PERFORMANCE OF ASSAYS FOR TESTING ANTIBODIES AGAINST PORCINE REPRODUCTIVE AND RESPIRATORY SYNDROME VIRUS IN SERA COLLECTED FROM SWINE FARMS IN A REGION WITH AN EXTREME VIRUS HETEROGENEITY

Porcine reproductive and respiratory syndrome is the most economically important viral disease in the swine industry worldwide. Porcine reproductive and respiratory syndrome virus (PRRSV) strains are classifi ed into two distinct genotypes, the European genotype and the North American genotype [1]. The European PRRSV genotype has been divided into three subtypes: a pan-European subtype 1 (the prototype is the Lelystad virus, LV) and East European subtypes 2 and 3 [2]. Recently, a previously unknown East European subtype 3 PRRSV strain (Lena) from Belarus has been isolated and its


INTRODUCTION
Porcine reproductive and respiratory syndrome is the most economically important viral disease in the swine industry worldwide.Porcine reproductive and respiratory syndrome virus (PRRSV) strains are classifi ed into two distinct genotypes, the European genotype and the North American genotype [1].The European PRRSV genotype has been divided into three subtypes: a pan-European subtype 1 (the prototype is the Lelystad virus, LV) and East European subtypes 2 and 3 [2].Recently, a previously unknown East European subtype 3 PRRSV strain (Lena) from Belarus has been isolated and its genetic, antigenic and pathogenic properties have been described [3].An immunoperoxidase monolayer assay (IPMA) and an enzyme-linked immunosorbent assay (ELISA) are commonly used to detect PRRSV-specifi c antibodies.It is considered that the IPMA is one of the most reliable serological tools [4].In addition, IDEXX PRRS 2XR Antibody Test (IDEXX 2XR ELISA) (IDEXX Laboratories, Westbrook, ME, USA), the enzyme-linked immunosorbent assay, became the standard for monitoring the serological status of swine herds [5].At present, IDEXX 2XR ELISA is the only test used in Belarus and some other East European countries to detect antibodies against PRRSV.An enzyme-linked immunosorbent assays, IDEXX PRRS X3 Antibody Test (IDEXX 3X ELISA) (IDEXX Laboratories, Westbrook, ME, USA) and INGEZIM PRRS Universal kit (INGEZIM ELISA) (INGENASA, Madrid, Spain) are also available on the market.All these ELISA kits are indirect assays based on viral nucleocapsid antigens belonging to the pan-European subtype 1 and North American PRRSV strains.An antigenic heterogeneity between pan-European subtype 1 and North American PRRSV strains has been described earlier [6][7][8].Some level of antigenic heterogeneity between East European subtype 3 and pan-European subtype 1 strains has also been reported [3].The aim of the present study was to evaluate several serological assays to test antibodies against PRRSV in serum samples collected from swine farms in a region with a high degree of PRRSV heterogeneity.Belarus became the country of choice for sample collection because different heterologous PRRSV strains of the all known European sub-types circulate in this region [2,9].Serum samples were tested applying IPMAs based on PRRSV Lena, LV and VR-2332 strains as antigens and commercial IDEXX 2XR ELISA, IDEXX 3X ELISA and INGEZIM ELISA.

