Kyoung-Jin Yoon 1, Bruce H. Janke, Rick W. Swalla, Gene Erickson

J Vet Diagn Invest 16:197 21 (24) Comparison of a commercial H1N1 enzyme-linked immunosorbent assay and hemagglutination inhibition test in detecting serum antibody against swine influenza viruses Kyoung-Jin Yoon 1, Bruce H. Janke, Rick W. Swalla, Gene Erickson Abstract. Recently a commercial enzyme-linked immunosorbent assay (ELISA) kit for detecting antibody against H1N1 swine influenza virus (SIV) has been made available to diagnosticians and veterinary practitioners. Because the hemagglutination inhibition (HI) test has been considered the standard test for SIV serology, diagnostic performance of the new ELISA was evaluated using positive (n 6) and negative (n ) serum samples from young pigs with known status of SIV infection and compared with that of the HI test. Both ELISA and HI test identified all negative animals correctly. None of the serum samples (n 64) from pigs inoculated with H3N2 SIV was positive by ELISA for SIV antibody. The H1N1 SIV antibody detectable by ELISA appears to develop more slowly in comparison with antibody detectable by HI test. Although antibody was detected by HI test in all inoculated animals (n 2) by day 7 postinoculation (PI), antibody was detected by ELISA in %, 75%, and 1% of the inoculated animals on days 7, 14, and 28 PI, respectively. Discrepancy in test results between the 2 serologic tests appeared to be because of differences in antibody isotypes detected by each test. Enzyme-linked immunosorbent assay mainly detected IgG antibody, whereas the HI test detects IgM antibody very efficiently as well as IgG antibody. Collectively, the commercial ELISA is highly specific for antibody to H1N1 SIV but may not identify positive animals at the early stage of infection as effectively as the HI test, particularly when SIV is introduced to a naïve swine population. Swine influenza, commonly known as swine flu, has become one of the economically significant respiratory diseases in pigs throughout the world because it was initially recognized in the early 19s. 5,6,11,16 The disease is caused mainly by influenza A viruses, which are enveloped RNA viruses with 8-segmented, singlestranded, negative-sense RNA molecules. 8 Although detection of swine influenza virus (SIV) or viral antigen in the lung or nasal secretions from clinically affected animals is considered as the definitive diagnosis of swine influenza, serologic testing is often used to detect animals that have been exposed to SIV because the disease has a very short course and the causative agent becomes undetectable in nasal secretions or lungs relatively quickly. 9,17 Serology is also used to assess immune status of pigs at various stages within an operation so that the level of herd immunity or timing of vaccination can be determined. Several serologic assays have been used for detecting antibody against SIV: hemagglutination inhibition (HI) test, serum virus neutralization test, and indirect From the Veterinary Diagnostic Laboratory, Iowa State University, Ames, IA 511 (Yoon, Janke), Murphy Family Farms, Algona, IA 5511 (Swalla), and Rollins Animal Disease Diagnostic Laboratory, Raleigh, NC 2765 (Erickson). 1 Corresponding Author: Kyoung-Jin Yoon, Associate Professor and Head of Virology, Veterinary Diagnostic Laboratory, College of Veterinary Medicine, Iowa State University, 16 South 16th Street, Ames, IA 511. fluorescent antibody test. 8,17 Among these assays, HI test has been used most commonly in veterinary diagnostic laboratories to detect anti-influenza virus antibody and is considered to be the standard test for international trade of animals by Office International des Épizooties. 12 The HI test is designed to detect antibody specific for hemagglutinin (H), which is 1 of 2 major envelope proteins on the surface of SIV and highly immunogenic. Because 15 different H subtypes are known to exist among influenza A viruses, 8 the HI test must be tailored for each subtype by using reference strains corresponding to individual subtypes in assays for measuring subtype-specific antibody. At present, H1 and H3 subtypes are of worldwide concern in the swine industry, 1,2,13 although pigs are known to be susceptible to all H subtypes. 12 The HI test is a relatively inexpensive serologic assay. In general, the presence of HI antibody in animals is indicative of protection against the subtype used in the test, and HI antibody titers appear to correlate with the level of protection. However, the labor intensiveness of the HI test is a major hindrance to its use on a large scale. Recently, a commercial enzyme-linked immunosorbent assay (ELISA) has been developed specifically for detecting antibody against SIV of H1N1 subtype. The following study was conducted to assess the diagnostic performance of the ELISA in comparison with the HI test using a set of serum samples from animals with known status of swine influenza. 197

