Development and evaluation of real-time TaqMan^® RT-PCR assays for the detection of avian nephritis virus and chicken astrovirus in chickens

Abstract

The development and preliminary evaluations of two TaqMan^®-based, real-time reverse transcriptase-polymerase chain reaction (rRT-PCR) assays for the quantitative detection of avian nephritis virus (ANV) and chicken astrovirus (CAstV) RNAs are described. The assays used amplicons generated from the 3′ untranslated region of the ANV genome and a conserved region of CAstV open reading frame 1b including its junction with open reading frame 2. High virus RNA levels (>10^5.99 viral copies) were detected for ANV and CAstV in 81% and 67% gut content samples from growth-retarded broiler flocks. Results from longitudinal surveys of two broiler flocks showed that ANV and CAstV RNAs were detected in most gut content and kidney samples collected at all time points from day 0 to day 35, with RNA levels of both astroviruses being higher in the gut contents than in the kidneys, and with the ANV RNA levels being greater than those of CAstV especially at early (days 7 and 14) time points. When the results obtained for the days 4/5 time-point samples from four broiler flocks with varying growth performances were compared, the two better-performing flocks had 100-fold to 1000-fold less ANV viral copies than the flocks that performed least well. Application of the rRT-PCR tests to samples collected from broiler chicks, which were experimentally infected with a crude gut content inoculum, demonstrated that ANV RNA could be detected in gut content and kidney samples at levels similar to those found at corresponding time points in longitudinal survey samples, whereas CAstV RNA was detected at lower levels than in the longitudinal survey samples, especially in kidney samples.

Introduction

Avian astroviruses were first identified in turkeys (turkey astrovirus) when particles with five-pointed and six-pointed star morphology were observed by electron microscopy in the faeces of UK poults suffering from diarrhoea and increased mortality (McNulty et al., 1980). Later, duck hepatitis virus type 2 (DHV-2) was also identified as an astrovirus by electron microscopy (Gough et al., 1984). Avian nephritis virus (ANV), which is associated with interstitial nephritis and growth retardation in chickens, was originally regarded as a picornavirus (Maeda et al., 1979) on the basis of virion size, morphology and physicochemical properties. However, following genome sequence determination and analysis, it is now classified as an avian astrovirus (Imada et al., 2000). More recently, isolates of a second astrovirus of chickens, named chicken astrovirus (CAstV), were recovered from growth-retarded broilers and were later identified as a novel avian astrovirus by partial sequence determination (Baxendale & Mebatsion, 2004).

The detection of ANVs in stunted broilers and the ability of some ANVs to cause growth depression following experimental infection of specific pathogen free chickens and broiler chickens has led to speculation that ANV may be important in causing growth problems in the field (McNulty et al., 1984a; Decaesstecker et al., 1986; Shirai et al., 1992). Additional evidence indicates that CAstVs may also be contributing to such problems. Antigenically different CAstVs, originally regarded as enterovirus-like viruses (McNulty et al., 1991; McNeilly et al., 1994), have recently been characterized (Todd et al., 2009a). One of these CAstVs, known as “FP3”, was isolated in the UK in 1984 from dead-in-shell chicks as part of an investigation into early broiler mortality (Spackman et al., 1984), and the second CAstV, known as “612” was isolated in South Africa from broilers with respiratory problems (McNeilly et al., 1994). Experimental infections of 1-day-old specific pathogen free and broiler chicks have demonstrated that these CAstVs can cause varying degrees of growth retardation (McNeilly et al., 1994, Smyth et al., 2007; Todd et al., unpublished results). The study by Smyth et al. (2007) showed that, following experimental infection with the FP3 isolate, histological lesions and virus-specific antigen were detected in the intestine, kidney and pancreas, indicating that, like ANV, CAstVs have the ability to infect internal organs.

Although ANV and CAstV have been implicated in growth depression including “uneven growth” and “runting-stunting syndrome”, the nature and extent of the disease problems that they may be causing remain unknown due to the absence of convenient diagnostic tests. Both viruses grow poorly in cell culture, making virus isolation difficult, and virus-specific antibodies are not commonly available for use in immunostaining-based diagnostic tests such as immunohistochemistry. Reverse transcriptase-polymerase chain reaction (RT-PCR) tests have recently been described for both ANV (Mandoki et al., 2006; Day et al., 2007; Todd et al., 2010) and CAstV (Pantin-Jackwood et al., 2006; Smyth et al., 2009). Using conventional RT-PCR tests, work in this laboratory has shown that both viruses were detected in very high (> 96%) proportions of gut content samples from growth-retarded broiler flocks, and were also commonly detected in pooled gut content samples collected in longitudinal surveys (days 0 to 42) of four broiler flocks with average and below-average performances (Smyth et al., 2009; Todd et al., 2010). These studies and earlier serological investigations with ANV (Imada et al., 1980; Connor et al., 1987) and CAstV (Todd et al., 2009b) showed that infections with these avian astroviruses were ubiquitous in broiler chicken flocks. As a consequence, while highly sensitive RT-PCR tests may be useful for detecting ANV and CAstV infections, they may be of limited use for differentiating subclinical infections from those that result in pronounced adverse effects on broiler growth performance. Previously we speculated that higher levels of astrovirus replication are likely to be associated with more severe pathogenic effects and that quantitative real-time RT-PCR (rRT-PCR) tests might be useful for demonstrating differences in virus levels between flocks that vary in their growth performance. In this paper we describe the development of the first TaqMan^®-based real time RT-PCR tests for ANV and CAstV and their application for quantifying virus loads in gut content and kidney samples. In the case of CAstV, the test involved the amplification of a 70 bp product from a highly conserved region of the CAstV genome located at the 3′ end of open reading frame (ORF) 1b, which encompassed the intergenic region between ORF1b and ORF2. In the case of ANV, an amplicon of 56 bp was produced from the highly conserved untranslated region at the 3′ terminus of the genome (3′ untranslated region [UTR]). Both tests were applied to gut content samples from growth-retarded broiler flocks from the UK and the USA and applied to gut content and kidney samples collected in longitudinal surveys of four broiler flocks and in experimental infections of broiler chickens.

