Duplex Real-Time RT-PCR Assays for the Detection and Typing of Epizootic Haemorrhagic Disease Virus

Epizootic haemorrhagic disease virus (EHDV) may cause severe clinical episodes in some species of deer and sometimes in cattle. Laboratory diagnosis provides a basis for the design and timely implementation of disease control measures. There are seven distinct EHDV serotypes, VP2 coding segment 2 being the target for serotype specificity. This paper reports the development and validation of eight duplex real-time RT-PCR assays to simultaneously amplify the EHDV target (S9 for the pan-EHDV real-time RT-PCR assay and S2 for the serotyping assays) and endogenous control gene Beta-actin. Analytical and diagnostic sensitivity and specificity, inter- and intra-assay variation and efficiency were evaluated for each assay. All were shown to be highly specific and sensitive.

To date, seven EHDV serotypes have been identified [6]. Some (EHDV-1 and EHDV-2) are endemic in the United States and Canada, causing severe clinical episodes with high mortality in some species of deer such as white-tailed deer [7,8]. Moreover, in 2006, an exotic strain of EHDV-6 was isolated from moribund and dead white-tailed deer (Odocoileus virginianus) in Indiana and Illinois. This EHDV-6 strain originates from a reassortment between co-circulating strains of serotypes 2 and 6 [9]. Several strains of EHDV-2, -6 and -7 can also lead to severe disease in cattle: in Japan (Ibaraki strain, serotype 2) [10]; in 2006 in Morocco, Algeria and Turkey (EHDV-6) [11]; in 2006 in Israel (EHDV-7) [12]; and in 2009 in Reunion Island (EHDV-6) [13]. More recently, EHDV-6 and -2 were reported in the Caribbean basin (Martinique, Guadeloupe Islands and French Guiana), while EHDV-1 was detected in French Guiana without clinical signs [14]. In western Kenya, a survey conducted between 2007 and 2010 showed the circulation of EHDV in indigenous calves [15]. Some positive samples could not be serotyped by seven serotype specific gel-based RT-PCRs [16]. The authors suggest that this is because the gel-based serotyping assays are less sensitive than the real-time group-specific RT-PCR, or because of mismatches in the VP2 primer sequences which were designed from the reference EHDV strains. Moreover, EHDV is regularly detected in areas where BTV also circulates (e.g. Reunion, French Guiana, Martinique and Kenya) [13][14][15]17]. The genome of these two viruses is very close, so it is important to have molecular tools able to both differentiate and serotype these viruses. In the last few years, real-time RT-PCR assays have been developed for the detection of all seven serotypes [18,19]. To our knowledge, only conventional RT-PCR assays were available up to now for virus serotyping [16,20].
This paper describes the development and validation (according to the methods described in chapter 1.1.5. of the World Organisation for Animal Health, OIE [21]) of eight duplex realtime RT-PCR assays for the detection and serotyping of each EHDV serotype (EHDV-1/-3, -2, -4, -5, -6, -7 and -8): a duplex pan-EHDV real-time RT-PCR assay that amplifies a portion of segment 9 of all EHDV strains coupled with a real-time RT-PCR that amplifies the beta-actin RNA of ruminant cells (housekeeping gene) and a set of seven duplex serotyping real-time RT-PCR assays that specifically amplify both the EHD segment 2 portion of each serotype and the beta-actin gene. The Ct (Cycle Threshold) value obtained with the real-time RT-PCR that amplifies the housekeeping gene is used to verify the integrity of the total RNA extracted and the absence of inhibitors that can interact during the RT-PCR steps for each biological sample tested. These assays were developed and validated to detect and serotype EHDV for application to biological samples from the field.