MATERIALS AND METHODS
Experimental serum samples were collected in a previous study (from 9 pigs after oronasal inoculation with 10 6 TCID 50 /pig of PRRSV Lena, [3]) and were tested applying commercial ELISAs.Afterwards, 269 fi eld samples from Belarus were tested in IPMAs and commercial ELISAs.Field samples from sows were derived from 15 Belarusian farms with a history of respiratory and reproductive disorders.Vaccination against PRRSV was never applied on those farms.Porcine reproductive and respiratory syndrome virus-specifi c antibodies were detected with the IPMA on MARC-145 cells infected with PRRSV as previously described [10].The reference pan-European subtype 1 PRRSV LV [4], East European subtype 3 PRRSV Lena and reference North American PRRSV VR-2332 [11] were used as IPMA antigens.All serum samples were tested in duplicate.The commercial IDEXX 2XR ELISA was used to detect PRRSV antibodies in experimental antisera and fi eld samples according to the manual.To test sera in the IDEXX 3X ELISA and INGEZIM ELISA, samples were sent to IDEXX (IDEXX Livestock and Poultry Diagnostics, Uppsala, Sweden) and INGEZIME (Inmunología y Genética Aplicada S.A., Madrid, Spain), respectively.The sensitivity and specifi city of serological diagnostic tests are based on a "gold standard".There is currently no recognized gold standard for the detection of PRRSV antibodies.In the present study, the IPMA status of samples was used as the stan-dard against which commercial ELISAs were evaluated.The selection of the diagnostic standard was based on several reasons.First, previous results obtained in the authors' study clearly demonstrate the high sensitivity and specifi city of IPMAs [3].All serum samples collected from 27 PRRSV-inoculated pigs were positive for virus-specifi c antibodies at 7-10 days post-inoculation.In contrast, all samples collected before or at three days post-inoculation were negative.Second, the IPMA allowed to use three different PRRSV strains as antigens (LV: pan-European subtype 1 PRRSV; Lena: East European subtype 3 PRRSV; and VR-2332: North American PRRSV) while all commercial ELI-SAs are based on only two PRRSV strains (pan-European subtype 1 PRRSV and North American PRRSV).Finally, the use of a primary serum dilution 1/10 in a dilution buffer for IPMA tests (in the present study serial sample dilutions 1/10, 1/40 and 1/160 were used) hypothetically allows a more sensitive detection of antibodies than ELISAs do.Test sample dilutions used for the IDEXX 2XR ELISA, IDEXX 3X ELISA and IN-GEZIM ELISA are 1/40, 1/40 and 1/100, respectively.Serum samples were regarded as IPMA-positive (IPMA + ) when it showed a specifi c antigen staining (Figure 1) in at least one of IPMAs (LV, Lena, VR-3223).Upon testing fi eld sera, the sensitivity and specifi city of the ELISAs were calculated as follows: sensitivity = [number of IPMA + sera] / [number of IPMA + sera + number of ELISA x false negative sera]; specifi city = [number of IPMA -sera] / [number of IPMA -sera + number of ELISA x false positive sera].) and incubated for 18h (37°C, 5% CO 2 ).Then, the culture medium was removed, and cells were washed in PBS and dried at 37°C for 1h.The plates were kept at -70°C until use.Plates were thawed and then fi xed in 4% paraformaldehyde for 10 min.The paraformaldehyde was removed, the cells were washed twice with PBS and a solution of 1% H 2 O 2 in methanol was added.Plates were washed twice with PBS and serial dilutions of the sera were added.Sera were incubated for 1h at 37°C.Plates were washed three times with PBS plus 1% Tween 80 and 50 μl of 1/250 rabbit anti-swine IgG HRP-conjugated antibodies (Dako) was added to each well.After incubation at 37°C for 1h, plates were washed three times and 50 μl of a substrate solution of 3-amino-9-ethylcarbazole in 0.05 M acetate buffer, pH5, with 0.05% H 2 O 2 was added to each well, and incubated at room temperature for 20 min.Then, the reaction was blocked by replacing the substrate by acetate buffer and the results were determined by examination with a microscope

RESULTS
Previously, the good sensitivity of ELISAs has been reported testing experimental pan-European subtype 1 PRRSV antisera [12,13].Therefore, prior to testing fi eld samples, experimental antisera from pigs inoculated with East European subtype 3 PRRSV Lena have been tested using three commercial ELISAs.All tests were able to determine PRRSV-specifi c antibodies during the course of infection (Figure 2).However, results differed at 7 days post-inoculation.At this time point, IDEXX 2XR ELISA was able to detect antibodies in 5 samples out of 9, IDEXX 3X ELISA in 1 sample out of 9, while INGEZIM ELISA was not able to detect antibodies.In our previous study, when same serum samples have been tested in the homologous Lena IPMA, 6 sera out of 9 were positive [3].One serum collected before inoculation of the pig gave a positive result in the IDEXX 2XR ELISA.In this sample the IPMA or IDEXX 3X ELISA and INGEZIM ELISA antibodies were not detected.The negative result of virus isolation and titration indicated that this animal had not previously been infected with PRRSV (false positive reaction).
The genetic heterogeneity of PRRSV strains in Belarus which might infl uence the performance of serological assays is well documented [2,9].In the present study, 249 fi eld serum samples out of 269 fi eld sera tested in IPMAs were positive in at least one of the assays.Out of 249 IPMA-positive fi eld samples, 15% were positive for antibodies against PRRSV LV antigens only and 4% were positive for antibodies against PRRSV Lena antigens only.There were no samples positive for antibodies against PRRSV VR-2332 antigens only.Therefore, it is possible that using LV, Lena or VR-2332 PRRSV strains alone as IPMA antigens for diagnostic purposes can provide false negative results.  1.The sensitivity of ELISAs varied from 80% to 86% (IDEXX 2XR ELISA-86%, IDEXX 3X ELISA-80% and INGEZIM ELISA-83%).A failure to detect antibodies against subtype 3 PRRSV Lena (54% of all false-negative ELISA tests were positive in IPMAs with PRRSV Lena as antigens), as well as higher serum sample dilutions (1/40 and 1/100 in ELISAs versus 1/10 in IPMAs) may both explain the lower sensitivity of commercial ELISAs.Among IPMA-negative samples, ELISA-positive sera were also detected (Table 1).The IDEXX 2XR ELISA, IDEXX 3X ELISA and INGEZIM ELISA had 6 (specifi city is 77%), 4 (specifi city is 83%) and 6 (specifi city is 77%) unexpected positive results (out of 20 IPMA -sera), respectively.