Materials and Methods

Viruses

Information relating to the source, propagation history and production of the five CAstV isolates (FP3, 11672, 612, 1009 and 11522), the DHV-2 and DHV-3 isolates and the G4260 serotype 1 ANV isolate (ANV-1) used in this investigation was previously described by Smyth et al. (2009).

Clinical samples

Gut contents, faecal matter, kidneys or cloacal swabs were sampled from chickens collected from problem flocks in the UK and the USA, and were stored at – 80°C. Samples from the UK were kindly provided by A. Fernandez-Gutierrez (Aviagen Ltd, Midlothian, UK), M. Alcorn (St David's Poultry Team Ltd, Exmouth, UK), B.H. Thorp (St David's Poultry Team at the Dick Vet, Easter Bush Veterinary Centre, Midlothian, UK), C. Prins, G. Hayes and A. Walker (Slate Hall Veterinary Practice Ltd, Cambridge, UK), P.W. Cargill (Wyatt Poultry Veterinary Services, Hereford, UK), and D. Pearson (VION UK Veterinary and Food Diagnostic Service, Grampian Country Chickens (Rearing) Ltd, West Lothian, UK). Twelve samples (799 MO/2005, 802 AR/2005, 812 DE/2005, 836 NC/2005, 840 AR/2005, 866 GA/2006, 883 MO/2006, 916 CA/2006, 1254 GA/2008, 1255 GA/2008, 1335 GA/2009 and 1340 GA/2009), prepared from diagnostic submissions from US broiler flocks, were provided as clarified 10% faecal extracts by Dr Michael Day (Southeast Poultry Research Laboratory, USDA, Agricultural Research Service, Athens, Georgia, USA).

Longitudinal surveys

Samples from four broiler flocks were tested for the presence of ANV and CAstV using real-time RT-PCR. These flocks were from different sites belonging to the same UK poultry organization and, based on recent performances, were predicted to exhibit average and below-average performances. Gut contents and kidneys from approximately 12 birds from each flock were sampled at each of 10 time points ranging from day 0 to day 42 for two flocks (Flocks A and B), and for the other two flocks (Flocks C and D) 12 birds from each flock were tested at each of two early time points (day 5 and day 7) . The overall performance of each flock was estimated after slaughter by calculating European production efficiency factor (EPEF) values, determined by the following equation:

EPEF values constitute standard measures of flock performance, and data relating to all birds in the flock are taken into consideration.Gut contents were processed as described previously (Smyth et al., 2009). Kidney samples were ground in a mortar and pestle and were diluted 1:10 in chilled phosphate-buffered saline. The organ homogenates were clarified by centrifugation at 4500 x g for 30 min at 4°C and the supernatants were retained for RNA extraction.

Experimental infection of broiler chickens

Fertile eggs were sourced from a grandparent flock, which had tested negative for antibodies specific to the 11672 isolate of CAstV using an indirect immunofluorescence test described by Todd et al. (2009b). Following incubation, the chicks were hatched and transferred into negative-pressure isolators. One group of 1-day-old broiler chicks was inoculated orally with 10% clarified crude inoculum (0.2 ml) prepared from the gut contents of all 22 birds sampled at days 4 and 7 in the longitudinal survey of Flock D. It is acknowledged that this crude inoculum may contain bacteria and other enteric viruses in addition to ANV and CAstV, which were detected in the longitudinal survey samples, from which the inoculum was prepared. Gut contents and kidney samples were removed from five birds at days 3, 7, 10, 14, 21 and 28 post infection (p.i.) and processed as described above. Twenty birds were weighed from the infected and uninfected groups at day 14 p.i.

RNA extraction

Viral RNA was extracted from 140 µl each supernatant using the QIAamp Viral RNA Mini Kit (Qiagen, Crawley, UK) according to the manufacturer's instructions. Each RNA was eluted in 40 µl RNase-free water. A negative extraction control, consisting of water instead of test sample supernatant, was included with each batch of test sample supernatants extracted. The inclusion of these controls was designed to indicate whether cross-contamination between positive test samples may have occurred.

