Evaluation of the Roche LightCycler parvovirus B19 quantification kit for the diagnosis of parvovirus B19 infections

Journal of Clinical Virology 31 (2004) 5–10

Evaluation of the Roche LightCycler parvovirus B19 quantification kit for the diagnosis of parvovirus B19 infections Sharleen Braham, Jayshree Gandhi, Stuart Beard, Bernard Cohen∗ Enteric, Respiratory and Neurological Virus Laboratory, Specialist and Reference Microbiology Division, Health Protection Agency, 61 Colindale Avenue, London NW9 5HT, UK Received 26 August 2002; received in revised form 12 September 2003; accepted 10 October 2003

Abstract Background: The rapid and quantitative detection of viral DNA is important in the diagnosis of parvovirus B19 infection in immunocompromised patients and in congenital infection. It is also valuable for monitoring progress following therapeutic interventions. Objectives: To evaluate the diagnostic sensitivity and specificity of the Roche LightCycler (LC) parvovirus B19 quantification kit in comparison with previously described nested PCR and dot blot hybridisation assays. Study design: Two hundred and twenty eight clinical samples and two standard B19 DNA sera were tested to assess the diagnositic performance of the Roche LC kit. Results: Ten clinical samples (4.3%) gave invalid LC results, including three of five bone marrow samples but only two of 165 serum samples. In the remaining 218 samples, the LC assay detected B19 DNA in 97.5% (79/81) samples that were positive by the nested PCR. The two samples (from the same patient) that were LC negative were sequenced in a 511-nucleotide region of the NS gene and 42 nucleotide changes were found. The Roche LC assay detected B19 DNA in 9.5% (13/137) samples that were negative by nested PCR. Analysis of the available clinical and serological data associated with these samples suggested that the LC results in the majority of these cases were true positive. In patients with resolving persistent infection, the LC assay remained positive for longer than nested PCR. Conclusions: The Roche LC assay was more sensitive than the nested PCR used in this study. The additional sensitivity and the quantitative DNA measurements were valuable for monitoring patients with persistent B19 infection. Practical advantages of the LC assay include a short running time and the possibility to automate the assay. The LC assay provides a controlled and standardised method for quantitative detection of viral DNA for the diagnosis and monitoring of parvovirus B19 infections but failed to detect a variant strain. © 2003 Elsevier B.V. All rights reserved. Keywords: Parvovirus B19; Real-time PCR; Diagnosis

1. Introduction Detection of parvovirus B19 DNA by PCR assay (Salimans et al., 1989; Clewley, 1989) has been available for a number of years and is valuable for diagnosis, especially in clinical settings where serology is unreliable eg, in immunocompromised patients. Qualitative PCR assays, however, do not distinguish high B19 viral load (>1012 genome copies/ml) present in the acute phase of self-limiting infection from low viral load that may be found months or years later (Clewley, 1989; Azzi et al., 1996). The need for quantitative B19 DNA assays extends to blood donor screening where recent studies have shown that removal of plasma donations containing B19 DNA at ∗ Corresponding author. Tel.: +44-208-200-4400; fax: +44-208-205-8195. E-mail address: [email protected] (B. Cohen).

1386-6532/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.jcv.2003.10.005

a level of >104 genome copies/ml is required to prepare plasma pools for preparation of blood products that are safe with respect to transmission of parvovirus B19 infection (Brown et al., 2001). Quantitative assays are therefore required to identify and remove plasma donations with high B19 viral load to prepare plasma pools that are safe with respect to transmission of parvovirus B19. In the diagnostic setting, quantitative measurement of B19 DNA is important for clinical management because it provides a means of monitoring the response to therapy of severe B19 infections in immunocompromised patients (Azzi et al., 1996; Harder et al., 2001). It also has a role in confirming congenital parvovirus B19 infection (Knöll et al., 2002). Although sensitive and specific, nested PCR assays are at best semi-quantitative unless tedious end-point titrations are performed. Quantification has been achieved by PCR-ELISA either in the format of competitive assay (Gallinella et al., 1997) or by detection of amplicons

