CASE REPORT

Liver disease in infancy caused by oxysterol 7α-hydroxylasedeficiency: successful treatment with chenodeoxycholic acid

Dongling Dai & Philippa B. Mills & Emma Footitt & Paul Gissen & Patricia McClean &

Jens Stahlschmidt & Isabelle Coupry & Julie Lavie & Fanny Mochel & Cyril Goizet &Tatsuki Mizuochi & Akihiko Kimura & Hiroshi Nittono & Karin Schwarz & Peter J. Crick &

Yuqin Wang & William J. Griffiths & Peter T. Clayton

Received: 12 December 2013 /Revised: 14 February 2014 /Accepted: 17 February 2014# SSIEM and Springer Science+Business Media Dordrecht 2014

Abstract A child of consanguineous parents of Pakistaniorigin developed jaundice at 5 weeks and then, at 3 months,irritability, a prolonged prothrombin time, a low albumin, andepisodes of hypoglycaemia. Investigation showed an elevatedalanine aminotransferase with a normal γ-glutamyl-transpeptidase. Analysis of urine by electrospray ionisation

tandem mass spectrometry (ESI-MS/MS) showed that themajor peaks were m/z 480 (taurine-conjugated 3β-hydroxy-5-cholenoic acid) and m/z 453 (sulphated 3β-hydroxy-5-cholenoic acid). Analysis of plasma by gas chromatography-mass spectrometry (GC-MS) showed increased concentra-tions of 3β-hydroxy-5-cholenoic acid, 3β-hydroxy-5-

Communicated by: K. Michael Gibson

Nomenclature and abbreviations Some of the bile acids and oxysterolsin this paper are referred to by the names in common use in the medicalliterature rather than the IUPAC recommended names. Thus 27-hydroxycholesterol is cholest-(25R)-5-ene-3β,26-diol / (25R)26-hydroxycholesterol; 7α,27-dihydroxycholesterol is cholest-(25R)-5-ene3β,7α,26-triol; 7α-hydroxycholesterol is cholest-5-ene-3β,7α-diol; 3β-hydroxy-5-cholestenoic acid is 3β-hydroxycholest-5-en-26-oic acid; 3β-hydroxy-5-cholenoic acid is 3β-hydroxychol-5-en-24-oic acid.Chenodeoxycholic acid (CDCA) is 3α,7α-dihydroxy-5β-cholan-24-oicacid; ursodeoxycholic acid (UDCA) is 3α,7β-dihydroxy-5β-cholan-24-oic acid; cholic acid is 3α,7α,12α-trihydroxy-5β-cholan-24-oic acid.

Electronic supplementary material The online version of this article(doi:10.1007/s10545-014-9695-6) contains supplementary material,which is available to authorized users.

D. Dai : P. B. Mills : E. Footitt : P. Gissen : P. T. Clayton (*)Clinical andMolecular Genetics Unit, UCL Institute of Child Health,London WC1N 1EH, UKe-mail: peter.clayton@ucl.ac.uk

D. DaiDepartment of Gastroenterology, Shenzhen Children’s Hospital,7019 YiTian Road, FuTian District, Shenzhen, China 518026

P. McClean : J. StahlschmidtChildren’s Liver Unit and Department of Histopathology, LeedsTeaching Hospitals NHS Trust, Leeds, UK

I. Coupry : J. Lavie : C. GoizetLaboratoire Maladies Rares: Génétique et Métabolisme (MRGM),EA 4576 2ème étage Ecole de Sage-Femmes, Hôpital Pellegrin,33076 Bordeaux Cedex, France

F. MochelInserm UMR S975/Department of Genetics/University Pierre etMarie Curie, Hôpital de La Salpêtrière, 47 Bd de l’Hôpital,75013 Paris, France

C. GoizetCHU Bordeaux, Service de Génétique Médicale, Bordeaux, France

T. Mizuochi :A. KimuraDepartment of Pediatrics and Child Health, Kurume UniversitySchool of Medicine, Kurume, Fukuoka, Japan

H. NittonoJunshin Clinic Bile Acid Institute, Meguro-ku, Tokyo, Japan

K. SchwarzNeonatal Unit, CRH, Salterhebble Road, Halifax HX3 0PD, UK

J Inherit Metab DisDOI 10.1007/s10545-014-9695-6

cholestenoic acid and 27-hydroxycholesterol, indicatingoxysterol 7α-hydroxylase deficiency. The patient was homo-zygous for a mutation (c.1249C>T) in CYP7B1 that alters ahighly conserved residue in oxysterol 7α-hydroxylase(p.R417C) - previously reported in a family with hereditaryspastic paraplegia type 5. On treatment with ursodeoxycholicacid (UDCA), his condition was worsening, but onchenodeoxycholic acid (CDCA), 15 mg/kg/d, he improvedrapidly. A biopsy (after 2 weeks on CDCA), showed a giantcell hepatitis, an evolving micronodular cirrhosis, andsteatosis. The improvement in liver function on CDCA wasassociated with a drop in the plasma concentrations and uri-nary excretions of the 3β-hydroxy-Δ5 bile acids which areconsidered hepatotoxic. At age 5 years (on CDCA, 6 mg/kg/d), he was thriving with normal liver function. Neurologicaldevelopment was normal apart from a tendency to trip.Examination revealed pes cavus but no upper motor neuronsigns. The findings in this case suggest that CDCA can reducethe activity of cholesterol 27-hydroxylase - the first step in theacidic pathway for bile acid synthesis.

Introduction

Inborn errors of bile acid synthesis can produce severe chole-static liver disease in infancy and progressive neurologicaldisease later in childhood or in adult life. Cholestatic liverdisease in 3β-hydroxy-Δ5-C27-steroid dehydrogenase defi-ciency andΔ4-3-oxosteroid 5β-reductase deficiency and neu-ropsychiatric disease in cerebrotendinous xanthomatosis(CTX) can often be treated very effectively with bile acidreplacement therapy and it is important to diagnose thesedisorders early (Clayton 2011). Synthesis of bile acids fromcholesterol occurs via two main pathways: the neutral path-way involving cholesterol 7α-hydroxylase (encoded byCYP7A1) and the acidic pathway, using a microsomaloxysterol 7α-hydroxylase encoded by CYP7B1 (Setchellet al 1998; Stiles et al 2009) (Fig. 1). Oxysterol 7α-hydroxylase converts 27-hydroxycholesterol to 7α,27-dihydroxycholesterol in the liver and in extrahepatic tissuesincluding the brain (Meaney et al 2007). The first describedpatient with oxysterol 7α-hydroxylase deficiency due to non-sense mutations in CYP7B1 (R388X/R388X) presented withsevere liver disease in early infancy and had marked accumu-lation of 27-hydroxycholesterol and bile acids derived from it(without 7α-hydroxylation) - 3β-hydroxy-5-cholestenoic acidand 3β-hydroxy-5-cholenoic acid (Setchell et al 1998). Theproposed mechanism of the liver disease was accumulation oftoxic 3β-hydroxy-Δ5 bile acids (Fig. 1). Conjugates of 3β-