Real-time RT-PCR assay design

A multiple alignment of the ANV 3′ UTR sequences was constructed from the G4260 serotype 1 isolate of ANV (ANV-1; accession number NC003790), a serotype 2 ANV isolate (ANV-2; accession number: AB046864) (Imada et al., 2000), and ANVs present in 20 field samples, which were sequenced in this laboratory (Todd, unpublished results). A highly conserved region within the 3′ UTR was identified as being suitable for the location of the TaqMan^® assay. A previous multiple sequence analysis of partial CAstV ORF1b and ORF2 sequences had revealed a relatively conserved sequence region in which to locate the CAstV TaqMan^® assay (Smyth et al., 2009). However, due to minor sequence variation across the two sets of aligned sequences, both ANV and CAstV forward primers contained degeneracy. The ANV assay amplified 56 bp within the 3′ UTR of the ANV genome, from nucleotides 6669 to 6721 in the ANV-1 genome sequence. The CAstV assay amplified 70 bp within the precapsid region, which comprised the 3′ end of ORF1b and the 24 bp intergenic region between ORF1b and ORF2. The primers and TaqMan^®-labelled probes with non-fluorescent minor groove binder (MGB) moieties were designed manually using the Primer Express^® software version 3.0 (Applied Biosystems, Warrington, UK) and were also synthesized by Applied Biosystems. The assay sequences were examined for specificity by nucleotide BLAST. The sequences of the assay components are presented in Table 1.

Table 1.  Primer and probe sequences used in ANV and CAstV rRT-PCR assays.

	ANV	CAstV
Forward primer	GTAAACCACTGGYTGGCTGACT	GCYGCTGCTGAAGAWATACAG
Probe	6-FAM-CAGCAACTGACTTTC-MGB	6-FAM-CAGAAGTCGGGCCC-MGB
Reverse primer	TACTCGCCGTGGCCTCG	CATCCCTCTACCAGATTTTCTGAAA

Real-time RT-PCR

Real-time RT-PCR reactions were set up in triplicate per sample with a total volume of 20 µl per replicate reaction. Each reaction comprised 10 µl AgPath-ID™ One-step RT-PCR 2x buffer (Applied Biosystems), 0.8 µl AgPath-ID™ One-step RT-PCR enzyme (Applied Biosystems), primers to a final concentration of 400 nM, probe to a final concentration of 120 nM, 2 µl sample or positive control RNA and nuclease-free distilled water (dH₂O) to 20 µl. The 2 µl sample RNA were replaced in the PCR negative control by 2 µl dH₂O or 2 µl negative extraction control. The reactions were conducted in a 7500 Real-Time PCR System (Applied Biosystems) starting with a reverse transcription stage at 45°C for 10 min, then an initial denaturation stage at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 15 sec, and then primer annealing and template amplification at 60°C for 45 sec. Fluorescence readings were taken during the amplification stage. During post-PCR analysis, cycle thresholds were set while the reactions were in true exponential phase prior to the linear phase.

Determination of assay sensitivity, efficiency and specificity

To determine the sensitivity of the ANV assay, an ANV-specific PCR product of 394 bp (Todd et al., 2010) was cloned into the pCR^®II-TOPO vector (Invitrogen, Paisley, UK). An 858 bp CAstV amplicon, derived from the CAstV isolate 11672, was similarly cloned (Smyth et al., 2009). Run-off RNA transcripts were produced and quantified as described previously (Smyth et al., 2009). Ten-fold serial dilutions of the in vitro transcribed RNAs were made and examined in triplicate by rRT-PCR as described above. Limits of detection were determined for both assays from the last dilution to give positive amplification. The standard curves, produced by plotting C_T values versus the viral copy numbers present in the 10-fold dilutions using logarithmic (to the base 10) scales, were used to calculate the viral RNA copy numbers in test samples for both assays. This was done by applying the following equation:

where C_T-Test is the C_T value of the test sample, C_T-_{Y intercept} is the intercept of the slope on the y (C_T) axis and M is the slope. PCR efficiencies were also determined from the 10-fold dilution series of in vitro transcribed RNA using the following equation:

The specificities of the assays to ANV or CAstV were tested using samples of DHV-2, DHV-3, ANV-1 and 5 CAstV isolates as well as field samples, which were shown to be positive for ANV and CAstV using previously described conventional RT-PCR tests (Smyth et al., 2009; Todd et al., 2010). An ANV and a CAstV rRT-PCR amplicon were each cloned and sequenced to ensure specificity.

Assay evaluation

The abilities of extraneous substances, which may be present in the RNA extracts prepared from the gut content samples, to inhibit RT-PCR were assessed using the TaqMan^® Exogenous Internal Positive Control (Exo IPC) Reagents kit (Applied Biosystems). This one-tube end-point analysis approach duplexes the FAM-labelled ANV or CAstV assay with the VIC-labelled Exo IPC assay. The Exo IPC assay was performed at a lower concentration than the test assay, and fluorescence readings were taken both during the amplification stages and at the end of the cycling in a final extension phase. The rRT-PCR reactions were performed in triplicate per sample with a total volume of 20 µl per replicate reaction. Each reaction comprised 10 µl AgPath-ID™One-step RT-PCR 2x buffer (Applied Biosystems), 0.8 µl AgPath-ID™One-step RT-PCR enzyme (Applied Biosystems), ANV or CAstV primers to a final concentration of 900 nM, ANV or CAstV probe to a final concentration of 200 nM, 2 µl of 10x Exo IPC mix (containing IPC assay and polymerase), 0.4 µl of 50x Exo IPC DNA, 2 µl RNA or nuclease-free distilled water (no template control) or 10x Exo IPC Block (no amplification control), and nuclease-free distilled water to 20 µl. The cycling and detection conditions were identical to the above method but ended with a final extension hold and fluorescence reading at 60°C for 1 min.

