Development and Validation of a Laboratory-Developed Multiplex Real-Time PCR Assay on the BD Max System for Detection of Herpes Simplex Virus and Varicella-Zoster Virus DNA in Various Clinical Specimens

Abstract

A multiplex real-time PCR (quantitative PCR [qPCR]) assay detecting herpes simplex virus (HSV) and varicella-zoster virus (VZV) DNA together with an internal control was developed on the BD Max platform combining automated DNA extraction and an open amplification procedure. Its performance was compared to those of PCR assays routinely used in the laboratory, namely, a laboratory-developed test for HSV DNA on the LightCycler instrument and a test using a commercial master mix for VZV DNA on the ABI7500fast system. Using a pool of negative cerebrospinal fluid (CSF) samples spiked with either calibrated controls for HSV-1 and VZV or dilutions of a clinical strain that was previously quantified for HSV-2, the empirical limit of detection of the BD Max assay was 195.65, 91.80, and 414.07 copies/ml for HSV-1, HSV-2, and VZV, respectively. All the samples from HSV and VZV DNA quality control panels (Quality Control for Molecular Diagnostics [QCMD], 2013, Glasgow, United Kingdom) were correctly identified by the BD Max assay. From 180 clinical specimens of various origins, 2 CSF samples were found invalid by the BD Max assay due to the absence of detection of the internal control; a concordance of 100% was observed between the BD Max assay and the corresponding routine tests. The BD Max assay detected the PCR signal 3 to 4 cycles earlier than did the routine methods. With results available within 2 h on a wide range of specimens, this sensitive and fully automated PCR assay exhibited the qualities required for detecting simultaneously HSV and VZV DNA on a routine basis.

INTRODUCTION

Herpes simplex viruses 1 and 2 (HSV-1 and -2, respectively) and varicella-zoster virus (VZV) are human double-stranded DNA viruses belonging to the Herpesviridae family. In immunocompetent hosts, benign vesicular eruption of the skin and mucosa occurs during either the primary infection or reactivation episodes (1). However, life-threatening manifestations, mainly related to virus neurovirulence, can be observed (2), HSV-1 and VZV being the first and second most common causes of viral encephalitis, respectively (3, 4). These viruses are associated with mortality and morbidity, especially when treatment with acyclovir is delayed (5). HSV-2 is more likely responsible for meningitis (1). HSV-2 and VZV are also implicated in neonatal infections (1, 6, 7).

Due to similarities in the clinical presentations of the diseases caused by these 3 herpesviruses and other pathogens, virological testing is often needed. Accurate and specific diagnosis is of special importance in cases of severe infection or in infection occurring in immunosuppressed patients who can present atypical clinical pictures (8). In replacement of cell culture and immune detection of viral antigens, molecular assays are currently the main tool used for the virological diagnosis of infections related to HSV and VZV. This trend is due to the multiple advantages of nucleic acid testing (NAT), including the rapid availability of results, its adaptation to a wide diversity of clinical specimens, the capacity for automation, and, mostly, its high sensitivity and specificity. Several FDA-approved or CE-marked in vitro diagnostic (IVD) kits are available for detecting these viruses in various specimens (9–11). However, all of them propose separate detection for HSV and VZV DNA, and only a few multiplex assays including both agents have been published as laboratory-developed tests (LDTs) (10, 12).

Although detection of viral DNA is the gold standard for the virological diagnosis of neurological severe infections (3), to date, no commercial fully automated PCR assay detecting both HSV and VZV is available. However, an FDA-approved assay for HSV detection in cerebrospinal fluid (CSF) was recently validated for clinical use (13, 14). As low viral loads (<500 copies/ml) are not infrequent in CSF of patients suffering encephalitis associated with HSV (15–20), it was recommended to reach a sensitivity of at least 200 copies/ml for the detection of this agent in CSF (21). The same need for sensitive tests is found with encephalitis associated with VZV (15, 22).

The recently commercialized BD Max system (Becton Dickinson, Pont-de-Claix, France) combines the extraction and amplification steps on the same instrument, shortening the hands-on time and allowing its use in emergency molecular diagnostic testing. Dedicated analyte kits (23–26) but also master mix from other companies (27) and laboratory-developed methods (28) can be used on this open, fully automated platform. Recently, the BD Max system was used to perform in 2 separate assays a real-time PCR test detecting HSV and VZV in skin and mucosal lesions (29).

The present work was aimed at developing on the BD Max instrument a laboratory-developed HSV/VZV PCR assay able to detect HSV and VZV DNA in the same reaction tube. The sensitivity of the BD Max assay was shown to be similar to that of PCR assays routinely used in the laboratory on various types of clinical specimens, including CSF.

