Journal of Inherited Metabolic Disease

, Volume 35, Issue 3, pp 521–530

Bile acid-CoA ligase deficiency—a new inborn error of bile acid metabolism

Authors

  • Catherine P. K. Chong
    • Clinical & Molecular Genetics UnitUCL Institute of Child Health
  • Philippa B. Mills
    • Clinical & Molecular Genetics UnitUCL Institute of Child Health
  • Patricia McClean
    • Leeds Teaching Hospitals NHS Trust
  • Paul Gissen
    • Clinical & Molecular Genetics UnitUCL Institute of Child Health
  • Christopher Bruce
    • School of Clinical and Experimental MedicineUniversity of Birmingham
  • Jens Stahlschmidt
    • Leeds Teaching Hospitals NHS Trust
    • Paediatric HistopathologySt James’s University Hospital Bexley Wing
  • A. S. Knisely
    • Institute of Liver Studies / HistopathologyKing’s College Hospital
    • Clinical & Molecular Genetics UnitUCL Institute of Child Health
    • Biochemistry Research Group, Clinical & Molecular Genetics UnitUCL Institute of Child Health
Original Article

DOI: 10.1007/s10545-011-9416-3

Cite this article as:
Chong, C.P.K., Mills, P.B., McClean, P. et al. J Inherit Metab Dis (2012) 35: 521. doi:10.1007/s10545-011-9416-3

Abstract

Born at 27 weeks gestation, a child of consanguineous parents of Pakistani origin required prolonged parenteral nutrition. She developed jaundice, with extensive fibrosis and architectural distortion at liver biopsy; jaundice resolved with supportive care. Serum γ-glutamyl transpeptidase values were within normal ranges. The bile acids in her plasma and urine were >85% unconjugated (non-amidated). Two genes encoding bile-acid amidation enzymes were sequenced. No mutations were found in BAAT, encoding bile acid-CoA : aminoacid N-acyl transferase. The patient was homozygous for the missense mutation c.1012C > T in SLC27A5, predicted to alter a highly conserved amino-acid residue (p.H338Y) in bile acid-CoA ligase (BACL). She also was homozygous for the missense mutation c.1772A > G in ABCB11, predicted to alter a highly conserved amino-acid residue (p.N591S) in bile salt export pump (BSEP). BACL is essential for reconjugation of bile acids deconjugated by gut bacteria, and BSEP is essential for hepatocyte-canaliculus export of conjugated bile acids. A female sibling born at term had the same bile-acid phenotype and SLC27A5 genotype, without clinical liver disease. She was heterozygous for the c.1772A > G ABCB11 mutation. This is the first report of a mutation in SLC27A5. The amidation defect may have contributed to cholestatic liver disease in the setting of prematurity, parenteral nutrition, and homozygosity for an ABCB11 mutation.

Introduction

Infants with inborn errors of bile acid synthesis often present with jaundice and malabsorption of fat and fat-soluble vitamins that may progress to cirrhosis and liver failure, with conjugated hyperbilirubinaemia and normal-range serum γ-glutamyl transpeptidase (GGT) values (Clayton 2006a, b). Several such errors have been identified (Bove et al. 2004; Clayton 2006a, b; Heubi et al. 2007), usually following detection of unusual bile acids or bile alcohols in blood or urine.

Infants with bile-acid amidation defects cannot convert unconjugated bile acids to conjugated bile acids. Two enzymes participate in the amidation of 24-carbon (C24) bile acids—bile acid-CoA ligase (BACL) and bile acid-CoA : aminoacid N-acyltransferase (BAAT) (Solaas et al. 2000; Mihalik et al. 2002; Doege et al. 2006; Hubbard et al. 2006). BACL, which converts chenodeoxycholic acid (CDCA) and cholic acid (CA) to their CoA esters chenodeoxycholoyl-CoA and choloyl-CoA, is encoded by SLC27A5. BAAT, which converts chenodeoxycholoyl-CoA and choloyl-CoA to the glycine and taurine conjugates of CDCA and CA, is encoded by BAAT.

