Progressive familial cholestasis type 2 is caused by a genetically determined absence or reduction in the activity of the bile salt export pump (BSEP). Reduction or absence of BSEP activity causes a failure of bile salt excretion, leading to accumulation of bile salts in hepatocytes and subsequent hepatic damage. Clinically, patients are jaundiced, suffer from severe intractable pruritus, and evidence progressive liver dysfunction. A low level of serum γ-glutamyl transpeptidase, when associated with the described signs and symptoms, is often an early identifier of this condition. Treatment options to date include liver transplantation and the use of biliary diversion. We report a multidrug regimen of 4-phenylbutyrate, oxcarbazepine, and maralixibat (an experimental drug owned by Shire Pharmaceuticals, Dublin, Republic of Ireland) that completely controlled symptoms in 2 siblings with partial loss of BSEP activity.
Patients A and B are male and female siblings respectively affected with progressive familial intrahepatic cholestasis type 2 (PFIC2).1 Patient A is the older of the children, currently 12 years of age. He presented in the neonatal period with failure to resolve presumed physiologic jaundice. At 3 months of age he developed a bleeding diathesis. Diagnostic testing revealed a decrease in hepatic-based procoagulants, elevated serum aspartate aminotransferase at 10 times the upper limit of normal (normal [range of values] 15–30 U/L) and alanine aminotransferase at 10 times (nl 20–37 U/L) the upper limit of normal, and an elevated conjugated bilirubin of 4.1 mg/dL. γ-glutamyl transpeptidase was in the low-normal range, raising the suspicion of PFIC2. Liver biopsy performed at 6 months of age revealed cholestasis with giant cell transformation. A portion of the biopsy specimen was sent to King’s College London, London, United Kingdom, for bile salt export pump (BSEP) staining. The results of the BSEP staining revealed an absolute reduction compared with normal controls but not complete absence of staining (A. Knisely, personal communication, 2016). By 1 year of age, prothrombin time had normalized without vitamin K supplementation and aspartate aminotransferase and alanine aminotransferase settled in at 3 times the upper limit of normal, whereas the γ-glutamyl transpeptidase fell to less-than-normal range. Total serum bile acid determinations were variable but always elevated and reached as high as 20 times the upper limit of normal (nl 0–19 μmol/L), but were more often 6 to 12 times the upper limit of normal. Patient A, who was irritable throughout infancy, began scratching himself by a year of age. Patient A’s subsequent course became dominated by intractable severe pruritus. Patient A’s scratching was assessed to be in the 4 range but was never <3 on the Whittington severity scale, with 1 being the least severe and 4 the most severe (Fig 1A).2 This scratching led to multiple and frequent impetiginized excoriations and episodes of cellulitis requiring antimicrobial therapy. Attempts to reduce the itching with topical and systemic antipruritic therapy including hydroxyzine, diphenhydramine, ondansetron, naltrexone, and topical corticosteroids were met with limited success. Cholestasis-specific therapy, including cholestyramine, phenobarbital, ursodiol, and rifampin also had little efficacy. A trial of peripheral nerve-stabilizing agents was also unsuccessful, although oxcarbazepine, when used at antiseizure dosing serum levels, appeared to help patient A sleep, perhaps if only by a sedative effect.
DNA testing on patient A at 3 years of age, performed at the Baylor College of Medicine Medical Genetic Laboratory, Houston, Texas, revealed 2 novel (at the time of the evaluation) heterozygous missense variants in ABCB11: c.470A>G (p.Y157C) and c.3892G>A (p.G1298R). ABCB11 is the gene loci associated with PFIC2 disease. Once the gene defect was identified, each parent was subsequently shown to carry 1 of the 2 missense mutations as part of a Triome evaluation. Patient B was born 3.5 years after her brother, patient A, and her early course, including cholestasis and bleeding diathesis, duplicated her brother’s problems in infancy but with somewhat less severity. Two months after her birth, genetic testing at Baylor Laboratory confirmed that patient B had the same 2 heterozygous missense mutations that are present in her older brother. The BSEP expression studies performed on patient A’s liver biopsy specimen were not performed in patient B because she had an identical genotype and similar phenotype to her older brother.
