Материал: Bovine Viral Diarrhea Virus Diagnosis, Management, and Control

(Pellerin et al., 1994; Ridpath et al., 1994). The two new BVDV 1 subtypes from southern Africa were designated 1c and 1d. The latter BVDV subtype was reported to be predominantly associated with respiratory tract disease and later shown experimentally to be able to cause primary respiratory disease (Baule et al., 2001). Vilˇcek et al. (2001) examined 76 BVDV 1 isolates from various countries and reported that these viruses could be separated into 11 phylogenetic groups. Recently, Flores et al. (2002) identified six South American strains and one North American BVDV 2 strain that cluster in a separate group from other BVDV 2. Thus, two subgenotypes of BVDV 2 (2a and 2b) have now been identified.

GENETIC RECOMBINATION AND

SPONTANEOUS AND POSTVACCINAL

MUCOSAL DISEASE

After the findings by Meyers et al. (1989) of the cellular insertion of ubiquitin-coding sequences in the NS2 gene of cytopathic Osloss strain of BVDV and another unidentified insertion (also identified later to be of cellular origin) in the same gene of the NADL strain, much of the research activity in the early and mid-1990s was devoted to the genetic analysis of cytopathic and noncytopathic virus pairs from cases of MD. This analysis revealed a number of genetic changes in cytopathic BVDV, including insertions and gene duplications and deletions, changes that resulted in the production of NS3 by creation of a cleavage site at its amino terminus (reviewed by Meyers and Thiel, 1996). In cases where viral genes were duplicated, an additional NS3 gene generated the NS3 protein. The genetic changes observed in cytopathic BVDV demonstrated that both viral-cellular and viral-viral recombination were the operating mechanisms for the generation of cytopathic BVDV from the persisting noncytopathic BVDV.

In 1995, Ridpath and Bolin (1995) characterized a noncytopathic (BVDV 2-125nc) and cytopathic (BVDV 2-125c) viral pair isolated from an animal suffering from late-onset pvMD. This animal had been vaccinated 3 months prior with a modified-live BVDV 1 (NADL strain) vaccine. Genetic sequencing of the BVDV 2 viral pair revealed that in comparison with BVDV 2-125nc, the BVDV 2-125c isolate contained a 366-nucleotide insertion in the NS2-3 gene (leading to the expression of the NS3 protein). The inserted sequence was found to have a 99% identity with sequences of the BVDV 1-NADL vaccine virus, indicating that the vaccine virus had recombined with the noncytopathic virus BVDV

2-125nc to produce the cytopathic BVDV 2-125c. The authors speculated that the prolonged time between vaccination and pvMD may have been a reflection of the time required for a single virus particle, generated by the recombination event, to replicate to a high enough titer to precipitate pvMD. The cytopathic BVDV 2-125c was one of the first isolated and characterized cytopathic BVDV 2 strains and has since been used by many laboratories as a challenge virus in serum neutralization assays for determination of neutralizing antibodies to BVDV 2.

Subsequently, Becher et al. (2001) also found, in two independent cases of pvMD, that genetic recombination between the persisting noncytopathic BVDV and a BVDV vaccine virus had occurred. In one of these cases, the cytopathic component of the virus pair was shown to have been generated by both recombination and deletion events and was actually comprised of five different cytopathic BVDV subgenomes (genomic BVDV RNA containing large internal deletions but able to express the NS3 protein).

LATE-ONSET MUCOSAL DISEASE

Several experimental studies were conducted on late-onset (or delayed-onset) MD in the midand late 1990s. Whereas in MD (or acute MD) the superinfecting cytopathic BVDV kills the persistently infected host within about 3 weeks, in late-onset MD there may be several months between challenge with the cytopathic virus and the onset of clinical signs of disease. Late-onset MD was first described by Brownlie et al. (1986), who infected six persistently infected calves with homologous cytopathic BVDV and six with heterologous cytopathic BVDV. All six animals superinfected with the homologous virus developed MD in the first 3 weeks, whereas none of the animals superinfected with a heterologous virus developed MD in this time period. However, two of the animals infected with the heterologous virus developed late-onset MD at 98 and 146 days postinoculation. Three others remained healthy until euthanized at 59–209 days postinoculation. The sixth animal appeared to be developing late-onset MD at about 80 days with signs of intermittent diarrhea over a 4-week period. Later, Westenbrink et al. (1989) also inoculated clinically healthy, persistently infected animals with three heterologous cytopathic BVDV, and MD occurred as late as 99 days postinoculation. It should be noted here that since spontaneous MD can occur in persistently infected animals that are not superinfected and are held in isolation (Brownlie and Clarke, 1993), it is possible

that some experimental cases of late-onset MD (and pvMD), could actually be cases of spontaneous MD. Thus, in experiments of this type, an analysis is needed (preferably genetic sequence analysis), beyond that of virus-neutralizing antibody responses, of the superinfecting input cytopathic BVDV, the persisting noncytopathic virus, and the output or MD-associated cytopathic BVDV.