Conjugation defects theoretically should manifest not with jaundice, but with steatorrhoea, effects of malabsorption of fat-soluble vitamins, and perhaps bile-acid diarrhoea (Hofmann and Strandvik 1988). This was confirmed in part in 1995 in a 14-year-old male and subsequently in a sibling boy and girl born to consanguineous Saudi Arabian parents (Setchell and O’Connell 2004); the girl was asymptomatic at diagnosis, but both boys had been jaundiced as infants. No genetic basis was identified for these patients’ disorder(s). Decreased absorption of fats and fat-soluble vitamins, found in these patients, was ascribed to detergent inefficiency of not only unconjugated bile acids but also bile-acid sulphates and glucuronides (which these patients could synthesise). Lipids not dispersed into micelles were likely absorbed poorly (Setchell and O’Connell 2004; Heubi et al. 2007).

Amish and Mennonite children in Pennsylvania with failure to thrive, chronic upper respiratory-tract infections, pruritus, fat malabsorption, and hypocoagulability (in one instance fatal) were found to have unconjugated hypercholanaemia associated with homozygosity for a c226A > G (p.M76V) mutation in BAAT (Carlton et al. 2003). Jaundice was seen in only one infant; it resolved spontaneously and has not recurred with menarche (personal observations, ASK). In our laboratory, urine and plasma bile-acid analyses undertaken on a girl born to consanguineous parents of Pakistani origin, who presented in infancy with jaundice, failure to thrive and rickets, indicated an amidation defect; homozygosity for a c.415C > T (p.R139X) mutation in BAAT was found (Dr L Bull, University of California San Francisco). She was treated with ursodeoxycholic acid (UDCA) but gained weight poorly and required fat-soluble vitamin supplements. At age 6 years she was asymptomatic with normal liver function tests.

Six patients with BAAT mutations manifest as growth delay (n = 3), neonatal jaundice (3) and fat-soluble vitamin deficiency (5) have been described in a preliminary communication (c.68C > T / p.R20X, n = 4; c.206A > T / p.D69V and c.250 > A / p.P84T, each n = 1; all in homozygous state). The only one with elevated serum transaminase activities was prematurely born and received parenteral nutrition, with disease that eventually required liver transplantation (personal communication, L Bull). In both other jaundiced patients, icterus resolved spontaneously (Heubi et al. 2009).

In the present patient and her sister, no deviation from BAAT consensus sequence was found. Instead both were homozygous for a missense mutation affecting a highly conserved residue in SLC27A5 (encoding BACL).

Patient

AK, born prematurely at 27 weeks’ gestation, is the second child of first cousins of Pakistani origin. A sibling was stillborn at 21 weeks’ gestation; a younger sister was born at term. All three pregnancies were uncomplicated. AK was ventilated for only 3 days but had two episodes of possible necrotising enterocolitis, treated with antibiotics and 35 days of parenteral nutrition, and developed jaundice with conjugated hyperbilirubinaemia, elevated serum transaminase values, and normal-range serum GGT values. Plasma concentrations of vitamins A and E were slightly below normal ranges. Evaluation at age 13 weeks excluded usual structural, infective and metabolic causes of infantile cholestasis. UDCA and fat-soluble vitamins were given, with resolution both of icterus and of abnormalities in results of usual clinical biochemistry tests. Microscopy of a liver-biopsy specimen obtained at age 31 weeks found architectural distortion, with inconspicuous bile ducts, portal-portal bridging fibrosis, parenchymal nodularity, and slight hepatocellular cholestasis (Fig. 1). At age 34 weeks UDCA was withdrawn. However, the severe fibrosis seen in the liver biopsy prompted further investigations, including tests for an inborn error of bile acid synthesis. When bile-acid analysis was undertaken, at age 8 months, AK was not jaundiced and serum transaminase values were in normal ranges. Aged 5 years, her only medication is a multivitamin preparation. She remains well, without clinical or usual clinical laboratory evidence of liver disease. Her weight and height track the 25th and 50th centiles respectively, and her plasma fat-soluble vitamin values are normal.
https://static-content.springer.com/image/art%3A10.1007%2Fs10545-011-9416-3/MediaObjects/10545_2011_9416_Fig1_HTML.jpg
Fig. 1

