Detection of Anti-Hepatitis B Virus Drug Resistance Mutations Basedon Multicolor Melting Curve Analysis

Yi Mou,a,b Muhammad Ammar Athar,a,b Yuzhen Wu,a,b Ye Xu,a,b Jianhua Wu,c Zhenxing Xu,c Zulfiqar Hayder,d Saeed Khan,e

Muhammad Idrees,f Muhammad Israr Nasir,g Yiqun Liao,a,h Qingge Lia,b

State Key Laboratory of Cellular Stress Biology, State Key Laboratory of Molecular Vaccinology and Molecular Diagnostics, Engineering Research Center of MolecularDiagnostics of the Ministry of Education, School of Life Sciences, Xiamen University, Xiamen, Fujian, Chinaa; Shenzhen Research Institute of Xiamen University, Shenzhen,Guangdong, Chinab; Xiamen Hospital of Traditional Chinese Medicine, Xiamen, Fujian, Chinac; Department of Pathology, Quid-e-Azam Medical College, Bahawalpur,Punjab, Pakistand; Department of Molecular Pathology, Dow University of Health Sciences, Karachi, Pakistane; Center for Applied Molecular Biology, University of thePunjab, Lahore, Pakistanf; Department of Molecular Pathology, Liaquat National Hospital, Karachi, Pakistang; School of Public Health, Xiamen University, Xiamen, Fujian,Chinah

Detection of anti-hepatitis B virus (HBV) drug resistance mutations is critical for therapeutic decisions for chronic hepatitis Bvirus infection. We describe a real-time PCR-based assay using multicolor melting curve analysis (MMCA) that could accuratelydetect 24 HBV nucleotide mutations at 10 amino acid positions in the reverse transcriptase region of the HBV polymerase gene.The two-reaction assay had a limit of detection of 5 copies per reaction and could detect a minor mutant population (5% of thetotal population) with the reverse transcriptase M204V amino acid mutation in the presence of the major wild-type populationwhen the overall concentration was 104 copies/�l. The assay could be finished within 3 h, and the cost of materials for each sam-ple was less than $10. Clinical validation studies using three groups of samples from both nucleos(t)ide analog-treated and -un-treated patients showed that the results for 99.3% (840/846) of the samples and 99.9% (8,454/8,460) of the amino acids were con-cordant with those of Sanger sequencing of the PCR amplicon from the HBV reverse transcriptase region (PCR Sangersequencing). HBV DNA in six samples with mixed infections consisting of minor mutant subpopulations was undetected by thePCR Sanger sequencing method but was detected by MMCA, and the results were confirmed by coamplification at a lower dena-turation temperature-PCR Sanger sequencing. Among the treated patients, 48.6% (103/212) harbored viruses that displayedlamivudine monoresistance, adefovir monoresistance, entecavir resistance, or lamivudine and adefovir resistance. Among theuntreated patients, the Chinese group had more mutation-containing samples than did the Pakistani group (3.3% versus 0.56%).Because of its accuracy, rapidness, wide-range coverage, and cost-effectiveness, the real-time PCR assay could be a robust tool forthe detection if anti-HBV drug resistance mutations in resource-limited countries.

Hepatitis B is caused by the hepatitis B virus (HBV), an envel-oped DNA virus that infects the liver, causing hepatocellular

necrosis and inflammation (1). Chronic hepatitis B (CHB) infec-tion affects approximately 248 million people worldwide and is aleading cause of liver-related morbidity and mortality, particu-larly in low- and middle-income countries (LMICs) (2). Patientswith CHB can be successfully treated using nucleos(t)ide analogs(NAs), but drug-resistant HBV mutants frequently arise, leadingto treatment failure and progression to liver disease (3). The de-velopment of drug resistance begins with mutations in the HBVpolymerase gene, followed by an increase in the viral load andserum alanine aminotransferase levels several weeks to monthslater (4). Detection of drug resistance mutations is thus critical inprompt decision making for new therapeutic regimes (5, 6). Amethod enabling detection of NA resistance mutations should bereliable, rapid, and, in particular, easy to use and cost-effectivewhen the aim is for it to be used in high-CHB-burden countries,which are often undeveloped and resource limited.

Many methods for the detection of anti-HBV drug resistancemutations have been developed. However, these methods are al-most exclusively performed by highly skilled technicians in well-equipped referral hospitals. This limitation restricts the use of an-ti-HBV drug resistance assays to only a small number of patients,which offers little help for CHB management program in LMICs.Sanger sequencing of the PCR amplicon from the HBV reversetranscriptase region (referred to here as PCR Sanger sequencing)

is currently the “gold standard” method for the detection of anti-HBV drug resistance mutations. It is advantageous for the accu-rate detection of all nucleotide mutations, including novel muta-tions which have never been reported before. However, the Sangersequencer is often unaffordable for the local hospitals that admit amajority of CHB patients in LMICs. Moreover, the sequencingprocedure is complex and lengthy and requires very careful oper-ations to avoid contamination from PCR amplicons. This is alsotrue for other sequencing-based assays, such as pyrosequencing(7), coamplification at a lower denaturation temperature (COLD)-

Received 27 February 2016 Returned for modification 12 May 2016Accepted 5 August 2016

Accepted manuscript posted online 17 August 2016

Citation Mou Y, Athar MA, Wu Y, Xu Y, Wu J, Xu Z, Hayder Z, Khan S, Idrees M,Nasir MI, Liao Y, Li Q. 2016. Detection of anti-hepatitis B virus drug resistancemutations based on multicolor melting curve analysis. J Clin Microbiol54:2661–2668. doi:10.1128/JCM.00439-16.