Moennig et al. (1993) inoculated a persistently infected bull with a heterologous cytopathic BVDV (TGAC), and MD was observed 15 weeks after inoculation. In this case, the neutralizing antibody response to the cytopathic TGAC virus was delayed 31 days postinfection and neutralizing antibody titers increased to a maximum on day 114 postinoculation (when the animal was euthanized), indicating prolonged persistence of the superinfecting cytopathic virus in this animal. However, only noncytopathic virus was recovered from buffy coat cells and nasal swabs. The cytopathic BVDV recovered from fecal samples (designated cpX) at the onset of disease was analyzed and found to have the same phenotype (monoclonal antibody reactivity pattern) as the noncytopathic persisting virus but the same genotype in the NS2-3 region as the TGAC cytopathic virus as observed by restriction enzyme digestion of PCR products (Fritzemeier et al., 1995). Thus, they concluded that cpX was a phenotypically altered variant of TGAC. Subsequent nucleotide sequencing (Fritzemeier et al., 1997) indicated that cpX arose by recombination of persisting noncytopathic virus and TGAC. This finding was consistent with the finding of Ridpath and Bolin (1995) of viral-viral recombination for a case of late-onset pvMD. Fritzemeier et al. (1997) also found that a superinfecting cytopathic BVDV in experimental acute MD (MD occurring in a 2–3-week time frame after inoculation) remained genetically unchanged from inoculation to the onset of MD. They concluded that acute MD and late-onset MD may occur by two different mechanisms: In the former, it is the superinfecting virus that causes the disease; whereas in the latter, genetic recombination between the superinfecting cytopathic virus and the persisting noncytopathic virus creates a new cytopathic virus that causes the disease. Subsequently, Fray et al. (1998) inoculated a persistently infected calf with culture of an uncloned heterologous cytopathic BVDV obtained from a natural case of MD (thus also containing a noncytopathic virus) and the animal developed MD at 145 days postinoculation. Of interest in this study was the occurrence of a prolonged nasal shedding and viremia of cytopathic BVDV before the

onset of MD. No genetic analysis of viral isolates was conducted, but neutralizing antibody data indicated that at least three antigenically distinct cytopathic viruses were isolated from the calf during the period between superinfection and postmortem examination. The authors speculated that the noncytopathic virus in the heterologous challenge may have influenced the survival of cytopathic BVDV in the animal in some manner, possibly by influencing the immune response of the host.

NORMAL CALVES FROM PERSISTENTLY

INFECTED COWS

Calves born to persistently infected cows are likewise persistently infected; thus these animals are unsuitable for breeding (McClurkin et al., 1984). Wentink et al. (1991), in an attempt to preserve the genetic material of a highly valued heifer that was developmentally normal but persistently infected, transferred a fertilized embryo of the heifer to an immunocompetent recipient cow. Before transfer, the embryo was treated and washed according to routine methods of the International Embryo Transfer Society (IETS). The result was a normal heifer calf with normal immunity to BVDV. Since then, several other researchers have repeated these results and have produced, by embryo transfer, normal calves from persistently infected cows (Bak et al., 1992; Brock et al., 1997; Smith and Grimmer, 2000).

It was noted however, that there was a lack of a superovulatory response following hormone stimulation in persistently infected cows and only low numbers of viable embryos could be obtained. Wentink et al. (1991) obtained only one viable embryo out of six, and Brock et al. (1997), who used seven persistently infected donors, obtained only nine transferable embryos from two of the females after 45 individual uterine flushes. In the study of Brock et al. (1997), only one pregnancy was obtained from the transfer of six embryos into seronegative recipients. None of the recipient cows showed seroconversion to BVDV leading the authors to conclude that neither horizontal nor vertical transmission of the virus occurs when recommended IETS embryo washing procedures are followed.

