Methods for the Detection of Viruses in Bovine Serum

JOURNAL OF CLINICAL MICROBIOLOGY, Feb. 1975, p. 212-218 Copyright 1975 American Societv for Microbiology Vol. 1, No. 2 Printed in U.S.A. Methods for the Detection of Viruses in Bovine Serum N. S. SWACK, C. K. Y. FONG, G. D. HSIUNG,* AND P. A. GROSS Department of Laboratory Medicine and Internal Medicine, Yale University School of Medicine, New Haven, Connecticut 06510, and Veterans Administration Hospital, West Haven, Connecticut 06516* Received for publication 18 November 1974 An evaluation of selected commonly used procedures for the recovery of endogenous viral contaminants in bovine serum was undertaken. Low speed centrifugation (25,000 x g) was found to be efficient for the recovery of bovine herpesvirus type 1 (BHV-1) and parainfluenza virus type 3 (PI-3) in bovine serum. Decreased infectivity titers were obtained when parainfluenza virus type 3, and to a lesser extent bovine herpesvirus type 1, were concentrated using high speed centrifugation (100,000 x g) for extended time periods. In neither case could infectious virus be recovered from serum containing sufficient titers of homologous neutralizing antibody, although electron microscopy examination revealed the presence of the viruses previously added. In the presence of homologous antibody, virus particles appeared to have a diffuse, poorly defined outer membrane. Neutralizing antibody titers to bovine herpesvirus type 1 and parainfluenza virus types were found in fetal, calf, and adult bovine sera. The prevalence and magnitude of the antibody titers to these viruses increased with the age of the animals examined. The incorporation of bovine serum as a nutritive supplement into most tissue or cell culture media has resulted in extensive investigation into the prevalence of endogenous bovine viruses. Fetal bovine serum has been implicated in the appearance of bovine virus diarrhea virus in cell cultures previously known to be virus free (4). Bovine herpesvirus type 1 in a concentration of 150 mean tissue culture infective doses (TCID50)/ml has been recovered as an endogenous contaminant in fetal bovine serum by Molander et al. (6). These investigators also isolated bovine virus diarrhea virus from bovine serum collected by cardiac puncture. Recently, a number of investigators (1, 5) have isolated bacteriophage from many lots of bovine serum. The present report is concerned with the evaluation of some of the commonly used procedures for the detection of viruses in bovine serum, together with studies of some of the conditions which might influence virus recovery. MATERIALS AND METHODS Viruses and cell cultures. The viruses used in this study were bovine herpesvirus type 1 (BHV-1), the bovine strain of parainfluenza virus type 3 (PI-3), and the simian virus strain of parainfluenza virus type 5 (PI-5). The BHV-1 and PI-3 viruses were obtained from the American Type Culture Collection and PI-5 was isolated in our laboratories from a primary rhesus monkey kidney cell culture. Primary bovine embryonic kidney cells were used for the propagation of BHV-1 and primary monkey kidney cell cultures were used for PI-3 and PI-5. The growth of the latter two viruses was determined by the hemadsorption of guinea pig erythrocytes onto infected cell culture monolayers, whereas observation of cultures for cytopathic effect was used to determine the infectivity of BHV-1 virus. Source of serum samples for antibody studies. Selected samples consisting of pooled fetal bovine or calf sera were purchased commercially. Single fetal bovine serum samples were obtained from commercial sources and calf and adult sera were drawn from individual animals in Rhode Island. All sera were inactivated at 56 C for 30 min and tested for the presence of neutralizing antibody to selected viruses, including BHV-1, PI-3 and PI-5. Antibody titers were determined by adding 100 TCID5O of virus to each serum dilution, allowing the virus-serum mixture to stand at room temperature for 30 min followed by inoculation onto susceptible cell cultures. The highest dilution of serum which completely inhibited the test dose of virus was considered to be the end point. Sedimentation of virus in serum by centrifugation. A variety of ultracentrifugation speeds and times were employed to evaluate the optimal conditions for sedimentation of known virus previously added to calf serum or fetal bovine serum. Clarification of' serum was done using a GSA rotor at 5,000 rpm for 20 min in a RC-2 Sorvall ultracentrifuge. After this, virus sedimentation was accomplished using a SS-34 rotor at 14,500 rpm (25,000 x g) in the RC-2 ultracentrifuge or a SW27 rotor at 27,000 rpm (100,000 x g) in a Beckman L2-65 ultracentrifuge. All sedimentation procedures were conducted at 4 C. 212

