canine distemper

Detection of Canine Distemper Virus Nucleoprotein RNA by Reverse Transcription-PCR Using Serum, Whole Blood, and Cerebrospinal Fluid from Dogs with Distemper

A. L. Frisk,¹ M. König,² A. Moritz,³ and W. Baumgärtner^1,*

Institut für Veterinär-Pathologie,¹ Institut für Virologie, Fachbereich Veterinärmedizin,² and Medizinische und Gerichtliche Veterinärklinik,³ Justus-Liebig-Universität Giessen, 35392 Giessen, Germany

Received 21 April 1999/Returned for modification 10 June 1999/Accepted 26 July 1999

ABSTRACT

Reverse transcription-PCR (RT-PCR) was used to detect canine distemper virus (CDV) nucleoprotein (NP) RNA in serum, whole blood, and cerebrospinal fluid (CSF) samples from 38 dogs with clinically suspected distemper. Results were correlated to clinical findings, anti-CDV neutralizing antibody titers, postmortem findings, and demonstration of CDV NP antigen by immunohistochemistry. The specificity of the RT-PCR was ensured by amplification of RNA from various laboratory CDV strains, restriction enzyme digestion, and Southern blot hybridization. In 29 of 38 dogs, CDV infection was confirmed by postmortem examination and immunohistochemistry. The animals displayed the catarrhal, systemic, and nervous forms of distemper. Seventeen samples (serum, whole blood, or CSF) from dogs with distemper were tested with three sets of primers targeted to different regions of the NP gene of the CDV Onderstepoort strain. Expected amplicons were observed in 82, 53, and 41% of the 17 samples, depending upon the primer pair used. With the most sensitive primer pair (primer pair I), CDV NP RNA was detected in 25 of 29 (86%) serum samples and 14 of 16 (88%) whole blood and CSF samples from dogs with distemper but not in body fluids from immunohistochemically negative dogs. Nucleotide sequence analysis of five RT-PCR amplicons from isolates from the field revealed few silent point mutations. These isolates exhibited greater homology to the Rockborn (97 to 99%) than to the Onderstepoort (95 to 96%) CDV strain. In summary, although the sensitivity of the RT-PCR for detection of CDV is strongly influenced by the location of the selected primers, this nucleic acid detection system represents a highly specific and sensitive method for the antemortem diagnosis of distemper in dogs, regardless of the form of distemper, humoral immune response, and viral antigen distribution.

INTRODUCTION

Canine distemper virus (CDV), which is closely related to measles virus and rinderpest virus, two other members of the genus Morbillivirus of the Paramyxoviridae family, is a devastating, highly contagious pathogen that occurs worldwide (10, 32). The host spectrum of CDV comprises dogs and many other carnivores and noncarnivores as well as marine mammals (1, 3, 7, 10, 27, 45). A possible link between Paget's disease of bone in humans and CDV infection was shown by epidemiological studies and was substantiated by detection of CDV RNA in affected tissues (17, 30). CDV is also discussed as a candidate that might play a role in the initiation of multiple sclerosis (35). Recently, a new member of the Paramyxoviridae family was isolated from an outbreak of fatal respiratory and nervous disease in horses and humans in Australia. This new isolate, first classified as a morbillivirus, most likely represents a new genus within the Paramyxovirinae subfamily (26, 46).

In dogs, CDV infection can result in subclinical infection, gastrointestinal signs, and/or respiratory signs, frequently with central nervous system (CNS) involvement (3, 4, 22). Nervous signs may also occur as a late manifestation of CDV infection without any other signs (7, 22, 33). Following aerosol infection (4), the virus replicates in macrophages and lymphoid cells of the upper respiratory tract (4, 22). Systemic dissemination is mediated by infected cells, such as lymphocytes, monocytes, and platelets, and/or occurs through non-cell-associated virus, leading to infection of various organs (5, 23, 44). Pathologic lesions are most prominent in the respiratory and gastrointestinal tracts, lymphoid tissues, and CNS (1, 2, 7, 14, 29).