Editor: Y.-W. Tang, Memorial Sloan-Kettering Cancer Center

Address correspondence to Yiqun Liao, yqliao@xmu.edu.cn, orQingge Li, qgli@xmu.edu.cn.

Y.M. and M.A.A. contributed equally to this article.

Supplemental material for this article may be found athttp://dx.doi.org/10.1128/JCM.00439-16.

Copyright © 2016, American Society for Microbiology. All Rights Reserved.

crossmark

November 2016 Volume 54 Number 11 jcm.asm.org 2661Journal of Clinical Microbiology

on Novem

ber 21, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

PCR Sanger sequencing (8–10), and next-generation sequencing(11, 12), as well as the high-end instrument-based assays, such asDNA microchip (13) and mass spectrometry (14) assays. Severalsimpler and more cost-effective assays for detecting anti-HBVdrug resistance mutations have been developed. For example, theline probe assay (HBV DR v.3; Fujirebio Europe, Ghent, Belgium)is a reverse line blot assay that detects 27 mutations at 11 aminoacid positions in the reverse transcriptase region of the HBV poly-merase gene. The HBV DR v.3 assay is technically simple, rapid,and amenable to high throughput (15). However, as an open-tubeassay, it requires multiple post-PCR hybridization and visualiza-tion steps, which are prone to amplicon contamination, whichleads to the generation of false-positive results. Real-time PCR isan excellent platform of choice for clinical diagnosis because of itsclosed-tube detection format, which makes it easy to use and rapidand which results in decreased amplicon contamination. A real-time amplification refractory mutation system (ARMS) PCR forthe detection of six HBV resistance mutations has been developed.Nevertheless, it is far from practical due to the limited coverage ofmutation types (16).

We previously reported a multiplex real-time PCR detectionstrategy based on multicolor melting curve analysis (MMCA) us-ing dually labeled, self-quenched probes (17). Up to 16 geneticmutations could be detected in a single reaction by this strategy(18). In the present study, we describe a two-reaction MMCAassay that can detect 24 anti-HBV drug resistance mutations at 10amino acid positions. We systematically evaluated its analyticalperformance, including its mutation detection accuracy, its ana-lytical sensitivity, and, in particular, its ability to detect minorvariants. We further evaluated its clinical performance by analyz-ing samples from a group of 212 CHB patients who had receivednucleos(t)ide analog treatment, samples from a second group of276 HBV-positive people who were selected from among thosereceiving a routine health examination, and a third group of 400HBV-positive samples collected from individuals in Pakistan. Theresults were compared with those of PCR Sanger sequencing andconfirmed by COLD-PCR Sanger sequencing.

MATERIALS AND METHODSClinical samples. In total, 888 serum samples were collected from thesame number of deidentified patients. These samples were labeled HBVpositive when they were received. The clinical samples were divided intothree groups. Samples in group A (212 samples) were collected from pa-tients who had received or were undergoing treatment with NAs at theXiamen Hospital of Traditional Chinese Medicine (Xiamen, China).Samples in group B (276 samples) were collected in the same hospitalfrom individuals who were being seen for a routine health examinationand had never received NA drugs. Samples in group C (400 samples) werecollected from four medical colleges or hospitals in Pakistan, includingQuid-e-Azam Medical College, Dow University of Health Sciences, Uni-versity of the Punjab, and Liaquat National Hospital. An exemption fromfull committee review for human subject studies was granted by the Xia-men University Research Ethics Committee since only coded samplescollected for other procedures were used in the present work.

HBV DNA was extracted from the serum samples using a Lab-Aid 820virus DNA isolation kit (Zeesan Biotech Ltd., Xiamen, China). The DNAextracts were used for both MMCA and PCR Sanger sequencing. Beforemutation detection, the presence of HBV in the sample was confirmed byuse of an HBV real-time PCR quantification kit (Zhijiang Biotech Ltd.,Shanghai, China).

Plasmid DNA. Wild-type plasmid DNA was prepared by use of anHBV-positive sample obtained from the Xiamen Hospital of Tradi-

tional Chinese Medicine. An amplicon of 828 bp from reverse transcrip-tase amino acid 14 (rt14) to rt289 was generated using forward primer5=-ATCAGGACTCCTAGGACC-3= and reverse primer 5=-TCGTTGACATACTTTCCAATC-3=. Sequencing analysis confirmed that the ampliconcorresponded to that from the HBV isolate with GenBank accession num-ber GQ377639.1 (http://www.ncbi.nlm.nih.gov/) with one non-drug re-sistance mutation (A ¡ G) on the third base of rt255. A plasmid carryingthe wild-type HBV sequence was constructed by cloning the ampliconinto the pMD18-T vector using a TA cloning kit (TaKaRa, Dalian, China).Plasmids carrying mutant DNA sequences were constructed from thewild-type plasmid by site-directed mutagenesis (19). In total, 24 plasmidswith 24 mutations, i.e., rtL80I, rtL80V, rtV173L, rtV173G, rt(L/V)180M,rtA181T, rtA181V, rtT184A, rtT184C, rtT184F, rtT184G, rtT184I,rtT184L, rtT184M, rtT184S, rtA194T, rtS202C, rtS202G, rtS202I,rtM204I, rtM204V, rtM204S, rtI233V, and rtN236T, were prepared. Thedrug resistance nucleotide mutation(s) corresponding to each amino acidalteration was determined (see Table S1 in the supplemental material).The plasmid DNA in 10 mM Tris-HCl (pH 8.0) containing 1 mM EDTA(TE buffer) was quantified using a NanoDrop ND-1000 spectrophotom-eter (NanoDrop Technologies, Montchanin, DE).

