Ultrasensitive Amplification Refractory Mutation System Real-Time PCR (ARMS RT-PCR) Assay for Detection of Minority Hepatitis B Virus-Resistant Strains in the Era of Personalized Medicine

Abstract

Resistance to antiviral treatment for chronic hepatitis B virus (HBV) has been associated with mutations in the HBV polymerase region. This study aimed at developing an ultrasensitive method for quantifying viral populations with all major HBV resistance-associated mutations, combining the amplification refractory mutation system real-time PCR (ARMS RT-PCR) with a molecular beacon using a LightCycler. The discriminatory ability of this method, the ARMS RT-PCR with molecular beacon assay, was 0.01 to 0.25% for the different HBV resistance-associated mutations, as determined by laboratory-synthesized wild-type (WT) and mutant (Mut) target sequences. The assay showed 100% sensitivity for the detection of mutant variants A181V, T184A, and N236T in samples from 41 chronically HBV-infected patients under antiviral therapy who had developed resistance-associated mutations detected by direct PCR Sanger sequencing. The ratio of mutant to wild-type viral populations (the Mut/WT ratio) was >1% in 38 (63.3%) of 60 samples from chronically HBV-infected nucleos(t)ide analogue-naive patients; combinations of mutations were also detected in half of these samples. The ARMS RT-PCR with molecular beacon assay achieved high sensitivity and discriminatory ability compared to the gold standard of direct PCR Sanger sequencing in identifying resistant viral populations in chronically HBV-infected patients receiving antiviral therapy. Apart from the dominant clones, other quasispecies were also quantified. In samples from chronically HBV-infected nucleos(t)ide analogue-naive patients, the assay proved to be a useful tool in detecting minor variant populations before the initiation of the treatment. These observations need further evaluation with prospective studies before they can be implemented in daily practice.

INTRODUCTION

Nucleos(t)ide analogues are a fundamental tool for the treatment of hepatitis B virus (HBV)-associated liver disease because of their potent antiviral activity in the absence of remarkable side effects and major contraindications (1). Long-term nucleos(t)ide analog therapy among patients significantly improves survival and reduces the risk of liver-related major complications, such as death, hepatocellular carcinoma (HCC), and liver decompensation (2). A major concern with nucleos(t)ide analogue therapy is the risk for the development of viral populations with resistance-associated mutations (3, 4). Standard direct PCR Sanger sequencing and point mutation assays are able to detect viral populations with resistance-associated mutations only when they become dominant, leaving a gap in the understanding of the dynamics in the development of resistance. Direct PCR Sanger sequencing detects, on average, mutations present at ratios of >20% of the circulating virus population (5). Clonal sequencing has a higher sensitivity for detecting low-prevalence HBV mutations, but it is costly and labor-intensive (6). Point mutation assays can detect specific variants at as low as 5% of the virus population, but only if well-established mutations are present, whereas it may miss mutations that are not yet associated with resistance (2). Other sequence-specific genotypic resistance tests are available or under development, such as restriction fragment length/mass polymorphism, mutation-specific real-time PCR, oligonucleotide microarray, and gene chip technology (7, 8). The use of novel technologies to sequence multiple genetic variants in a heterogeneous pool of amplified DNA molecules, such as those from a virus quasispecies, by massively parallelizing PCR amplification, represents the next-generation sequencing (NGS) that is able to quantitatively detect minority variants at levels as low as 1 to 2% (9, 10).

Our aim here was to develop a highly sensitive and reproducible method for quantifying viral populations with all major HBV resistance-associated mutations combining the amplification refractory mutation system real-time PCR (ARMS RT-PCR) with molecular beacon biotechnology, even if they represent minor quasispecies.

MATERIALS AND METHODS

DNA extraction.

Viral DNA was extracted from 500 μl of serum by using a QIAamp DNA blood minikit (Qiagen NV, Netherlands) according to the standard protocol. The extracted products from the clinical samples were stored at −80°C.

Primers and molecular beacon.

