Abstract
Background and Aims
The aim was to determine if liver biochemistry indices can be used as biomarkers to help differentiate patients with neonatal Dubin–Johnson syndrome (nDJS) from those with biliary atresia (BA).
Methods
Patients with genetically-confirmed nDJS or cholangiographically confirmed BA were retrospectively enrolled and randomly assigned to discovery or verification cohorts. Their liver chemistries, measured during the neonatal period, were compared. Predictive values were calculated by receiver operating characteristic curve analysis.
Results
A cohort of 53 nDJS patients was recruited, of whom 13 presented with acholic stools, and 14 underwent diagnostic cholangiography or needle liver biopsy to differentiate from BA. Thirty-five patients in the cohort, with complete biochemical information measured during the neonatal period, were compared with 133 infants with cholangiographically confirmed BA. Total and direct bilirubin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bile acids, alkaline phosphatase, and gamma-glutamyl transferase were significantly lower in nDJS than in BA. In the discovery cohort, the areas under the curve for ALT and AST were 0.908 and 0.943, respectively. In the validation cohort, 13/15 patients in the nDJS group were classified as nDJS, and 10/53 in the BA control group were positive (p<0.00001) with an ALT biomarker cutoff value of 75 IU/L. Thirteen of 15 patients were classified as nDJS and none were classified positive in the BA group (13/15 vs. 0/53, p<0.00001) with an AST cutoff of 87 IU/L.
Conclusions
Having assembled and investigated the largest cohort of nDJS patients reported to date, we found that nDJS patients could be distinguished from BA patients using the serum AST level as a biomarker. The finding may be clinically useful to spare cholestatic nDJS patients unnecessary invasive procedures.
Keywords: Jaundice, ABCC2, Cholestasis, MRP2, Acholic stools
Introduction
Hyperbilirubinemia II, or Dubin–Johnson syndrome (DJS; OMIM 237500), was first reported by Dubin and Johnson in 1954.1 It is an autosomal recessive disorder caused by variations in the ATP-binding cassette subfamily C member 2 (ABCC2) gene that result in a decrease in the production or loss of function of multidrug resistance-associated protein 2 (MRP2).2,3 The disorder presents in adolescence and is characterized by a low-grade elevation of conjugated bilirubin in the blood, and histologically with the accumulation of dark, coarsely granular, melanin-like pigment in centrilobular hepatocytes, with no other signs of hepatic injury.4,5 It has been reported that DJS can present with severe cholestasis and hepatomegaly in neonates (nDJS).6–8 However, with only a limited number of reported cases, the clinical symptoms, pathologic signs, and genetic features of nDJS remain poorly described.
Biliary atresia (BA) is the leading cause of pediatric liver transplantation worldwide. It presents as an obliterative cholangiopathy with neonatal jaundice and pale stools.9 Prompt diagnosis and the Kasai procedure improve the odds of native liver survival. However, differentiation from other causes of neonatal cholestasis is difficult because of the low specificity of imaging studies, including ultrasonography, hepatobiliary iminodiacetic acid scans, and magnetic resonance cholangiopancreatography, among others. Quite often, liver biopsy or surgical cholangiography is needed to confirm the diagnosis.9
As nDJS patients usually do not require specific treatment, but BA needs early surgical intervention,10 early differentiation is critical. A previous study in Japan by Togawa et al.,11 reported that eight of ten nDJS patients underwent liver biopsy during neonatal cholestasis because they were suspected of having BA. The primary aim of this study is to better define the clinical and pathologic features of nDJS and to determine whether convenient biomarkers can be identified for helping to distinguish nDJS from BA.
Methods
Subjects
This was a single-center retrospective study. The subjects were Chinese children who were admitted to the Children’s Hospital of Fudan University, between 2011 and 2021 for the study of neonatal cholestasis with consent under a protocol (ethical approvals 2015-178 and 2020-402) in accordance with the ethical guidelines of the 1975 Declaration of Helsinki. nDJS was defined as onset of cholestasis before 3 months of age, with confirmed homozygous or compound-heterozygous variants of uncertain significance, or with pathogenic/likely pathogenic ABCC2 variants, according to American College of Medical Genetics /Association for Molecular Pathology guidelines,12 and after exclusion of other causes such as infection, hemolytic jaundice, endocrine disease, bile acid synthesis disorders, and other metabolic conditions. Of 53 genetically-confirmed nDJS patients (Table 1), the 35 patients with complete liver biochemistry data obtained during the neonatal period were selected as the test subjects. A group of 133 age- and sex-matched cholangiography-confirmed BA patients, with available pre-operation liver biochemistry indices, were used as the control group. Their liver chemistry indices were compared and analyzed by receiver operating characteristic (ROC) curves. A total of 80 patients with BA and 20 with nDJS were randomized to the discovery cohort; in addition, 53 patients with BA and 15 with nDJS were included in the validation cohort.
Table 1. Clinical and genetic characteristics of the study subjects.
