Materials and Methods

Sample population—Serum samples were obtained during a 6-month period in 1999 and 2000 from cattle in 9 states (Alaska, Hawaii, Idaho, Maryland, Montana, Nevada, Oregon, Utah, and Washington) and submitted to the WADDL in Pullman, Wash. Samples were assigned an accession number at the time of arrival at the WADDL. Samples were subjected to various diagnostic and screening procedures, and remaining serum was frozen and placed in storage for 6 months.

Identifying information (except for the species indication) was removed from these samples, and a new identification number was assigned to ensure investigators would not be aware of the source of each sample. Serum samples were then shipped to the Laboratory for Calicivirus Studies at Oregon State University for testing.

After vesivirus results were determined, the data for each sample were used to construct an analytic database by use of a computer spreadsheet.^b Information included in the database for each submitted sample was telephone area code, postal zip code, city, and state for each premises; sex, age, and breed of each animal; herd type (dairy or beef); and reasons for initial sample submission, which included abortion, respiratory tract disease, and a number of other less common conditions. Some sera were from sick cattle, whereas others were from healthy cattle (typically obtained for surveillance purposes [eg, status for BLV or presale health assessment]).

Laboratory accession numbering at the WADDL revealed that some samples came from cattle in the same herd. For the purposes of analysis, we defined a herd as being a group of >4 samples with a common accession number prefix and successive accession number suffixes. The database also included the diagnostic tests requested and whether testing yielded positive or negative results for that sample. Diagnostic tests conducted at the WADDL included detection of antibodies against BLV (276 samples), Mycobacterium avium subsp paratuberculosis (50), Neospora caninum (223), bluetongue virus (24), BVDV (388), anaplasma (15), bovine respiratory syncytial virus (163), BHV 1 (341), parainfluenza virus type 3 (158), Leptospira spp (206), and Brucella abortus (216). Sera were also submitted, but less frequently, to be tested for detection of antigen for N caninum (7 samples), BVDV (122), bovine respiratory syncytial virus (2), BHV 1 (4), Mannheimia haemolytica (1), Hemophilus somnus (30), Staphylococcus epidermidis (1), and coronavirus (1).

Vesivirus serologic screening—A recombinant vesivirus antigen was used in an indirect ELISA for antibody detection. This ELISA was developed from an 882-bp SMSV5 capsid gene clone selected for expressing broadly cross-reactive antigen against multiple (>30) vesivirus-neutralizing antibody serotypes.^c A commercial laboratory^d processed the clone to produce a fusion peptide in an Escherichia coli expression vector, which was purified as a 293-amino acid peptide (ie, D3A).^1,9,a Stock antigen was then adjusted to a concentration of 1.2 mg/mL.

The D3A antigen was used in the ELISA at a final concentration of 1 μg/mL. Fifty microliters of D3A antigen in coating buffer was incubated for 2 hours at 37°C in 96-well, flat-bottomed ELISA plates.^e Wells were then washed twice with TBS solution, and nonspecific binding was blocked by use of 3 washes (200 μL/well) with blocking buffer,^f which was followed by the addition of 100 μL of test serum diluted 1:100 in neutral buffer (bovine serum albumin in TBST) and incubation for 2 hours at 37°C. Wells were washed 6 times with TBST, and 100 μL of monoclonal anti-bovine IgG conjugated to alkaline phosphatase^g diluted 1:40,000 in neutral buffer was added to each well; wells were then incubated for 2 hours at 37°C. Warm chromogenic substrate^h was added (100 μL/well), and plates were incubated in the dark for 3 hours at 37°C.

The OD for each sample was determined at 620 nm by use of an ELISA plate reader.ⁱ Net OD values were calculated as the values for the OD of the sample tested against 3DA antigen minus the OD of control wells treated the same in all aspects except for the addition of test serum. A cutoff value for net OD of 0.290 for the ELISA was established after calculating the mean ± 2SD (0.104 ± [2 × 0.092]) value for the OD results of 41 sera obtained from juvenile cattle (<18 months old) in the study population.¹² These sera were selected for comparison because the cattle were young, healthy, and part of a field survey of BLV. Specificity of the D3A antigen for detecting antibody against diverse serotypes of vesivirus was spot-checked by use of hyperimmune neutralizing typing antisera for 14 vesivirus serotypes. Positive and negative control samples were included for each assay.