A total of 67 HBV full-length sequences representing all available human HBV genotypes (A to H) were downloaded from the GenBank sequence database (http://www.ncbi.nlm.nih.gov) and aligned using the CLUSTAL W program (version 1.81) (11). The most common point mutation was selected, especially regarding genotype D, for each one of the studied resistance-associated mutations. Primers were designed according to the ARMS technique so as to be refractory for the PCR amplification of the nonmatching target sequences (12–14). A series of experiments where mutant primers with four different n−1 substitutions have been used under the same conditions of temperature and time allowed us to choose the best set of primers that gave sufficient discriminatory ability without decreasing the sensitivity and specificity of the method. In cases where neighboring mutations were present, e.g., L180M, A181V, and T184A, we applied the same reverse primer designed in a stable area of the targeted sequence and different forward primers specifically designed for each mutation. In the positions where different bases might have been expected according to the alignment of all possible combinations and expected variations attributed to different mutations, we inserted ambiguity codes in the forward primer's design. The insertion of ambiguity codes did not affect the primer binding, nor did it decrease the sensitivity of the ARMS RT-PCR with molecular beacon assay in any way. Single-stranded target sequence synthesis and high-purity clearance for primers was performed by Invitrogen Laboratories (Invitrogen Corp., Carlsbad, CA). The molecular beacon was synthesized by Biolegio Laboratories (Biolegio BV, Netherlands). The primers and the molecular beacon we finally applied can be found in Table S1a in the supplemental material.

WT and Mut laboratory target sequences.

A single-stranded DNA target sequence containing 326 bases was designed as to include codons encoding for the wild-type (WT) virus amino acids at positions 173, 180, 181, 184, 204, and 236. Based on the WT target sequence a mutant (Mut) target sequence was constructed so that the most common point-mutational codon for each mutation would be included. The laboratory designed target of 326 bases was introduced to the plasmid by a standard cloning protocol. The target was ligated into the pCR2.1 TOPO-TA vector. The ligated plasmids were transformed into Escherichia coli XL-1 competent cells according to the manufacturer's instructions (Invitrogen Corp., Carlsbad, CA). Transformed clones were selected using a blue-white selection system, and plasmids were purified using a Miniprep DNA purification kit (TaKaRa Bio, Inc., Otsu, Japan). Insertion was verified by EcoRI digestion and sequencing (ABI3100 Sequencer; Abbot Diagnostics, IL) using M13 primers. Plasmid-derived Mut and WT target sequences contained all six studied codons.

ARMS RT-PCR with molecular beacon reaction.

Quant-iT PicoGreen double-stranded DNA reagent was applied to quantitate the PCR amplification product with real-time measurements using the LightCycler 2.0 system after the WT and Mut laboratory-designed target sequences were converted to double-stranded DNA forms (Invitrogen and Roche Applied Science [Indianapolis, IN]). After the initial concentrations were calculated for both WT and Mut target sequences, appropriate dilutions were performed in order to obtain an initial concentration of 10⁶ copies per reaction for each target. The LightCycler system was used for the ARMS RT-PCR (Roche Diagnostics Corp., Roche Applied Science). The final concentrations applied were as follows: 2 μl per reaction of 10× LC FastStart DNA Master HybProbe, 1 μl per reaction of each of the appropriate for the target sequence forward and reverse set of primers from a stock concentration of 30 μM, 5 μl per reaction of MgCl₂ from a stock concentration of 25 mM, 0.5 μl per reaction of the molecular beacon from a stock concentration of 20 μM, and 1 μl per reaction of heat-labile uracil-DNA glycosylase from a stock concentration of 1 U/μl. Next, 20-μl PCR capillary tubes were used for mixing 10.5 μl of Master Mix with 10 μl of the HBV DNA target sequences. ARMS RT-PCR conditions and different annealing temperatures applied are summarized in Table 1. Serial dilutions were used for WT and Mut target sequences, ranging from 10⁶ to 10 copies per reaction. The two sets of primers, one specifically designed for the WT target and the other designed for the Mut target, were applied to the relevant targets to construct a DNA standard curve. The overall workflow of the ARMS RT-PCR with molecular beacon assay is shown in Fig. 1. In early experiments, WT and Mut target sequences were mixed in predetermined concentrations in order to resemble clinical samples. However, in the experiments performed in order to determine the discriminatory ability of the assay, WT and Mut target sequences were used separately with the appropriate sets of primers.