198 Yoon et al. Materials and methods Serum samples. Known positive serum samples (n 6) used in the study were collected from a group of 2 young crossbred pigs inoculated experimentally with 1 of 2 H1N1 SIV field isolates through nebulization of 2 ml of homogenate of lung collected from gnotobiotic pigs infected with the virus. The inoculum was determined to contain SIV at a titer of 1 1 1 EID (egg infectious dose) 5 /.2 ml. One isolate was antigenically similar to classical H1N1 SIV (A/ Sw/ID/88) on the basis of its reactivity to a panel of monoclonal antibodies specific for H protein of H1N1 SIV 14 and designated A/Sw/IA/4776/92. Using the same technique, the other isolate was considered to be atypical in comparison with the classical H1N1 SIV and designated A/Sw/NE/969/ 92. After inoculation, infection was confirmed by monitoring changes in rectal temperature, respiratory clinical signs (nasal discharge, sneezing, coughing), pathological evaluation of selected animals, and virus isolation or detection in lungs. Sera were collected from all animals on days 7, 14, and 28 postinoculation (PI). In addition to sera positive for antibody against H1N1 SIV, 64 serum samples were obtained through sequential bleeding from 24 pigs exposed to 1 of 2 H3N2 SIV isolates designated A/Sw/IA/4135/98 and A/Sw/NC/ 35922/98, 18 respectively, at a rate of 1 1 7 EID 5 /.2 ml. Known negative serum samples (n 56) were collected through sequential bleeding from 4 sham-inoculated control pigs in the study described above and 1 additional pigs used in another animal trial as sham-inoculated controls. These animals were sampled, 7 or 8, 14, and 21 or 28 days after inoculation with either virus-free cell culture medium or egg fluid. Additional known negative sera (n 132) were collected from 132 pigs known to be free of SIV infection and used for the study. Enzyme-linked immunosorbent assay. Enzyme-linked immunosorbent assay a was performed according to the protocol recommended by the manufacturer. Briefly, samples to be tested were diluted 1:5 in the diluent provided in the kit. One-hundred microliters of each control or diluted sample were transferred to a well and incubated for 3 min. After incubation, the controls and samples were discarded from wells, and plates were washed using a programmable automated microtiter plate washer. One-hundred microliters of peroxidase-labeled anti-igg secondary antibody, which was also provided in the kit, were dispensed to each well and incubated for another 3 min. After rinsing, the presence of antigen antibody complexes was observed by adding 1 l of substrate to each well. Optical density (OD) of each well was measured using a microtitration plate reader at 45 nm. OD values were then converted into sample-to-positive (S/ P) ratios using a formula provided by the manufacturer. Samples with S/P ratio equal to or greater than.4 were considered to be positive for antibody against H1N1 SIV. All procedures were done at ambient temperature (2 25 C). All samples were run in duplicate. The commercial ELISA kit was also used to assess SIVspecific IgM antibody response. The test was performed in the identical manner as described above with 2 exceptions. One was that peroxidase-labeled goat anti-swine IgM antibody was used instead of anti-swine IgG antibody conjugate Table 1. Diagnostic performance of commercial H1N1 SIV ELISA and HI test for detecting virus-specific antibody in young pigs with known status of infection. Test Infected Uninfected Total HI ( ) HI ( ) 6* 6 ELISA ( ) ELISA ( ) 35 25 35 213 Total 6 248 * Samples showing HI activity at 1:1 or higher dilutions were determined to be positive. Samples with S/P ratio of.4 or higher were determined to be positive according to the manufacturer s recommendation. provided in the kit. The other was that OD values at 65 nm were used to determine the presence of SIV-specific IgM in samples instead of calculating S/P ratio because no reference serum for IgM antibody was available. Samples with OD value higher than mean OD value of negative samples plus 3 SD were considered to be positive for IgM antibody against H1N1 SIV. Hemagglutination inhibition test. Hemagglutination inhibition tests were performed using standard protocol established in our laboratory which is based on the procedure recommended by Centers for Disease Control and Prevention. 