Low-power images of liver core biopsy (H&E and elastica stain, both ×100 magnification) showing moderate portal inflammation with bridging fibrosis and parenchymal nodularity. Note the inconspicuous cholestasis at this magnification

Methods

Urine bile acids

Cholanoids (bile acids and bile alcohols) in urine were analysed by negative ion electrospray ionization mass spectrometry (ESI-MS; Quattro Micro, Micromass, Waters, UK) as described (Mills et al. 1998). A mass-to-charge ratio (m/z) range of 350–700 was used for both the MS direct scan and the parent ion scans. Details of ESI-MS/MS conditions are shown in Table 1. Amidation defects were diagnosed when concentrations of unconjugated bile acids were increased and concentrations of conjugated bile acids were markedly reduced.
Table 1

MS/MS settings for the analysis of cholanoids by LC-ESI-MS/MS

Scanning modes

Direct scan

Parents of m/z 74 scan

Parents of m/z 80 scan

Parents of m/z 97 scan

Parents of m/z 85 scan

Capillary voltage (kV)

3.70

3.70

3.70

3.70

3.30

Cone voltage (V)

65.0

65.0

85.0

135.0

49.0

Collision energy (eV)

Not applicable

40

62

44

50

Parent of m/z 74 Scan for glycine conjugates, parents of m/z 80 scan for taurine conjugates, parents of m/z 97 scan for sulphate conjugates, parents of m/z 85 scan for glucuronide conjugates

Plasma bile acids

Plasma bile acid analysis by gas chromatography-mass spectrometry (GC-MS; Agilent UK) was undertaken as described (Clayton and Muller 1980; Clayton et al. 1987a, 1995), with and without deconjugation of amidated bile acids using cholyglycine hydrolase, and total and unconjugated bile-acid concentrations were thus determined. Control ranges for plasma bile-acid concentrations in normal infants have been documented (Clayton 1983; Clayton et al. 1987b).

Sequencing of genes encoding enzymes involved in bile-acid amidation

Genomic DNA was extracted from venous blood by a modified version of the ammonium acetate salting out method (Miller et al. 1988; Davies et al. 1993).

As mutations in BAAT have been described in patients with cholestasis (Carlton et al. 2003; Heubi et al. 2009), BAAT was screened initially. Polymerase chain reaction (PCR) intronic primers (Table 2) were designed; the three exons of BAAT, with intron/exon boundaries, were amplified by PCR. A typical PCR reaction using 50 ng of genomic DNA contained 25 pmol of each primer, 1 × NH4 reaction buffer (Bioline, London, UK), 0.2 mmol/l deoxynucleotide triphosphates, and 0.5 μl (2.5 units) BioPro DNA polymerase (Bioline; added after a ‘hot start’). Annealing temperatures and MgCl2 concentrations used are provided in Tables 2 and 3. Cycling conditions were typically 96°C for 10 min, followed by 35 cycles of 30 s at 96°C, 30 s at 52–55°C, 30 s at 72°C, and a final extension at 72°C for 10 min. PCR products were directly sequenced using the BigDye Terminator v. 3.1 Cycle Sequencing Kit (Applied Biosystems, Warrington, UK) and the MegaBACE capillary DNA sequencer (Amersham Biosciences UK, Chalfont St. Giles, UK)
Table 2

Primers and PCR conditions used for the amplification of the BAAT gene

 

Primers

Product size (bp)

[MgCl2] (mmol/l)

Annealing temperature (°C)