A variety of clinical parameters and different types of assays have been suggested for use for the definitive antemortem diagnosis of distemper. However, due to the unpredictable and variable course of distemper, e.g., length of viremia, organ manifestation, and a lack of or delayed humoral and cellular immune responses, the final diagnosis for most animals remains uncertain. Various specimens including conjunctival and vaginal imprints, urinary epithelium cells, skin and stomach biopsy specimens, cells from tracheal washings, blood smears, and cerebrospinal fluid (CSF) taps have been used for an etiological diagnosis (1, 6, 42). In addition, inoculation of canine primary (lung macrophages or fibroblasts) or permanent cell lines with organ suspensions or cell explants from diseased animals, the ferret inoculation test, immunofluorescence, antigen immunocapture enzyme-linked immunosorbent assay, immunocytochemistry, and in situ hybridization have been used for detection of CDV antigen and CDV RNA (3, 4, 6, 16, 40). However, the majority of these methods are laborious and time-consuming, and, more importantly, they are of limited usefulness when they are applied to clinical specimens. Although immunohistochemistry represents a highly sensitive and specific method for detection of CDV antigen in tissue obtained postmortem, it is suitable only within limits for the diagnosis of distemper in living animals (6). The determination of CDV neutralizing antibodies in serum or CSF may be helpful in some animals with of chronic CNS infection, but again, the results are variable and depend on the stage of the disease. In addition, a vaccine-induced immune response or the presence of maternally derived antibodies cannot always be excluded (3, 24).

In summary, none of the methods mentioned above fulfills the requirements of a sensitive and specific CDV detection assay. Recent developments in molecular techniques revealed the suitability of these methods for diagnostic purposes as well as pathogenic and epidemiological studies (15, 39). In a recent investigation, a reverse transcription (RT)-PCR was used for detection of CDV RNA in peripheral blood mononuclear cells from dogs with suspected distemper (41). However, only 53% of the animal were positive by RT-PCR, and the diagnosis of distemper was not confirmed by or correlated with the results of other methods, including immunohistochemistry, histopathology, and in vitro virus isolation methods. To further investigate the suitability of RT-PCR for the detection of CDV RNA in clinical specimens, serum, whole blood, and CSF from dogs with spontaneous CDV infection were used as a source of viral RNA in the present study and results were correlated with the clinical, pathologic, serological, and immunohistochemical findings.

MATERIALS AND METHODS

Animals, tissue samples, and viruses. Three healthy dogs (dogs 1 to 3) and 38 dogs with suspected CDV infection (dogs 4 to 41) were used in this study (Table 1). Tissue specimens from CNS, respiratory tract, spleen, and urinary and gastrointestinal tracts were collected at necropsy from 38 animals (dogs 4 to 41), fixed in 10% nonbuffered formalin, embedded in paraffin, and investigated for CDV antigen by routine histology and immunohistochemistry techniques (7). Depending on the size of the animals, approximately 250 to 3,000 µl of serum (n = 38), 250 to 10,000 µl of whole blood (n = 22), and 250 to 3,000 µl of CSF (n = 22) were collected by venipuncture from living animals and/or from the left ventricle or vena cava and by puncture of the atlanto-occipital joint during necropsy. In addition, serum and whole-blood samples were obtained from three healthy dogs (dogs 1 to 3) one day prior to and 2 and 16 days after vaccination (SHLT+P Candur; Rockborn strain, Hoechst, Marburg, Germany). Blood smears were taken from 15-infected CDV dogs before the dogs were killed (Table 1). Serum, whole blood (clotted, without the use of anticoagulants), and CSF (without centrifugation) were stored at 80°C until they were used.

View this table: [in this window] [in a new window] TABLE 1. Age, sex, vaccination record, clinical form of distemper and histological and immunohistological findings for dogs with naturally occurring distempera

For in vitro studies, the following CDV strains were used: Onderstepoort (Ond-CDV; kindly provided by A. E. Metzler, Institut für Virologie, Universität Zürich, Zürich, Switzerland), R252-CDV (kindly provided by S. Krakowka, Ohio State University, Columbus, Ohio), Convac (kindly provided by C. Örvell, Central Microbiological Laboratory of Stockholm County Council, Stockholm Sweden), Rockborn, and four different field isolates (isolates 2582/90, 2015/91, 1052/93, and 98/91 [1]). In addition, a porpoise morbillivirus virus (kindly provided by B. K. Rima, Medical Biology Centre, University of Belfast, Belfast, United Kingdom), canine parainfluenza virus type 2 (8), and the Edmonston strain of measles virus (kindly provided by C. Örvell, Central Microbiological Laboratory of Stockholm County Council) were used.

RNA extraction. RNA was extracted from serum (150 µl), whole blood (250 mg), and CSF (150 µl) with the RNaid PLUS KIT (Dianova, Hamburg, Germany) according to the manufacturer's instruction. Briefly, cells were lysed with guanidinium thiocyanate, followed by RNA extraction with acid phenol and chloroform-isoamyl alcohol (24:1). RNA, which was present in the top aqueous phase, was purified by adsorption to an RNA matrix. Negative controls for carryover contamination included RNA extracted from noninfected African green monkey kidney (Vero) cells between the extraction of RNA from each sample from the dogs. Vero cells infected with Ond-CDV served as a positive control.

