Deep sequencing analysis of hepatitis B virus in patients with incomplete response to tenofovir alafenamide fumarate treatment

Abstract

BACKGROUND

Tenofovir alafenamide fumarate (TAF) is one of the first-line treatments used to treat chronic hepatitis B patients; TAF has strong antiviral activity and a high barrier to resistance. Although virological breakthroughs in patients during TAF treatment are rare, patients with incomplete responses to TAF are occasionally observed.

AIM

To investigate responsible mutations in the reverse transcriptase region of hepatitis B virus (HBV) for TAF-incomplete responses.

METHODS

Thirteen chronic hepatitis B patients who received TAF monotherapy were included. A TAF-incomplete responder was defined as one who was continuously positive for HBV DNA over 2 years after TAF treatment initiation. The emergences of mutations in TAF-incomplete responders were evaluated before, one year after, and two years after treatment by deep sequencing of HBV DNA and RNA.

RESULTS

Two patients were continuously positive for HBV DNA over two years. The rtL269I mutation, one of the CYEI mutations linked to tenofovir resistance, was detected in both patients by direct sequencing. The deep sequencing analysis revealed that a combination of rtT118A and rtL220I mutations and the rtL269I mutation were predominantly detected in HBV DNA even when these mutations were barely detected in HBV RNA. This suggests a superior replication capability of the HBV variants with these mutations under TAF treatment.

CONCLUSION

The deep sequencing analysis of HBV DNA and RNA and comparing the detection rates of mutations were useful for estimating responsible mutations for TAF-incomplete responses. Such analysis is needed to evaluate the association between mutations that emerge during TAF treatment and incomplete responses to TAF.

Key Words: Tenofovir; Tenofovir alafenamide fumarate; Tenofovir disoproxil fumarate; Resistance; Mutation; Deep sequence

Core Tip: Tenofovir alafenamide fumarate (TAF) is a potent treatment for chronic hepatitis B with low resistance rates. In this study, two TAF-incomplete responders were investigated. In these patients, mutations of rtL269I and combinations of rtT118A/rtL220I were detected in hepatitis B virus (HBV) DNA predominantly as associated with incomplete response to TAF treatment. The deep sequencing of infected HBV DNA and RNA effectively identified responsible mutations, emphasizing its utility in understanding the association between emerged mutations and incomplete responses to TAF and guiding treatment strategies for incomplete responders.

INTRODUCTION

Hepatitis B virus (HBV) is a major causative agent of chronic hepatitis. HBV infection has spread worldwide, and approximately 300 million patients are chronically infected with HBV[1]. In chronic hepatitis B patients, liver inflammation leads to cirrhosis and hepatocellular carcinoma (HCC) development[2]. Currently, two kinds of anti-HBV agents are available for use in the clinic: Interferon (IFN)-α and nucleoside/nucleotide analogs (NAs)[3]. IFN-α inhibits HBV at multiple points in its life cycle by exerting direct antiviral and immunomodulatory effects. However, IFN-α is associated with a limited response rate and undesirable side effects[4]. NAs exert potent antiviral effects by targeting reverse transcription and inhibiting HBV DNA production from pregenomic RNA (pgRNA). Although NAs ameliorate liver diseases, prevent disease progression to liver fibrosis, and inhibit HCC development, they cannot eliminate the covalently closed circular DNA of HBV in infected hepatocytes[5]. Therefore, chronic hepatitis B patients require lifelong NA administration, and long-term NA treatment occasionally leads to the emergence of resistance-associated mutations (RAMs) in the infecting HBV[6]. Among the available NAs, tenofovir disoproxil fumarate (TDF) and tenofovir alafenamide fumarate (TAF) are used as first-line treatments for chronic hepatitis B. Both are prodrugs of tenofovir and are phosphorylated to form the active metabolite tenofovir diphosphate. These compounds are known to have strong antiviral activity with a high barrier to resistance. Although the emergence of RAMs during treatment with these NAs is rare, RAMs have been reported in a few patients[7-21]. First, the amino acid substitution of alanine to threonine at amino acid (aa) 194 in the reverse transcriptase (RT) region (rtA194T) was observed in two HBV/human immunodeficiency virus-coinfected patients[7]. Then, two quadruple mutations were reported to be associated with viral breakthroughs during TDF treatment. One quadruple mutation, including serine to cysteine at aa106 (rtS106C), histidine to tyrosine at aa126 (rtH126Y), aspartic acid to glutamic acid at aa134 (rtD134E), and leucine to isoleucine at aa269 (rtL269I), was detected in two cases of viral breakthroughs during TDF treatment[17]. These mutations were designated the CYEI mutations. The other quadruple mutation was a combination of leucine to methionine at aa180 (rtL180M), threonine to leucine at aa184 (rtT184 L), alanine to valine at aa200 (rtA200V), and methionine to valine at aa204 (rtM204V)[18]. These mutations were detected in a patient who experienced virological breakthrough during TDF treatment and were designated MLVV mutations. These RAMs have been shown to increase IC₅₀ values and reduce sensitivity to tenofovir in vitro. However, other groups have reported contradictory data and have failed to confirm the contributions of these RAMs to tenofovir resistance[22-24]. Therefore, the impact of these RAMs on sensitivity to tenofovir remains controversial. In this study, we evaluated the emergence of amino acid mutations in two patients with incomplete responses to TAF via direct and deep sequencing and investigated the associations between the detected mutations and incomplete responses to TAF.