Exon 1

S: 5′-AGTGGTCTGACTGCTGATTCAA-3′

801

1.0

52

A/S: 5′-GCAGGAAGCAGTGCTCTAGG-3′

Exon 2

S: 5′-ACCTTAATTCTGCCCCAGGT-3′

585

1.5

55

A/S: 5′- TCTGGCCAAGTGTCAGATCA-3′

Exon 3

S: 5′-TGCTTTGTTCTCATTAGCCAGT-3′

839

1.5

55

A/S: 5′-TCAGGTTTTTACCCTATGCTCTTT-3′

S Sense primer, A/S antisense primer

Table 3

Primers and PCR conditions used for the amplification of the human SLC27A5 gene

 

Primers

Product size (bp)

[MgCl2] (mmol/l)

Annealing temperature (°C)

Exon 1

S: 5′-GCCAGTGATGGAAAGGGTTA-3′

1,040

1.0

60

A/S: 5′-GCCTGTGAAAGTAGGCAGGT-3′

Exon 2

S: 5′-GAGCCAGTGGAGAATGGAAG-3′

969

1.0

60

A/S: 5′-ATCAAGTCTGGGAGCAGTGG-3′

Exon 3

S: 5′-ATTCAGCCTGTGAACCCAAC-3′

452

1.5

60

A/S: 5′-GTGGGTGTCCCTTCTCAAAA-3′

Exon 4

S: 5′-GGCCAAGGACACAAAGATGA-3′

541

1.0

60

A/S: 5′-GAGGCTCAAGGTAGCACAGC-3′

Exons 5 & 6

S: 5′-CCAAGGTAAGGACCCAGGAT-3′

746

1.0

60

A/S: 5′-CCGTCTTCCACAGAAAGCTC-3′

Exon 7

S: 5′-GGTGCTGGGTGATGAAGAGT-3′

409

1.0

60

A/S: 5′-AAGTTCCGCCCTCTTACGAC-3′

Exon 8

S: 5′-GACTGGGCCTCCTGATCC-3′

321

1.0

60

A/S: 5′-GCTGTGAGAGCCAGCAAGTC-3′

Exons 9 & 10

S: 5′-GGGCCCAGAGAGGTTTAGAG3′

1,035

1.0

60

A/S: 5′-CTAGCCTCAGTCCAGCAACC-3′

S Sense primer, A/S antisense primer

The 10 exons, with intron/exon boundaries, of SLC27A5 were sequenced secondarily using similar conditions (intronic-primer sequences, Table 3, with annealing temperatures and MgCl2 concentrations as shown). The observed variation from consensus SLC27A5 gene and transcript sequences (ENSG00000083807 and ENST00000263093 respectively; http://www.ensembl.org) was numbered with +1 as the A of the ATG initiation codon.

Restriction-enzyme digest testing

Restriction-enzyme digestion was used to confirm observed sequence changes. All restriction enzymes and buffers were from New England Biolabs (Hitchin, UK). For the SLC27A5 sequence change detected in AK, the restriction enzyme Pm1I was chosen as cutting wild-type but not mutated sequence. Exon 3 of SLC27A5 was PCR-amplified in the patient, her parents and 258 anonymised, ethnically matched control chromosomes. PCR products were incubated at 37°C with 20 units of enzyme, 1 μl of 10x NEBuffer and 100 μg/ml bovine serum albumin overnight in a final volume of 10 μl. The digestion products were separated by electrophoresis on 2.5% agarose gels that contained ethidium bromide alongside a 1 kb plus ladder (Invitrogen, Paisley, UK).

Gels were viewed under UV light, and the sizes of PCR products were determined. To confirm that the change detected by sequencing was not a common polymorphism, it was sought in an anonymised control cohort of 129 individuals of Pakistani origin living in the UK.