Table 1.

Final cycling conditions for each major HBV resistance-associated mutation

Reaction step	Final cycling conditions for indicated mutation
	V173L		L180M		A181V		T184A		M204I		N236T
	Temp (°C)	Time (min)	Temp (°C)	Time (min)	Temp (°C)	Time (min)	Temp (°C)	Time (min)	Temp (°C)	Time (min)	Temp (°C)	Time (min)
Denaturation	94	10	94	10	94	10	94	10	94	10	94	10
50 RT-PCR cycles
Initial temp	94	0	94	0	94	0	94	0	94	0	94	0
Fluorescence measurement	52	0.02	52	0.02	52	0.02	52	0.02	52	0.02	52	0.02
Annealinga	58	0.10	58	0.10	56	0.10	58	0.10	58	0.10	56	0.10
Final extension	72	0.10	72	0.10	72	0.10	72	0.10	72	0.10	72	0.10

^aThe final cycling conditions applied for each of the major HBV resistance-associated mutations according to the ARMS RT-PCR with a molecular beacon assay are given. Different annealing temperatures were required for each mutation. The fluorescence measurement was performed in a different step of the reaction.

Fig 1.

ARMS RT-PCR with molecular beacon assay results. Each sample was quantitated twice; point k represents the sample quantification with the master mix containing the set of primers specifically designed to detect the WT sequence, and point l represents the sample quantification with the master mix containing the set of primers specifically designed to detect the Mut sequence. In the final graph, the x axis represents the sample concentration calculated by the ARMS RT-PCR with molecular beacon reaction, while the y axis represents the cycle and the fluorescence channel on which the sample has been detected by the reaction. The X₂/X₁ ratio represents the Mut/WT ratio applied in data analysis. The procedure was repeated six times for the six different mutations.

Evaluation of ARMS RT-PCR with molecular beacon assay in clinical samples.

Based on our previous study, the biological cutoff the ARMS RT-PCR assay has been determined to be 0.04%, meaning that the assay was able to detect the viral subpopulation in a clinical sample if it was represented in at least 4 copies per 10⁴ copies of the total viral population (14). For this reason, we decided that most of our clinical samples should have a viral load >10⁴ copies per reaction after dilution in order to be included in our study. The reference method for the viral load of HBV in the patient samples was a method developed in-house (15). A total of 41 plasma or serum samples from chronically HBV-infected patients who failed antiviral treatment were selected for the evaluation of the ARMS RT-PCR with molecular beacon assay. Eligible patients included those that had shown antiviral resistance, proved by detectable HBV DNA in two sequential measurements and the development of resistant mutations when tested by direct PCR Sanger sequencing. Sixty samples from chronically HBV-infected nucleos(t)ide analogue-naive patients were selected for measurement by the ARMS RT-PCR with molecular beacon assay. The results assessed the presence of resistant quasispecies in samples from HBV-naive patients for each of the six mutational sites.

Bioethical approval.

The Bioethical Committee of the Medical School, Athens University, approved the conduct of the study under the title “Quantification of HBV Mutations Associated with Antiviral Resistance” (reference 37/26-4-06).

Statistical analysis.

Constant variables were expressed as means, standard deviations, medians, and percentiles. Pearson and Spearman correlation coefficients were applied for correlations between sequential variables. McNemar's test was used to assess discordance and differentiation for the two methods. Sensitivity and 95% confidence intervals of the ARMS RT-PCR with molecular beacon assay were calculated with the SAS 9.0 (2004; SAS Institute, Inc.) statistical program (16).