hydroxy-5-cholenoic acid induce cholestatic liver disease inrats (Mathis et al 1983). The failure of production of bile acidsby the acidic pathway in the infant reported by Setchell et al(1998) was not associated with increased activity of choles-terol 7α-hydroxylase (the rate-limiting step for the classicpathway); 7α-hydroxycholesterol was actually undetectablein plasma, raising the possibility of a secondary inhibition ofcholesterol 7α-hydroxylase or an inability to upregulateCYP7A1 expression in early infancy. Treatment withursodeoxycholic acid (UDCA) and then cholic acid, wereineffective at preventing disease progression or reducing con-centrations of the 3β-hydroxy-Δ5 bile acids. A second patientwith fatal infantile liver disease due to CYP7B1 nonsensemutations (R112X/R112X) was reported by Ueki et al in2008. This patient failed to respond to UDCA treatment. Athird patient with CYP7B1 mutations (R112X/R417C) failedto respond to treatment with UDCA but underwent a success-ful liver transplant (Mizuochi et al 2011).

In 2008, Tsaousidou et al, using linkage analysis, identifiedmutations in CYP7B1 as a cause of hereditary spastic paraple-gia (HSP). HSP is a heterogeneous group of neurodegenera-tive diseases characterized by progressive spasticity in thelower limbs due to degeneration of upper motor neurons(Fink 2006). All modes of inheritance have been reported:autosomal dominant, autosomal recessive, or X-linked, andeach is associated with multiple genes or loci. In autosomalrecessive SPG5 (OMIM 603711), Tsaousidou et al demon-strated mutations in CYP7B1. The age of onset ranged from 1to 40 years and upper motor neuron signs were accompaniedby posterior column impairment (diminished vibration sensa-tion and proprioception and bladder dysfunction). The authorssuggested that this phenotype was the true expression of lossof function of the CYP7B1 gene product and that patients withneonatal onset liver disease probably had additional mutationsin CYP7A1 (hence the low 7α-hydroxycholesterol). (Theymade this suggestion despite the fact that Setchell et al hadshown there were no mutations in the coding regions ofCYP7A1.) The cause of the neuropathology in SPG5 wasnot immediately clear but Tsaousidou et al suggested twopossibilities:- i) Impairment of oxysterol - liver X receptor(LXR)-α signalling in the central nervous system. The authorsstate that the LXRα knockout mouse develops lower motorneuron disease, however, the paper they refer to describesmotor neuron degeneration in the LXRβ knockout mouse(Andersson et al 2005); ii) Impaired conversion of dehydro-epiandrosterone (DHEA) to 7α-hydroxy-DHEA in the brain(Fig. 1) with loss of the neuroprotective function of the lattercompound.

Further investigation of the role of LXR receptors hassuggested that motor neuron survival is reduced by the weakLXR ligand, 3β-hydroxy-5-cholestenoic acid and enhancedby another ligand 3β,7α-dihydroxy-5-cholestenoic acid. InSPG5 patients, levels of 3β-hydroxy-5-cholestenoic acid are

P. J. Crick :Y. Wang :W. J. GriffithsInstitute of Mass Spectrometry, College of Medicine, SwanseaUniversity, Swansea SA2 8PP, UK

J Inherit Metab Dis

increased and levels of 3β,7α-dihydroxy-5-cholestenoic acidare reduced (Theofilopoulos et al JCI, under review).

The fact that some patients with CYP7B1 mutations havepresented with fatal liver failure and other patients have pre-sented later in life with HSP cannot be explained by genotype.Homozygous R112X and R388X mutations have been de-scribed for both presentations (Stiles et al 2009).

This paper describes the first UK case presenting with liverdisease as a result of oxysterol 7α-hydroxylase deficiency dueto CYP7B1 mutations. Previous patients have failed to re-spond to treatment with UDCA or cholic acid but, in this case,a posit ive response to bile acid replacement(chenodeoxycholic acid) was documented. Furthermore,while the patient had a mutation previously implicated as acause of HSP, the patient is currently free from signs of uppermotor neuron dysfunction at 5 years. Further research will beneeded to determine whether CDCA can consistently reverseliver disease and prevent upper motor neuron damage inpatients with CYP7B1 mutations.

Patient

This boy, the second child of consanguineous parents ofPakistani origin, had a birth weight and length between the50th and 75th centiles. He presented to his local hospital at5 weeks with mainly unconjugated hyperbilirubinaemia andthen, over the next 2 months the conjugated fraction rose. At3–4 months his prothrombin time (PT) was prolonged, then,

after a set of immunizations he decompensated with irritabil-ity, deranged clotting and low albumin, and was referred to theChildren’s Liver Unit at Leeds Teaching Hospitals. On phys-ical examination, he was still on the 75th centile for length buthis weight had fallen to the 25th centile. He was alert anddeveloping normally. He was mildly icteric with a normal sizeliver and spleen and normal muscle tone. An abdominalultrasound scan (USS) was normal. He was found to behaving episodes of hypoglycaemia and after an initial periodof 2 weeks when his liver function tests and clotting timesremained static they started to deteriorate further.

Laboratory investigations showed no evidence of bacteri-al or viral infection and no endocrine abnormality. Metabolictesting showed no evidence of a disorder of amino acid orcarbohydrate metabolism or of mitochondrial disease. Hisbilirubin was 130 μmol/L, ALT 60–70 U/L, and GGTnormal at 15–20 U/l. His albumin was low at 28 g/L andprothrombin time prolonged at 26 seconds. Fat-soluble vita-min levels were all low. On day 2 of his admission in Leedshis blood sugar fell to 2.0 mM, with plasma cortisol 673nM, and lactate 2.64 mM (mildly elevated). The acyl carni-tine profile was normal.

Analysis of a urine sample by ESI-MS/MS showed that themajor urinary bile acid was 3β-hydroxy-5-cholenoic acid,present principally as the taurine conjugate (m/z 480), andthe sulphate (m/z 453). Analysis of non-sulphated plasma byGC-MS showed the presence of increased amounts of 27-hydroxycholesterol, 3β-hydroxy-5-cholestenoic acid and3β-hydroxy-5-cholenoic acid.