VOL. 1, 1975 DETECTION OF VIRUSES IN BOVINE SERUM 213 The specific time and speed used in each experiment are indicated in the results. In certain experiments, pellets obtained by high speed centrifugation were resuspended by brief sonic treatment. Recovery of known virus added to bovine serum. Various concentrations of BHV-1 or PI-3 viruses were added to bovine serum, and their recovery was determined under different conditions. In general, the virus suspension was added to a 100-ml serum sample from which 1 ml was removed for virus assay, and the remainder was divided into 33-ml aliquots and subjected to centrifugation. After careful removal of the supernatant from each tube, the pooled pellets in a volume of approximately 1 ml were considered to represent a 100 times concentration of the added virus as compared to the virus titers in serum before centrifugation. Virus recovery from serum with or without homologous neutralizing antibody was evaluated. Infectivity titrations and examination of specimens in the electron microscope were employed to compare these procedures fordetecting each virus. Detection of virus in bovine sera by virus isolation and by electron microscopy. Virus isolation studies on serum samples were performed as follows: before centrifugation 0.1 ml was inoculated into 10 to 15 culture tubes each of bovine embryonic kidney, monkey kidney (green or rhesus monkey), and rabbit kidney. Pooled serum samples were further tested by inoculating the same primary cell cultures with 0.1 ml of the 100 times concentrated pellets obtained by centrifugation as mentioned above. All inoculated cultures were kept for observation for 30 days without subculturing. At the end of the incubation period the monkey kidney cell cultures were tested for hemadsorption, and bovine embryonic kidney and rabbit kidney cell cultures were fixed and stained for microscopy examination for virus-induced inclusions. Methods for the preparation of serum samples for electron microscopy have been described in detail in the accompanying paper (2). Briefly, clarified sera were pelleted in a Beckman L2-65 ultracentrifuge at 27,000 rpm in SW27 rotor for 90 min. The pellets were fixed with 2% glutaraldehyde and then centrifuged directly into Beem capsules at 30,000 rpm in a SW65 rotor for 60 min. The pellets in Beem capsules were postfixed with osmium tetroxide and embedded in Epon 812. Thin sections were stained with uranyl acetate and lead citrate and were examined with a Philips EM 300 electron microscope. RESULTS Effect of speed and time of centrifugation on the recovery of known virus from bovine serum. The effect of centrifugation upon the recovery of known virus from serum at high and low speeds was compared. Representative experiments are shown in Table 1. A sedimentation procedure using a low speed of centrifugation (25,000 x g) appeared to be an efficient method for the recovery of both BHV- 1 and PI-3 virus. After high speed centrifugation at 100,000 x g, recovery of BHV-1 virus was not appreciably higher than that obtained at 25,000 x g. In those instances where the pellet of BHV-1 obtained from high speed centrifugation was subjected to a brief period of sonic treatment, the resultant virus titers were somewhat higher than those from the untreated sample suggesting some aggregation of virus. No significant amount of infectious virus was obtained from the supernatants. When PI-3 was used in these experiments, titers were observed to be reduced significantly during the prolonged period of high speed centrifugation. This decrease appeared to progress as the time of treatment was increased. The cause of this decrease has not been determined. When smaller amounts of BHV-1 were added to bovine serum samples at concentrations most likely to be encountered in contaminated bovine serum, virus recovery appeared to be almost complete when centrifuged at 100,000 x g for 90 min (Table 2). It was apparent that infectious virus was readily recovered at such high speed centrifugation even when the concentrations of virus in the original serum samples were less than 1 log10 TCID5,Jml of serum. Effect of antibody on recovery of known virus from bovine serum. The presence of homologous neutralizing antibody in the serum markedly reduced the efficiency of the procedure used to recover infectious virus added to the serum samples (Table 3). Serum with an antibody titer of 1:2 masked more than 10-fold BHV-1 as indicated by the decreased infectivity titers, and infectious virus was not recovered when the homologous antibody titer was 1:16. Similar results were obtained with PI-3 in TABLE 1. Recovery of BHV-1 and PI-3 from antibody-free bovine serum after high and low speed centrifugation Virus titers (log TCID,dml of pellet) Centrifugation time gravity BHV-1 PI-3 (min) Expt 1 Expt 2 Expt 1 Expt 2 25,000 x Oa 6.4 6.3 5.5 5.5 90 6.8 7.1 7.0 6.5 180 8.2 8.0 6.2 6.7 360 7.4 7.2 6.5 6.0 100,000 x 0 4.8 5.6 5.4 5.5 90 6.0 6.1 5.9 5.7 180 5.9 6.2 5.2 4.7 360 6.0 6.9 4.0 4.2 avirus titer in serum samples before centrifugation.