ATYPICAL PERSISTENT INFECTION

In 1998, Voges et al. (1998) described for the first time a persistent infection of BVDV in an immunecompetent animal. The bull in this case was not viremic but continuously shed virus in its semen over a period of 11 months until slaughtered. The bull ap-

peared healthy and its growth rate and testicular development were unremarkable. Semen quality was considered good. Limited pregnancy data suggested that the in vivo fertility of the bull’s sperm was not compromised. The animal had a consistently high titer of serum neutralizing antibodies to the standard test virus ( 1:4096). Against the homologous virus, the animal had an extraordinarily high serum neutralizing antibody titer of >1:100,000.

The virus titer in the semen was relatively low (<103.3 TCID50/ml) when compared to semen obtained from typical, persistently infected (viremic) bulls (approximately 105–107 TCID50/ml) (Barlow et al., 1986; Meyling and Jensen, 1988; Kirkland et al., 1991). On postmortem, BVDV was isolated from testicular tissues but not from any other tissue examined. This long-term shedding of BVDV in a non-viremic bull is reminiscent of the situation in stallions infected with equine arteritis virus (Timoney and McCollum, 1993). In this infection, stallions undergo a transient viremia resulting in an immune response that eliminates the virus except for that in the genital tract. Chronic shedding of the virus in the semen occurs in many infected stallions and may persist for years, although most horses eventually clear the infection. These carrier stallions also have high serum neutralizing antibody titers.

As hypothesized by Voges et al. (1998), the bull with the atypical persistent infection was likely infected near the age of puberty, shortly before the blood-testis barrier becomes fully functional. This allowed infection of testicular tissues but excluded ensuing antibodies from the site. The high level of serum neutralizing antibodies suggests continual exposure of the immune system to the virus. This atypical persistent infection of BVDV in bulls appears to be rare, but at least one other case has been reported (van Rijn, 1999). Nevertheless, this has changed the testing requirements for bulls entering artificial insemination (AI) centers. Previously, testing for viremia was all that was required to identify typical persistently infected bulls and to restrict their entry; semen was rarely tested. Now the World Organization for Animal Health (Office International des Epizooties) requires the testing of semen from each bull before entry into an AI center.

ADVANCES IN MOLECULAR BIOLOGY

In the latter half of the 1990s, infectious cDNA clones were constructed for the type 1, cytopathic BVDV strains cp7, and NADL (Meyers et al., 1996; Vassilev et al., 1997; Mendez et al., 1998). Subsequently, 40 years after its initial discovery (Gillespie

et al., 1960), an infectious cDNA clone was constructed from the genome of the type 1, cytopathic Oregon strain (Kümmerer and Meyers, 2000). The availability of infectious cDNA clones has begun to have, and will continue to have, a large impact on our understanding of the molecular biology of BVDV as these cDNA clones are manipulated and used in studies to decipher the function of genes and genetic elements.

Meyers et al. (1996) removed an insertion of 27 nucleotides in the NS2 gene of the infectious clone of the cytopathic cp7, which is not present in the genome of cp7’s noncytopathic counterpart, ncp7 (the two viruses representing a homologous viral pair from a case of MD). In removing the insertion, a noncytopathic BVDV was recovered, proving that the insertion was responsible for the cytopathogenicity of the cp7 virus. Similarly, Mendez et al. (1998) deleted a 270-nucleotide insertion in the NS2 gene of the infectious clone of NADL and also produced a virus that was no longer cytopathic. In both cases, the altered infectious clones of cp7 and NADL failed to produce NS3, showing that these insertions led to processing at the NS2/NS3 site, production of NS3, and cytopathic effect.

Some strains of cytopathic BVDV, such as the Oregon and Singer strains, do not contain insertions or genetic rearrangements and, thus, did not arise by recombination but by another mechanism. Kümmerer et al. (1998) and Kümmerer and Meyers (2000) used chimeric cDNA constructs in a transient expression system and alterations in an infectious cDNA clone to show that processing at the NS2/ NS3 site in the Oregon strain was the result of point mutations within the NS2 protein. They noted that about 40 cytopathic pestiviruses have thus far been analyzed and the majority of these have been generated by recombination. The remaining strains have apparently arisen by point mutations generated during replication. However, since these cytopathic viruses occur less frequently than those that arise by recombination (itself a rare event) these researchers hypothesized that this second mechanism of generating cytopathic BVDV was not one of sequential accumulation of point mutations, but rather the simultaneous introduction of a set of point mutations.