Statistical analysis

Real-time RT-PCR results, for each gut content/kidney, Flock A/B and CAstV/ANV combination separately, were analysed in GenStat (VSNi Ltd, Hemel Hempstead, UK) using a one-way analysis of variance to assess the effect of time point. If any of these were significant (at the 5% level), then pair-wise differences were determined by least-significant differences. Prior to analysis all variables were transformed by taking logarithms to the base 10. The same method of analysis was employed to look for differences between four broiler flocks for CAstV/ANV in gut content/kidney samples at day 4/5 and day 7 separately by rRT-PCR. In a third experiment, ANV and CAstV rRT-PCR results for gut content and kidney were again analysed using a one-way analysis of variance.

Results

Assay sensitivity and efficiency

The detection limit and efficiency of the ANV rRT-PCR assay were determined using C_T values obtained from a 10-fold dilution series of run-off RNAs, which had been in vitro transcribed from a cloned PCR product of 394 bp. A limit of detection of approximately 180 copies was estimated for the assay, based on the last reproducibly detectable dilution, which had a C_T value <35. Using a similar approach, also involving a 10-fold dilution series of in-vitro-transcribed RNA derived from a cloned 858 bp CAstV11672 fragment, the CAstV assay was estimated to have a limit of detection of approximately 105 copies. For both real-time assays, standard curves—produced by plotting C_T values versus the viral copy numbers present in the 10-fold dilutions using logarithmic (to the base 10) scales—were used to determine C_T-_{y intercept} (intercept of the slope on the y axis) and M (slope) values from which the viral copy numbers could be calculated. For the ANV assay, the C_T-_{y intercept} and M values were 42.146 and – 3.344, respectively, and in the case of the CAstV assay the C_T-_{y intercept} and M values were 40.732 and – 3.385, respectively. PCR efficiencies were estimated to be 99.00% for the ANV assay and 97.98% for the CAstV assay. In both assays, the R ² value was 0.999. For convenience, the viral RNA copy numbers were transformed to their logarithm (to the base 10) values, hereafter termed “log values”. Appropriate dilutions of viral RNAs, which were extracted from cell culture grown pools of CAstV11672 and ANV-1, were used as positive control samples in all rRT-PCR experiments. A 10^–3 dilution of CAstV11672 RNA extract produced a C_T value of 29 (±0.5), while a 10^–5 dilution of the ANV-1 RNA extract gave a C_T value of 29.5 (±0.5). When the TaqMan^® Exo IPC assay was applied to RNA extracts from 20 randomly selected gut content samples, no PCR inhibition was observed for any of the samples.

Assay specificity

The CAstV rRT-PCR test was positive for all five CAstV isolates, with a maximal log value of 8.23 being obtained with RNA extracted from an undiluted sample of CAstV612 that had been grown in chorioallantoic membranes of embryonated chicken eggs. The CAstV assay was negative when applied to ANV-1, DHV-2 and DHV-3 virus samples. The ANV rRT-PCR assay was negative when applied to the five CAstV isolates and the DHV-2 and DHV-3 samples, but was positive when applied to RNA extracted from an ANV-1 cell culture pool. An amplicon of CAstV and one of ANV were cloned and sequenced using the M13 forward and reverse primers and were found to have sequences specific to their respective viruses.

Detection of ANV and CAstV RNAs in diagnostic samples from broiler flocks

Both assays were also assessed using RNAs extracted from a panel of 36 field samples that originated in the UK and the USA (Table 2). These comprised samples prepared from gut contents (n=29), kidneys (n=3) and cloacal swabs (n=4). The majority (27 of 29) of the gut content samples came from broiler flocks with enteritis and/or growth retardation problems. For the CAstV assay, 35/36 field samples were positive with log values ranging from 2.66 to 7.44. Eighteen out of 27 (67%) gut content samples from growth-retarded broilers were considered to have high (>5.99) log values, while two samples (7%) had comparatively low (<4.00) log values. One of the two samples from a healthy pedigree flock was negative (VF06-02/3) and the other had a low log value of 3.14 (VF06-02/1). The CAstV RNA log values for the three kidney samples ranged from 2.75 to 4.87 and were less than their counterpart gut content samples (range 6.04 to 6.46). None of the swab samples, two of which were from broilers with “wet litter”, had high log values.

Table 2.  Real-time RT-PCR detection of CAstV and ANV RNAs in field samples.

