduction of 21-deoxypregnanes dominates 20-reduction. Fraction 4 (37–55 ml) contained primarily pregnane3,6,20-triols and no evidence was found for the presence of
dehydro counterparts. An important steroid detected was what we believe to be a 7-, or 8-pregnenetetrol, but further studies failed to confirm a structure. The molecular ion (Fig. 4) was probably at m/z 638 and base peak at m/z 431 (M-117–90). A fully saturated steroid with homologous mass spectrum was found in urine from a normal pregnancy, but this also was unknown. The presence of m/z 117 and apparent losses of 117 mass units strongly suggest these
Fig. 4. Unidentified pregnenetetrol. (A) Illustration a steroid provisionally identified as a 3,X,17,20-pregnenetetrol (M⫹638). It is not present in normal pregnancy urine, where it is replaced by an unidentified pregnanetetrol (B) of identical retention index.
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
451
Fig. 5. Proposed alternative pathways illustrating origin of maternal urinary 5-pregn-7-ene-3␣,17␣,20␣-triol. Similar schemes could be drawn for the 16␣-hydroxylated pregnenetriols and for equivalent ⌬8 steroids. Although the intermediates proposed have not been identified, steroids identical to or similar to those in the fetal compartment have been identified in neonatal urine and plasma from SLOS patients.
steroids are pregn-7(or 8)-ene-3,X,17,20-tetrol (SLOS pregnancy) and pregnane-3,X,17,20-tetrol (normal pregnancy). Remaining Sephadex LH-20 fractions (55– 80 and 80 – 160) contained primarily conventional metabolites of cortisol, i.e. tetrahydrocortisone, the tetrahydrocortisols, cortols, and cortolones. No evidence was found for the presence of unsaturated homologues of these steroids. However, in addition, unsaturated estriol was found in this fraction [12].
4. Discussion We have identified 7- or 8-dehydropregnanetriols of the 3,16,20 and 3,17,20 series in maternal urine obtained at midgestation from a woman carrying a SLOS fetus. These steroids were not detected in urine from six normal pregnancies, and therefore appear to be specific markers for the SLOS condition. These novel pregnenetriols presumably originated in the fetus, and that 7- or 8-unsaturation is retained suggests that the placenta and maternal organs do not have a 7-reductase capable of acting on their fetal precursors. That the 3,16,20triol and 3,17,20-triol structures were the only ones containing ring B unsaturation is not surprising in that we identified 7- and 8-dehydro homologues of 16␣-hydroxypregnenolone, 5-pregnene-3,16␣,20␣-triol, and 5-pregnene3,17␣,20␣-triol as the dominant steroids in neonatal urine from a SLOS neonate [7]. Because these steroids would
pass to the placenta and mother during pregnancy, they would act as natural precursors for the steroids we have identified in maternal urine (Fig. 5). It is significant that 16␣-hydroxypregnadienolones, 5-pregnene-3,16␣,20␣triol, and 5,7(and 8)-pregnadiene-3,17␣,20␣-triols were also major steroids in pooled neonatal and infantile plasma from SLOS patients aged 3 days to 6 months (unpublished data). In this pooled sample, the ratio of ⌬7 to ⌬8 varieties of the steroids was ⬇4:1. Unless the placental and maternal compartments have active 7,8-isomerase, the maternal urinary metabolites we identified are likely to be predominantly 7-unsaturated. The progesterone metabolites, namely, pregnanolones, 6␣-hydroxypregnanolones, pregnanediols, and pregnane3,6,20-triols, were excreted fully saturated. We found no evidence for dehydro homologues. This affirms that, although progesterone is synthesized fetoplacentally, it must use maternal cholesterol as exclusive precursor [13]. These new markers of SLOS-affected pregnancies are not the only ones in maternal urine. In a technically simpler study, we readily found 7-dehydro estriol in pregnancy urine [12] and proposed a 7-dehydroestriol/ estriol ratio may be diagnostic of the condition. The urine estrogen ratio, now backed by the presence of these new pregnenetriol markers, should provide reliable noninvasive discriminants for early confirmation of an affected fetus. If these maternal urinary steroid abnormalities can be reproduced in a prospective study of SLOS pregnan-
452
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
cies, then the use of this noninvasive diagnostic test for SLOS should be routinely undertaken. Finally, other studies have shown that 5-pregn-7-ene-3␣,17␣,20␣triol is present in urine from all children and adults with SLOS, and so is useful for diagnosis of the disorder throughout life.
