Abstract
Virulent systemic feline calicivirus (VS-FCV) is a novel, emerging pathogen with mortality up to 67% even in previously healthy adult cats; VS-FCV has resulted in at least six epidemics since 1998. Affected cats have systemic vascular compromise and hemorrhagic-fever like signs in part due to viral invasion of epithelium and endothelium, coupled with host cytokine responses. Affected skin tissues had, on average, 3.8 elevated cytokines compared with control tissue, with prominent upregulation in IL-10, TNF-α, and MIP-1α. Sequencing of most of the genomes of two VS-FCV strains documented patterns of virus relatedness and implicated changes in the capsid gene in the emerging phenotype, possibly through initiation of immune mechanisms manifest in the cytokine changes. Understanding the features contributing to the emergence of this disease is critical for management and prevention of this and similar outbreaks attributable to RNA viruses in animals and humans.
Genomic instability in RNA viruses, coupled with ineffectual or counterproductive host responses, can result in novel emerging infections. One emerging virus with high mortality is virulent systemic feline calicivirus (VS-FCV), a novel pathogen with up to 67% mortality even in healthy adult cats that has erupted in at least six epidemics since 1998 with epidemiological and clinical similarities to hemorrhagic fever virus epidemics in humans (Pedersen et al 2000, Hurley et al 2004). All outbreaks were characterized by rapid onset and spread, with enigmatic, gradual or abrupt conclusion. Most cats had generalized facial and limb swelling, edema consistent with systemic vascular compromise, high fever (39.4–42.4°C), multiple-organ dysfunction and sudden death.
Adult age and feline calicivirus vaccination (although confounded) were implicated as risk factors in severe disease, suggesting a possible immunopathological mechanism to some lesions in addition to possible direct cellular damage induced by the virus. Understanding the epidemiology, pathogenesis and underlying features contributing to the emergence of this disease is critical for management and prevention of this and similar outbreaks attributable to RNA viruses in animals and humans. Here, we describe immunopathological evaluation of affected cats as well as results of sequencing of most of the genomes of two VS-FCV strains.
Materials and methods
Cats
Seven cats with VS-FCV infection were evaluated. All met the case definition of having a consistent exposure history (ie, from an affected practice or with contact with a confirmed case), clinical signs of VS-FCV (fever, facial or limb edema, and multiple-organ dysfunction), and an FCV strain was recovered by culture or PCR with capsid hypervariable region sequences identical to those from a known case strain in the same epidemic. Three cases were kittens of 8–16 weeks of age, one was 8 months old, and three were described as adults over 2 years of age. Four were female and three were male; all adults were neutered. All were from a previously described outbreak in Los Angeles County in 2002 (Hurley et al 2004), with index case strains designated FCV-Kaos.
FCV culture
Virus culture was performed using EDTA-anticoagulated blood, oropharyngeal secretion, or spleen and lung specimens collected at the time of necropsy. Specimens were cultured on a confluent monolayer of Crandall-Reese feline kidney (CRFK) cells at 37°C in air with 5% CO2 in 1:1 Liebovitz L-15 medium and Eagle's minimum essential medium with 10% fetal bovine serum, 100 U of penicillin G/ml, and 100 μg of streptomycin/ml. Infection was confirmed by the presence of characteristic cytopathic effects within 12–52 h followed by FCV-specific RT-PCR (below).
Cytokine evaluation
Cytokines including IL-1β, IL-4, IL-6, IL-10, IL-12p40, IL-18, IFN-γ, IFN-α, TNF-α, MIP-1α, and RANTES were evaluated by TaqMan PCR of cDNA as described previously (Leutenegger et al 1999, Foley et al 2004), with modifications. Formalin-fixed skin tissue corresponding to areas with light microscopic VS-FCV lesions from seven VS-FCV case cats was excised from paraffin blocks and deparaffinated with xylene. Matched control tissue was obtained from unaffected skin (based on lack of gross and microscopic lesions) from the same cats. From all samples, RNA was extracted using a kit (Qiagen Tissue Kit, Valencia, CA) and 10 μl of tRNA pre-incubated with 600 ng random hexadeoxyribonucleotide (pd(N)6) primers (random hexamer; Promega). Reverse transcription was performed in a 20 μl volume containing 50 mM Tris–HCl (pH 8.3), 40 mM KCl, 6 mM MgCl2, 0.5 mM dNTP's, 40 U RNase inhibitor (Gibco BRL, Life Technologies, Grand Island, NY), 5 mM dithiothreitol (DTT) and 200 U SuperScript II (Gibco BRL) at 42°C for 50 min. After 5 min at 95°C, 80 μl of DEPC water was added. PCR reactions contained 400 nM of each primer, 80 nM of the TaqMan probe and mastermix (TaqMan Universal PCR Mastermix, Applied Biosystems) and 5 μl of the diluted cDNA sample in a final volume of 25 μl; samples were amplified in a combined thermocycler/fluorometer (ABI Prism 7700 Sequence Detection System, Applied Biosystems) for 2 min at 50°C, 10 min at 95°C, and then 40 cycles of 15 s at 95°C and 60 s at 60°C. Final quantitation was done using the comparative C T method (Leutenegger et al 1999) and is reported as relative transcription or the n-fold difference relative to an internal calibrator (GAPDH). Cytokine levels between affected samples and unaffected controls were compared by paired t-test, with P≤0.05 as a cut off for establishing statistical significance.
