INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The human erythrovirus B19, a member of the family Parvoviridae, causes a broad and seemingly expanding spectrum of disorders (7, 12, 23, 44). The clinical picture depends on the immune status and age of the patient (14, 36, 43). Parvovirus B19 shows a remarkable tropism for erythroid progenitor cells in the bone marrow, which is partly based on binding to its receptor, the blood group P antigen (6). Viral replication, and possibly also the induced immune response, interferes with physiological functions and loss of the erythroid progenitor cells, resulting in a usually subclinical reticulocytopenia (8). In patients with an underlying hematologic disorder and high blood cell turnover (e.g., hemolytic anemia), transient aplastic anemia may ensue. Immunocompromised patients, who have an increased risk of developing a persistent B19 infection, are threatened by a chronic reticulocytopenic anemia, also known as pure red cell anemia (2, 8).

Diagnosis of uncomplicated cases of acute B19 infection (fifth disease or arthropathy) is usually clinically based and can be accomplished by detection of specific immunoglobulin M (IgM) antibodies except in immunocompromised patients, who are prone to persistent infection and who may generate IgM-specific B19 antibodies less reliably (7, 19). Likewise, specific IgG is not a reliable marker for discriminating a reconvalescent status from chronic persistent infection (35), although recent data indicate that IgG antibodies specific for nonstructural protein 1 (NS-1) of B19 are more frequently associated with persistent infection (24). Detection of DNA by hybridization or PCR has been reported to be superior in the diagnosis of prenatal B19 infections and in children with oncologic or hematologic disorders (5, 11, 34). Several qualitative and quantitative PCR methods targeting different regions of the parvovirus genome have been published (1, 4, 9, 10, 13, 15, 17, 22, 25, 40). These PCR approaches are based on conventional block thermocycling and combined single-round PCRs with subsequent oligohybridization or use a nested format.

The recently developed LightCycler (LC) DNA amplification technology (Roche Diagnostics, Mannheim, Germany) combines rapid glass capillary thermal cycling with real-time microvolume fluorescence monitoring (47). Detection of amplicons is achieved during the run in real time. Melting point analysis of the amplicons at the end of the run is used as a specificity control when the fluorochrome SYBR green is used for detection of double-stranded DNA. Alternatively, specificity can be tested during the run by using two target-specific hybridization probes which utilize fluorescence resonance energy transfer (FRET) to generate a measurable signal. In the latter case, two oligonucleotide probes (HybProbe) bind to immediately adjacent regions of the respective amplicon. The upstream probe (referred to as the probe) is labeled at its 3' end with fluorescein, while the 5' end of the downstream probe (referred to as the anchor) is labeled with either of the fluorochromes LC-Red 640 and LC-Red 705. To avoid elongation by the Taq polymerase, the anchor is 3' phosphorylated. When it is ensured that the probe and anchor are spaced no more than five nucleotides apart, simultaneous binding to the specific target generates an amplified signal which is detected fluorometrically.

In the present study, we introduced a rapid and sensitive PCR assay for the quantification of parvovirus B19 DNA, which is based on LC-FRET technology. Up to 25 samples can be quantitatively analyzed within 45 min. The quantification limit was found to be theoretically 5 genome equivalents (geq) per assay, which is equivalent to 250 geq per ml of serum, as calculated on the basis of an external plasmid standard. Optionally, the PCR efficiency can be controlled by an internal amplification control (IC). The benefit of the new method was demonstrated in a child with underlying systemic onset of juvenile idiopathic arthritis (JIA) and relapsing B19 infection, who required an attenuation of immunosuppressive therapy in addition to repeated doses of immunoglobulin to eliminate the virus.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Patients and serum samples. Immunocompetent patients were grouped on the basis of their B19-specific serostatus irrespective of their disease status: (i) IgM and IgG negative (n = 30), (ii) IgG positive and IgM negative (n = 52), or (iii) IgM positive (n = 27). The sera had been tested for B19-specific IgM and IgG antibodies by an enzyme immunoassay (Medac, Wedel, Germany) which utilized a mixture of baculovirus-expressed recombinant VP-1 and VP-2 B19 proteins. In addition, serum samples (n = 10; IgG positive, IgM negative) obtained over a 6-month period from an immunocompromised child with a relapsing B19 infection were examined (see "Case report"). Furthermore, serum samples obtained 15 months before and after the intense 6-month observation period were examined.

