Sensitive and accurate quantitation of hepatitis B virus DNA using a kinetic fluorescence detection system (TaqMan PCR)

Journal of Virological Methods 85 (2000) 75 – 82 www.elsevier.com/locate/jviromet

Sensitive and accurate quantitation of hepatitis B virus DNA using a kinetic fluorescence detection system (TaqMan PCR) Klaus M. Weinberger *, Elisabeth Wiedenmann, Stephan Bo¨hm, Wolfgang Jilg Institute for Medical Microbiology and Hygiene, Uni6ersity of Regensburg, Franz-Josef-Strauß-Allee 11, D-93053 Regensburg, Germany Received 29 July 1999; received in revised form 7 October 1999; accepted 8 October 1999

Abstract The laboratory diagnosis of hepatitis B virus (HBV) infection is based mainly on serological assays. Yet the detection and quantitation of viral DNA is necessary when addressing directly the question of infectivity or when monitoring the viral load during therapy. Standard hybridization assays allow for exact quantitation, but their sensitivity is limited to 105 –106 viral genomes per ml of serum. The most sensitive tests for HBV DNA are nested PCR systems, which recognize virtually one molecule of the target DNA per reaction. However, these assays only provide very coarse quantitative statements. To take advantage of both methods, a new assay for HBV DNA is described based on the commercial TaqMan© system. This assay is capable of quantifying HBV DNA from the theoretical lower limit up to 1010 genome equivalents per ml of serum and, thus, covers the complete range of naturally occurring states of infections. The method was calibrated on the basis of serial plasmid dilutions and compared with a well-established nested PCR system. More than 100 HBV positive sera and serial dilutions of the Eurohep standard for both ad and ay subtypes were analyzed. The assay reliably detected all HBV positive samples. It shows minimal run-to-run deviations, allows for quantitation that covers eight orders of magnitude, and finally, completely avoids the risk of cross-contamination by PCR products. Thus, this technique combines the sensitivity of PCR amplification and the quantitation potential of hybridization tests and it is time efficient and safer. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Diagnostics; Infectivity; Viral load; Therapy monitoring

1. Introduction * Corresponding author. Tel.: +49-941-944-6429; fax: + 49-941-944-6402. E-mail address: [email protected] (K.M. Weinberger)

Hepatitis B virus (HBV) is one of the most efficient parenterally transmitted viruses (Tabor et al., 1983) and requires special care regarding diag-

0166-0934/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 6 - 0 9 3 4 ( 9 9 ) 0 0 1 5 4 - 8

76

K.M. Weinberger et al. / Journal of Virological Methods 85 (2000) 75–82

nostic assays as well as safety measures for blood and blood products (Hoofnagle, 1990). High sensitivity is, therefore, one major goal for all HBV detection systems. On the other hand, the virus replicates to very high titers; 109 or even more viral particles per ml of serum are seen frequently in patients with acute or chronic infection (Zyzik et al., 1986; Kann and Gerlich, 1998). Quantitative assays should, therefore, cover a wide range of concentrations since the absolute viral load, as well as its time course, are valuable prognostic and therapeutic parameters (Decker et al., 1988; Pastore et al., 1992; Lok, 1994; Steele, 1995). The standard laboratory techniques for the diagnosis of HBV are serological assays (mainly modified EIA systems) for the viral surface antigen (HbsAg) and a secreted non-structural antigen (HBsAg), as well as for antibodies against the surface (anti-HBs), core (anti-HBc) and HBe antigens (anti-HBe). Although the combination of these assays is well-established, there are instances where ambiguous or contradictory results need to be clarified by direct detection of the viral DNA (Kaneko et al., 1990; Gerlich et al., 1995). The discrepancy between the competing goals of sensitivity and quantitation appears most obvious. The classical hybridization techniques are very reliable and allow for the direct quantitation of HBV DNA without any non-linear amplification steps. Yet, the lower limit of detection for these assays is approximately 105 – 106 genome equivalents per ml of serum (Kessler et al., 1998), and this threshold is far too high to exclude the risk of infectivity with reasonable certainty. Moreover, during therapy with interferon or polymerase inhibitors, such as lamivudine or famciclovir, many patients have readily negative hybridization results although they do not eliminate the virus from the liver or the peripheral blood. In these cases, the time course of the viral load can no longer be monitored (Kessler et al., 1998; Bo¨hm et al., unpublished data). During the last 10 years the polymerase chain reaction (PCR) has proved to be useful for the diagnosis of HBV (Sumazaki et al., 1989; Zaaijer et al., 1994). Used in combination with a nested set of primers (nested PCR), it is able to detect virtually one single molecule of the target DNA