(The data were previously presented at the 30th Clinical Virology Symposium [April 2014, Daytona Beach, FL].)

MATERIALS AND METHODS

Positive controls used for sensitivity determination.

Commercially available quantified standards of HSV-1 and VZV claimed to contain 10,000 copies/ml of each target (Qnostics, Glasgow, United Kingdom) were used to check the empirical sensitivity of the new assay. Given that the HSV-2 control was not convenient due to the fact that the strain sequence is not referenced in GenBank, a clinical strain of HSV-2 was chosen as a positive control. For the three controls, serial dilutions were performed in a pool of CSF samples that tested negative for HSV and VZV before DNA extraction. Their viral load, expressed as copies per milliliter, was measured using the HSV-1, HSV-2, and VZV R-gene kit (bioMérieux SA, Verniolles, France) (30).

QCMD panels.

The 2013 HSV and VZV panels of the Quality Control for Molecular Diagnostics organization (QCMD; Glasgow, United Kingdom) were tested retrospectively. Each panel contained 9 presumably positive specimens and 1 negative specimen.

Clinical specimens.

For clinical validation, 180 specimens of various origins were retrospectively analyzed by the routine and BD Max assays, including 38 CSF samples. All were residual specimens previously analyzed in the laboratory for diagnosis of HSV or VZV infection and stored at −20°C. This research was conducted in accordance with the Helsinki declaration of the rights of patients.

Primers and probes.

For HSV, both the routine and BD Max assays targeted the viral DNA polymerase gene and were able to detect HSV-1 and HSV-2 without type differentiation, as previously described (31). For VZV, the BD Max assay targeted the viral DNA polymerase gene (32) whereas the routine assay targeted glycoprotein 19 (11). For the BD Max multiplex assay, primers targeting the sample process control (SPC) were added to the PCR master mix; this internal control consists of a fragment of Drosophila melanogaster genome (GenBank sequence accession no. AC246436; nucleotides 35779 to 35978) cloned in a pUC119 derivative (GenBank sequence accession no. U07650.1). The sequences of primers and probes used for the BD Max multiplex assay are shown in Table 1.

TABLE 1.

Sequences of primers and probes used in this study for BD Max multiplex assay^a

Target	Length of product (bp)	Sequence (5′ to 3′)	Tm, °C (% CG)
HSV	92	Forward: CATCACCGACCCGGAGAGGGAC	66.9 (68)
		Reverse: GGGCCAGGCGCTTGTTGGTGTA	69.0 (63)
		Probe: FAM-CCGCCGAACTGAGCAGACACCCGCGC-BHQ1	79.8 (73)
VZV	63	Forward: CGGCATGGCCCGTCTAT	56 (64.7)
		Reverse: TCGCGTGCTGCGGC	50 (78.6)
		Probe: YY-ATTCAGCAATGGAAACACACGACGCC-BHQ1	54.4 (50)
SPC	149	Forward: GGATCTAGCCGTGTGCCCGCT	66.8 (66)
		Reverse: GGCATGGAGGTTGTCCCATTTGTG	64.5 (54)
		Probe: ATTO647n-TTGATGCCTCTTCACATTGCTCCACCTTTCCT-BHQ3	67.8 (45)

^aAbbreviations: HSV, herpes simplex virus; VZV, varicella-zoster virus; YY, Yakima Yellow; SPC, sample process control; BHQ, black hole quencher; FAM, 6-carboxyfluorescein; T_m, melting temperature; ATTO, German company that manufactures the corresponding fluorophore.

Routine assay for detection of HSV genome.

The extraction step was performed on 200 μl of sample using the NucliSens easyMAG instrument with the Specific B protocol, as recommended by the manufacturer (bioMérieux SA, Marcy l'Etoile, France). For CSF specimens, 100 μl of Basematrix diluent (InGen, Chilly-Mazarin, France) was added during the lysis step in order to increase the extraction efficiency, as previously recommended (33). The elution volume was 50 μl. Five microliters of extract was mixed with 10 pmol of each HSV primer, 4 pmol of the HSV probe (Table 1), and the LightCycler FastStart DNA Master HybProbe (Roche, Meylan, France) in a glass capillary. The cycling conditions on LightCycler 1.0 were as previously published (31); a total of 55 cycles was performed, and every positive signal was interpreted as positive without a threshold cycle (C_T) cutoff value. No internal control was present for this assay.

Routine assay for detection of VZV genome.