Attempts to manage pruritus in both children continued with limited efficacy. Serum bile acids were always elevated, and the children were always beset by unremitting pruritus, although their bile acid levels did not correlate with the degree of pruritus. In addition to the management of itch and occasional fat-soluble vitamin–deficiency support, regular abdominal ultrasounds and serum tumor markers (α-fetoprotein and Ca 19-9) were obtained every 4 months. These were obtained as surveillance for possible liver tumor development, a known complication of PFIC2.3,4 The serial ultrasounds in patient A revealed slowly increasing echogenicity, suggesting worsening hepatic fibrosis as well as cholelithiasis. Repeat liver biopsy on patient A at 6 years of age revealed significant fibrosis with some bridging as well as occasional binucleated hepatocytes. Serial ultrasound results in patient B remained normal. Tumor markers in both children remained normal. The children slept fitfully, such that the parents developed a strategy of sleeping apart with the children so they could scratch them back to sleep when they awoke. The children began to show reduced growth velocity, presumably from lack of an overnight growth hormone surge (a consequence of interrupted sleep), although somatotropin levels were never obtained.5 The children grew along the 10% until age 8 (patient A) and age 6 (patient B), then fell off the growth curve, reaching <3% (patient A) and at the 3% (patient B). The parents elected to list the children for liver transplantation as opposed to biliary diversion.6
On the basis of the reported effectiveness of 4-phenylbutyrate (4PB) as an enhancer of cell-surface expression of the transport capacity of BSEP and a clinical trial using 4PB in a patient suffering from PFIC2, we initiated the sodium preparation of 4PB at 500 mg/kg per day.7,8 The parents noted an improvement in their children’s itching after only 10 days of treatment with this new drug therapy. Despite the clinical improvement, no significant or persistent drop in total serum bile acids was noted. Bile acids remained in the 200 μmol/L range for both children.
Unfortunately, although the pruritus improved, it was not eliminated and continued to have a significant effect on the children’s lives. In 2014, an experimental drug, LUM001, developed by Lumena Laboratories (San Diego, CA), became available under study circumstances. The drug LUM001, now referred to as maralixibat or SHP625 (because its rights were obtained by Shire Pharmaceuticals), attaches to and blocks the apical sodium–dependent bile acid transporter receptor in the small bowel, which normally captures bile salts for enterohepatic recirculation. The children were entered into the maralixibat trial. Informed consent was provided in keeping with the drug study. Also, as required by the study, all predrug trial therapies were maintained for at least 6 months after initiation of the study drug.
Four weeks into the maralixibat trial, pruritus ceased, and the children were sleeping through the night. The skin excoriations cleared (Fig 1B), and the children’s concentration in school improved, according to their respective classroom teachers. The parents returned to a normal sleeping arrangement in which the children slept separately from the parents. Total serum bile acids dropped from over 200 μmol/L for both the older and younger child to 18 and 14 μmol/L, respectively. Two months into the trial drug, all other laboratory values had also normalized. One month later, 4PB became temporarily unavailable for 6 weeks. The children were now on just maralixibat and adjuvant antipruritic therapies, including oxcarbazepine, hydroxyzine, ursodiol, and ondansetron, and experienced a return of severe pruritus. Bile acids increased into the high-normal range as symptoms exacerbated. When 4PB became available, the symptoms again resolved. At 6 months into maralixibat treatment, we began removing the adjuvant anti-itch drugs serially. When oxcarbazepine was discontinued, symptoms recurred, and bile acids increased for both the older and younger child from normal to 32 and 34 μmol/L, respectively. Restarting a single, once-daily, small dose of oxcarbazepine was associated once again with resolution of symptoms. During the past two years the combination of 5 mg/kg per day as a single dose of oxcarbazepine at bedtime, 500 mg/kg per day divided into 4 doses of 4PB, and 1 mL of maralixibat in the morning before eating had shown efficacy; the children remained symptom free, liver enzymes were within normal limits, bile acids were 4 and 6 μmol/L for the older (patient A) and younger (patient B) child, respectively, and they had made catch-up growth, returning to their previous growth percentiles. Six months ago patient A developed recurrent urticaria. He was streptococcal throat culture negative had normal thyroid studies without elevated thyroid antibodies and did not appear to have any apparent triggers of the urticaia. Once again oxcarbmazapine was withdrawn though this time without recurrance of pruritis. We have subsequently discontinued the oxcarbmazapine on patient B without recurrance of pruritis.