and is, therefore, the most sensitive assay for detecting HBV in many different specimens, such as serum or even solid tissue. The techniques for isolating viral nucleic acids from these sources have become the limiting factor for the detection threshold of the systems. However, PCR requires a number of particular safety measures to avoid cross-contamination (Kwok and Higuchi, 1989) and trained personnel in order to benefit completely from the method’s potential. The reaction products have to be detected by analytical agarose gel electrophoresis or secondary tests, such as Southern blotting, which always represent end-point methods. Since the amount of PCR products, especially in nested reactions, is usually limited by the concentration of any of the reaction components (primers, nucleotides), this end-point analysis shows only a very indirect correlation to the amount of template DNA. However, a number of groups have evaluated quantitative assays using signal amplification (Watanabe et al., 1993; Khakoo et al., 1996), one-step PCR with different secondary tests (Ha¨mmerle et al., 1997; Payan et al., 1997; Simpson et al., 1997; Kessler et al., 1998) or semi-nested PCR (Matsumoto et al., 1997). To achieve accurate quantitation of the template, the kinetics would have to be observed, but with standard detection systems this would mean interrupting the PCR and analyzing samples at different times during the reaction. To combine the highest sensitivity with an exact quantitation, a TaqMan PCR assay for hepatitis B virus DNA was developed based on kinetic fluorescence detection of the amplified products during PCR.

2. Material and methods

2.1. Serology All sera used for this study were sent to the Institute for Medical Microbiology and Hygiene for diagnostic purposes. They were either processed immediately or stored at − 20°C. Qualitative serological tests for HBsAg, anti-HBc, anti-HBc-IgM, HBeAg, and anti-HBe, and the quantitation of anti-HBs, were carried out using

K.M. Weinberger et al. / Journal of Virological Methods 85 (2000) 75–82

standard commercial microparticle enzyme immunoassays (AxSym, Abbott Laboratories, North Chicago, IL).

77

copies per reaction; for sequence reference see GenBank entry X51970).

2.3. TaqMan assay and quantitation 2.2. DNA isolation and con6entional amplification Viral nucleic acid was prepared from serum samples using the QiaAmp blood and tissue kit (QiaGen, Hilden, Germany), as described previously (Weinberger et al., 1999). Briefly, 200 ml of serum were used for the lysis reaction with chaotropic salts, detergents and a final proteinase K concentration of 1 mg/ml. After the addition of 2-propanol to decrease solubility, the DNA was bound to spin columns and washed twice before elution in 100 ml 10 mM Tris – HCl, pH 8.0. It was either used immediately or stored at − 20°C. This preparation (10 ml) served as the template for both conventional and TaqMan PCR assays. The primers and reaction conditions used for amplifying specifically a fragment of the HBV surface antigen gene (HBs-gene) have been described previously (Jilg et al., 1995) and are listed in Table 1. First-round products were hybridized to a digoxigenin-labelled probe generated with the nested pair of primers and detected using biotincoupled alkaline phosphatase with CSPD as the bioluminescence substrate. Semi-quantitative results were achieved by comparison with a 10-fold dilution series of plasmid pHBV991 (106 to 1

