Clinical reactivation during lamivudine treatment correlates with mutations in the precore/core promoter and polymerase regions of hepatitis B virus in patients with anti-hepatitis B e-positive chronic hepatitis


Dr A. Marrone, Internal Medicine and Hepatology, Second University of Naples, Via Cotugno, 1, c/o Ospedale Gesù e Maria, 80135 Napoli, Italy.


Background : Drug-resistant mutants may emerge in patients with chronic hepatitis B receiving lamivudine therapy.

Aim : To evaluate whether different viral mutational patterns may be associated with clinical reactivation during lamivudine treatment in patients with chronic B hepatitis.

Methods : Eight anti-hepatitis B e-positive patients with (group A) and 14 patients without clinical exacerbation (five anti-hepatitis B e-positive, group B1; nine hepatitis B e antigen-positive, group B2) during lamivudine treatment were investigated.

Results : ‘Polymerase region’: M204V/I variants were found in all group A patients, but in none of group B1 (P = 0.0007) and in four of nine of group B2 (44%; P = 0.02) patients. The L180M substitution was detected in four of eight (50%) of group A and in none of groups B1 and B2. ‘Core promoter’: the double basic core promoter (A1762T/G1764A) variant was detected in seven of eight (87%) of group A and in one of five (20%; P = 0.03) of group B1 and one of nine (11%; P = 0.002) of group B2 patients. ‘Precore’: the G1896A stop codon mutation was present in seven of eight (87%) of group A and in zero of five (P = 0.004) of group B1 and one of nine (11%; P = 0.002) of group B2.

Conclusions : Different mutational patterns were observed in the lamivudine-treated patients with and without exacerbation. There was an association of the basic core promoter and stop codon mutations with lamivudine resistance in patients with disease exacerbation.


Lamivudine is a nucleoside analogue that is effective in inhibiting viral polymerase/reverse transcriptase (rt) of the hepatitis B virus (HBV). It is well-tolerated and several trials have demonstrated its efficacy in treating both hepatitis B e antigen (HBeAg)-positive and -negative chronic hepatitis B patients. Treatment leads to a decrease in HBV-DNA levels, anti-HBe seroconversion and alanine aminotransferase (ALT) normalization.1–6 During lamivudine treatment patients may experience a mild or severe clinical reactivation, which can be frequently associated with drug resistance caused by the appearance of mutations in the YMDD (tyrosine, methionine, aspartate, aspartate) motif of the viral polymerase.7–9 However, the emergence of lamivudine-resistant mutants is not always accompanied by liver disease exacerbation, suggesting that additional host and viral factors may contribute to different clinical outcomes.10–12

According to the recently proposed nomenclature for anti-viral-resistant HBV mutations in the rt domain of the polymerase region, the most common amino acid substitutions are found at positions M204V/I (previously M552V/I, subdomain C) and L180M (previously L528M, subdomain B).13, 14 As the HBV envelope open reading frame overlaps the polymerase gene, mutations in the latter can lead to changes in the immunodominant ‘a’ determinant of the hepatitis B surface antigen (HBsAg).9 Reduced antigenicity of HBsAg as a result of polymerase sequence changes selected during lamivudine treatment has been reported.15 Furthermore, it has been shown by in vitro studies that compensatory mutations in the fingers subdomain of the HBV polymerase can restore the replicative competence of lamivudine-resistant mutants.16

Chronic HBeAg-negative hepatitis associated with the presence of the precore (PC) stop codon mutation, which abrogates the ‘e’-antigen expression, is responsible for the majority of chronic HBV cases in the Mediterranean area and other parts of the world.6 HBV mutations that affect the HBeAg expression have been detected in cases of severe or fulminant hepatitis and in patients less responsive to interferon treatment.17–19 Clinical studies show that lamivudine resistance arises at a similar rate in HBeAg-positive and -negative patients, but the emergence of drug-resistant mutants is more frequently associated with a clinical reactivation in anti-HBe-positive patients.4, 20 Hadziyannis et al. reported a clinical exacerbation, with ALT values higher than eight times the upper limit of normal (UNL) in about 75% of HBeAg-negative Greek patients receiving treatment with virological breakthrough.6 Chen et al. using a recombinant HBV baculovirus system have demonstrated that the presence of the G1896A PC mutation restores the replication competence of lamivudine-resistant mutants.21 Furthermore, two different studies have described the evolution of PC and core promoter (CP) mutants during lamivudine treatment in association with changes in the YMDD domain and in some cases a reversion from G1896A PC mutants to wild type.20, 22

In the present study, we analysed polymerase, surface, PC and CP sequences from HBeAg-negative patients who experienced clinical reactivation during lamivudine treatment to see if the flares were triggered by different genotypic changes; two groups of HBeAg-positive and -negative patients without clinical reactivation were used as controls.