Resequencing-chip “cholestasis gene” screening

As cholestatic liver disease in AK was more severe than expected for amidation deficiency, we used a BRUM1 resequencing chip (Bruce et al. 2010) to see if mutations were present in other genes implicated in neonatal cholestatic liver disease, looking particularly at those characterised by normal serum GGT activity—ABCB11 and ATP8B1—but also including NPC1 and NPC2. Deviations from consensus sequence, when suggested, were confirmed as described (Bruce et al. 2010).

Immunostaining of liver tissue

Sections of formalin-fixed, paraffin-embedded liver from AK and from persons without known bile-acid amidation or secretion defects were immunostained using antibodies to BAAT [rabbit polyclonal anti-BAAT antibody (ab97455; Abcam, Cambridge, UK)], BACL [rabbit affinity-purified anti-SLC27A5 (−BACL) antibody (HPA007292; Sigma, St Louis, MO)], and bile salt export pump [BSEP; rabbit affinity-purified anti-BSEP, encoded by ABCB11 (HPA019035; Sigma-Aldrich, Gillingham, UK)], with EnVision reaction development (DAKO UK, Ely, Cambs) and haematoxylin counterstaining (Evason et al. 2011). Reaction patterns were assessed by light microscopy.

Investigation of sister

In 2008 SK was born at term, a full sibling to AK. Mild neonatal jaundice lasted only 24 h. SK fed and gained weight satisfactorily: at age 6 months, length was between the 9th and 25th centiles and weight between the 2nd and 9th centiles. She was not jaundiced and did not have excoriated skin or exhibit behaviour suggesting pruritus. Liver size was normal. Steatorrhoea was not identified. Evidence of hepatobiliary injury or fat-soluble vitamin malabsorption was not found on clinical biochemistry testing. Urine bile acids were assayed by ESI-MS. Peripheral-leucocyte DNA was subjected to BRUM1-chip screening and evaluated for the SLC27A5 mutation found in AK.

Results

Urine bile acids

Analysis of spectrograms of urine from AK subjected to ESI-MS/MS (Fig. 2c) found a predominant peak with m/z 407, corresponding to unconjugated CA. Unconjugated CDCA, m/z 391, was also present, as were sulphate and glucuronide conjugates of dihydroxy- and trihydroxy-cholanoic acids (m/z 471, 487, 567, 583). However, peaks attributable to the glycine and taurine conjugates of CDCA and CA (m/z 448, 464, 498, 514) were lacking, in contrast to urine samples from other infants with cholestasis (Fig. 2b).
https://static-content.springer.com/image/art%3A10.1007%2Fs10545-011-9416-3/MediaObjects/10545_2011_9416_Fig2_HTML.gif
Fig. 2

Urinary cholanoid profiles from AK and normal and cholestatic controls

Plasma bile acids

A plasma sample from patient AK subjected to GC-MS proved to contain a high proportion of unconjugated bile acids. Table 4 shows the concentrations of the unconjugated bile acids CDCA, CA and UDCA in plasma as obtained (non-amidated bile acids only) and in plasma treated with the deconjugating enzyme cholylglycine hydrolase (both originally non-amidated and newly deconjugated bile acids; total bile acids). Originally non-amidated bile acids constituted >85% (normal <25%) of bile acids in the sample from AK. The total (amidated plus non-amidated) plasma CDCA concentration was mildly elevated at 20.9 μM (normal 0.22–12.4); plasma total CA was normal at 3.25 μM (0.05–4.55).
Table 4

Results of analysis of plasma bile acids from patient AK by GC-MS

Bile acid

Total bile acid concentrationa (μM) [normal range]

Unconjugated bile acid concentrationb (μM)

Unconjugated (%)

Unconjugated in controls (%)

Chenodeoxycholic acid

20.9 [0.22–12.4]

18.4

88%

<25%

Cholic acid

3.25 [0.05–4.55]

2.95

91%

<25%

Ursodeoxycholic acid

4.07 [0–2.09]

3.75

92%

<25%

aConcentration of the bile acid after enzymatic deconjugation

bConcentration of the free bile acid (measured without enzymatic deconjugation)