Case report. Informed consent was obtained from the parents of the patient prior to publication. An 11-year-old girl had suffered from systemic onset of JIA (also called rheumatoid or chronic arthritis) since the age of 5. Her disease was controlled with a combined immunosuppressive therapy consisting of oral prednisone (0.15 mg/kg daily), subcutaneous methotrexate (20 mg weekly), and oral cyclosporine (5 mg/kg daily). She had a polyarticular bout triggered by an undefined upper respiratory tract infection. To control her symptoms, her dose of prednisone was increased to 1.25 mg/kg daily. One month later, while the prednisone dose was being tapered, the patient contracted another febrile disease. She complained of a dry cough, pain on inspiration, and upper abdominal pain. Her polyarticular symptoms recurred. Later, she developed exertional dyspnea and fatigue. On examination, the patient was afebrile and appeared pale and cushingoid. Her heart rate was 92/min. The edge of the liver was palpable 2 cm below the right costal margin; the spleen was not enlarged. Movement of the cervical spine was limited in all directions, while examination of all other joints revealed no abnormalities.

DNA extraction. DNA was extracted by a commercial silica-based method (blood kit; Qiagen, Hilden, Germany) from 200 µl of serum samples which had been stored for up to 1 year at 20°C. DNA was eluted with 50 µl of ultrapure water (concentration factor, 4:1). The serum samples had been used before in serological assays; in order to minimize the risk of cross contaminations, care was taken to select sera negative for B19 virus-specific IgM which had not been analyzed in batch with samples that tested IgM positive.

Block cycler-dependent amplification of B19 DNA. Primers for PCR were selected from highly conserved genome regions encoding NS-1 and the structural VP proteins (Table 1). Primers were chosen with the help of Prime software, which is embedded in the GCG package, and the MeltCalc program (41). Amplifications were performed in 50-µl reaction mixtures containing 50 pmol of each primer, 200 µM concentrations of each deoxynucleoside triphosphate, 2 U of HotStar polymerase (Qiagen), and 1.5 mM MgCl₂ in the appropriate buffer. The cycling conditions for a T3 block cycler (Biometra, Göttingen, Germany) are shown in Table 1. Amplicons were detected by conventional ethidium bromide-agarose gel electrophoresis.

LC-dependent amplification of B19 DNA. Cycling conditions for primers NS-1a and NS-1a' were adjusted using SYBR green for detection of amplicons in real time; the specificity of amplicons was evaluated by melting point analysis (FastStart SYBR green kit; Roche Diagnostics). Hybridization probes to be used with the HybProbe LC-FRET technology were selected with the aid of the MeltCalc software (41): probe, 5'-gCA AAA gCC ATT TTA ggC ggg CA-fluorescein; anchor, 5'-LC-Red 640-CAC CAg ggT AgA TCA AAA AAT gCg Tgg A-PO₄ (Fig. 1A). Reaction mixtures of 20 µl were prepared with the FastStart HybProbe kit (Roche Diagnostics) using primers at 20 pmol each and an MgCl₂ concentration of 3.0 mM. Probe and anchor were added at 3 pmol each. Sample DNA was added in a volume of 5 µl per reaction. Cycling started with an initial denaturation of 6 min at 96°C followed by 40 cycles of 95°C for 4 s, 56°C for 16 s, and 72°C for 12 s. The temperature for transition from annealing to elongation was lowered from 20 to 2°C per s. Specific hybridization, indicated by the LC-Red 640 signal, was monitored during the annealing phase with the channel setting F2/F1. Fluorimeter gains of the LC were universally set to values of 5 (channel 1), 15 (channel 2), and 30 (channel 3). Melting points for the probe-anchor-target duplexes were checked following completion of the cycling procedure. Hybrids were denatured at 95°C for 10 s, allowed to reanneal at 55°C for 20 s, and then heated at 0.2°C per s to 75°C. Specific products showed a T_m of 64.0°C (±0.5°C) in melting point analysis. The assay was completed within 45 min.

