Implementation and Evaluation of a Fully Automated Multiplex Real-Time PCR Assay on the BD Max Platform to Detect and Differentiate Herpesviridae from Cerebrospinal Fluids

A fully automated multiplex real-time PCR assay—including a sample process control and a plasmid based positive control—for the detection and differentiation of herpes simplex virus 1 (HSV1), herpes simplex virus 2 (HSV2) and varicella-zoster virus (VZV) from cerebrospinal fluids (CSF) was developed on the BD Max platform. Performance was compared to an established accredited multiplex real time PCR protocol utilizing the easyMAG and the LightCycler 480/II, both very common devices in viral molecular diagnostics. For clinical validation, 123 CSF specimens and 40 reference samples from national interlaboratory comparisons were examined with both methods, resulting in 97.6% and 100% concordance for CSF and reference samples, respectively. Utilizing the BD Max platform revealed sensitivities of 173 (CI 95%, 88–258) copies/ml for HSV1, 171 (CI 95%, 148–194) copies/ml for HSV2 and 84 (CI 95%, 5–163) copies/ml for VZV. Cross reactivity could be excluded by checking 25 common viral, bacterial and fungal human pathogens. Workflow analyses displayed shorter test duration as well as remarkable fewer and easier preparation steps with the potential to reduce error rates occurring when manually assessing patient samples. This protocol allows for a fully automated PCR assay on the BD Max platform for the simultaneously detection of herpesviridae from CSF specimens. Singular or multiple infections due to HSV1, HSV2 and VZV can reliably be differentiated with good sensitivities. Control parameters are included within the assay, thereby rendering its suitability for current quality management requirements.

PCR based diagnostics allow for a fast and sensitive detection of an increasing range of infectious agents and specimens [10]. Several commercially available and certified test-systems based on diverse protocols are available for separate HSV1, HSV2 and VZV detection from various clinical specimens [12]. In-house multiplex assays covering the simultaneous detection of up to 8 herpesviridae have been published [13][14][15][16]. For all these tests, numerous time consuming and error-prone manual working steps are necessary, i.e. preparation of working solutions, DNA extraction, transfer of DNA and mastermixes to the amplification devices.
These technical drawbacks can be circumvented by the usage of fully automated PCR platforms. Recently, protocols for monoplex real-time PCR testing of HSV1, HSV2 and VZV from skin and mucosal lesions as well as duplex testing of HSV1/2 and VZV from various clinical specimens have been published for the BD Max system (Becton Dickinson, Heidelberg, Germany) [17,18].
The present work was aimed to implement and evaluate a multiplex PCR assay on the BD Max platform to simultaneously detect and specifically differentiate HSV1, HSV2, and VZV in CSF specimens while continuously fulfilling quality management requirements.

Clinical specimens
For clinical validation 123 CSF specimens were analyzed for the presence of HSV1, HSV2, and VZV genomes. Samples were routinely obtained by the attending physician according to clinical indications as well as standard diagnostic algorithms established in this hospital. Samples were collected between 2013-2015 and stored at -80°C. 11 CSF samples previously found to be negative were spiked 1:7 v/v with HSV1 (IC accession: 363055, undiluted = 151500 copies/ml) or HSV2 (IC accession: 363043, undiluted = 28850 copies/ml) positive reference materials from Instand e.V. interlaboratory comparisons (S1 Table).

Ethics statement
All clinical investigation has been conducted according to the principles expressed in the Declaration of Helsinki. The study was approved by the institutional review board of Rostock University Hospital (Ethikkommission an der Medizinischen Fakultät der Universität Rostock; approval no: A 2015-0161) in line with national and ICH-GCP guidelines. All patients provided written informed consent.