Sample	Age (days)	CAstV log value	ANV log value	Sample	Age (days)	CAstV log value	ANV log value
VF04-01/2	U	7.15	5.79	VF07-13/9^k	17	2.75	4.17
VF04-01/6	U	5.13	6.26	VF08-05/8^s	U	5.72	6.57
VF0401/11	U	7.44	5.69	VF08-05/9^s	U	3.00	5.13
VF05-01/3	6	7.41	8.23	VF08-05/21^s	U	3.00	4.38
VF0501/14	10	6.72	5.88	VF08-05/24^s	U	2.66	4.38
VF06-01/2	13	4.51	7.60	VF08-07/2	10 to 17	6.45	6.02
VF06-01/3	11	6.42	7.53	799 MO/2005	7	6.70	7.55
VF06-02/1	25	3.14	6.16	802 AR/2005	7	3.07	7.59
VF06-02/3	39	Negative	3.22	812 DE/2005	10	6.05	7.84
VF06-02/9	42	4.06	Negative	836 NC/2005	8	6.41	6.91
VF06-07/1	10	6.59	7.50	840 AR/2005	5	6.56	7.49
VF07-04/1	U	2.91	5.39	866 GA/2006	14	4.60	7.08
VF07-04/2	U	4.94	NT	883 MO/2006	7	6.10	7.52
VF07-13/1	14	6.29	7.64	916 CA/2006	12	5.16	6.14
VF07-13/1^k	14	4.87	4.39	1254 GA/2008	7	7.08	8.80
VF07-13/7	14	6.46	7.63	1255 GA/2008	4	6.45	8.42
VF07-13/7^k	14	4.17	5.67	1335 GA/2009	9	6.25	7.92
VF07-13/9	17	6.04	7.02	1340 GA/2009	9	4.75	6.86
Log values relate to the virus RNA copy numbers expressed as logarithmic values (to the base 10). U, age of the bird sampled is unknown. All samples are from gut contents, except: ^kfrom kidney; and ^sfrom cloacal swabs.

The ANV assay detected the presence of ANV in 34 out of 35 samples tested, with a similar broad log value range (3.22 to 8.80) being observed. The majority of samples (23/34; 68%) were considered to have high (>5.99) RNA log values, including three samples with log values >7.99. Of the 26 gut content samples tested from growth-retarded broilers 21 (81%) had high log values. Although one (VF06-02/9) of the 26 samples tested negative, none of the four remaining positive samples had low (<4.00) log values. The ANV RNA log values for the three kidney samples ranged from 4.17 to 5.67, and these were less than the values (range 7.02 to 7.63) obtained for gut content samples collected from the same birds.

Detection of ANV and CAstV RNAs in longitudinal survey samples of broiler flocks

In the longitudinal surveys of two broiler flocks (Flocks A and B), the gut content and kidney samples from approximately 12 birds, collected between days 0 to 42, were tested for CAstV and ANV using the rRT-PCR tests. Results obtained with the day 0 samples showed that ANV and CAstV RNAs were detected in very few chickens and resulted in very low mean log values. At most times after day 0, CAstV and ANV RNAs were detected in all 12 birds, or in the majority of gut content and kidney samples collected (Table 3 ).

Table 3.  Real-time RT-PCR detection of ANV and CAstV RNAs in gut content and kidney samples collected in longitudinal surveys of Flocks A and B.

	Day 0	Day 5	Day 7	Day 14	Day 21	Day 28	Day 35	SEM	P value
Flock A
CAstV gut	0.00 (1)	4.44^AB (12)	3.76^AC (12)	2.61^D (11)	5.24^B (12)	3.48^CD (12)	4.19^AC (12)	0.323	<0.001
CAstV kidney	0.00 (0)	4.38 (12)	2.79^A (9)	1.90^AB (9)	2.69^A (12)	0.97^B (6)	2.49^A (11)	0.390	<0.001
ANV gut	0.00 (1)	4.67^AB (12)	7.27^C (12)	7.02^CD (12)	6.43^D (12)	5.21^A (12)	4.35^B (12)	0.244	<0.001
ANV kidney	0.00 (2)	1.02 (4)	3.62^A (12)	4.01^A (12)	4.00^A (12)	2.82 (10)	1.94 (3)	0.267	<0.001
Flock B
CAstV gut	0.00 (0)	5.00^A (12)	3.98^B (12)	4.73^A (12)	3.87^BC (12)	3.77^BC (12)	3.26^C (12)	0.221	<0.001
CAstV kidney	0.00 (0)	3.94^A (12)	3.02^AB (11)	2.08^BC (9)	1.50^C (7)	1.75^C (8)	1.44^C (8)	0.385	<0.001
ANV gut	0.00 (1)	8.09^A (12)	8.00^A (12)	6.84^B (12)	6.26^B (12)	5.08^C (12)	5.16^C (12)	0.211	<0.001
ANV kidney	0.00 (0)	4.71^AB (12)	4.94^A (12)	4.31^AB (12)	3.89^B (12)	2.10^C (8)	1.41^C (6)	0.320	<0.001
Virus RNA copy numbers are shown as their logarithm values (to the base 10). Data in parentheses are the number of samples (out of 12 tested) that were positive for virus RNA. The mean values were calculated using the values of all 12 samples. Within a row, means with a common uppercase superscript letter are not significantly different. The day 0 values were not considered in the one-way analysis of variance. SEM, standard error of the mean.

In both surveys, the ANV RNA levels detected in the gut content samples were substantially greater than those obtained with CAstV. For example, log value ranges of 4.35 to 7.27 (Flock A; mean of time points 5.82) and of 5.08 to 8.09 (Flock B; mean of time points 6.57) were obtained with ANV, whereas log value ranges of 2.61 to 5.24 (Flock A; mean of time points 3.95) and 3.26 to 5.00 (Flock B; mean for time points 4.05) were obtained for CAstV. The ANV RNA levels were also generally higher in the kidney samples, although the differences between the ANV and CAstV RNA levels were smaller than those observed with the gut content samples (Table 3). It was noted that the levels of ANV RNA in the gut contents and kidneys were considerably greater at early time points (days 7 and 14) than those at later times (days 28 and 35). Thus, in Flock A, the ANV RNA log values for the gut content samples at days 28 and 35 were 5.21 and 4.35, respectively, whereas those at days 7 and 14 were 7.27 and 7.02, respectively. In Flock B, the ANV log values for the kidney samples were 2.10 and 1.41 at days 28 and 35, respectively, whereas values of 4.94 and 4.31 were obtained for kidney samples at days 7 and 14, respectively. The contrast between early and late time points was less evident with the CAstV RNA levels, particularly regarding those detected for gut content samples. For example, in Flock A, CAstV RNA log values of 3.48 and 4.19 were observed for the gut content samples at days 28 and 35, while values of 3.76 and 2.61 were observed for the day 7 and day 14 samples. Comparatively high CAstV RNA levels were detected in the gut content (log value 5.24) and kidney (log value 2.69) samples collected from Flock A at day 21. These values were significantly greater than those detected at day 14. Similar increases were not observed when the ANV RNA levels at the day 14 and day 21 samples were compared.