RT-PCR and restriction enzymes. The oligonucleotides used for amplification of the CDV nucleoprotein (NP) gene sequences are shown in Fig. 1 and Table 2. Positions are indicated according to the positions of Sidhu et al. (42), which are available from the GenBank-EMBL data bank under accession nos. AF014953, L13194, and L13195. The sequences of all CDV primers except the antisense primer at positions 1610 to 1587 were localized in the highly conserved region of the NP gene of the Ond-CDV strain, which shows great homology among morbilliviruses (11, 34, 36, 42). The expected amplicon lengths are 287, 260, and 900 bp for primer pair I (PP-I), PP-II, and PP-III, respectively. RNA integrity was ensured by amplification of a sequence from a housekeeping gene that encodes glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (primers were kindly provided by T. J. Rosol, Ohio State University, Columbus, Ohio) (Table 2). The amplification product has a length of 229 bp (18).

View larger version (16K): [in this window] [in a new window] FIG. 1. Schematic drawing of the CDV genome and mRNA with location of the primers used for PCR. P/V/C, phosphoprotein; M, matrix protein; F, fusion protein; H, hemagglutinin; L, large protein; nt, nucleotide. Arrows indicate directions of primers. Numbers are molecular sizes (in base pairs). Moderate, high, and little or no, sequence homology of the NP gene within the genus morbillivirus.

View this table: [in this window] [in a new window] TABLE 2. Nucleotide sequence and position of primer pairs used for RT-PCR

The total amount of RNA isolated from whole blood varied between 400 and 1,000 ng/µl, and the RNA concentration in the CSF samples was between 20 to 400 ng/µl. The total RNA in the serum samples was not measured. The isolated RNA was transcribed into cDNA followed by PCR amplification with the RNA PCR Core Kit (Perkin-Elmer, Weiterstadt, Germany) according to the manufacturer's instruction (12, 18, 19). Briefly, RT was performed at 42°C for 15 min with 2.5 U of murine leukemia virus reverse transcriptase and 50 µM random hexamers. After inactivation of the murine leukemia virus reverse transcriptase, the PCR master mixture (0.15 µM each CDV oligonucleotide primer) was added, followed by denaturation at 94°C for 1 min and 40 cycles consisting of denaturation at 94°C for 1 min, annealing at 59.5°C for 2 min, extension at 72°C for 1 min, and final extension at 72°C for 5 min in a thermocycler (Biomed TC 60/2). The PCR products were analyzed on a 2% agarose gel after staining with ethidium bromide.

DNA restriction enzymes AluI (Advanced Biotechnologies Ltd., Hamburg, United Kingdom) and BsiMI (Angewandte Gentechnologie Systeme GmbH, Hamburg, Germany) were used for further characterization of the amplicons (Table 3) (37).

View this table: [in this window] [in a new window] TABLE 3. DNA restriction enzyme endonucleolytic cleavage sites and fragment sizes for three different RT-PCR products derived from CDV NP RNA

Statistical analysis. By assuming that the immunohistochemistry method represents a well-characterized and highly specific method for detection of CDV antigen, the RT-PCR results were correlated with the results obtained with this protein detection system. Therefore, sensitivity refers to the number of RT-PCR-positive probes for the group of immunohistologically CDV-positive dogs, whereas specificity expresses the number of RT-PCR-negative probes for the group of animals that were negative for CDV by the immunohistochemistry method. The computation of the 95% confidence limits for specificity and sensitivity of the RT-PCR was performed by using the statistical program package BiAS.

Sequence analysis of PCR products. The 287-bp DNA fragment (PP-I) was extracted from the agarose gel and was directly sequenced from samples from five animals (animals 13 [isolate 833-Gi95], 23 [isolate 1127-Gi95], 29 [isolate 2852-Gi95], 30 [isolate 2153-Gi95], and 40 [isolate 2495-Gi95]) and the Rockborn CDV strain. Briefly, after electrophoresis the DNA bands were visualized under UV light (360 nm) and were excised from the agarose gel. Purification of the DNA was performed according to the standard protocol with the QIAEX II DNA extraction kit (QIAGEN GmbH, Hilden, Germany). Direct sequencing was performed with the Vent cycle sequencing kit by using p1 or p2. The DNA fragments were separated on a 5% polyacrylamide gel (Sequa gel; Biozym, Oldendorf, Germany) in 1× TBE (Tris-borate-EDTA) buffer. After fixation (10% acetic acid) and drying, the gels were exposed to a Biomax X-ray film (Kodak, Berlin, Germany) for 1 day. Analysis of sequence data was performed by using the GCG package (Genetics Computer Group, Inc., Madison, Wis.).