214 SWACK ET AL. serum samples containing homologous antibody. Infectious virus could not be recovered from serum samples with antibody titers of 1:40 and 1:160. As mentioned above, PI-3 appeared to be easily inactivated when subjected to centrifugation at 100,000 x g for 90 min since no increase in the infectivity titer was observed in the antibody negative serum sample after this concentration procedure (Table 3, bottom section). Efficiency of recognizing viruses in bovine serum by electron microscopy. Neither BHV-1 nor PI-3 could be detected by measurement of the infectivity titers when these viruses were added to serum samples which contained homologous antibody (Table 3). However, the presence of virus particles of each type was confirmed by electron microscopy examination of thin sections (Table 3, last column). Figure TABLE 2. Expt no. Recovery of small concentrations of BHV-1 by high speed centrifugation Infectivity titers of BHV-1 (log TCID,Jml) Input Pelleta 1 3.5 5.0 1.7 4.5 0.7 2.9 <0.1 2.1 2 6.5 7.2 3.4 4.2 2.5 3.1 1.5 3.5 3 4.5 6.2 3.5 5.8 0.5 2.8 a Pellets obtained after centrifugation at 100,000 x g for 90 min. TABLE 3. J. CLIN. MICROBIOL. la illustrates a mixture of BHV-1 and PI-3 virus particles in a fetal bovine serum sample. Each virus type can be identified by its morphology and can be compared with each virus alone in other bovine serum samples (Fig. lb and c and 2). Complete, enveloped, typical herpesvirus particles were easily seen in the thin sections (Fig. lb and c) When PI-3 virus was added to calf serum containing specific homologous antibody, a layer of electron dense fuzzy material, presumably antibody molecules, was seen on the viral envelopes (Fig. 2a and b). These myxovirus particles appeared to be disintegrating when compared to those seen in samples prepared in the absence of neutralizing antibody (Fig. 2c and d). Attempts to isolate endogenous viruses from bovine sera and determination of neutralizing antibody. No virus was isolated in cell cultures inoculated with 21 lots of pooled fetal bovine and eight lots of pooled calf serum obtained from commercial laboratories, and 43 serum samples from bleedings obtained from individual calves and 37 adult cattle. There was no evidence of virus-induced cytopathic effect or hemadsorption, nor were any virus-induced inclusions found in the stained preparations. Subcultures and blind passages of inoculated cultures were not done in these studies. Although virus was not isolated from any of the fetal bovine and calf serum pools which were subjected to the high speed sedimentation procedures mentioned above, virus-like particles were observed by electron microscopy as reported in the accompanying paper (2). Since no virus was isolated from the above serum samples, they were then examined for the presence of neutralizing antibody to some commonly found bovine viruses (Table 4). It was found that antibody to BHV-1 virus was prevalent in the serum pools purchased commer- Recovery of BHV-1 and PI-3 viruses from antibody-positive and -negative bovine sera after high speed centrifugation Virus titer (log TCID,Jml) Bovine serum Electron Virus neutralizing Virus microscopic antibody titer addeda Before concn After concn5 examination BHV-1 < 1:2 4.0 4.0 5.5 + 1:2 3.3 1.7 2.5 + 1:16 3.7 < 1.0c <1.0 + PI-3 < 1:5 5.5 5.6 4.2 + 1:40 2.7 <1.0 < 1.0 + 1:160 4.3 <1.0 <1.0 + a Virus titer in 100-ml serum samples. b Virus titer in 1-ml pellets after centrifugation at 100,000 x g for 90 min. c <1:10 = no infectious virus detected.