Primers and Probes
Primers and probes were used as previously described [19,20]. Briefly, HSV1 primers and probe were directed to the genomic junction between the glycoprotein G and J genes (nt 137680-137742, accession no. X14112) [20], HSV2 primers and probe to the glycoprotein G gene (nt 139824-139896, accession no. NC_001798) [20] and VZV primers and probe to the glycoprotein B gene (nt 57634-57689, accession no. JN704710). The VZV probe contains incorporated locked nucleic acids (LNAs) to improve the sensitivity and specificity of this short probe [21][22][23]. All primers and probes were purchased from an accredited commercial provider (Tib Molbiol Syntheselabor GmbH, Germany) ( Table 1). In silico analysis (Genbank, NCBI) of the primer-probe-sets showed no cross-homology to non-targeted DNA-sequences.
Routine diagnostic assay for the detection of HSV1, HSV2, VZV The routine diagnostic assay is carried out in the DAkkS (Deutsche Akkreditierungsstelle GmbH) accredited laboratory of the Institute of Medical Microbiology, Virology and Hygiene of the University Hospital of Rostock. HSV 1, HSV2 and VZV are detected together with EBV, CMV, HHV6 and a Sample Process Control (SPC) within one test run that is partitioned into two qualitative duplex PCRs (HSV1/HSV2, VZV/HHV6) and one quantitative triplex PCR (EBV/CMV/SPC). The standard extraction step is performed with 200 μl aliquots of patient samples using the NucliSens easyMAG instrument with the Specific B protocol as stated by the manufacturer (BioMérieux, Nürtingen, Germany). The elution volume of extracted sample is 50 μl from which 10 μl are subsequently applied to individual PCR assays. For amplification/ detection the LightCycler 480 Probes Master and the LightCycler 480/II are used (Roche, Mannheim, Germany). Amplification mixtures and cycling conditions were adopted from previous publications with minor modifications [19,20,[24][25][26][27]. This method has an LoD of about 200 copies/ml. It has been validated and accredited and is now used for routine diagnostics of various clinical specimen including CSF. ROX, X-rodamin; BBQ, black berry quencher; T m /°C melting temperature; GC/%, relative guanidine and cytosine content; +C, +T, +A, locked nucleic acids (LNAs).

BD Max assay
The assay was designed as a fully automated multiplex real-time PCR test including DNA extraction, PCR and results interpretation. DNA was extracted from 400 μl patient samples using the BD MAX™ ExK™ DNA-2 extraction Kit. The BD Max settings were as follows: The "Master Mix Format" was "Type 2 using BD MMK or BD MMK (SPC) with liquid primers and probes" (BD, Heidelberg, Germany). The BD MMK (SPC) master mix, including all required ingredients as well as a Drosophila melanogaster sample process control (SPC), was used with standard extraction parameters. Standard color compensation settings were used. PCR was performed as follows: initial single step.: 600 s at 98°C, type "hold"; cycle process.: 5s at 98°C and 41.9 s at 60°C type "2-temperature", 45 cycles. Detection parameters for channel 475/520, 530/565 and 680/715 were adjusted in the following way: "gain": 40, "Threshold": 200. For channel 585/630 "gain" was adjusted to 30 and "Threshold": 200. Signals with intensities below 200 were assessed as "negative", intensities above 200 were evaluated as "positive".
The primers/probes mix consisted of 20 pmol of each primer, 10 pmol of each probe, 16% v/v BD Max Diluent (BD, Heidelberg, Germany) and was adjusted to 12.5 μl with nuclease free water (Promega, Madison, WI USA). The mix was subsequently aliquoted in BD Max snap-in tubes, sealed using the Axygen Plate-Max semiautomated plate sealer (Axygen Scientific, Union City, CA, USA) (8s at 180°C), and stored at -20°C until use. The mixture was shown to be stable for at least 3 months but was generally used within 2 weeks.