References [1] Irons M, Elias ER, Salen G, Tint GS, Batta AK. Defective cholesterol biosynthesis in Smith-Lemli-Opitz syndrome. Lancet 1993;341:1414. [2] Cunniff C, Kratz LE, Moser A, Natowicz MR, Kelley RI. Clinical and biochemical spectrum of patients with RSH/Smith-Lemli-Opitz syndrome and abnormal cholesterol metabolism. Am J Med Genet 1997; 68:263–9. [3] Kelley RI. Smith-Lemli-Opitz syndrome. In: Moser HW, editor. Handbook of clinical neurology. Amsterdam: Elsevier Science BV, 1996. pp. 581–97. [4] Batta AK, Tint GS, Shefer S, Abuelo D, Salen G. Identification of 8-dehydrocholesterol (cholesta-5,8-dien-3-ol) in patients with SmithLemli-Opitz syndrome. J Lipid Res 1995;36:705–13.
[5] Irons MB, Tint GS. Prenatal diagnosis of Smith-Lemli-Opitz syndrome. Prenat Diagn 1998;18:369 –72. [6] Kratz LE, Kelley RI. Prenatal diagnosis of the RSH/Smith-Lemli-Opitz syndrome. Am J Med Genet 1999;82:376 – 81. [7] Shackleton CHL, Roitman E, Kelley R. Neonatal urinary steroids in Smith-Lemli-Opitz syndrome associated with 7-dehydrocholesterol reductase deficiency. Steroids 1999;64:481–90. [8] Shackleton CHL, Whitney JO. Use of Sep-pak® cartridges for urinary steroid extraction: evaluation of the method for use prior to gas chromatographic analysis. Clin Chim Acta 1980;107:231– 43. [9] Ja¨nne O, Vihko R, Sjo¨vall J, Sjo¨vall K. Determination of steroid monoand di-sulphates in human plasma. Clin Clin Acta 1969;23:405–12. [10] Shackleton CHL. Mass spectrometry in the diagnosis of steroid-related disorders and in hypertension research. J Steroid Biochem Mol Biol 1993;45:127– 40. [11] Setchell KDR, Shackleton CHL. The group separation of plasma and urinary steroids by column chromatography on Sephadex LH-20. Clin Chim Acta 1973;47:381– 8. [12] Shackleton CHL, Roitman E, Kratz LE, Kelley RI. Equine type estrogens produced by a pregnant woman carrying a Smith-Lemli-Opitz syndrome fetus. J Clin Endocrinol Metab 1999;84:1157–9. [13] Carr BR, Simpson ER. Cholesterol synthesis in human fetal tissues. J Clin Endocrinol Metab 1982;55:447–52.
Rapid communication
Midgestational maternal urine steroid markers of fetal Smith–Lemli– Opitz (SLO) syndrome (7-dehydrocholesterol 7-reductase deficiency)夞 Cedric H.L. Shackletona,*, Esther Roitmana, Lisa E. Kratzb, Richard I. Kelleyb a
Children’s Hospital Oakland Research Institute, 747 52nd Street, Oakland, CA 94609, USA Division of Neurogenetics, Kennedy Krieger Institute, The Johns Hopkins University, Baltimore, MD, 21215 USA
b
Received 26 January 1999; received in revised form 17 February 1999; accepted 18 February 1999
Abstract Smith–Lemli–Opitz syndrome (SLOS) is a malformation syndrome associated with 7-dehydrocholesterol (7DHC) 7-reductase deficiency. Although SLOS can be detected in an affected fetus before midpregnancy by measurement of 7DHC levels in amniotic fluid or chorionic villus cells, a noninvasive, more routine method is needed. Accordingly, this study was instigated to search for specific steroids in maternal urine in an affected pregnancy that reflect the 7-reductase deficiency of the fetus, ie, steroids retaining 7,8-unsaturation. Steroids were characterized by gas chromatography/mass spectrometry after urinary extraction, conjugate separation, and derivatization. Most steroids in maternal urine from a patient carrying a SLOS fetus were identified as progesterone metabolites, and these were entirely conventional, showing no evidence of additional unsaturation. Unsaturated homologues of the cortisol metabolites were also not detected. However, unsaturated homologues