6

S. Braham et al. / Journal of Clinical Virology 31 (2004) 5–10

with sensitive labelling methods (Sato et al., 2000; Daly et al., 2002). However, it was not until the development of real-time PCR techniques that convenient and rapid quantification of B19 DNA became feasible. Glass capillary thermal cycling (LightCyler, LC) methods with detection by either DNA-binding dye (Manaresi et al., 2002) or hybridisation probes (Harder et al., 2001) have been described as well as hydrolysis probe-based assays (TaqMan, Aberham et al., 2001). A commercial LC kit for parvovirus B19 (Roche Diagnostics) has recently become available that, as an additional feature, includes an internal control to detect any interference from inhibitory substances in the specimen. In the study described here, the Roche LC kit was evaluated for diagnostic testing and compared to manual nested PCR and hybridisation assays for detection of parvovirus B19 DNA in clinical specimens.

2. Materials and methods 2.1. Clinical specimens A total of 228 clinical samples were tested comprising 165 sera; 15 amniotic fluids; 11 plasmas; 12 fetal tissue homogenates; eight fetal sera; five bone marrows; five fetal ascitic fluids; four fetal pleural fluids; two fetal chest fluids; one placental swab. All samples were received for parvovirus B19 diagnosis during 2000 and 2001 and stored at −20 ◦ C prior to testing in the Roche LC assay. They had previously been tested for B19 DNA by manual nested PCR using primers from the NS gene (Clewley, 1993) and by ‘dot blot’ hybridsation assay using a 4.5 kb PCR-generated DNA probe representing 80% of the B19 genome (Hicks et al., 1995). Where clinically indicated the samples had also been tested for B19 IgM by ‘in-house’ ELISA (Brown et al., 1989) and for B19 IgG by a commercial ELISA (Biotrin International, Dublin, Ireland). 2.2. Sequential sera Seventeen of the 165 sera were sequential samples from two immunocompromised patients with prolonged parvovirus B19 infection. Seven of these sera, collected over a period of 6 months, were from a 21-year-old female lung transplant patient (patient 1) who presented with relapsing red cell aplasia and was treated with intravenous immunoglobulin. Ten sera, collected over a period of 9 months, were from a 6-year-old leukaemia patient (patient 2) who presented with rash and persistent B19 viraemia.

dards and Controls, South Mimms, Herts, UK) which has been assigned a concentration of 106 international units (IU)/ml, equivalent to 2 × 106 genome copies/ml (Saldanha et al., 2002); and an ‘in-house’ standard serum, NAN, obtained by screening blood donors (Hicks et al., 1998). B19 DNA was isolated from the ‘NAN’ serum (Hicks et al., 1998) and its concentration, estimated by optical density measurement at a wavelength of 260 nm, was 1011 genome copies/ml. The concentration of B19 DNA in five standards included in the Roche kit for constructing a calibration curve ranged from 7.1 × 101 to 6.2 × 105 genome copies/5 ␮l in approximately 10-fold increments. 2.4. Roche LC parvovirus B19 quantification kit The LightCycler-parvovirus B19 quantification kit (Roche Diagnostics, Lewes, Sussex, UK) was used according to the manufacturer’s instructions. Automated sample extraction was performed with the MagNA Pure LC instrument and the MagNA Pure LC total nucleic acid isolation kit (Roche Diagnostics). The internal control was added with the lysis buffer. “Post elution” stages (addition of extracted DNA and reaction mix to glass capillaries) were performed either manually or using the MagNA Pure LC instrument and a cooling block carousel. Results from samples failing to give a signal with either the B19 or the internal control target were considered invalid as described in the kit insert. Quantitative results from the kit are obtained in terms of genome copies per 5 ␮l extracted DNA added to the PCR reaction. For this report, genome copies per 5 ␮l were transformed into IU/ml taking into account the sample volume (200 ␮l), the elution volume (50 ␮l), the volume (5 ␮l) of extracted DNA added to the PCR reaction and the kit lot-specific conversion factor of 1 IU = approximately three genome copies. 2.5. Sequencing and phylogenetic analysis A 511-base pair region of the NS gene was amplified using primers ‘I’ and ‘F’ (Clewley, 1993). PCR products were excised from a 2% agarose gel and purified using Geneclean II (BIO 101,Vista,CA 92083). Nucleotide sequences were then determined using the ‘DyeDeoxy Terminator’ method (Applied Biosystems, Perkin-Elmer, Warrington, UK). Primers were used at 3.2 pmol per 312 ng of template DNA. Phylogenetic analysis was performed using the neighbour-joining algorithm in Clustal (Megalign version 1.03, DNASTAR, London, UK).