RESULTS

ARMS RT-PCR with molecular beacon assay has been evaluated using serial dilutions of WT and Mut target sequences from 10⁶ copies per reaction to 10 copies per reaction with the appropriate set of primers for every one of the HBV resistance-associated mutations (14). Repeated measurements showed that the assay successfully detected WT and Mut target sequences in concentrations as low as 10 copies per reaction an analytical sensitivity of 90 to 100% for each one of the studied mutations. Representative examples of experiments performed in target sequences in order to determine the discriminatory ability of the ARMS RT-PCR with molecular beacon are shown in Fig. 2. The discriminatory ability of the ARMS RT-PCR with molecular beacon assay was 0.01 to 0.25% for the different HBV resistance-associated mutations, as shown in Table 2. Discriminatory ability has been evaluated in repeated experiments (n = 5); however, measurements are based on representative experiment data. The methodology of the ARMS RT-PCR with molecular beacon assay focuses on the detection of Mut variants when they represent minor viral populations. The experiments for the discriminatory ability were performed at the lowest concentrations for the WT (10⁶ to 10 variants) variants, suggesting that they apply also at higher values. The reaction has proven to be efficient to detect higher and lower concentrations of both Mut and WT targets, and the Mut/WT ratio was kept constant for each of the tested mutations regardless of the initial concentration of the dominant target sequence.

Fig 2.

Representative results of experiments performed in target sequences in order to determine the discriminatory ability of the ARMS RT-PCR with molecular beacon assay. The discriminatory ability represents the ratio (%) of the lowest Mut target sequence concentration that can be detected with the ARMS RT-PCR with molecular beacon assay without interference by the WT target sequence using Mut primers divided by the WT target concentration. It was operationally defined as the lowest detected Mut variant concentration within the window of the assay performed when WT and Mut primers were applied to the WT target sequence. In the example of the resistance-associated mutation V173L (A), there is a nine-cycle difference window between the matching and nonmatching primers which enables the quantitative detection of Mut target sequences with concentrations of 10³, 5 × 10², and 10² copies per reaction with the set of primers specifically designed for the detection of the Mut target sequence (blue lines). However, a Mut target sequence concentration of 10 copies per reaction cannot be detected with the matching set of primers within the red-line window, yielding a discriminatory ability of 0.01% for this mutation. In the example of the resistance-associated mutation A181V (B), blue lines represent the detection of Mut target sequences in concentrations 10⁵, 10⁴, 10³, and 2.5 × 10³copies per reaction with the matching Mut set of primers within the 11-cycle difference window between the matching and nonmatching primers for the WT target sequence, yielding a discriminatory ability of 0.25% for this mutation.

Table 2.

ARMS RT-PCR discriminatory ability and cycle difference for each HBV resistance-associated mutation

Mutation	Discriminatory abilitya (%)	Cycle difference (no. of differences)
V173L	0.010	9
L180M	0.010	16
A181V	0.250	11
T184A	0.125	8
M204I	0.125	16
N236T	0.125	9

^aThe discriminatory ability and cycle difference of the ARMS RT-PCR with a molecular beacon assay for each of the HBV resistance-associated mutations (V173L, L180M, A181V, T184A, M204I, and N236T) is presented. Discriminatory ability was defined as the lowest detected Mut variant concentration within the window of the assay performed when WT and Mut primers were applied on the WT target sequence.

Evaluation of ARMS RT-PCR with molecular beacon assay in samples from chronically HBV-infected patients receiving antiviral therapy.