of pregnane-3,16,20-triols and pregnane-3,17,20-triol were found. Most likely, these are 7,8-unsaturated homologues, but 8,9-unsaturation is also possible because of the known activity of ⌬7-⌬8-isomerase on 7DHC, which results in 8DHC being a prominent sterol in SLOS. Among these novel human steroids, the following were provisionally characterized: 5-pregn-7(or 8)-ene3␣,17␣,20␣-triol, 5-pregn-7(or 8)-ene-3␣,16␣,20␣-triol, and 5␣-pregn-7(or 8)-ene-3,16␣,20␣-triol. Confirmation of the position of unsaturation will require steroid synthesis. These novel steroids are not present in normal pregnancy urine and, therefore, are valuable for prenatal diagnosis of SLOS. In addition, separate studies have shown that 5-pregn-7(or 8)-ene-3␣,17␣,20␣-triol is present in urine of children and adults with SLOS, and so is a useful analyte for confirmation of the disorder throughout life. © 1999 Published by Elsevier Science Inc. All rights reserved. Keywords: Smith–Lemli–Opitz syndrome; Pregnancy urine steroids; ⌬7-⌬8-Pregnenetriols; Gas chromatography/mass spectrometry
1. Introduction Smith–Lemli–Opitz syndrome (SLOS) is an autosomal recessive malformation disorder associated with 7-dehydrocholesterol (7DHC) 7-reductase deficiency [1,2]. Because this enzyme catalyzes the final step in cholesterol synthesis, a primary feature of the disorder is cholesterol deficiency [2,3] with concomitant buildup of 7DHC. Presence of an active ⌬7,⌬8-isomerase also causes production of 8-dehydrocholesterol (8DHC) [4]. SLOS is a severe disorder, and 夞
C.H.L.S. was supported by Children’s Hospital Oakland Research Institute, and R.I.K. and L.E.K. are grateful for support from the National Organization of Rare Disorders. * Corresponding author. Tel.: ⫹1-510-428-3604; fax: ⫹1-510-4283608. E-mail address: [email protected] (C.H.L. Shackleton)
frequently a choice is made to terminate pregnancies associated with affected fetuses. There remains a critical need for noninvasive midgestation diagnostic tests for affected pregnancies, because current methodology relies on measurement of 7DHC/cholesterol ratios in amniotic fluid or chorionic villus cells [5,6]. Fetuses affected by SLOS use 7DHC or 8DHC precursors to produce steroid metabolites with a C7 or C8 double bond in addition to the conventional one at C5. This we have deduced from analysis of neonatal SLOS urine and serum where such metabolites are still dominant [7]. We hypothesized that in an affected fetus, fetal ⌬5,7- or ⌬5,8-pregnadienes would reach the placenta where they would be converted to 3-oxo-4,7- or 3-oxo-4,8(9)-dienes before being transferred to the maternal compartment, subjected to further metabolism, glucuronidation, and excretion in maternal urine. In normal pregnancy, characterizing nonestrogenic
0039-128X/99/$ – see front matter © 1999 Published by Elsevier Science Inc. All rights reserved. PII: S 0 0 3 9 - 1 2 8 X ( 9 9 ) 0 0 0 2 6 - 4
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
metabolites of fetal precursors in maternal urine is a needlein-a-haystack problem, because metabolites of fetal origin are quantitatively dwarfed by identical or structurally similar metabolites of placental progesterone. However, in the SLOS condition the metabolites sought should retain the 7 or 8 double bond, even after losing the 5 double bond through placental metabolism. The object of this study was to attempt characterization of pregnene metabolites of fetal origin in maternal urine. Metabolites such as these could be used as diagnostic markers of SLOS-affected pregnancies.
