Carla S. Coffin was supported by the American Association for the Study of Liver Diseases through the 2006 Advanced Hepatology Fellowship Award and the 2010 Jan Albrecht Clinical and Translational Research Award.
John P. Roberts,
Departments of Surgery, University of California San Francisco, San Francisco, CA
The use of hepatitis B immune globulin (HBIG) and oral nucleos(t)ide analogue (NA) therapies has significantly improved patient outcomes: the risk of serologically evident hepatitis B virus (HBV) reinfection has been reported to be less than 10% in most published studies.1, 2 The factors associated with recurrent HBV in the current era of prophylaxis appear to be high serum levels of HBV before transplantation3, 4 and the presence of mutations in the HBV surface (S) and polymerase (P) genes before transplantation.2, 5 In a recent study by Cheung et al.,6 occult HBV infections after liver transplantation [LT; ie, persistent HBV DNA in the liver despite negative findings for hepatitis B surface antigen (HBsAg) in serum] were shown to originate from both the donors for patients who received antibody to hepatitis B core antigen (anti-HBc)–positive grafts and the recipients despite prolonged NA prophylaxis. However, the persistence of intrahepatic HBV covalently closed circular DNA (cccDNA) post-LT was linked to the persistence of HBV cccDNA within the liver of an anti-HBc–positive donor. As the efficacy of NA therapy has improved, there have been more reports of prophylaxis protocols using a truncated course of HBIG with long-term NA therapy as the primary form of prophylaxis post-LT.7, 8 Additionally, according to a recent report of LT recipients with undetectable serum HBV DNA levels at transplantation and with no evidence of intrahepatic and cccDNA after at least 3 years of HBIG and NA therapy, the complete withdrawal of HBV prophylaxis was possible without evidence of recurrence in the majority of the patients with off-treatment follow-up of approximately 2 years.9 Notably, before prophylaxis was available, natural history studies showed that not all patients developed recurrent HBV infections.10 For the individualization of prophylaxis protocols, knowledge of both the potential sources of viral recurrence [ie, the liver and peripheral blood mononuclear cells (PBMCs)] and the ability of extrahepatic reservoirs to harbor drug-resistant variants is important.
HBV replicates via an error-prone reverse transcriptase and exists within the host at any given time as a collection of closely related but distinct genomes or quasispecies.11 In addition, because of overlapping open reading frames, changes within the HBV P gene can also alter the HBV S gene sequence and potentially lead to HBIG treatment failure.12, 13 These viral variants are subjected to the selective pressure of the host immune system as well as antiviral therapy directed toward the HBV reverse transcriptase/P. In the current era of highly effective antiviral therapy for HBV, many patients have received sequential drug therapy, and this has possibly led to the selection of complex mutants.14 The fitness of these mutants in a defined environment is considered to be a key factor in the development of drug resistance, but the molecular evolution of these variants in the context of immunosuppression after LT is unknown.
Viral reservoirs such as the lymphoid cell compartment (particularly those harboring drug-resistant HBV) may be particularly important in the optimization of prophylaxis post-LT. Our previous studies in patients on long-term anti-HBV therapy have shown that the PBMC reservoir is more likely to harbor the wild-type (WT) virus; this is possibly related to the lower level of antiviral therapy infiltration into the lymphoid cell environment and the limitations of the replication space (ie, the pool of cells in which HBV can establish an infection).15 We hypothesize that a similar situation occurs in LT recipients and that HBV persists, albeit at low levels, despite highly suppressive antiviral therapy after transplantation.
The aims of the current study are to characterize occult HBV infection in the liver and extrahepatic compartments (i.e., lymphoid cells) and to determine the molecular diversity of HBV quasispecies, the prevalence of antiviral resistant mutants and changes within the overlapping HBV S gene in patients on long-term suppressive antiviral therapy after liver transplantation.
ADF, adefovir; anti-HBc, antibody to hepatitis B core antigen; C, HBV core gene; cccDNA, covalently closed circular DNA; dN, nonsynonymous distance; ETV, entecavir; FTC, emtricitabine; Gr, grade; HBeAg, hepatitis B e antigen; HBIG, hepatitis B immune globulin; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HDV, hepatitis delta virus; LMV, lamivudine; LMVr, lamivudine resistance; LT, liver transplantation; P, HBV polymerase gene; NA, nucleos(t)ide analogue; NAH, nucleic acid hybridization; nt, not tested; PBMC, peripheral blood mononuclear cell; PCR, polymerase chain reaction; S, HBV surface gene; SEM, standard error of the mean; st, stage; TDF, tenofovir; WT, wild type; X, HBV X gene.