MMCA assay. The MMCA assay has two reactions that detect 24 mu-tations in HBV that confer resistance against the five FDA-approved anti-HBV NA drugs, including lamivudine (LMV), adefovir (ADV), entecavir(ETV), tenofovir (TDF), and telbivudine (LdT). Reaction A detects 10mutations, including rtL80I, rtL80V, rt(L/V)180M, rtA181T, rtA181V,rtA194T, rtM204I, rtM204V, rtM204S, and rtN236T. Reaction B detects14 mutations, including rtV173L, rtV173G, rtT184A, rtT184C, rtT184F,rtT184G, rtT184I, rtT184L, rtT184M, rtT184S, rtS202C, rtS202G,rtS202I, and rtI233V.

Each MMCA assay reaction mixture contains 2.5 �l of 10� Taq HSbuffer (TaKaRa), 0.2 �l of Taq HS (5 units/�l), 0.2 �l of deoxynucleosidetriphosphates (dNTPs; 25 mM), 0.1 �l of primer-probe mix, 17 �l ofdeionized H2O, and 5 �l of the HBV DNA template. PCR and meltingcurve analysis were performed in a Bio-Rad CFX-96 thermocycler (Bio-Rad, Hercules, CA). PCR was started with denaturation at 95°C for 5 minand was followed by 45 cycles of 95°C for 15 s, 52°C for 15 s, and 72°C for20 s. Melting curve analysis began with denaturation at 95°C for 60 s andhybridization at 30°C for 60 s, followed by a gradual temperature increasefrom 30°C to 85°C in increments of 0.5°C/s. The fluorescence from the6-carboxyfluorescein (FAM; absorbance � [�abs] and emission � [�em],495 nm and 520 nm, respectively), hexachloro-6-carboxyfluorescein(HEX; �abs and �em, 521 nm and 536 nm, respectively), 6-carboxy-X-rhodamine (ROX; �abs and �em, 586 nm and 610 nm, respectively), andindodicarbocyanine 5 (Cy5; �abs and �em, 647 nm and 667 nm, respec-tively) channels was recorded.

PCR and COLD-PCR for Sanger sequencing. Each 25-�l PCR mix-ture was prepared as described above and contained 2.5 �l of 10� Taq HSbuffer, 0.2 �l of Taq HS, 0.2 �l of dNTPs, 0.1 �l of primer mix, 17 �l ofdeionized H2O, and 5 �l of HBV DNA. The primer mix was composed of20 pM each forward primer 5=-ATCAGGACTCCTAGGACC-3= and re-verse primer 5=-TCGTTGACATACTTTCCAATC-3=. Conventional PCRwas carried out in a T3 thermocycler (Biometra, Göttingen, Germany)using a program of 95°C for 5 min, followed by 45 cycles of 95°C for 15 s,52°C for 15 s, and 72°C for 40 s.

COLD-PCR was performed essentially as previously described (10).Briefly, 5 �l of the HBV DNA template was added to a 20-�l PCR mixturecontaining 2.5 �l of 10� Taq HS buffer, 0.2 �l of Taq HS (5 unit/�l), 0.1�l of primer mix (50 �M for each), 0.2 �l of dNTPs (25 mM), and 17 �lof deionized H2O. Amplification was carried out in a RotorGene Q real-time PCR system (Qiagen, Hilden, Germany) under the following condi-tions: 95°C for 10 min; 10 cycles of 95°C for 15 s, 50°C for 30 s, and 72°Cfor 1 min; 72°C for 7 min; 95°C for 2 min; and finally, 30 cycles of 95°C for15 s, 70°C for 1 min, 77°C for 5 s, 50°C for 30 s, and 72°C for 1 min. Theamplified products were sent out for bidirectional Sanger sequencing(Sangon Inc., Shanghai, China).

Mou et al.

2662 jcm.asm.org November 2016 Volume 54 Number 11Journal of Clinical Microbiology

on Novem

ber 21, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

RESULTS

We evaluated the analytical performance of the MMCA assay byusing a plasmid carrying wild-type HBV DNA (wild-type plas-mid) and 24 plasmids carrying mutant HBV DNA (mutant plas-mids) as mimics of the respective wild-type and mutant HBVDNA samples. For each plasmid, we prepared serial 10-fold dilu-tions in TE ranging from 106 to 100 copies/�l. The MMCA assayshowed that for each plasmid, a unique melting temperature (Tm)value could be obtained regardless of the concentration. Figure 1shows the melting curves of all the plasmids at 10 copies/�l. Tenrepeated detections at this concentration provided the average Tm

and standard deviation (SD) for each plasmid (Table 1). All theTm values for the mutant plasmids had a �4°C difference from theTm value for the wild-type plasmid, and the SD values were lessthan 1°C. As the Tm resolution of the MMCA assay was 0.5°C, allmutant plasmids investigated could be unequivocally differenti-ated from the wild-type plasmid.

The analytical sensitivity of the assay was obtained through 10repeated detections of each plasmid at 10 copies/�l and 1 copy/�l.The results showed that all plasmids at 1 copy/�l or 5 copies perreaction could be detected (Fig. 2), and the limit of detection of theMMCA assay was thus 1 copy/�l, or 5 copies per reaction.