Recently, an infectious cDNA clone for a virulent, noncytopathic type 2 BVDV (New York 93 strain) was constructed (Meyer et al., 2002). Infectious virus derived from the cDNA clone retained virulence when used to infect cattle. These animals developed fever, leukopenia, and clinical signs, including respiratory symptoms and gastrointestinal

disorders. Alteration of single histidine codons (at positions 300 and 349) in the sequence of the Erns gene in the infectious cDNA clone led to two new infectious clones with RNase-negative phenotypes. Virus derived from the infectious clone having a deletion of the histidine codon at position 349 was used in animal experiments and was found to have an attenuated phenotype. None of the calves infected with this mutant virus had body temperatures above 39.5°C, nor did they develop diarrhea; only mild respiratory signs were observed. Leukopenia occurred but with an early recovery of leukocyte numbers. Viremia and nasal shedding also occurred but for a shorter duration than in animals infected with virus derived from the unaltered infectious cDNA clone. As demonstrated by this study, the availability of an infectious cDNA clone in which virulent BVDV can be recovered allows new approaches to the study of BVDV-induced disease and identification of genetic markers of virulence and attenuation.

DIAGNOSIS BY VIRUS ISOLATION

An essential element of control of BVDV in herds is the detection and elimination of persistently infected calves. The classical method of detecting persistent infections is virus isolation in cell culture using serum or leukocytes as the test sample. In testing young calves, a major concern with the use of virus isolation, especially from serum, was the presence of colostral antibodies to BVDV, which could interfere with the sensitivity of the test. Palfi et al. (1993) studied the decline of BVDV colostral antibodies and the detectability of BVDV in young, persistently infected calves and found that viremia was not detectable in the serum of seven persistently infected calves with colostral antibody titers, which were as low as 1:16 to 1:24.

In a later study, only one of four persistently infected calves, which had the highest viral titer (106.5 TCID50/ml) before ingestion of colostrum, was detected as viremic after colostrum ingestion when both serum and leukocytes were tested (Brock et al., 1998). Palfi et al. (1993) found that the half-life of colostral antibodies in persistently infected calves (5 to 11 days) was much shorter than in non-persist- ently infected calves (approximately 3 weeks). Presumably, the continuous high virus production in persistently infected calves is responsible for the rapid clearance of colostral antibodies. Palfi et al. (1993) found that clearance of these antibodies in persistently infected calves may occur by 8 weeks of age, at which time viremia can be detected.

However, it has been generally recommended that, if calves under 3 months of age have been tested by virus isolation, they be retested at 3 months of age (Brock et al., 1998; Saliki et al., 1997). Dubovi (2002) recommended that, for young calves, in addition to the standard washing of leukocytes to remove colostral antibody, fresh, rather than freezethawed, leukocytes be used for virus isolation since interference, presumably by residual antibodies, may occur upon freeze-thawing.

Although persistently infected cattle generally have very high BVDV titers (104–105 TCID50/ml of serum), Brock et al. (1998) found that the level of viremia in one of seven persistently infected animals became undetectable over several test dates when serum was tested. In this apparently rare case, the animal developed neutralizing antibodies to the persisting virus although the infection was never cleared. During periods when the virus was undetectable in serum because of neutralizing antibody, virus was isolated at low concentrations from blood leukocytes.

Rae et al. (1987) examined the viability of BVDV in serum and plasma collected from persistently infected cattle. Storage of samples in the dark at room temperature (17–26°C) for up to 5 days had no significant effect on virus titer. They concluded that successful isolation of BVDV from persistently infected animals was unlikely to be compromised by a delay of up to 5 days between sample collection and testing.

MONOCLONAL ANTIBODY-BASED TESTS

With the emergence of BVDV 2, existing panels of BVDV 1 monoclonal antibodies were examined for cross-reactivity with these new isolates, and monoclonal antibodies to BVDV 2 were also produced. Ridpath et al. (1994) tested 29 E2-specific monoclonal antibodies produced against BVDV 1 strains and found that most of these failed to react with BVDV 2 isolates. However, two E2-specific monoclonal antibodies were reactive with all 15 BVDV 2 isolates tested. Deregt and Prins (1998) also determined that one BVDV 1 E2-specific monoclonal antibody was reactive to all 21 BVDV 2 isolates tested, whereas another E2-specific monoclonal antibody was determined to be BVDV 1-specific. The epitopes of these two monoclonal antibodies were mapped, the first to an immunodominant, type-common epitope (Paton et al., 1992; Deregt et al., 1998a) and the latter to a conformational epitope containing a critical amino acid deleted in BVDV 2 isolates (Deregt et al., 1998a). E2-specific monoclonal antibodies

against BVDV 2 that were reactive with neutralizing epitopes in three antigenic domains were also produced (Deregt et al., 1998b). Most of these monoclonal antibodies were unreactive or only weakly reactive with BVDV 1. One monoclonal antibody had a type-common specificity, and two others were entirely BVDV 2-specific.