Fig. 1 Simplified scheme of the two major pathways for the synthesis of bile acids and the pathway for synthesis of neurosteroids from cholesterol.Mutations in CYP7B1 lead to over-production of 27-hydroxycholesterol, 3β-hydroxy5-cholestenoic acid and 3β-hydroxy-5-cholenoic acid

J Inherit Metab Dis

He had been treated with ursodeoxycholic (UDCA) from4 months (before his blood samples for bile acid analysis weretaken), but he was slowly worsening, so at 4.5months he wasstarted on CDCA 15 mg/kg/d, and within two days he startedto improve. By 6.5months his weight was back up to the 50thcentile.

A liver biopsy was undertaken 2 weeks after startingCDCAwhen his clotting had returned to normal. This showedadvanced fibrosis with an evolving micronodular cirrhoticpattern (Fig. 2). The lobules showed features of a giant cellhepatitis with lobular disarray and liver cell rosettes. Therewas also widespread micro and macrovesicular fatty changeand paraseptal copper-associated protein.

Liver function tests were essentially normal by 7 monthswhen the UDCA was stopped, and remained so thereafter.Followed up to 5 years old, on CDCA (now 6 mg/kg/d), hehad developed well. He had no jaundice and no stigmata ofchronic liver disease. His weight and height were on the50th centile and he had normal liver function tests.However, he was noticed to have a symptom that could bedue to central nervous system dysfunction; he tripped a lotwhen he ran. Examination by a paediatric neurologistshowed he had pes cavus but normal power, tone andreflexes in both arms and legs; there was no evidence ofspastic paraparesis. His tendency to trip improved when hewas given insoles. There had been no further motor prob-lems when he was seen at 6.5years.

Sibling

The family decided to have a third child and have him/hertested after birth. A daughter born at term had a normal urinebile acid spectrum. She has remained asymptomatic with nosigns of liver dysfunction.

Methods

Urine bile acids

Cholanoids (bile acids and bile alcohols) in urine wereanalysed by negative ion electrospray ionization mass spec-trometry (ESI-MS and ESI-MS/MS; Quattro Micro,Micromass, Waters, UK) as described previously (Mills et al1998). A mass-to-charge ratio (m/z) range of 350–700 wasused for both the MS direct scan and the parent ion scans.

Plasma bile acids

Plasma bile acid analysis by gas chromatography–mass spec-trometry of methyl ester trimethylsilyl TMS ether derivativeswas undertaken as described (Clayton and Muller 1980;Clayton et al 1987). Control ranges for plasma bile-acidconcentrations in normal infants have been documented(Clayton 1983; Clayton et al 1996). The method as employedroutinely in our laboratory does not involve cleavage of sul-phates by solvolysis and so measures the non-sulphated frac-tion. The recovery of a monohydroxy bile acid (lithocholicacid) was 88 % in the original version of this method (Clayton1983), however, replacement of benzene by toluene for thealumina chromatography step may have reduced this and therecoveries of 3β-hydroxy-5-cholenoic acid and 3β-hydroxy-5-cholestenoic acid have not been checked.

Plasma oxysterols

Plasma oxysterols and 3β-hydroxy-Δ5 bile acids wereanalysed by enzyme-assisted derivatisation for sterol analysis(EADSA) (Griffiths et al 2013). In this procedure 3β-hydroxy-Δ5 oxysterols and bile acids are converted to thecorresponding 3-oxo-Δ4 compounds which are then

Fig. 2 Histological features of the liver biopsy at 5 months of age. a)Lobular disarray with hepatocellular giant cell and rosette formations.There is advanced portal bridging fibrosis. Interlobular bile ducts are

preserved (x200, H&E). b) Increased paraseptal copper associated protein(x400, Shikata stain)

J Inherit Metab Dis

derivatised with Girard P hydrazine and analysed by LC-MSn.No saponification step was performed so sterols esterifiedwith fatty acids were not measured. Two samples wereanalysed from the patient described above; one while he wason UDCA therapy and prior to starting CDCA treatment andthe other at age 6.5years after 6 years of CDCA therapy.Stored plasma samples were also analysed from the patientsdescribed byUeki et al (2008) andMizuochi et al (2011) at thesame time using the same method. These samples had alsobeen taken when the patients were on UDCA therapy.

Sequencing of gene encoding oxysterol 7α-hydroxylase(CYP7B1)

Genomic DNAwas extracted from venous blood by a modi-fied ammonium acetate salting out method (Miller et al 1988;Davies et al 1993). The coding sequence and exon–intronboundaries of the six exons of CYP7B1 (Genbank accessionnumber NM_000008) were amplified by polymerase chainreaction (PCR) as described previously (Goizet et al 2009).Sequencing of CYP7A1 was not undertaken.

Restriction-enzyme digest testing

Restriction-enzyme digestion with Dpn1 (New EnglandBiolabs, Hitchin, UK) was used to confirm the observedsequence change. Dpn1 cuts the wild-type but not the mutatedsequence. Exon 6 of CYP7B1 was PCR-amplified for thepatient, his parents and his sibling. PCR products were incu-bated at 37 °C with 20 units of enzyme, 1 μl of 10x NEBufferand 100 μg/ml bovine serum albumin overnight in a finalvolume of 10 μl. The digestion products were separated byelectrophoresis on 2.5 % agarose gels containing ethidiumbromide alongside a 1 kb plus ladder (Invitrogen, Paisley,UK). Gels were viewed under UV light, and the sizes ofPCR products determined.

Results

Urine bile acids

The first urine sample was obtained from the patient at4 months of age, prior to starting UDCA. The negative ionmass spectrum revealed intense ions in the mass range corre-sponding to singly-charged bile acids and their conjugates(m/z 350–700) (Fig. 3). There were strong similarities to theprofile produced by negative ion fast atom bombardment massspectrometry in neonatal liver failure due to oxysterol 7α-hydroxylase deficiency (Setchell et al 1998) but also somedifferences. Our patient showed the profile shown in Fig. 3with (suggested identities in brackets): m/z 480 (taurine con-jugated 3β-hydroxy-5-cholenoic acid) >453 (sulphated 3β-

hydroxy-5-cholenoic acid) >462 (glycine conjugatedtrihydroxycholenoate[s]) >514 (taurine conjugatedtrihydroxycholanoates) >446 (glycine conjugateddihydroxycholenoates) >497 (sulphated cholestenetriol[s])>510 (sulphated and glycine conjugated 3β-hydroxy-5-cholenoate). In contrast the spectrum shown by Setchell et al(1998) showed m/z 453>510>480=479(?)>462>497. Thusalthough the same abnormal bile acids were present, in ourpatient these bile acids were less extensively sulphated (majorbile acid, m/z 480, not-sulphated) and in our patient somenormal bile acids were present (m/z 514).