447
2.2. Materials Reference steroids were obtained from Steraloids (Wilton, NH, USA) or Sigma (St. Louis, MO, USA). Sephadex LH-20 was a product of Pharmacia Biotech AB (Uppsala, Sweden) and Sep-pak cartridges a product of Waters Corp (Milford, MA, USA). -Glucuronidase/sulfatase was obtained from Sigma and Boehringer (Mannheim, Germany). Trimethylsilylimidazole was obtained from Pierce Chemicals (Rockford, IL, USA) and methoxyamine hydrochloride and sodium bismuthate were obtained from Sigma. Lipidex 5000 was obtained from the Packard Instrument Company (Meriden, CT, USA).
2. Materials and methods 2.3. Steroid extraction and analysis 2.1. Patient Urine was collected from a patient suspected of carrying a fetus with SLOS. The pregnancy, the second for the healthy 35-year-old woman, was studied at 17 weeks’ gestation because of the sonographic findings of growth retardation, bilateral polydactyly of the hands, and an unidentifiable aortic arch. The maternal serum unconjugated estriol level was undetectable (⬍0.15 ng/ml; multiples of median, MoM ⬍0.23). Based on the physical anomalies and the undetectable maternal serum estriol level, a fetal diagnosis of SLOS was entertained. Amniotic fluid sent for sterol analysis showed a markedly increased level of 7DHC of 19 974 ng/ml (normal, 6 ⫾ 3 ng/ml; n ⫽ 89), consistent with the diagnosis of SLOS. The pregnancy was terminated at 18 weeks’ gestation and the diagnosis confirmed by sterol analysis of the fetal tissues, which showed a 7DHC/cholesterol ratio of 9.64 (normal, ⬍0.03).
Ten milliliters of the urine were extracted by Sep-pak C18 cartridge [8] and the extract was fractionated into free ⫹ glucuronide, monosulfate, and disulfate fractions by using the Sephadex LH-20 methodology of Ja¨nne et al. [9]. We reasoned that potential steroids of interest would be in the free ⫹ glucuronide fraction, so this was studied in depth. Although the fetal precursors reach the placenta principally as steroid sulfates, placental desulfation, conversion to 3-oxo-4-ene steroids, and maternal hepatic metabolism would ensure that the pregnadienes would be excreted principally as urinary steroid glucuronides. The separated steroid conjugate fractions were hydrolyzed with Helix pomatia -glucuronidase and sulfatase [10], and after extraction the freed steroids were fractionated by Sephadex LH-20 chromatography by using the method of Setchell and Shackleton [11]. Six-gram columns were used with cyclohexane/ethanol (4:1) as eluant. The
Table 1 Major components of Sephadex LH-20 chromatographic fractions of free ⫹ glucuronidated steroids: Presence of ring B unsaturated homologues indicated Fraction
Steroid
Fraction 2 12–22 ml
Androsterone Etiocholanolone Pregnanolones Pregnanediols
Fraction 3 22–37 ml
11-Oxo-etiocholanolone 11-Oxo-androsterone 11-Hydroxyetiocholanolone 16␣-Hydroxypregnanolone 17␣-Hydroxypregnanolone 6␣-Hydroxypregnanolones Pregnane-3,6,20-triols Pregnane-3,16,20-triols 5-Pregnane-3␣,17␣,20␣-triol 11-Oxo-Pregnanediol
⌬7 or ⌬8 Homologue
Fraction
Steroid
Fraction 4 37–55 ml
Hydroxyandrosterone Androstane-3,16,17-triols Pregnane-3,6,20-triols 11-Oxo-pregnanetriol Pregnane-3,20,21-triol Pregnanetetrol
Fraction 5 55–80 ml
Tetrahydrocortisone ␣-Cortolone -Cortolone
Fraction 6 80–160 ml
Tetrahydrocortisol 5␣-Tetrahydrocortisol ␣-Cortol -Cortol 1-Hydroxycortolone 6␣-Hydroxycortolone Estriol
⌬7 or ⌬8 Homologue
⫹
(⫹)
⫹ ⫹
⫹
448
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
following six fractions were collected: 0 –12, 12–22, 22–37, 37–55, 55– 80, and 80 –160 ml. A portion of each fraction was derivatized to form methyloxime trimethylsilyl (MOTMS) ethers, which were analyzed by gas chromatography/ mass spectrometry [10]. Certain fractions suspected of containing ⌬7-or ⌬8-unsaturated pregnenes were further subjected to bismuthate oxidation before derivatization as TMS and MO-TMS derivatives. 2.3.1. Bismuthate oxidation This was undertaken to oxidatively remove the side chains of 17-hydroxy pregnenes to provide 17-oxosteroids for use in confirmatory identification. To a 20% aliquot of the free ⫹ glucuronide fraction was added 0.1 ml of acetic acid, 0.1 ml of water, and 5 mg of sodium bismuthate. After sonication, the reaction tube was left in the dark for 2 h, diluted with 2 ml of 0.1 M acetate buffer, and extracted twice with 3 ml of methylene chloride. The methylene chloride extracts were dried under nitrogen and derivatized. 2.3.2. Gas chromatography/mass spectrometry conditions A Hewlett-Packard 5971 instrument was used with 15-m DB-1 capillary columns. Samples were injected splitless at 50°C and, after a 3-min delay, the oven was taken to a starting temperature of 210°C at a rate of 30°C/min. The temperature was programmed to 300°C at 3°C/min. After a 12-min delay, repetitive scanning was performed every 1.3 s over the mass range 90 to 650 amu.