The ARMS RT-PCR with molecular beacon assay was applied for the quantitative detection of Mut HBV variants in 41 samples from chronically HBV-infected patients receiving antiviral therapy. The presence of HBV resistance-associated mutations was previously detected by Sanger sequencing. The ratio of Mut to WT viral populations in each sample (Mut/WT ratio) was determined. The results from the quantitative detection of the WT HBV quasispecies were compared to the overall viral load for each sample, and the correlations were statistically significant apart from the mutation N236T (see Table S2a in the supplemental material). Measurements of each of the samples from the HBV-infected patients with the ARMS RT-PCR with molecular beacon assay were compared to the results obtained using direct PCR Sanger sequencing. The ARMS RT-PCR with molecular beacon assay showed a sensitivity of 100% for the detection of mutant variants A181V, T184A, and N236T. The sensitivity of the assay for the remaining mutations was 75%, 96.3% and 93.3% for the V173L, L180M and M204I, respectively (Table 3). In most cases, the Mut/WT ratio tested by the ARMS RT-PCR assay was >99% when a resistance-associated mutation was determined by direct PCR Sanger sequencing. The total number of mutations detected by Sanger sequencing was 57. Except for three mutations in different samples (mutations V173L, L180M, and M204I) detected by Sanger sequencing, the rest were quantitatively detected by ARMS RT-PCR with a molecular beacon assay as dominant populations, leading to an overall sensitivity for the ARMS RT-PCR assay of 54 of 57 (94.7%), a sensitivity comparable to other available ultrasensitive assays. However, the total number of mutations quantitatively detected by the ARMS RT-PCR assay was 77, although not always as dominant populations. Upon reexamination of the three samples, the viral load after the dilution was very low when the samples were tested by ARMS RT-PCR (lower than the described biological cutoff for the assay), whereas the samples had been tested using sequencing in higher concentrations before being diluted; therefore, the inability of the ARMS RT-PCR to detect the Mut variants was due to the low concentration of HBV variants following the dilution. Moreover, sequencing results for the V173L mutation gave a mixed population of V173V/L for the sample where ARMS RT-PCR failed to detect a Mut/WT ratio. Apart from the dominant viral populations, minor variant populations were also detected with the ARMS RT-PCR with a molecular beacon assay that might have contributed to the overall resistance profile of each patient (see Table S3a in the supplemental material).

Table 3.

Comparison of results obtained by ARMS RT-PCR with a molecular beacon assay and Sanger sequencing for HBV resistance-associated mutations in 41 samples from chronic HBV patients under antiviral therapy in whom HBV had developed resistance-associated mutations

Mutation	No. of samples (n = 41)		Sensitivitya (%) (95% CI)
	ARMS RT-PCR	Sequencing
V173L	4	4	75b (19.4–94.4)
L180M	30	27	96.3 (81–99.9)
A181V	7	5	100 (47.8–100)
T184A	9	2	100 (15.8–100)
M204I	16	15	93.3 (68.1–99.8)
N236T	11	4	100 (39.8–100)

^aSensitivity was defined as the percentage of mutations detected by Sanger sequencing that were also quantitatively detected by the ARMS RT-PCR assay. CI, confidence interval.

^bIn the V173L mutation, although the ARMS RT-PCR assay and Sanger sequencing detected the same numbers of positive samples, one sample that was detected by Sanger sequencing was missed by the ARMS RT-PCR assay, whereas another sample tested negative for the V173L mutation with Sanger sequencing. ARMS RT-PCR quantitatively detected a viral subpopulation with the V173L mutation in a Mut/WT ratio of 3.3%.

Evaluation of ARMS RT-PCR with a molecular beacon in samples from chronically HBV-infected, nucleos(t)ide analogue-naive patients.