| P | S | ABCC2 variation | Acholic stools | Jaundice | Liver below the rib | Liver below the xiphoid | Spleen below the rib | Cholangiography | Liver biopsy | Additional observations |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | M | c.1627C>T (p.Gln543Ter); c.3825C>G (p.Tyr1275Ter) | − | + | − | − | − | |||
| 2 | M | c.3825C>G (p.Tyr1275Ter); c.2755T>A (p.Ser919Thr), c.4239_4240dupTC (p.His1414LeufsTer18) | + | |||||||
| 3 | M | c.2303G>A (p.Arg768Gln); c.3476T>G (p.Ile1159Ser) | + | jaundice free after infant, recurrent jaundice at 8.5 years | ||||||
| 4 | F | c.3825C>G (p.Tyr1275Ter); c.1399G>A (p.Val461Ile) | + | − | − | − | ||||
| 5 | M | c.298C>T (p.Arg100Ter); c.3928C>T (p.Arg1310Ter) | + | − | − | − | persistent jaundice at 10 years | |||
| 6 | M | c.3825C>G (p.Tyr1275Ter); c.1322_1325dupTGTG (p.Trp442CysfsTer37) | + | − | − | − | transient prolonged prothrombin time (international normalized ratio [INR]) | |||
| 7# | M | c.1177C>T (p.Arg393Trp); c.1177C>T (p.Arg393Trp) | + | 2 cm | 1 cm | − | + | jaundice free at 7.5 months | ||
| 8# | F | c.1177C>T (p.Arg393Trp); c.4024T>C (p.Ser1342Pro) | − | + | − | − | − | subclinical jaundice at 4.3 months | ||
| 9# | F | c.1177C>T (p.Arg393Trp); c.4024T>C (p.Ser1342Pro) | + | + | 2 cm | − | − | + | persistent jaundice at 6.2 months, subclinical jaundice at 7.3 months and 9.4 months | |
| 10 | F | c.1815+3_1815+6dup; c.1813delinsCAGGT (p.Ala606ValfsTer17) | + | − | − | − | ||||
| 11 | F | c.1281T>G (p.Asp427Glu); c.4025C>A (p.Ser1342Tyr) | + | − | − | − | ||||
| 12# | M | c.1177C>T (p.Arg393Trp); c.2980delA (p.Ile994LeufsTer29) | + | persistent jaundice at 8.9 months | ||||||
| 13 | M | c.3825C>G (p.Tyr1275Ter); c.1177C>T (p.Arg393Trp) | + | + | mother had itching during pregnancy, jaundice free at 11.0 months | |||||
| 14# | M | c.632+2_632+5delTAGG; c.4239_4240dupTC (p.His1414LeufsTer18) | + | |||||||
| 15# | M | c.2302C>T (p.Arg768Trp); c.4024T>C (p.Ser1342Pro) | + | + | 4 cm | 2 cm | − | + | premature delivery, subclinical jaundice at 7.4 months | |
| 16# | M | c.3825C>G (p.Tyr1275Ter); c.4146+1G>T | − | + | 2 cm | 2 cm | − | persistent jaundice at 2.2 years | ||
| 17# | F | c.1177C>T (p.Arg393Trp); c.4145A>G (p.Gln1382Arg) | + | jaundice free at 8.0 months, subclinical jaundice at 1.0 year | ||||||
| 18# | F | c.2821delG (p.Glu941LysfsTer5); c.4313+1G>T | − | + | 3 cm | − | − | low muscle tension | ||
| 19 | F | c.1018C>A (p.Pro340Thr); c.1968-1G>A | + | 2.5 cm | 2.5 cm | − | ||||
| 20# | M | c.1177C>T (p.Arg393Trp); c.2063T>C (p.Met688Thr) | + | + | 2.3 cm | − | 2.3 cm | + | + | jaundice free at 5 months |
| 21# | M | c.4250delC (p.Ser1417PhefsTer14); c.4256T>G (p.Val1419Gly) | + | jaundice free at 4.1 months | ||||||
| 22 | M | c.1177C>T (p.Arg393Trp); c.3660delC (p.Ser1222GlnfsTer3) | + | 1 cm | − | − | jaundice free at 4.7 months | |||
| 23# | F | c.4384G>A (p.Glu1462Lys); c.4384G>A (p.Glu1462Lys) | + | + | ||||||
| 24# | M | c.2755T>A (p.Ser919Thr); c.2883+1G>A, c.4239_4240dupTC (p.His1414LeufsTer18) | − | + | subclinical jaundice at 3.3 years | |||||
| 25# | F | c.1177C>T (p.Arg393Trp); c.1177C>T (p.Arg393Trp) | + | + | 3 cm; | 2.5 cm | − | + | + | |
| 26# | F | c.3259-1G>A; c.3191T>A (p.Ile1064Asn) | + | |||||||
| 27# | M | c.3825C>G (p.Tyr1275Ter); c.1177C>T (p.Arg393Trp) | + | 2 cm | 3 cm | − | + | + | persistent jaundice at 4.9 months | |
| 28# | M | c.338T>C (p.Leu113Pro); c.338T>C (p.Leu113Pro) | + | + | 1 cm | − | − | + | + | premature, hypoglycemia, convulsion, atrial septal defect, retarded brain development |
| 29 | M | c.2443C>T (p.Arg815Ter); c.2078G>A (p.Gly693Glu) | + | − | − | − | persistent jaundice at 7.0 months | |||
| 30# | M | c.1177C>T (p.Arg393Trp); c.1177C>T (p.Arg393Trp) | − | + | premature delivery; hypoglycemia; atrial septal defect | |||||
| 31# | M | c.2302C>T (p.Arg768Trp); c.3988-3C>G, c.1457C>T (p.Thr686Ile) | + | − | − | − | jaundice free at 9.6 months, subclinical jaundice at 2.5y | |||
| 32# | M | c.2755T>A (p.Ser919Thr); c.4239_4240dupTC (p.His1414LeufsTer18) | + | + | 2.5 cm | 1.5 cm | − | jaundice free at 1.4 years | ||
| 33# | F | c.3825C>G (p.Tyr1275Ter); c.2201T>C (p.Leu734Pro) | + | + | ||||||
| 34# | M | c.2366C>T (p.Ser789Phe); c.2362_2363delCT (Leu788ValfsTer13) | − | + | transient prolonged prothrombin time (INR) | |||||
| 35 | M | c.4239_4240dupTC (p.His1414LeufsTer18), c.1462C>T (p.Gln488Ter), c.2755T>A (p.Ser919Thr) | + | + | 4 cm | 3.5 cm | − | |||