Southern blotting. To ensure the specificities of the RT-PCR products, Southern blotting was performed with each amplicon obtained by RT-PCR for CDV by using a digoxigenin (DIG)-labeled double-stranded DNA (dsDNA) probe (13). Briefly, DIG-11-dUTP (DIG-dUTP) was incorporated during PCR by using the PCR DIG Labeling mix (Boehringer Mannheim, Mannheim, Germany) and PP-I, resulting in a 287-bp dsDNA probe. For Southern hybridization, a standard capillary blot was applied (37). Prehybridization and hybridization were performed at 42°C under constant, gentle agitation. The hybridization buffer contained 6 ng of dsDNA probe in 10 ml of prehybridization buffer (2% [wt/vol] blocking stock solution, 50% [vol/vol] formamide, 0.1% N-lauroyl sarkosine-NaCl, 0.02% [wt/vol] sodium dodecyl sulfate, and 5× SSC [SSC is 750 mM NaCl plus 75 mM sodium citrate]). After washing under stringent conditions, the membrane was incubated with an anti-DIG-alkaline phosphatase antibody (Boehringer Mannheim). To visualize the hybridization reaction, a colorimetric detection system (nitroblue tetrazolium chloride, X-phosphate) was used.

Histology and immunohistochemistry. For histological examination, tissue sections were cut to a thickness of 2 to 4 mm and were stained with hematoxylin-eosin. In addition, CNS sections were stained with luxol fast blue-cresyl-violet to determine the loss of myelin. Immunohistochemically, viral protein was demonstrated by the avidin-biotin complex method with a CDV NP-specific monoclonal antibody (monoclonal antibody NP-2; clone 3991) kindly provided by C. Örvell, Central Microbiological Laboratory, Stockholm County Council (7). The degree of immunoreactivity was scored semiquantitatively as follows: (+), single positive cells; +, single focus of immunopositive cells; ++, moderate number of immunopositive cells; and +++, numerous immunopositive cells.

Serum and CSF microneutralization test. To investigate the presence of anti-CDV neutralizing antibodies in serum and CSF, a standard serum microneutralization test was performed in 96-well microtitration plates (8). Prior to use, serum and CSF samples were heat inactivated. Twofold serum dilutions of 50 µl were prepared (starting dilution, 1:10 in Eagle's minimum essential medium with 10% fetal calf serum) and were tested in quadruplicate. A total of 50 µl of the Eagle's minimum essential medium with 100 median tissue culture infective doses of the Ond-CDV strain was added to each well. Serum-virus mixtures were incubated at 37°C for 1 h. A total of 100 µl of the Vero cell suspension was added to each well, and the titration plates were incubated at 37°C in 5% CO₂ for 3 to 5 days. The neutralizing capacity of the sera was determined by inhibition of the Ond-CDV-induced cytopathogic effect (giant cell formation) and the neutralization titer was calculated by the Reed and Muench method (8).

RESULTS

Histological and immunohistochemical findings. According to the immunohistological findings the necropsied animals were divided into CDV antigen-negative animals (group I; dogs 4 to 12) and CDV antigen-positive animals (group II; dogs 13 to 41).

Animals in group I displayed a variety of changes including anemia, bronchopneumonia, septicemia, cardiac dilatation, subaortic stenosis, and subdural hemorrhage in the spinal cord. CNS lesions were absent from three animals (dogs 4 to 6), and the remaining dogs suffered from granulomatous meningoencephalitis, lymphohistiocytic meningitis, purulent choroiditis, nonsuppurative encephalitis, or a meningioma.

The immunohistochemical and most important histological observations for animals in group II are summarized in Table 1. Five dogs lacked microscopic brain lesions. The remaining 24 animals displayed white matter lesions characteristic of acute to chronic distemper (Table 1 and Fig. 2). Interstitial pneumonia and/or purulent bronchopneumonia and lymphocytic depletion in the spleen were also observed. Cytoplasmic and intranuclear inclusion bodies were found in the CNS, epithelium cells of the gastric mucosa, urinary bladder, renal pelvis, bronchi, and bronchioles of various animals.

