Hepatitis B after liver transplantation (LT) greatly affects patient outcome. In principle, 2 modes of infection are possible: (1) endogenous reinfection in patients who were transplanted for hepatitis B, and (2) exogenous or de novo infection, mostly as a result of nosocomial transmission of hepatitis B virus (HBV) through the liver allograft, blood products, and other exogenous sources. Although the prevalence of de novo infection with HBV is only 3.5%1 and has a good prognosis in most cases, endogenous reinfection is frequent and may lead to severe hepatitis.2 A prerequisite for endogenous reinfection after LT is the presence of active hepatitis B before transplantation. The incidence of reinfection after LT depends on the level of pretransplant viremia and the prophylactic measures taken.3–6 Few cases of reinfection have been described in patients with serologic signs of previous hepatitis B.7–9 These patients were negative for hepatitis B surface antigen (HBsAg) and positive for anti–hepatitis B core antigen (HBc), but negative for anti-HBs. In recipients who are positive for both anti-HBs and anti-HBc antibodies before LT, to our knowledge, clinically apparent reinfection has not been described yet. By definition, these patients have resolved hepatitis B and can be divided into 2 clinically important groups in the transplantation setting. First, the vast majority of these individuals are negative for serum HBV DNA by polymerase chain reaction (PCR). In the second group, serum HBV DNA can be detected by PCR and these patients are referred to as occult HBV carriers. The outcome of LT in such carriers has yet been investigated in only few studies. None of the study patients experienced clinically apparent reinfection with HBV,1, 7–10 although viral reactivation may occur after LT, possibly after apparent de novo hepatitis B.1
We present the case of a patient, who underwent LT due to cryptic cirrhosis with serologic signs of previous hepatitis B. He did not receive hepatitis B prophylaxis and experienced acute hepatitis B after the second retransplantation. Retrospective analysis revealed that the patient was an occult, unidentified HBV carrier before his initial LT. Only the fact that the infecting HBV exhibited an unusual insertion within the HBV polymerase/HBsAg gene enabled retrospective detection of this HBV strain by variant-specific PCR in blood and native liver samples obtained before all transplantations.
HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; LT, liver transplantation; HBVINS+, HBV with frame insertion of CAG motif within HBV polymerase; HBc, hepatitis B core; PCR, polymerase chain reaction; CP, core promoter.
MATERIALS AND METHODS
The case patient was a 62-year-old man from Saudi Arabia in whom Child C liver cirrhosis had been diagnosed 3 months earlier. Further workup revealed no definitive cause of his liver cirrhosis. Investigation of serologic parameters for HBV at that time showed a constellation compatible with past exposure to hepatitis B. The patient was positive for anti-HBs (11 IU/mL) and anti-HBc IgG and negative for HBsAg and anti-HBc IgM. There were no antibodies detectable to hepatitis C virus, other hepatitis viruses, or human immunodeficiency virus.
The patient underwent living related LT in September 2004. He received 2 partial liver grafts obtained from 2 of his sons. Because of hepatic artery thrombosis, a retransplantation with a full cadaveric liver graft was performed 10 days after the initial LT. Immunosuppressive medication consisted of basiliximab, cyclosporine, and corticosteroids. During outpatient follow-up regular serologic HBV testing showed negative results for HBsAg and a steady decrease of anti-HBs titers from 77 to 5 IU/mL (measured in monthly to bimonthly intervals). A second re-LT that used a full-size cadaveric donor liver was performed in April 2005 because of recurrent hepatic artery thrombosis with progressive ischemic type biliary lesions. Two days afterward, the patient simultaneously was positive for HBsAg and anti-HBs (1,466 IU/mL); PCR analysis showed a low viral load of 4 × 104 copies/mL (Fig. 1). Clinically, the patient recovered well and liver function tests showed a steady decline. However, on postoperative day 52, a sudden increase in the values of his liver function tests occurred (alanine aminotransferase 428 U/L; aspartate aminotransferase 764 U/L; glutamate lactate dehydrogenase 23 U/L; gamma glutamyl transferase 47 U/L; total bilirubin 1.9 mg/dL). HBV testing at that time revealed an increased viral replication (7 × 106 copies/mL) with positive results for HBsAg and anti-HBs (63 IU/mL). Treatment with lamivudine (100 mg/d) was initiated. Shortly afterward, Kaposi sarcoma affecting the skin of the left lower limb and the gastric mucosa was diagnosed. The immunosuppressive medication was changed to low-dose sirolimus monotherapy. After 6 weeks of lamivudine treatment, HBV replication had decreased below detection limit of the PCR assay (<1 × 102 copies/mL), and with the exception of gamma glutamyl transferase (478 U/L) values, results of the liver function tests had returned to normal values (alanine aminotransferase 31 U/L; aspartate aminotransferase 65 U/L; glutamate lactate dehydrogenase 10 U/L; total bilirubin 0.8 mg/dL).