Statistical analysis—Because the age distribution or prevalence of antivesivirus antibody in cattle was unknown before the start of the study reported here, we analyzed net OD values for the group without prior selection. Samples were scored as positive or negative on the basis of whether the net OD value exceeded the cutoff value of 0.290. Overall seroprevalence for all sera collected was determined from the fraction scored positive. Data were then compared by use of a statistical program^j to determine the distribution of positive values on the basis of specific attributes (eg, age) by use of the Fisher exact test, Yates corrected χ² test, or Mantel χ² test for trend as appropriate. Those variables that had a significant difference in proportion for these univariate analyses then were subjected to logistic regression to test independence of statistical associations with a positive ELISA result in accordance with the techniques of Hosmer and Lemeshow.¹³ Finally, to estimate whether various groups had higher titers against vesivirus, we analyzed the net OD value of samples with ELISA-positive results on the basis of group by use of an ANOVA or the Mann-Whitney test when the variances differed significantly for the Bartlett test. For each of these comparisons, a value of P < 0.05 was considered significant.

Results

Evaluation of the ELISA—A randomly selected OD-negative serum sample was used as a negative control sample for each of the 30 ELISA plates (mean net OD of negative control sample, 0.074; SD, 0.021; range, 0.052 to 0.095; mean ± SE, 28%). The positive control sample for each plate was a randomly selected serum sample that yielded a positive result (mean net OD of positive control sample, 1.159; SD, 0.155; range, 0.906 to 1.359; mean ± SE, 13%). Specificity of the D3A antigen for detecting antibody against diverse serotypes of vesivirus was spot-checked by use of hyperimmune neutralizing typing antisera for 14 vesivirus serotypes from diverse origins (each serotype was not neutralized by 100 antibody units of any other serotype), all of which had strong positive reactions against the D3A antigen (Table 1).

Overall results and univariate comparisons—The ELISA results were determined for the 693 bovine sera (Table 2). Overall, 105 (15.2%) of the samples tested had values higher than the cutoff point. Sera were received from 9 US states, and all 693 samples had a state of origin recorded. Sera from cattle in Utah (n = 1), Maryland (4), Montana (10), and Nevada (5) were excluded from further geographic analysis because of a small number of samples. Mean seroprevalence for vesivirus was 15% (103/673) in the cattle in the remaining states (Alaska, 4/14 [39%]; Hawaii, 4/25 [16%]; Washington, 73/467 [16%]; Oregon, 9/66 [14%]; and Idaho, 13/101 [13%]). There was no significant difference in seroprevalence among states, telephone area codes, and postal zip codes. In other univariate comparisons, older age (P = 0.009), reason for testing (P = 0.007), and herd type (P = 0.007) differed significantly among seropositive and seronegative groups. No significant difference was observed for seropositive status on the basis of sex (628 samples), breed (294), whether samples were part of a herd submission (441), whether samples were submitted separately (ie, 4 or fewer samples from a source; 252), and herd size. Positive ELISA results were significantly (P = 0.045; Fisher exact test) associated with positive results when samples were tested to detect antibodies against M avium subsp paratuberculosis in that 2 of 2 samples with positive results for antibodies against M avium subsp paratuberculosis also had positive results for antibodies against vesivirus, whereas 9 of 48 samples with negative results for antibodies against M avium subsp paratuberculosis had positive results for antibodies against vesivirus. This likely represented a type I error. Positive results for antibodies against vesivirus were not significantly associated with positive results for any of the other serologic tests conducted at the WADDL.

Assessment of independent factors associated with seropositive results for vesivirus—Factors found to be significantly associated with antibody against vesivirus were assessed for independence by use of logistic regression. For this analysis, data were reclassified. Three age groups were classified as 0 (0 to 6 months), 1 (7 to 12 months), and 2 (13 months and older). Because seroprevalence did not differ significantly between samples submitted for testing for respiratory tract disease or abortion (ie, reason for testing), these groups were consolidated for logistic regression and classified as 0, whereas all other reasons for testing were classified as 1. Samples from beef herds were classified as 0, whereas samples from dairy herds were classified as 1. Logistic regression revealed that age (P = 0.01) and reason for testing (P = 0.025) were independently associated with a positive test result for antibodies against vesivirus, whereas the type of herd (ie, beef or dairy) was not independently associated.

Differences in seroprevalence by herd—A total of 441 samples were categorized as herd samples, and herds were in distinct geographic locations. Herds with high and low seroprevalence were identified (Figure 1). Although cattle from herds with abortion and respiratory tract disease had a high seroprevalence (up to 33% seropositive), no significant difference existed among herds. In contrast, herds with low and high prevalence within the group categorized on the basis of other reasons for testing had significant differences. One such herd from Washington contained dairy cattle that had been screened only for antibodies against BLV; 11 of 28 (39%) were seropositive for BLV, but none were seropositive for vesivirus. Another herd of adult dairy cows from the same location in Washington yielded 42 sera that were tested for antibodies against BLV; 12 of 42 (29%) were seropositive for BLV, and 15 of 42 (36%) also were seropositive for vesivirus. The seroprevalence for vesivirus differed significantly (P = 0.001) between these 2 herds.