PFIC2 is an autosomal recessive condition that occurs because of failure of normal BSEP function. BSEP is an adenosine triphosphate–binding cassette transmembrane transporter located on the canalicular membrane of the hepatocyte that transfers bile salts synthesized in the hepatocyte into the biliary canalicular lumen as part of bile secretion. Reduction or loss of BSEP causes an elevated total serum bile acids level, cholestasis (because of reduced osmotic activity of bile), and jaundice in the first year of life. The natural course of PFIC2 in patients with complete loss of BSEP function is progressive hepatic injury, eventual liver failure, as well as potential development of hepatocellular carcinoma.
PFIC2 has not had effective treatment to date. In our patients with this specific genotype, which results in reduced but not absent BSEP activity, it appears that for complete resolution of symptoms to occur, use of all 3 drugs (maralixibat, 4PB, and oxcarbazepine) in concert was necessary. Serendipitously, it became apparent that stopping any 1 of the 3 drugs in this cocktail resulted in less effective therapy. It is likely that the 4PB enhanced the BSEP expression, increasing bile salt secretion and reducing symptoms as a single agent as previously reported.8,9 Unfortunately, this was not adequate for complete symptom resolution. However, we believe that once there was a significant bile salt load in the bile being secreted because of 4PB’s effect in enhancing BSEP activity, subsequently blocking enterohepatic reuptake of bile salts with maralixibat caused depletion of the bile salt pool more quickly and more completely. When the 4PB was temporarily withheld, there was too little bile salt in the bile for maralixibat to be clinically helpful. As a potent, selective, minimally absorbed inhibitor of apical sodium–dependent bile acid transporter, maralixibat blocks bile acid reabsorption in the terminal ileum, increasing fecal bile acid excretion and thereby reducing bile acid recirculation to the liver.
Why did the discontinuation of oxcarbazepine initially result in loss of symptom control yet when discontinuation occurred over a year and a half later no such recudesence of symptoms occurred? Oxcarbazepine had been added because of its effect as a peripheral nerve stabilizer and hopefully an inhibitor of pruritus. We believe that it is the enzyme induction caused by oxcarbazepine that plays the critical role. The 4PB is actually a prodrug and requires metabolism to phenylacetate and then conjugation with glutamine to phenylacetylglutamine, the active form. Whereas the active form is the primary metabolite excreted, a significant percent (perhaps higher in those with impaired liver function) of 4PB is excreted in its prodrug (inactive) form. We believe that oxcarbazepine, by enzyme induction (it is an inducer of the CYP34A enzymes complex, uridine diphosphate glucuronosyltransferases,10 and a subset of the P450 system, the P450 IIIA11), increases the percent of drug in its active or effective form, leading to more profound enhancement of BSEP expression and resolution of symptoms, which had been lost when the oxcarbazepine was withdrawn. Assuming it is this indirect effect of oxcarbazepine that is critical to symptom control, we restarted the drug as a single dose (one-third of the previous dosing), not attempting to maintain anticonvulsant levels (8–35 μg/mL) as we had when we were administering the drug before 4PB and maralixibat were initiated. Oxcarbazepine drug levels of only 2 to 3 μg/mL on this single daily dose were obtained. This low, subtherapeutic (with regard to anticonvulsant effect) dose was adequate to recapture the symptom-free state. Months later in response to a new problem, chronic urticaria, we again trialed patient A off oxcarbazapine and this time symptoms did not return. We believe this happen because we had at this point reduced to bile salt pool to the point that less effectiveness of the 4PB was not detrimental and the unenhance 4PB therapy was adequate for bile salt homeostasis. We do recognize that we might have alternatively provided more 4PB and that might have also led to more drug activity. However, 4PB side effects, such as gastrointestinal distress, are more likely to occur at larger doses. Because the 4PB dose was already large and difficult to tolerate, we opted to restart oxcarbazepine first.