The sequences of the probe and the primers were chosen carefully from conserved regions of the hepatitis B virus surface gene, using the Primer Express software supplied by the manufacturer of the TaqMan Sequence Detection System (Perkin-Elmer –Applied Biosystems, Foster City, CA). They fulfil the standard requirements, such as similar melting temperatures of the primers (maximum DTm = 2°C), shortness of the amplicon, and suggested sequence properties of the probe (low content of G). All three oligonucleotides are identical with more than 95% of all published HBV-gene sequences (as contained in GenBank release 110.0, 12/98), as well as the 3%-half of both primers and the 5%-half of the probe with more than 99%. The fluorogenic probe is 5%-labelled with FAM (6-carboxyfluoresceine) and 3%-labelled with TAMRA (6-carboxytetramethylrhodamine), which serves as a quenching dye (sequences are listed in Table 1). After several steps of optimization the components of the amplification reaction were chosen at the following concentrations: 50 mM KCl; 10 mM Tris–HCl, pH 8.3; 5 mM MgCl2; 400 nM dUTP; 200 nM dATP; 200 nM dCTP; 200 nM dGTP;

Table 1 Oligonucleotides used as primers and probes Oligo

Function

Sequence (5%–3%)

Positiona

HBV-S1

Conventional PCR first round forward primer Conventional PCR first round reverse primer Conventional PCR nested forward primer Conventional PCR nested reverse primer TaqMan PCR forward primer TaqMan PCR reverse primer TaqMan PCR probe

CTC GTG TTA CAG GCG GGG TTT TTC

191–214

CAT CAT CCA TAT AGC TGA AAG CCA AAC A

748–721

TTG TTG ACA AGA ATC CTC ACA ATA CC

215–240

GCC CTA CGA ACC ACT GAA CAA ATG G

710–686

CAA CCT CCA ATC ACT CAC CAA C ATA TGA TAA AAC GCC GCA GAC AC FAM-TCC TCC AAT TTG TCC TGG TTA TCG CTTAMRA

321–342 401–379 349–374

HBV-S2 HBV-S3 HBV-S4 HBV-Taq 1 HBV-Taq 2 BS-1

a

Nucleotide position according to Genbank entry X51970.

78

K.M. Weinberger et al. / Journal of Virological Methods 85 (2000) 75–82

Fig. 1. Example of the mathematical fitting of the serial plasmid dilutions. The threshold cycle (average of triplicates 9S.D.) is plotted against the decadic logarithm of the number of plasmid molecules per reaction. Linear regression yields a theoretical ordinate value for one template molecule (y0), the gradient in cycles per order of magnitude (m) and a coefficient of reliability (r 2) close to the optimum of 1.

300 nM oligonucleotide primers HBV-Taq1 and HBV-Taq2; 200 nM TaqMan probe BS-1; 0.5 U uracyl-N-glycosylase (UNG); 1.25 U AmpliTaq gold DNA polymerase and 10 ml of template DNA preparation per 50 ml reaction (all reagents Perkin-Elmer, Branchburg, NJ). After an initial incubation at 50°C for 2 min to activate the UNG and a denaturation phase of 10 min at 95°C, the temperature profile followed a two-step cycle pattern with a combined annealing and primer extension phase at 60°C for 1 min and a short denaturation at 95°C for 15 s. Fifty cycles were carried out routinely and a threshold of fluorescence was defined in order to exclude samples which were clearly negative as seen from the shape of their curves. Specific amplification is characterized by sigmoid curves, with a sharp exponential rise and reaching of a plateau soon after. For each run triplicates of a serial dilution of plasmid pHBV991 containing 1 – 108 copies of the HBV genome per reaction were analyzed. Based

on the mean threshold cycles for each dilution, a linear regression was carried out with the threshold cycle as a function of the decadic logarithm of the number of template molecules per reaction (for example see Fig. 1). On the basis of this regression analysis, the content of template molecules per reaction was calculated for each serum sample, again using the median of triplicates. Samples with threshold cycles greater than the calculated ordinate value for one plasmid molecule per reaction were defined as negative. In cases, where the S.D. had the same order of magnitude as the calculated result, the results were defined as unreliable and had to be repeated.

3. Results

3.1. Plasmid dilutions After the optimization experiments, plasmid

K.M. Weinberger et al. / Journal of Virological Methods 85 (2000) 75–82

concentrations ranging from 1 to 108 copies per reaction were tested in triplicates. Down to ten copies per reaction, all single values were positive. At the one molecule level, 54.2% of the reactions were positive, corresponding well to the expected portion. The consistency between the different runs is shown in Fig. 2, where seven such experiments are compared in one diagram. Between concentrations of 102 and 108 DNA molecules per reaction, two dilution steps were differentiated typically by more than 30 S.D.