To test the ability of MMCA assay to detect minor mutantsubpopulations, we prepared a series of mixtures containing plas-mids carrying the wild type and the rtM204V mutant. The overalltemplate concentration was set to 104 copies/�l, and the percent-age of the population that was the rtM204V mutant was 0%, 5%,10%, and 20%. These mixtures were subjected to MMCA, PCRSanger sequencing, and COLD-PCR Sanger sequencing in paral-lel. The results showed that both the MMCA assay and COLD-PCR Sanger sequencing could detect the rtM204V mutant sub-population when it was present at proportions of 5% and above,whereas PCR Sanger sequencing method could detect the sub-population when it was present at proportions of 10% and above

(Fig. 3), demonstrating that MMCA has a sensitivity for the detec-tion of minor mutant subpopulations equal to that of COLD-PCRSanger sequencing, which is more sensitive than PCR Sanger se-quencing.

To evaluate the clinical performance of MMCA, we used threedifferent groups of patient samples. The results showed that allsamples in group A were confirmed to be HBV positive. The mu-tations detected are listed in Table 2. Of the 212 samples, 103(48.6%) were found to contain anti-HBV drug resistance muta-tions, indicating the widespread prevalence of drug resistance inthis group. An LMV monoresistance mutation (70.9%, 73/103)was the most frequently detected mutation, followed by an ADVmonoresistance mutation (21.4%, 22/103), an LMV and ETV re-sistance mutation (4.9%, 5/103), and an LMV and ADV resistancemutation (2.9%, 3/103) (Fig. 4A). MMCA detected 6 more sam-ples with mutations than PCR Sanger sequencing did, and thesewere found to carry minor mutant subpopulations that escapeddetection by PCR Sanger sequencing. These results were con-firmed by COLD-PCR Sanger sequencing.

In group B, 274 samples (99.3%, 274/276) were confirmed tobe HBV positive, and all of these gave MMCA results. Nine sam-ples (3.3%, 9/274) were found to contain anti-HBV drug resis-tance mutations, including 7 LMV monoresistance mutations and2 ADV monoresistance mutations, indicating the lower preva-lence of drug-resistant HBV in this group than group A (Fig. 4B).The results were confirmed by PCR Sanger sequencing.

In group C, 360 samples (90%, 360/400) were confirmed to beHBV positive, and all of them gave MMCA results. Two samples(0.56%, 2/360) were found to contain an LMV monoresistancemutation, indicating the rare occurrence of drug-resistant HBV inthis group compared with its rate of occurrence in groups A and B(Fig. 4C). The results were again confirmed by PCR Sanger se-quencing.

Collectively, 114 of all 846 HBV-positive samples were found

FIG 1 Readout for a wild-type plasmid and 24 mutant plasmids using the MMCA method. (Top) Typical readout for reaction A; (bottom) typical readout forreaction B. Black lines, melting curve of wild-type plasmid; gray lines, melting curves of 24 mutant plasmids. WT, wild type; �dF/dt, negative derivative offluorescence over temperature; TET, tetrachloro-6-carboxyfluorescein.

Real-Time PCR Detection of Drug-Resistant HBV

November 2016 Volume 54 Number 11 jcm.asm.org 2663Journal of Clinical Microbiology

on Novem

ber 21, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

to contain anti-HBV drug resistance mutations by the MMCAassay. The concordance with PCR Sanger sequencing was 99.3%(840/846), with a kappa value of 0.968 (95% confidence interval,0.972 to 0.996). The concordance for the samples containing mu-tant and wild-type HBV DNA was 94.7% (108/114) and 100%(732/732), respectively.

Finally, we estimated the turnaround time and cost for theMMCA assay. One round of the assay, which can process 48 sam-ples, required 10 min for template addition, 2 h and 10 min forPCR and melting analysis, and 10 min for mutation reading. Thus,the entire assay could be finished within 3 h when HBV DNA andthe PCR master mix are ready for use. Two rounds of MMCA

TABLE 1 Tm for each plasmid and �Tm between wild-type plasmid and mutant plasmid

Resistance Amino acid alteration

Tm � 2SDs (°C)

��Tmc � 2SDs (°C) (n 10)WTa (n 10) MTb (n 10)

LMV and LdT rtL80I 52.5 � 1.10 58.1 � 0.42 �5.65 � 1.06rtL80V 52.5 � 1.10 58.1 � 0.48 �5.70 � 1.26

LMV rtV173G 59.1 � 0.32 51.6 � 0.32 7.50 � 0.47rtV173L 59.1 � 0.32 55.1 � 0.32 4.00 � 0.47

LMV and LdT rtL180M 60.5 � 0.32 65.0 � 0.32 �4.60 � 0.42

ADV and LdT rtA181T 60.5 � 0.32 55.5 � 0.47 4.95 � 0.57rtA181V 60.5 � 0.32 52.5 � 0.32 7.90 � 0.42

ETV rtT184A 58.6 � 0.42 54.0 � 0.32 4.65 � 0.48rtT184S 58.6 � 0.42 54.6 � 0.63 4.00 � 0.81rtT184L 58.6 � 0.42 49.0 � 0.32 9.65 � 0.72rtT184I 58.6 � 0.42 47.5 � 0.32 11.15 � 0.48rtT184C 58.6 � 0.42 46.5 � 0.63 12.00 � 0.48rtT184F 58.6 � 0.42 46.5 � 0 12.10 � 0.42rtT184M 58.6 � 0.42 46.5 � 0.32 12.05 � 0.56rtT184G 58.6 � 0.42 45.5 � 0.32 13.05 � 0.32