The extraction conditions are similar to those described above for the HSV routine assay. Ten microliters of extract was added to 15 μl of ready-to-use VZV R-gene master mix before cycling on an ABI7500fast system (Applied Biosystems) as recommended by bioMérieux. An internal control was included through extraction with the sample.

BD Max assay.

The PCR master mix consisted of 10 pmol of each primer and 4 pmol of the probe for HSV and VZV and 8 pmol of each primer and 4 pmol of the probe for the SPC (Table 1), in a final volume of 7.5 μl containing the RealTime Ready DNA Probe Master (Roche) for one test. This mixture of reagents, considered a batch, was prealiquoted in BD Max snap-in tubes, sealed using colored cover foil with the Axygen PlateMax semiautomated plate sealer (Axygen Scientific, Union City, CA, USA; 1 time for 8 s at 180°C), and stored at −20°C until use. It was shown to be stable for at least 12 weeks (data not shown), which allows testing of the positive and negative controls at the beginning and end of each batch use, without intermediate controls for each run.

The BD Max HSV/VZV/SPC multiplex assay was fully automated. A 200-μl volume of sample, together with 550 μl of DNA-free water, was added to the sample buffer tube. A snap-in tube containing the master mix was thawed and snapped into the BD Max ExK DNA-1 reagent strip containing all the materials (pipette tips and waste) and reagents (wash buffer and elution buffer) necessary for the DNA extraction step, as well as the SPC template that was coextracted and coamplified with the viral targets for controlling the full process. After the sample buffer tube and the strip were placed on the BD Max instrument, the automated method was started. The sample lysis was performed onboard at 62°C for 9 min. The cycling conditions consisted of denaturation for 60 s at 98°C followed by 55 cycles of 5.7 s at 98°C and 33.3 s at 62°C. The results were available within 1.5 h for fewer than 5 samples; 6 to 9 min was necessary for testing each series of 4 additional samples. Up to 24 specimens can be tested in the same run for a total time of 130 min.

The C_T cutoff value of SPC that was used for considering a sample to be invalid was 29.6, which corresponds to the mean of the SPC of the 87 negative clinical specimens plus 2 standard deviations, in accordance with a previous study (34).

Statistical analysis.

The limit of detection (LoD) of the assays was estimated by PriProbit analysis (version 1.63) (35). The LoD of the assay corresponded to the value of viral DNA load detected with 95% probability and the estimation of its 95% confidence interval (95% CI). The Spearman coefficient was used for correlation analyses.

RESULTS

Reproducibility of BD Max assay.

Five replicates of each serial dilution of positive controls for HSV-1, HSV-2, or VZV were tested to evaluate the intra- and interassay reproducibility of the multiplex BD Max assay (Table 2). Reproducibility of the C_T value for the SPC was also satisfactory with a coefficient of variation (CV) of <5% (data not shown). No cross-contamination was observed when alternating highly positive and negative samples in the same run (data not shown).

TABLE 2.

Reproducibility of BD Max assay for known copy numbers of HSV-1, HSV-2, and VZV DNA diluted in a pool of HSV- and VZV-negative CSF specimens^a

Viral load (copies/ml)	Interrun		Intrarun
	No. of positive replicates	Mean CT ± SD (% coefficient of variation)	No. of positive replicates	Mean CT ± SD (% coefficient of variation)
HSV-1
3,333	5	35.66 ± 0.62 (1.74)	5	35.08 ± 0.85 (2.42)
1,111	5	37.88 ± 0.96 (2.53)	5	38.76 ± 1.05 (2.72)
370	5	42.00 ± 2.26 (5.39)	5	43.64 ± 1.72 (3.94)
HSV-2
39,800	5	30.16 ± 1.39 (4.60)	5	29.66 ± 0.56 (1.89)
3,980	5	33.64 ± 1.19 (3.53)	5	32.38 ± 0.37 (1.14)
398	4	36.38 ± 0.90 (2.47)	4	35.58 ± 0.54 (1.51)
VZV
3,333	5	31.42 ± 0.68 (2.15)	5	31.32 ± 0.58 (1.85)
1,111	5	33.24 ± 0.95 (2.86)	5	32.98 ± 0.71 (2.15)
370	5	36.70 ± 1.78 (4.84)	5	38.98 ± 2.45 (6.28)

^aEach sample was tested in five replicates.

Sensitivity of BD Max assay.

The empirical sensitivity of the BD Max assay was determined by using serial dilutions of positive controls from 10 independent determinations as described above. The LoD of the BD Max multiplex assay was 195.65 copies/ml (95% CI, 117.41, 828.07), 91.80 copies/ml (95% CI, 37.90, 1,499.20), and 414.07 copies/ml (95% CI, 234.80, 1,303.80) for HSV-1, HSV-2, and VZV, respectively.