MATERIALS AND METHODS

Patients

Thirteen chronic hepatitis B patients (male/female ratio: 8/5; median age: 40 years) who began TAF treatment at Seizankai Kiyokawa Hospital (Tokyo, Japan) from March 2018 to October 2018 were included (Table 1). These patients were clinically diagnosed with chronic hepatitis and were positive for hepatitis B surface antigen (HBsAg) and HBV DNA for at least 6 months prior to the initiation of treatment. None of the patients were infected with hepatitis C virus or diagnosed with cirrhosis or HCC. All the patients received 25 mg of TAF for more than one year but no other NA treatment. This study was approved by the Ethics Committees of our institutions (approval numbers: 0167 and 1081 from Seizankai Kiyokawa Hospital and the National Institute of Infectious Diseases, respectively), and written informed consent was obtained from each patient.

Table 1 Baseline characteristics of patients.

Patients	n = 13
Sex (male/female)	8/5
Age (year), median (range)	40 (30–62)
Genotype (A/B/C)	1/0/12
HBsAg (IU/mL), median (range)	5037.79 (605.77-139448.20)
Positive for HBeAg	8 (61.5%)
ALT (U/L), median (range)	64 (24–591)
HBV DNA (log IU/mL), median (range)	6.6 (1.6–9.1)

ALT: Alanine aminotransferase; HBsAg: Hepatitis B surface antigen; HBeAg: Hepatitis B e antigen; HBV: Hepatitis B virus.

Virological markers of HBV

HBsAg and hepatitis B e antigen (HBeAg) levels were measured via a chemiluminescent immunoassay with a commercial assay kit (Abbott Japan Co., Ltd., Tokyo, Japan). The HBV genotype was determined via an enzyme immunoassay (IMMUNIS HBV Genotype EIA, Institute of Immunology Co., Ltd., Tokyo, Japan). HBV DNA titers were determined with the Cobas 6800/8800 system HBV (Roche Diagnostics K.K., Tokyo, Japan); the lower limit of detection of this assay was 10 IU/mL[25].