3 In brief, sera were heat inactivated at 56 C for 3 min and treated with an equal amount of 1% chicken red blood cells (RBC) for 3 min at ambient temperature before conducting HI tests for antibody to SIV of H1 subtype. To assess antibodies to H3, SIV samples were pretreated with receptor destroying enzyme (RED) b at 37 C for 12 18 hr after heat inactivation. Each sample was then serially 2-fold diluted in.1 M of phosphate buffered saline (ph 7.4) and tested for its HI activity against 8 hemagglutination (HA) units of H1 (A/Sw/IA/73) c or H3 (A/Sw/4135/IA/98) reference strain of SIV. The presence of HI antibody in each sample against H1 or H3 SIV was observed by adding.5% rooster RBC and turkey RBC, respectively, and incubating for 45 min at ambient temperature. HI antibody titer of each sample was determined as the reciprocal of highest dilution in which no hemagglutination was observed. Samples with no HI activity at 1:1 dilution were considered negative. Results All known negative samples (n ) were negative by both commercial ELISA (S/P ratio.4) and HI test ( 1:1); therefore, the apparent diagnostic specificities of both ELISA and HI test were 1% (Table 1). As illustrated in Fig. 1, 97.3% of the negative samples had S/P ratio less than.2, suggesting that cutoff S/P ratio of.4 for positive that was set by the manufacturer would give distinct separation between positive and negative animals. All serum samples (n 64) collected at different points of time from pigs inoculated with H3N2 SIV isolates were also negative by the ELISA, demonstrating no cross-reaction of antigens in ELISA plates with

Test performance of commercial H1N1 SIV ELISA 199 Figure 1. Distribution of S/P ratios in ELISA for antibody specific for SIV among known negative animals. Samples with S/P ratio less than.4 are considered to be negative for antibody against H1N1 SIV. Each sample was run in duplicate. antibody to H3N2 SIV. The 2 H3N2 viruses (i.e., IA isolate vs. NC isolate) that induced the antibody in these sera were antigenically variants. As shown in Fig. 2, these sera were positive by HI test using homologous H3N2 SIV isolate but exhibited no or little cross reactivity with the heterologous isolate, demonstrating that results of HI test for the same subtype can be affected by the virus strain used in the test. Antibody responses of animals to H1N1 SIV infection over time as determined by ELISA and HI test are summarized in Fig. 3. Using S/P ratio of.4 as cutoff for positive, none of 2 animals had antibody detectable by ELISA on day 7 PI, regardless of challenge virus, although average S/P ratio of inoculated pigs was significantly higher than that of negative control animals. Some of inoculated animals started to seroconvert to SIV on day 14 PI, and by 28 days PI all inoculated animals had seroconverted to H1N1 SIV by ELISA (S/P ratio.4). Overall the diagnostic sensitivity of ELISA was estimated to be 58% (Table 1). In comparison, all animals already had HI antibody to both H1N1 SIV isolates on day 7 PI, ranging from 1: 32 to 1:64 and were still seropositive at termination of the study, although HI antibody levels of exposed animals tended to gradually decline after 7 days PI, resulting in 1% diagnostic sensitivity (Table 1). As illustrated in Fig. 4, IgM antibody specific for H1N1 SIV developed initially in inoculated animals in correlation with development of HI antibody to H1N1 SIV. After 14 days PI, no significant level of IgM antibody was detected in the animals. Discussion The present study demonstrated that the new H1N1 SIV ELISA kit possesses an excellent diagnostic spec- Figure 2. Antibody responses of young pigs to H3N2 SIV infection as determined by 2-way HI tests. Circles ( ) and triangles ( ) represent pigs inoculated with North Carolina isolate (A/Sw/ NC/35922/98) and Iowa isolate (A/Sw/IA/4135/98), respectively. Solid ( ) and open ( ) symbols represent HI antibody titers against homologous and heterologous viruses, respectively. Symbols are mean of observations and bars standard error of mean at a given day PI. Figure 3. Antibody responses of young pigs to H1N1 SIV infection as determined by ELISA (top panel) and HI test (bottom panel). Samples with S/P ratio.4 are considered to be positive for antibody against H1N1 SIV. Each sample was run in duplicate. Symbols are mean of observations and bars standard error of mean at a given day PI.