View larger version (14K): [in this window] [in a new window]   FIG. 1.   Amplification and detection of parvovirus B19-specific DNA by using LC-FRET technology. Primers were chosen from the gene encoding NS-1. For primer sequences (NS-1a and NS-1a') and locations, see Table 1. (A) Detection of the B19-specific amplicon by use of a hybridization probe-and-anchor pair. FL, 3' fluorescein dye label of the probe; LC640, 5' LC-Red 640 dye label of the anchor. (B) IC amplicon constructed for coamplification with the specific fragment. S and B, unique SphI and a BglII sites, respectively, introduced by site-directed mutagenesis. The original anchor sequence between these restriction sites was replaced with a semisynthetic IC sequence (white bar) which can be detected using the IC anchor labeled with LC-Red 705.

Construction of an IC. An IC was constructed for use in a competitive PCR protocol with the LC-FRET technology. For this purpose, the 229-bp PCR fragment, defined by primers NS-1a and NS-1a', was cloned into the T/A plasmid vector pCR2.1TOPO (Invitrogen, Groningen, The Netherlands) to give the plasmid pPB19N. The fragment was mutated using the QuickChange kit (Stratagene, La Jolla, Calif.) to accept unique SphI and BglII sites flanking the sequence of the anchor probe (Fig. 1B). The mutated fragment was transferred into the EcoRI site of the vector pGEM 9Zf(+) (Promega, Madison, Wis.), which does not contain SphI or BglII sites. The original HybProbe anchor sequence was replaced, after SphI/BglII double digestion, with a synthetic sequence (plasmid pPBIC). Detection of IC amplicons generated with primers NS-1a and NS-1a' was then carried out by means of the B19-specific fluorescein-labeled probe described above and the unique IC anchor 5'-LC-Red 705-CAC gCT CAC AgC gCA gTA gAT CT-PO₄ measured at channel 3 of the LC (band-pass, 710 nm).

DNA quantification. Quantities and copy numbers of plasmid pPB19N and pPBIC were calculated following UV spectrophotometry (260 nm) and by gel electrophoresis against a standardized size ladder (SmartLadder; Eurogentech, Seraing, Belgium).

Parvovirus B19 isolate V9. Plasmid DNA which harbored a large part of the entire genomic DNA of the B19 isolate V9 (38), including the NS-1 and VP gene sequences targeted by our PCRs, was kindly supplied by Q. T. Nguyen, Institut Pasteur, Paris, France.

Statistical analysis. The strength of agreement between the results of different assays was measured by robust kappa () statistics (18), implemented, e.g., in the EpiInfo 6.0 software distributed by the World Health Organization and Centers for Disease Control and Prevention. According to convention, a  coefficient of >0.6 indicated good agreement, whereas a  of >0.8 signaled very good agreement.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Performance characteristics of the B19 LC PCR. A standard dilution of B19-specific sequence copies was prepared using plasmid pPB19N, which contains a 229-bp fragment of the NS-1 gene generated by PCR using the primer pair NS-1a-NS-1a' (Table 1). Serial 10-fold dilutions were made in diethyl pyrocarbonate-treated water containing 2 mg of denatured herring sperm DNA liter¹ as a stabilizer. Aliquots of 5 µl were analyzed for B19 DNA.

View larger version (39K): [in this window] [in a new window]   FIG. 2.   Performance characteristics of the LC-FRET-dependent amplification of an NS-1-specific fragment of parvovirus B19 DNA. A standard log10 dilution series was prepared from plasmid pPB19N containing the 229-bp NS-1-specific target sequence. Results of up to 14 runs per dilution step are represented. The Spearman rank correlation coefficient (rs) and the 95% confidence level (CI) are indicated.