2.3. Control samples

3. Results

To assess the analytical sensitivity of the Roche LC assay the following control samples were used: the first World Health Organisation international standard for parvovirus B19 DNA (99/800, National Institute for Biological Stan-

3.1. Control samples The mean of 5 quantitative determinations of the WHO international standard was 7.1 × 105 IU/ml (S.D. 4.4 × 105 ).

S. Braham et al. / Journal of Clinical Virology 31 (2004) 5–10 Table 1 A comparison of LC and nested PCR qualitative results

(on 8/6/2000 and 21/6/2000, respectively) from the same renal transplant patient with evidence of persistent B19 infection, gave LC negative/nested PCR positive results (samples 14 and 15, Table 2). The nucleotide sequence of B19 DNA from these two samples was identical in the 511-base pair region of the NS gene that was analysed (Fig. 1) but showed 42 changes (8.2%) compared to the sequence of the reference B19 strain, STU (Hicks et al., 1995).

Nested PCR result

Roche LC result

Positive Negative

Positive

Negative

79 2

13 124

7

The expected value is 106 IU/ml. The mean of 12 quantitative measurements on the in-house standard serum,‘NAN’, estimated from the serum diluted 10−5 in water, was 2.0 × 1011 IU/ml (S.D. 1.5 × 1011 ).

3.3. Sequential sera from immunocompromised patients with prolonged parvovirus B19 infection 3.3.1. Patient 1 (26 years, female, lung transplant) At presentation, bone marrow examination revealed arrest of red cell maturation and a high level B19 viraemia was detected. The initial 2 sera, taken about 3 weeks apart, gave Roche LC assay results of 1.4 × 1010 and 1.2 × 1011 IU/ml and were B19 DNA dot blot positive (Fig. 2). One day after immunoglobulin therapy, a 2 log10 fall in the LC quantification and a negative dot blot result were recorded. The B19 DNA concentration fell further in the following months to 8.9 × 103 IU/ml but after approximately 15 weeks there was a clinical relapse marked by a fall in haemoglobin and the absence of reticulocytes. A rise in LC quantification (to 1.3 × 109 IU/ml) and a positive dot blot indicated the return of high levels of B19 viraemia. After further immunoglobulin therapy there was again a clinical recovery with a fall in B19 DNA levels to 1.7 × 104 genome copies/5 ␮l and a dot blot negative result.

3.2. Clinical samples Ten samples (two serum, one amniotic fluid, two plasma, one fetal serum, one fetal ascitic fluid and three bone marrow) gave invalid results. The results for the remaining 218 clinical samples are shown in Table 1. Seventy nine samples gave concordant positive and 124 concordant negative results by LC and nested PCR assay. The mean concentration of B19 DNA in 79 samples giving concordant positive results was high, 1.3 × 109 (range 2.1 × 102 –3.2 × 1011 ) IU/ml. Of 15 samples giving discrepant results, 13 were LC positive/nested PCR negative (Table 2). The quantitative LC results in these samples were low (<1.5 × 106 IU/ml) with the exception of one sample, an amniotic fluid with >108 IU/ml (sample 11, Table 2). Serological and clinical evidence associated with 8 of the 13 samples in this category (samples1, 2, 3, 4, 6, 9, 12 and 13, Table 2) supported the diagnosis of either recent acute or resolving persistent B19 infection. Only two samples, collected within two weeks

3.3.2. Patient 2 (6 years, male, leukaemia) This patient presented with rash and high levels of B19 viraemia which were initially detected by dot blot as well

Table 2 Details of samples giving discrepant B19 DNA results in LC and nested PCR assays Sample Sample no. type

Roche LC assay (IU/ml)

In-house assays

Patient details

Dot blot

PCR

Age (years)

Sex B19 IgM

B19 IgG

Clinical information Leukaemia, rash; later samples showed IgM/IgG seroconversion Pregnant, fetal ascites; later samples showed IgM/IgG seroconversion Pregnant Pregnant, fetal ascites Lymphoma, pancytopenia 2 months after “flu’-like illness” with rash and arthropathy HIV+, anaemia Renal transplant, anaemia Kawasaki disease, on immunoglobulin therapy; previous sample B19 PCR+ SCID 26/40 gestation