Patients and methods

We studied eight chronic anti-HBe-positive patients treated with lamivudine who had been admitted to our Unit during clinical reactivation (group A). Five anti-HBe-positive (group B1) and nine HBeAg-positive patients (group B2) receiving lamivudine treatment, without clinical exacerbation were used as controls. Serum samples were collected at the time of reactivation in group A and at a comparable point during lamivudine treatment in groups B1 and B2. Liver biopsy had been performed before starting lamivudine treatment in all patients.

A clinical reactivation was defined on the basis of an increase in the transaminase level to at least twice the UNL after ALT normalization during therapy. Virological reactivation was defined as a rise in the serum HBV-DNA values >1 log during therapy after initial negativization. All patients were hepatitis C virus (HCV)-, hepatitis D virus (HDV)- and HIV-negative before starting treatment and during follow-up. The clinical characteristics of the patients are presented in Table 1.

Table 1.  Patient clinical characteristics at the time of exacerbation in group A and the corresponding median time of lamivudine treatment in groups B1 and B2 (no clinical reactivation)
PatientsM/FMedian ageMedian duration of therapy before reactivationMedian ALT* (range)Median HBV-DNA (cps/mL)Median HAI (range)
  1. M/F, male/female; *UNL, upper normal limit; HAI, histological activity index; NR, no reactivation; ALT, alanine aminotransferase; HBV, hepatitis B virus.

Group A
 87/14918.5 months10 (2.5–20)1.3 × 1069.5 (5–12)
Group B1
 54/143NR1.01.6 × 1024 (4–8)
Group B2
 96/319NR1.08.9 × 1062 (1–5)

Serological tests

Markers for HBV, HCV, HDV and HIV were tested in serum using commercially available immunoenzymatic assays (EIA; Abbott Laboratories, North Chicago, IL, USA and Ortho Diagnostic Systems, Raritan, NJ, USA). HBV-DNA was assayed by Affigene HBV-VL (Sangtec, Bromma, Sweden).

Molecular analysis

Core promoter, PC, core and polymerase regions were analysed in group A patients during ALT flare up; the same regions were studied in groups B1 and B2 patients, who showed persistent normal ALT during lamivudine treatment.

Nucleic acid extraction and amplification

For genome analysis, HBV-DNA was extracted from 100 μL of serum mixed with 100 μL of a digestion solution containing 25 mm sodium acetate, 2.5 mm ethylenediaminotetracetic acid, 1% sodium dodecyl sulphate, 2 mg/mL proteinase K and 10 μg/mL of yeast transfer RNA as carrier. Digestion was performed at 37 °C overnight and the digests were extracted twice with phenol/chloroform, once with chloroform, precipitated with ethanol and resuspended in 10 μL of water.

DNA was used for the separate amplification by polymerase chain reaction (PCR) of the surface/polymerase, CP/PC and core regions. For amplification of the overlapping surface/polymerase region primers S (5′-GGTTATCGCTGGATGTGT-3′, sense, nt 367) and AS (5′-ACCCAGAGACAAAAGAAAA-3′, antisense, nt 826) were used for the first round, and S3 (5′-CTCTTCATCCTGCTGCTATGCC-3′, sense, nt 406) and S4 (5′-CAGACTTGGCCCCCAATACC-3′, antisense, nt 770) for the second round of nested-PCR. For the amplification of CP/PC regions primers BP1 (5′-TCTGTGCCTTCTCATCTG-3′, sense, nt 1554) and BP2 (5′-AATGCTCAGGAGACTCTAAG-3′, antisense, nt 2044) were used for the first, and BP5 (5′-ACTTCGCTTCACCTCTGCACG-3′, sense, nt 1586) and BP6 (5′-TTCCCGATACAGAGCTGAGGC-3′, antisense, nt 2020) for the second round of nested-PCR. For the amplification of the core region primers M3 (5′- CTGGGAGGAGTTGGGGGAGGAGATT-3′, sense, nt 1720) and Pol8 (5′- AGGATAGAATCTAGCAGGC-3′, antisense, nt 2265) were used for first, and M3 and 3C (5′- CTAACATTGAGATTCCCGAGA-3′, antisense, nt 2458) for the second semi-nested PCR. The reaction was performed in 100 μL containing 10 mm Tris-HCl pH 8.3, 50 mm potassium chloride, 200 μmol of each dNTP, 2.5 mm magnesium chloride and 1 unit of Taq polymerase (Applied Biosystems, PE, Italia). Samples were subjected to 35 rounds of denaturation at 95 °C for 1 min, annealing at 55 °C for 1 min and extension at 72 °C for 1 min. About 2 μL of the first round PCR product were used as template for the second round of PCR under the same reaction conditions, but performing 20 cycles only. The PCR products were electrophoresed in 1.5% agarose gels and visualized by UV transillumination, following staining with ethidium bromide.