A repeat urine bile acid analysis undertaken when ourpatient was 4 years old and taking CDCA (8 mg/kg/d),showed very low levels of bile acid excretion with the follow-ing profile: m/z 448 (glycodihydroxycholanoates) >>510(sulphated and glycine conjugated 3β-hydroxy-5-cholenoate)=567 (glucuronidated dihydroxycholanoates)>583 (glucuronidated trihydroxycholanoates).

Plasma bile acids analysis

When the first plasma sample was taken the patient wasreceiving UDCA, so the chromatogram was complicated bythe presence of UDCA metabolites. However, the trace wasvery clearly abnormal with larger than normal peaks due to(non-sulphated) 3β-hydroxy-5-cholestenoic acid, 3β-hydroxy-5-cholenoic acid, and 25- and 27-hydroxycholesterol. This first sample was later re-analysedby EADSA so the results for 3β-hydroxy-Δ5 bile acids couldbe compared. For 3β-hydroxy-5-cholestenoic acid, the resultwas 10.1 μM by EADSA and 1.12 μM by GC-MS; for 3β-hydroxy-5-cholenoic acid, the result was 0.95μMby EADSAand 0.22 μM by GC-MS. The EADSA method has beenchecked by addition of standard 3β-hydroxy-5-cholestenoicacid. These results indicate that the GC-MS method seriouslyunderestimates the plasma concentration of the 3β-hydroxy-5-cholestenoic acid and 3β-hydroxy-5-cholenoic acid. Thus,while the method can show that levels are above normal andfollow the response to treatment, the values are not accurate.

In the plasma samples taken when the patient was receivingUDCA, the normal bile acids were within or minimally abovethe normal range (i.e. low for a child with cholestatic jaundice)and there was a very substantial UDCA peak. Three resultsobtained during the few days the patient was on UDCAmonotherapy are shown in Table 1 together with two resultson plasma samples obtained at 9 months and 13 monthsrespectively after starting CDCA therapy. The samples takenwhile the patient was being treated with CDCA, and followingresolution of his liver disease, showed lower levels of 3β-hydroxy-5-cholestenoic acid than the samples taken duringUDCA therapy when the liver disease was uncontrolled. Thelevel of non-sulphated 3β-hydroxy-5-cholenoic acid was notsubstantially lower (despite the fact that the urine analysis

J Inherit Metab Dis

showed a dramatic fall to almost undetectable excretion of thesulphated forms of this bile acid). The plasma CDCA concen-tration on CDCA treatment was 1.5-5x the upper limit ofnormal; cholic acid remained in or below the normal range.

Plasma oxysterol analysis

The results of analysis of free oxysterols in plasma usingthe EADSA method are shown in Table 2 and in detail in

Supplementary materials. The plasma sample from ourpatient taken prior to starting CDCA showed massive ele-vation of 27-hydroxycholesterol, 25-hydroxycholesterol,24S-hydroxycholesterol but, in addition mild elevation, of7α-hydroxycholesterol, 7β-hydroxycholesterol and 7-oxocholesterol (Table 2). The analysis confirmed massiveelevation of 3β-hydroxy-5-cholenoic acid and 3β-hydroxy-5-cholestenoic acid as well as the presence of cholest-5-enetriol, and cholest-5-enediol sulphates (Supplementary

Fig. 3 Urinary cholanoid profile of the first UK case of oxysterol 7α-hydroxylase deficiency presenting with liver disease in infancy; sampletaken at 4 months of age, prior to any bile acid treatment. Major peaks are

m/z 480 (taurine-conjugated 3β-hydroxy-5-cholenoic acid) and m/z 453(sulphated 3β-hydroxy-5-cholenoic acid)

Table 1 Analysis of non-sulphated plasma bile acids by GC-MS. Three blood samples were taken during 5 days of ursodeoxycholic acid treatmentadministered shortly after presentation. A further two samples were taken at 9 months and 13 months after starting chenodeoxycholic acid therapy

Plasma concentration of non-sulphated compound (μmol/L)

Age 4 mo 4 mo 4 mo 13.5mo 17.5mo Normal range*

Treatment UDCA (1) UDCA (2) UDCA (3) CDCA for 9 mo CDCA for 13 mo

3β-OH-5-cholestenoic acid** 1.12 1.46 1.16 0.48 0.47 0.09-0.15

3β-OH-5-cholenoic acid** 0.22 0.15 0.09 0.10 0.11 0-0.05

Chenodeoxycholic acid 3.3 13 3.0 54 20 0.22-12.4

Cholic acid 1.03 0.47 1.06 0.32 n.d 0.05-4.6

Ursodeoxycholic acid 74.2 116 67.4 n.d 0.06 0-2.1

n.d. = not detected; CDCA = chenodeoxycholic acid; UDCA = ursodeoxycholic acid; mo = months

*Age-matched controls (n=50)

**The concentrations of these bile acids are very significantly underestimated by the GC-MS method

J Inherit Metab Dis

materials). The sample taken after 6 years of CDCA ther-apy when liver function was normal, showed substantialfalls in the concentration of all the oxysterols and 3β-hydrox-Δ5 bile acids,e.g. 27-hydroxycholesterol (4.34 to0.44 μM), 3β-hydroxy-5-cholestenoic acid (10.1 to0.13 μM ) and 3β-hydroxy-5-cholenoic acid (0.95 to0.066 μM). The samples from the two Asian patients withoxysterol 7α-hydroxylase deficiency taken while they wereon UDCA treatment showed a similar profile to our patientwith marked elevation of 27-, 25- and (24S)-hydroxycholesterols and of 3β-hydroxy-5-cholestenoic and5-cholenoic acids. 7α-Hydroxycholesterol was elevated inthe case reported by Ueki et al (2008) but not the casereported by Mizuochi et al (2011).

Sequencing of gene encoding oxysterol 7α-hydroxylase

CYP7B1 sequencing of the DNA sample from our patientdemonstrated a homozygous missense mutation in exon 6,c.1249C>T; p. Arg417Cys (Fig. 1, Supplementary materials).A heterozygous mutation was found in the DNA of both hisfather and his mother. Sequence alignment showed that thismutation affects an amino acid that is highly conserved acrossspecies, suggesting it is important for protein activity.