Although control of pruritus and stabilization of liver disease progression was our first priority, we saw the additional benefits of complete symptom control. Growth velocity has shown a clear increase since the drug regimen was begun (from 2.4 cm/year pretreatment to 10.2 cm/year after treatment for patient A, and from 1.8 cm/year pretreatment to 10.8 cm/year after treatment for patient B). We believe this is due to normalization of sleep and its attendant nighttime somatotropin surge6 (Fig 2). The children’s short stature would not have been expected given their midparental height (the father is 6′5″) and was not explained by constitutional delay (bone ages were normal for both children). The improvement in their growth velocity correlated precisely with control of their chronic disease and its associated symptoms. The children continue to cross growth percentiles upward on their current treatment regimen. Although we did not measure somatotropin surges in our patients, it is hoped that others who employ this regimen in patients with ABCB11 disease and growth failure might assess this possible hormone effect prospectively. There have been no additional skin infections or requisite treatment with antimicrobial agents. The children have had an improvement in school performance (as determined by teachers’ comments). Although the children are remarkably well adjusted, a tribute to extraordinary parental care, the older child has anxiety at times and was old enough to have become a peculiarity to his schoolmates before symptom control. These issues also are improving. A follow-up liver biopsy on the older child, obtained over a year after treatment began, revealed a less altered histology with elimination of binucleated hepatocytes and less overall fibrosis.
PFIC2, when associated with reduced but not absent BSEP expression, might be managed with symptom reduction or elimination (or both) by using a 3-drug cocktail of maralixibat, 4PB, and oxcarbazepine. This drug combination warrants a larger trial in which each ABCB11 mutation associated with incomplete loss of BSEP function is recognized and cataloged and the response to this therapy is evaluated.11,–13 As a Case Report, it is possible that this genotype’s natural history is related to symptom resolution over time, although the dramatic response correlated in time to the treatment of the children (despite their significant differences in age and sex), and the associated recapture of the symptom-free state occurred when the 3-drug regimen was restarted. Caution in interpreting an observation in a Case Report is always prudent. Continued vigilance for tumor occurrence should be maintained, even in those with apparent resolution of symptoms and amelioration of disease.
Dr Malatack conceived of the idea of care for the children and wrote the paper; and Dr Doyle assessed the children’s growth reduction and wrote the information related to the children’s growth response.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: Dr Malatack serves as Case Reports editor for Pediatrics but did not participate in the peer review process of this Case Report as a result of being an author; and Dr Doyle has indicated he has no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURE: The content of this article is based solely on the observations of Drs Malatack and Doyle as the treating physicians for 2 patients with PFIC2. The investigational compound developed by Lumena Laboratories as LUM001 (name recently changed to maralixibat, or SHP625, by current product owner Shire) under current product rights is currently undergoing phase 2 trials, it has not been approved by regulatory authorities for use, and it is premature to evaluate safety or efficacy for use. Neither Shire nor Lumena, an indirect, wholly-owned subsidiary of Shire North American Group, Inc, had any role in or made any financial contribution to the authorship of this article, and they are not responsible for any of the content herein.