Direct sequencing of HBV

The patient serum was collected at the time points indicated in Figure 1, and total DNA was isolated from 200 µL of patient serum with the QIAamp Blood Mini Kit (Qiagen, Hilden, Germany). The entire genome of HBV was amplified in two fragments, A and B, with the following primers: For the 1st round PCR of fragment A: Forward: 5’-ATTCCACCAAGCTCTGCTAGATCCCAGAGT-3’ and reverse: 5’-GGTGCTGGTGAACAGACCAATTTATGCCTA-3’; for the 2nd round PCR of fragment A: Forward: 5’-CCTATATCTTCCTGCTGGTGGCTCCAGTTC-3’ and reverse: 5’-TAACCTAAT CTCCTCCCCCAACTCCTCCCA-3’; for the 1st round PCR of fragment B: Forward: 5’-ACGTCGCATGGAGACCACCGTGAACGCCCA-3’ and reverse: 5’-AAGTCCACCACGAGTCTAGACTCTGTGGTA-3’; and for the 2nd round PCR of fragment B: Forward, 5’-CATGGTCTTGCC CAAGGTCTTGCATAAGAG-3’ and reverse, 5’-CCCGCCTGTAACACGA GCAGG GGTCCTAGG-3’). The amplified products were sequenced with primers that were used in the 2^nd round PCR.

Figure 1 Clinical courses of two tenofovir alafenamide fumarate-incomplete responders. Serum alanine aminotransferase levels (open circles), hepatitis B virus (HBV) DNA titers (open squares), HBV RNA titers (closed bars), and status of HBeAg are indicated. The lower limit of detection of HBV DNA is indicated by the dotted line. The serum samples were collected at the time points indicated by open arrows with numbers. ALT: Alanine aminotransferase; HBV: Hepatitis B virus; TAF: Tenofovir alafenamide fumarate; HBeAg: Hepatitis B e antigen.

HBV RNA extraction and quantification

Total RNA was isolated from 140 µL of patient serum with the QIAamp Viral RNA Mini Kit (Qiagen). The extracted RNA was further treated with TURBO DNase (Thermo Fisher Scientific, Waltham, MA) and purified with RNeasy Mini Kit (Qiagen) combined with on-column DNase digestion (RNase-Free DNase; Qiagen). cDNA was synthesized from HBV RNA with a Superscript VILO cDNA Synthesis Kit (Thermo Fisher Scientific). HBV RNA levels were quantified by real-time PCR with a primer and probe set that targets the HBs region, as previously described[26].

Deep sequencing of the HBV RT region

The extracted HBV DNA and the cDNA that was synthesized from HBV RNA were amplified via PCR with primers that covered the RT region[27]. The deep sequencing of amplicons was conducted via the Illumina MiSeq platform (Illumina, San Diego, CA). The amplified and analyzed region was aa445-aa625 in the polymerase (aa98-aa279 in the RT domain).

RESULTS

Patients with incomplete responses to TAF treatment

A TAF-incomplete responder was defined as one who was continuously positive for HBV DNA over 2 years after TAF treatment initiation. Among the 13 chronic hepatitis B patients who were treated with TAF, 2 exhibited continuous positivity for HBV DNA even after 2 years of TAF administration. Both patients were infected with HBV genotype C and positive for HBsAg and HBeAg (Table 2). For both patients, good adherence to TAF treatment, as well as the absence of any concomitant medications or surgical history, was confirmed. The HBV DNA level in each patient was 9.1 log IU/mL at the start of TAF treatment. These patients exhibited a partial virological response, and viral breakthrough was not observed throughout the clinical course (Figure 1). For patient B, the alanine aminotransferase (ALT) value increased and then gradually decreased 1 year after the initiation of TAF treatment. An increase in the HBV DNA level did not accompany this increase in the ALT level.

Table 2 Profile of two tenofovir alafenamide fumarate-incomplete responders.

Patients	Patient A	Patient B
Genotype	C	C
HBsAg (IU/mL)1	139448.20	62210.94
HBeAg	Positive	Positive
ALT (U/L)1	52	47
HBV DNA (log IU/mL)1	9.1	9.1
Direct sequencing of infected HBV1
BCP (nt1762; A:T/nt1764; G:A)	T/G	T/A
Precore stop mutation (nt1896; G:A)	G	G

¹Values at the initiation of tenofovir alafenamide fumarate treatment.

ALT: Alanine aminotransferase; HBsAg: Hepatitis B surface antigen; HBeAg: Hepatitis B e antigen; HBV: Hepatitis B virus; BCP: Basal core promoter.