Although the astrovirus RNA levels observed at the early (days 5 and 7) time points were generally higher than those observed when older birds were investigated, the ANV RNA levels observed in the Flock A gut content (log value 4.67) and kidney (log value 1.02) samples at day 5 were notable exceptions, being substantially less than those obtained with Flock A samples at day 7 (gut content: 7.27; kidney: 3.62) and at day 14 (gut content: 7.02; kidney: 4.01), and those obtained with day 5 samples from Flock B (gut content: 8.08; kidney: 4.71). The variation in ANV and CAstV RNA levels present in gut content and kidney samples at these early time points was further investigated by testing day 4/5 and day 7 samples from two additional broiler flocks (Table 4). With Flocks C and D, samples were collected at day 4 and not at day 5, as was the case for Flocks A and B. For the purposes of the present study, the results obtained with the four flocks were compared at the day 4/5 and the day 7 time points. With regard to the CAstV RNA levels in the gut content and kidney, no significant differences were observed between the four flocks at day 4/5 and day 7. In contrast, significant differences were observed when the four flocks were compared in terms of their ANV RNA levels detected at these time points in both the gut content and kidney samples. Thus, at day 4/5 the ANV RNA levels in samples from Flock A (gut content: 4.67, kidney: 1.02) and Flock C (gut content: 4.69; kidney: 2.92) were significantly lower than the levels detected in Flock B (gut content: 8.09; kidney: 4.71) and Flock D (gut content: 7.69; kidney: 5.12). In addition, the Flock A day 5 ANV RNA level detected in the kidney was significantly lower than that detected in the day 4 kidney sample from Flock C.

Table 4.  Real-time RT-PCR detection of ANV and CAstV RNAs in gut content and kidney samples collected at early time points from four broiler flocks with different performance values.

	Flock A	Flock B	Flock C	Flock D	SEM	P value
Day 4/5
CAstV gut	4.44	5.00	5.02	5.06	0.193	0.091
CAstV kidney	4.38	3.94	4.07	4.65	0.283	0.295
ANV gut	4.67^A	8.09^B	4.69^A	7.69^B	0.289	<0.001
ANV kidney	1.02	4.71^A	2.92	5.12^A	0.383	<0.001
Day 7
CAstV gut	3.76	3.98	3.94	4.66	0.261/0.239^a	0.082
CAstV kidney	2.79	3.02	2.55	3.35	0.565/0.515^a	0.748
ANV gut	7.27	8.00^A	8.48	7.89^A	0.172/0.157^a	<0.001
ANV kidney	3.62^A	4.94^B	4.69^B	3.98^A	0.226/0.206^a	<0.001
EPEF^b	327 (male)	308 (male)	315 (male)	238 (female)
Virus RNA copy numbers are shown as their logarithm values (to the base 10). Within a row, means with a common uppercase superscript letter are not significantly different. ^aStandard error of the mean (SEM) presented for minimum/maximum replication as the number of birds from each flock differs. The mean values were calculated by using the values for all 12 samples. ^bEPEF=[live weight (kg) x liveability (%) x 100]/[age at depletion (days) x feed conversion rate].

In contrast to the large differences observed between flocks at day 4/5, the ANV RNA levels detected in the day 7 gut contents (log value range: 7.27 to 8.48) and kidney (log value range: 3.62 to 4.94) samples were much closer in value, although some differences were considered significant (Table 4). For example, the Flock A day 7 ANV RNA levels in kidney were significantly lower than those detected in corresponding samples from Flocks C and D, and the Flock A day 7 ANV RNA level in gut content was significantly less than that detected in the corresponding sample from Flock C. The EPEF values obtained for the three male broiler flocks were 327 (Flock A), 308 (Flock B) and 315 (Flock C), while an EPEF value of 238 was estimated for Flock D, the only female broiler flock investigated.

Detection of ANV and CAstV RNAs in experimental infection samples

One-day-old broiler chicks were infected orally with a clarified crude inoculum, prepared from pooled gut content samples that were collected at days 4 and 7 from Flock D. At 14 days p.i., the weights of the inoculated birds (n=20) were 21.4% less than those of the control, non-infected birds. Application of the ANV and CAstV rRT-PCR tests to samples that were collected from groups of five experimentally infected chickens at different days p.i. showed that ANV RNA was detected in 30/30 (100%) gut content samples and in 25/30 (83%) kidney samples that were collected up to day 28 p.i. (Table 5). ANV RNA levels were high (log values: 6.85 to 8.07) in the gut content samples collected up to day 14 p.i., while substantially reduced virus RNA levels were detected at days 21 (log value 4.84) and 28 (log value 4.05). A similar trend was observed with the kidney samples, although the ANV RNA levels were markedly less (by 2 to 3 log values) at most times p.i. In terms of both the numbers of positive samples and the amounts of virus RNA detected, the levels of CAstV RNA detected were substantially less than those of ANV RNA. Although most gut content samples were positive up to day 14 p.i., with mean log values ranging from 3.04 to 4.64, 4/5 and 5/5 birds were CAstV-negative at days 21 and 28, respectively. Overall, 20/30 (67%) gut content samples and 9/30 (30%) kidney samples were positive for CAstV RNA throughout the sampling period.