2 Yoon et al. Figure 4. Response of IgM antibody specific for H1N1 SIV as determined by ELISA. Samples with OD value.31 (mean OD value of negative samples 3 SD) were considered to be positive for IgM antibody against H1N1 SIV. Symbols are mean of observations and bars standard error of mean at a given day PI. ificity, i.e., no false positives (Table 1). The test was also shown to be highly specific for H1N1 SIV, diminishing a concern about false positive results due to cross-reactivity with SIV of other subtypes because of common internal antigens. However, it raises the need for a separate ELISA test for H3N2 SIV because both subtypes are known to cocirculate in swine populations. 4 Or it would be beneficial to develop a serologic assay that can detect antibody to SIV regardless of subtype of the virus affecting pigs in a herd. In contrast to its excellent specificity, the new commercial SIV ELISA was not as sensitive as expected for an ELISA-based test. In this study, among 6 sera from animals exposed to 1 of 2 H1N1 SIV isolates, 35 samples were positive by ELISA for SIV-specific antibodies, whereas all 6 samples were positive by HI test. From this observation, diagnostic sensitivity of the commercial H1N1 SIV ELISA kit was estimated to be 58% for pigs infected recently with SIV (Table 1). Most, if not all, of false negative results occurred in early stage of infection (Fig. 3). However, if cutoff S/P ratio had been set at.2, most of inoculated animals were considered seropositive for SIV at 7 days PI. It remains yet to be determined whether cutoff S/ P ratio should be adjusted when using a large number of field samples. Collectively, the commercial SIV ELISA is highly specific for antibody to H1N1 SIV but may not identify positive animals at the early stage of infection as effectively as the HI test, particularly when the virus is introduced to a naïve swine population. In a primary viral infection, the first antibody to appear in the blood circulation is of the IgM class, and it disappears shortly after the IgG antibodies are developed. 15 The commercial ELISA used in this study was designed to primarily detect IgG specific for H1N1 SIV. In comparison, the HI assays can detect both IgM and IgG antibodies to the virus. Such a difference in the isotype of antibody to be detected by 2 tests may explain discrepancies in results obtained with the 2 assays on samples collected early after initial exposure. Results of IgM assay (Fig. 4) in the present study certainly supported this explanation, suggesting that an IgM-based assay may be useful in detecting animals exposed recently to SIV. Otherwise, incorporation of secondary antibody specific for both porcine IgM and IgG antibodies might be an alternative to increase the sensitivity of the SIV ELISA. Although the sensitivity of SIV ELISA may need improvement, ELISA-based serologic assays are known to have several merits over HI test, such as its ability to deal with a large number of samples at a time, a good turnaround time, less labor intensiveness, and higher analytic sensitivity. In addition, the commercial SIV ELISA appeared to detect seroconversion to H1N1 SIV much better than HI test when pigs (finishers and adult females) with immunological memory for H1N1 SIV from previous infection or vaccination were reexposed to H1N1 SIV (Jackson T, et al.: 22, Monitoring serologic responses in finishing pigs to swine influenza vaccine using hemagglutination inhibition and ELISA techniques. Proc Annu Meet Am Assoc Vet Lab Diagn, p 43). 