Rituximab as therapy for the recurrence of bile salt export pump deficiency after liver transplantation
Henry C. Lin,
Departments of Pediatrics, Northwestern University Feinberg School of Medicine, Chicago, IL
Address reprint requests to Henry C. Lin, M.D., Division of Gastroenterology, Hepatology, and Nutrition, Children's Hospital of Philadelphia, 34th Street and Civic Center Boulevard, Philadelphia, PA 19103. Telephone: 267-426-4957; FAX: 215-590-5326; E-mail: email@example.com
Progressive familial intrahepatic cholestasis type 2 (PFIC2) results from recessive mutations in the adenosine triphosphate–binding cassette B11 (ABCB11) gene, which encodes for bile salt export pump (BSEP). Children with PFIC2 develop progressive cholestasis in association with normal serum gamma-glutamyltransferase (GGT) activity (ie, low-GGT PFIC). The liver disease is often refractory to medical and surgical management and progresses to its end stage often within the first decade of life. In addition, PFIC2 patients are at high risk for hepatobiliary malignancies.[1-3] Liver transplantation (LT) is offered to PFIC2 patients with end-stage liver disease or evidence of malignancy.
BSEP protein is expressed on the hepatocyte canalicular membrane and provides the main export pathway for primary bile acids. A paucity of functional BSEP leading to deficient bile salt–dependent bile formation is thought to be the primary mechanism of cholestasis development in PFIC2. LT is considered to be curative for PFIC2 because BSEP has no known function outside the liver.[3-7] However, recent reports have described recurrent cholestasis in PFIC2 patients after LT.[8-10] These cases have been associated with immunoglobulin G (IgG) antibodies to BSEP. High-titer anti-BSEP antibodies appear to correlate with episodes of cholestatic graft dysfunction. IgG gains access to bile presumably through a paracellular pathway in which molecular sieving determines its concentration.[11, 12] The antibodies bind to an epitope in the intracanalicular BSEP domain, which appears to have a critical function in bile salt transport. The effect is a blockade of the bile salt transport function of the protein, which induces cholestasis by the same mechanism as BSEP deficiency. Posttransplant recurrent BSEP disease has proven to be as recalcitrant to therapy as the original disease and may lead to rapid graft failure.
There is no established paradigm for treating antibody-mediated posttransplant BSEP disease. It appears to be refractory to changes in immunosuppressant medications that would typically be effective in treating allograft rejection.[8-10, 13] Taking what is known about its pathophysiology, we designed a treatment consisting of plasmapheresis for the removal of existing BSEP antibodies followed by high-dose intravenous immunoglobulin (IVIG) and then rituximab, a chimeric monoclonal anti-CD20 antibody, to prevent the synthesis of new anti-BSEP antibodies. Using this approach, we report the successful management of 2 patients with antibody-mediated recurrence of PFIC2 after LT.
PATIENTS AND METHODS
Detection and Quantitation of BSEP Antibodies
Indirect immunofluorescence staining was performed on rat liver sections essentially as previously described. Posttransplant serum samples from the 2 patients were screened at dilutions ranging from 1:20 to 1:5120 with commercially prepared sections from rat livers. Bound antibodies were detected with the use of fluorescein isothiocyanoate–conjugated rabbit anti-human IgA, IgG, and IgM antibodies (DakoCytomation, Glostrup, Denmark). Serum samples from 2 PFIC2 children with and without anti-BSEP antibodies after LT were tested in parallel as positive and negative controls, respectively.
Membrane lysates (30 μg of protein) from Sf9 cells expressing human BSEP (Sigma-Aldrich, Inc., Saint Louis, MO) or human multidrug resistance protein 1 (BD Biosciences, Woburn, MA) were fractionated by 6% sodium dodecyl sulfate–polyacrylamide gel electrophoresis, transferred onto polyvinylidene fluoride membranes, and cut into strips. Strips were probed with sera from the patients (1:100 dilution) or with a goat polyclonal anti-BSEP antibody (Santa Cruz Biotechnology, Santa Cruz, CA; 1:200 dilution). Immune complexes were detected with horseradish peroxidase–conjugated goat anti-human IgG (Nordic Immunological Laboratories, Tilburg, NE; 1:16,000 dilution) or rabbit anti-goat Ig (DakoCytomation, Glostrup, Denmark; 1:6000 dilution). The signal was developed with an ECL Advanced western blotting detection kit (GE Healthcare, Buckinghamshire, United Kingdom). These study protocols were approved by the appropriate institutional review committee.