Detection of the genetically diverse human erythrovirus isolate V9. Recently, a genetically markedly distinct erythrovirus isolate (termed V9) was detected in a child with transient aplastic anemia (38). The authors reported an estimated sequence divergence from standard B19 strains of at least 15% in some regions. A plasmid containing large parts of the genomic DNA of the V9 isolate was tested in our various PCR formats. Specific amplicons were obtained with the NS-1-specific block cycler PCR and with the LC-FRET assays but not with the seminested VP-1 PCR (data not shown).

Comparative analysis of patient sera for B19 DNA. A total of 109 sera from 102 immunocompetent patients and 12 sera from the immunocompromised child with JIA were examined in three different PCRs. The results are summarized in Fig. 3 (Venn diagram). All samples from immunocompetent patients which yielded a positive PCR result that was concordantly positive between at least two formats (grey portions of the Venn diagram) also proved to be positive for B19 IgM. These patients presented with typical symptoms of acute B19 infection. By nested NS-1 PCR, six additional sera from immunocompetent patients tested positive. Three of these also tested IgM positive, while the remaining three came from the IgG⁺ IgM cohort. None of the seronegative patients tested positive in either PCR format. An additional eight sera from the immunocompromised child (IgG⁺ IgM) with relapsing B19 infection tested concordantly positive in all three PCR formats, while one sample from this child was positive by nested NS-1 PCR only, and three sera were concordantly negative. Interrater analysis (kappa statistics [18]) indicated a very good agreement (conventional definition,  > 0.8) between the LC-FRET PCR and the nested NS-1 as well as the seminested VP PCR ( = 0.85 and 0.98, respectively). Therefore, the LC-FRET PCR can be used as a substitute for both the nested NS-1 and the seminested VP-1 PCRs.

View larger version (44K): [in this window] [in a new window]   FIG. 3.   Detection of parvovirus B19 DNA in patient serum samples by three different DNA amplification methods. LC, LC-FRET assay specific for a 229-bp fragment of the NS-1 locus; VP, seminested PCR specific for a 174-bp fragment of the VP-1 locus; NS1, nested PCR specific for a 99-bp fragment of the NS-1 locus. Shaded areas of the Venn diagram indicate samples which were concordantly positive by at least two PCR formats. Numbers in parentheses indicate serum samples from an immunocompromised child; other samples originated from immunocompetent individuals.

B19 DNA loads and clinical course of a relapsing infection in an immunocompromised child. Based on the detection of B19 DNA in serum, the diagnosis of an acute parvovirus B19 infection (Fig. 4) with aplastic crisis was made in a child who required immunosuppressive therapy for underlying systemic onset of JIA. Initially, the B19 DNA load peaked at 3.2 × 10⁹ geq per ml serum (Fig. 4, day 0). Circulatory compromise occurred and the patient received two packs of red blood cells. The ongoing severe systemic inflammatory activity did not allow discontinuation of the immunosuppressive therapy at that point. To specifically treat the B19 infection, the patient received a 5-day-course of polyvalent intravenous (i.v.) IgG (400 mg/kg daily). The reticulocytopenic anemia resolved within a week. The polyarticular symptoms subsided shortly after the diagnosis of acute parvovirus infection was made and the prednisone dose was tapered (day 10). B19 DNA loads decreased below the detection level (by LC-FRET PCR) in the serum within 2 months (day 60), suggesting a gradual clearance of the virus from these compartments (Fig. 4). The serum obtained at day 60 tested positive for B19 DNA only by the nested NS-1 PCR.