Controls
To ensure the adequacy of the test runs, SPC and plasmid based positive controls were used. The SPC, already included within the extraction stripes and the BD MMK (SPC) Mastermix, confirms the quality of the extraction procedure and the suitability of the analyzed patient material. Negative samples lacking the SPC signal were assessed to contain inhibitory factors resulting in non evaluable PCR results. A plasmid based positive control (PC) containing 10 5 copies/ml of commercially available plasmids carrying amplification targets of each virus was employed to ensure the functional stability of the frozen primer/probes (GenExpress Gesellschaft für Proteindesign mbH, Germany, Table 2). It was included in each test run using an additional reagent strip.

Measurement of test duration
Total time as well as hands-on time was recorded from the beginning of the procedures until the final evaluation of the results (Fig 1). Time was averaged from the work of different technicians (Fig 1). Only test runs containing the same amount of samples were incorporated into the analysis. The preparation of master mixes as well as primer/probes mixes were not included into the work flow analysis since these steps (p1, p2, p3) are inherent part of the analysis protocol.

Reproducibility
Reproducibility of the results was checked by intra-and interassay analysis. Tests including 3 positive samples of each specimen were performed 3 times within the same run and 3 times in different runs.

Sensitivity/LoD, Statistical analysis
Serial dilutions of the Instant e.V. reference samples in previously negative tested CSF were used to determine the analytical sensitivity (LoD) of the BD Max assay. Statistical analyses of CI and COV were performed from determined C T values. For means interval estimates in terms of 95%-confidence intervals were specified to provide an information about extent of uncertainty inherent in results derived from data that are themselves only a randomly selected subset of a population. If necessary, lower bound was stated as zero.

Inter-species cross reactivity
To ensure specificity of the HSV1, HSV2, VSV assay, 25 viral, bacterial and fungal species (S2 Table) typically involved in systemic as well as CNS infections were examined with the above described test format. All species were purchased from ATCC or Instand e.V. Interlaboratory comparison (S2 Table).

Results
The study focused on methodic benefits on routine diagnostic workflow, reproducibility, sensitivity, as well as the specificity of the individual species-directed PCR assays and compared the new method with an established accredited laboratory protocol employing. Turn around and hands-on time reduction In comparison to the PCR protocol utilizing the Nuclisens easyMAG and the Roche LightCycler, the new BD Max analysis protocol was found to be about 27%-39% faster, depending on sample numbers (Fig 1). The analysis of 24 samples employing the BD Max system takes about 190 min in average compared to 260 min when utilizing the conventional procedure, for analysis of single samples, 122 min and 200 min, respectively (Fig 1). Again depending on sample numbers, hands-on time on the BD Max was found to be reduced for about 60% to 74% (24 samples: 45 min vs. 111 min, 1 sample 10 min vs. 38 min) as compared to the conventional procedure. Work steps could be reduced from 13 to 5 steps using the BD Max protocol (Fig 1).

Reproducibility
The c T values of all specimens confirmed reproducibility within a coefficient of variation between 0.4%-2.1% in intra-assay and between 0.6%-3.2% in inter-assay analysis ( Table 3).

Specificity
No cross reaction was observed between HSV1, HSV2, and VZV positive samples investigating patient-and reference materials (S1 and S2 Tables). Inter-species cross reactivity could be excluded for 25 common viral, bacterial and fungal species (S2 Table).

Clinical validation
123 clinical CSF specimens were tested utilizing both, the BD Max system and the established method using the easyMAG in combination with the Lightcycler 480/II. In all, the observed concordance between the two methods was 97.6% for CSF specimens. 3 samples showed negative SPC detection resulting in non-evaluable PCR data. (Table 5, S2 Table).  (Table 6).