Strains and samples
Reference strains. The different EHDV reference strains were used to check the initial specificity of the assays.
All isolates were obtained from IEMVT-CIRAD and the Pirbright Institute. Biological samples and field strains. For the diagnostic specificity evaluation, blood samples from 234 animals known to be uninfected were collected in EHDV-free areas of Metropolitan France.
All the infected biological samples used in this study were taken either from naturally infected domestic cattle or buffaloes, or from experimentally infected animals [13,14,17,22]: 40 blood samples infected by EHDV-1 (Reunion Island and French Guiana); 32 blood samples infected by EHDV-6 (Reunion Island and French Guiana); 24 blood samples infected by EHDV-6 (experimental infection).
All the Biological samples used in this study were taken during our own previous studies. Samples from experimentally infected animals were taken during a previous study published by Breard et al [17]. All experimental protocols were reviewed by a state ethics commission and have been approved by the competent authority (Comité d'Ethique en Expérimentation Animale Val de Loire-CEEAVdL reference: 2011-09-1).
Naturally infected and non-infected samples were collected between 2009 and 2013 as part of routine veterinary investigations carried out by qualified veterinarians in the respective area of origin.
All field strains used in the study were obtained during our own previous studies and no specific passage in cells culture or in ECE were achieved for this work. Further ethical approval was not therefore needed.
The two EHDV-6 (Morocco) and EHDV-7 (Israel) field strains were provided by Biopharma and the Pirbright Institute respectively.

Primers and Probes
EHDV Target. Primers and TaqMan minor groove binder (MGB) probes were designed by multiple sequence alignment of segment 9 (available in Genbank) using MegAlign 8.0.2 (Lasergene DNASTAR) for the detection of all seven serotypes and the sequences of segment 2 for serotyping. The primers and probes were selected and evaluated using Primer Express software (Primer Express Software v3.0.1 by Lifetechnologies).

Nucleic Acid Sample Preparation
Total RNA was extracted from 100μl of blood samples or culture cell supernatants using the QIAcube robot (QIAGEN) or the Kingfisher 96 robot (Life Technologies) with the QIAamp Viral kit (QIAGEN; reference 52906) or the MagVet Universal isolation kit (Life Technologies; reference MV384) according to the manufacturer's instructions.

Duplex One-Step RT-PCR
The different primers and probe sets were used to amplify a part of the targeted gene using the AgPath One-Step RT-PCR Kit (Life Technologies; reference 4387424). The final composition of the mix used for the eight real-time RT-PCRs was optimised.
After RNA denaturation with 10% DMSO and heating at 95°C for 3 minutes, 5 μl of denaturated RNA was added to the mix and the amplifications were carried out using the same cycling parameters for all of the assays: 45°C for 10 min, 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Fluorescence was measured during the 60°C annealing/extension step. Cycle threshold (Ct) values were measured at the point at which the sample fluorescence signal crossed a threshold value (the background level). Negative results (for assays that did not exceed this signal level) are reported as 'No Ct'.

Analytical Specificity
A panel of different EHDV strains including all seven serotypes was tested in each of the eight assays to check specificity. Serotypes of bluetongue, African horse sickness virus (AHSV) and other pathogens infecting cattle were also tested to ensure that there was no cross-reaction.

Analytical Sensitivity
PCR detection limit (LOD PCR ). The detection limit of the PCR (LOD PCR ) is the smallest number of copies of the target nucleic acid per unit volume that can be detected in 95% of cases. It was determined by three independent sessions that were conducted by testing eight replicates of three dilutions around the detection limit.
In order to evaluate the LOD PCR , a synthetic single strand RNA corresponding to the nucleotide sequence of each targeted amplification product was produced by in vitro transcription using a T7 promoter. An RT-PCR was performed with forward primers modified by addition of a T7 promoter sequence (ATT AAT ACG ACT CAC TAT AGG) [24] at the extreme 5' end of the upstream primer (no modification for the reverse primers).
The PCR product was then purified with a QIAquick PCR Purification Kit (QIAGEN). In vitro transcription was performed at 37°C for 4 h with a MEGAscript T7 kit (Lifetechnologies). After the transcription reaction, the template DNA was degraded by addition of 1μl of RNasefree DNase I for 15 min at 37°C. The transcribed RNA was purified with a MEGAclear kit (Lifetechnologies) and dosed by spectrophotometry.
Method detection limit. The method detection limit (LOD method ) is used to estimate the smallest quantity of biological target that must be initially present in the sample for it to be detected. Two independent sessions of four extractions were performed on three quantities of synthetic RNA around the detection limit added to a negative lysed blood sample.
Efficiency. The efficiency of real-time RT-PCR assays was estimated on the basis of a standard curve plotting Ct values against the corresponding log ssRNA copy number per reaction. Efficiency was calculated by the formula: E% = (10 -1/slope -1)x100.

