Introduction

Biliary atresia represents a frequent cause of neonatal cholestasis. The disease is defined as occlusive panductular cholangiopathy affecting both intra- and extrahepatic bile ducts [1]. Aetiology of biliary atresia is still unknown; genetic background might play a role in some cases [2–5]. Rapid diagnosis and therapy of biliary atresia is essential because the efficiency of surgical reconstruction of the extrahepatic biliary tract by portoenterostomy (the Kasai procedure) is limited by the age of the patient [6]. Most patients require liver transplantation at later age.

Alagille syndrome (OMIM #118450), a complex multisystem autosomal dominant disorder with incomplete penetrance, may be indistingushible from biliary atresia in the neonatal period. The disease is defined clinically by the association of at least three of the five major features—chronic cholestasis, congenital heart disease, skeletal anomalies (typically butterfly vertebrae), ocular abnormalities (primarily posterior embryotoxon) and peculiar face—and genetically by heterozygous state for mutations in JAG1 [7, 8] or rarely in NOTCH2 [9]. Mutations in JAG1 have also been found in a subset of patients with biliary atresia [4]; however, their contribution to clinical symptoms and the disease course remains unclear. Moreover, some patients with Alagille syndrome are likely misdiagnosed as biliary atresia due to variable expression of clinical features of the syndrome in early infancy, which may result in minimal or subclinical disease that can be missed [10].

In this study we focused on patients with biliary atresia. Our aim was to identify and characterize individuals carrying JAG1 mutations among 72 patients with biliary atresia and to confirm the diagnosis in 4 patients with suspected Alagille syndrome having intrahepatic cholestasis without biliary atresia.

Materials and Methods

Results

Five null mutations in JAG1 have been detected in five unrelated patients presenting initially as biliary atresia (patients 1–5 in Table 1). Four of them are novel whereas the fifth (p.Tyr320*, Human Gene Mutation Database accession No. CM124879) has been reported recently [12]. Only one of the mutations (p.Cys438Serfs found in patient 4) was inherited from the patient´s mother characterized by craniofacial dysmorphia. The remaining four mutations are of de novo origin. All five mutation carriers are single heterozygotes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Mutations in JAG1 found in patients with biliary atresia and Alagille syndrome.

https://doi.org/10.1371/journal.pone.0143939.t001

Even after careful re-examination of these five “biliary atresia”patients prompted by their genetic findings, none of them met the diagnostic criteria for Alagille syndrome at time of hospitalization for neonatal cholestasis. All had systolic murmur and one patient had foramen ovale apertum but neither structural anomalies of the heart nor stenosis of branches of pulmonary artery could be documented (Table 2). Moreover, in wedge liver biopsy specimens obtained during the Kasai surgery, portal expansion with ductular proliferation and occasional bile ducts but no bile duct paucity was observed. On long-term follow-up, all five patients developed clinical signs sufficient for clinical diagnosis of Alagille syndrome before 3 years of age (Table 2). Three patients were enlisted for liver transplantation with the end stage liver disease at 3 years of age. Two of them (patients No. 1 and 5 in Tables 1 and 2) underwent successful liver transplantation whereas the third patient (No. 4) died from variceal bleeding.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Clinical features present in carriers of JAG1 mutations at the time of hospitalization for neonatal cholestasis and at 3 years of age.

https://doi.org/10.1371/journal.pone.0143939.t002

No predictably pathogenic mutations in JAG1 have been detected in the remaining 67 patients with biliary atresia by Sanger sequencing and the multiplex ligation-dependent probe amplification assay (S2 Table).

In contrast, three novel mutations were found in three index subjects and one sibling with typical features of Alagille syndrome at the age of six months (patients 6–9 in Table 2). All mutations present in heterozygous state are of maternal origin. The otherwise healthy mothers had craniofacial dysmorphic features. The two frameshift mutations create early stop codons and, provided that the mutated mRNA is not eliminated by nonsense-mediated decay, result in non-functional truncated proteins lacking the membrane and cytoplasmic domain (Fig 1). The missense mutation c.402G>T (p.Leu134Phe) affects exon 3, which encodes a part of the DSL domain of JAG1 (Fig 1). The domain is highly conserved between different species and is essential to activate Notch receptors [13]. The substitution of 134Leu with phenylalanine is predicted as pathogenic using the PredictSNP 1.0 classifier, the reliability of the prediction is 87% (S3 Table). The frameshift mutation p.Asp684fs found in patient 2 is also present in his symptomatic brother (patient 8 in Table 2). Both siblings differ in their clinical presentation (see Table 2); however, this is well in line with the known highly variable expressivity of the disease [14].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schematic representation of the Jagged1 protein and spliced JAG1 mRNA with mutations found in our patients.

https://doi.org/10.1371/journal.pone.0143939.g001

Location in the spliced mRNA and protein sequence of all 8 mutations we have found in our patients is depicted in Fig 1. Positions of the nonsense and frameshift mutations indicate the exact or approximate ends of the hypothetical truncated non-functional proteins. Pathogenicity of the mutations is further supported by their unique character: None of them has occurred in the GenBank SNP database, 1000 Genomes project, Exome Aggregation Consortium database and the Exome Sequencing Project (S3 Table).