View larger version (205K): [in this window] [in a new window] FIG. 2. Cervical spinal cord of dog 38 showing chronic myelitis with malacia, demyelination, and moderate perivascular lymphohistiocytic cuffs. Hematoxylin and eosin stain was used. Magnification, ×35.

Widespread distribution of CDV antigen indicating early CNS infection was found in endothelial, meningeal, and ependymal cells, choroid plexus epithelium, and occasionally, Purkinje's cells and astrocytes of eight dogs (Fig. 3). In 20 animals (dogs 20, 22, and 24 to 41), CDV antigen was found predominantly in lesions, although some chronic lesions were completely devoid of viral antigen. At extracerebral sites, viral antigen was detectable in bronchial epithelium cells, bronchial glands, and alveolar macrophages of the respiratory tract (Table 1). CDV antigen was also observed in gastrointestinal and urinary tract epithelium cells, splenic lymphocytes, and interdigitating follicular cells. Detection of virus antigen in vascular endothelium cells and/or intravascular leukocytes from of 13 dogs indicated ongoing viremia (Fig. 3 and Table 1). Although blood smears were available for seven of these dogs, no virus antigen-positive cells were detected in these preparations (Table 1). The only animal (dog 13) with CDV antigen-positive cells in the blood smear showed no evidence of viremia in tissue sections, underlining the highly variable and unpredictable course of virus dissemination in animals with distemper.

View larger version (175K): [in this window] [in a new window] FIG. 3. Cerebellum of dog 32 showing a strong positive signal for NP antigen in endothelial cells and intravascular lymphocytes interpreted as ongoing viremia. The avidin-biotin complex method was used. Magnification, ×350.

Clinical findings. The mean age of the two female and seven male animals (animals 4 to 12) in group I was 7.3 months (age range, 2 to 18 months). One dog was vaccinated with unknown vaccines, and the vaccination records for the remaining eight animals were not available. Clinically, two dogs (dogs 7 and 12) presented with neurological dysfunction, including partial and generalized seizures, hind-leg ataxia, and rhythmic tonic-clonic movements. Four dogs (dogs 6 and 8 to 10) showed gastrointestinal and/or respiratory tract disease, and three dogs (dogs 4, 5, and 11) displayed nervous system and gastrointestinal signs.

The mean age of the 17 female and 12 male dogs in group II with different vaccination histories was 7.2 months (age range, 2 to 36 months) (Table 1). According to their clinical findings 10, 14, and 4 animals (Table 1) suffered from the catarrhalic, systemic, or nervous form of distemper, respectively. The nervous form of distemper was characterized by seizures, hind-leg ataxia, and rhythmic tonic-clonic movements. Dogs suffering from the catarrhalic form of distemper showed gastrointestinal and/or respiratory tract disease, whereas animals with the systemic form of distemper displayed a mixture of both the nervous and catarrhalic forms, including nervous signs, fever, mucopurulent conjunctivitis and rhinitis, and multifocal erosive dermatitis.

Serological results. Four of six animals in group I had a virus neutralization antibody titer higher than 1:100. Two of three healthy control animals seroconverted 16 days after vaccination. Among the animals in group II (Table 4), the virus neutralizing antibody titers in seven dogs were >1:100, those in seven animals were between 1:40 and 1:100, and those in the remaining nine dogs were <1:40.> virus neutralization antibody titer in CSF samples from 10 dogs with confirmed CDV infection was >1:40 in only 1 dog.

View this table: [in this window] [in a new window] TABLE 4. Detection of CDV NP nucleic acid by RT-PCR and virus-specific neutralizing antibody titers in dogs with distemper

RT-PCR results and restriction enzymes. The GAPDH-specific amplification product was demonstrated in all whole-blood samples, whereas amplification of GAPDH was possible for only 63% of the CSF samples. Amplification from the GAPDH housekeeping gene was not possible for two and four CSF samples from animals in group I (dogs 4 and 8) and group II (dogs 15, 19, 33, and 34), respectively.