| 36 | F | c.4384delG (p.Glu1462ArgfsTer8); c.298C>T (p.Arg100Ter) | + | − | − | − | + | N/A | ||
| 37 | M | c.2026G>C (p.Gly676Arg); c.1939G>T (p.Glu647Ter) | + | + | − | − | − | + | jaundice free after infancy, recurrent subclinical jaundice at 3.2 years after fever | |
| 38# | F | c.3825C>G (p.Tyr1275Ter), c.2755T>A (p.Ser919Thr), c.4239_4240dupTC (p.His1414LeufsTer18) | + | − | − | − | ||||
| 39 | F | c.3928C>T (p.Arg1310Ter); c.1882C>T (p.Arg628Ter) | − | + | 4 cm | 2.5 cm | − | |||
| 40# | F | c.1177C>T (p.Arg393Trp); c.1177C>T (p.Arg393Trp) | + | + | 2.5 cm | 3.5 cm | 2 cm | + | transient prolonged prothrombin time (INR) | |
| 41# | F | c.338T>C (p.Leu113Pro); c.2855_2858delTTAA (p.Ile952ArgfsTer5) | + | 3 cm | − | − | jaundice free at 5.1 months | |||
| 42# | F | c.4239_4240dupTC (p.His1414LeufsTer18); c.2755T>A (p.Ser919Thr), c.4343G>A (p.Gly1448Asp) | − | + | 1.5 cm | 3 | − | hypoglycemia, persistent jaundice at 7.5 months | ||
| 43# | F | c.1177C>T (p.Arg393Trp); c.2980delA (p.Ile994LeufsTer29) | + | − | − | − | + | + | subclinical jaundice at 3.9 months | |
| 44# | F | c.3825C>G (p.Tyr1275Ter); c.2077G>A (p.Gly693Arg) | − | + | 2 cm | − | − | |||
| 45# | F | c.3928C>T (p.Arg1310Ter); c.116delA (p.Tyr39SerfsTer40) | − | + | 2 cm | 2 cm | − | |||
| 46# | M | c.1939G>T (p.Glu647Ter); c.4384delG (p.Glu1462ArgfsTer8) | − | + | 3 cm | − | − | |||
| 47# | F | c.3196C>T (p.Arg1066Ter); c.2439+5G>A | − | + | 2 cm | − | − | |||
| 48# | M | c.3825C>G (p.Tyr1275Ter); c.3825C>G (p.Tyr1275Ter) | − | + | − | − | − | |||
| 49 | M | c.4465_4473delinsGGCCCACAG (p.Ile1489_Ile1491delinsGlyProGln); c.2302C>T (p.Arg768Trp) | − | + | 2.5cm | 2.5 cm | − | |||
| 50# | M | c.389G>A (p.Trp130Ter); c.3196C>T (p.Arg1066Ter) | + | + | 2.5 cm | 2 cm | − | |||
| 51 | M | c.1939G>T (p.Glu647Ter); c.3928C>T (p.Arg1310Ter) | − | + | − | − | − | |||
| 52# | M | c.2830G>T (p.Glu944Ter); c.4465_4473delinsGGCCCACAG (p.Ile1489_Ile1491delinsGlyProGln) | − | + | 2 cm | − | 2 cm | |||
| 53# | M | c.1078G>T (p.Gly360Ter), c.4239_4240dupTC (p.His1414LeufsTer18), c.2755T>A (p.Ser919Thr) | + | + | 3 cm | 4 cm | − | + | + |
The screening criteria for BA included: (1) Onset of clinical manifestations of BA, such as prolonged jaundice with high GGT levels and acholic stools during the neonatal period; (2) A cholangiogram showing that the bile ducts were not patent during the cholangiography; (3) Biopsy specimens evaluated by two pathologists to ascertain the diagnosis of BA; (4) Exclusion of other causes of neonatal cholestasis through appropriate investigations, such as Alagille syndrome, citrin deficiency, progressive familial intrahepatic cholestasis type 3 and other metabolic conditions.
Ninety-four age-matched patients with neonatal intrahepatic cholestasis were chosen as cholestasis controls (Supplementary Table 1). They included four patients with bile acid synthesis defects, three with citrin deficiency, one with progressive familial intrahepatic cholestasis type 2, 21 with Alagille syndrome, and 65 with an unknown etiology, i.e. testing for neonatal hemochromatosis, viral hepatitis, and others. failed to identify an etiology.
Pathogenicity of newly discovered missense ABCC2 variants
The pathogenicity of newly discovered missense variants was predicted by in silico tools including the MutationTaster (http://www.mutationtaster.org/), Polymorphism Phenotyping v2 (Polyphen-2, http://genetics.bwh.harvard.edu/pph2/index.shtml), Sorting Intolerant From Tolerant (SIFT, http://sift.jcvi.org), Protein Variation Effect Analyzer (PROVEAN, http://provean.jcvi.org/index.php), Mendelian Clinically Applicable Pathogenicity (http://bejerano.stanford.edu/MCAP/), and Functional Analysis Through Hidden Markov Models (http://fathmm.biocompute.org.uk/). The pathogenicity of noncanonical splicing variants and nonframeshift insertion/deletion variants was predicted by MutationTaster. Default settings were used for all in silico tools.
Liver histopathology
Liver biopsy specimens were obtained from one patient with neonatal Dubin–Johnson syndrome during evaluation of the cause of cholestasis and from seven during laparoscopic cholangiography because BA was suspected. The specimens were processed routinely, and 4 µm sections were cut from formalin-fixed, paraffin-embedded blocks and stained with hematoxylin and eosin for routine histologic evaluation of collagen, bile, and melanin-like pigment deposits in hepatocytes.