3. Results Results will be presented according to the steroids found in the six fractions from Sephadex LH-20 chromatography. A nearly complete list of steroids characterized in the individual fractions is shown in Table 1. Fraction 1 (0 –12 ml) contained minor amounts of monosubstituted sterols and contaminants, and was not further studied. Fraction 2 (12–22 ml) contained di-substituted C19 and C21 steroids of which the principal constituents were pregnanolone and pregnanediol isomers. The steroids were entirely conventional and there was no evidence for the presence of unsaturated ⌬7 or ⌬8 homologues. Fraction 3 (22–37 ml) contained hydroxypregnanolones, certain pregnanetriols, and tri-substituted C19 steroids. Components included 6␣-hydroxypregnanolones, 16␣-hydroxypregnanolones, 17-hydroxypregnanolone, and pregnane-3,16,20-triols. A 16-hydroxydehydropregnanolone was tentatively identified, and because this was not 3,16␣dihydroxy-5-pregnen-20-one from retention time and mass spectrum, a ⌬7 or ⌬8 steroid must be considered. Pregnanetriol (5-pregnane-3␣,17␣,20␣-triol) was identified at the correct retention time but superimposed was a compound whose mass spectrum had pregnanetriol-like
Fig. 1. The identification of 5-pregn-7-ene-3␣,17␣,20␣-triol. The upper panel (A) is the spectrum obtained at the retention time of 5-pregnane3␣,17␣,20␣-triol tris-trimethylsilyl (TMS) ether in the SLOS pregnancy. The ion components m/z 253, 243 should not be present as illustrated in the spectrum of pregnanetriol in normal pregnancy urine (B, mirror image spectrum). Spectrum C was obtained by subtraction of (B) from (A) and represents the spectrum of steroid identified as the dehydropregnanetriol, probably ⌬7.
fragment ions of m/z values 2 amu less (Fig. 1). This is illustrated in the “mirror-image” spectra shown in Fig. 1A and B. The upper panel (A) is a spectrum of the pregnanetriol chromatographic peak from the SLOS pregnancy. The presence of the second component with fragment ions of 2 mass units less is clearly illustrated. The lower (reverse) spectrum (B) shows a pregnanetriol spectrum from the equivalent steroid fraction of a normal pregnancy, and the absence of the second component is clearly evident. As superimposition of pregnanetriol and an unsaturated pregnanetriol had never been seen in studies of normal pregnancy steroids, there was a strong likelihood that the new component was a ⌬7 or ⌬8 homologue of pregnanetriol. From the identity of retention time it was also likely that the unsaturated compound shared 3␣- and 20␣-hydroxyl
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
449
Fig. 2. 17-Ketosteroids produced by sodium bismuthate oxidation of the pregnanetriol fraction. (A) The spectrum of the methyloxime trimethylsilyl (MO-TMS) derivative of the steroid identified as etiocholanolone (M⫹391). (B) The MO-TMS derivative of 7-dehydroetiocholanolone (M⫹389). (C) The TMS derivative of etiocholanolone (M⫹362). (D) The TMS derivative of 7-dehydroetiocholanolone (M⫹360).