Number of cattle seropositive (black bars) or seronegative (white bars) for vesivirus as detected in samples submitted to the WADDL. Samples were submitted for routine testing or monitoring (control herds) or for abortion, respiratory tract disease, or reasons other than abortion or respiratory tract disease. Each bar on the x-axis represents a herd; a herd was considered a laboratory submission of serum samples from >4 cattle at the same location.

Citation: American Journal of Veterinary Research 67, 1; 10.2460/ajvr.67.1.114

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Number of cattle seropositive (black bars) or seronegative (white bars) for vesivirus as detected in samples submitted to the WADDL. Samples were submitted for routine testing or monitoring (control herds) or for abortion, respiratory tract disease, or reasons other than abortion or respiratory tract disease. Each bar on the x-axis represents a herd; a herd was considered a laboratory submission of serum samples from >4 cattle at the same location.

Citation: American Journal of Veterinary Research 67, 1; 10.2460/ajvr.67.1.114

• Download Figure
• Download figure as PowerPoint slide

Number of cattle seropositive (black bars) or seronegative (white bars) for vesivirus as detected in samples submitted to the WADDL. Samples were submitted for routine testing or monitoring (control herds) or for abortion, respiratory tract disease, or reasons other than abortion or respiratory tract disease. Each bar on the x-axis represents a herd; a herd was considered a laboratory submission of serum samples from >4 cattle at the same location.

Citation: American Journal of Veterinary Research 67, 1; 10.2460/ajvr.67.1.114

• Download Figure
• Download figure as PowerPoint slide

A herd of male beef cattle from Alaska that were screened for export to Canada was represented by 7 sera that were seronegative for BVDV and bluetongue virus, but 3 of 7 (43%) were seropositive for vesivirus. A third herd of adult dairy cows from Washington (at a location distant from the 2 aforementioned dairy herds in Washington) was represented by 5 sera that were screened for antibodies against BVDV, BLV, and M avium subsp paratuberculosis, and all 5 were uniformly seronegative for all 3 agents; however, 4 of 5 (80%) samples were seropositive for vesivirus, which was a significant (P = 0.024; 1-sided Fisher exact test) seroprevalence when compared with seroprevalence for the other 3 pathogens. The seroprevalence for vesivirus with these latter 2 herds (from Alaska and Washington) differed significantly (P = 0.004 and P < 0.001, respectively), compared with the seroprevalence for vesivirus in the aforementioned 2 dairy herds that were in Washington.

Estimates of differences in antibody titer for seropositive cattle among groups—Samples with positive results when tested for vesivirus were identified, and the net OD for those samples was analyzed on the basis of various factors. In these analyses, the net OD values of the groups whose samples were submitted for respiratory tract disease or other reasons for testing were statistically comparable and consolidated into a single group. Net OD values of the samples submitted because of abortion were significantly (P = 0.015; Mann-Whitney test) higher (mean, 0.911; SD, 0.739; 26 samples) than among the group of samples submitted for respiratory tract disease or other reasons (mean, 0.596; SD, 0.472; 79 samples). Although not significantly different, older cattle (>13 months old) that were seropositive for vesivirus had a higher mean net OD value (0.852), compared with the net OD value for the 2 seropositive younger cattle (7 to 12 months old; mean net OD, 0.330).

Discussion

The study reported here involved testing of a large collection of sera from a diagnostic laboratory to estimate the point prevalence for antibodies against vesivirus, which was 15.2% for the entire collection. We explored available data to assess differences in seroprevalence by subgroup and found increasing seroprevalence with older age, for samples submitted for reasons other than abortion or respiratory tract disease, and in dairy herds. Logistic regression revealed that older age and samples submitted for reasons other than abortion or respiratory tract disease remained independently associated with higher seroprevalence for vesivirus. Seroprevalence for vesivirus differed among herds and also between vesivirus and other pathogens.

Cattle were selected for this test population because this livestock species is appropriately characterized, is affected by diseases with a high percentage of unknown causes, typically has extensive animal data available, and is an important food source. Furthermore, several reports^1,2,6,9 exist of isolation of vesivirus from cattle. The 693 bovine sera tested represent a small subset of the total population of approximately 9 million cattle in the 9 states represented in the collection that ranged geographically from 7 of the 48 contiguous states as well as Alaska and Hawaii.¹⁴ If the true seroprevalence in the larger population was also 15%, then 1.4 million cattle in these 9 states would have been infected by vesivirus. Laboratory-based surveillance is notable for bias, so interpretation of these observations is limited.