Mutations detected by direct sequencing in TAF-incomplete responders

HBV was isolated from serum samples that were collected from two TAF-incomplete responders at three time points: At the beginning of treatment (time point 1), 1 year after the initiation of TAF treatment (time point 2), and 2 years after the initiation of TAF treatment (time point 3; Figure 1). The entire genomes of the isolated HBVs were directly sequenced, and the signature amino acid polymorphisms in the core promoter and the precore were evaluated. In Patients A and B, the amino acid polymorphisms in the core promoter (nt1762/nt1764) were T/G and T/A, respectively. The amino acid polymorphism in the precore (nt1896) was the wild-type in both patients (Table 2). The amino acids that were detected in the RT region were also compared with the reference sequence of the genotype C strain (Figure 2). Consequently, in patient A, the rtL269I, which is one of the CYEI mutation, was detected at time point 3. In Patient B, the mutations of threonine to alanine at aa118 (rtT118A) and leucine to isoleucine at aa220 (rtL220I), which have not yet been reported to be associated with NA resistance, were detected before the initiation of TAF treatment (time point 1), were still detected at time point 2, but were not observed at time point 3. Instead, the rtL269I was detected at that time point. The detected amino acid mutations of the RT region in these patients were not associated with mutations in HBsAg.

Figure 2 Alignment of reverse transcriptase regions of hepatitis B virus strains detected in tenofovir alafenamide fumarate-incomplete responders. The reverse transcriptase regions of the hepatitis B virus strains detected in tenofovir alafenamide fumarate (TAF)-incomplete responders before treatment (pre), 1 year after TAF treatment (1 year), and 2 years after TAF treatment (2 year) were directly sequenced, and deduced amino acids were aligned with the reference sequence of genotype C (accession number: GQ872210). The inverted triangle indicates the positions of the reported resistance-associated mutations to tenofovir disoproxil fumarate and TAF. A box with a dotted line indicates the positions of the mutations detected in this study. RT: Reverse transcriptase.

Mutations detected by deep sequencing in TAF-incomplete responders

To evaluate fluctuations in the detection rates of mutations that were detected in direct sequencing, deep sequencing analysis of HBV DNA and RNA in samples from the two TAF-incomplete responders was performed. HBV RNA was extracted from the patient’s serum at three time points. The HBV RNA titers were quantified via real-time PCR after the extracted RNA was subjected to DNase treatment. In both patients, the HBV RNA titer was approximately 7 log copies/mL at the start of TAF treatment, and the titer remained at nearly similar levels throughout the clinical course (Figure 1). According to deep sequencing analysis of patient A, fluctuations in the rates of mutation detection were observed only for rtL269I (Figure 3). At the start of TAF treatment, the wild-type sequence (leucine; L) at aa269 (rtL269wt) was predominant in HBV DNA, although the rate of rtL269I detection in HBV RNA was nearly half. After the initiation of TAF treatment, the rate of rtL269I detection temporarily decreased to almost one-fourth at time point 2 but then increased to 100% at time point 3. In contrast, analysis of HBV RNA revealed that the rate of rtL269I detection remained at half throughout the observation period. For patient B, deep sequencing analysis showed fluctuations in the detection rates for three amino acids, namely, rtT118A, rtL220I, and rtL269I, in both HBV DNA and RNA. At the initiation of TAF treatment, the rtT118A and rtL220I mutations were detected in HBV DNA and were predominant, while the detection rates of this mutation in HBV RNA were nearly half. At time point 2, the rates of rtT118A and rtL220I mutation detection in HBV DNA reached 100%. In contrast, these mutations were scarcely detected in HBV RNA. At time point 3, rtT118A and rtL220I mutations disappeared in HBV DNA, although these mutations remained detectable in HBV RNA. Instead, another mutation previously undetected in patient B, rtL269I, was detected in HBV DNA, accounting for 99%, but it was not detected in HBV RNA. As a control, the two responders to TAF were randomly selected, and deep sequencing analysis of HBV DNA in samples before TAF treatment was performed. The mutations of rtT118A, rtL220I, and rtL269I were not detected in both patients (data not shown).