A number of monoclonal antibody-based tests were developed in the 1990s. Competitive (blocking) enzyme-linked immunosorbent assays (ELISAs) for the detection of antibodies to BVDV were constructed using NS3 monoclonal antibodies (Lecomte et al., 1990; Paton et al., 1991). Paton et al. (1991) used two NS3 monoclonal antibodies reactive to 157 different pestiviruses for development of a blocking ELISA. They examined the ability of the assay to detect antibodies against pestiviruses in cattle, sheep, and swine. The relative sensitivity of the blocking ELISA compared to serum neutralization was high for bovine and ovine sera (94.7% and 99.1%, respectively) but lower for swine (76%), whereas the specificity was high in each case ( 96%).

Several capture ELISAs for detection of the NS2- 3 protein in persistently infected cattle were also developed and commercialized (Brinkhof et al., 1996). The NS3/NS2-3 proteins are highly immunogenic and the NS2-3 protein is produced in large amounts in persistently infected cattle. Brinkhof et al. (1996) evaluated four commercial capture ELISAs employing NS3 monoclonal antibodies in which blood leukocyte preparations were tested. These assays demonstrated high sensitivity (94–97%) and specificity (100%). The capture ELISA was considered to be the test of choice for eradication programs where many animals need to be screened and monitored. In the above study and in another study (Shannon et al., 1992), false negative results were obtained with capture ELISAs for calves with high levels of colostral antibodies in their blood. Thus, as for virus isolation, it was recommended that if young calves under 3 months of age are tested by capture ELISA for persistent infections they be retested again at 3 months of age.

Haines et al. (1992) developed a monoclonal anti- body-based immunohistochemical method for detecting BVDV antigen in formalin-fixed tissues using the monoclonal antibody, 15C5. This monoclonal antibody, with specificity for the Erns protein, was the only monoclonal antibody out of 32 tested that was reactive against antigen preserved in these tissues, suggesting that most protein epitopes of BVDV are conformational in nature. The mono-

clonal antibody 15C5 was found to be particularly useful for this application because its epitope was shown to be highly conserved, reacting with all of the 70 BVDV isolates tested (Corapi et al., 1990a).

Pooled monoclonal antibodies were employed in immunoperoxidase monolayer assays (IPMA) (Saliki et al., 1997; Deregt and Prins, 1998) as well as in a monolayer ELISA (m-ELISA) for microtiter virus isolation (Saliki et al., 1997). Saliki et al. (1997) utilized the Erns monoclonal antibody 15C5 and a NS3 monoclonal antibody, whereas Deregt and Prins (1998) utilized a pool of E2 and NS3 monoclonal antibodies for detection of BVDV 1 and BVDV 2. The m-ELISA uses a spectrophotometer for reading samples and is a more rapid test than the IPMA. The microtiter IPMA and m-ELISA, compared to conventional virus isolation, were shown to have a relative sensitivity of 100% for samples from cattle greater than 3 months of age suspected of being persistently infected and 85% when samples of cattle with acute infections were included (Saliki et al., 1997).

Flow cytometry has also been investigated for use as a diagnostic assay. Qvist et al. (1990, 1991) evaluated its use in identification of persistently infected cattle and found it to be more sensitive than virus isolation. Using a fluorescence-activated cell sorter to analyze fluorescent antibody-bound, infected leukocytes, BVDV-specific antigens were shown to occur in 3–21% (mean 11%) of mononuclear leukocytes of persistently-infected cattle.

POLYMERASE CHAIN REACTION

The first of many reverse transcription-polymerase chain reaction (RT-PCR) assays for BVDV was developed in 1990 (Schroeder and Balassu-Chan, 1990). Although validation of this assay was limited, it was reported to be as analytically sensitive as the IPMA for virus detection. Since the development of this assay, many more sophisticated PCR assays were developed for BVDV in the 1990s, which matched the development of PCR in general for other viruses. These included nested PCR assays with two rounds of PCR to increase analytical sensitivity, multiplex assays with species-specific primers for genotyping, and real-time PCR assays that utilize fluorescent signals from oligonucleotide probes to detect amplification.

With the discovery of BVDV 2 and the existence of two genotypes of BVDV in the mid 1990s, several PCR-based assays for typing BVDV were developed. Harpin et al. (1995) used an indirect method, following RT-PCR with restriction endonu-