Sequencing analysis

Amplicons were purified using Qiaquick spin columns (Qiagen, Hilden; Germany) according to the manufacturer's instructions and directly sequenced using an automated capillary sequence reader (Abi Prism 310, Applied Biosystem). The sequences were analysed with the dnasis package for Windows 95 and the amino acid alignments with PROSIS (Hitachi, Japan). The reference strain for HBV genotype D used was that defined by Galibert et al.23 HBV genotypes were determined by reverse hybridization line probe assay (INNO-LIPA HBV genotyping, Innogenetics N.V.; Ghent, Belgium).

Statistical analysis

Statistical analysis of non-parametric data were done using Fisher's exact test. Differences among groups A, B1 and B2 in histological activity index (HAI) and HBV-DNA serum levels were compared with the Mann–Whitney U-test. A P-value below 0.05 was considered significant.


Clinical and virological characteristics of patients

Group A consisted of eight anti-HBe-positive patients (M/F, 7/1) with a median age of 49 years (range: 25–63). Group B1 consisted of five anti-HBe-positive (M/F, 4/1) with a median age of 43 years (range: 26–60) and group B2 of nine HbeAg-positive (M/F, 6/3; median age 19 years, range: 12–26).

Liver histology.  The HAI values were significantly higher in group A patients than group B2 patients (P = 0.004), but not significantly higher than group B1 (P > 0.06). No differences in fibrosis and steatosis scores were observed among the groups.

The median time to reactivation was 12 months (range: 11–30). The ALT values were determined every month. During reactivation, the median ALT values rose to 10 × UNL (range: 2.5–20) and decreased to 4.5–6 × UNL later on in group A. In groups B1 and B2 the ALT levels were within the normal range in all patients.

No alterations in bilirubin and albumin levels, or prothrombin time were observed among patients with exacerbation. All patients had HBV genotype D and were reactive for HBsAg by commercial assay.

The median serum HBV-DNA levels were 1.3 × 106 cps/mL (range: 4.2 × 105–3 × 107) in group A, 1.6 × 103 cps/mL (range: 0–3.8 × 105) in anti-HBe-positive patients (group B1) and 8.9 × 106 cps/mL (range: 2.9 × 104–2.5 × 107) in HBeAg-positive patients (group B2). The HBV-DNA levels in patients in group A were significantly higher than those in anti-HBe-positive patients (group B1; P < 0.06), but not in those who were HBeAg-positive (group B2; P = 0.266).

Sequencing analysis

Amino acid substitutions in the polymerase/surface/core regions and nucleotide substitutions in the CP/PC HBV regions are reported in Tables 2 and 3 and in Figure 1.

Table 2.  Amino acid substitution patterns in the reverse transcriptase (rt) and surface region in group A (with clinical reactivation) and groups B1 and B2 (no clinical reactivation)
 Group A, n (%)Group B1, n (P-value)Group B2, n (P-value)
Rt region
 M204V/I8/8 (100)0/5 (0.0007)4/9 (0.02)
 L180M4/8 (50)0/5 (0.09)0/9 (0.02)
 L122F1/8 (12.5)4/5 (0.03)7/9 (0.01)
 P130Q1/8 (12.5)3/5 (0.1)8/9 (0.002)
 Y135S8/8 (100)0/5 (0.09)2/9 (0.001)
 Y135F0/84/5 (0.006)5/9 (0.02)
 V142E1/8 (12.5)4/5 (0.03)4/9 (0.03)
 T127P8/8 (100)0/5 (0.0007)2/9 (0.001)
 T127S0/83/5 (0.03)6/9 (0.006)
 Y134N1/8 (12.5)3/5 (0.1)5/9 (0.08)
 I195M4/8 (50)0/5 (0.09)1/9 (0.1)
 W196L4/8 (50)0/5 (0.09)4/9 (0.3)
Table 3.  Core promoter and precore nucleotide mutations, and core amino acid substitution patterns in group A (with clinical reactivation) and groups B1 and B2 (no clinical reactivation)
 Group A, n (%)Group B1, n (P-value)Group B2, n (P-value)
Core promoter
 T1753C7/8 (87)0/5 (0.004)0/9 (0.0004)
 A1762T/G1764A7/8 (87)1/5 (0.03)1/9 (0.002)
 G1896A7/8 (87)0/5 (0.004)1/9 (0.002)
 T12S5/8 (62.5)1/5 (0.1)0/9 (0.009)
 Y38F3/8 (37.5)0/5 (0.1)0/9 (0.08)
 E77D2/8 (25)0/5 (0.3)0/9 (0.2)
 P79Q4/8 (50)0/5 (0.09)0/9 (0.02)
Figure 1.