Diagnostic Specificity and Sensitivity
Diagnostic specificity is the proportion of samples from animals known to be uninfected that test negative in the assay. It was estimated using the blood samples from cattle sampled in EHDV-free areas of Metropolitan France. Between 92 and 138 samples were analysed for the different EHDV serotyping assays, and 234 for the pan-EHDV assay.
Diagnostic sensitivity is the proportion of samples from animals known to be infected that test positive in the assay. It was estimated using EHDV-positive blood samples from 96 animals collected in French Guiana and Reunion Island or during animal experimentation. Ninety-six samples were analysed for the pan-EHDV assay, 40 for the EHDV-1 serotyping assay and 32 for the EHDV-6 serotyping assay. No infected biological samples were available for the EHDV-2, -4, -5, -7 and -8 serotyping assays. Confidence intervals were calculated using binomial distribution [25].

Repeatability
The repeatability of the eight real-time RT-PCR assays was evaluated by determining the intraand inter-assay coefficient of variation (CV). A panel of eight positive samples with different viral loads (cycle threshold range 23-35) was tested in triplicate in two independent sessions. The CV was calculated for each sample [26].
Intra-assay CV = (average of each session's Ct standard deviation (SD)/average of each session's Ct average) x100.
Inter-assay CV = (SD of the Ct averages for each session/average of the averages for each session) x100.

Primer and Probe Design
A comparison of segment 9 nucleotide sequences (available in GenBank) from EHDV strains isolated around the world, and representing the seven EHDV serotypes, showed a conserved sequence in the 5' end region that allowed the design of two primers and one probe for the pan-EHDV real-time RT-PCR assay (Table 1).
By aligning the segment 2 sequences of each EHDV serotype, it was possible to design two primers and one probe for EHDV-4, -5 and -8. There was greater sequence divergence among the different strains of serotypes EHDV-1, -2, -6 and -7, so more than two primers were designed for the specific real-time RT-PCR assays in order to ensure sensitivity.

Real-time RT-PCR Mix
The composition of the mix after optimisation is presented in Table 2.

Analytical Specificity
All EHDV strains (including all seven serotypes) from infected cell cultures tested positive using the pan-EHDV real-time RT-PCR assay (Table 3). For the seven serotyping RT-PCR assays, only the homologous serotypes gave positive results (Table 3). No amplification was observed (no Ct) for BTV, AHSV or other pathogens that infect cattle (Table 4).

Analytical Sensitivity
The PCR and method detection limits (LOD PCR and LOD Method ) and efficiency values for the eight real-time RT-PCR assays are presented in Table 5. The LOD PCR and LOD Method varied between 1 to 50 and 10 to 1,000 copies respectively. Efficiency rates were determined for the

Repeatability
The intra-and inter-assay coefficients of variation are presented in Table 5. The lowest and highest percentages are shown for each assay. All inter-and intra-assay CVs are less than 5%.