ARMS RT-PCR with a molecular beacon was used to quantify Mut HBV variants in 60 samples from chronically HBV-infected, nucleos(t)ide analogue-naive patients. The ratio of Mut to WT viral populations in each sample (i.e., the Mut/WT ratio) was also determined. The results for the WT HBV quasispecies were compared to overall viral load for each sample and the log HBV DNA (as calculated by the HBV DNA real-time PCR standard protocol) showed a good correlation with the log WT (as calculated by the ARMS RT-PCR with a molecular beacon assay) (see Table S4a in the supplemental material). The ratio of mutant to wild-type viral populations in each sample (Mut/WT ratio) was >1% in 38 (63.3%) of 60 samples from chronically HBV-infected, nucleos(t)ide analogue-naive patients. Combinations of resistance-associated mutations with a Mut/WT ratio >1% were detected in 65 of 209 (31%) measurements in which a Mut/WT ratio could be estimated. In some cases, the Mut/WT ratio was >99%. The results were categorized in the following groups: below the cutoff or a Mut/WT ratio of <1%, 1 to 5%, 5 to 20%, or >20% for each one of the HBV resistance-associated mutations, as shown in Table 4. The T184A mutation showed a higher proportion of Mut/WT ratios that were >1% (22 samples), followed by N236T, A181V, L180M, and M204I (18, 15, 5, and 5 samples, respectively), whereas no Mut viral population was detected over the predetermined ratio of 1% for V173L mutation. Interestingly, combinations of mutant quasispecies were found in 19 (50%) of the 38 samples in which a Mut/WT ratio of >1% was detected. The most common combinations were T184A-N236T and A181V-N236T (eight samples), followed by the combination A181V-T184A (six samples). The percentage of mutant populations was higher at L180, A181, and T184 resistant sites, probably due to the high fitness cost of major resistance sites (e.g., M204 or N236) in the absence of therapy. Analytical results can be found in Table S5a in the supplemental material. We present there a comparison of the results obtained with ARMS RT-PCR with a molecular beacon assay with two different thresholds—one at a Mut/WT ratio of >1% and one at Mut/WT ratio of >5%—to show the relative benefits in the quantitative detection of viral variants with HBV resistance-associated mutations with an ultrasensitive assay (Fig. 3).

Table 4.

Mut/WT ratio in 60 samples from chronically HBV-infected nucleos(t)ide analogue-naive patients^a

Mutation	No. of samples with the indicated Mut/WT ratio
	Below the cutoff	<1%	1–5%	5–20%	>20%
V173L	49	11	0	0	0
L180M	46	9	1	2	2
A181V	18	27	5	4	6
T184A	2	36	11	6	5
M204I	31	24	0	0	5
N236T	5	37	6	6	6

^aQuantitative detection of the Mut/WT ratio was performed in 60 samples obtained from chronically HBV-infected nucleos(t)ide analogue-naive patients using ARMS RT-PCR with a molecular beacon assay. Each mutation was quantitatively detected with Mut and WT primers, and the Mut/WT ratio was calculated. Results are grouped as follows: a Mut/WT ratio below the cutoff or a Mut/WT ratio of <1%, 1 to 5%, 5 to 20%, or >20% for each of the HBV resistance-associated mutation (V173L, L180M, A181V, T184A, M204I, and N236T).

Fig 3.

Frequencies of observed combinations of mutant quasispecies in samples from HBV-naive patients in which the Mut/WT ratio was >1% when detected by the ARMS RT-PCR with molecular beacon assay. The results are compared to those obtained with a Mut/WT ratio of >5%.

DISCUSSION

Our data show that the ARMS RT-PCR with molecular beacon assay is an ultrasensitive method for the quantitative detection of HBV quasispecies with nucleos(t)ide analogues resistance-associated mutations before they emerge as dominant populations. This assay showed high sensitivity and discriminatory ability compared to the direct PCR Sanger sequencing currently used in clinical practice in identifying resistant viral populations in chronically HBV-infected patients under antiviral therapy. The ARMS RT-PCR with molecular beacon assay yielded results comparable to those of other novel ultrasensitive methods, such as next-generation sequencing (NGS), for the quantitative detection of quasispecies, even in cases where combinations of resistance-associated mutations exist. Moreover, ARMS RT-PCR provides a low-cost, user-friendly method that can be easily implemented in clinical laboratories with no access to NGS technologies.

Studies in patients with HBV and/or HIV infection, both drug-naive and treated, showed that NGS application succeeded in detecting clinically relevant minority drug resistance mutants even when they represented <1% of the total viral population (17, 18). The ARMS RT-PCR with molecular beacon assay quantitatively detected, apart from the dominant clones, minor variant populations, even when they represented 0.01 to 0.25% of the total viral population, depending on the mutation tested. One of the major limitations of NGS technologies is the different types of errors that can be introduced during the PCR amplification step, such as the nucleotide misincorporation due to the inaccuracy of DNA polymerases (19). In the ARMS RT-PCR assay, genome amplification is not necessary since the quantification of specific variants is accomplished due to differences in hybridization between the WT and Mut alleles and primers targeting specific positions. These characteristics may be the reason why the ARMS RT-PCR with molecular beacon assay is more sensitive in the detection of viral quasispecies. Moreover, the ARMS RT-PCR with molecular beacon assay can be easily adopted for the quantitative detection of novel or other well-established HBV resistance-associated mutations, with only slight modifications in the primers' design and in the annealing temperatures of the reaction.