Validation of BD Max assay by using external quality controls.

The 20 specimens of the 2013 QCMD panels were tested by the routine and BD Max assays. As shown in Table 3, all of the specimens were reliably detected by both assays, including those exhibiting very low viral DNA loads (<500 copies/ml). Specimens expected to be negative for HSV or VZV DNA tested negative by both assays.

TABLE 3.

Detection of HSV and VZV DNA in specimens of 2013 QCMD panels using routine and multiplex BD Max assays

QCMD specimen	Expected result from QCMD		CT by assay
			Routine assayb		BD Max assay
	DNA content (origin)	Sample concn (copies/ml)a	HSV	VZV	HSV	VZV
HSVDNA13-07	HSV-1 (MacIntyre strain)	9,840	34.82	NTc	30.24	Negd
HSVDNA13-01	HSV-1 (95/1906)	9,795	34.96	NT	30.03	Neg
HSVDNA13-08	HSV-1 (95/1906)	490	38.45	NT	34.72	Neg
HSVDNA13-04	HSV-1 (MacIntyre strain)	189	40.19	NT	38.00	Neg
HSVDNA13-02	HSV-2 (MS)	6,353	32.71	NT	29.06	Neg
HSVDNA13-03	HSV-2 (09-015681)	4,335	32.36	NT	28.24	Neg
HSVDNA13-05	HSV-2 (MS)	249	39.74	NT	33.97	Neg
HSVDNA13-09	HSV-2 (09-015681)	91	44.46	NT	34.13	Neg
HSVDNA13-10	VZV	NAe	Neg	29.30	Neg	28.69
HSVDNA13-06	Negative		Neg	NT	Neg	Neg
VZVDNA13-04	VZV (OKA)	25,586	NT	29.11	Neg	28.66
VZVDNA13-05	VZV (63/1444)	2,576	NT	33.47	Neg	30.91
VZVDNA13-01	VZV (Ellen strain)	2,249	NT	33.56	Neg	31.32
VZVDNA13-06	VZV (OKA)	1,140	NT	33.65	Neg	31.46
VZVDNA13-07	VZV (9/84)	783	NT	34.91	Neg	32.00
VZVDNA13-08	VZV (63/1444)	226	NT	37.14	Neg	34.64
VZVDNA13-10	VZV (Ellen strain)	118	NT	37.76	Neg	34.84
VZVDNA13-02	VZV (9/84)	61	NT	39.44	Neg	35.15
VZVDNA13-03	HSV-1	10,000	35.12	Neg	30.47	Neg
VZVDNA13-09	Negative		NT	Neg	Neg	Neg

^aAs determined by QCMD.

^bThe routine tests that were used as comparators were a laboratory-developed test for HSV DNA on LightCycler and a test using a commercial master mix for VZV DNA on ABI7500fast.

^cNT, not tested.

^dNeg, negative.

^eNA, not available.

Clinical validation of BD Max assay.

From the 180 tested clinical samples, no amplification of the SPC target was observed for 2 CSF samples (1.11% with comparison to the total number of samples and 7.14% with comparison to the number of CSF samples) using the BD Max HSV/VZV/SPC assay, suggesting the presence of PCR inhibitors; despite an additional cycle of freezing/thawing, the SPC target remained undetected. These 2 samples that displayed no atypical characteristics in terms of aspect or consistency were found negative for HSV and VZV by the routine assays, with a correct amplification of the internal control present in the VZV assay.

All of the remaining 178 specimens were found concordant by both assays for the two viruses, corresponding to a total agreement of 100% for both HSV and VZV (kappa coefficient of 1 with a 95% confidence interval of 0.974 to 1 for both targets). Table 4 depicts the distribution of the results obtained with different kinds of specimens using the multiplex BD Max assay. As shown in Fig. 1, the C_T values of specimens found positive for HSV (n = 61 specimens) or VZV (n = 31 specimens) were highly correlated by comparing the BD Max and routine assays. Of note, the BD Max assay detected the PCR signal 3 to 4 cycles earlier than did the routine methods.

TABLE 4.

Clinical analysis of the 180 specimens tested in the study using multiplex HSV/VZV/SPC assay on BD Max platform

Type of specimen	No. of specimens with result by BD Max assay:
	HSV positive	VZV positive	HSV and VZV positive	Negative	Invalid
CSF	9	1	0	26	2
Neonatal swab	0	0	0	11	0
Genital swab	10	0	0	6	0
Oral or nasal swab	11	0	0	7	0
Cutaneous swab	8	26	0	6	0
Ocular swab	1	0	0	11	0
Respiratory secretion	17	2	1	12	0
Amniotic fluid	0	0	0	8	0
Serum	4	1	0	0	0

FIG 1.