groups, and a 5-hydrogen. The spectra of 3,17,20-pregnanetriols are quite distinctive. They have important ions at m/z 117, representing C20 and C21, but the major fragments are those formed by the loss of 117 mass units. The small parent ion of pregnanetriol is at m/z 552, but a major ion (base peak) is at m/z 435 (M-117), and other important fragments are m/z 345 (M-117–90) and m/z 255 (M-117– 90-90). By subtracting the normal pregnancy ‘pregnanetriol,’ a spectrum of a purported ⌬7 or 8 pregnanetriol is obtained (Fig. 1C). This pregnenetriol spectrum has a lowintensity ion at m/z 433 (M-117) but strong ions at m/z 343 (M-117–90) and m/z 253 (M-117–90-90). Although this pair of compounds comprised the major pregnanetriol peak, an earlier eluting pair was present in a lesser amount. These probably represented the 3␣-hydroxy-5␣ epimers. To obtain confirmatory information on these structures, Fraction 3 was subjected to sodium bismuthate oxidation to remove the side chain of these pregnanetriols. Any di-functional C19 steroids that would subsequently be found in this fraction had to come from pregnanetriols, as none were present in Fraction 3 before oxidation. The bismuthate oxidation products were analyzed as MO-TMS and TMS ethers. Two pairs of steroids were found equivalent to androsterone and etiocholanolone and their unsaturated homologues. Representative spectra of etiocholanolone as MO-TMS ether and its ⌬7 or 8 homologue are shown in Fig. 2A and B, and spectra of the TMS derivatives are shown in Fig. 2C and D. The earlier
eluting androsterone and dehydroandrosterone pair showed essentially identical spectra to the etiocholanolones. In the mass spectrum of the MO-TMS derivative of etiocholanolone, the molecular ion is almost unseen at m/z 391, but major fragment ions are present at m/z 360 (M-31) and m/z 270 (M-31–90). This steroid had identical mass spectrum and retention time to authentic etiocholanolone. For the mass spectrum of the MO-TMS derivative of the steroid identified as dehydroetiocholanolone (Fig. 2B), the molecular ion is at m/z 389 and loss of the oxime is represented by the ion at m/z 358 (M-31). All these ions are small in comparison with the series representing loss of trimethylsilanol, i.e. m/z 299 (M-90), m/z 284 (M-90 –15), and m/z 268 (M-31–90). An ion of high abundance is seen at m/z 253 (M-31–90-15). The mass spectrum and retention time of etiocholanolone TMS ether were identical to the authentic compound. The molecular ion was at m/z 362 and important fragments at m/z 347 (M-15), m/z 272 (M-90), m/z 257 (M-15–90), and m/z 244 (M-118). The 118 mass unit fragment probably represents -90 –28; 28 mass units being CO or C2H4. The mass spectrum of the TMS derivative of ⌬7 or 8 etiocholanolone is remarkably similar to that of the fully saturated analogue, the intensity of the homologous ions (m/z and m/z ⫺2) being almost identical. The molecular ion of dehydroetiocholanolone is at m/z 360, base peak at m/z 270, and a prominent ion at m/z 242 (M-118). From the spectra of the native pregnane- and pregnene-
450
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
Fig. 3. Mass spectrum of steroid identified as 5-pregn-7-ene-3␣,16␣,20␣-triol (M⫹550). Two other steroids with essentially identical mass spectrum were found with different retention times.
