Next‐generation Sequencing Facilitates the Diagnosis in a Child With Twinkle Mutations Causing Cholestatic Liver Failure

Abstract

In many disorders, overlapping clinical phenotypes and locus heterogeneity can significantly hamper making a clinical or molecular diagnosis. In both acute and cholestatic liver failure, making such a diagnosis has significant potential to alter the direction of therapy. In several large series, the majority of children with acute liver failure have no identifiable cause. It is the hope that new technologies may allow for rapid detection of all of the genetic causes of liver failure (1). Mitochondria are essential for the survival of eukaryotic cells. Their major function is to generate adenosine triphosphate, the energy for supporting cellular activities. Mitochondria contain their own DNA and machinery for transcription and translation. The replication, repair, transcription, and translation of mitochondrial DNA (mtDNA) depend on the protein components encoded by the nuclear genome (2,3). The C10orf2 gene encodes the mtDNA helicase TWINKLE, which is one of the proteins essential for mtDNA maintenance. Dominant mutations cause multiple mtDNA deletions and progressive external ophthalmoplegia (Online Mendelian Inheritance in Man no. 609286), and recessive mutations can lead to mtDNA depletion and infantile onset spinocerebellar ataxia (IOSCA) (Online Mendelian Inheritance in Man no. 271245). The clinical features of IOSCA include hypotonia, athetosis, ataxia, ophthalmoplegia, hearing deficit, sensory neuropathy, and epileptic encephalopathy. Previous cases have described patients with either homozygous (Y508C) or compound heterozygous (Y508C and A318T) TWINKLE mutations. A 20-year follow-up on 23 patients with IOSCA demonstrated that the common symptoms include refractory status epilepticus, migraine-like headaches, and severe psychiatric symptoms (4). We report on a child with compound heterozygous recessive TWINKLE mutations who presented at an early age with acute liver failure, subsequent neurologic decompensation, and Fanconi syndrome. CLINICAL PRESENTATION A full-term infant was born at 38 weeks’ gestation following an uncomplicated pregnancy. Her birth weight was 2.8 kg, which was at the 12th percentile. At 4 days of age, this formula-fed infant was transferred from a newborn nursery to the neonatal intensive care unit at another institution. At presentation, she was lethargic, hypotonic, hypoglycemic, and hypothermic. Initial laboratory studies showed a white blood cell count of 7.9 cells/mm3, hemoglobin 9.5 g/dL, lactic acid 24 mg/dL, aspartate aminotransferase 116 IU/L, alanine aminotransferase 16 IU/L, total bilirubin 6.9 mg/dL, conjugated bilirubin 0.8 mg/dL, glucose level 7 mg/dL, and elevated arterial ammonia 170 μmol/L. Her hypoglycemia responded to intravenous dextrose infusion. She was also started on sodium phenylacetate/sodium benzoate and arginine, which resulted in a rapid decrease in the ammonia level. The remainder of the workup, including a sepsis workup, plasma amino acids, urine organic acids including urine ketone profile, newborn screen, and brain MRI were normal. Family history was reported by the mother to be negative with 5 healthy siblings and she passed her newborn hearing screen. After 1 week of hospitalization, tolerating regular cow's-milk infant formula orally and maintaining blood glucose level, she was discharged. Because of the resolution of symptoms and normal metabolic workup, the discharge diagnosis was transient hyperammonemia of the newborn. At her 2-month routine visit at her primary care physician's office, she was thriving but had been switched to a soy-based formula because of spitting up. At 10 weeks of age, she presented to the emergency department with increased work of breathing and jaundice. Physical examination was significant for dyspnea and hepatomegaly without ascities. Initial laboratory studies revealed a hemoglobin 7.8 g/dL, platelet count 527 cells/mm3, total bilirubin 10.9 mg/dL, conjugated bilirubin 7.6 mg/dL, aspartate aminotransferase 1033 IU/L, alanine aminotransferase 634 IU/L, γ-glutamyl transpeptidase 77 IU/L, albumin 2.5 g/dL, lactate dehydrogenase 2633 IU/L, alkaline phosphatase 1084 IU/L, ammonia 29 μmol/L, lactic acid 31 mg/dL, international normalized ratio 15.22, prothrombin time 60 seconds, fibrinogen <60 mg/dL, D-dimer 10.12 μg/dL, ferritin 973 ng/mL, serum glucose 67 mg/dL, and α-fetoprotein 84,800 