1 2

Serum Serum

2.7 × 103 4.8 × 104

− −

− −

5 29

M F

neg neg

neg neg

3 4 5 6 7 8 9

Serum Serum Serum Serum Plasma Serum Serum

8.4 6.0 2.0 1.7 1.7 4.4 3.3

× × × × × × ×

103 103 103 103 103 104 103

− − − − − − −

− − − − − − −

26 27 nk 41 37 48 3

F F nk F M M M

pos pos neg pos neg neg neg

pos pos pos pos nt nt pos

Serum Amniotic fluid Serum Serum Serum Plasma

2.3 × 103 1.5 × 108

− −

− −

<1 36

F F

neg nt

nt nt

2.4 × 105 5.9 × 102 – –

− − + −

− − + +

30 6 33 33

F M M M

neg neg neg neg

pos pos neg neg

10 11 12 13 14 15

Heart/lung transplant, anaemia; chronic B19 infection (resolving) Leukaemia, anaemia; chronic B19 infection (resolving) Renal transplant; 2 samples from same patient

nk: not known; nt: not tested; SCID: severe combined immunodeficiency; neg: negative; pos: positive.

8

S. Braham et al. / Journal of Clinical Virology 31 (2004) 5–10

Fig. 1. Nucleotide sequence of the 511-base pair region of the B19 NS gene analysed from samples 14 and 15 (Table 2) and from reference strain, STU.

as LC and nested PCR assays (Table 3). In subsequent samples the concentration of B19 DNA declined and the B19 DNA tests became negative in the following sequence: dot blot; first round of nested PCR; second round of nested PCR and finally LC assay. The patient remained B19 DNA positive with low quantitative LC results (6.9 × 102 and 5.9 × 102 IU/ml) for 2 months after the nested PCR became negative. The patient eventually became B19 DNA negative by all methods and rising levels of B19 IgG were de-

tected, indicating viral clearance and a mounting immune response.

4. Discussion In terms of qualitative detection of B19 DNA, the Roche LC kit performed well, correctly identifying the B19 DNA status of most of the clinical samples previously tested by a

Table 3 Results of parvovirus B19 tests in patient 2 Specimen date

1 February 2001 26 February 2001 12 March 2001 24 May 2001 24 May 2001 16 August 2001 18 August 2001 13 September 2001 12 October 2001 08 November 2001

B19 DNA dot blot

pos pos neg neg neg neg neg neg neg neg

B19 nPCR First round

Second round

pos nt pos nt nt neg neg neg neg neg

pos nt pos nt nt pos pos neg neg neg

T/CO: test/cut off ratio; nt: not tested; neg: negative; pos: positive.

Roche LC (IU/ml)

B19 IgM EIA T/CO

B19 IgG EIA T/CO

1.7 × 9.1 × 2.1 × 2.1 × 7.5 × 4.8 × 2.0 × 6.9 × 5.9 × neg

neg neg neg neg pos neg neg neg neg neg

neg neg neg neg neg pos pos pos pos pos

1011 108 106 104 103 103 103 102 102

0.6 0.8 1.1 0.7 2.3 0.4 0.5 0.5 0.5 0.3

0.5 0.4 0.4 0.8 0.8 1.7 1.6 1.8 1.5 5.0

S. Braham et al. / Journal of Clinical Virology 31 (2004) 5–10

9

Fig. 2. Parvovirus B19 DNA, IgM and IgG in heart/lung transplant patient with relapsing red cell aplasia treated with immunoglobulin (patient 1).