triols and their C19 steroid oxidation products, we characterize the major compound as 5-pregn-7(or 8)-ene3␣,17␣,20␣-triol and the minor as 5␣-pregn-7(or 8)-ene3␣,17␣,20␣-triol. Confirmation of the location of the double bond would require isolation of a substantial amount of the materials to allow use of other physico-chemical techniques or synthesis of authentic compounds. In addition to the pregnene-3,17,20-triols characterized, three steroids with nearly identical spectra consistent with pregnene-3,16,20-triols were detected. In contrast to the pregnane-3,17,20-triol homologue, these steroids were separable from the conventional pregnancy steroids 5-pregnane-3␣,16␣,20␣-triol, 5␣-pregnane3␣,16␣,20␣-triol, and 5␣-pregnane-3,16␣,20␣-triol. A representative spectrum of the TMS ether is illustrated in Fig. 3. The molecular ion (unseen) is at m/z 550 and prominent ions are formed by loss of trimethylsilanol (-90), ie, m/z 460 (M-90), m/z 370 (M-90 –90), and m/z 280 (M-90 –90-90). Loss of 117 mass units is a minor fragmentation in pregnane-3,16,20-triols compared with pregnane-3,17,20-triols, although such an ion is seen at m/z 343 (M-90 –117). The m/z 117 ion representing the C-20,21 side chain is significant, proving the 21-deoxy nature of the steroid, but the most distinctive ion is to the base peak at m/z 143. This ion probably represents C16, C17, C20, C21 with a single remaining silylated hydroxyl. Further evidence that these steroids had a 16- rather than 17-hydroxyl comes from their stability to bismuthate oxidation, i.e. C19 steroids were not formed. By analogy to their saturated counterparts, we believe these steroids (in order of elution) have the structures 5-pregn-7(or 8)-ene3␣,16␣,20␣-triol, 5␣-pregn-7(or 8)-ene-3␣,16␣,20␣-triol, and 5␣-pregn-7(or 8)-ene-3,16␣,20␣-triol. 16␣-Hydroxylation appears to be exclusive for pregnanes in humans and 20␣reduction of 21-deoxypregnanes dominates 20-reduction. Fraction 4 (37–55 ml) contained primarily pregnane3,6,20-triols and no evidence was found for the presence of
dehydro counterparts. An important steroid detected was what we believe to be a 7-, or 8-pregnenetetrol, but further studies failed to confirm a structure. The molecular ion (Fig. 4) was probably at m/z 638 and base peak at m/z 431 (M-117–90). A fully saturated steroid with homologous mass spectrum was found in urine from a normal pregnancy, but this also was unknown. The presence of m/z 117 and apparent losses of 117 mass units strongly suggest these
Fig. 4. Unidentified pregnenetetrol. (A) Illustration a steroid provisionally identified as a 3,X,17,20-pregnenetetrol (M⫹638). It is not present in normal pregnancy urine, where it is replaced by an unidentified pregnanetetrol (B) of identical retention index.
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
451
Fig. 5. Proposed alternative pathways illustrating origin of maternal urinary 5-pregn-7-ene-3␣,17␣,20␣-triol. Similar schemes could be drawn for the 16␣-hydroxylated pregnenetriols and for equivalent ⌬8 steroids. Although the intermediates proposed have not been identified, steroids identical to or similar to those in the fetal compartment have been identified in neonatal urine and plasma from SLOS patients.
steroids are pregn-7(or 8)-ene-3,X,17,20-tetrol (SLOS pregnancy) and pregnane-3,X,17,20-tetrol (normal pregnancy). Remaining Sephadex LH-20 fractions (55– 80 and 80 – 160) contained primarily conventional metabolites of cortisol, i.e. tetrahydrocortisone, the tetrahydrocortisols, cortols, and cortolones. No evidence was found for the presence of unsaturated homologues of these steroids. However, in addition, unsaturated estriol was found in this fraction [12].