Statistical analysis
Statistical analysis was performed with SPSS 19.0 (IBM Corp., Armonk, NY, USA) to determine differences between the nDJS patients and control groups. Kolmogorov-Smirnov tests were performed to check if the data were normally distributed. Student’s t-test was performed when the data had a normal distribution. A nonparametric test, the Mann-Whitney U test, was performed when the data did not have a normal distribution. Data were reported as medians and interquartile range, or means ± SD. ROC curve analysis was used to calculate the areas under the curve (AUC) with 95% confidence intervals (CIs).
Results
Clinical features of nDJS patients
During the study period, about 7,000 patients were referred to us for investigation of the causes of neonatal cholestasis. Of those 53 (31 men and 22 women) were diagnosed with nDJS (Table 1). Cholestasis subsided or improved during follow-up in all nDJS patients, and none failed to thrive. Of 28 patients with a recorded date of onset of jaundice during the neonatal period, 25 visited a pediatric gastroenterologist because of jaundice shortly after birth (0–10 days after birth, with an average of 3.6±1.9 days. Another three patients presented with jaundice at the first, second, and third month after birth. Peak jaundice was detected at 2–79 (27.9±19.9) days after birth, during the first 3 months. Peak total bilirubin (TB) ranged from 3.5–29.3mg/dL (13.9±6.3mg/dL) and peak direct bilirubin (DB) ranged from 2.3–22.2mg/dL (6.2±4.5mg/dL). Twenty-two patients were followed-up at our center. Jaundice disappeared in seven patients (P7, P13, P20, P21, P22, P32, P41), declined to subclinical levels (TB levels between 1–2mg/dL) in five (P8, P9, P15, P24, P43), and persisted in six (2–4mg/dL; P5, P12, P16, P27, P29, P42) at the last follow-up. Jaundice in patients P3, P17, and P31 disappeared at follow-up but relapsed as subclinical jaundice of unknown cause. P37 experienced a recurrence of subclinical jaundice at 3.2 years of age after a fever lasting for 2 days.
The liver and spleen were palpable 2–3 cm below the costal margin or xiphoid in some cases, but were not palpable in others. Fourteen patients had cholangiography or liver biopsy or both, to verify the presence of BA. Thirteen of 30 patients with descriptions of stool color were reported to have acholic stools. All infants diagnosed with DJS were full-term except for patients P15, P28, and P30 (Table 1). Three (P28, P30, P42) presented with transient hypoglycemia, and three (P6, P34, P40) presented with transient prolonged prothrombin time responsive to vitamin K1 injection. One patient (P18) had low muscle tension. The mother of P13 reported itching during pregnancy, and P28 had retarded brain development. The characteristics of BA patients and cholestasis controls are shown in Supplementary Material 1.
ABCC2 variants
Fifty-five ABCC2 variants were identified in 53 nDJS patients (Table 1). c.1177C>T (p.Arg393Trp), c.3825C>G (p.Tyr1275Ter), c.2755T>A (p.Ser919Thr), c.4239_4240dupTC (p.His1414LeufsTer18), c.3928C>T (p.Arg1310Ter) occurred in seventeen, twelve, seven, seven, and four alleles, respectively; c.1939G>T (p.Glu647Ter), c.2302C>T (p.Arg768Trp), c.338T>C (p.Leu113Pro), c.4024T>C (p.Ser1342Pro) were each found in three alleles, while c.2980delA (p.Ile994LeufsTer29), c.298C>T (p.Arg100Ter), c.3196C>T (p.Arg1066Ter), c.4384delG (p.Glu1462ArgfsTer8), c.4384G>A (p.Glu1462Lys), c.4465_4473delinsGGCCCACAG (p.Ile1489_Ile1491delinsGlyProGln) occurred two times each.
Identification of novel ABCC2 variants and in silico assessment
Twenty-two of the variations we observed had been previously reported,10,11,13–16 and thirty-three were novel (Table 2). Of the novel variants, six were canonical splicing variants, two were noncanonical splicing variants, five were nonsense variants, thirteen were missense variants, and seven were frameshift variants. No significant correlations were observed between the type of variant and the presence of acholic stools (Supplementary Material 2, Supplementary Table 2).
Table 2. Allele frequencies and in silico identification of 33 ABCC2 variants.