Sequencing of genes encoding enzymes involved in bile-acid amidation

No mutation in BAAT was found in AK. She was, however, homozygous for the substitution mutation c.1012C > T (p.H338Y) in SLC27A5 (Fig. 3). Sequence alignment (Fig. 4) showed that this mutation is in a gene region highly conserved across species, suggesting importance for protein activity.
https://static-content.springer.com/image/art%3A10.1007%2Fs10545-011-9416-3/MediaObjects/10545_2011_9416_Fig3_HTML.gif
Fig. 3

Electropherograms showing the homozygous mutation (c.1012C > T; H338Y) in the SLC27A5 gene of patient AK

https://static-content.springer.com/image/art%3A10.1007%2Fs10545-011-9416-3/MediaObjects/10545_2011_9416_Fig4_HTML.gif
Fig. 4

Sequencing alignment showing that H338Y is in a highly conserved area of the SLC27A5 gene

Restriction-enzyme digest testing

Analysis of PCR products subjected to Pm1I digestion and gel electrophoresis confirmed that the c.1012C > T sequence change in SLC27A5 was not present in 129 ethnically matched anonymous DNA samples (258 control chromosomes), suggesting that the change in AK is not a polymorphism. Each parent of patient AK proved heterozygous for the c.1012C > T mutation (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs10545-011-9416-3/MediaObjects/10545_2011_9416_Fig5_HTML.gif
Fig. 5

Restriction enzyme digest test for the H338Y mutation in the SLC27A5 gene. H338Y confirmed by digestion of the PCR product for exon 3 with enzyme Pm1I. Digestion of wild-type results in fragments of 276 and 176 bp. H338Y abolishes the restriction enzyme recognition site. Parents are heterozygous. AK is homozygous

Resequencing-chip “cholestasis gene” screening

No mutation in ATP8B1, NPC1 or NPC2 was found in AK. She was, however, homozygous for the sequence change c.1772A > G (p.N591S) in ABCB11. This has been described in the heterozygous state in one patient with intrahepatic cholestasis of pregnancy and is assessed as probably pathogenic (Pauli-Magnus et al. 2004).

Immunostaining of liver tissue

No abnormality in expression of BACL was found in liver from AK (Fig. 6). Expression of BAAT and BSEP were also unremarkable (not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs10545-011-9416-3/MediaObjects/10545_2011_9416_Fig6_HTML.jpg
Fig. 6

Immunostaining for bile acid-CoA ligase in the liver biopsy from patient AK and a control paediatric liver biopsy. The enzyme/bile acid transporter is expressed in the cytoplasm of hepatocytes. There is no reduction in the amount of immunoreactive protein in the patient

Investigation of sister

Bile-acid profiles in urine from the sisters were identical, consistent with both harbouring the amidation defect. The sister proved homozygous for the c.1012 > T (p.H338Y) mutation in SLC27A5 and heterozygous for the c.1772A > G (p.N591S) sequence change in ABCB11.

Discussion

Biosynthesis and recycling of bile acids involve multiple enzymes, organelles and tissues (Mihalik et al. 2002; Kelly 2003). In the recycling pathway, unconjugated C24 bile acids (deconjugated by gut bacteria) returned to the liver in the enterohepatic circulation, are conjugated with either glycine or taurine (amidated) and secreted into bile by BSEP. Two separate enzymes catalyse this conjugation. The first-acting, BACL, converts CDCA, CA and bile acids produced by bacteria, such as deoxycholic acid, to their CoA thioesters (Mihalik et al. 2002; Hubbard et al. 2006). BACL is encoded by SLC27A5. The second-acting, BAAT, catalyses the amidation of a bile acid-CoA thioester to form a glycine or taurine conjugate. BAAT is encoded by BAAT. In de novo bile-acid synthesis, the 27-carbon (C27) bile acids 25(R) 3alpha,7alpha-dihydroxy-5beta-cholestanoic acid (DHCA) and 25(R) 3alpha,7alpha,12alpha-trihydroxy-5beta-cholestanoic acid (THCA) are converted to their CoA esters. Peroxisomal ß-oxidation yields chenodeoxycholoyl-CoA and choloyl-CoA, which can be amidated by BAAT. Two CoA ligases are thought to act on the C27 bile acids–very-long-chain acyl-CoA synthase (VLCS) and BACL (Mihalik et al. 2002). Failure to convert unconjugated bile acids to conjugated bile acids constitutes an amidation defect. Patient AK is the first person identified with mutations in SLC27A5 [although a Slc27a5 knockout mouse has been described (Doege et al. 2006; Hubbard et al. 2006)].