An infant boy presented with a clinical and biochemical picture that was characterized by an elevated serum bilirubin level and persistently normal GGT levels and was suspicious for an intrahepatic cholestatic disease. Fast atom bombardment mass spectrometry of urine excluded an inborn error of bile acid synthesis. Genetic testing (Hereditary Liver Disease Genetic Testing, Children's Hospital Medical Center, Cincinnati OH) revealed a homozygous frameshift mutation in the ABCB11 gene (c.2787_2788insGAGAT) that resulted in p.Lys930GlufsX79 and predicted an absence of BSEP protein. The patient developed end-stage liver disease and underwent LT at 10 months of age with a deceased donor left lateral segment. A histological examination of the explanted liver revealed micronodular cirrhosis with hepatocanalicular cholestasis, giant cell transformation, and interstitial fibrosis in the residual parenchyma and regenerative nodules (Fig. 1A). Immunohistochemistry with BSEP antibodies (catalogue number HPA109035, Sigma, St. Louis, MO) was abnormal with no canalicular or hepatocyte staining, and this indicated an absence of BSEP protein (Fig. 1B); however, immunohistochemistry with canalicular multispecific organic anion transporter antibodies demonstrated a pattern of staining consistent with intact canaliculi replete with the membrane protein (Fig. 1C).
Transplant immunosuppression was induced with prednisone and tacrolimus. His posttransplant course was complicated by biopsy-proven acute rejection (endothelialitis and lymphocytic portal inflammatory infiltrates but no duct involvement; Banff score = 5), which occurred 15 months after transplantation and transiently responded to treatment with additional corticosteroids. At the time of the acute rejection, his GGT level was 9 IU/L. Thereafter, his serum bilirubin and aminotransferase levels fluctuated, but his GGT levels remained normal. A liver biopsy sample taken 20 months after transplantation showed features of PFIC2, including severe canalicular cholestasis, hepatocyte cholate injury, Mallory's hyaline, multinucleated giant cell transformation, scattered hepatocyte necrosis, and some loss/atrophy of interlobular bile ducts (Fig. 1D). The patient was medically managed with pulse steroids and a change in the primary immunosuppressant to sirolimus, but his graft function progressively deteriorated. Repeat liver biopsies showed worsening bile duct loss and cholestasis. His GGT levels remained normal, and this suggested recurrence of the primary disease as the cause of his progressive cholestatic injury. His IgG levels were elevated at 1680 mg/dL and were suggestive of a B cell–mediated process.
He received a second transplant at 3.5 years of age with a left lateral segment from a deceased donor. Immunosuppression was induced with solumedrol, mycophenolate mofetil (MMF), and cyclosporine. Five years after the second transplant, graft dysfunction recurred (aspartate aminotransferase level = 2837 IU/L, alanine aminotransferase level = 1622 IU/L, total bilirubin level = 20.2 mg/dL, direct bilirubin level = 12.9 mg/dL, GGT level = 35 IU/L, and international normalized ratio = 1.5). A liver biopsy sample showed hepatocanalicular cholestasis, giant cell transformation, and mild interstitial and pericentral fibrosis, but there was no evidence of allograft rejection (Fig. 1E). The clinical course of low-GGT cholestasis was highly suggestive of antibody-mediated recurrent PFIC2. An indirect immunofluorescence assay with the patient's serum showed anti-canalicular antibodies at a 1:2560 titer (Fig. 2A), and a western blot analysis showed that the patient's serum contained anti-human BSEP IgG antibodies (Fig. 2B). These findings were considered diagnostic for antibody-mediated recurrent BSEP disease.