VOL. 1, 1975 DETECTION OF VIRUSES IN BOVINE SERUM 215... 1 -r.' -^.~. ps A Ap #k.' i, '*, "-4.:^ -i.c.a.9- C~KA.A Ar A*e FIG. 1 (a) Thin section of a pellet of a fetal bovine serum sample to which a mixture of BHV-1 (single arrow) and PI-3 (double arrow) viruses was added. x64,800. (b) Thin section of a pellet of a fetal bovine serum sample to which BHV-1 (arrows) virus was added. x48,000. (c) Higher magnification of the same sample as shown in (b). x100,800. cially. Of the 21 fetal bovine sera tested, two showed antibody to BHV-1 and four of the seven calf serum pools tested were positive. One of the 33 fetuses, none of the six calves and only two of the 36 adult cattle contained BHV-1 antibody. The mean titers of neutralizing antibody to BHV-1 were low (1:2 to 1:9). Of the two parainfluenza viruses tested, fetal bovine sera showed the lowest percentage of antibody-positive animals with a mean titer of 1:12 to PI-3 in the pooled samples. No antibody was detected in sera obtained from individual fetuses. Most of the sera collected from calves and adult cattle were antibody positive to PI-3 and had higher titers than those obtained from the fetal samples. The pooled calf sera had a mean antibody

a *k A.i I Aj ~.. I C...,., O:'' X X - ;-; '*~~~ 4 -r"a? ; #F ts.'n.<x' 't at,, F..:,, 3-4. A * *.'.x; s;f ~ ~ ~ ~ ~ ~ * '' * :,8. '. J A~~~~~~~~~~~I a K.;... *,- s '',s,..,,.1,-.1..i -Ii.,..: r '; '. FIG. 2 (a) Thin section of a pellet of a PI-3 antibody-positive calf serum sample to which homologous virus was added, showing fuzzy surface of PI-3 virus particles. x48,000. (b) Higher magnification of another section of the same sample as shown in (a). x 79,200. (c) Thin section of a pellet of PI-3 virus in Hanks balanced salt solution showing many PI-3 virus particles with sharply outlined envelope. x48,000. (d) Higher magnification of another section of the same sample as shown in (a). x 100,800. 216

VOL. 1, 1975 TABLE 4. DETECTION OF VIRUSES IN BOVINE SERUM Neutralizing antibody status of bovine serum samples Neutralizing antibody statusa Source of Serum BHV-1 PI-3 PI-5 serum type No. % Mean No. % Mean No. % Mean Ma positive! positive! Ma positive! Ma tosti/ Positive titer toal Positive titer total Positive titer Commercial pools Fetal 2/21 10 1:2 5/19 26 1:12 3/20 15 1:5 Calf 4/7 57 1:2 5/5 100 1:80 2/10 20 1:5 Individual animals Fetal 1/33 3 1:5 0/34 0-0/34 0 - Calf 0/6 0 4/6 67 1:34 2/6 33 1:5 Adult 2/36 6 1:9 37/37 100 1:107 23/35 66 1:6 a Antibody titer 1:2 or higher for BHV-1 and 1:5 or higher for PI-3 and PI-5 were considered to be antibody positive. Samples studied represent individual lots obtained from commercial laboratories or individual serum samples from both commercial and local sources. 217 titer of 1:80 to PI-3, whereas the four individual calf sera had a mean titer of 1:34 to this virus. The mean antibody titer to PI-3 in adult cattle was greater than 1:100 and all of the adult sera tested were positive to PI-3. Although the incidence of PI-5 antibody-positive animals followed the same pattern as PI-3, the actual neutralizing antibody titers were markedly lower than those observed for PI-3. DISCUSSION The widespread use of bovine sera as a supplement to tissue culture media makes it imperative that this product be free of contaminating viruses and other microbial agents. Through the selective use of ultracentrifugation involving variations of speed and time, optimal conditions for recovery of each type of virus varied. It was found that both BHV-1 and PI-3 were less likely to be aggregated or inactivated by centrifugation at 25,000 x g. Centrifugation at the low speed for long periods of time may enhance the efficiency of recovery of these viruses in bovine serum. While bovine viruses have been isolated from bovine serum by previous investigators (6; A. J. Kniazeff, L. J. Wopschall, and H. E. Hopps, In Vitro 9:354, 1974), the present study demonstrated that knowledge of the