DISCUSSION
Since different heterologous PRRSV strains of the all known European subtypes circulate in Belarus, this region can be considered as a territory with an extreme PRRSV heterogeneity.Also, bacterial infections (Streptococcus suis, Escherichia coli) are endemic on many Belarussian swine farms.As a result, abortions, birth of mummifi ed, dead and weak piglets, high mortality rate before weaning, respiratory disorders and mortality in growing pigs (up to 70%) are common on swine farms (information from local veterinarians).In addition, a highly pathogenic East European subtype 3 PRRSV strain Lena (and probably other, yet unknown strains) circulates in Belarus [3].Under these circumstances, testing of pigs for antibodies against PRRSV has a particular importance.According to manufacturers, all commercial ELISAs tested in the present study are developed as a herd-screening tool.Thereunder, all swine farms included in the present study were ELISA-positive on the herd level, which coincides with the IPMA results.Consequently, all IPMAs and ELISAs provided valuable information about the farm PRRSV serological status.However, recognition of the individual animal serological status can be important during PRRSV eradication [14].Applying IPMAs with only LV or only Lena PRRSV strains as antigens, or any of ELISAs resulted in missing of some positive fi eld samples.Therefore, in the present study, none of the commonly used serological tools for PRRSV diagnosis may guarantee a fl awless detection of antibodies at the individual animal level.Taking into account the heterogeneity of recently isolated PRRSV strains in different European countries [2,9,[15][16][17][18], the problem might also be relevant in these territories.Immunoperoxidase monolayer assays which combine anti- gens of locally circulating PRRSV subtypes may provide the best performance.Some unexpected positive ELISA results were observed among fi eld sera which were negative in all IPMAs.Most probably, the results can be attributed to non-specifi c ELI-SA reactions, since the IPMAs are more sensitive upon testing experimental, as well as fi eld sera.Field and experimental reports of non-specifi c reactions of the IDEXX 2XR ELISA are known [5,19].Also in the present study, one serum collected from a negative pig prior PRRSV inoculation was positive in the IDEXX 2XR ELISA.Non-specifi c reactions of the IDEXX 3X ELISA and INGEZIM ELISA have not been reported yet.These fi ndings have practical implementations, for example, if only a small number of sera is available from a farm, or only few positives are observed among tested samples, the ELISA results need to be cautiously interpreted.In this case, the fi ndings can be confi rmed by retesting with IPMAs.
In conclusion, none of the commonly used serological tools for PRRSV diagnosis can guarantee the fl awless detection of antibodies at the individual animal level in regions with the extreme PRRSV heterogeneity.This fact should be taken into account during PRRS prevention and virus eradication, testing incoming pigs and monitoring PRRSVnegative herds in regions under a risk of being infected.Using IPMAs based on antigens of locally circulating PRRSV subtypes may provide the best performance under these circumstances.

Figure 1 .
Figure 1.IPMA staining.PRRSV antigens (brown staining) within the cytoplasm of MARC-145 cells.Cells were seeded in 96-well cell culture plates, inoculated with 50 μl of PRRSV (Lelystad, Lena or VR-2332) and incubated for 18h (37°C, 5% CO 2 ).Then, the culture medium was removed, and cells were washed in PBS and dried at 37°C for 1h.The plates were kept at -70°C until use.Plates were thawed and then fi xed in 4% paraformaldehyde for 10 min.The paraformaldehyde was removed, the cells were washed twice with PBS and a solution of 1% H 2 O 2 in methanol was added.Plates were washed twice with PBS and serial dilutions of the sera were added.Sera were incubated for 1h at 37°C.Plates were washed three times with PBS plus 1% Tween 80 and 50 μl of 1/250 rabbit anti-swine IgG HRP-conjugated antibodies (Dako) was added to each well.After incubation at 37°C for 1h, plates were washed three times and 50 μl of a substrate solution of 3-amino-9-ethylcarbazole in 0.05 M acetate buffer, pH5, with 0.05% H 2 O 2 was added to each well, and incubated at room temperature for 20 min.Then, the reaction was blocked by replacing the substrate by acetate buffer and the results were determined by examination with a microscope

Figure 2 .
Figure 2. Cross-reactivity of PRRSV Lena experimental antisera with IDEXX 2XR ELISA, IDEXX 3X ELISA and INGEZIM ELISA PRRSV antigens.Symbols represent mean titres, whiskers above and below are standard deviations.IDEXX 2XR ELISA and IDEXX 3X ELISA s/p values lower than 0.4 were considered to be negative.An INGEZIM ELISA s/p value lower than 0.35 was considered to be negative

Table 1 .
Comparison of the IPMA and ELISA results (fi eld sera)