Treatment for recurrent PFIC2 was initiated with plasmapheresis and 3 daily doses (1 g/kg/dose) of IVIG, which were followed by monthly IVIG infusions (1 g/kg/dose; see Table 1 for the course of his disease with treatment). Symptoms of cholestasis and pruritus continued, and his total serum bile acid levels were elevated at 296 μmol/L (normal level < 20 μmol/L). A partial cutaneous biliary diversion was created as follows: halfway along the biliary limb of the Roux-en-Y, roughly 20 cm from the bilioenteric anastomosis, the intestine was brought out through the skin as a loop jejunostomy with the expectation that it would divert approximately half the bile. This diversion procedure had no effect on the clinical status. The serum bile acid level remained elevated at 160.4 μmol/L, and there was constant high-volume, intermittently bloody drainage requiring a takedown of the loop ileostomy. Because of the pathogenesis of the recurrent PFIC2 phenotype and because his disease remained refractory to treatment, rituximab therapy was initiated 6 months after the initial IVIG treatment. His rituximab regimen consisted of weekly 375 mg/m2 doses for 6 weeks and then doses every 6 weeks. In all, the patient received rituximab for 6 months and IVIG for 12 months. His serum anti-BSEP antibody levels decreased to 1:320, and his cholestasis was resolved by the end of this course of therapy. The patient remained well with no recurrence of cholestasis for 3 years after his rituximab treatment. His most recent anti-BSEP antibody titer was 1:80, and his total serum bile acid level was 7.9 μmol/L.
Table 1. Serum BSEP Antibody Titers and Measures of Allograft Function and Injury Before, During, and After Treatment
BSEP Antibody Titer
Total Serum Bile Acid (μmol/L)
Total Bilirubin (mg/dL)
Alanine Aminotransferase (IU/L)
Treatment was initiated with plasmapheresis (×3), which was followed by monthly IVIG. Partial external biliary diversion was performed 1 month later.
A Pakistani boy born to first cousins presented at 20 months of age with vitamin K and D deficiencies. A subsequent workup revealed an elevation of serum aminotransferases (maximum aspartate aminotransferase level = 875 U/L, alanine aminotransferase level = 612 U/L) and elevated levels of total bilirubin (2.8 mg/dL) and direct bilirubin (1.4 mg/dL), with a normal GGT (9 IU/L). A liver biopsy sample taken at 28 months of age showed a paucity of interlobular bile ducts, cholestasis, and mild interstitial, pericentral, and periportal fibrosis suggestive of PFIC. Genetic testing (Hereditary Liver Disease Genetic Testing, Children's Hospital Medical Center) revealed a homozygous splice site mutation in the ABCB11 gene [c.1639(−2) A>C] predictive of a complete absence of the BSEP protein. A partial external biliary diversion was performed at 29 months of age.[14, 15] He continued to have elevations of serum aminotransferases, growth impairment (height and weight < 5th percentile for age), and elevated serum bile acids (119 μmol/L). He received a transplant with a whole liver allograft from a deceased donor at 5 years of age. The hepatic explant histology demonstrated micronodular cirrhosis with marked hepatocanalicular cholestasis and small but generally intact interlobular bile ducts. Immunohistochemistry with BSEP antibodies (performed by Pierre Russo, Department of Pathology, Children's Hospital of Philadelphia) was abnormal with no canalicular or hepatocyte staining, and this indicated an absence of BSEP protein (Fig. 3A). Transplant immunosuppression was induced with tacrolimus, MMF, and corticosteroids. The patient did well after transplantation, but at 17 years of age (12 years after LT), he presented with pruritus and hyperbilirubinemia (the total bilirubin level peaked at 22.5 mg/dL, and the direct bilirubin level peaked at 18.8 mg/dL) with a normal GGT level. A liver biopsy sample showed lobular cholestasis with mild portal inflammation. The total serum bile acid level was 412 μmol/L. Repeat liver biopsy showed some apoptotic bodies and hepatocanalicular cholestasis but insufficient elements to diagnose rejection (Fig. 3B). A suspicion that the cholestatic allograft dysfunction was antibody-mediated recurrent PFIC2 led to testing that demonstrated anti-BSEP antibodies in the patient (Fig. 2) with a serum titer of 1:1280. Immunohistochemistry with BSEP antibodies showed normal expression of canalicular BSEP and indicated that although the BSEP antibody blocked the function of BSEP, it did not eliminate the protein from the canaliculus (Fig. 3C).