| cDNA change (NM_000392.5) | Protein change (NP_000383.2) | gnomAD | gnomAD_EAS | MuT | SIFT | PPH2 | PROVEAN | M-CAP | FATHMM | ACMG |
|---|---|---|---|---|---|---|---|---|---|---|
| c.338T>C | p.Leu113Pro | 0.00004950(0/4/282,822) | 0.0007016(0/14/19,954) | D | D | D | D | D | D | VUS |
| c.389G>A | p.Trp130Ter | 0.000003978(0/1/251,358) | 0(0/0/18,394) | NMD | / | / | / | / | / | P |
| c.632+2_632+5delTAGG | / | – | – | D | / | / | / | / | / | LP |
| c.1018C>A | p.Pro340Thr | – | – | D | D | D | D | D | D | VUS |
| c.1078G>T | p.Gly360Ter | – | – | NMD | / | / | / | / | D | P |
| c.1281T>G | p.Asp427Glu | – | – | D | D | D | D | D | D | VUS |
| c.1322_1325dupTGTG | p.Trp442CysfsTer37 | – | – | NMD | / | / | / | / | / | P |
| c.1399G>A | p.Val461Ile | 0.000003981(0/1/251,204) | 0.00005440(0/1/18,384) | D | N | N | N | D | D | VUS |
| c.1462C>T | p.Gln488Ter | – | – | NMD | / | / | / | / | / | P |
| c.1627C>T | p.Gln543Ter | – | – | NMD | / | / | / | / | / | P |
| c.1813delinsCAGGT | p.Ala606ValfsTer17 | – | – | NMD | / | / | / | / | / | P |
| c.1815+3_1815+6dup | / | – | – | D | / | / | / | / | / | LP |
| c.1968-1G>A | / | 0.000003979(0/1/251,348) | 0(0/0/18,392) | D | / | / | / | / | / | P |
| c.2063T>C | p.Met688Thr | – | – | D | D | D | D | D | D | VUS |
| c.2821delG | p.Glu941LysfsTer5 | – | – | NMD | / | / | / | / | / | P |
| c.2830G>T | p.Glu944Ter | – | – | NMD | / | / | / | / | / | P |
| c.2883+1G>A | / | – | – | D | / | / | / | / | / | P |
| c.2980delA | p.Ile994LeufsTer29 | – | – | NMD | / | / | / | / | / | P |
| c.3191T>A | p.Ile1064Asn | – | – | D | N | D | D | D | D | VUS |
| c.3259-1G>A | / | – | – | D | / | / | / | / | / | P |
| c.3476T>G | p.Ile1159Ser | – | – | D | D | D | D | D | D | VUS |
| c.3660delC | p.Ser1222GlnfsTer3 | – | – | NMD | / | / | / | / | / | P |
| c.3988-3C>G | / | – | – | D | / | / | / | / | / | P |
| c.4024T>C | p.Ser1342Pro | – | – | D | D | D | D | D | D | VUS |
| c.4025C>A | p.Ser1342Tyr | – | – | D | D | D | D | D | D | VUS |
| c.4146+1G>T | / | – | – | D | / | / | / | / | / | P |
| c.4238_4239dupCT | p.His1414LeufsTer18 | – | – | NMD | / | / | / | / | / | P |
| c.4256T>G | p.Val1419Gly | – | – | D | D | D | D | D | D | VUS |
| c.4313+1G>T | / | – | – | D | / | / | / | / | / | P |
| c.4343G>A | p.Gly1448Asp | – | – | D | D | D | N | D | D | VUS |
| c.4384delG | p.Glu1462ArgfsTer8 | – | – | NMD | / | / | / | / | / | P |
| c.4384G>A | p.Glu1462Lys | – | – | D | D | D | D | D | D | VUS |
| c.4465_4473delinsGGCCCACAG | p.Ile1489_Ile1491delinsGlyProGln | – | – | D | / | / | / | / | / | LP |
ACMG: American College of Medical Genetics and Genomics; FATHMM, Functional Analysis Through Hidden Markov Models (http://fathmm.biocompute.org.uk/); –, not reported/not applicable; D, disease-causing; gnomAD and gnomAD_EAS, allele frequencies of corresponding variants in all populations and in East Asian populations in gnomAD (http://gnomad-old.broadinstitute.org/), respectively; LP, likely pathogenic variants; M-CAP, Mendelian Clinically Applicable Pathogenicity (http://bejerano.stanford.edu/MCAP/); MuT: MutationTaster (http://www.mutationtaster.org); N, no effect; na, not available; NMD, nonsense-mediated mRNA decay; P, pathogenic; PPH2: PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/index.shtml); PROVEAN: Protein Variation Effect Analyzer (http://provean.jcvi.org/index.php); SIFT: Sorting Intolerant From Tolerant (http://provean.jcvi.org/index.php); VUS, variants of uncertain significance.
Biochemical features
Total protein and albumin values were within the normal range in DJS patients (data not shown). We evaluated serum TB, DB, total bile acids (TBAs), alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), and gamma-glutamyl transferase (GGT) in 35 patients during neonatal cholestasis between 16 and 93 days of age. In those nDJS patients the median and interquartile range was determined for TB [7.2 (5.3, 9.1) mg/dL], DB [4.5 (3.3, 5.8) mg/dL], ALT [29.7 (16.5, 45) IU/L], AST [33.5 (23.1, 68.8) IU/L], ALP [386 (326, 524) IU/L], GGT [149 (84.5, 216.5) IU/L], and TBA [72.2 (30.6, 109.1) µmol/L]. TB, DB, ALT, AST, GGT, ALP, and TBA levels were all significantly lower in infants with nDJS than in infants with BA. DB, ALT, AST, and ALP levels were also significantly lower in nDJS patients than in cholestasis controls (Fig. 1). No significant correlations were observed between the type of variant and the levels of TB, DB, ALT, and AST (Supplementary Material 2, Supplementary Fig. 1). In addition, TB, DB, ALT, AST, TBA, ALP, and GGT levels had no significant linear correlation with days after birth in any of the patient groups. (Supplementary Fig. 2).
Fig. 1. Biochemical features of neonatal Dubin-Johnson syndrome (nDJS) patients.
Total bilirubin (TB), direct bilirubin (DB), total bile acids (TBA), alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transferase (GGT) and alkaline phosphatase (ALP) values in 53 neonatal DSJ patients were compared with 133 biliary atresia (BA) patients and 94 cholestasis controls.
Pathological findings
Liver biopsy specimens were obtained from eight patients at 1–3 months of age, six of whom had undergone cholangiography and one with a needle liver biopsy to verify the presence of BA. Only one patient was reported to have a dark red liver during the procedure, at 40 days of age. Histopathologically, the presence of fibrosis, cirrhosis, or melanin-like pigment deposits in hepatocytes was not observed, except fibrosis in P20. However, giant-cell transformation of hepatocytes/hepatocyte ballooning (6/8), steatosis (4/8), and slight intracanalicular and intracytoplasmic cholestasis, i.e. bile plugs (7/8) were observed. The results of five patients with liver biopsies at our center are shown in Figure 2.
Fig. 2. Histopathologic findings in the livers of five patients.