Early detection of minor viral populations before they emerge to dominant clones may have an impact in the clinical outcome of chronically HBV-infected patients. Several studies have been performed to determine whether mutational patterns of the HBV genome are related to clinical outcomes after the emergence of drug-resistant variants during nucleos(t)ide analog therapy (20). Enomoto et al. (21) reported that an additional L180M mutation was detected in 4 of the 9 patients who had the M204I mutations alone at the time of first detection. Elevated alanine transaminase (ALT) activity was observed when the L180M mutation was detected in addition to the mutation M204I. The worst outcome was attributed to the close interaction between position 180 in the 173-189 helix and position 204 of the YMDD loop, as shown by three-dimensional structural modeling (22). In the example of adefovir dipivoxil, initially used for the treatment of patients with lamivudine resistance mutations, a more complex dynamic than previously thought raised concerns regarding the clinical significance of early detection of minor viral quasispecies (23). Salvage therapy in patients with lamivudine resistance achieves more rapid and higher rates of virological response if adefovir dipivoxil is initiated at an early phase of lamivudine resistance with low levels of viral replication, suggesting that early detection of mutant viral quasispecies using a sensitive technique may be crucial for patients under long-term antiviral treatment in order to detect resistance and initiate salvage therapy promptly (24).

Sensitive HBV resistance assays may be useful even in the current era of entecavir and tenofovir, since there are still patients who are not nucleos(t)ide naive but have been previously exposed to other antiviral agents. The long-term efficacy of tenofovir monotherapy for HBV-monoinfected patients after the failure of nucleos(t)ide analogues has been clearly shown, but entecavir therapy is associated with increasing rates of resistance (25, 26). The resistance rates to entecavir in lamivudine-experienced patients are highest in cases with lamivudine resistance mutations at the onset of entecavir, intermediate in patients with a history of lamivudine resistance, and relatively lower, but still much higher than the entecavir resistance rates in naive patients, in patients with exposure but no resistance to lamivudine, at least according to standard resistance assays (27). Novel sensitive HBV resistance assays might be helpful in identifying lamivudine-exposed patients with lamivudine-resistant HBV strains as minor variant populations, who will have an increased risk of developing resistance if they receive long-term entecavir monotherapy. Combination therapy with entecavir and tenofovir has been used as salvage therapy in chronically HBV-infected patients with multidrug viral resistance patterns or a partial antiviral response to preceding therapies with promising results (28).

In conclusion, our results suggest that genotyping analysis, based on the next-generation ARMS RT-PCR with molecular beacon assay, is suitable for characterization of genetic diversity and detection of minor viral quasispecies that may influence therapeutic decisions in both nucleos(t)ide-naive and -experienced HBV-infected patients. The cost of the ARMS RT-PCR with molecular beacon assay is low compared to other novel molecular techniques, and future applications in clinical practice may lead to less-frequent HBV DNA level measurements and so reduce the overall burden for the follow-up of chronically HBV-infected patient under therapy. Moreover, this assay leads to a more personalized approach, which is especially important for chronically HBV-infected patients since they are candidates for developing resistance-associated mutations and fail to be cured (29). Our data may have a clinical significance when planning a treatment strategy in chronically HBV-infected patients in order to avoid the potential development of HBV resistance-associated mutations; HBV viral quasispecies with resistance-associated mutations were detected in a high proportion of our nucleos(t)ide-naive chronically HBV-infected patients, suggesting that the initiation of inappropriate therapy may trigger the emergence of viral resistance. Novel, sensitive, and accurate methods that enable the early detection of minor viral quasispecies before they emerge as dominant populations could lead to better targeted antiviral therapy and increased chances of success, but this hypothesis remains to be prospectively confirmed by clinical studies before it can be applied to everyday practice.