The urinary cholanoid profiles of patient AK completely lacked glycine and taurine conjugates of CDCA and CA (m/z 448, 464, 498, 514). The predominant peak was m/z 407 (unconjugated CA). Other analytes detected in lesser abundance were unconjugated dihydroxycholanoic acid (m/z 391) and sulphate and glucuronide conjugates of dihydroxy- and trihydroxy-cholanoic acids (m/z 391, 471, 487, 567, 583). These results resembled those in amidation-deficient patients of undetermined genetic basis, described by Setchell and co-workers (Setchell and O’Connell 2004) and in the first infant (patient AZ) in whom we identified an amidation defect (with BAAT mutation). The urinary cholanoid profile observed in AK cannot be explained by mutation in ABCB11. We have analysed over 20 urine samples from infants with severe BSEP deficiency and documented ABCB11 mutation, and these show increased urinary excretion of amidated bile acids as found in Fig. 2b.

Plasma total CA concentrations in AK were normal, and total CDCA concentrations were only mildly elevated. Plasma bile acids were >87% unconjugated. This pattern resembles that in the Slc27a5 knockout mouse (Hubbard et al. 2006). In patients with BAAT deficiency, total plasma bile-acid concentrations are often elevated, and plasma bile acids are 100% unconjugated. Some conjugated bile acids occur in patients/mice with BACL/Bacl deficiency because BACL is not essential for de novo synthesis of conjugated bile acids. The production of some conjugated bile acids may explain both a degree of cholestasis (reflected by total plasma bile acid concentrations) which was milder in AK than has been observed in BAAT deficiency and the absence of cholestasis in her sister.

If AK had been investigated by measurement of total plasma bile acids using the 3α-hydroxysteroid dehydrogenase assay, total bile acids would have been interpreted as very slightly elevated; it is low levels of 3α-hydroxy-bile acids in conjugated hyperbilirubinaemia that usually evoke suspicion of a bile-acid synthesis defect.

Patient AK’s plasma bile-acid profile showed no accumulation of THCA and DHCA (precursors of CA and CDCA respectively), suggesting that she has no defect in de novo synthesis of bile acids at the level of synthesis of THCA-CoA and DHCA-CoA. It seems likely that VLCS in peroxisomes and in endoplasmic reticulum (Mihalik et al. 2002) suffices for the generation of CoA thioesters of THCA and DHCA.

In mice, Bacl at the basal plasma membrane of hepatocytes participates in uptake of long-chain fatty acids; livers of Bacl knockout mice contain subnormal concentrations of free fatty acids and triglycerides (Doege et al. 2006). We could not study hepatic lipid composition in AK, a point interesting to explore in future patients.