Amino acid substitutions in the hepatitis B virus (HBV) polymerase region of patients from group A (with clinical reactivation) and groups B1 and B2 (without clinical reactivation) during lamivudine treatment. Group A patients are numbered from 1 to 8; group B1 from 9 to 13 and group B2 from 14 to 22. Subdomains B and C are highlighted.

Polymerase region.  M204V/I changes were present in eight of eight patients (100%) in group A, in zero of five in group B1 (P = 0.0007) and in four of nine (44%) in group B2 (P = 0.02). The L180M change was detected in four of eight patients (50%) in group A vs. zero of five and zero of nine, respectively in groups B1 (P = 0.09) and B2 (P = 0.02). P130Q was present in one of eight (12.5%) in group A and in eight of nine (88%) in group B2 (P = 0.002). Y135S was detected in eight of eight (100%) in group A and in two of nine (22%) in group B1 (P = 0.001), while Y135F was revealed in zero of eight of group A and in four of five (80%) of group B1 (P = 0.006).

Surface region.  A T127P change was detected in eight of eight (100%) patients in group A, in zero of five in group B1 (P = 0.0007) and in two of nine (22%) in group B2 (P = 0.001). A T127S mutation was present in zero of eight patients in group A and in six of nine (66%) in group B2 (P = 0.006).

Core region.  Amino acid substitutions were observed in the T-cell epitope at position T12S in five of eight (62%) patients in group A, in one of five (20%) in group B1 and in zero of nine in group B2 (P = 0.009).

Nucleotide substitutions in the core promoter and precore regions

The double A1762T/G1764A CP mutation was present in seven of eight (87%) patients from group A, in one of five (20%) in group B1 (P = 0.03) and in one of nine (11%) in group B2 (P = 0.002). The T1753C mutation was detected in seven of eight patients (87%) in group A and in none from group B1 (P = 0.004) or group B2 (P = 0.0004). Furthermore, the G1896A PC stop codon mutation was detected in seven of eight (87%) patients in group A, in zero of five patients in group B1 (P = 0.004) and in one of nine (11%) in group B2 (P = 0.002).

Follow-up.  Two of eight patients in group A stopped lamivudine after 6 months of reactivation and ALT values dropped soon after treatment withdrawal. In one of these, the M204V and L180M mutations disappeared, but the other surface/polymerase and core promoter/PC/core mutations persisted. The second patient, who during reactivation had only the M204I substitution without the L180M, maintained the same mutational pattern.

Three others in group A switched to adefovir and the ALT levels normalized. Sequencing data are unknown because sera were not available.

Finally, the last three patients continued lamivudine therapy for another 8 months despite reactivation; after stopping treatment they still had mild ALT elevations and the mutational pattern was unmodified. Groups B1 and B2 patients maintained normal ALT and HBV-DNA positivity during the same follow-up period.


Long-term lamivudine treatment in patients with chronic hepatitis B can lead to the emergence of M204V/I and L180M mutations in the polymerase region of HBV, which are responsible for drug resistance. The selection of these mutants is often responsible for mild clinical reactivation and occasionally for severe hepatitis flares. However, the emergence of these mutants does not always correlate with liver disease recurrence.2, 8 In the present study, sequence analysis revealed a distinct mutational pattern in the groups of patients treated with lamivudine who presented with a clinical reactivation during therapy and those without.