TDF rtA194T 68.2 � 0.48 64.1 � 0.42 4.05 � 0.74

ETV rtS202C 60.0 � 0.47 53.6 � 0.32 6.45 � 0.56rtS202G 60.0 � 0.47 53.1 � 0.42 6.90 � 0.63rtS202I 60.0 � 0.47 54.1 � 0.42 5.90 � 0.63

LMV and LdT rtM204I 63.6 � 0.32 53.5 � 0.32 10.10 � 0.42rtM204S 63.6 � 0.32 49.6 � 0.42 13.95 � 0.56rtM204V 63.6 � 0.32 56.0 � 0.32 7.60 � 0.42

TDF rtI233V 53.9 � 0.42 58.3 � 0.84 �4.40 � 0.79ADV rtN236T 52.2 � 0.67 56.2 � 0.67 �4.00 � 1.05a WT, wild type.b MT, mutant type.c �Tm, Tm for wild type � Tm for mutant type.

FIG 2 Readout of TET channel in reaction A for serial dilutions of wild-type plasmids and rtM204V mutant plasmids. Red lines, melting curve for plasmids with5 copies per reaction; black lines, melting curves for plasmids with concentrations of 5 � 106, 5 � 105, 5 � 104, 5 � 103, 5 � 102, and 5 � 101 copies per reaction;gray lines, negative control.

Mou et al.

2664 jcm.asm.org November 2016 Volume 54 Number 11Journal of Clinical Microbiology

on Novem

ber 21, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

assay could easily be completed in one working day, during which96 samples could be processed with a single real-time PCR instru-ment. The cost of materials for the MMCA assay was estimated tobe $4 for each sample. The overall cost of materials for one samplewould be less than $10 when other consumables, e.g., DNA extrac-tion reagents and tips, are included. The cost of instrumentationfor our assay depends on the thermocycler used. In this study, weused a Bio-Rad CFX-96 thermocycler, which costs approximately$30,000. In comparison, the cost of instrumentation for otherHBV drug resistance assays currently available is much higher. Forexample, the instrument used to perform the line probe assay(Auto-LIA 48; Fujirebio Europe, Ghent, Belgium) is close to$100,000, and the cost of the genetic analyzer used for Sangersequencing (Applied Biosystems 3500 genetic analyzer; Thermo

Fisher Scientific, Waltham, MA, USA) is at least $150,000. Clearly,the short turnaround time and low financial barrier of the MMCAassay would facilitate its implementation in LMICs.

DISCUSSION

We examined a two-reaction MMCA assay for anti-HBV drugresistance mutations. The MMCA assay could reliably detect 24mutations at 10 amino acid positions in the reverse transcriptaseregion of the HBV polymerase gene. The assay had a limit of de-tection of 5 copies per reaction. The ability of MMCA to detectminor variant populations was better than that of PCR Sangersequencing and equivalent to that of COLD-PCR Sanger sequenc-ing. Clinical validation studies using 846 valid serum samples ob-tained from different patient populations demonstrated the accu-

FIG 3 Comparison of PCR Sanger sequencing, COLD-PCR Sanger sequencing, and the MMCA method for minority rtM204V mutation detection. (Left) Thereadouts for minor subpopulations of the rtM204V variant obtained using PCR Sanger sequencing (left), COLD-PCR Sanger sequencing (middle), and theMMCA assay (right) are shown. Green curves in right panels, melting curves for wild-type plasmid; black curves in right panels, melting curves for minorityrtM204V mutant plasmid.

Real-Time PCR Detection of Drug-Resistant HBV

November 2016 Volume 54 Number 11 jcm.asm.org 2665Journal of Clinical Microbiology

on Novem

ber 21, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

racy, suitability, and feasibility of the MMCA assay in clinicalsettings.

The establishment and successful implementation of assays todetect anti-HBV drug resistance in resource-limited settings de-pend on a well-planned process of adaptation to and integrationinto relevant national strategies and guidelines (20). HBV DNAquantification is important for decisions on the initiation of anti-viral therapy and monitoring of individuals during antiviral ther-apy (21). HBV DNA viral load assays based on real-time PCR havebeen increasingly available in LMICs (22). Thus, detection of anti-HBV drug resistance mutations could be facilitated by utilizationof the same platform. In this regard, the MMCA assay could bedirectly added to current settings for HBV viral load detection andis a preferred choice for CHB management over assays that needextra instruments. Moreover, the MMCA assay is a closed-tubeprocedure that requires no post-PCR manipulations, which notonly lowers the risk of false-positive results and simplifies the op-erating procedure but also saves the costs for the facility and stafftraining.