HBsAg, HBeAg, anti-HBs, anti-HBe, and anti-HBc as well as anti-HAV, anti-HCV, and anti-HIV were determined by enzyme immunoassay (Axsym, Abbott Laboratories, Wiesbaden, Germany). A retrospective analysis was performed on all serum samples that had tested negative for HBsAg during routine clinical testing by performing an independent HBsAg enzyme-linked immunosorbent assay (HBsAg 5.0, Behring, Marburg, Germany).
Hematoxylin and eosin, periodic acid–Schiff, Masson-Goldner, and iron staining were performed on sections of routinely prepared paraffin-embedded native and transplanted liver tissue samples. In addition, paraffin sections were analyzed by immunohistochemistry for HBsAg and HBcAg applying the Envision detection system with the marker enzyme peroxidase following the instructions of the manufacturer (Dako Cytomation, Hamburg, Germany). HBsAg was detected by a monoclonal antibody (AbNo364M, dilution 1:1,500, Biogenex, San Ramon, CA, USA) and HBcAg by a polyclonal rabbit antibody (Code B0586, dilution 1:3,000, Dako Cytomation) without further antigen demascing before treatment of the sections.
PCR and Direct Sequencing
HBV copy numbers were quantified by real-time PCR (LightCycler-DNA Master SYBRGreenI, Roche Diagnostics, Basel, Switzerland) that targeted a conserved region of the HBV genome that overlaps the genes encoding the X protein and DNA polymerase. Primer pairs HBV1F (5′-CCGTCTGTGCCTTCTCATCTG-3′) and HBV1R (5′-AGTCCAAGAGTYCTCTTATGYAAGACCTT-3′) were used, and amplicons were quantified by fluorescent activities compared with serial dilutions of an external standard preparation (cloned HBV, plasmid pHBV991), which was calibrated with 2 reference samples containing HBV DNA of genotypes A and D, respectively. The detection limit of the PCR was shown to be 102 HBV genomes/mL serum (20 WHO U/mL serum) for both genotypes with a dynamic range up to 109 genomes/mL.11
To investigate HBV DNA in serum and liver samples a second, nested PCR assay was used. A part of the HBsAg was amplified with primers 252 (5′-AGACTCGTGGTGGACTTCTCT-3′)/1,309 (5′-AGAATGTTTGCTCCAGACC-3′) as external primers and 377 (5′-GGATGTGTCTGCGGCGTTT-3′)/840 (5′-ACCCCATCTTTTTGTTTTGTTAGG-3′) as internal primers spanning the HBsAg from codon 75 to codon 236. The reaction mixture for each nested PCR assay consisted of 5 μL of sample DNA, 35.5 μL water, 15 pmol of each primer, 5 μL PCR buffer, 1 μL dNTPs (10 mM), and 0.4 μL (2 units) Taq polymerase. After an initial melting step for 2 minutes at 95°C, 35 cycles were run (30 seconds at 95°C, 45 seconds at 55°C, 45 seconds at 72°C) with a final extension step at 72°C for 5 minutes. The amplified region was used for the determination of the patient's HBV genotype. Both strands of the amplification products were sequenced by BigDye termination chemistry (Applied Biosystems) with an automated sequencer (ABIPrism). A phylogenetic analysis of all sequences was performed by MEGA 2.1 with reference sequences of different HBV genotypes (GenBank accession numbers: genotype A: V00866; genotype B: D23677; genotype C: D00630; genotype D: V01460; genotype E: X75657; genotype F: X75658; genotype G: AF160501). Also, sequences of HBV strains obtained from patients followed at our institution were included. This cohort consisted of 53 consecutive, chronic HBV carriers who were treated with antiviral drugs or transplanted for chronic hepatitis B. The identical region and procedure was used for the determination of HBV resistance mutations within the overlapping gene of the HBV polymerase.