Sequencing of genomic DNA showed that AK was homozygous for a single base substitution, c.1012C > T, in SLC27A5 (19q13.43). This is predicted to generate the amino-acid residue substitution H338Y in BACL. BACL belongs to the ATP-dependent AMP-binding enzyme family and to the greater family of acyl-CoA synthetases. Sequence annotation (http://www.uniprot.org/uniprot/Q9Y2P5) indicates that the site of substitution is not in the AMP-binding domain of BACL. However, sequence alignment of human SLC27A5 with orthologues in other species (cow, mouse, rat and fowl) shows that this site is in a highly conserved region, suggesting that the mutation very likely disrupts protein function. In addition, the histidine residue in question is highly conserved in many acyl-CoA synthetases. Hisanaga et al. (2004) described this residue as part of a G- (or “gate”) motif: When ATP binds to the enzyme, the gate opens, allowing the fatty acid (or in this case, the bile acid) to enter a binding pocket or tunnel. Watkins et al. (2007) reported that this histidine is invariant in all long-chain and very-long-chain acyl-CoA synthetases as well as in BACL. Substitution of a tyrosine residue generated a catalytically inactive variant. That immunostaining of liver-biopsy materials from AK showed normal amounts of immunoreactive protein, normally distributed, suggests that the variant protein is both stable and normally trafficked.

PM1I restriction-enzyme digestion testing results showed that the patient’s parents were heterozygous for the mutation (consistent with autosomal-recessive inheritance) and that c.1012C > T is not a common SLC27A5 polymorphism among persons of Pakistani descent.

Unexceptionally for an infant born at 27 weeks’ gestation, AK required parenteral nutrition to achieve adequate weight gain. Whilst the Slc27a5 knockout mouse fails to gain weight on a high-fat diet, this is due to decreased food intake and increased energy expenditure rather than to fat malabsorption (Hubbard et al. 2006). As AK’s younger sister, SK, also homozygous for the c1012C > T mutation but born at term, gained weight satisfactorily, failure to thrive is not an inevitable consequence of this BACL amidation defect.

Like patients with BAAT deficiency, AK presented in early infancy with self-limiting cholestasis manifest as conjugated hyperbilirubinaemia, elevated serum transaminase activity, and normal serum GGT activity. Cholestasis persisted until age 31 weeks. By age 49 weeks, clinical biochemistry test result abnormalities had resolved. We suspect, but cannot state definitively, that transient cholestasis was caused principally by prolonged exposure to parenteral nutrition (with a contribution from hepatic immaturity). That AK’s sister SK, who has the amidation defect, did not develop neonatal cholestasis indicates that the defect can be present without causing cholestatic liver disease.

AK was homozygous for a missense mutation in ABCB11. This sequence change, c.1772A > G, has been encountered, in heterozygous state, in association with intrahepatic cholestasis of pregnancy and is assessed as potentially pathogenic (Pauli-Magnus et al. 2004). Our immunohistochemical findings and clinical observations in AK demonstrate that in homozygous state this mutation does not ablate BSEP expression and does not lead to chronic cholestasis; it may have contributed to her transient cholestasis. Heterozygosity for the c.1772A > G ABCB11 mutation in conjunction with a homozygous c.1012C > T SLC27A5 mutation was seen in AK’s sister, SK, and she has shown no signs of cholestasis. Of interest is that these girls’ mother, who demonstrated heterozygosity for this mutation, did not experience symptomatic cholestasis while pregnant.

In patients with BAAT deficiency, symptoms such as fat malabsorption, failure to thrive and coagulopathy reportedly respond to treatment with UDCA (Morton et al. 2000). Glycocholic acid treatment may improve growth and some aspects of fat-soluble vitamin malabsorption (Heubi et al. 2009). Patient AK’s cholestasis resolved concomitant with UDCA treatment; whether UDCA played a role in the resolution of cholestasis is uncertain. Aged 5 years, without UDCA, she is growing normally, without hypovitaminaemia or coagulopathy.

Conclusion

Homozygosity for a missense mutation that alters a highly conserved histidine residue in SLC27A5, encoding bile acid CoA ligase, causes a failure of C24 bile-acid amidation. This failure can be asymptomatic but may contribute to cholestasis and/or malabsorption of fat and fat-soluble vitamins.

Acknowledgements

P.T.C. is funded by Great Ormond Street Children’s Charity. P.G. is funded by a Wellcome Trust Senior Fellowship.

Copyright information

© SSIEM and Springer 2011