The reliability of the MMCA assay was well characterized by itsanalytical performance. First, the MMCA assay covers almost all

important anti-HBV drug resistance mutations. It covers 24 mu-tations at 10 amino acid positions, excluding 3 mutations[rtM250(I/V/L)] at residue rtM250 detected by the HBV DR v.3assay (15). The rtM250(I/V/L) variation is a member of multiplemutations (T184, S202, or M250) that confer ETV resistance in anLVD-resistant HBV background with changes that have alreadyoccurred at L180M and M204(I/V) (23). However, rtM250(I/V/L)is rare compared with the incidence of mutations at T184 andS202 (23, 24), and its accurate detection is challenging due to thepresence of various polymorphic sites in the neighboring nucleo-tides. For example, Degertekin et al. (15) reported that among 240detected mutations, the HBV DR v.3 assay detected M250L inthree samples, but sequencing analysis or follow-up studies couldnot confirm this result for any of them. Thus, omission of rtM250would not compromise the overall performance of the MMCAassay. Second, the MMCA assay could detect minor variant sub-populations with a sensitivity that was equal to that of COLD-PCRSanger sequencing but better than that of PCR Sanger sequencing.The plasmid DNA experiments showed that both MMCA andCOLD-PCR Sanger sequencing could reliably detect rtM204Vwhen it was present at a prevalence of as low as 5% of the total

TABLE 2 Mutations detected by MMCA and PCR Sanger sequencing in 212 clinical samples from NA therapy population in China

Amino acid alteration

No. of mutations detected by: Agreement (%)

MMCA Sequencing

Mutation TotalWTa MTb WT MT

rtL80I/V 192 20 194 18 90.0 99.1rtV173L 211 1 211 1 100 100rtV173G 211 1 211 1 100 100rt(L/V)180M 172 40 172 40 100 100rtA181T 206 6 206 6 100 100rtA181V 203 9 203 9 100 100rtT184(A/C/F/G/I/L/M/S) 208 4 208 4 100 100rtA194T 212 0 212 0 100 100rtS202(C/G/I) 210 2 211 1 50.0 99.5rtM204I 158 54 158 54 100 100rtM204V 185 27 185 27 100 100rtM204S 212 0 212 0 100 100rtI233V 212 0 212 0 100 100rtN236T 197 15 200 12 80.0 98.6a WT, wild type.b MT, mutant type.

FIG 4 Mutant HBV strains detected by MMCA from three different groups of patients. (A) Mutant HBV strains detected from NA therapy patient populationin China; (B) mutant HBV strains detected from health examination population in China; (C) mutant HBV strains detected from health examination populationin Pakistan. Only results for those HBV-positive samples confirmed by this study (total number, 846) are included.

Mou et al.

2666 jcm.asm.org November 2016 Volume 54 Number 11Journal of Clinical Microbiology

on Novem

ber 21, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

population, while conventional PCR Sanger sequencing could de-tect rtM204V only when it was present at a prevalence of greaterthan 10%. When it was applied to clinical samples, MMCA de-tected mutations in 6 more samples from the group A populationthan PCR Sanger sequencing, and all of them were confirmed byCOLD-PCR Sanger sequencing. Although no systematic compar-ison with COLD-PCR Sanger sequencing was made in this work,our results demonstrate that MMCA can be used for the earlydetection of the emergence of anti-HBV drug resistance muta-tions. Third, MMCA had a high reproducibility for the determi-nation of Tm values. Indeed, in these experiments, MMCA pro-vided consistent Tm values regardless of the template amount (25).In this study, the SD values for Tm were smaller than 1°C, while thechange in Tm (�Tm) values were larger than 4°C, ensuring theunambiguous detection of the mutations. Polymorphic nucleo-tides might occur in the proximity of the mutation and exert aninfluence on the Tm values. By using flexible design strategies inprobe design, such an influence could be eliminated (26). Finally,the high analytical sensitivity of the MMCA assay ensured its ap-plicability to samples with varied amounts of HBV DNA.

The robustness of the MMCA assay was demonstrated by test-ing three groups of clinical samples. Of the total of 846 HBV-positive samples, 840 (99.3%) samples gave results concordantwith those of PCR Sanger sequencing. Specifically, of the overall8,460 amino acids analyzed, 8,454 (99.9%) amino acids gave re-sults identical to those obtained by PCR Sanger sequencing. Fur-thermore, the 6 samples with inconsistent results were found tocontain minor mutant subpopulations undetected by PCR Sangersequencing, and their presence was confirmed by COLD-PCRSanger sequencing. These results are in line with those of ourprevious studies that MMCA never missed a single mutation de-tected by PCR Sanger sequencing (18). The three groups of sam-ples were representative of those from two types of patients,treated (group A) and untreated (groups B and C). Among thetreated patients, 48.6% were found to be infected with isolates thatcontained drug resistance mutations. In contrast, among the un-treated group, isolates from only 1.7% contained drug resistancemutations. This result was expected, as most anti-HBV drug re-sistance mutations are produced by long-term NA treatment. Thefrequency of mutation was as follows: LMV monoresistance �ADV monoresistance � LMV and ETV resistance � LMV andADV resistance. This order is understandable, as LMV was the firstNA drug cleared by FDA in 1998 (27), followed by ADV (28) andETV (29). Also, LMV and ADV resistance can be caused by a singlemutation, whereas ETV resistance is caused by multiple muta-tions. Notably, among the untreated patients, the Chinese groupshad more mutation-containing samples than the Pakistani group(3.3% versus 0.56%), indicating that Chinese people have agreater chance of being infected by drug-resistant HBV than Pak-istani people. This result reflects the fact that NAs are more com-monly used to treat CHB in China than in Pakistan.

In conclusion, the MMCA assay could reliably detect 24 anti-HBV drug resistance mutations. It was more sensitive than PCRSanger sequencing in the detection of minor variant subpopula-tions. Because the MMCA assay is technically simple, rapid, inex-pensive, and amenable to high throughput, it could be recom-mended as a tool of choice for detection of anti-HBV resistancemutations in LMICs with high CHB burdens.