Table 5.  Real-time RT-PCR detection of ANV and CAstV RNAs in gut content and kidney samples collected from experimentally infected broiler chickens (infected at day 0) at selected times post infection.

Virus	Sample	Day 3	Day 7	Day 10	Day 14	Day 21	Day 28	SEM	P value
CAstV	Gut	3.04^AB (4)^2	4.23^AC (5)	4.17^AC (5)	4.64^C (5)	1.55^B (1)	0.00 (0)	0.515	<0.001
CAstV	Kidney	0.87 (2)	0.00 (0)	1.30 (2)	1.75 (3)	1.04 (2)	0.00 (0)	0.589	0.247
ANV	Gut	8.07^A (5)	7.80^A (5)	7.68^AB (5)	6.85^B (5)	4.84^C (5)	4.05^C (5)	0.321	<0.001
ANV	Kidney	5.36^A (5)	5.0^A (5)	4.6^A (5)	4.0^A (5)	1.09^B (2)	1.61^B (3)	0.500	<0.001
Virus RNA copy numbers are shown as their logarithm values (to the base 10). Within a row, means with a common uppercase superscript letter are not significantly different. Data in parentheses are the numbers of samples (out of five tested) that were positive for virus RNA. The mean values were calculated using the values of all five samples.

Discussion

Conventional, non-quantitative RT-PCR tests have been previously described for CAstV (Pantin-Jackwood et al., 2006; Smyth et al., 2009) and ANV (Mandoki et al., 2006; Day et al., 2007; Todd et al., 2010). Earlier work in our laboratory with conventional tests demonstrated that the vast majority of broiler chickens become infected with these astroviruses and led us to speculate that many of the infections may be subclinical. If such was the case, highly sensitive RT-PCR tests will be of limited use for differentiating non-affected flocks from flocks in which CAstV and ANV infections may be causing growth depression problems. We further speculated that, as with other virus infections, the severities of the pathogenic effects caused by ANV and CAstV may correlate with the levels of virus replication and that virus RNA levels as determined by rRT-PCR tests might prove useful indicators of clinical importance. This paper describes the first developments of TaqMan^®-based rRT-PCR assays for ANV and CAstV and their application for the quantitation of virus RNAs in gut content and kidney samples collected from broiler flocks with growth problems, in longitudinal survey samples and in samples from experimentally infected chickens. The TaqMan^® primer and probe sets used in both assays were located in conserved regions of their respective astrovirus genomes. These were identified following comparison of partial sequences of at least 20 viruses of each astrovirus species (Smyth et al., 2009; Todd et al., 2010). Our work has shown that the coding regions of CAstV and ANV show considerable sequence diversity, which makes it difficult to identify conserved regions of the required length (50 to 100 nucleotides) with which to design TaqMan^® primers and probe. With ANV, a region of approximately 56 nucleotides from within the 3′ UTR has been exploited, whereas the CAstV assay uses a highly conserved 70-nucleotide region, located at the 3′ end of ORF1b, which incorporates the 24-nucleotide intergenic sequence between ORF1b and ORF2. It is possible that this CAstV region displays high-level sequence conservation because it encompasses or overlaps with the promoter region, which is used to produce the subgenomic RNA encoding the CAstV capsid protein gene. The lack of positive RT-PCR signals obtained when heterologous avian astroviruses were tested confirmed the specificities of both rRT-PCR assays. Application of the Exo IPC RT-PCR assay to RNAs that had been extracted from 20 randomly selected gut content samples demonstrated that PCR inhibitors were not commonly found in the extracts used in the present investigation. Das et al. (2009) have recently reported the occurrence of PCR inhibitors in 18.4% of cloacal swab samples investigated for the presence of avian influenza virus. Although the effect of PCR inhibition is likely to be more critical when the levels of virus RNA being detected are very low, which was not the case in our study, the use of alternative extraction methods should be considered for future use with gut content samples. For convenience, the mean C_T values obtained with each test sample RNA were converted to viral copy numbers using standard curves prepared from 10-fold serial dilutions of in-vitro-transcribed RNAs, and these copy number values were presented as logarithm (to the base 10) values (log values).