RT-PCR with CDV RNAs from various CDV strains resulted in amplicons of the expected length for each primer pair. The specificities of PCR amplification products were ensured by restriction enzyme digestion and positive hybridization by Southern blotting. Porpoise morbillivirus RNA was amplified only by PP-I and not by PP-II and PP-III. No amplification products or hybridization signals were detected with canine parainfluenza type 2 and the Edmonston strain of the measles virus (Fig. 4 and 5). To investigate the influence of the selected primer pairs on the RT-PCR results, samples from 17 CDV antigen-positive animals were tested with the three primer pairs. For samples from 14 (82%), 9 (53%), and 7 (41%) dogs, specific RT-PCR bands were observed with PP-I, PP-II, and PP-III, respectively, by use of serum, whole blood, and CSF (Table 5). Although the number of RT-PCR-positive animals was not increased by using all three primer pairs for amplification of CDV RNA in the same body fluid, the number of positive animals was increased when all three body fluids from one animals were used. By using PP-I in RT-PCR tests with the remaining tissues, CDV NP RNA was detected in 25 serum samples (sensitivity, 86%; 95% confidence interval, 68 to 96%) and 14 whole-blood and CSF samples (sensitivity, 88%; 95% confidence interval, 62 to 98%) (Fig. 6 and Table 4). CDV RNA was not detected in serum samples (specificity, 100%; 95% confidence interval, 72 to 100%) or whole-blood and CSF samples (specificity, 100%; 95% confidence interval, 61 to 100%) from immunohistologically CDV negative dogs. All samples from 11 (85%) of 13 dogs with virus antigen in the vascular endothelium and/or in the intravascular space showed specific RT-PCR products. CSF samples from 2 of 22 animals with a systemic antigen distribution lacked amplification products, but a specific amplicon was detected in serum and whole-blood samples. Samples from animals in which the virus antigen distribution restricted to the CNS showed variable RT-PCR results (Table 4). In one animal (dog 37) a strong hybridization signal was obtained by Southern blotting, even though no band was visible in the ethidium bromide-stained agarose gel. Although CDV RNA was detected in most samples, it appeared that negative results or only weak bands were more frequently found in gels for serum samples from dogs with nervous distemper and that virus antigen expression was restricted to the CNS (Table 4). All RT-PCR products were cleaved with the AluI restriction enzyme. Surprisingly, digestion with the restriction enzyme BsiMI was observed in only six dogs (dogs 15, 17, 23, 28, 33, and 40), indicating nucleotide substitutions in most isolates between positions 975 and 980.

View larger version (29K): [in this window] [in a new window] FIG. 4. RT-PCR results (top panels) obtained with PP-I from CDV laboratory strains, CDV field isolates, Edmonston strain of measles virus, porpoise morbillivirus, and canine parainfluenza virus RNA and results obtained by Southern blot analysis (bottom panels). (A) Lanes 1, CDV-Convac; lanes 3, canine parainfluenza virus type 2; lanes 5, porpoise morbillivirus; and lanes 7, R252-CDV. (B) Lanes 1, Ond-CDV; lanes 3, Edmonston strain of measles virus; lanes 5, field isolate 98/91; lanes 7, field isolate 2582/90. Lanes with even numbers, negative controls (noninfected Vero cells); lanes M1, DIG-labeled molecular size marker; lane M2, molecular size marker (100-bp ladder). Numbers on the left and right are molecular sizes (in base pairs).

View larger version (26K): [in this window] [in a new window] FIG. 5. RT-PCR results (top panels) with PP-III from CDV laboratory strains, CDV field isolates, the Edmonston strain of the measles virus, porpoise morbillivirus, and canine parainfluenza virus RNAs and results obtained by Southern blotting analysis (bottom panels). (A) Lanes 1, CDV Convac; lanes 3, canine parainfluenza virus type 2; lanes 5, porpoise morbillivirus; and lanes 7, CDV R252. (B) Lanes 1, Ond-CDV; lanes 3, Edmonston strain of measles virus; lanes 5, field isolate 98/91; and lanes 7, field isolate 2582/90. Lanes with even numbers, negative controls (noninfected Vero cells); lanes M1, DIG-labeled molecular size marker; lanes M2, molecular size marker (100-bp ladder). Numbers on the left and right are molecular sizes (in base pairs).

View this table: [in this window] [in a new window] TABLE 5. Detection of CDV NP RNA in serum, CSF, and whole blood by RT-PCR with three different primer pairs in 17 dogs with immunohistologically confirmed CDV infection

View larger version (41K): [in this window] [in a new window] FIG. 6. RT-PCR amplification of CDV NP nucleic acid (A) and Southern blot analysis (B) of CSF, serum, and whole blood from dog 32 with PP-I. Lanes 1, 3, and 5, amplicons in CSF, serum, and whole blood, respectively; lanes 2, 4, and 6, controls (noninfected Vero cells); lanes M1, DIG-labeled molecular size marker; lanes M2, molecular size marker (100-bp ladder). Numbers on the left and right are molecular size markers.