Figure 3 Deep sequencing data of serum hepatitis B virus DNA and RNA from tenofovir alafenamide fumarate-incomplete responders. The amino acid mutations that changed in the clinical course of each patient are shown. The time points at which the serum samples were collected were the same as those indicated in Figure 1. The detection rates at which mutations were detected in hepatitis B virus DNA and RNA and the number of total reads are indicated. HBV: Hepatitis B virus.

DISCUSSION

NAs are commonly used in the treatment of chronic hepatitis B patients, and the administration of NAs has potent anti-HBV effects. TDF, which is an NA, is an ester prodrug of tenofovir and is used to treat chronic hepatitis B patients. TDF has strong anti-HBV effects with a low emergence rate of drug-resistant HBV, but there is concern that TDF is associated with reduced renal function and bone density. Owing to these safety issues, TAF, which is a phosphonate prodrug of tenofovir, was developed and used for the treatment of chronic hepatitis B patients[28]. The pharmacological characteristics of TAF allow a reduction in dosage, resulting in bone and renal safety, although TAF maintains high virological efficacy and a low incidence of virological breakthrough[13]. In Japan, TAF has been used for the treatment of chronic hepatitis B since 2017 with few cases of resistance reported. However, patients who exhibit HBV DNA continuously positive even after 2 year of TAF treatment are occasionally observed. In this study, we investigated 2 TAF-incomplete responders and detected responsible mutations.

In two TAF-incomplete responders the amino acid mutation rtL269I, which is one of the CYEI mutations reported to be involved in tenofovir resistance[17], was detected after 2 years of TAF treatment via direct sequencing. Other mutations, such as those included in MLVV mutations[18], were not detected in these patients. Interestingly, the rtL269I mutation was one of the RAMs that we previously reported to have emerged in a patient treated with entecavir[29]. Although the association between rtL269I and incomplete responses to TAF treatment was suspected, it was not apparent in the direct sequencing data. Therefore, to indicate the involvement of this mutation in incomplete responses, we used deep sequencing analysis and compared the detection rates of mutations in DNA and RNA of infecting HBV.

In patient A, rtL269I was already present in HBV DNA as a minor population at the initiation of TAF treatment. The rate of rtL269I detection in HBV DNA fluctuated during TAF treatment and reached 100% after 2 years of TAF treatment. In contrast, the rate of rtL269I detection in HBV RNA remained nearly half throughout the observation period. The HBV RNA detected in patient sera is primarily capsidized pgRNA, and the HBV polymerase included in the capsid is considered to have amino acids that correspond to the sequence of the pgRNA, as demonstrated by a recent study that examined mechanisms for cis-preferential reverse transcription[30]. Therefore, from these data, in HBV-infected hepatocytes, HBV RNA with the rtL269wt and rtL269I mutations was presumed to be present and encapsidated in almost equal amounts. However, only the DNA of HBV with rtL269I was synthesized via RT. The predominance of HBV DNA with rtL269I mutation, despite the presence of almost equal amount of wild-type and mutated pgRNA, indicates that the HBV polymerase with rtL269I mutation functions more efficiently than the wild-type under TAF treatment. These observations indicate that HBV strains with rtL269I overcome the TAF-mediated inhibition of the RT reaction.