Treatment with corticosteroids (1 mg/kg/day for 7 days) failed to improve his clinical status. Because of the success with case 1, this patient was treated with a regimen of plasmapheresis, IVIG, and rituximab (his course with treatment is shown in Table 2). Specifically, he received 5 cycles of plasmapheresis, with each followed by IVIG (1 g/kg), over a 2-week period, and this was followed by a single dose of rituximab (375 mg/m2). Six months after the rituximab treatment, he was clinically doing well with normal liver enzyme and bilirubin levels. His most recent total serum bile acid level was 12 μmol/L. Titers of anti-BSEP antibodies were not repeated.
Table 2. Management of the Recurrence of the PFIC2 Phenotype After LT
Plasmapheresis (×3), monthly IVIG (×12), rituximab every 6 weeks (×5)
Plasmapheresis (×5), IVIG (×5), rituximab (×1)
Tacrolimus, corticosteroids, ursodiol
Tacrolimus, methylprednisolone, MMF
Plasmapheresis (×1), rituximab (×5)
Cyclosporine, prednisone, IVIG (×1)
Cyclosporine, azathioprine, prednisone
Immunosuppression after treatment
Prednisone, tacrolimus, azathioprine
Normal 33 months later
Normal 3 months later
Disappearance of jaundice 4 months later
Normal 1 month later
Normal 3 months later
Cardiac arrest, death while awaiting second LT
Normal 17 years later
The subjects of this report received LT for progressive cholestatic liver disease due to severe ABCB11 gene mutations. They developed recurrent cholestatic disease after transplantation and were found to have high-titer anti-BSEP IgG antibodies. They met the definition of having recurrent BSEP disease due to antibody formation.[8-10, 13] Their courses after the development of recurrent BSEP disease differed, we think, because of the therapy that they received. Patient 1 developed progressive liver disease in his first allograft that was unresponsive to enhanced immunosuppressive therapy, including a change to sirolimus due to concerns about chronic rejection contributing to allograft dysfunction, and this ultimately necessitated a second transplant. He subsequently developed proven antibody-mediated recurrent BSEP disease in the second allograft, which in retrospect strongly suggested that the first allograft was similarly affected. A therapeutic strategy was developed for his antibody-mediated recurrent BSEP disease, and he responded nicely with a full and sustained resolution of his cholestatic liver disease. Patient 2 was treated with the therapy used on the second occasion to treat patient 1 with minor modifications, and his cholestatic disease was promptly resolved. Both patients have maintained normal graft function since then. This experience leads us to recommend aggressive treatment with plasmapheresis, IVIG, and rituximab immediately upon the diagnosis of antibody-mediated recurrent BSEP disease.
PFIC2 is unique among the many genetic diseases for which LT is performed in its development of antibody-mediated recurrence. The mechanism for the antibody-mediated recurrence of this disease is not entirely clear. The domain against which BSEP antibodies develop resides within the canalicular space and can be considered sequestered from the immune system. It is reasonable to assume that the adaptive immune system must be exposed in some way to this domain in order to initiate an antibody response. Biliary injury, whether due to acute rejection, a bile duct obstruction, or another reason, might lead to paracellular flux of canalicular bile into plasma, as is the case for bile lipids in the production of lipoprotein X during obstructive cholestasis. Similarly, biliary injury might lead to bits of canalicular membrane or exfoliated membrane proteins appearing in plasma and thus provide the exposure to BSEP needed to incite a B cell response. Therefore, BSEP must be considered conditionally sequestered. The conditions needed for sensitization are such that individuals undergoing transplantation for PFIC2 might never develop BSEP antibodies. In addition, this conditional sequestration might explain the highly variable interval between transplantation and disease recurrence in those patients who do develop antibodies. The canalicular space has another feature that makes antibody-mediated disease possible: IgG readily enters the canaliculus by way of the paracellular pathway. Thus, conditional exposure of BSEP to the immune system provides for the variable development of IgG anti-BSEP antibodies, which, if they develop, find access to the canaliculus to bind BSEP. The other question is whether the antibodies are directed against an epitope that is involved in bile salt transport or, when it is bound by antibodies, structurally changes the transporter to the degree that bile salt transport is impaired. It is likely that antibodies to only certain epitopes are disease-producing.