Restriction-enzyme digest test

Analysis of PCR products subjected to Dpn1 digestion andgel electrophoresis confirmed that our patient was homozy-gous for the c.1249C>T sequence change in CYP7B1. Hisparents and the younger sister (who had a normal urine bileacid profile) were heterozygous (Fig. 2, Supplementarymaterial).

Discussion

The diagnosis of oxysterol 7α-hydroxylase deficiency in thisboy is beyond doubt. He is homozygous for a missensemutation (p.Arg417Cys) in a highly conserved region of theCYP7B1 gene; this mutation has previously been found inpatients with SPG5 (Stiles et al 2009; Goizet et al 2009). Hehad cholestatic liver disease with a normal γ-GT, a biopsyshowing a giant cell hepatitis and severe fibrosis, and thecharacteristic metabolic profile of oxysterol 7α-hydroxylasedeficiency - increased plasma concentrations of 27- and 25-hydroxycholesterols and increased plasma concentration andurinary excretion of 3β-hydroxy-5-cholestenoic acid and 3β-hydroxy-5-cholenoic acid. It is important to recognize that theurine bile acid spectrum can vary from the profile described bySetchell et al in 1998 largely because of differences in conju-gation of the major metabolites. While differences in analyt-ical technique could affect urine cholanoid spectra andSetchell et al use fast atom bombardment mass spectrometry(FAB-MS) whereas we have changed to ESI-MS, in practice,spectra that we have obtained by ESI-MS and those obtainedby FAB-MS by ourselves or published by the Setchell groupare very similar for other disorders such as 3β-hydroxy-Δ5-C27-steroid dehydrogenase deficiency, Δ4-3-oxo-steroid 5β-reductase deficiency, CTX and amidation defects. The onetechnique that leads to substantially different spectra isnanospray ESI-MS in which many bile acid and oxysterolconjugates are seen as doubly-charged species (Alvelius et al2001; Griffiths et al 2013).

In addition to cholestatic liver disease, our patient hadepisodes of hypoglycaemia. While these may have been dueto severe liver disease, it is important to remember thatoxysterols and bile acids have effects on carbohydrate metab-olism, in particular, raised levels of LXR ligands such as 27-hydroxycholesterol can inhibit gluconeogenesis (Cao et al

Table 2 Analysis of oxysterols by LC-ESI-MSn following enzyme-assisted derivatisation for sterol analysis (EADSA). Data is for freeoxysterols as no saponification step was carried out. Columns 2 and 3show results for the patient described in this paper: Plasma samples wereobtained on ursodeoxycholic acid therapy (column 2) and after 6 years of

chenodeoxycholic acid therapy (column 3). Columns 4 and 5 showresults obtained from previously reported patients: Ueki et al 2008(column 4) and Mizuochi et al 2011 (column 5). ND = not detected. Fullresults shown in Supplementary materials

Oxysterol Our patient prior toCDCA (μmol/L)

Our patient after 6 yrsCDCA (μmol/L)

Case of Ueki et al2008 (μmol/L)

Case of Mizuochi et al2011 (μmol/L)

Normal range(μmol/L)*

27-Hydroxycholesterol* 4.34 0.44 2.65 2.87 0.025-0.050

25-Hydroxycholesterol 0.99 0.15 0.41 1.11 <0.015

24S-Hydroxycholesterol 0.43 0.17 0.13 0.46 0.015-0.032

7α-Hydroxycholesterol 0.025 0.009 0.24 0.0036 <0.005

7β-Hydroxycholesterol 0.052 0.0015 0.37 0.010 <0.005

7-Oxocholesterol 0.027 0.11 0.16 ND <0.012

3β-Hydroxy-5-cholestenoic acid 10.1 0.13 6.11 4.72 0.089-0.200

3β-Hydroxy-5-cholenoic acid 0.95 0.066 0.27 0.22 0.0027-0.0054

* Adult and paediatric controls

J Inherit Metab Dis

2003). Our patient also had significant hepatic steatosis; this isalso seen when mice are given an LXR agonist (Gao and Liu2013).

The main questions that this case poses are: i) Why didHSP patients, with the same mutations as our patient, notpresent with liver disease in infancy?; ii) Why did the liverdisease respond to treatment with CDCA when the casesreported by Setchell et al (1998) and Ueki et al (2008) failedto respond to UDCA and cholic acid? iii) Will treatment withCDCA prevent HSP?

Liver disease in oxysterol 7α-hydroxylase deficiency hasbeen attributed to the accumulation of toxic 3β-hydroxy-Δ5

bile acids (Setchell et al 1998). 3β-Hydroxy-5-cholenoic in-duces cholestatic liver disease (Mathis et al 1983). In all fourpatients with CYP7B1 mutations and liver disease, the majorurinary bile acid was 3β-hydroxy-5-cholenoic acid; whereFAB-MS or ESI-MS of urine bile acids was undertaken itwas shown to be present as the sulphate and taurine conjugate.The patient in this report and the Asian patients described byUeki et al (2008) and by Mizuochi et al (2011) had a plasma3β-hydroxy-5-cholestenoic acid concentration (measured bythe EADSA method) of 4–10 μM whereas the plasma con-centrations in adults with SPG5 were 0.5 – 1.7 μM (normal<0.2 μM) (Theofilopoulos et al, under review). In the patientwith neonatal onset liver disease described by Setchell et al(1998) the concentration was >4 μM. The very high levels of3β-hydroxy-Δ5 bile acids in the patients with liver diseasesuggests that the degree of accumulation of 3β-hydroxy-Δ5

bile acids may determine whether a patient has cholestaticliver disease or not. The tendency to accumulate these bileacids and to suffer hepatotoxic effects may be determined byage. In young infants, the acidic pathway is probably moreactive than the neutral pathway. This makes them more likelyto produce large amounts of 3β-hydroxy-Δ5 bile acids in thepresence of a block at the level of oxysterol 7α-hydroxylase.Young infants may also suffer more hepatotoxic effects be-cause the low activity of cholesterol 7α-hydroxylase meansthat they have limited ability to use the neutral pathway toproduce CDCA and cholic acid which could counteract thecholestatic action of the 3β-hydroxy-Δ5 bile acids and fuelbile flow. Our patient showed a remarkable reduction in theurinary excretion of 3β-hydroxy-5-cholenoic acid when bileacid spectra recorded at 4 months (prior to any treatment or onUDCA) and at the age of 4 years (on CDCA treatment) werecompared. The plasma concentration of 3β-hydroxy-5-cholestenoic acid was also reduced (Tables 1 and 2).However the CDCA treatment may not be the only factorresponsible for this change, the effects of age and resolution ofcholestasis could be important, although, in adults, cholestasisper se does not appear to lead to a significant rise in plasma3β-hydroxy-5-cholestenoic acid (Axelson et al 1989) and thisis also our experience of determining plasma bile acids inchildren with a wide range of cholestatic disorders.