Sequence analysis. RT-PCR products from samples from five animals (dogs 13, 23, 29, 30, and 40) were sequenced. For three of these animals (dogs 13, 29, and 30), endonucleolytic cleavage sites for BsiMI were lacking. The alignment of the nucleotide sequences revealed a small number of nucleotide substitutions (2 to 12) compared to the numbers for the Rockborn and Ond-CDV (42) strains; the substitutions were most frequently observed at the third positions of the codons (Fig. 7). In two sequences the first base of the codon was substituted. A transversional substitution was observed in only one sequence; the remaining substitutions were transitional replacements. None of these resulted in a change in the deduced amino acid sequence. Nucleotide sequence analyses revealed 97 to 99% homology with the Rockborn CDV strain and 94 to 95% homology with the Ond-CDV strain (36, 42). A substitution of cytosine for thymine at position 977 compared to the sequence of Ond-CDV was found in the sequence of the RT-PCR products from samples from animals which were not cleaved with BsiMI.

View larger version (17K): [in this window] [in a new window] FIG. 7. Alignment of the nucleotide sequence of the NP gene of different distemper virus strains (five CDV isolates and the Rockborn strain) compared to that of the NP gene of Ond-CDV GenBank-EMBL data bank accession numbers are as follows: strain 13, AF166268 (isolate 833-Gi95); strain 23, AF166269 (isolate 1127-Gi95); strain 29, AF166270 (isolate 2852-Gi95); strain 30, AF166271 (isolate 2153-Gi95); strain 40, AF166272 (isolate 2495-Gi95); CDV Rockborn (CDV-ROCK), AF166273.

DISCUSSION

The present study confirms and extends previous observations on the usefulness of RT-PCR as a fast, sensitive, and specific method for the diagnosis of CDV infection in dogs. CDV RNA was detected by RT-PCR in 86% of serum samples and 88% of whole-blood and CSF samples from dogs with immunohistochemically confirmed distemper. The nucleic acid detection system applied in the present study proved to be highly sensitive and specific, regardless of clinical signs, pathological findings, neutralizing antibody titers, and virus antigen distribution. However, the sensitivity of the RT-PCR varied between selected primers, depending on their position in the gene.

The three different primer pairs investigated recognized various CDV strains but not closely related morbilliviruses and paramyxoviruses, such as the Edmonston strain of measles virus and canine parainfluenza virus type 2. Surprisingly, despite congruent results with laboratory strains, the three primer pairs differed in their sensitivities when they were applied in tests with clinical specimens, indicating that isolates from clinical specimens might display higher degrees of nucleotide substitutions. RT-PCR with all three primer pairs did not increase the sensitivity of the assay. However, by use of RT-PCR with all three different body fluids (serum, CSF, and whole blood), sensitivity was increased, showing a heterogeneous distribution of CDV RNA in different body compartments. Similar results with respect to the role of the selected primer pair for the sensitivity of RT-PCR for detection of CDV has been described by others (41). Furthermore, the activities of endogenous RNases and lack of accessibility of partially degraded RNA may influence the sensitivity of RT-PCR. Serum, whole blood, and CSF appeared to be equally suitable as substrates for the RT-PCR. We found no evidence of inhibition of the Taq DNA polymerase by hemoglobin, as described elsewhere (31). The false-negative results with whole blood from two animals from group II were not due to inadequate RNA isolation, as demonstrated by GAPDH amplification. Amplification of GAPDH was not possible for two and four CSF samples from animals in groups I and II, respectively, indicating a lack of cells in these preparations. However, CDV RNA could still be amplified from the CSF of two of these dogs, suggesting that CDV RNA might not be always associated with CSF cells; alternatively, single cells carry very large loads of CDV RNA. CDV RNA was not detected in immunohistochemically CDV antigen-negative animals or in dogs following vaccination, supporting previous observations that a previous vaccination does not cause false-positive results (43). In contrast, Shin et al. (41) obtained positive RT-PCR results until 10 days after vaccination, indicating that under certain circumstances vaccination may cause false-positive results for dogs. To rule out false-positive results due to cDNA contamination, DNase treatment prior to RT (data not shown) or PCR without preceding RT was performed in some cases; however, there was no evidence of CDV cDNA, as has been described for murine lymphocytic choriomeningitis virus infection (21). The negative RT-PCR results for four serum samples and for whole blood and CSF samples from four and two animals, respectively, might be due to a complete lack of CDV RNA or to the presence of only low levels of CDV RNA in the samples. Autolytic degradation of CDV RNA due to released endogenous RNases should be considered a possible source of false-negative results; however, in the present and previous studies, CDV transcripts were also found in animals with advanced autolytic changes (12, 13, 19), indicating that postmortem changes play an inferior role as a cause of false-negative results.