The majority of field samples from growth-retarded broiler flocks contained high levels (log values >5.99) of ANV and CAstV RNAs (Table 2). However, it would be premature to suggest that high astrovirus levels were responsible for the clinical effects observed since other enteric viruses, including reoviruses and rotaviruses, and bacteria, which might have been present in these samples, may also have been contributing to causing growth retardations. In the absence of additional results from normal broiler flocks, it is difficult to assess the clinical significance of the high virus RNA levels that we have observed in samples from clinically-affected broiler flocks. Nevertheless, the results obtained with samples from longitudinal surveys of two broiler flocks have provided some indication about the range of astrovirus RNA levels that can be detected in broiler birds throughout their lifetimes. The main findings from the two longitudinal surveys were that: ANV and CAstV RNA were detected in the gut contents and kidneys of most of the sampled birds and at all time points; the levels of ANV RNA were generally higher than those of CAstV RNA; and the ANV RNA levels within gut content and kidney samples were substantially (2 to 3 log values) higher at early times (e.g. day 7) than those detected at late time points (e.g. days 28, 35), whereas much less variation with age was observed with CAstV RNA levels. At this stage in our investigations, it is unknown whether the detection profiles that we have observed have arisen due to single, long-lasting infections or multiple infections with different ANV and CAstV types. Broiler flocks are known to become sequentially infected with a number of serologically different rotaviruses (McNulty et al., 1984b) and this may also be the case with ANV, which is known to have at least two serotypes (Shirai et al., 1992), and CAstV, which is also known to comprise antigenically different variants (Todd et al., 2009a). The large increase in the Flock A CAstV RNA levels from day 14 to day 21 may be due to the introduction and spread of a different CAstV within this flock (Table 3). Sequence determinations of RT-PCR-generated capsid gene fragments produced at early and late time points would resolve whether flocks are infected with single or multiple ANV and CAstV types.

The most notable difference between the four longitudinally-surveyed broiler flocks related to the ANV RNA levels detected at days 4 and 5 in gut content and kidney samples (Table 2). When the ANV RNA levels of the day 4/5 samples were compared, significant differences were observed between the better-performing flocks (Flocks A and C) and Flocks B and D, which performed less well as indicated by their EPEF values . The ANV RNA levels detected in the day 7 samples were more closely related, although the smaller differences observed between the levels of the best-performing male flock (Flock A) and worst-performing male flock (Flock C) were significant. These findings indicated that, while all four flocks had high ANV levels at day 7, higher proportions of birds in the poorly-performing flocks (Flocks C and D) had higher virus loads at the early day 4/5 time points. On this basis, we can speculate that very young (1 to 4 days) chicks will be more severely affected by substantial ANV challenges than older (7 days and older) chicks. Earlier studies have shown that experimental ANV infections of specific pathogen free chicks were more pathogenic following infection at 1 day old than infection at 14 days old (Imada et al., 1981). In the field situation, the appearance of clinical signs such as diarrhoea, depressed condition and even growth arrest as early as 6 to 7 days in flocks that go on to develop severe stunting problems, suggest that the chicks have received a severe infectious challenge before this time. With the possibility that in the best-performing flocks the astrovirus RNA levels will remain low at day 7 and beyond, investigations of additional broiler flocks, including flocks with optimal performances (EPEF values >350), are required to determine whether broiler performance can be correlated with early-time-point astrovirus RNA levels. It is noteworthy that none of the CAstV RNA levels detected in the longitudinal survey were considered high (>5.99 log values), which was the case for 67% of the samples obtained from growth-retarded broilers (Table 2). This finding and the absence of significant differences between the day 4/5 and day 7 CAstV RNA levels of the four broiler flocks suggested that CAstV infections may not have been contributing to the differences in broiler performance observed with these flocks. However, although the CAstV infections detected in the surveyed flocks may have been less severe in terms of virus RNA levels detected, the possibility that these infections were contributing to the below-average performances of the four flocks cannot be ruled out.

The rRT-PCR tests for ANV and CAstV also proved useful for monitoring virus replication in experimentally infected chickens. In the reported experiment, 1-day-old broiler chicks that were inoculated with a crude inoculum prepared from day 4 and day 7 gut content longitudinal survey samples showed a 21% weight reduction at 14 days p.i. in comparison with uninfected controls. It must be remembered that this crude inoculum may contain bacteria and other enteric viruses, which may have contributed to the growth retardation observed in addition to any possible effects caused by ANV and CAstV. The ANV RNA detection profile with time p.i. obtained in this experimental infection was very similar to those obtained in the longitudinal surveys, with substantial RNA levels being obtained for gut content (7.00 to 8.00 log value) and kidney (>4.00 log values) samples respectively up to day 14 p.i., followed by substantially reduced values at days 21 and 28. However, considerably lower levels of CAstV RNA were detected, especially in the kidney, in comparison with the CAstV RNA levels seen in the longitudinal surveys. One possible explanation is that the experimentally infected broiler chickens may have had maternal antibodies to the CAstV type present in the inoculum. Although maternal antibodies may not prevent replication within the intestine, they may limit spread of the virus to the kidney, thereby accounting for the low levels of RNA detected in the kidney samples. The rRT-PCR tests for ANV and CAstV will be of use in pathogenesis studies based on experimental disease reproduction, with the ability to conveniently measure virus loads in wide tissue ranges being considered advantageous in comparison with existing methods.

In summary, we have described the development of sensitive rRT-PCR tests for detecting ANV and CAstV, and have demonstrated how these tests can be used to detect differences in the virus RNA levels present in field, survey and experimental samples. It is now planned to use these tests to investigate the roles of ANV and CAstV in causing growth-related, performance problems in broiler chickens.

Acknowledgements

The present work was funded in part by the Biotechnology and Biological Sciences Research Council, Department of the Environment, Food and Rural Affairs, and the Department of Agriculture and Rural Development for Northern Ireland.