PATIENTS AND METHODS
Twelve LT recipients were recruited with a median follow-up of 18 months after LT (range = 12-180 months), and biological samples were collected at the end of each patient's follow-up period (Table 1). All patients provided written, informed consent for participating in this investigation, and the study received approval from the local institutional ethics review board according to the guidelines of the 1975 Declaration of Helsinki. No donor organs were obtained from executed prisoners or other institutionalized persons. Six males and 6 females were enrolled (Table 1). The median age was 63.5 years (range = 51-74 years). One patient received an anti-HBc–positive graft (case 1); in 1 case, the donor anti-HBc status was unknown (case 2); and the other 10 patients received grafts from anti-HBc–negative donors. The indications for LT were decompensated cirrhosis in 42% (5/12) and hepatocellular carcinoma (HCC) in 58% (7/12). One patient had a hepatitis delta virus (HDV) coinfection (case 2), and another with a pretransplant diagnosis of HCC was found to have cholangiocarcinoma in the explant liver (case 5). Immunosuppression consisted of tacrolimus (8/12), cyclosporine (1/12), or sirolimus (1/12), mycophenolate mofetil (8/12), and low-dose prednisone (6/12). Antiviral resistance before LT was based on evidence of virological breakthrough during treatment after confirmation of compliance. Resistance testing was not performed. The recurrence of an HBV infection post-LT was defined as the reappearance of HBsAg in serum.
Table 1. Summary of the Clinical and Virological Data for Patients on Suppressive Anti-HBV Therapy Who Were Followed for 1 to 15 Years Post-LT
Plasma sensitivity <10 vge/ milliliter of plasma HBV DNA
Pre-LT Plasma HBV DNA (Copies/mL)‡/ Donor Anti-HBc Status
HBV cccDNA sensitivity <100 vge/microgram of total Liver DNA†
HBV cccDNA by sensitivity <100 vge/ microgram total PBMC DNA†
Known Clinical HBV Recurrence Post-LT
NOTE: In cases with no known clinical HBV recurrence post-LT, 90% (9/10) liver samples were HBV DNA–positive; 28.5% (2/7), PBMC samples were HBV DNA–positive; and 33% (3/9) plasma samples were HBV DNA–positive.
Antiviral drug resistance was assumed because of increasing serum HBV DNA levels (ie, virological flare) after the confirmation of compliance with daily oral NA drug therapy.
In house nested PCR/Nucleic Acid Hybridization Assay. For HBV DNA assay sensitivity of <10 virus genome equivalents (vge) per ml of plasma; <10 vge per microgram of total liver or total PBMC DNA. For HBV cccDNA assay sensitivity of <100 vge per microgram of total liver or total PBMC DNA.
Commercial PCR assay with a sensitivity of 160 copies/mL.
Liver tissue was collected from all 12 patients, plasma was collected from 11 patients, and PBMCs were isolated from 9 cases at the end of the clinical follow-up period (Table 1). Whole blood for the isolation of plasma and PBMCs was collected at the same time as the liver biopsy samples. PBMCs were isolated by density gradient centrifugation on Ficoll-Histopaque Plus (Pharmacia, Uppsala, Sweden) and were stored at −80°C. Liver biopsy specimens were immediately snap-frozen in liquid nitrogen and stored at −80°C. All biopsy samples were scored for hepatitis and rejection according to the Banff schema and for disease activity according to the Ludwig-Batts staging system.