Interestingly, CDV RNA was also found in serum, whole blood, or CSF from animals with subacute and chronic distemper encephalitis. By immunohistochemistry, in some of these animals, viral antigen was restricted only to the CNS. Whether the detected CDV RNA represents intracellular degradation products or a mechanism of virus spread and persistence in these animals remains to be determined. Similarly, the measles virus genome was detected in plasma, peripheral blood mononuclear cells, and CSF from patients with subacute sclerosing encephalitis (SSPE) and measles encephalitis (28, 38).

So far, confirmation of suspected canine distemper virus infection in living dogs was unrewarding, mainly because of the low level of sensitivity of the available methods (3, 6, 16). In the present study, detection of virus antigen in vascular endothelial cells and/or intravascular leukocytes was observed in 13 dogs indicating that these animals were still in the stage of ongoing viremia. Notably, in seven of these dogs no virus antigen was demonstrated in blood smears, supporting the observation of the low sensitivity of this assay (6). Interestingly amplification of CDV RNA in five of seven serum samples and one of three whole-blood samples from animals in which virus antigen expression was restricted to the CNS was possible. These findings cannot readily be explained. Whether detection of CDV RNA in the absence of viral protein is the result of a restrictive infection as described for oligodendrocytes and neurons remains to be clarified in future studies (29, 48).

Detection of neutralizing antibodies did not correlate with the form of distemper, antigen distribution, or RT-PCR results, indicating the noncontributory role of neutralizing antibody titers for the etiological diagnosis of distemper. Furthermore, neutralizing activity against the Onderstepoort strain may not correspond to neutralizing activity against field isolates and, therefore, may not be protective (25).

Comparison of the sequences of selected isolates to the sequences of vaccine strains demonstrated distinct silent point mutations. Katayama et al. (20) found significant nucleotide substitutions with changes in the deduced amino acids among measles and SSPE viruses in a highly conserved region of NP in brain tissues. Similar to SSPE virus, nucleotide transitions were more frequent than transversions in CDV NP RNA (9). The CDV isolates showed greater nucleotide sequence homology to the Rockborn strain than to the Ond-CDV. Although the database is too small and the significance of the observed silent mutations remains unclear, most nucleotide substitutions were found in CDV RNA from a dog with chronic brain lesions. It is tempting to speculate that these findings suggest that there could be a correlation between an altered NP RNA sequence and viral persistence. However, the possibility that the observed mutations might have an impact on virus translation and virus persistence remains speculative and needs to be substantiated by further studies. Similarly, the biological significance of these mutations needs to be substantiated by further studies by using a broader database and by including CDV genomic regions with known variability, such as the hemagglutinin protein (10). Yoshida et al. (47) found that CDV field isolates in Japan had one cluster of nucleotide substitutions that distinguished them from the laboratory Onderstepoort strain. However, they also found no correlation between sequence substitutions and differences in distemper pathology.

In summary, RT-PCR for detection of CDV represents a sensitive and specific method for the early and safe antemortem diagnosis of distemper by using serum, whole blood, and/or CSF regardless of clinical signs, pathological findings, neutralizing antibody titers, and virus antigen distribution.

ACKNOWLEDGMENTS

This study was supported by grants from the Deutsche Forschungsgemeinschaft (grants Ba 815/3-1 and Ba 815/3-2) and the Gemeinnützige Hertie-Stiftung.

We thank Sandra Heinz and Annette Artelt for excellent technical assistance; Ute Zeller for photographic support; Paul Becker, Institut für Virologie des Fachbereichs Veterinärmedizin der Justus-Liebig-University, Giessen, Germany, for kind help and support with the analysis of the sequence data; and K. Failing and H. Heiter, Arbeitsgruppe Biomathematik und Datenverarbeitung des Fachbereichs Veterinärmedizin der Justus-Liebig-Universität Giessen, for performing the statistical analysis. We also thank A. Lemmer for providing the serum and whole-blood samples from three of his dogs.

FOOTNOTES

^* Corresponding author. Mailing address: Institut für Veterinär-Pathologie, Justus-Liebig-Universität Giessen, Frankfurter Strasse 96, 35392 Giessen, Germany. Phone: 49 (0)641 99 38202. Fax: 49 (0)641 99 38209. E-mail: wolfgang.baumgaertner@vetmed.uni-giessen.de .