Core promoter (CP) and precore sequences of the samples were amplified by a nested PCR as described previously.11 The sequences were analyzed for CP mutations with nucleotide exchanges at positions A1762T and G1764A and for precore mutations at codon 28 (M2 stop codon mutant with the nucleotide exchange G1896A) and codon 29 (M4 mutant with the nucleotide exchange G1899A/G1900C).
Cloning and Sequencing
The HBsAg region was amplified from 2 serum samples and the liver before first transplantation and as described above. Amplicons were cloned into Escherichia coli by means of the TOPO-TA cloning system (Invitrogen, Karlsruhe, Germany), and 8 clones were sequenced and analyzed by automatic sequencing (Beckmann CEQ2000).
A nested PCR assay was established that specifically detected the insertion within the HBsAg gene of the reinfecting HBV variant (HBVINS+). Primers 252/1309 were used for the first round of amplification, followed by primers INS (5′-GAATCTGCAAACACCAGCAGCAG-3′)/840 within the second round of amplification, resulting in an amplicon of 376 base pairs. The CAG motif at the 3′ end of primer INS represented the insertion. HBVINS+ obtained from samples after reinfection served as a positive control. Water instead of template DNA and HBV DNA of 2 independent patients infected with HBV genotype D without insertion (D:450 and D:480) served as negative controls. The variant-specific PCR assay was carried out by the identical protocol as described above.
Investigation of Liver Donors
Sera of all 4 liver donors (2 living donors for the dual graft living related LT, and 2 cadaveric donors for the first and second re-LT) were investigated retrospectively by HBV PCR. These sera were negative for HBsAg and anti-HBc as determined before the respective LT. HBV PCR yielded negative results for serum HBV DNA in all 4 donors.
Two initial blood samples taken from the patient before the first LT were HBsAg negative and positive for anti-HBs and anti-HBc. These samples were retested for HBsAg when it became clear that the patient was infected with HBV after his third LT. Two independent assays gave negative results for HBsAg in the pretransplantation sera. HBeAg was also negative. During the time between first and third LT, the concentrations of anti-HBs ranged between 5 and 152 mU/mL. HBsAg or anti–HBc-IgM was never detected. Two days after the third LT, HBsAg became positive in both antigen assays with a low concentration below the detection limit of an immune-electrophoresis assay. As a result of mass transfusion, the concentration of anti-HBs rose to 1,466 mU/mL immediately after the third LT and slowly decreased to 3 mU/mL in August 2005.
Histologic examination of the explanted native liver revealed piecemeal necroses in the pseudolobulus area. Inflammatory infiltrates were seen together with hepatocellular single-cell necroses. Hepatocytes showed hyperchromatic nuclei without signs of cellular atypia. The slides stained with antibodies to HBsAg and HBcAg initially yielded negative results. A pathologic diagnosis of complete cirrhosis of unknown etiology was made.
After occurrence of HBV infection after the third LT, more slides of the native liver were stained retrospectively with anti-HBs and anti-HBc and evaluated for signs of HBV infection. Now, in few areas of some slides, grouped hepatocytes could be detected that showed cytoplasmic reactivity for HBsAg. HBcAg was not demonstrated.