If the hypothesis that high levels of 3β-hydroxy-Δ5 bileacids and low levels of CDCA and cholic acid cause chole-stasis and liver damage is correct, then effective treatment islikely to involve suppression of the production of the toxic3β-hydroxy-Δ5 bile acids and enhancing the supply of CDCAand cholic acid. There are two potential ways that the lattercould be achieved - giving one or both as medication orstimulating their production via the neutral pathway byboosting the activity of the rate-limiting step - cholesterol7α-hydroxylase. That the latter may be important is suggestedby the fact that our patient, who responded to treatment withCDCA had an elevated plasma concentration of 7α-hydroxycholesterol (indicating some activity of the neutralpathway prior to starting treatment) whereas the patient de-scribed by Setchell et al (1998) who failed to respond to cholicacid had undetectable plasma 7α-hydroxycholesterol. Thispatient was shown to have no mutations in the CYP7A1 genebut no cholesterol 7α-hydroxylase activity in the liver, how-ever, age-matched controls also had no activity with theenzyme assay that was used.

The factors determining cholesterol 7α-hydroxylase activ-ity in a patient with CYP7B1mutations are probably complex(Fig. 4). Transcription of CYP7A1 encoding cholesterol 7α-hydroxylase would normally increase during the first year oflife. Lack of CDCA and cholic acid in the liver should in-crease transcription of CYP7A1 because of reduced activationof the FXR/SHP/LRH pathway. Lack of these bile acids in theintestine should also increase transcription of CYP7A1 via theFXR/FGF 15/FGFR4/β Klotho pathway. However, these ef-fects may be lost in oxysterol 7α-hydroxylase deficiency asthe accumulating metabolites, 27-hydroxycholesterol, 25-hydroxycholesterol and 3β-hydroxy-5-cholestenoic acid, canalso act as FXR ligands (Nishimaki-Mogami et al 2004).Moreover, these three intermediates are also LXR ligands,and in man, unlike in rat, activation of LXRα in hepatocytesreduces CYP7A1 expression (Goodwin et al 2003). Finally,failure to synthesis bile acids by both the major pathways maylead to a build-up of unesterified cholesterol in the liver. Thiscan be converted non-enzymatically to 7-oxocholesterol byreactive oxygen species. 7-Oxocholesterol is a potent inhibitorof cholesterol 7α-hydroxylase (Song et al 1998). Treatmentwith CDCAwill need to be undertaken with care as too muchmight inhibit cholesterol 7α-hydroxylase to an undesirabledegree.

The presence of a normal or elevated plasma concentrationof 7α-hydroxycholesterol, indicating some activity of choles-terol 7α-hydroxylase may be an important determinant of thecapacity of the liver to respond to treatment. However, itseems likely that successful treatment must depend on thesuppression of synthesis of the toxic bile acids - 3β-hydroxy-5-cholestenoic acid and 3β-hydroxy-5-cholenoic ac-id. For this to occur, bile acid replacement therapymust reducethe activity of the first step in the acidic pathway - cholesterol

J Inherit Metab Dis

27-hydroxylase. The majority of studies suggest that bileacids, particularly hydrophobic bile acids such as CDCA canreduce cholesterol 27-hydroxylase activity (Twisk et al 1995;Vlahcevic et al 1996; Segev et al 2001), however, there areclear species differences and, in man, modest changes in bileacid levels such as those induced by administration of chole-styramine or bile acids do not have a very major effect oncholesterol 27-hydroxylase levels in the liver (Björkhem andEggertsen 2001). On the other hand, the changes in CDCAlevels in the liver that occur on giving CDCA to a patient withliver disease due to CYP7B1 mutations are likely to be muchmore dramatic than those that occur in normal adults givencholestyramine or CDCA.

In our patient, and the other three patients described in theliterature, UDCA had no beneficial effect on the course ofliver disease. UDCA can protect cholangiocytes and hepato-cytes against cytotoxicity of hydrophobic bile acids, stimulatehepatobiliary secretion and inhibit apoptosis induced by toxicbile acids (Paumgartner and Beuers 2004). It can also modu-late the immune response, by inhibiting inflammatory cyto-kines or by enhancing chemotaxis in inflammatory cells.

However, in the context of treatment of liver disease due toCYP7B1 mutations, it is probably the ability to suppress thesynthesis of hepatotoxic 3β-hydroxy-Δ5 bile acids that isimportant. UDCA does not reduce the synthesis of abnormalmetabolites (bile alcohol glucuronides) in cholesterol 27-hydroxylase deficiency (CTX) because it does not inhibitcholesterol 7α-hydroxylase (Koopman et al 1984).However, that could be an advantage in liver disease due toCYP7B1mutations. The important question is, “Can it reducethe activity of sterol 27-hydroxylase and hence synthesis ofhepatotoxic 3β-hydroxy-Δ5 bile acids?” There is no data inthe literature that addresses this question, however, Twisk et al(1995) suggested that it is the hydrophobic bile acids such asCDCA and deoxycholic acid that can reduce transcription ofsterol 27-hydroxylase in cultured rat hepatocytes; hydrophilicbile acids such as β-muricholic acid are ineffective. UDCA isa hydrophilic bile acid. Overall the data suggest that CDCAmay be more effective than either UDCA or cholic acid in thetreatment of liver disease due to oxysterol 7α-hydroxylasedeficiency. Unfortunately the knockout mouse is not a usefulmodel for the human disease. Disruption of Cyp7b1 in mice