ng/mL. Urine analyses showed trace ketones and trace protein. She had undetectable acetaminophen and salicylic acid levels. She was admitted to the pediatric intensive care unit for further management of her acute liver failure. She was intubated for worsening respiratory distress, stabilized, and started on N′-acetylcysteine for her acute liver failure (5,6). She had negative testing for cytomegalovirus, toxoplasmosis, rubella, herpesvirus, and syphilis. She had normal α-1-antitrypisin phenotyping and a normal lip biopsy, which ruled out hemochromatosis. She had normal extended toxicology screens, urine organic acids, plasma amino acids, and fractionated bile acid testing; however, after extubation and weaning of her sedatives, she was noted to have a dysconjugate gaze. Seizure activity was also suspected. Electroencephalography and a repeat brain magnetic resonance imaging with magnetic resonance spectroscopy were normal, but she failed visual and auditory evoked potential testing. She was only able to tolerate 0% to 40% of her enteral feedings because of emesis and abdominal distention; consequently, she was supplemented with parenteral nutrition without intralipids. She had persistent anemia, thrombocytopenia, coagulopathy, hypoalbuminemia, cholestasis, and elevation of serum transaminase levels. She was also found to have persistent hypokalemia and hypophosphatemia requiring high doses of parenteral supplementation. Her renal ultrasound was normal, but urinary studies indicated that she had significant urinary phosphate wasting. She had a normal anion gap acidosis, which was consistent with Fanconi syndrome. Liver wedge biopsy performed at 12 weeks of age showed active cirrhosis with bile ductular proliferation, marked cholestasis, diffuse hemosiderosis, and patchy macrovesicular steatosis (Fig. 1). mtDNA content in the liver was found to be 17% of the mean value of the age- and tissue-matched controls, but no mutations in POLG, MPV17, or DGUOK were detected.FIGURE 1: Liver wedge biopsy performed at 12 weeks of age with (A) hematoxylin and eosin stain showing marked cholestasis and hemosiderosis, ballooning hepatocyte degeneration, bile ductular proliferation, and mild portal inflammation, (B) Gomori trichrome stain highlighting (in blue) bridging fibrosis delineating cirrhotic nodules, and (C) anticytokeratin-19 immunohistochemical stain showing (in brown) bile ductular proliferation.Concurrent with her clinical evaluation, because of the mtDNA depletion, consent was obtained, in line with the Declaration of Helsinki, and she was enrolled in an institutional review board–approved research study evaluating the utility of genomic sequencing in diagnosing mitochondrial disease. DNA was extracted from peripheral leukocytes and sample preparation followed the protocol outlined in Figure 2. The initial step in sequencing was enrichment of the target sequence. We achieved this by long-range PCR of target genes (including DGUOK, MPV17, C10orf2 (TWINKLE), POLG, POLG2, TYMP, TMEM70, and SUCLA2) using site-specific primers amplifying 7- to 15-kb products with the Roche Expand Long Range dNTPack (Roche Applied Science, Indianapolis, IN; primers and polymerase chain reaction conditions available upon request). After shearing, samples were tagged using a slightly modified version of the protocol published by Meyer et al (7). The amplified DNA was then sequenced on the Roche GS-FLX system (Roche Applied Science) using the Titanium upgrade reagents. Variants were then annotated using in-house software as previously described (8) and analyzed following the American College of Medical Genetics’ recommendations on standards for interpretation of sequence variations (9). Only 2 significant variants were detected, both in C10Orf2, a novel truncating mutation c.85C>T (p.R29X) and the previously reported c.1523A>G (p.Y508C); these were confirmed by clinical Sanger sequencing at an outside laboratory.FIGURE 2: Schematic representation of next-generation sequencing process used in this case.Given the abnormal neurologic examination and sequence-based confirmation of a primary mtDNA depletion disorder and previous published experience, a decision was made that she would not be an appropriate candidate for liver transplantation (10–13). She died at 6 months of age from sepsis and multiorgan failure. Autopsy revealed no histopathologic changes in the brain, spinal cord, skeletal muscle, peripheral nerves, or heart (data not