View larger version (23K): [in this window] [in a new window]   FIG. 4.   Kinetics of the parvovirus B19 DNA load in serum, hemoglobin values, and reticulocyte counts in an immunosuppressed child with underlying systemic onset of JIA (++, severe symptoms; +, moderate symptoms; , no symptoms) and relapsing B19 infection. Day 0 indicates the first PCR-based diagnosis of an active B19 infection. B19 genome equivalents were calculated on the basis of an external plasmid standard. Treatments include packed red cells (PRC), prednisone (Pred; arrows indicate an increasing or decreasing dose), i.v. IgG containing antibodies specific for B19 VP-1, i.v. methotrexate (MTX), and discontinuation of cyclosporine (CSA%), as part of the immunosuppressive regimen.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We have developed a rapid quantitative PCR assay, based on LC-FRET technology, for the detection of B19 DNA. Quantification of viral DNA loads in serum was possible over a dynamic range of 7 log steps down to the limit of theoretically 250 geq per ml of serum. Lower virus loads were detected unreliably and by melting point analysis only. False-negative results can be controlled by coamplification of an internal standard, albeit at the expense of a loss of sensitivity of 1 log step (Fig. 2).

LC-FRET was used to detect B19 DNA in serum samples from immunocompetent patients selected on the basis of their B19-specific serostatus, and the results were compared to those of two conventional block cycler PCR formats (nested NS-1-specific and seminested VP-specific PCR). The nested NS-1 PCR proved to be more sensitive, since an additional six sera tested positive by this assay only (Fig. 3). This reflects the performance characteristics of the nested NS-1 PCR when tested with the positive control plasmid pPBN19 (data not shown). Samples which tested concordantly positive by at least two different PCR formats were also positive for B19 IgM (Fig. 3). Both LC-FRET and the VP PCR failed to identify six sera of which three also tested positive for B19 IgM. An additional three sera which tested DNA positive by the nested NS-1 PCR exclusively, however, came from immunocompetent IgM-negative IgG-positive patients with no clinical signs of B19 infection. It remains to be clarified whether these results are due to a prolonged viremic but asymptomatic infection or represent false-positive results. Despite these discrepancies, the kappa coefficient revealed a very good correlation between the LC-FRET PCR and the block cycler PCR formats. The sensitivity of the LC-FRET PCR is in the range of other published B19 DNA quantification methods, i.e., approximately 5 geq per assay (1, 4, 10, 17, 22, 25, 40). For immunocompetent patients, LC-FRET PCR also correlated favorably with IgM serostatus ( = 0.92).

Genetic variation between B19 wild-type isolates has been reported to be less than 5%, and few genetic changes accumulate longitudinally during prolonged chronic infection in individual patients (27). Our primers and probes were positioned in regions which were completely conserved between at least five different isolates for which sequence data were available from the databases. DNA derived from the human erythrovirus isolate V9, which was reported to be genetically divergent from standard B19 wild-type strains (38), was readily detected by our NS-1-specific PCRs (LC and nested NS-1) but not by the VP-specific seminested PCR.

The reported case of the immunosuppressed child with underlying systemic onset of JIA and relapsing parvovirus B19 infection is an example of the new method's value in therapeutic decision-making. The child experienced two separate episodes of active B19 replication, which could be quantitatively monitored and clearly discerned only by the LC-FRET PCR (Fig. 4) but not by the nested NS-1 PCR, which also yielded a positive signal for the serum obtained at day 60. Therapeutic interventions were guided by the quantification of viral loads in the serum in association with the hematological values. Treatment with i.v. IgG, combined with an attenuated T-cell-suppressive regimen, finally led to a stable hematologic state with viral loads below the detection level of the LC-FRET PCR.