After retransplantation, the explanted livers from the first 3 donors (2 live and 1 deceased donor) were analyzed and were histologically negative for HBsAg and HBcAg. Also, HBV DNA could not be detected by nested PCR, either by the S gene or by variant-specific PCR assay.
Virologic Analysis of Source of Infection
In retrospect, it turned out that the blood samples taken before the first transplantation were positive for HBV DNA by PCR. The concentrations were log 3.78 and log 2.7 HBV DNA copies/mL (Fig. 1). Both HBsAg/polymerase and HBV core sequences could be amplified from these sera. Also, the tissue of the explanted, native liver was positive for both regions of the HBV genome. Sequencing of the S gene showed an HBV strain genotype D, which was, however, completely distinct from other genotype D sequences of patients in our cohort (Fig. 2). At the time after the last LT, the HBV strain isolated from the blood exhibited an unusual insertion (CAG) within the HBV polymerase gene, leading to an additional glutamine at codon rt127. Also, the mutation rtV207L was detectable. The overlapping S gene was affected between codons 118/119 with an amino acid exchange to serine at codon 118 and an additional arginine between codons 118 and 119. The G145R mutation within the S gene was not detected. This variant is further referred to as HBVINS+. The sequence has been submitted to GenBank and was assigned the accession number DQ282130.
HBVINS+ was detected in 4 samples during further follow-up. Phylogenetic analysis revealed that HBVINS+ clustered with the HBV genotype D reference sequence but showed considerable heterogeneity with the patient's HBV being prevalent in the pretransplantation samples (Fig. 2). The nucleotide homology between the patient's HBV strains before and after LT was 91.6% across the analyzed region, resulting in an amino acid homology of 85% across the S gene and 86% across the polymerase gene. Also, independent genotype D sequences obtained from our cohort clustered differently (Fig. 2).
Sequencing of the CP-/precore gene revealed a stop codon mutation at nucleotide 1896 together with the G1899A mutation for both HBVINS+ and the HBV before LT. Also, the CP mutations at positions A1762T and G1764A were detectable in both virus populations. The nucleotide homology between HBVINS+ and wild-type HBV before LT was 98.6% across the CP-/precore gene.
To investigate the source of HBVINS+, we amplified and cloned HBsAg/polymerase sequences and analyzed the distribution of HBV variants within the native liver, the pretransplantation sera, and the initial serum after HBV infection after the third LT. There was 100% homology between all clones analyzed from the native liver and the pretransplantation sera. Eight of 8 clones from the serum and 4 of 4 clones from the native liver lacked the insertion and the rtV207L mutation. On the other hand, we did not detect the pretransplantation HBV in the sample after the emergence of HBsAg, indicating that HBVINS+ was the major population with a proportion of at least 80%. Taken together, the cloning results indicated that a major population of HBV was prevalent before the first LT that was distinct from HBVINS+.
From the cloning results, we estimated that the maximum detection limit was only 12.5% for HBV minor species. Therefore, we established an insertion-specific PCR assay to detect HBVINS+ with higher sensitivity. Known sequences of HBV genotype D without insertion were used as negative controls and yielded negative results (Fig. 3). HBVINS+ served as a positive control. With this assay, HBVINS+ could be demonstrated clearly in blood and liver samples obtained before the first LT (Fig. 3). In the 2 explanted donor livers, neither wild-type HBV nor HBVINS+ could be detected by either routine or insertion-specific PCR. The amplicons obtained by insertion-specific PCR from the native liver and the pretransplantation sera were sequenced and showed 100% homology to HBVINS+ that circulated after the third LT. These results confirmed endogenous reinfection with HBVINS+ after LT.
The present case is important for several reasons. For the first time, we describe a liver transplant patient with serologic signs of past exposure to hepatitis B who developed symptomatic hepatitis after LT.