7 In conclusion, the commercial SIV ELISA can be a useful tool for serodiagnosis of H1N1 SIV infection; however, the ELISA should be used with caution because the present study demonstrated that the test could not detect antibody against SIV of subtypes other than H1N1 and missed animals exposed recently. Acknowledgements The authors thank Brenda Schwarz, Raquel Hansel, and Serology staff in the Veterinary Diagnostic Laboratory at Iowa State University for technical assistance in virus preparation and laboratory testing. The authors acknowledge Drs. J. David Schneider and Caryn Paulin for assistance in animal work while they were enrolled in the College of Veterinary Medicine at Iowa State University. The study was supported in part by funding from Iowa Livestock Health Advisory Council and Iowa Healthy Livestock Initiative Research Grant. Sources and manufacturers a. HerdCheck Swine Influenza Virus H1N1 Antibody Test Kit, IDEXX Laboratories, Inc., Westbrook, ME. b. Bio Whittaker, Walkersville, MD. c. National Veterinary Services Laboratories, USDA/APHIS, Ames, IA. References 1. Brown IH: 22, Influenza A viruses in pigs in Europe. In: Trends in emerging viral infections of swine, ed. Morilla A,

Test performance of commercial H1N1 SIV ELISA 21 Yoon K-J, Zimmerman JJ, pp. 29 36. Iowa State University Press, Ames, IA. 2. Brown IH: 2, The epidemiology and evolution of influenza viruses in pigs. Vet Microbiol 74:29 46. 3. Centers for Disease Control: 1982, Concepts and procedures for laboratory-based influenza surveillance. Centers for Disease Control, U.S. Department of Health and Human Services, Washington, DC. 4. Choi YK, Goyal SM, Joo HS: 22, Prevalence of swine influenza virus subtypes on swine farms in the United States. Arch Virol 14:129 122. 5. Chun J: 1919, Influenza including its infection among pigs. Nat Med J China 5:34 44. 6. Dorset M, McBryde CN, Niles WB: 1922, Remarks on Hog Flu. J Am Vet Med Assoc 62:162 171. 7. Erickson G: 22, Transitioning from HI to ELISA when evaluating herd status and time of vaccination for H1N1 swine influenza. Focus on swine health and performance, vol. 5. no. 1. Schering-Plough Animal Health Corporation, Union, NJ. 8. Esterday BC, Van Reeth K: 1999, Swine influenza. In: Diseases of swine, ed. Straw BE, D Allaire S, Mengeling WL, Taylor DJ, pp. 277 29. Iowa State University Press, Ames, IA. 9. Janke BH: 2, Diagnosis of swine influenza. Swine Health Prod 8:79 84. 1. Kida H, Ito T, Yasuda J, et al.: 1994, Potential for transmission of avian influenza viruses to pigs. J Gen Virol 75:2183 2. 11. Koen JS: 1919, A practical method for field diagnosis of swine diseases. Am J Vet Med 14:468 47. 12. OIE Standards Commission: 2, Manual of standards for diagnostic tests and vaccines, 4th ed. Office International des Épizooties, Paris, France. 13. Olsen CW: 22. Emergence of novel strains of swine influenza virus in North America. In: Trends in emerging viral infections of swine, ed. Morilla A, Yoon K-J, Zimmerman JJ, pp. 37 43. Iowa State University Press, Ames, IA. 14. Olsen CW, McGregor MW, Cooley AJ, et al.: 1993, Antigenic and genetic analysis of a recently isolated H1N1 swine influenza virus. Am J Vet Res 54:163 1636. 15. Rothbarth PH, Groen J, Bohnen AM, et al.: 1999, Influenza virus serology a comparative study. J Virol Methods 78:163 169. 16. Shope RE: 1931, Swine influenza. Filtration experiments and etiology. J Exp Med 54:373 385. 17. Yoon K-J, Janke BH: 22, Swine influenza virus: evolution, epidemiology and diagnosis. In: Trends in emerging viral infections of swine, ed. Morilla A, Yoon K-J, Zimmerman JJ, pp. 23 28. Iowa State University Press, Ames, IA. 18. Zhou NN, Senne DA, Landgraf JS, et al.: 1999, Genetic reassortment of avian, swine, and human influenza a viruses in American pigs. J Virol 73:8851 8856.