4. Discussion We have identified 7- or 8-dehydropregnanetriols of the 3,16,20 and 3,17,20 series in maternal urine obtained at midgestation from a woman carrying a SLOS fetus. These steroids were not detected in urine from six normal pregnancies, and therefore appear to be specific markers for the SLOS condition. These novel pregnenetriols presumably originated in the fetus, and that 7- or 8-unsaturation is retained suggests that the placenta and maternal organs do not have a 7-reductase capable of acting on their fetal precursors. That the 3,16,20triol and 3,17,20-triol structures were the only ones containing ring B unsaturation is not surprising in that we identified 7- and 8-dehydro homologues of 16␣-hydroxypregnenolone, 5-pregnene-3,16␣,20␣-triol, and 5-pregnene3,17␣,20␣-triol as the dominant steroids in neonatal urine from a SLOS neonate [7]. Because these steroids would
pass to the placenta and mother during pregnancy, they would act as natural precursors for the steroids we have identified in maternal urine (Fig. 5). It is significant that 16␣-hydroxypregnadienolones, 5-pregnene-3,16␣,20␣triol, and 5,7(and 8)-pregnadiene-3,17␣,20␣-triols were also major steroids in pooled neonatal and infantile plasma from SLOS patients aged 3 days to 6 months (unpublished data). In this pooled sample, the ratio of ⌬7 to ⌬8 varieties of the steroids was ⬇4:1. Unless the placental and maternal compartments have active 7,8-isomerase, the maternal urinary metabolites we identified are likely to be predominantly 7-unsaturated. The progesterone metabolites, namely, pregnanolones, 6␣-hydroxypregnanolones, pregnanediols, and pregnane3,6,20-triols, were excreted fully saturated. We found no evidence for dehydro homologues. This affirms that, although progesterone is synthesized fetoplacentally, it must use maternal cholesterol as exclusive precursor [13]. These new markers of SLOS-affected pregnancies are not the only ones in maternal urine. In a technically simpler study, we readily found 7-dehydro estriol in pregnancy urine [12] and proposed a 7-dehydroestriol/ estriol ratio may be diagnostic of the condition. The urine estrogen ratio, now backed by the presence of these new pregnenetriol markers, should provide reliable noninvasive discriminants for early confirmation of an affected fetus. If these maternal urinary steroid abnormalities can be reproduced in a prospective study of SLOS pregnan-
452
C.H.L. Shackleton et al. / Steroids 64 (1999) 446 – 452
cies, then the use of this noninvasive diagnostic test for SLOS should be routinely undertaken. Finally, other studies have shown that 5-pregn-7-ene-3␣,17␣,20␣triol is present in urine from all children and adults with SLOS, and so is useful for diagnosis of the disorder throughout life.
References [1] Irons M, Elias ER, Salen G, Tint GS, Batta AK. Defective cholesterol biosynthesis in Smith-Lemli-Opitz syndrome. Lancet 1993;341:1414. [2] Cunniff C, Kratz LE, Moser A, Natowicz MR, Kelley RI. Clinical and biochemical spectrum of patients with RSH/Smith-Lemli-Opitz syndrome and abnormal cholesterol metabolism. Am J Med Genet 1997; 68:263–9. [3] Kelley RI. Smith-Lemli-Opitz syndrome. In: Moser HW, editor. Handbook of clinical neurology. Amsterdam: Elsevier Science BV, 1996. pp. 581–97. [4] Batta AK, Tint GS, Shefer S, Abuelo D, Salen G. Identification of 8-dehydrocholesterol (cholesta-5,8-dien-3-ol) in patients with SmithLemli-Opitz syndrome. J Lipid Res 1995;36:705–13.
[5] Irons MB, Tint GS. Prenatal diagnosis of Smith-Lemli-Opitz syndrome. Prenat Diagn 1998;18:369 –72. [6] Kratz LE, Kelley RI. Prenatal diagnosis of the RSH/Smith-Lemli-Opitz syndrome. Am J Med Genet 1999;82:376 – 81. [7] Shackleton CHL, Roitman E, Kelley R. Neonatal urinary steroids in Smith-Lemli-Opitz syndrome associated with 7-dehydrocholesterol reductase deficiency. Steroids 1999;64:481–90. [8] Shackleton CHL, Whitney JO. Use of Sep-pak® cartridges for urinary steroid extraction: evaluation of the method for use prior to gas chromatographic analysis. Clin Chim Acta 1980;107:231– 43. [9] Ja¨nne O, Vihko R, Sjo¨vall J, Sjo¨vall K. Determination of steroid monoand di-sulphates in human plasma. Clin Clin Acta 1969;23:405–12. [10] Shackleton CHL. Mass spectrometry in the diagnosis of steroid-related disorders and in hypertension research. J Steroid Biochem Mol Biol 1993;45:127– 40. [11] Setchell KDR, Shackleton CHL. The group separation of plasma and urinary steroids by column chromatography on Sephadex LH-20. Clin Chim Acta 1973;47:381– 8. [12] Shackleton CHL, Roitman E, Kratz LE, Kelley RI. Equine type estrogens produced by a pregnant woman carrying a Smith-Lemli-Opitz syndrome fetus. J Clin Endocrinol Metab 1999;84:1157–9. [13] Carr BR, Simpson ER. Cholesterol synthesis in human fetal tissues. J Clin Endocrinol Metab 1982;55:447–52.