The foregoing suggests that not all individuals undergoing LT for PFIC2 can be expected to develop antibody-mediated recurrence. In a 2010 review by Siebold et al., 6 of 36 children (17%) with PFIC2 who underwent LT developed recurrent BSEP disease. A mutation analysis suggested a complete absence of BSEP protein in these 6 patients, but this was unavailable for the other 30 children.[8-10, 13] It cannot be said with certainty that a complete absence of BSEP is a necessary precondition for developing antibodies after transplantation, but available data suggest that this may be the case. Thus, when we are constructing a posttransplant management plan for individuals with PFIC2, it seems that special attention should be paid to patients with mutations expected to result in a complete absence of BSEP protein. Immunohistochemistry for BSEP, as in these cases, can confirm the effect of mutations on canalicular BSEP expression and should be performed if at all possible. Monitoring for the development of BSEP antibodies after transplantation is problematic. Routine periodic measurements of antibody titers are not currently feasible because such testing is not commercially available. Furthermore, there is little understanding of the relationship between the development of any BSEP antibodies and the development of recurrent disease. Routine periodic measurements of serum bile acid levels might provide insight into the effects of antibody development. Good allograft function is accompanied by normal total serum bile acid levels, as demonstrated for the patients after treatment. Rising levels would point to an impairment of BSEP function and might occur well before the development of clinical cholestasis. A carefully performed prospective study of PFIC2 patients undergoing transplantation would be required to fully answer many of the remaining questions about this unique condition. For the moment, LT for PFIC2 should carry with it a heightened awareness of the possibility of recurrent disease, and this complication should assume a place at the top of the list of differential diagnoses should cholestasis develop at any time after transplantation.
Standard immunosuppression for LT appears not to be effective in preventing recurrent BSEP disease. Furthermore, most cases of recurrent disease fail to respond to enhanced transplant immunosuppression that would normally be effective in treating allograft rejection, as shown in Table 2, which reports on 3 of the four previously published case series.[8, 10] Symptoms have persisted in 6 of 10 reported cases despite enhanced immunosuppression. Most cases progress to liver failure and need retransplantation, as did our patient 1. Aggressive therapy specifically directed against humoral immunity is likely to be required for a satisfactory outcome. One reported case was treated with a single dose of IVIG, and the outcome was poor. Another patient was treated with plasmapheresis plus 5 courses of rituximab, and the resolution of the disease was incomplete. The patients reported here achieved full resolution of recurrent BSEP with a regimen of plasmapheresis followed by multiple courses of IVIG and rituximab. In patient 1, the biochemical measures of cholestasis worsened after therapy with plasmapheresis and IVIG, and this suggests the possibility that the continued production of anti-BSEP antibodies might have been involved. The persistent cholestasis led to the addition of rituximab to the therapy and, ultimately, to the resolution of the disease. Patient 2 received a more complete and compressed treatment course of plasmapheresis with IVIG followed by rituximab over the course of 2 weeks on the basis of the experience with patient 1. Rituximab is a chimeric monoclonal antibody against CD20, a B cell surface protein that is primarily found on immature B cells. Its effect on the production of specific antibodies might be delayed because it has little if any immediate effect on populations of plasma cells. Thus, it might not be an appropriate single-agent therapy. The rapid and complete resolution of disease in response to the combined therapy applied to patient 2 suggests that this regimen is an appropriate approach to treating the condition of antibody-mediated recurrent BSEP disease. Rituximab has been reported to lead to prolonged hypogammaglobulinemia, and this should be kept in mind when one is using this approach.
From this limited experience and the literature, we conclude that care providers should maintain a high level of suspicion for antibody-mediated recurrent disease in posttransplant PFIC2 patients. In the event that cholestasis develops, recurrent disease should be highly suspected. Serum bile salts, measured at regular intervals (2-4 times a year), could serve as a potential screening tool. Rising levels or clinical cholestasis in the absence of another graft pathology should be followed by antibody measurements. However, the measurement of BSEP antibodies is not commercially available. Thus, in the clinical arena, if antibody-mediated disease is strongly suspected, treatment with plasmapheresis, IVIG, and rituximab might be considered even without antibody measurements because of the consequences of not providing directed therapy.
The authors thank Pierre Russo, M.D., for performing the BSEP immunohistochemistry for patient 2.