Correlation of the C_T values of the clinical specimens found positive by routine and multiplex BD Max HSV/VZV/SPC assays. (A) Specimens found positive for HSV (Spearman correlation coefficient of 0.91, P < 0.0001). (B) Specimens found positive for VZV (Spearman correlation coefficient of 0.98, P < 0.0001).

DISCUSSION

The laboratory-developed HSV/VZV/SPC BD Max assay presented here is a multiplex assay aimed at detecting both HSV and VZV DNA, allowing the use of only one test in routine settings for the clinical diagnosis of these 2 infectious agents in various clinical specimens, including CSF. Although not differentiating between HSV-1 and HSV-2, it was shown to be at least as sensitive as the molecular assays routinely used in the laboratory for separately detecting HSV and VZV DNA.

In addition to HSV and VZV targets, an internal control, named SPC, was incorporated into the new assay in order to prevent false-negative results. As SPC is included in the extraction strip, this DNA is extracted and amplified concomitantly with the viral targets, allowing the control of the full process on the BD Max platform. The presence of PCR inhibitors has been reported in several types of specimens, including whole blood, CSF, and amniotic fluid. Of note, this control is absent in the HSV test routinely used in the laboratory. As described with other HSV molecular assays (36–39), the prevalence of PCR inhibition was low in this study (1.1%) and concerned CSF specimens only. Approximately 7% of the tested CSF samples were found invalid; the test would need to be validated on a larger number of CSF samples for verifying whether this observation is anecdotal or corresponds to a recurrent phenomenon with this type of sample, which could be a limitation for the test. In the case of inhibition, the specimen should be retested after dilution or a freeze-thawing cycle.

The level of analytical sensitivity for detecting HSV and VZV DNA in CSF specimens is critical. The lack of sensitivity of most of the HSV DNA assays tested during an international external quality assessment performed more than 10 years ago conducted by Schloss et al. recommends that the analytical sensitivity of NAT for the diagnosis of HSV encephalitis should be close to 200 genome copies/ml (22). Several studies confirmed that the HSV DNA load can be low in CSF in the case of encephalitis (15–20). The VZV DNA load in CSF can also be low (15), especially in the case of VZV-associated vasculopathy (22), meningitis (40, 41), and encephalitis (21), leading to the recommendation of combining the titration of VZV antibodies in CSF with the direct detection of viral DNA (3, 42). Few of the assays published in the literature (10) have been tested for LoD estimation, and only one met the criterion of 200 copies/ml for HSV in CSF specimens (43); another one exhibited a correct sensitivity but was validated for cutaneous and mucous swabs only (44). Concerning the test described in the present report, the LoDs were shown to be less than 200 and 500 copies/ml for HSV and VZV, respectively. This high sensitivity was confirmed on the 2013 QCMD panels; it validates the use of the present assay for detecting simultaneously HSV and VZV DNA in CSF and in other specimens exhibiting low viral loads, such as peripheral blood specimens or swabs taken from late cutaneous or mucosal lesions (45).

A limit of the current assay is the lack of distinction between HSV-1 and HSV-2. Although HSV typing could be useful in certain clinical situations, notably in chronic infection occurring in immunosuppressed patients or in severe diseases such as encephalitis and neonatal infections (46–48), the therapeutic issue is not dependent upon the HSV serotype in most cases (49–51). Consequently, HSV typing can be performed in a further step by using specific primers able to distinguish between the two types (10).

The BD Max platform combines the extraction and amplification steps with reduced hands-on time and full automation. Moreover, the master mix can be prepared as a batch that is validated once before use and is stable for approximately 12 weeks, avoiding the daily use of controls in addition to the clinical specimen, thus reducing the cost. The result can be obtained in less than 2 h, allowing the use of this assay for rapid testing done by non-molecular diagnostic staff in case of emergency, especially for the etiological diagnosis of viral encephalitis. In the future, other targets could be implemented in the assay either by using the two remaining channels of detection or by introducing a second master mix in the new extraction strips available from BD, which could allow the additional detection of other neurotropic pathogens such as enteroviruses, Streptococcus pneumoniae, or Neisseria meningitidis.

In conclusion, this new test targeting HSV-1, HSV-2, and VZV, together with an internal control (SPC), is the first multiplex laboratory-developed assay for the simultaneous detection of DNA of these viruses in various clinical specimens, including CSF, on the fully automated BD Max platform; due to its high sensitivity and reliability, it is suitable for implementation in emergency medicine laboratories.