3.2. Patients’ sera To test the reliability of the new method and to compare it with an in-house semi-quantitative PCR system, 111 HBsAg-positive, seven antiHBc-only positive sera and 50 negative control samples were examined. In the group of HBsAgpositives, all 111 were viraemic with an average titer of 5.4× 107 genome equivalents per ml of serum (maximum= 1.7 × 109). Seven sera were positive for HBeAg, four of which also had anti-

79

HBe-IgM. The mean titer in this group was 3.6× 108 per ml (range= 1.1× 103 –1.7×109). For the anti-Hbe positive sera (n=104, 92.5%), the average was only slightly lower, with 4.3× 107 per ml (range= 52–7.7× 108). Of the seven anti-HBconly positive sera, three were DNA-positive with an average of 200 copies of the viral genome per ml. Fig. 3 summarizes graphically the results of this comparison and demonstrates that a linear regression coincides nicely with the data. Yet, the range of the conventional PCR is obviously too narrow to cover the naturally occurring virus titers.

3.3. Eurohep standard Triplicates of the Eurohep standards (Gerlich et al., 1995) for HBV subtypes ad and ay were tested undiluted (109 genome equivalents per ml of plasma) and in 10-fold dilutions, down to ten copies per ml. The sensitivity and accuracy for this series is shown in Fig. 4. As expected for a theoretical detection limit of 50 copies per ml, all samples down to a concentration of 100 genome

Fig. 2. Comparison of the plasmid standard curves from seven different runs. In the range 10 – 108 template molecules per reaction, the curves show good linearity and only minimal run-to-run deviations.

80

K.M. Weinberger et al. / Journal of Virological Methods 85 (2000) 75–82

Fig. 3. Comparison with the in-house semi-quantitative PCR system (n =100). The results from the TaqMan assay are plotted against the discrete values of the diagnostic test. Linear regression of the data (r 2 =0.8341) shows a gradient that is slightly too low, i.e. the range of the conventional system is too narrow to cover all virus titers found in patients.

equivalents per ml yielded positive results. The quantitation was almost linear over the complete range of seven orders of magnitude and showed no significant difference in the detection of both subtypes.

4. Discussion The aim of the present study was to apply the technical features of the TaqMan PCR system (Heid et al., 1996), as published recently for bacterial and viral pathogens (Kennedy et al., 1998; McGoldrick et al., 1998; Berg et al., 1999; Kawai et al., 1999; Pusterla et al., 1999), to establish a highly sensitive and accurate assay for HBV DNA that should be able to monitor the wide range of virus concentrations found in acute or chronic carriers of HBV. Extensive studies based on plasmid dilutions, on panels of viraemic patients, and on the Eurohep reference plasma samples clearly demon-

strated that amplification techniques, such as the Amplicor™ PCR monitor (Kessler et al., 1998) or ‘in-house’ PCR assays with different detection strategies (Payan et al., 1997; Simpson et al., 1997), are able to yield similar quantitation accuracy, when compared with conventional hybridization tests (e.g. Digene™ hybrid capture). In contrast, other groups either reported assays with narrow ranges of linearity (Ha¨mmerle et al., 1997) or with very broad estimates of the quantity of viral DNA (Matsumoto et al., 1997). Systems based on signal amplification instead of template amplification (e.g. branched DNA assays) should guarantee optimal linearity; however, they still lack sufficient sensitivity when compared with PCR (Khakoo et al., 1996). All PCR methods described, however, have one additional problem in common: after amplification, highly concentrated solutions of specific DNA fragments must be handled for detection by electrophoresis, Southern blotting (Southern, 1979), ELOSA (Mallet et al., 1993) or scintillation

flK.M. Weinberger et al. / Journal of Virological Methods 85 (2000) 75–82

81

Fig. 4. Analysis of the Eurohep standard samples. The calculated result of the TaqMan assay (average of triplicates 9 S.D.) is plotted against the concentration of the serial dilution for both HBV subtypes; the theoretical curve is given in dashes. The linearity of the quantitation is obvious and there is no significant difference in the accuracy or sensitivity between subtypes ad and ay.