Discussion
In order to obtain good performance for the pan-EHDV assay, we decided to target VP6 encoding gene S9. In silico studies of S9 sequences available in GenBank showed that nucleotide variations ranged between 1.2 and 33.9%. However, a conserved region (in the 5' end region) appeared to be of interest for designing the primers and probe. Only one set of two primers and one probe was designed for this assay, whereas the previously reported pan-EHDV real-time RT-PCR assays use several primer pairs and/or probes targeting the NS3 or NS1 encoding genes (S10 and S5) [18,19]. The cost of the reaction can therefore be significantly reduced using this new assay. All the reference or field strains were detected and no amplification was generated with nucleic acid from the similar Bluetongue virus African Horse Sickness virus or various pathogens infecting cattle. Good diagnosis sensitivity (99%) and specificity (98%) rates were observed when testing samples from the field. Analytical sensitivity (LOD PCR ) is very close to that reported by Schroeder et al (<100 copies) [26]. These authors targeted segment 5 (encoding NS1) when developing a real-time RT-PCR assay to detect all seven serotypes. Seven real-time serotyping RT-PCR assays have also been developed and validated that target VP2 encoding gene S2, which is the primary determinant of virus serotype [6,27]. In silico analyses have revealed high variability of up to 30-35% between the segment 2 nucleotide sequences of different strains belonging to the same serotype (e.g. a 32% difference between the EHDV-6 of Reunion Island (Accession number HQ848379) and of Guadeloupe (Accession number HQ848380); a 32% difference between EHDV-7 Australia (Accession number AM745048) and Israel (Accession number HM156731)). Consequently, several selected primer pairs/probes were added into the same mix to be able to amplify all the different strains of EHDV serotypes 1, 2, 6 and 7 (Table 1). All seven serotype-specific assays showed good results in terms of sensitivity and specificity.
The high variability of the S2 gene underlines the need to verify for news strains that the primers/probe binding site is suitable. It also shows the need to sequence the RNA segments targeted by these real-time RT-PCRs in order to supply and complete the data bank. It might be necessary to design a new set of primers and probe.
To date, the traditional method for identifying EHDV from clinical samples is viral isolation on culture cells or embryonated chicken eggs, followed by a vironeutralisation test (VNT) using reference sera [28]. Unfortunately, reference sera are not available for each serotype. Moreover, these procedures are very time consuming (at least two weeks). To improve EHDV diagnosis, various RT-PCR assays have been developed. In the literature, only gel-based RT-PCRs are described for serotyping EHDV [16,20]. In a previous study, we described a method to easily identify EHDV serotypes using both conventional RT-PCR and sequencing [13]. To our knowledge, this is therefore the first report of real-time serotyping RT-PCR assays.
These molecular tools have been specifically developed and validated to quickly detect EHDV directly from ruminant blood. The diagnostic sensitivity and specificity of the pan-EHDV and the EHDV-1 and -6 real-time serotyping RT-PCR assays were evaluated using characterized biological samples collected during epidemics, epidemiological surveys [13,14,22] or naïve animals sampled in an EHDV-free area. The results, ranging from 98 to 100% for specificity and 99 to 100% for sensitivity, demonstrate that these tools were efficient in detecting EHDV and determining its serotype from a field blood sample. Moreover, the simultaneous amplification of an endogenous control gene (Beta-actin from wild or domestic ruminants) as an internal control attests to the quality of the analysed sample and the absence of PCR inhibitors [23]. If no Ct was obtained for the endogenous control gene when testing a ruminant blood sample, negative results for the EHDV target were not taken into account. In 2007, we already reported the use of this endogenous reporter gene [23]. However, beta-actin was not amplified in the same tube as the pathogen target (BTV). Duplexing is a real improvement for real-time RT-PCR assays. It optimises experiment efficiency by decreasing the volume of pipetted sample required and the cost of reagents, and by increasing the number of samples tested in the same run. The primers/probe set for all of the pan-EHDV and serotyping assays described above were also determined for use with the same cycling parameters, and can therefore be used in the same run.
EHDV infection is widespread in the USA, where it causes severe epidemics in wild Cervidae, particularly white-tailed deer (Odocoileus virginianus) [29]. According to Allison [9], EHDV is one of the most important viral diseases of white-tailed deer worldwide. In the last decade, EHDV has also been reported in Maghreb, in the Middle East [11], Africa [15], the Caribbean basin [14] and Reunion Island [13,22,29] especially associated with disease in cattle. In most infected areas, Bluetongue and EHD virus co-circulate, and mixed infections are reported. These reports highlight the need for efficient diagnostic tools to quickly detect and identify the causal agent. The pan-EHDV assay developed in this work was able to detect all the strains and biological samples included in the study, representing all seven EHDV serotypes, without any cross reaction with Bluetongue strains. Moreover, the seven serotyping assays gave good results in terms of analytical/diagnostic specificity and sensitivity, making them reliable tools for serotype identification.
In conclusion, the development of such molecular assays for EHDV diagnosis is very useful for disease control. It is crucial to have a better knowledge of the worldwide distribution of this virus and to add viral sequences to the data bank.