HBV DNA and cccDNA Detection
Total DNA was isolated from 200 μL of plasma, from approximately 2 × 106 PBMCs, or from 2 μg of liver tissue with a standard proteinase K digestion and phenol-chloroform extraction procedure as previously described.16, 17 HBV DNA was detected with a highly sensitive direct and nested polymerase chain reaction (PCR) method; 4 sets of oligonucleotide primers specific for nonoverlapping regions of the S, core (C), P, and X genes were applied. The primer sequences and amplification conditions were reported previously.15, 18, 19 The specificity of the amplified virus sequence and the validity of the negative and positive controls were routinely verified by nucleic acid hybridization (NAH; ie, Southern blot hybridization analysis); radiolabeled complete recombinant HBV DNA was used as a probe. For the PCR detection of HBV cccDNA, total DNA was digested with S1 nuclease (Invitrogen, Burlington, Canada) according to the manufacturer's instructions to eliminate single-stranded DNA sequences, and then it was amplified by PCR with primers spanning the nick region of the HBV genome as previously described.15, 18 For nested PCR, 10 μL of the direct PCR product was amplified with appropriate negative and positive controls. The PCR/NAH sensitivity was determined by the amplification of serial dilutions of recombinant HBV DNA and by a densitometer analysis of the resulting hybridization signals. For detection of HBV DNA the assay sensitivity was <10 virus copies/virus genome equivalents (vge) per ml of plasma; <10 vge per microgram of total liver or total PBMC DNA. For detection of HBV cccDNA the assay sensitivity was <100 vge per microgram of total liver or total PBMC DNA.
HBV P Gene Sequencing and Analysis of the P Gene and the Overlapping S Gene
Amplified fragments of the HBV P gene in liver, PBMC, and/or plasma samples from 3 cases (cases 1, 2, and 12) were cloned with the TOPO TA cloning system (Invitrogen, Burlington, Canada). Ten to 20 clones per sample were sequenced bidirectionally by universal priming with a 3730 XL sequencing instrument (Applied Biosystems, Foster City, CA). The HBV P sequence data and the overlapping HBV S region were analyzed with Sequencher version 4.7 (Gene Codes Co., Ann Arbor, MI). Phylogenetic and evolutionary changes within the HBV reverse transcriptase and the HBV quasispecies complexity were analyzed with MEGA version 4.0. The overall diversity of the HBV quasispecies in the liver and PBMC compartments was compared to previously published sequences from a cohort of nontransplant patients with chronic HBV who were on long-term anti-HBV therapy15 (see Supporting Information Table 1).
Posttransplant Clinical Outcomes
All 12 patients received post-LT prophylaxis with HBIG in combination with NAs (Table 1). The pre-LT viral load was known in 10 of the 12 patients, the HBV DNA level was undetectable (<160 copies/mL) in 5 patients, and the median HBV DNA level in those with detectable viremia before LT was 3.6 × 104 copies/mL (5.5 × 102 to 106 copies/mL) according to PCR (Amplicor assay, Roche, Pleasanton, CA). Both patients without HBV DNA results were hepatitis B e antigen (HBeAg)–positive before LT. Post-LT, 10 of the 12 patients remained serum HBsAg–negative for a median follow-up period of 15.5 months (range = 12-96 months) after transplantation. In 2 patients with follow-up periods of 144 and 180 months, HBV recurred with the reappearance of HBsAg in their serum 3 and 7 months post-LT, respectively. These 2 patients were on lamivudine (LMV) and HBIG at the time of recurrence. Case 1 was subsequently treated with LMV and adefovir (ADF), and case 2 (coinfected with HDV) was treated with LMV alone for approximately 10 years and then with LMV and tenofovir (TDF). The remaining 10 patients (median follow-up = 15 months, range = 12-96 months) without recurrence (HBsAg-negative) received HBIG plus an NA for at least a year, which was followed by NA therapy. Four patients had lamivudine resistance (LMVr) before LT; both patients with recurrence had LMVr before LT. No patient suffered from HCC recurrence during follow-up.
HBV in Liver Tissue, PBMCs, and Plasma
As noted previously, 2 patients had a remote history of clinically recurrent, serum HBsAg–positive HBV disease. The evaluation of the liver biopsy tissue revealed the presence of HBV DNA with 3 or 4 HBV genome–specific primers (C, S, X, and P; Table 1 and Supporting Fig. 1). In addition, HBV cccDNA was detected in both liver biopsy samples (Supporting Fig. 2). HBV DNA was also detectable in the plasma and PBMCs of both patients. In the remaining 10 cases without clinical recurrence post-LT, HBV genome fragments were detected (with at least 1 HBV genome–specific primer) in 90% of the liver biopsy specimens (9/10). For only 1 patient (case 5) was the virus undetectable in the liver with all primers (Table 1). HBV genomes were detected in 3 of 9 plasma samples (33%) with HBV S primers, HBV C primers, or both and in 2 of 7 PBMC samples (28.5%) with HBV C, S, X, or P primers (including 1 sample with HBV cccDNA in PBMCs; Table 1 and Supporting Information Figs. 1 and 2).