Discussion

Confirmation of the clinical diagnosis of Alagille syndrome in patients without biliary atresia by finding mostly novel pathogenic mutations in JAG1 was not surprising. In contrast, the role of JAG1 mutations in aetiology of biliary atresia is still unclear. Since defective Jagged1/Notch-2 signalling is responsible for transdifferentiation of hepatoblasts to biliary epithelia [15] and mutations in either JAG1 or NOTCH2 are known to cause Alagille syndrome type 1 and 2, respectively, association of JAG1 with biliary atresia reported by Kohsaka et al. [4] would be acceptable. These authors found 9 missense mutations present in 11 out of 102 patients with biliary atresia where 28 of them underwent liver transplantation before 5 years of age. None of the mutation carriers developed typical clinical features of Alagille syndrome before the age of 5 years. According to the first hypothesis presented in [4] patients with biliary atresia and JAG1 deficiency could represent atypical Alagille syndrome with not fully expressed clinical features of Alagille syndrome. Alternatively, mutated Jagged1 protein may affect inflammatory processes in the liver via the Jagged1/Notch-2 pathway-mediated regulation of cytokine expression.

Neither we in this study nor any other group to the best of our knowledge, observed a carrier of JAG1 mutation with biliary atresia and no clinical features of Alagille syndrome similar to the 11 carriers of missense mutations reported by the Kohsaka´s group [4]. Association of these missense mutations with Alagille syndrome has not been reported as well. Such discrepancy may be attributed to differences in ethnicity of biliary atresia patients studied in Japan, Europe and USA. Likewise, the role of the JAG1 mutations in genetic aetiology of biliary atresia may be questioned. Eight of the 9 mutations reported by Kohsaka et al. [4] were identified in sporadic cases. Since JAG1 is not a strong biliary atresia candidate disease gene concurrent with the sporadic incidence of the disease, the reported occurrence of rare JAG1 missense mutations does not univocally prove the genotype—phenotype correlation. The correlation is further doubted by the fact that pathogenicity prediction of all but two mutations reported in [4] has been evaluated in silico as neutral by several prediction programs exploited by the PredictSNP 1.0 classifier (S3 Table) and the predictably pathogenic mutation p.Pro871Arg (dbSNP rs35761929) is classified as benign in the ClinVar database owing to its high frequency in the 1000 Genomes project, Exome Aggregation Consortium database and the Exome Sequencing Project (S3 Table).

In our cohort of 72 patients with biliary atresia of which 27 were enlisted for liver transplantation before 5 years of age, we found 5 carriers of JAG1 null mutations but no carrier of a single unique potentially pathogenic missense mutation. All these patients who initially presented as biliary atresia developed sufficient number of clinical signs typical for Alagille syndrome before 3 years of age.

One of the main diagnostic features of Alagille syndrome is bile duct paucity, which is more common later in infancy and childhood [16–18]. Ductular proliferation is present in a small number of infants with Alagille syndrome, leading to significant diagnostic confusion. Because of the variability in the early histopathology of the liver in Alagille syndrome, a number of patients have been misdiagnosed as having biliary atresia [16, 19, 20]. Surgical reconstruction of extrahepatic bile duct system in patients with Alagille syndrome does not correct for the loss of the bile ducts within the liver and liver transplantation is preferred. This is a reason why early molecular diagnostics could be considered in selected cases of overlapping Alagille syndrome and biliary atresia.

A question arises about how the candidates for accelerated JAG1 sequencing should be selected. None of our 5 patients with mutations in JAG1 presenting as biliary atresia in early infancy met the diagnostic criteria for Alagille syndrome at the critical time of hospitalisation for neonatal cholestasis. However, all had pathological findings on endoscopic retrograde cholangiopancreatography (see Table 2) and systolic murmur but no signs of pulmonary stenosis on echocardiography. Therefore we suppose that cardiological phenotype (typical pulmonary stenosis) can develop quite slowly. Since craniofacial dysmorfia and embryotoxon posterior are difficult to find in newborns or patients in early infancy and liver histology may be inconclusive, we suggest that patients with neonatal cholestasis due to biliary atresia and either systolic murmur or any single clinical feature characteristic for Alagille syndrome should be elicited for mutational analysis of JAG1 before surgery.

In conclusion, our findings do not support association of biliary atresia with JAG1 mutations. In addition to liver histology, mutational analysis of JAG1 could be useful for diagnosis of the “grey zone”patients with Alagille syndrome presenting initially as biliary atresia in early infancy.

Supporting Information

Acknowledgments

The authors are thankful for support of the families of the biliary atresia and Alagille syndrome patients, and would also like to thank Lucie Budišová for technical assistance and Mariia Lunova, Ph.D., for critical reading the manuscript.