Recent investigations of liver transplant patients with evidence of past exposure to HBV have already shown the possibility of HBV reinfection after LT.7–9, 12 However, positive results for HBV in these patients were mainly obtained by PCR analysis of post-LT liver tissue samples. Summarizing all patients reported in the literature, 4 of 15 tested positive for HBV DNA in post-LT liver tissue samples, and 4 of 24 were found to be HBV positive by sensitive PCR assays in post-LT serum samples. Only 1 of 49 patients investigated had temporary antigenemia (HBsAg positivity), and in no case was clinical reactivation of HBV after LT reported. In the case presented here, clinical reactivation was even more unexpected because the patient was not only positive for anti-HBc but also positive for anti-HBs antibodies before transplantation. Two studies have reported that posttransplant viremia was less likely in recipients who were positive for both anti-HBc and anti-HBs before transplantation compared with patients with sole anti-HBc antibodies before LT.7, 9
Compared with recipients of solid organ transplants, clinical HBV reactivation has been described in a number of patients with hematological malignancies harboring occult HBV infection.9 This difference is generally attributed to the more profound immunosuppression experienced by hematological patients during treatment.9 It might be speculated that the patient described here may well have been in a more immunosuppressed condition than the average liver transplant recipient because of his prolonged and complicated clinical course. The fact that the patient developed Kaposi sarcoma with gastric and cutaneous manifestations might serve as an indicator of severe impairment of T cell function.
The second point is the distinctive diagnostic feature in this patient. In occult carriers, Bréchot et al.13 observed some heterogeneity in the distribution of HBV genomes within the liver. This is in line with the initially negative results of HBsAg staining performed from the histology of the patient's native liver. Only the retrospective immunohistochemical analysis of several additional slides revealed hepatocytes positive for HBV antigens within the native liver. These positive results indicate that the binding affinity of the monoclonal antibody used for the detection of HBsAg in the liver was not reduced as a result of alterations of the S gene of the patient's wild-type HBV strain. However, the insertion within HBVINS+ may have led to a decrease in the binding affinity of this monoclonal antibody. A polyclonal antibody mixture might have had increased the sensitivity to detect HBsAg of HBVINS+. However, the results of the cloning experiments and the PCR assays showed that HBVINS+ represented a minor virus population of less than 20% of the total HBV population before the first LT. This suggests that HBVINS+ had a secondary clinical importance for the progressive liver disease before all transplantations as compared with wild-type HBV, which was the major virus population at this time. Nevertheless, our results emphasize the need for sensitive HBV DNA assays before LT in patients positive for anti-HBs/anti-HBc.
In general, the analysis of sources of HBV infection is still a big problem.14 In a retrospective study, posttransplantation hepatitis B could be elucidated in only 7 of 20 patients.8 For the patient after LT, however, the source of infection is important to know because the clinical course of hepatitis B is commonly believed to be much more benign after de novo infection as compared with endogenous reinfection,8, 12 although severe courses of de novo hepatitis B have been described.15, 16 Our results show that phylogenetically distinct HBV may occur in a single patient. This is in line with previous reports.17, 18 The course in our patient further indicates that minor populations of HBV may replace the formerly circulating major virus population during immunosuppression. The fact that neither HBV DNA nor HBsAg could be demonstrated in any of the explanted donor livers and in none of the consecutive serum samples drawn during the time period between the first and third LT suggests a long-lasting persistence of infectious HBVINS+ in extrahepatic tissue such as peripheral blood mononuclear cells19 and vascular endothelium.20 This minor strain was then capable of establishing a clinically relevant hepatitis. This has great impact for the analysis of sources of HBV infection after LT. In our patient, only the knowledge about the insertion within the polymerase/surface gene of HBVINS+ enabled specific and sensitive detection of this variant in pretransplant samples. Without this insertion, reinfecting HBV would have been missed in samples before LT and de novo infection would have been incorrectly diagnosed. Therefore, cloning and sequencing of a considerable number of subclones appears to be necessary in these cases in order to increase the sensitivity for minor HBV strains in pretransplant samples.