Discussion
HSV1, HSV2 and VZV are common pathogens causing severe CNS infections [1][2][3]. Numerous in-house protocols and commercially available "CE"-certified platforms such as the "Lyra 1 Direct HSV1+2/VZV assay" (Quidel Germany GmbH, Germany), the "RealStar 1 alpha Herpesvirus PCR" Kit (Altona Diagnostics, Germany) and the "HSV-1 HSV-2 VZV Rgene 1 kit" (Argene/bioMerieux, France) are available for their detection, but all of these tests are either not approved for CSF or cannot be run with full automation [12,[28][29][30][31][32]. The introduced, fully automated multiplex real time PCR assay on the BD Max platform allows for the simultaneous detection and specific differentiation of HSV1, HSV2 and VZV from cerebrospinal fluids. Quality parameters, demanded for example by quality management systems, are covered by including sample process and positive controls. Manual preparation steps, hands-on time and total processing time could be significantly decreased as compared to a conventional Nuclisens easyMAG and Roche LightCycler based protocol, facilitating usage in routine diagnostics. The simplified workflow enables a more variable routine work plan especially in cases of additionally requested emergency diagnostics. This in combination with the complete automation could reduce flaws in sample preparation inevitable present when manual handling is performed. Prior studies pointed out low HSV1-, HSV2-and VZV-DNA concentrations in CSF in cases of acute encephalitis and meningitis [12,[33][34][35][36][37][38][39][40][41][42]. Therefore, HSV1/2 and VZV detection  levels of approximately 200 copies/ml are considered to be necessary for adequate diagnostics [43]. The LoD of the present PCR protocol clearly fulfills these postulated sensitivity criteria. The comparison of 123 CSF specimens prospectively analyzed with the established routine PCR and retrospectively investigated with the new protocol revealed a 97.6% concordance. In comparison to other methodical studies the discordance of 2.4% is within the range normally found when comparing different herpesviridae PCR protocols [17,40,[44][45][46].
Cardenas and colleagues established a HSV1, HSV2 and VZV assay on the BD Max platform for specimens from skin and mucosal lesions as a proof of principle to demonstrate the BD Max to be a suitable one-step open-access-platform [18]. In contrast to the present work they established two different protocols to detect HSV1 and HSV2 as duplex-and VZV separately as a singleplex PCR. This approach does not cover the technical potential of the BD Max to analyze up to 5 PCR targets within one test, and thus, increases the material costs and complicates the workflow. Furthermore, the required amounts of patient material plays a minor role when using skin or mucosal swabs, but could be a major obstacle for CSF diagnostics.
Pillet and colleagues published an excellent protocol to detect HSV and VZV from various specimens including 38 CSFs utilizing the BD Max platform [17]. In comparison to their protocol the present method contains some features leading to improved analytical value and feasibility, like specific differentiation between HSV1 and 2, usage of ready-made mastermixes and positive controls. Despite similar clinical manifestations, the discrimination between HSV1 and HSV2 is important since CNS infection site, potential complications, responses and resistance rates to antivirals, as well as reactivation rates on mucosal surfaces vary between both virus types. Therefore, a HSV-type-specific test result leads to a more detailed statement on the infection status and its prognosis [47,48]. The described utilization of the commercially available pre-assembled BD MMK(SPC) mastermix leads to a decreased labor input for sample preparation and simplifies the quality management processes since the tubes are delivered with lyophilized aliquots and with information on batches as well as expiry dates. The additional use of plasmid based PC improves validity of negative results ensuring the functionality of the inhouse prealiquoted primer/probes mixes.
Recently, FilmArray Meningitis/Encephalitis Panel, a cartridge based fully automated multiplex PCR-system was introduced allowing for the detection of 14 pathogens, including HSV1, HSV2 and VZV, requiring little hands-on and turn-around time (BioFire, USA). In comparison to the presented BD Max assay, the LODs achieved with the FilmArray system for these viruses range from 1290 to 1660 copies/ml. This is approximately 7.5-19.7 times higher than the LODs achieved by the presented method. This potentially could lead to false negative results, especially when dealing with low viral loads.
In conclusion, this new method is the first fully automated multiplex assay for the simultaneous detection and differentiation of HSV1, HSV2, and VZV from CSF specimens on the BD Max platform. The combined usage of a positive and a sample process control covers relevant quality interests. Complete method validation including 123 clinical CSF specimens gives a profound and reliable overview about the test performance in routine settings.
Supporting Information S1