The overall seroprevalence for vesivirus in cattle ranged between 10% and 20%. This range was similar to that observed for other common animal pathogens, such as bluetongue virus, M avium subsp paratuberculosis, N caninum, and various species or serovars of Leptospira organisms when there is no history of vaccination. The seroprevalence for B abortus was even less.

Serologic data often are used to associate pathogens with illness. For example, in follow-up testing of paired serum samples obtained from 9 cows from a herd with abortions in Washington, 8 had positive results for N caninum but only 1 of 9 had an increasing antibody titer for N caninum, which was the official diagnosis for the abortion problem. The 1 cow that had negative results for N caninum was the only one that had a positive response for vesivirus (OD, 1.032). Because acute serum for this same cow was no longer available, the timing of seroconversion to vesivirus could not be determined or associated with the time of abortion. Another example of possible vesivirus-induced abortion involved 3 cows from Washington that had aborted 1, 4, and 14 days before sera were submitted. The second and third cows had positive results when tested for BVDV, BHV 1, and N caninum, whereas the first cow was seropositive for only BVDV and BHV 1. As a follow-up, convalescent sera were submitted 18 days after the acute sera, and testing revealed no substantial change in antibody titer. The final diagnosis for cause of abortion was N caninum infection for the second and third cows. All serum samples except the acute sera of the first cow were tested for antibody against vesivirus. Although the convalescent serum sample of the first cow was negative for vesivirus (OD, 0.041) and no change in antibody titer was detected for the second cow (OD, 0.039 and 0.033 for the acute and convalescent serum samples, respectively), the third cow had a >3-fold increase in OD (0.543 on day 14 after abortion to 1.895 on day 32 after abortion), clearly indicating evidence of a recent vesivirus infection at approximately the time of abortion. One cow from Idaho with idiopathic abortion had negative results when tested for B abortus, Leptospira organisms, and BHV 1 and a low positive result when tested for BVDV, but this cow had a high seropositive value (OD, 1.656) for vesivirus. Finally, a 6-year-old cow from Washington with an idiopathic late-term abortion of twins was seronegative for N caninum, B abortus, and Leptospira organisms and weakly positive for BVDV and BHV 1 (most likely attributable to vaccination) but had a high seropositive value (OD, 1.920) for antibodies against vesivirus. These findings in combination with results for experimentally induced vesivirus disease in cattle,² several case series,^1,2,9 and the widespread prevalence of antibodies against vesivirus in cattle⁶ provide additional evidence that vesiviruses are pathogenic for and can be abortigenic in cattle.

The test antigen D3A is recognized by most of the hyperimmune typing sera used to serotype vesivirus strains. Because it is a group antigen, the testing reported here would not differentiate vesivirus variants that differ in infection outcome (Table 1). Furthermore, we cannot draw any inference from these data about the number of vesivirus types that circulated in the study population; however, other point-prevalence studies that used neutralizing antibody tests have revealed cattle with type-specific neutralizing antibodies against 13 of the 14 serotypes tested,² and 4 serotypes have been isolated from cattle with signs of illness.^1,2,9

The significant difference among herds, some of which were represented by a large number of serum samples, suggests that exposure to vesivirus is not uniform. Exploration of risk factors important for exposure of swine or mink, such as contaminated food supplements or fish,^11,15 may be a justifiable starting point for exploring these observed differences. An alternative and more direct exploratory approach would be to evaluate cattle for vesivirus when they have clinical syndromes that have been associated with vesivirus infections in case studies,^{1,2,7–9,11,k} such as hepatitis, abortion, diarrhea, encephalitis, vesicular lesions, hemorrhagic disease, and myocarditis. Approximately 75% of all aborting cattle in the United States remain without a definitive etiologic diagnosis,^16,17 which suggests the need to examine causal relationships for new and emerging abortigenic pathogens such as vesivirus.^2,8,9 The ELISA-positive samples submitted because of abortion had a higher net OD than samples submitted for respiratory tract disease or other reasons, and in some of the aforementioned cases, the laboratory evidence supporting a diagnosis of abortion attributable to vesivirus was more robust than the laboratory evidence supporting established abortigenic agents that were reported as the etiologic diagnosis for the abortions. The lack of statistical association of seroprevalence for vesivirus with seroprevalence for other pathogens assessed at the WADDL suggests that the routes of transmission of the various pathogens are distinct.

The laboratory-based study reported here provides evidence of widespread vesivirus infections in cattle across a large area of the United States. The clinical, zoonotic, and other implications of this finding in a major food animal species warrants further investigation.