Hematoxylin and eosin staining showed giant-cell transformation of hepatocytes and hepatocyte ballooning, steatosis (arrow) and slight intracanalicular and intracytoplasmic cholestasis with bile plugs, circles). Presence of inflammation was observed in P20, as confirmed by Masson staining (line 2). No inflammation, fibrosis, cirrhosis, or melanin-like pigment deposits in hepatocytes were observed in P25, P27, P28, and P53.
ROC curve analysis
Performance of discriminatory features of biomarkers for nDJS diagnosis: 35 nDJS patients and 133 age-matched subjects with BA were enrolled in the study (Supplementary Fig. 3). To investigate the predictive values of discriminatory features (TB, DB, TBA, ALT, AST, GGT, and ALP levels) for nDJS, we calculated the AUC of each feature in the discovery cohort (Fig. 3). The AUC for ALT in nDJS was 0.908 (95% CI: 0.834–0.982; p<0.0001) and that of AST was 0.943 (95% CI: 0.861–1.000; p<0.0001; Fig. 3E). The sensitivity and specificity for ALT were 81.25% and 90%, respectively, at a cutoff value 75 IU/L. The optimal threshold cutoff for AST in nDJS, was 87 IU/L with a sensitivity of 95% and a specificity of 90%. We tested the discriminating power of ALT and AST using a double-blind strategy in a validation cohort of 68 subjects, 15 with DJS and 53 with BA (Table 3). Thirteen of the 15 nDJS were classified as positive and 10 of the 53 BA patients were classified as positive using ALT as the biomarker. With an AST cutoff value of <87 IU/L, 13 of 15 patients were classified as positive in the nDJS group, and 0 of the 53 BA patients were classified as positive.
Fig. 3. Receiver operating characteristic curve analysis of discriminatory features in 20 neonatal Dubin-Johnson syndrome (nDJS) patients and 80 biliary atresia (BA) patients in the discovery cohort.
The areas under the curve (AUC) for serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) are >0.9 for differentiating nDJS from BA. The AUCs are shown with 95% CIs.
Table 3. Performance of ALT (cutoff value <75 U/L) and AST (cutoff value <87 IU/L) in patients in the validation cohort with nDJS and BA.
| Positive | Negative | Total | ||
|---|---|---|---|---|
| ALT | ||||
| nDJS | 13 | 2 | 15 | |
| BA | 10 | 43 | 53 | p<0.00001 |
| AST | ||||
| nDJS | 13 | 2 | 15 | |
| BA | 0 | 53 | 53 | p<0.00001 |
ALT, alanine aminotransferase; AST, aspartate aminotransferase; BA, biliary atresia; nDJS, neonatal Dubin–Johnson syndrome.
Discussion
Accurate diagnosis of DJS in neonates is important to rule out other hepatobiliary disorders such as BA, which is the most common cause of neonatal cholestasis and usually progresses rapidly, with severe liver injury leading to liver transplantation.5,10 nDJS is a relatively mild, autosomal recessive disorder that does not develop into severe liver injury, but manifests with severe cholestasis and acholic stools during the neonatal period, a presentation that sometimes overlaps with BA. The Kasai hepatic portoenterostomy procedure for re-establishing bile flow is the most effective surgical intervention for survival in BA. The liver is preserved, and the procedure is time-sensitive, with best outcomes if performed before 60 days of age.17 Thus, patients suspected of BA often undergo invasive liver biopsy/cholangiography in order to obtain a clear diagnosis. However, the similarity in presentation between nDJS and BA, and the importance of a successful differential diagnosis, has not attracted much attention and is not a topic discussed in the guidelines for evaluating cholestatic jaundiced infants by the North American and European Societies for Pediatric Gastroenterology, Hepatology, and Nutrition.18 The lack of diagnostic differentiation often leads to nDJS patients being sent for liver biopsy to exclude BA, as shown in our study cohort and in a previous Japanese study cohort.11
To address that issue, we assembled the largest cohort of nDJS patients reported to date at our center to characterize their clinical symptoms, liver chemistries, and pathologic and genetic features. We investigated whether a convenient, noninvasive, biomarker for the early diagnosis of nDJS, differentiating it from BA, could be found. We report here that the liver biochemistry indices of nDJS were generally lower than those in BA. ROC curve analysis of various enzymes revealed that serum ALT and AST levels were robust differentiators of DJS and BA in neonates. The use of an AST cutoff of <87 IU/L was particularly helpful in distinguishing DJS from BA (13/15 nDJS vs. 0/53 BA patients). The finding has potential implications for clinical practice, as liver chemistries are routine and rapid. We believe that such information may be particularly useful in community clinics and nontertiary centers to spare patients invasive procedures such as liver biopsy or cholangiography for diagnosis when nDJS is suspected.5 This study found that an invasive cholangiography can be avoided for the differentiation of BA in patients with cholestasis, acholic stools and AST levels consistently <80 IU/L. We recognize that elevation of ALT and AST can be affected by a variety of factors, including infection, alcohol, and drugs.19,20 AST levels might also be affected by an improper blood collection process or hemolysis. Caregivers must pay attention to those factors when using ALT and AST as biomarkers for the diagnosis of nDJS. Only Chinese patients were enrolled in this study. It would be of interest to determine if the observations made in this report are seen in children from other ethnic groups.