ACKNOWLEDGMENTS

We thank Xinlin Zhao for critical reading of the manuscript.The assay described has been transferred to Zeesan Biotech from Xia-

men University. Q.L. holds equity interest in Zeesan Biotech.This work was supported in part by the National Natural Science

Foundation (no. 81271929 to Q.L.) and the Key Project of CooperationProgram for University and Industry of Fujian Province (no. 2014Y4007to Y.L.).

FUNDING INFORMATIONThis work, including the efforts of Yiqun Liao, was funded by Key Projectof Cooperation Program for University and Industry of Fujian Province(2014Y4007). This work, including the efforts of Qingge Li, was funded byNational Natural Science Foundation of China (NSFC) (81271929).

REFERENCES1. Ganem D, Prince AM. 2004. Hepatitis B virus infection—natural history

and clinical consequences. N Engl J Med 350:1118 –1129. http://dx.doi.org/10.1056/NEJMra031087.

2. Wang Y-J, Yang L, Zuo J-P. 2016. Recent developments in antiviralsagainst hepatitis B virus. Virus Res 213:205–213. http://dx.doi.org/10.1016/j.virusres.2015.12.014.

3. Hoofnagle JH, Doo E, Liang TJ, Fleischer R, Lok AS. 2007. Managementof hepatitis B: summary of a clinical research workshop. Hepatology 45:1056 –1075. http://dx.doi.org/10.1002/hep.21627.

4. Zoulim F, Locarnini S. 2009. Hepatitis B virus resistance to nucleos(t)ideanalogues. Gastroenterology 137:1593–1608.e1-2. http://dx.doi.org/10.1053/j.gastro.2009.08.063.

5. Lok AS, Zoulim F, Locarnini S, Bartholomeusz A, Ghany MG, Paw-lotsky JM, Liaw YF, Mizokami M, Kuiken C. 2007. Antiviral drug-resistant HBV: standardization of nomenclature and assays and recom-mendations for management. Hepatology 46:254 –265. http://dx.doi.org/10.1002/hep.21698.

6. Brahmania M, Feld J, Arif A, Janssen HL. 2016. New therapeutic agentsfor chronic hepatitis B. Lancet Infect Dis 16:e10 – e12. http://dx.doi.org/10.1016/S1473-3099(15)00436-3.

7. Lindström A, Odeberg J, Albert J. 2004. Pyrosequencing for detection oflamivudine-resistant hepatitis B virus. J Clin Microbiol 42:4788 – 4795.http://dx.doi.org/10.1128/JCM.42.10.4788-4795.2004.

8. Li J, Wang L, Mamon H, Kulke MH, Berbeco R, Makrigiorgos GM.2008. Replacing PCR with COLD-PCR enriches variant DNA sequencesand redefines the sensitivity of genetic testing. Nat Med 14:579 –584. http://dx.doi.org/10.1038/nm1708.

9. Li J, Milbury CA, Li C, Makrigiorgos GM. 2009. Two-round coampli-fication at lower denaturation temperature-PCR (COLD-PCR)-basedSanger sequencing identifies a novel spectrum of low-level mutations inlung adenocarcinoma. Hum Mutat 30:1583–1590. http://dx.doi.org/10.1002/humu.21112.

10. Wong DK-H, Tsoi O, Huang F-Y, Seto W-K, Fung J, Lai C-L, YuenM-F. 2014. Application of coamplification at lower denaturation temper-ature-PCR sequencing for early detection of antiviral drug resistance mu-tations of hepatitis B virus. J Clin Microbiol 52:3209 –3215. http://dx.doi.org/10.1128/JCM.00343-14.

11. Solmone M, Vincenti D, Prosperi MCF, Bruselles A, Ippolito G, Ca-pobianchi MR. 2009. Use of massively parallel ultradeep pyrosequencingto characterize the genetic diversity of hepatitis B virus in drug-resistantand drug-naive patients and to detect minor variants in reverse transcrip-tase and hepatitis BS antigen. J Virol 83:1718 –1726. http://dx.doi.org/10.1128/JVI.02011-08.

12. Margeridon-Thermet S, Shulman NS, Ahmed A, Shahriar R, Liu T,Wang C, Holmes SP, Babrzadeh F, Gharizadeh B, Hanczaruk B. 2009.Ultra-deep pyrosequencing of hepatitis B virus quasispecies from nucleo-side and nucleotide reverse-transcriptase inhibitor (NRTI)-treated pa-tients and NRTI-naive patients. J Infect Dis 199:1275–1285. http://dx.doi.org/10.1086/597808.

13. Tran N, Berne R, Chann R, Gauthier M, Martin D, Armand M-A,Ollivet A, Teo C, Ijaz S, Flichman D. 2006. European multicenterevaluation of high-density DNA probe arrays for detection of hepatitis Bvirus resistance mutations and identification of genotypes. J Clin Micro-biol 44:2792–2800. http://dx.doi.org/10.1128/JCM.00295-06.

14. Kim H-S, Han K-H, Ahn SH, Kim E-O, Chang H-Y, Moon M, Chung

Real-Time PCR Detection of Drug-Resistant HBV

November 2016 Volume 54 Number 11 jcm.asm.org 2667Journal of Clinical Microbiology

on Novem

ber 21, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

H, Yoo W, Kim S-O, Hong SP. 2005. Evaluation of methods for moni-toring drug resistance in chronic hepatitis B patients during lamivudinetherapy based on mass spectrometry and reverse hybridization. AntivirTher 10:441– 449.