In patient B, HBV strains with rtT118A and rtL220I were predominant at the initiation of TAF treatment. The rates of rtT118A and rtL220I detection in HBV RNA were nearly half at this time point. One year after TAF treatment, HBV strains with both rtT118A and rtL220I mutations accounted for 100% of the HBV DNA, but strains with these mutations were barely detectable in the HBV RNA. Thus, at this time point, the HBV variants with rtT118A and rtL220I mutations successfully synthesized HBV DNA even in the presence of TAF, and the HBV strains with these mutations were revealed to be associated with the incomplete responses to TAF. The increase in ALT values may be related to a change in the predominant strain to one with these mutations. After two years of TAF treatment, the HBV strains with rtT118A and rtL220I mutations became almost undetectable in HBV DNA, and another HBV strain with the rtL269I mutation emerged in HBV DNA but was not detected in HBV RNA. These data indicate that the HBV strain with the rtL269I mutation also has some advantages in incomplete responses to TAF or HBV replication efficiency, overcoming the population of HBV strains with rtT118A and rtL220I mutations. Taken together, deep sequencing analysis and comparison of the detection rates of mutations in HBV DNA and RNA revealed that two sets of mutations of rtT118A + rtL220I and rtL269I are associated with incomplete responses to TAF. Furthermore, the changes of the predominant strain in HBV DNA to another strain suggest that these mutations are involved in incomplete responses to TAF. Similar observations of the differences in the detection rates of mutations between HBV DNA and RNA have been shown in patients treated with lamivudine or entecavir[31].

Previous reports have shown that complete viral responses without viremia relapse for 48 weeks were achieved and maintained by TAF monotherapy even in patients infected with HBV with the rtL269I mutation[27]. These data indicate that HBV with the rtL269I mutation is susceptible to TAF, which may be inconsistent with our observations. One explanation for this discrepancy is that mutation-mediated resistance to TAF is HBV strain-dependent. It has been reported that the effects of introducing the rtA181V and rtN236T mutations on susceptibility to tenofovir and adefovir depend on the strain of HBV, even within the same genotype[32]. Besides, we previously reported that the emergence of the rtL180M and rtM204V mutations, known as RAMs to lamivudine, confers resistance to entecavir in an HBV strain-dependent manner[29]. Thus, HBV sequence diversity may cause different resistance profiles to NAs even if it is outside the RT region. Another possibility is the host factors of the patient who exhibited incomplete responses to TAF. The patients in our study may have some disadvantages in terms of host factors associated with suppressing HBV DNA replication, such as IFN signaling. From these observations, even mutations that are characterized as not being associated with resistance to NAs require clinical attention because they may be associated with resistance or treatment refractoriness, depending on the HBV strain that carries the mutation and/or the host that is infected with HBV. The impacts of HBV strains and host factors on differences in resistance to NA via RAMs need to be further investigated.

Importantly, neither of these two patients experienced virological breakthroughs during their clinical course. Although several tenofovir-refractory cases have been reported, the number of cases accompanied by virological breakthroughs is small[19]. These data may indicate that mutations that are associated with incomplete responses to TAF including those identified in our experience, do not confer enough replication efficiency to cause virological breakthroughs. However, depending on a patient's condition or genetic background, HBV strains with mutations that were previously shown to be unrelated to virologic breakthroughs may cause an increase in HBV titers. Additionally, even if an emerged mutation alone is not sufficient to cause a virologic breakthrough, the addition of other mutations may cause a virologic breakthrough. Moreover, prolonged infection due to an incomplete response to TAF may affect clinical outcomes, such as the progression of fibrosis or the development of HCC. Therefore, it will be necessary to identify mutations associated with incomplete responses to NAs via the methods used in this study and collect more information on the RAMs. At the same time, we cannot rule out the possibility that the difference in detection rates between DNA and RNA mutations may result from other strain- or mutation-dependent factors, such as bias in reverse transcription efficiency or RNA encapsidation, selection pressure after reverse transcription, or difference in degradation. Further investigation will be necessary to clarify this point in the future.

CONCLUSION

We detected two sets of mutations that are associated with incomplete responses to TAF, a combination of rtT118A and rtL220I mutations and the rtL269I mutation. The association of these mutations to incomplete responses to TAF was demonstrated by comparing the detection rates of mutations in HBV DNA and RNA, and the involvement of these mutations in incomplete response to TAF treatment was indicated by the changes of the predominant strain in HBV DNA to another strain in the clinical course. The strategy used in this study will be useful in detecting mutations responsible for NA-refractory or incomplete response and is needed to evaluate the association between mutations that emerge during treatment with NAs and drug resistance.