It should be noted that increased urinary excretion of coproporphyrin isomer I is a characteristic feature of adult DJS21 and has been proposed by Junge and Norman10 as a potential biomarker to differentiate nDJS from BA. They concluded that urinary coproporphyrin analysis is a fast and reliable diagnostic tool for differentiating nDJS from BA in a comparison of four nDJS patients and twenty-six BA patients. We did not include a urinary coproporphyrin isomer in this study since because we did not have access to a commercial test. The same was true for the 10 nDJS patients reported from the study in Japan.11
Acholic stools are an important diagnostic feature of BA, but their presence in nDJS patients has also been reported by several centers.11,15,22 In a previous study in Japan, eight of ten nDJS patients underwent a liver biopsy because of a suspicion of BA.11 In this study, 13 of the 30 patients, with a described stool color, were reported as having acholic stools, indicating that acholic stool is a relatively common feature of nDJS. Our data suggest that acholic stools are not a reliable differentiator of nDJS and BA.
In adult patients, black livers, or melanin-like pigment. are typical histological characteristics of DJS, but not all nDJS patients present with black livers or melanin-like pigment.11,14,23 In this study, no melanin-like pigments were observed in any of the eight patients we biopsied, which indicates that a black liver, or accumulation of melanin-like pigment deposits in hepatocytes, is not a reliable characteristic in nDJS patients. That might be because the infants were too young for accumulation of the pigments, as confirmed by a reported case with prolonged cholestasis in early infancy, and an accumulation of specific pigment granules in the liver cells by 6 years of age.23 The absence of histological accumulation of melanin-like pigment increases the difficulty of diagnosing DJS during the neonatal period.
This study reports 33 new DJS variants among the 55 identified in the patients. That substantially expands the current ABCC2 mutation spectrum. Some alleles [c.1177C>T (p.Arg393Trp), c.3825C>G (p.Tyr1275Ter), c.4239_4240dupTC (p.His1414LeufsTer18), c.2755T>A (p.Ser919Thr), c.3928C>T (p.Arg1310Ter), c.298C>T (p.Arg100Ter), c.3196C>T (p.Arg1066Ter), c.2302C>T (p.Arg768Trp), c.1939G>T (p.Glu647Ter)] occurred multiple times, suggesting that they are frequent variants in the population. No significant correlations of patient genotype and the presence of acholic stools were found in the study cohort (Supplementary Material 2). It is reasonable to assume that there exists some other factors affecting the disease phenotype. In addition, we observed some novel clinical features in some patients, including three with transient hypoglycemia and three with transient prolonged prothrombin time (INR), which might be worthy of future investigation.
In conclusion, we assembled the largest reported cohort of neonatal Dubin–Johnson syndrome patients that has been evaluated. The most likely negative outcome of their condition is invasive and unnecessary testing to exclude a diagnosis of BA. For that reason, we explored the diagnostic value of liver biochemistry, which is convenient and routine, and found that serum AST level has the potential to act as a sensitive biomarker for the differentiation of suspected nDJS from a possible misdiagnosis of BA in neonates. A formal diagnosis of nDJS would then follow once confirmatory DNA sequencing at the ABCC2 locus was carried out.
Supporting information
No significant correlations were observed with different genotype groups except for TBA (biallelic non-null variants vs. biallelic-null variants) and GGT (biallelic non-null variants vs. single-null variant) levels.
No linear correlations of liver enzyme levels and days after birth were observed in these patients.
A total of 168 infants, from 1–3 months of age, were enrolled in the study, and 35 were diagnosed with DJS and 133 with BA.
Acknowledgments
We would like to thank Xiao-Jun Zhu from Université Laval for his great help in the analysis of genetic information.
Abbreviations
- AUC
area under the curve
- ABCC2
ATP-binding cassette subfamily C member 2
- ALT
alanine aminotransferase
- AST
aspartate aminotransferase
- ALP
alkaline phosphatase
- BA
biliary atresia
- DB
direct bilirubin
- DJS
Dubin–Johnson syndrome
- nDJS
neonatal Dubin-Johnson syndrome
- GGT
gamma-glutamyl transferase
- INR
international normalized ratio
- MRP2
multidrug resistance-associated protein 2
- ROC
receiver operating characteristic
- TB
total bilirubin
- TBA
total bile acids
Data sharing statement
The detailed data used to support the findings of this study are available from the corresponding author at liuteng@fudan.edu.cn upon request.
References
- 1.Dubin IN, Johnson FB. Chronic idiopathic jaundice with unidentified pigment in liver cells; a new clinicopathologic entity with a report of 12 cases. Medicine. 1954;33(3):155–197. doi: 10.1097/00005792-195409000-00001. [DOI] [PubMed] [Google Scholar]
- 2.Paulusma CC, Kool M, Bosma PJ, Scheffer GL, ter Borg F, Scheper RJ, et al. A mutation in the human canalicular multispecific organic anion transporter gene causes the Dubin-Johnson syndrome. Hepatology. 1997;25(6):1539–1542. doi: 10.1002/hep.510250635. [DOI] [PubMed] [Google Scholar]
