Successful therapy of hepatitis B with tenofovir in HIV-infected patients failing previous adefovir and lamivudine treatment

Three HIV-infected patients with chronic hepatitis B (genotype A) were switched to adefovir therapy after unsuccessful lamivudine treatment. Surprisingly, adefovir therapy failed, although none of the virus isolates displayed mutations known to be associated with adefovir resistance (A181V, N236T). In two isolates we identified hepatitis B virus DNA polymerase mutation L217R, in one case we found multiple frameshifts in the same region. In all cases adefovir was replaced by tenofovir, resulting in a significant drop in the viral load. Current therapies for hepatitis B virus (HBV) infections are based on the application of interferon, lamivudine or combinations of both drugs. For lamivudine, the development of resistant HBV strains has been observed in 25% of treated patients per year. Moreover, in HBV/HIV co-infected patients the development of lamivudine resistance is more frequent than in HBV mono-infected patients, making a therapy alternative to lamivudine indispensable [1–4]. Therefore, the new drug adefovir serves as an alternative therapy for the treatment of chronic HBV infection. Adefovir passed clinical studies successfully [5], and so far two mutations mediating resistance to adefovir have been described [6,7]. The frequency of the mutation A181V is approximately 2.5% per year with hitherto unkown relevance, whereas 1.7–2.5% per year of adefovir-treated patients carry the resistance mutation N236T [6,7]. Moreover, adefovir is also active against HIV, but HIV-DNA polymerase mutations selected during treatment with adefovir did not influence the sustained response in viral load for 6–12 months [8]. Studies by Perrillo et al. [9] and Peters et al. [10] demonstrated that 8–15% of patients infected with lamivudine-resistant HBV exhibit initial non-response to adefovir. Unfortunately, the reasons for adefovir non-response in those studies remained unclear. In line with those observations, we detected three out of 20 patients chronically infected with HBV and co-infected with HIV, who despite good compliance did not respond to adefovir dipivoxil with a decrease in HBV DNA. Patient 1 was a 39-year-old white man with a long-term HIV infection [Centers for Disease Control and Prevention (CDC) stage B2] and chronic hepatitis B e antigen-positive hepatitis B. Antiretroviral therapy (ART) was started in 1996; current ART consisted of lamivudine/abacavir/efavirenz with an HIV viral load below 25 copies/ml and good immunological recovery (CD4 cell count 672 cells/μl; 30%). Liver function tests (LFT) fluctuated but never exceeded twice the upper limit of normal (alanine aminotransferase; ALT). On ultrasound examination the liver appeared normal. Because of a high HBV load while receiving lamivudine-containing ART, adefovir treatment was started and was continued for 6 months without any response in the HBV load (Fig. 1). Adefovir was replaced by tenofovir, resulting in a significant drop in the HBV load after 4 months and maintained virological control of HIV infection.Fig. 1.: Course of hepatitis B viral load in three chronically HIV/hepatitis B virus-infected patients (adefovir-non- responders) during antiretroviral therapy; antiretroviral therapy was initially combined with adefovir and was switched to combination with tenofovir after no significant decrease of hepatitis B viral load was observed as a result of adefovir. All patients share the genotype A and identical lamivudine resistance patterns (L180M/YVDD). Sequencing analysis of the isolates revealed mutations at amino acid 217 (patients 1 and 3) or 215–226 (patient 2) (see main text). ABC, Abacavir; ADF, adefovir; ANR, adefovir non-responders; BL, baseline; D4T, stavudine; EFV, efavirenz; IDV, indinavir; TDF, tenofovir; 3TC, lamivudine; ZDV, zidovudine.Patient 2 was a 34-year-old white man with chronic hepatitis B e antigen-positive hepatitis B and HIV infection diagnosed on the occasion of a Pneumocystis jeroveci pneumonia and cryptosporidial diarrhea (CDC stage C3). ART was started initially with zidovudine/lamivudine/indinavir/ritonavir and subsequently with zidovudine/lamivudine/efavirenz, with