assays and represent a major source of cross-contamination. As the kinetic fluorescence detection of the TaqMan system is carried out through ‘optical caps’ during the reaction without opening the tubes, this risk is avoided completely. Furthermore, the initial incubation with uracyl-N-glycosylase degrades amplification products and allows only amplification of authentic template DNA. The strict need for a three room organization of PCR laboratories could, therefore, be reduced to two rooms. This, of course, does not abolish the necessity to optimize further the process of DNA isolation from clinical specimens; because of the very high concentrations found in patients, careful handling to guarantee laminar flow of all fluids by slow pipetting, thus avoiding the formation of aerosols, is particularly important with samples suspected to be positive for HBV. In comparison with commercially available and different ‘in-house’ quantitation systems for HBV DNA, the assay described above offers a number of advantages, such as sensitivity at the theoreti-

cal threshold, a wide range of linearity, high accuracy for different HBV subtypes, and reliability shown by minimal run-to-run deviations. Such an assay is certainly not considered as replacing the fully automated serological screening tests, but could serve as the single diagnostic detection method for HBV DNA.

Acknowledgements Plasmid pHBV991 and the Eurohep standard samples were kindly provided by Prof R. Thomssen, University of Go¨ttingen. The authors wish to thank Barbara Hottentra¨ger for excellent technical assistance in performing the serological assays.

References Berg, T., Mu¨ller, A.R., Platz, K.P., et al., 1999. Dynamics of GB virus C viremia early after orthotopic liver transplanta-

82

K.M. Weinberger et al. / Journal of Virological Methods 85 (2000) 75–82

tion indicates extrahepatic tissues as the predominant site of GB virus C replication. Hepatology 29, 245–249. Decker, R.H., Kuhns, M.C., Brawner, T.A., Mimms, L.T., 1988. Future advanced diagnostic techniques for hepatitis B. In: Zuckerman, A.J. (Ed.), Viral Hepatitis and Liver Disease. Alan R. Liss, New York, pp. 231–236. Eurohep Group, Gerlich, W.H., Heermann, K.H., Thomssen, R., 1995. Quantitative assays for hepatitis B virus DNA: standardization and quality control. Viral Hepat. Rev. 1, 53– 57. Ha¨mmerle, T., Himmelspach, M., Dorner, F., Falkner, F.G., 1997. A sensitive PCR assay system for the quantitation of viral genome equivalents: human immunodeficiency virus type 1 (HIV-1) and hepatitis B virus (HBV). Arch. Virol. 142, 1297 – 1306. Heid, C.A., Stevens, J., Livak, K.J., Williams, P.M., 1996. Real time quantitative PCR. Genome Res. 6, 986–994. Hoofnagle, J.H., 1990. Postransfusion hepatitis B. Transfusion 30, 384 – 386. Jilg, W., Sieger, E., Zachoval, R., Scha¨tzl, H., 1995. Individuals with antibodies against hepatitis B core antigen as the only serological marker for hepatitis B infection: high percentage of carriers of hepatitis B and C virus. J. Hepatol. 23, 14 – 20. Kaneko, S., Miller, R.H., Di Bisceglie, A.M., Feinstone, S.M., Hoofnagle, J.H., Purcell, R.H., 1990. Detection of hepatitis B virus DNA in serum by polymerase chain reaction. Gastroenterology 99, 799–804. Kann, M., Gerlich, W.H., 1998. Hepatitis B. In: Collier, L., Balows, A., Sussman, M. (Eds.), Microbiology and Microbial Infections: Virology, vol. 1. Topley and Wilson, London. Kawai, S., Yokosuka, O., Kanda, T., Imazeki, F., Maru, Y., Saisho, H., 1999. Quantification of hepatitis C virus by TaqMan PCR: comparison with HCV Amplicor monitor assay. J. Med. Virol. 58, 121–126. Kennedy, M.M., O’Leary, J.J., Oates, J.L., et al., 1998. Human herpes virus 8 (HHV-8) in Kaposi’s sarcoma: lack of association with bcl-2 and p53 protein expression. Mol. Pathol. 51, 155 – 159. Kessler, H.H., Pierer, K., Dragon, E., et al., 1998. Evaluation of a new assay for HBV DNA quantitation in patients with chronic hepatitis B. Clin. Diagn. Virol. 9, 37–43. Khakoo, S.I., Soni, P.N., Brown, D., Dusheiko, G.M., 1996. A clinical evaluation of a new method for HBV DNA quantitation in patients with chronic hepatitis B. J. Med. Virol. 50, 112 – 116. Kwok, S., Higuchi, R., 1989. Avoiding false positives with PCR. Nature 339, 237–238. Lok, A.S.F., 1994. Treatment of chronic hepatitis B. J. Viral Hepat. 1, 105 – 124. Mallet, F., Hebrard, C., Brand, D., et al., 1993. Enzymelinked oligosorbent assay for detection of polymerase chain