- 3.Wada M, Toh S, Taniguchi K, Nakamura T, Uchiumi T, Kohno K, et al. Mutations in the canilicular multispecific organic anion transporter (cMOAT) gene, a novel ABC transporter, in patients with hyperbilirubinemia II/Dubin-Johnson syndrome. Hum Mol Genet. 1998;7(2):203–207. doi: 10.1093/hmg/7.2.203. [DOI] [PubMed] [Google Scholar]
- 4.Sprinz H, Nelson RS. Persistent non-hemolytic hyperbilirubinemia associated with lipochrome-like pigment in liver cells: report of four cases. Ann Intern Med. 1954;41(5):952–962. doi: 10.7326/0003-4819-41-5-952. [DOI] [PubMed] [Google Scholar]
- 5.Talaga ZJ, Vaidya PN. Dubin Johnson Syndrome. Treasure Island (FL): StatPearls; 2020. [PubMed] [Google Scholar]
- 6.Kondo T, Yagi R, Kuchiba K. Letter: Dubin-Johnson syndrome in a neonate. N Engl J Med. 1975;292(19):1028–1029. doi: 10.1056/nejm197505082921913. [DOI] [PubMed] [Google Scholar]
- 7.Nakata F, Oyanagi K, Fujiwara M, Sogawa H, Minami R, Horino K, et al. Dubin-Johnson syndrome in a neonate. European journal of pediatrics. 1979;132(4):299–301. doi: 10.1007/BF00496853. [DOI] [PubMed] [Google Scholar]
- 8.Lee JH, Chen HL, Chen HL, Ni YH, Hsu HY, Chang MH. Neonatal Dubin-Johnson syndrome: long-term follow-up and MRP2 mutations study. Pediatr Res. 2006;59(4 Pt 1):584–589. doi: 10.1203/01.pdr.0000203093.10908.bb. [DOI] [PubMed] [Google Scholar]
- 9.Lakshminarayanan B, Davenport M. Biliary atresia: A comprehensive review. J Autoimmun. 2016;73:1–9. doi: 10.1016/j.jaut.2016.06.005. [DOI] [PubMed] [Google Scholar]
- 10.Junge N, Goldschmidt I, Wiegandt J, Leiskau C, Mutschler F, Laue T, et al. Dubin-Johnson Syndrome as Differential Diagnosis for Neonatal Cholestasis. J Pediatr Gastroenterol Nutr. 2021;72(5):e105–e111. doi: 10.1097/MPG.0000000000003061. [DOI] [PubMed] [Google Scholar]
- 11.Togawa T, Mizuochi T, Sugiura T, Kusano H, Tanikawa K, Sasaki T, et al. Clinical, Pathologic, and Genetic Features of Neonatal Dubin-Johnson Syndrome: A Multicenter Study in Japan. J Pediatr. 2018;196:161–167.e161. doi: 10.1016/j.jpeds.2017.12.058. [DOI] [PubMed] [Google Scholar]
- 12.Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405–424. doi: 10.1038/gim.2015.30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Fu H, Zhao R, Jia X, Li X, Li G, Yin C. Neonatal Dubin-Johnson syndrome: biochemical parameters, characteristics, and genetic variants study. Pediatr Res. 2021 doi: 10.1038/s41390-021-01583-7. [DOI] [PubMed] [Google Scholar]
- 14.Kim KY, Kim TH, Seong MW, Park SS, Moon JS, Ko JS. Mutation spectrum and biochemical features in infants with neonatal Dubin-Johnson syndrome. BMC Pediatr. 2020;20(1):369. doi: 10.1186/s12887-020-02260-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Corpechot C, Barbu V, Chazouilleres O, Broue P, Girard M, Roquelaure B, et al. Genetic contribution of ABCC2 to Dubin-Johnson syndrome and inherited cholestatic disorders. Liver Int. 2020;40(1):163–174. doi: 10.1111/liv.14260. [DOI] [PubMed] [Google Scholar]
- 16.Kimura A, Ushijima K, Kage M, Mahara R, Tohma M, Inokuchi T, et al. Neonatal Dubin-Johnson syndrome with severe cholestasis: effective phenobarbital therapy. Acta Paediatr Scand. 1991;80(3):381–385. doi: 10.1111/j.1651-2227.1991.tb11867.x. [DOI] [PubMed] [Google Scholar]
- 17.Serinet MO, Wildhaber BE, Broue P, Lachaux A, Sarles J, Jacquemin E, et al. Impact of age at Kasai operation on its results in late childhood and adolescence: a rational basis for biliary atresia screening. Pediatrics. 2009;123(5):1280–1286. doi: 10.1542/peds.2008-1949. [DOI] [PubMed] [Google Scholar]
- 18.Fawaz R, Baumann U, Ekong U, Fischler B, Hadzic N, Mack CL, et al. Guideline for the Evaluation of Cholestatic Jaundice in Infants: Joint Recommendations of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition and the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition. J Pediatr Gastroenterol Nutr. 2017;64(1):154–168. doi: 10.1097/MPG.0000000000001334. [DOI] [PubMed] [Google Scholar]
- 19.Kwo PY, Cohen SM, Lim JK. ACG Clinical Guideline: Evaluation of Abnormal Liver Chemistries. Am J Gastroenterol. 2017;112(1):18–35. doi: 10.1038/ajg.2016.517. [DOI] [PubMed] [Google Scholar]
- 20.Moseley RH. Evaluation of abnormal liver function tests. Med Clin North Am. 1996;80(5):887–906. doi: 10.1016/s0025-7125(05)70472-7. [DOI] [PubMed] [Google Scholar]
- 21.Koskelo P, Toivonen I, Adlercreutz H. Urinary coproporphyrin isomer distribution in the Dubin-Johnson syndrome. Clin Chem. 1967;13(11):1006–1009. [PubMed] [Google Scholar]
- 22.Okada H, Kusaka T, Fuke N, Kunikata J, Kondo S, Iwase T, et al. Neonatal Dubin-Johnson syndrome: novel compound heterozygous mutation in the ABCC2 gene. Pediatr Int. 2014;56(5):e62–64. doi: 10.1111/ped.12404. [DOI] [PubMed] [Google Scholar]
- 23.Shieh CC, Chang MH, Chen CL. Dubin-Johnson syndrome presenting with neonatal cholestasis. Arch Dis Child. 1990;65(8):898–899. doi: 10.1136/adc.65.8.898. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
No significant correlations were observed with different genotype groups except for TBA (biallelic non-null variants vs. biallelic-null variants) and GGT (biallelic non-null variants vs. single-null variant) levels.
No linear correlations of liver enzyme levels and days after birth were observed in these patients.
A total of 168 infants, from 1–3 months of age, were enrolled in the study, and 35 were diagnosed with DJS and 133 with BA.