15. Degertekin B, Hussain M, Tan J, Oberhelman K, Lok AS. 2009. Sensi-tivity and accuracy of an updated line probe assay (HBV DR v.3) in de-tecting mutations associated with hepatitis B antiviral resistance. J Hepatol50:42– 48. http://dx.doi.org/10.1016/j.jhep.2008.08.020.

16. Ntziora F, Paraskevis D, Haida C, Manesis E, Papatheodoridis G,Manolakopoulos S, Elefsiniotis I, Karamitros T, Vassilakis A, HatzakisA. 2013. Ultrasensitive amplification refractory mutation system real-time PCR (ARMS RT-PCR) assay for detection of minority hepatitis Bvirus-resistant strains in the era of personalized medicine. J Clin Microbiol51:2893–2900. http://dx.doi.org/10.1128/JCM.00936-13.

17. Huang Q, Liu Z, Liao Y, Chen X, Zhang Y, Li Q. 2011. Multiplexfluorescence melting curve analysis for mutation detection with dual-labeled, self-quenched probes. PLoS One 6:e19206. http://dx.doi.org/10.1371/journal.pone.0019206.

18. Xiong F, Huang Q, Chen X, Zhou Y, Zhang X, Cai R, Chen Y, Xie J,Feng S, Wei X. 2011. A melting curve analysis-based PCR assay forone-step genotyping of -thalassemia mutations: a multicenter valida-tion. J Mol Diagn 13:427– 435. http://dx.doi.org/10.1016/j.jmoldx.2011.03.005.

19. Higuchi R, Krummel B, Saiki R. 1988. A general method of in vitropreparation and specific mutagenesis of DNA fragments: study of proteinand DNA interactions. Nucleic Acids Res 16:7351–7367. http://dx.doi.org/10.1093/nar/16.15.7351.

20. World Health Organization. 2015. Guidelines for the prevention, careand treatment of persons with chronic hepatitis B infection. World HealthOrganization, Geneva, Switzerland.

21. Chu CJ, Hussain M, Lok AS. 2002. Quantitative serum HBV DNA levelsduring different stages of chronic hepatitis B infection. Hepatology 36:1408 –1415. http://dx.doi.org/10.1002/hep.1840360617.

22. Abe A, Inoue K, Tanaka T, Kato J, Kajiyama N, Kawaguchi R, Tanaka

S, Yoshiba M, Kohara M. 1999. Quantitation of hepatitis B virus genomicDNA by real-time detection PCR. J Clin Microbiol 37:2899 –2903.

23. Baldick CJ, Tenney DJ, Mazzucco CE, Eggers BJ, Rose RE, PokornowskiKA, Yu CF, Colonno RJ. 2008. Comprehensive evaluation of hepatitis Bvirus reverse transcriptase substitutions associated with entecavir resis-tance. Hepatology 47:1473–1482. http://dx.doi.org/10.1002/hep.22211.

24. Tenney DJ, Rose RE, Baldick CJ, Pokornowski KA, Eggers BJ, Fang J,Wichroski MJ, Xu D, Yang J, Wilber RB. 2009. Long-term monitoringshows hepatitis B virus resistance to entecavir in nucleoside-naïve patientsis rare through 5 years of therapy. Hepatology 49:1503–1514. http://dx.doi.org/10.1002/hep.22841.

25. Hu S, Li G, Li H, Liu X, Niu J, Quan S, Wang F, Wen H, Xu Y, Li Q.2014. Rapid detection of isoniazid resistance in Mycobacterium tubercu-losis isolates by use of real-time-PCR-based melting curve analysis. J ClinMicrobiol 52:1644 –1652. http://dx.doi.org/10.1128/JCM.03395-13.

26. Margraf RL, Rong M, Wittwer CT. 2006. Masking selected sequencevariation by incorporating mismatches into melting analysis probes. HumMutat 27:269 –278. http://dx.doi.org/10.1002/humu.20290.

27. Lai C-L, Chien R-N, Leung NW, Chang T-T, Guan R, Tai D-I, Ng K-Y,Wu P-C, Dent JC, Barber J. 1998. A one-year trial of lamivudine forchronic hepatitis B. N Engl J Med 339:61– 68. http://dx.doi.org/10.1056/NEJM199807093390201.

28. Peters MG, Hann HW, Martin P, Heathcote EJ, Buggisch P, Rubin R,Bourliere M, Kowdley K, Trepo C, Gray D. 2004. Adefovir dipivoxilalone or in combination with lamivudine in patients with lamivudine-resistant chronic hepatitis B. Gastroenterology 126:91–101. http://dx.doi.org/10.1053/j.gastro.2003.10.051.

29. Sherman M, Yurdaydin C, Sollano J, Silva M, Liaw YF, Cianciara J,Boron-Kaczmarska A, Martin P, Goodman Z, Colonno R, Cross A,Denisky G, Kreter B, Hindes R, AI463026 BEHoLD Study Group. 2006.Entecavir for treatment of lamivudine-refractory, HBeAg-positivechronic hepatitis B. Gastroenterology 130:2039 –2049. http://dx.doi.org/10.1053/j.gastro.2006.04.007.

Mou et al.

2668 jcm.asm.org November 2016 Volume 54 Number 11Journal of Clinical Microbiology

on Novem

ber 21, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from