antibody status of the serum being tested is essential for success of the isolation procedures. The failure to isolate viral agents from the bovine sera in the present study is somewhat surprising. Kniazeff et al., using similar procedures for sedimentation of serum samples, found that 36% of 39 commercially available fetal bovine serum samples contained endogenous viruses (In Vitro 9:354, 1974). The blind passage of cells and serial subcultivation of inoculated cultures described by these investigators are apparently essential steps in the demonstration of the presence of minimal amounts of viral contaminants in bovine sera (6). In addition, some of the sera tested in the present study were found to contain significant levels of neutralizing antibody to the viruses expected including BHV-1 and PI-3, therefore the likelihood of isolating virus from these bovine serum samples became even less. The data presented here also illustrated the advantage of recognizing viruses in bovine serum by electron microscopy, especially in the presence of homologous antibody (Table 2). Whether the virus-antibody complex in bovine sera is capable of inducing any undesirable effects on cell cultures is yet to be determined. Other investigators have shown that bovine sera contain antibody to several viruses. Schell et al. (8), studying individual calves, showed that antibodies to BHV-1, PI-3, and BVD were quite prevalent in animals from widely different geographical areas. Commercially available pools of sera have also been reported to contain neutralizing antibody to these bovine viruses (7) as well as to others, such as PI-5 (3). The finding that 15 to 20% of commercial bovine sera contained neutralizing antibody to PI-5 virus might account for the reports of the low prevalence of this virus as a latent endogenous contaminant in primary cell cultures since bovine sera have been incorporated in the culture media used (9). These findings are extremely important to the virologist employing tissue cultures for virus isolations from clinical specimens. ACKNOWLEDGMENTS This work was supported by research contracts NIH- DBS-72-2105 and FDA 233-74-1035 from the Department of Health, Education and Welfare, and by Public Health Service research grant AI-08648 from the Institute of Allergy and Infectious Diseases.

218 The technical assistance of Deborah Winograd, David Potocik, Barbara Meek, and JoEllyn Bradley is greatlv appreciated. SWACK ET AL. LITERATURE CITED 1. Chu, F. C., J. B. Johnson, H. C. Orr, P. G. Probst, and J. C. Petricciani. 1973. Bacterial virus contamination of fetal bovine sera. In Vitro 9:31-34. 2. Fong. C. K. Y., P. A. Gross, G. D. Hsiung, and N. S. Swack. 1975. Use of electron microscopy for detection of viral and other microbial contaminants in bovine sera. J. Clin. Microbiol. 1:219-224. 3. Hsiung, G. D. 1963. Studies on parainfluenza viruses. Discussion on the relative role of identifiable agents in respiratory diseases. Am. Rev. Resp. Dis. 88:189-192. 4. Kniazeff, A. J., V. Rimer, and L. Gasta. 1967. y-globulin J. CLIN. MICROBIOL. in fetal bovine sera: significance in virology. Nature (London) 214:805-806. 5. Merril, C. R., T. B. Friedman, A. F. M. Attallah, M. R. Geier, K. Krell, and R. Yarkin. 1972. Isolation of bacteriophages from commercial sera. In Vitro 8:91-93. 6. Molander, C. W., A. J. Kniazeff, C. W. Boone, A. Paley, and D. T. Imagawa. 1972. Isolation and characterization of viruses from fetal calf serum. In Vitro 7:168-173. 7. Rossi, C. R., and G. K. Kiesel. 1974. Antibody to viruses affecting cattle in commercial tissue culture grade fetal calf serum. Appl. Microbiol. 27:114-117. 8. Schell, K., R. P. Sanderson, J. W. Whalen, and J. L. Bittle. 1972. The antigenicity of multivalent vaccines for bovine respiratory disease. Cornell Vet. 162:101-109. 9. Swack, N. S., and G. D. Hsiung. 1974. Endogenous agents in primary cell cultures with special reference to latent viruses. In Vitro 10:260-267.