manual nested PCR. The Roche LC assay, however, detected B19 DNA for longer than the nested PCR assay in a series of samples from a patient with resolving persistent infection (Table 3). This suggested that the LC assay was more sensitive than the nested PCR assay used in this study. Moreover, the clinical and serological data associated with samples giving positive LC but negative nested PCR results (Table 2) indicated that in the majority of cases the LC results were consistent with other information indicating recent acute or resolving chronic infection and that the nested PCR results were false negative. The sensitivity (97.5%) and specificity (90.5%) calculated by comparison with results of the nested PCR therefore underestimated the true performance of the LC assay. In only two specimens (14 and 15, Table 2), both from the same patient, the LC results appeared to be inconsistent with other data on B19 status. Nucleotide sequencing showed that these two specimens contained a B19 virus strain that had an 8.2% variation in the 511-nucleotide region of the NS gene that was analysed compared to a reference B19 virus strain. The B19 virus strain in this patient was not recognised by the primers used in the Roche kit. A precedent for this is the false negative results for HIV type 1 in the Roche Amplicor kit due to sequence variation in the target region for one of the primers (Barlow et al., 1997). Although far less variable than the genome of HIV-1, B19 genome variation has been reported to be greater in strains from patients with persistent

infection (Hemauer et al., 1997). The primer sites for the B19 LC assay were not available (commercial information), so it is not possible to comment further on why this strain was not detected. However, it would be interesting in future work to extend the analysis of the viral DNA from the patient with persistent infection and a ‘variant’ B19 strain and to compare it with other B19 variants that have recently been described (Servant et al., 2002). Quantitative LC results for the international B19 DNA standard (4.7 × 105 IU/ml) and the in-house standard (2.0 × 1011 IU/ml) were close to the expected values calculated by other methods of 106 IU/ml and 1011 genome copies/ml respectively. Although the level of sensitivity of the dot blot hybridisation was sufficient to detect high concentrations of B19 DNA and to demonstrate the initial response to immunoglobulin therapy, it was possible to monitor viral clearance quantitatively only with the LC assay. It would be helpful for standardisation and comparison of results between laboratories if the Roche kit controls and quantitative results were given in terms of International Units. Invalid results, where no signal was obtained with either the B19 target or the internal control, were obtained with 4.4% (10/228) of the samples overall. However, only 2 of 165 sera gave invalid results compared to three of five bone marrow and two of 11 plasma samples. The presence of red blood cells in these types of sample may be inhibitory and alternative methods for their extraction might be necessary.

10

S. Braham et al. / Journal of Clinical Virology 31 (2004) 5–10

Although it is unhelpful for the clinical laboratory to issue a report stating “invalid result”, it is preferable to reporting a false negative result. The internal standard with the LC assay therefore serves to increase the reliability of B19 DNA testing. A variety of clinical samples in addition to serum and plasma were used in this study including amniotic fluids, fetal ascitic and pleural fluids and fetal tissue homogenates. This extends the range of sample types that have been tested using the MagNA Pure total nucleic acid extraction kit. By using the MagNA Pure instrument for sample extraction and for post elution addition of extracted samples and reaction mixture to glass capillaries, it was possible to automate the LC assay, thus helping to minimise operational errors. Using the automated procedure an assay running time of 4–5 h was obtained which would enable the laboratory to respond to urgent requests for B19 DNA detection eg, in pre-natal testing. The throughput of approximately 27 samples and five controls per assay run is adequate for a diagnostic laboratory but may be a constraint for the application of the LC test to blood donor screening. The value of the quantitative LC assay was illustrated in two immunocompromised patients, one of whom received immunoglobulin therapy (patient 1, Fig. 1). Although the initial diagnosis and response to therapy was also demonstrated by the simple and insensitive dot blot hybridisation assay, the LC assay provided a quantitative and more refined way to monitor the clinical events. Similarly, in patient 2 (Table 3), the LC results showed how viral clearance could be monitored quantitatively. LC assay for B19 DNA has also been used to monitor the outcome where attenuating immunosuppression of the underlying disease, allowing the patient to mount an immune response, was used in the management of relapsing parvovirus B19 infection in an immunocompromised patient (Harder et al., 2001). In summary, the Roche LC assay for quantitative detection of B19 DNA performed well in comparison with existing diagnostic methods. The assay has practical benefits for the laboratory and meets a clinical need for rapid quantitative determinations of B19 DNA required for management of patients in haematology/oncology units and in fetal care centres. However, the failure to detect a variant B19 strain may be a limitation for diagnostic detection of B19 virus.

Acknowledgements We thank Prof. Paul Boseley, Head of Scientific Development, Public Health Laboratory Service and Dr. Jane Sefton of Roche Diagnostics for funding this study. We also thank Dr. Kirstin Edwards and Dr. Jon Clewley of Central Public Health Laboratory for helpful advice.