Immunosuppressive therapy, e.g., in autoimmune diseases, is considered a risk factor for persistence of the parvovirus B19 (7). Immunocompromised patients have difficulties eliminating the virus from their bone marrow, and bone marrow insufficiency may persist for up to 10 years (31). Chronic infections without underlying immunodeficiencies are rarely described (16, 35, 37). Protection from parvovirus B19 is thought to be mediated by neutralizing antibodies (32, 49). Specific IgM antibodies are detectable as early as day 5 after infection and persist for months, while IgG antibodies are detectable from the second week postinfection and have lifelong persistence (49). Neutralizing antibodies are mounted against both capsid proteins VP-1 and VP-2 (3, 21, 39, 48). Antibodies against VP-1, the minor capsid protein, are found predominantly in convalescent-phase sera (32, 49). Antibodies against VP-2, the major capsid protein, are found predominantly during the acute disease and early convalescence (26, 32, 42), especially when denatured (linear) antigen is used in the test system. However when native (conformational) VP2 antigen is used, the antibodies can also be detected in convalescent-phase sera (28, 42). A deficiency in generating specific antibodies either qualitatively (e.g., patients with congenital or acquired immunodeficiencies) or quantitatively (e.g., patients with leukemia) is considered responsible for persistent B19 infections (32, 49). Consequently, supplementation of specific immunoglobulins is the accepted therapy of choice (29, 30, 32, 33, 49). Commercial preparations of immunoglobulins contain specific parvovirus B19 antibodies, which are deemed responsible for clinical effectiveness. We tested one lot of the immunoglobulins given to our patient (Sandoglobin; Novartis, Nürnberg, Germany) and detected B19 VP-specific antibodies by enzyme-linked immunosorbent assay (Medac). The standard dose of i.v. IgG is 400 mg/kg daily for 5 to 10 days (29, 49). In our patient, i.v. IgG treatment coincided with declining viral loads in the serum (Fig. 4) but obviously failed to eliminate the virus from the bone marrow, as was evident from the rebound of viral DNA loads associated with reticulocytopenia while the patient was under intensified immunosuppressive treatment. In AIDS patients with relapses of B19 infection, i.v. IgG is given at monthly intervals at a dosage of 1 g/kg daily for 1 to 2 days (30). We adapted the protocol accordingly. Again, i.v. IgG treatment led to gradually declining viral loads in serum, but only when the T-cell-suppressive regimen was attenuated by tapering the prednisone and discontinuing cyclosporine, in addition to repeated doses of immunoglobulins, did we observe a sustained hematological normalization and undetectable viral loads in serum.

The patient's course allows speculations about the relative importance of the humoral and cellular immune responses in persistent parvovirus infection. In general, neutralizing antibodies are thought to be the mainstay of acquired immunity (32, 49). Only recently has it become evident that cellular immunity plays a significant role in overcoming acute infection and providing long-term memory (20, 46, 45). When our patient had her first episode of parvovirus infection, a standard dose of immunoglobulins suppressed the DNA levels below the detection limit but did not eliminate the virus. The infection relapsed after the prednisone dose, a potent inhibitor of T-cell function, was increased. Immunoglobulin therapy was reinstituted. Four weeks after reinstitution of i.v. IgG therapy, the viral load declined by a factor of 4 × 10³. Immunoglobulin therapy after the first episode had led to a decline of viral DNA load by a factor of 5 × 10⁶. We therefore reasoned that immunoglobulin therapy might not be sufficient to clear the virus, and we attenuated the immunosuppressive regimen. In parallel, i.v. IgG therapy was continued. This two-pronged therapeutic approach allowed the patient to finally clear the virus from circulation. After more than 15 months, B19 DNA remained undetectable in serum. In retrospect, it is not possible to ascribe the effect of viral elimination to either attenuation of T-cell suppressive therapy or to repeated doses of immunoglobulins.

In conclusion, the LC-FRET PCR specific for B19 DNA can substitute for conventional nested PCR formats for the diagnosis of an active B19 infection. The advantages of the LC-FRET PCR are its rapidity (45 min versus 5 h), ease of preparation (one tube, including, if desired, an IC), diminished risk of cross contamination, high specificity, and the option of quantification. Given the reduced hands-on time, implementation of this format would appealingly allow cost reduction in the clinical diagnostic laboratory. The LC-FRET PCR can guide the therapeutic options of repeated immunoglobulin therapy versus a reduction of immunosuppressive regimens in patients with persistent parvovirus infections.