.

reaction-amplified human immunodeficiency virus type 1. J. Clin. Microbiol. 31, 1444 – 1449. Matsumoto, C., Nishioka, K., Oguchi, T., et al., 1997. Detection and quantitation of HBV DNA by semi-nested PCR in donated blood: comparison with HBV serological markers. J. Virol. Methods 66, 61 – 69. McGoldrick, A., Lowings, J.P., Ibata, G., Sands, J.J., Belak, S., Paton, D.J., 1998. A novel approach to the detection of the classical swine fever virus by RT-PCR with a fluorogenic probe (TaqMan). J. Virol. Methods 72, 125 – 135. Pastore, G., Santantonio, T., Milella, M., et al., 1992. AntiHBe-positive chronic hepatitis B with HBV-DNA in the serum response to a 6-month course of lymphoblastoid interferon. J. Hepatol. 14, 221 – 225. Payan, C., Ve´al, N., Crescenzo-Chaigne, B., Be´lec, L., Pillot, J., 1997. New quantitative assay of hepatitis B and C viruses by competitive PCR using alternative internal sequences. J. Virol. Methods 65, 299 – 305. Pusterla, N., Huder, J.B., Leutenegger, C.M., Braun, U., Madigan, J.E., Lutz, H., 1999. Quantitative real-time PCR for detection of members of the Ehrlichia phagocytophila genogroup in host animals and Ixodes ricinus ticks. J. Clin. Microbiol. 37, 1329 – 1331. Simpson, P.R., Yu, X.H., Redza, Z.M., Anson, J.G., Chan, S.H., Lin, Y., 1997. Quantitation of hepatitis B virus DNA using competitive PCR and a scintillation proximity assay. J. Virol. Methods 69, 197 – 208. Southern, E., 1979. Gel electrophoresis of restriction fragments. Methods Enzymol. 68, 152 – 176. Steele, F.R., 1995. Viral load: to treat or not to treat? Nat. Med. 1, 980. Sumazaki, R., Motz, M., Wolf, H., Heinig, J., Jilg, W., Deinhardt, F., 1989. Detection of hepatitis B virus in serum using amplification of viral DNA by means of the polymerase chain reaction. J. Med. Virol. 27, 304 – 308. Tabor, E., Purcell, R.H., London, W.T., Gerety, R.J., 1983. Use of and interpretation of results using inocula of hepatitis B virus with known infectivity titers. J. Infect. Dis. 147, 531 – 534. Watanabe, Y., Hino, K., Yasuda, K., Iino, S., Suzuki, H., 1993. Detection of HBV-DNA by bDNA probe assay and it’s clinical relevance. Med. Pharmacol. 30, 1217 – 1226. Weinberger, K.M., Zoulek, G., Bauer, T., Bo¨hm, S., Jilg, W., 1999. A novel deletion mutant of hepatitis B virus surface antigen. J. Med. Virol. 58, 105 – 110. Zaaijer, H.L., Ter Borg, F., Cuypers, H.T.M., Hermus, M.C.A.H., Lelie, P.N., 1994. Comparison of methods for detection of hepatitis B virus DNA. J. Clin. Microbiol. 32, 2088 – 2091. Zyzik, E., Gerlich, W.H., Uy, A., Ko¨chel, H., Thomssen, R., 1986. Assay of hepatitis B virus genome titers in sera of infected subjects. Eur. J. Clin. Microbiol. 5, 330 – 335.