Clonal Sequence Analysis of WT and Drug-Resistant HBV in the Liver and PBMCs
HBV clonal sequencing analysis was performed for 3 patients (cases 1, 2, and 12; Table 2). As noted, cases 1 and 2 had a previous history of HBV clinical recurrence post-LT but were on NA suppressive therapy at the time of sampling. For case 1, the sequence analysis revealed that 100% of the clones in the plasma and 70% of the clones in the liver showed a mutation (M204I) associated with LMVr. However, the PBMC sample carried mostly the WT virus (70% of the clones); only 30% showed the M204I LMV mutation. Case 2 (coinfected with HDV) was on LMV monotherapy at the time of the sequence analysis, which showed that the majority of the clones in the liver (81%) and 100% of the clones in PBMCs had a WT virus sequence. The third patient (case 12) had no known clinical recurrence of HBV or antiviral resistance and was on long-term prophylaxis with entecavir (ETV) and HBIG. Clonal sequencing revealed that 100% of the HBV sequences were WT in the liver and PBMCs.
Table 2. Summary of the Sequence Analysis of the HBV P Gene Fragments and the Amino Acid Changes in the Overlapping S Gene in the Liver, PBMCs, and Plasma
All patients received HBIG for at least 1 year post-LT. An analysis of the overlapping amino acid sequence encoded by the S gene did not show the well-described immune escape substitutions (ie, G145R, D144A, and P120S) or any minor amino acid changes leading to a loss of cysteine residue that could affect the secondary structure of the virus, the loss of conformational epitopes, and the antigenicity.13 Interestingly, however, all cases exhibited minor amino acid substitutions from the consensus sequence, and a variation of at least 30% was noted in amino acid changes between the 3 virus compartments in each patient.
Comparison of HBV Quasispecies Diversity in LT Recipients and Nontransplant Recipients on Long-Term Antiviral Therapy
The phylogenetic analysis of the sequenced P gene region from nontransplant patients with HBV infections who were on suppressive antivirals (n = 9; Supporting Table 1) were compared with the sequences of post-LT patients (n = 3; Figs. 1 and 2). A bootstrap analysis revealed that the virus sequences did clearly separate for each patient in the post-LT group, but they were more uniformly distributed in these patients. This can be attributed to the retrieval of more WT virus sequences from the different tissues and cell compartments of the posttransplant patients. Furthermore, we observed that the diversity of HBV quasispecies was significantly (P < 0.05) lower in the post-LT patients in either the PBMCs or the liver compartment; this was reflected in the nonsynonymous distance (dN) values (ie, distances based on amino acid–changing mutations; Fig. 2). A similar pattern was observed for the overall distance (data not shown).