References Aberham C, Pendl C, Gross P, Zerlauth G, Gessner M. A quantitative, internally controlled real-time PCR assay for the detection of parvovirus B19 DNA. J Virol Meth 2001;92:183–91. Azzi A, Zakrzewska K, Bertoni E, Guidi S, Salvadori M. Persistent parvovirus B19 infections with different clinical outcomes in renal transplant recipients: diagnostic relevance of polymerase chain reaction (PCR) and of quantification of B19 DNA in sera. Clin Microbiol Infect 1996;2:105–8. Barlow KL, Tosswill JHC, Parry JV, Clewley JP. Performance of the Amplicor Human Immunodeficiency Virus Type 1 PCR and analysis of specimens with false-negative results. J Clin Microbiol 1997;35:2846– 53. Brown KE, Buckley MM, Cohen BJ, Samuel D. An amplified ELISA for the detection of parvovirus B19 IgM using monoclonal antibody to FITC. J Virol Meth 1989;26:189–98. Brown KE, Young NS, Alving BM, Barbosa LH. Parvovirus B19: implications for transfusion medicine. Summary of a workshop. Transfusion 2001;41:130–5. Clewley JP. Polymerase chain reaction assay of parvovirus B19 DNA in clinical specimens. J Clin Microbiol 1989;27:2647–51. Clewley JP. PCR detection of parvovirus B19. In: Persing DH, Smith TF, White TJ, editors. Diagnostic molecular microbiology. Principles and applications. Washington (DC): American Society for Microbiology; 1993. p. 367–73. Daly P, Corcoran A, Mahon BP, Doyle S. High-sensitivity PCR detection of parvovirus B19 in plasma. J Clin Microbiol 2002;40:1958–62. Gallinella G, Zerbini M, Musiani M, Venturoli S, Gentilomi G, Manaresi E. Quantitation of parvovirus B19 DNA sequences by competitive PCR: differential hybridization of the amplicons and immunoenzymatic detection on microplate. Mol Cell Prob 1997;11:127–33. Harder TC, Hufnagel M, Zahn K, et al. New LightCycler PCR for rapid and sensitive quantification of parvovirus B19 DNA guides therapeutic decision-making in relapsing infections. J Clin Microbiol 2001;39:4413–9. Hicks KE, Beard S, Cohen BJ, Clewley JP. A simple and sensitive DNA hybridisation assay used for the routine diagnosis of human parvovirus B19 infection. J Clin Microbiol 1995;33:2473–5. Hicks KE, Beard S, Cohen BJ, Clewley JP. Diagnosis of parvovirus B19 DNA by dot-blot hybridisation. In: Stephenson JR, Warnes A, editors. Molecular medicine, vol. 12. Diagnostic virology protocols. Totowa (NJ): Humana Press Inc.; 1998. p. 173–88. Knöll A, Louwen F, Kochanowski B, et al. Parvovirus B19 infection in pregnancy: quantitative viral DNA analysis using a kinetic fluorescence detection system (TaqMan PCR). J Med Virol 2002;67:259–66. Manaresi E, Gallinella G, Zuffi E, Bonvicini F, Zerbini M, Musiani M. Diagnosis and quantitative evaluation of parvovirus B19 infections by real-time PCR in the clinical laboratory. J Med Virol 2002;67:275–81. Saldanha J, Lelie N, Yu MW. Heath A and the B19 Collaborative Study Group. Establishment of the first World Health Organisation International Standard for human parvovirus B19 DNA nucleic acid amplification techniques. Vox Sang 2002;82:24–31. Salimans MMM, Holsappel S, van de Rijke FM, Jiwa NM, Raap AK, Weiland HT. Rapid detection of human parvovirus B19 DNA by dot-hybridisation and the polymerase chain reaction. J Virol Meth 1989;23:19–28. Sato K, Mastuda E, Kamisango K, Iwasaki H, Matsubara S, Matsunaga Y. Development of a hypersensitive detection method for human parvovirus B19 DNA. J Clin Microbiol 2000;38:241–1243. Servant A, Laperche S, Lallemand F, Marinho V, de Saint Maur G, Merinet JF, Garbarg-Chenon A. Genetic diversity within human erythroviruses: identification of three genotypes. J Virol 2002;76:9124–34.