good virological and immunological response (HIV viral load less than 25 copies/ml, CD4 cell count 487 cells/μl, 21%). Initially on treatment with lamivudine, LFT were normal/minimally elevated and the HBV load was 1000 particles/ml, but increased 2 years after the initiation of lamivudine therapy in parallel with LFT (ALT up to four times the upper limit of normal). On ultrasound examination the liver was moderately enlarged. Adefovir was started, but no decrease in the HBV load or LFT was observed over 5 months (Fig. 1). ART was changed to tenofovir/lamivudine/efavoremz. Two months later the hepatitis B viral load and ALT decreased significantly. The third patient was a 37-year-old white man with HIV infection. ART started with didanosine/stavudine/ritonavir when the patient became symptomatic from HIV infection (CDC stage B2), with good virological and immunological response. During ART the patient acquired acute hepatitis B infection and ART had to be interrupted. Later, ART with zidovudine/lamivudine/efavirenz and IFN-α (12 months) was started. Six months later zidovudine was substituted by stavudine because of neutropenia. During this ART, HBV-DNA became negative, whereas hepatitis B surface and e antigen remained positive. However, HBV-DNA rebounded, with the patient still being on a lamivudine-containing ART with an HIV load below 25 copies/ml. LFT (ALT) ranged between twice and five times the upper limit of normal. On ultrasound examination, the spleen and liver were enlarged, with increased echogenicity of the liver. Although adefovir therapy was started, no decrease in the hepatitis B viral load was observed after 3 months (Fig. 1). Under treatment with tenofovir the HBV load dropped by more than 3 log. ALT levels decreased in all three patients after tenofovir treatment was started (data not shown). Adefovir was useless in all of the three patients. Sequencing analysis [11] revealed no amino acid exchange at position 236 nor at position 181, the position described to mediate resitance to adefovir [6,7]. All patients shared the same HBV genotype (genotype A) with the identical genotypical lamivudine resistance pattern (YVDD, L180M) as determined by line probe assay (Inno-LiPA HBV–DR, Innogenetics, Gent, Belgium) upon baseline. Sequencing analysis of the isolates revealed mutations at amino acid 217 (patients 1 and 3) or 215–226 (patient 2). Sequences obtained from sera collected immediatly before the initiation of adefovir therapy showed that patients 1 and 3 displayed the mutation before the treatment started, whereas patient 2 developed the domain exchange during ongoing adefovir therapy. Two further patients not described here included in a tenofovir study exhibited the L217R mutation, and responded well to tenofovir. Four control HBV strains responding to adefovir sequenced in our laboratory exhibited wild-type sequence L217. No stop codons but single amino acid exchanges were introduced in the overlapping hepatitis B surface open-reading-frame by the polymerase gene mutations. Based on GenBank data, the hepatitis B surface amino acid exchanges did not have a significant impact on the outcome of the hepatitis. The observation that differences in genotypic pattern in HCV and HIV correlate with subsequent virological response to treatment [12,13] in combination with our data suggest similar phenomena for HBV, of which its polymerase shares high homology with the HIV polymerase. On the basis of our observations we (carefully) conclude that: (1) the polymerase domain amino acids 215–226 might mediate adefovir resistance; (2) the combination of the lamivudine resistance pattern L180M/YVDD plus HBV genotype A may predetermine resistance to adefovir; and most important (3) in analogy to earlier observations [2–4], therapy changes to tenofovir should be considered for adefovir non-responders. Although an in-vitro assay confirming the hypothesis that the amino acid region 215–226 might be responsible for resistance is missing, this assumption is confirmed by the recent patent application of Bartholomeusz et al. [14]. In summary, at this stage, we recommend tenofovir for the treatment of adefovir-non-responders, probably also in HBV-mono-infected patients.