shown). The most striking abnormality was marked biliary cirrhosis. The histologic appearance of the liver at autopsy, as compared with the antemortem biopsy, showed marked progression during the 10-week interval with increased fibrosis, increased neocholangiogenesis, and a greatly reduced volume of viable hepatic parenchyma (Fig. 3). Jaundice, splenomegaly, and ascites were also seen; peritoneal fluid cultures were positive for Klebsiella pneumoniae.FIGURE 3: At autopsy, the liver (A and B) was grossly shrunken, cholestatic, and cirrhotic with nodular parenchyma. By light microscopy with hematoxylin and eosin staining (C), there was disease progression compared with the previous wedge biopsy, with increased fibrosis, cholestasis, and decreased volume of viable hepatocytes. Gomori trichrome stain (D) highlights fibrous bands (in blue) separating small islands of residual hepatocytes.DISCUSSION The mtDNA depletion syndromes are a clinically heterogeneous group of disorders caused by molecular defects in the nuclear genes involved in the mtDNA biosynthesis and the maintenance of the deoxynucleotide pools. mtDNA depletion syndromes cause a reduction in cellular mtDNA content. At the present time, mutations in at least 9 genes have been found to cause mtDNA depletion. In liver failure, the treatment of choice is liver transplantation (14); however, in at least 1 form of mtDNA depletion, in the presence of neurologic disease, transplantation has been shown to be futile (10), whereas in related disorders, case reports have alluded to similar poor long-term outcomes (11–13). Conversely, when considering liver disorders as a whole, specific disorders may respond to specific treatment while not being diagnosed on analytic testing (15,16). In either situation, a molecular diagnosis may prevent a child from undergoing an inappropriate liver transplant. Although present Sanger sequencing for the depletion genes alone costs in excess of $10,000 because liver transplantation costs on average $500,000 (16), ordering this testing would be cost-effective if it only establishes a cause in 1 in 50 cases. With whole genome sequencing now clinically available for <$20,000 (17), there is a more cost-effective method to screen for all of these disorders and many others in 1 test. It is clear that mitochondrial disorders play a role in cholestasis in children who may eventually progress to needing a transplant (13). Using whole genome sequencing to recognize these children and other rare treatable causes of liver disease likely reduces transplantation rates further in the cholestatic population (15,16) as well as informing health surveillance for extra hepatic manifestations such as vascular malformations in Alagille syndrome and hearing loss in PFIC1. To our knowledge, this is the first report of a recessive TWINKLE mutation presenting with fulminant acute liver failure. This patient initially presented in the newborn period with a brief episode of hyperammonemia, with resolution. Indeed, she had no other symptoms in the 2 months of life before her second presentation. She had normal growth and development documented at her routine well-child visit at her primary care physician's office, similar to several of the reported cases of DGUOK deficiency (13). Previously reported cases of recessive TWINKLE mutations presented at an older age with a gradual onset of neurologic symptoms, which progressed to intractable epilepsy. The initial symptoms of ataxia, muscle hypotonia, athetoid movements, and loss of deep tendon reflexes were usually seen around 1 year in homozygotes and at 6 months in the compound heterozygotes (4). Liver involvement has been reported in 2 siblings with recessive mutations who presented at earlier ages of 5 months and 6 months, respectively, with initial neurologic symptoms. Both cases had mild-to-moderate elevation of serum transaminase levels with normal bilirubin levels. Of note, both siblings were receiving anticonvulsive therapy (18). Renal involvement was not documented in any other cases. In conclusion, our report adds a new spectrum of symptoms (acute liver failure, acute onset developmental regression, and Fanconi syndrome) associated with recessive TWINKLE mutations. It demonstrates the utility of rapid molecular diagnosis in informing transplantation decision making in acute liver failure and suggests that larger molecular screens for molecular etiology may be cost-effective in pediatric liver disease.