Identifying patients for whom prophylaxis can be discontinued or, conversely, patients who are at risk of prophylaxis failure is important for long-term graft survival. The serum HBV DNA level at the time of LT is the most reliable predictor of HBV recurrence post-LT,3, 4 and this supports the goal of achieving undetectable HBV DNA levels before LT. However, a further refinement of risk groups may be possible by the evaluation of HBV DNA in other compartments (specifically the liver and lymphoid cell reservoirs) post-LT. In this study, we have characterized the frequency of HBV replication in serum, PBMC, and liver samples from patients on long-term prophylaxis. The presence of detectable and replicating virus in 1 or more sites in the majority of the patients suggests a persistent risk of clinical HBV recurrence (positive findings for HBsAg and/or recurrent hepatitis) and a need for ongoing prophylactic therapy. Other studies of LT recipients on prophylaxis have reported similarly high rates of intrahepatic HBV detection and indicate that these patients are at risk for recurrent HBV.6, 22, 23
Lenci et al.9 reported the successful weaning of 25 of 30 low-risk patients (ie, undetectable HBV DNA in the serum at the time of LT, negative findings for HBeAg, and undetectable intrahepatic total DNA and cccDNA at least 3 years after LT) from all HBV prophylaxis; PBMCs were not evaluated. In that study, after staged weaning from HBV prophylaxis, 5 patients suffered from HBV recurrence; in subsequent protocol liver biopsy samples, 1 patient was found to have detectable intrahepatic cccDNA, and 5 of the 5 patients had intrahepatic HBV DNA. In comparison, we found that 9 of 10 patients with no known recurrence (ie, they were HBsAg-negative) still had detectable levels of intrahepatic HBV DNA, and 2 patients with a remote history of recurrence, although they were HBsAg-negative at follow-up (144 and 180 months post-LT), had intrahepatic cccDNA. Additionally, 1 of 10 patients with no recurrence had cccDNA within the PBMC compartment 21 months post-LT. Although Lenci et al. found that cccDNA negativity in a baseline liver biopsy sample predefined a low risk of recurrence, this did not prove to be an absolute measure. We hypothesize that this may have been due to the existence of PBMC viral reservoirs and/or the sensitivity limitations of the assays for the detection of cccDNA. The assessment of intrahepatic HBV genomes (HBV DNA and cccDNA) is limited by the need for invasive liver biopsy and by the lack of a widely accepted and rigorous standardized laboratory assay. In our study, the sensitivity of the total DNA and cccDNA assays for HBV was comparable to (if not greater) than the sensitivity of the assays reported by other investigators because of 2-round PCR amplification and NAH of amplicons with a radiolabeled recombinant HBV DNA probe. This may account in part for the higher detection of intrahepatic HBV DNA in our patients.
Whether HBV replication within lymphoid cells can be a source of HBV for reinfecting the liver graft is controversial.24-26 Previous studies have shown that the recurrence of HBV post-LT can be demonstrated first in PBMCs and then in the liver compartment,25 and the predominant HBV S gene sequence in PBMCs before LT is the major sequence identified after LT. In the woodchuck model of hepatitis B, lymphoid cells are infected early in the course of the infection and during the infection with low levels of the invading virus (ie, <1000 virions).20, 21, 27 Woodchuck hepatitis virus is predisposed to target the lymphatic system before the liver is engaged. Thus, low doses of HBV initially invade and propagate in lymphoid cells, and there is subsequent spreading to the liver (graft). We found that the HBV detected within PBMCs was predominantly WT; this was also found in our previous study of chronic hepatitis B carriers on long-term antiviral therapy (Supporting Table 1).15 We hypothesize that the predominance of WT HBV and the lower quasispecies diversity within the lymphoid cells were due to the lower rates of viral replication and the replication space; in this case, the most replication-fit virus (ie, the WT virus) would have predominated. Unfortunately, we lack PBMCs from patients before LT for a comparison of viral quasispecies before and after LT.
An analysis of the P gene and overlapping S gene sequences in our patients revealed the presence of LMVr variants in liver samples and in some PBMC samples, but no predicted changes in HBV antigenicity at the immunodominant amino acid positions affecting virus conformational epitopes that could lead to a loss of antigenicity. Interestingly, all patients had S amino acid changes (Table 2), which differed between the liver, PBMC, and plasma compartments; this pointed to undefined factors affecting the evolution of the virus. Additionally, the HBV diversity was lower in LT recipients versus non-LT control patients. This suggests that host immunity may be a dominant factor in viral molecular evolution. Different selective forces, including immune-mediated positive selection and virus-mediated negative selection, likely shape viral population dynamics within a host.28
In summary, we have shown that the majority of patients on effective prophylaxis (negative findings for HBsAg in serum) have detectable HBV DNA in the liver and/or PBMCs. This indicates ongoing replication at very low levels despite prophylaxis and implies a risk for recurrent HBV disease if the prophylaxis is stopped. However, for those few patients who are negative for HBV DNA and cccDNA in all compartments, the discontinuation of therapy may be an option, as recently shown.9 Going forward, we believe that additional longitudinal studies are needed to characterize the changes in HBV in different compartments and the ways in which these changes are related to outcomes with prophylaxis or prophylaxis withdrawal. Such studies will yield a more individualized approach to the prevention of HBV recurrence post-LT.