27-Hydroxycholesterol

3β-Hydroxy-5-cholestenoic acid

LXR/SHP

Cholesterol CYP7A1 7α-OH-cholesterol

FXR/SHP/LRH

7-Oxocholesterol

FXR/FGF15/FGFR4/β Klotho

Chenodeoxycholic acid

Cholic acid

---

-

+

+

+

+

24S-and 25-Hydroxycholesterol

Fig. 4 The determinants of cholesterol 7α-hydroxylase (CYP7A1) ac-tivity. Suppression of bile acid synthesis by the neutral pathway ismediated by two mechanisms. The first involves binding of bile acidsto the farnesoid X receptor (FXR) in the liver, which then, with its shortheterodimeric partner (SHP), inhibits liver receptor homologue (LRH). Asecond mechanism involves binding of bile acids to FXR in theenterocyte, which leads to the production of fibroblast growth factor 15(FGF 15), which binds to a heterodimeric receptor composed of FGFR4

and βKlotho in the liver, and suppresses synthesis of cholesterol 7α-hydroxylase. Intermediates that accumulate in oxysterol 7α-hydroxylasedeficiency such as 25- and 27-hydroxycholesterol and 3β-hydroxy-5-cholestenoic acid are ligands for both FXR and the liver X receptor(LXR). In man (unlike in rat) activation of LXRα represses CYP7A1activity. In addition cholesterol can be converted non-enzymatically byreactive oxygen species to 7-oxocholesterol which is an inhibitor ofcholesterol 7α-hydroxylase

J Inherit Metab Dis

does not show the severe phenotype as seen in Cyp7a1 −/−

mice (Li-Hawkins et al 2000). As expected, 25- and 27-hydroxycholesterol are markedly elevated, but bile acid me-tabolism and serum cholesterol and triglyceride levels arenormal and the mice do not develop severe liver disease(Setchell et al 1998).

We do not know whether our patient will develop spasticparaparesis ± dorsal column signs. He started to trip a lot whenhe ran from 5 years old. He was found to have pes cavus butno upper motor neuron signs in the legs and improved whenhe was given insoles. Pes cavus is one of the signs occasion-ally observed in SPG5 patients by Goizet et al (2009). In theirpatients with CYP7B1 mutations, age of onset of spasticparaparesis varied from 5y to 47y. In two siblings with thesame genotype as our patient, the age of onset was at 10y and12y respectively. Both had pes cavus when examined at >50ybut they also had severe lower limb spasticity with a positiveBabinski sign (Goizet et al 2009).

The mutations in CY7B1 causing SPG5 include nonsensemutations and missense mutations (Tsaousidou et al 2008;Goizet et al 2009). Sequencing of genomic DNA showed thatour patient was homozygous for a missense mutation,c.1249C>T, p.R417C. R417 resides in the meander region(the conserved PERF region) just before the cysteine pocket.The meander region has been implicated in heme binding andin interactions with redox partners for P450 reduction (Siamet al 2012; Lewis 2004; Pikuleva and Waterman 2013; Cuiet al 2013). This residue may have a catalytic and/or structuralfunction; as such R417Cwould be predicted to severely affectenzyme function. Sequence alignment of CYP7B1orthologues (Fig. 1, Supplementary material) shows that thissite is in a highly conserved region, suggesting that the muta-tion disrupts protein function.

Acknowledgments The authors thank the family members for theirparticipation. PTC and PBM were funded by Great Ormond StreetChildren’s Charity. PG is funded by aWellcome Trust Senior Fellowship.Work in Swansea was supported by BBSRC. Work in Bordeaux wassupported by the Agence Nationale de la Recherche (ANR) (Project2010BLAN112601/LIGENAX), the Association Française contre lesMyopathies (AFM) (14879/MNM2 2012), the Conseil Régionald’Aquitaine (CRA) (2011-0151/LIGENAX), The AssociationStrumpell-Lorrain (ASL) (2011–0135), and the Pôle de compétitivitéProd’Innov.

Competing interest Dongling Dai, Philippa Mills, Emma Footitt, PaulGissen, Patricia McClean, Jens Stahlschmidt, Isabelle Coupry, JulieLavie, Fanny Mochel, Cyril Goizet, Tatsuki Mizuochi, AkihikoKimura, Hiroshi Nittono, Karin Schwarz, Peter Crick, Yuqin Wang,William Griffiths and Peter Clayton declare that they are not paid per-sonally, or for research, by any company involved in the manufacture,distribution or marketing of CDCA for medicinal use, nor do they holdshares in any such company. Cyril Goizet makes the following declara-tion: The work under consideration was financially supported by AgenceNationale de la Recherche (ANR) (Project 2010BLAN112601/LIGENAX), the Association Française contre les Myopathies (AFM)(14879/MNM2 2012), the Conseil Régional d’Aquitaine (CRA) (2011-

0151/LIGENAX), The Association Strumpell-Lorrain (ASL) (2011–0135), and the Pôle de compétitivité Prod’Innov. This support wasaddressed to Pr Cyril Goizet. Pr Cyril Goizet received consulting feesfrom Raptor, Genzyme, Actelion and Shire. Pr Cyril Goizet receivedfinancial support for research activities from TKT5S, Shire, Genzyme,Association Française contre les myopathies (AFM), Association contreles maladies mitochondriales (AMMi), Association Strumpell-Lorrain(ASL), Connaitre les Syndromes Cérébelleux (CSC), Conseil régionald’Aquitaine (CRA), Agence Nationale pour la Recherche (ANR),Programme Hospitalier de Recherche Clinique (PHRC), and Registryof the European Huntington Disease Network (EHDN). Inscriptionsand travels for congresses of Dr Cyril Goizet were funded by Shire,Genzyme, Actelion, Takeda. William Griiffiths, Peter Crick and YuqunWang declare that the development of the analytical methods used inSwansea for the analysis of cholesterol metabolites was funded by a grantfrom the UIK Research Council, BBSRC (grant no BB/I001735/1). PeterClayton receives a salary from Great Ormond Street Hospital Children’sCharity and declares the following relevant financial activities outside thesubmitted work: Grant from Actelion for investigator led project ondiagnosis and monitoring of Niemann-Pick C, fees for teaching oncourses from Orphan Europe/Recordati Foundation for Rare Diseases,fees for lectures / consultancy fromMerck Corp USA, Actelion, shares inWaters, Abbott, Abbvie.