All patients with ALT elevation had the M204V/I mutation, whilst the L180M mutation was present in 50% of them. In contrast, none of the anti-HBe-positive patients (group B1) and 44% of the HBeAg-positive patients (group B2) with normal ALT had the M204V/I substitution. This is in accordance with several previous reports and confirms the role of these polymerase mutations in the emergence of lamivudine resistance.12 However, in the same region, we found additional mutations that seem to differ significantly among the groups of patients. The Y135S change, in particular, was always present in patients with flares while L122F, P130Q and Y135F were predominant in patients with no clinical reactivation; these changes are in the A–B interdomain region of the polymerase. As a consequence of these sequence changes, a T127P substitution appeared in the ‘a’ determinant in the overlapping HBsAg region, in all cases with exacerbation, while T127S and Y134N were found in the majority of those with normal ALT. Furthermore, the I195M and W196L mutations emerged as a result of the M204V/I mutations in the overlapping region. Lamivudine M204V/I mutant strains are replication defective and the association with the L180M change appears to restore replication competence, although at a lower efficiency than the wild type. Several in vitro and in vivo studies have shown that additional mutations occur during lamivudine treatment in the polymerase and surface regions that can act as compensatory mutations and restore viral replication fitness.9, 15, 16, 24 This correlates frequently with a clinical reactivation and in some cases with severe progression of disease. In the present study, we observed two polymerase/surface patterns of additional mutations between the groups of patients with a clinical reactivation and those without; interestingly the patients with flares also had increased HBV-DNA levels. These mutations are located in the ‘A–B interdomain’ region of the polymerase that overlaps with the major hydrophilic region of HBsAg spanning amino acids 99–169, and more specifically loop 1 of the ‘a’ determinant.25 Changes in this region relate to subtype variations, but have also been observed in the vaccinated, in patients infected with serologically negative virus and after treatment with monoclonal antibody or hepatitis B immunoglobulin (HBIG) therapy.25 In the present study, all patients were HBsAg reactive with the commercial assays used.

The HBV CP and PC sequence analysis revealed that almost all patients with clinical exacerbation under lamivudine treatment harboured mutations in these regions. Specifically, the T1753C and the double CP A1762T/G1764A substitutions were both present in 87% of patients with flares vs. only 14% of those with normal transaminases, and the G1896A PC stop mutation was observed in 87% vs. 7%, respectively. Accordingly, Lok et al. reported that the double CP A1762T/G1764A mutation was more often associated with the selection of lamivudine-resistant mutants in a group of HBeAg-negative patients.20 Chen et al. have clearly demonstrated by a replication competent baculovirus system based on genotype D that the G1896A PC stop codon mutation increased the replication efficacy of lamivudine-resistant viruses;21 our patients were all genotype D. Tacke et al. confirmed the above viral replication restoring effect of the G1896A mutant in lamivudine-resistant HBV genotype A and showed that the double CP A1762T/G1764A mutation further increases replicative efficiency.26 Our results support these in vitro data and confirm that the PC and the basic CP mutations have an important role in restoring viral fitness of the lamivudine-resistant mutants. However, it is of interest that four HBeAg-positive patients had no clinical reactivation despite the M204V/I substitutions. These patients did not have the L180M, PC and CP mutations, and had viral reactivation with increased HBV-DNA levels but no ALT flares.

In the present study, a threonine to serine substitution at position 12 in the core region was observed in 62% of group A vs. 20% of group B1 and none of group B2 patients. This mutation, clustering in the T-helper cell epitope, has been reported to emerge in anti-HBe/HBV-DNA-positive patients with an active disease.27 Furthermore, E77D and P79Q, clustering in a B-cell epitope, and Y38F were present in 25%, 50% and 37%, respectively, of cases with clinical reactivation and in none of those with no flares. Although these differences are not statistically significant it is reasonable to argue that changes occurring in T- and B-cell epitopes, together with other mutations responsible for viral reactivation, may be associated with a clinical reactivation.

Despite the selection of lamivudine-resistant mutants, the clinical reactivation in our patients was not severe; the transaminase increases were not associated with liver disease decompensation. Three patients switched to adefovir treatment and three continued lamivudine for another 8 months with no worsening of disease. Interestingly, the remaining two patients stopped lamivudine and ALT dropped to normal values. Severe clinical exacerbation and, in some cases, fulminant hepatitis have been reported after discontinuation of lamivudine.28, 29 However, recent studies show that severe reactivation is more frequent in patients with more advanced liver disease and in patients with long-standing lamivudine-resistant mutations;20, 30 no patient in the present study had advanced liver disease. It is reasonable to argue that the continuation of lamivudine after the selection of drug-resistant mutants can lead in time to the emergence of compensatory mutations and the constitution of a more aggressive quasispecies swarm with higher replicative capacity.

In conclusion, clinical exacerbation during lamivudine treatment seems to be related to the selection of specific genetic mutational patterns that confer a replicative advantage to the virus, leading to increased viraemia. The detection of these mutants and HBV-DNA quantification may permit early identification of flares because of lamivudine resistance. However, further in vitro and in vivo longitudinal studies are needed to functionally characterize mutant viral strains and determine their clinical implications.


Authors thank Dr Loredana Costagliola and Mrs Geltrude Fiorillo for technical assistance.