References

Alvelius G, Hjalmarson O, Griffiths WJ et al (2001) Identification ofunusual 7-oxygenated bile acid sulfates in a patient with Niemann-Pick disease, type C. J Lipid Res 42:1571–1577

Andersson S, Gustafsson N, Warner M, Gustafsson JA (2005)Inactivation of liver X receptor beta leads to adult-onset motorneuron degeneration in male mice. Proc Natl Acad Sci U S A 102:3857–3862

Axelson M, Mörk B, Aly A et al (1989) Concentrations of cholestenoicacids in plasma from patients with liver disease. J Lipid Res 30:1877–1882

Björkhem I, Eggertsen G (2001) Genes involved in initial steps of bileacid synthesis. Curr Opin Lipidol 12:97–103

Cao G, Liang Y, Broderick CL et al (2003) Antidiabetic action of a liver xreceptor agonist mediated by inhibition of hepatic gluconeogenesis.J Biol Chem 278:1131–1136

Clayton PT (1983) The validation and application of an assay for themeasurement of bile acids in the plasma of infants and children. MDThesis. University of Cambridge

Clayton PT (2011) Disorders of bile acid synthesis. J Inherit Metab Dis34:593–604

Clayton PT, Muller DP (1980) A simplified gas–liquid chromatographicmethod for the estimation of non-sulphated plasma bile acids. ClinChim Acta 105:401–405

Clayton PT, Lake BD, Hall NA et al (1987) Plasma bile acids inpatients with peroxisomal dysfunction syndromes; analysis bycapillary gas chromatography – mass spectrometry. Eur J Pediatr146:166–173

Clayton PT, Mills KA, Johnson AW et al (1996) Delta 4-3-oxosteroid 5beta-reductase deficiency: failure of ursodeoxycholic acid treatmentand response to chenodeoxycholic acid plus cholic acid. Gut 38:623–628

Cui YL, Zhang JL, Zheng QC et al (2013) Structural and dynam-ic basis of human cytochrome P450 7B1: a survey of sub-strate selectivity and major active site access channels. Chemistry19:549–557

Davies JP, Winchester BG, Malcolm S (1993) Sequence variations in thefirst exon of alpha-galactosidase A. J Med Genet 30:658–663

J Inherit Metab Dis

Fink J (2006) Hereditary spastic paraplegia. Curr Neurol Neurosci Rep 6:65–76

Gao M, Liu D (2013) The liver X receptor agonist T0901317 protectsmice from high fat diet-induced obesity and insulin resistance.AAPS J 15:258–266

Goizet C, Boukhris A, Durr A et al (2009) CYP7B1mutations in pure andcomplex forms of hereditary spastic paraplegia type 5. Brain 132:1589–1600

Goodwin B,WatsonMA, KimH et al (2003) Differential regulation of ratand human CYP7A1 by the nuclear oxysterol receptor Liver Xreceptor-α. Mol Endocrinol 17:386–394

Griffiths WJ, Crick PJ, Wang Y et al (2013) Analytical strategies forcharacterization of oxysterol lipidomes: liver X receptor ligands inplasma. Free Radic Biol Med 59:69–84

Koopman BJ, Wolthers BG, van der Molen JC et al (1984) Capillary gaschromatographic determinations of urinary bile acids and bile alco-hols in CTX patients proving the ineffectivity of ursodeoxycholicacid treatment. Clin Chim Acta 142:103–111

Lewis DF (2004) 57 varieties: the human cytochromes P450.Pharmacogenomics 5:305–318

Li-Hawkins J, LundEG, Turley SD, Russell DW (2000)Disruption of theoxysterol 7alpha-hydroxylase gene in mice. J Biol Chem 275:16536–16542

Mathis U, Karlaganis G, Preisig R (1983)Monohydroxy bile salt sulfates:tauro-3 beta-hydroxy-5-cholenoate-3-sulfate induces intrahepaticcholestasis in rats. Gastroenterology 85:674–681

Meaney S, Heverin M, Panzenboeck U et al (2007) Novel route for theelimination of brain oxysterols across the blood–brain barrier: con-version into 7α-hydroxy-3-oxo-4-cholestenoic acid. J Lipid Res 48:944–951

Miller SA, Dykes DD, Polesky HF (1988) A simple salting out procedurefor extracting DNA from human nucleated cells. Nucleic Acids Res16:1215

Mills KA, Mushtaq I, Johnson AW et al (1998) A method for thequantitation of conjugated bile acids in dried blood spots usingelectrospray ionization-mass spectrometry. Pediatr Res 43:361–368

Mizuochi T, Kimura A, Suzuki M et al (2011) Successful heterozygousliving donor liver transplantation for an oxysterol 7α-hydroxylasedeficiency in a Japanese patient. Liver Transpl 17:1059–1065

Nishimaki-Mogami T, Une M, Fujino T et al (2004) Identification ofintermediates in the bile acid synthetic pathway as ligands for thefarnesoid X receptor. J Lipid Res 45:1538–1545

Paumgartner G, Beuers U (2004) Mechanisms of action and therapeuticefficacy of ursodeoxycholic acid in cholestatic liver disease. ClinLiver Dis 8:67–81

Pikuleva IA, WatermanMR (2013) Cytochromes p450: roles in diseases.J Biol Chem 288:17091–17098

Segev H, Honigman A, Rosen H, Leitersdorf E (2001) Transcriptionalregulation of the human sterol 27-hydroxylase gene (CYP27) andpromoter mapping. Atherosclerosis 156:339–347

Setchell KD, Schwarz M, O'Connell NC et al (1998) Identification of anew inborn error in bile acid synthesis: mutation of the oxysterol7alpha-hydroxylase gene causes severe neonatal liver disease. J ClinInvest 102:1690–1703

Siam A, Brancale A, Simons C (2012) Comparative modeling of 25-hydroxycholesterol-7α-hydroxylase (CYP7B1): ligand binding andanalysis of hereditary spastic paraplegia type 5 CYP7B1 mutations.J Mol Model 18:441–453

Song W, Chen J, Dean WL et al (1998) Purification and characterizationof hamster liver microsomal 7alpha-hydroxycholesterol dehydroge-nase. Similarity to type I 11beta-hydroxysteroid dehydrogenase. JBiol Chem 273:16223–16228

Stiles A, McDonald J, Bauman D, Russell D (2009) CYP7B1: onecytochrome P450, two human genetic diseases, and multiple phys-iological functions. J Biol Chem 284:28485–28489

Tsaousidou M, Ouahchi K, Warner TT et al (2008) Sequence alterationswithin CYP7B1 implicate defective cholesterol homeostasis inmotor-neuron degeneration. Am J Hum Genet 82:510–515

Twisk J, de Wit EC, Princen HM (1995) Suppression of sterol 27-hydroxylase mRNA and transcriptional activity by bile acids incultured rat hepatocytes. Biochem J 305:505–511

Ueki I, Kimura A, Nishiyori A et al (2008) Neonatal cholestatic liverdisease in an Asian patient with a homozygous mutation in theoxysterol 7alpha-hydroxylase gene. J Pediatr Gastroenterol Nutr46:465–469

Vlahcevic ZR, Jairath SK, Heuman DM et al (1996) Transcriptionalregulation of hepatic sterol 27-hydroxylase by bile acids. Am JPhysiol 270:G646–G652

J Inherit Metab Dis