cDNA sequencing
RT-PCR was performed from virus tissue culture fluid as described previously to obtain a 235 nt amplicon in the viral capsid hypervariable region for sequencing (Pedersen et al 2000). Amplification was performed in a thermal cycler (MJ Research, Watertown, MA, USA) and PCR products visualized in 2% agarose gels stained with ethidium bromide. Products were sequenced directly, ie, without cloning, following preparation in Microcon spin columns (Amicon, Beverly, MA, USA). Dye terminator cycle DNA sequencing was performed with the Ready Reaction Kit using an AmpliTaq DNA polymerase FS (ABI Prism, Foster City, CA, USA). Reactions were run on a 4.25% acrylamide/bisacrylamide gel using an ABI Prism 377 DNA Sequencer. Products were analyzed with the ABI Prism Sequencing 2.1.1 software and corrected by comparing forward and reverse for each product. Amplicons from Los Angeles (2002 southern California outbreak) were compared to sequences from VS-FCVs from northern California (FCV-Ari strain), North Carolina, and Massachusetts, the vaccine strain F9, and miscellaneous field strains using the program SeqWeb (GCG, Madison, WI). A dendrogram was generated by sequence comparison using the Cantor–Jukes correction, followed by distance calculations using the UPGMA algorithm.
In order to compare calicivirus genomes, RT-PCR primers were designed by comparison of the FCV consensus sequences among five strains published on Genbank (National Center for Biotechnology Information). Accession numbers for strains used for consensus were M86379, AF479590, AF109465, L40021, and FCLF4. Primer pairs, given in Table 1, amplified fragments no longer than 500 nt. Amplicons were directly sequenced as described above, correcting using forward and reverse products. If sequences were not obtained for any product, amplicons were cloned for sequencing into competent Escherichia coli cells using the TA-cloning kit (Invitrogen, San Diego, CA, USA), according to manufacturer's instructions. Plasmid inserts were verified by restriction endonuclease digestion of plasmids followed by agarose gel electrophoresis to determine insert size, and by PCR on plasmid DNA using specific primers. Reliability of sequences from clones was confirmed by resolving forward and reverse sequences from two randomly chosen clones for each product. Products were assembled using the program SeqWeb; this program also was used to generate predicted translation and peptide structure. Dendrograms were generated using distance and maximum likelihood algorithms with all available complete FCV sequences using Cantor–Jukes correction in the program Paup.
PCR primer pairs used in generating amplicons and sequences of VS-FCV strains as described in text
Results
Evaluation of cytokines in skin samples of cats with VS-FCV infection was performed to understand contributory roles of cytokines and indirectly some possible cellular effectors of injury and disease. On average, 3.8 of the 9 cytokines tested were elevated in affected compared with control tissue, particularly IL-10, TNF-α, and MIP-1α (P=0.04, Table 2). The mean level of IL-10 relative transcription in affected versus control tissues was 306.4 versus 0 (P=0.05) and mean MIP-1α relative transcription was 116.7 in affected tissue versus 20.8 in controls (P=0.05). The increase in TNF-α in affected tissues of 235.7 compared to 0 in controls was suggestive but not statistically significant (P=0.09). No differences were detected among affected and control tissues in the cytokines IL-1β, IL-4, IL-6, IL-12p40, IL-18, IFN-γ, IFN-α, and RANTES.
Mean value±standard error of cytokine transcription levels (TaqMan cycle threshold, normalized to GAPDH internal control) for cat skin tissue with and without histological evidence of VS-FCV infection
The rapid emergence of a novel FCV-associated clinical syndrome suggests that new genetic FCV variants could be responsible. When 235 nt amplicons in the viral capsid hypervariable region from northern and southern California VS-FCV outbreak strains were compared with North Carolina and Massachusetts VS-FCVs, F9 (the vaccine strain), and miscellaneous field strains, Los Angeles virulent strains clustered within a clade, although field strains from Los Angeles were intermingled within the tree (Fig 1); Los Angeles VS-FCV each contained a three base pair deletion, but this deletion was missing in VS-FCVs from other outbreaks. VS-FCV strains from different regions were scattered among VS-FCV and FCV field strains (although isolates from within outbreaks were closely related).
Dendrogram of a 233 nt hypervariable region of the feline calicivirus capsid region from VS-FCV strains from southern California (Kaos strains), non-virulent FCV strains from LA outbreak associated with FCV-Kaos, VS strains from Massachusetts and North Carolina, an FCV vaccine, and FCV field strains from naturally infected cats with URI.
The majority of the genomes of FCV-Ari and FCV-Kaos were sequenced and compared with other genomes previously reported, including the vaccine. At this level, FCV-Ari continued to cluster with the vaccine and was only 80.3% homologous with FCV-Kaos. VS-FCV strains did not form a clade and were inter-mixed with less virulent field strains. The predicted amino acid translation for the three reading frames of FCV was compared. In ORF 1, there were three VS-FCV-specific changes (ie, occurring in all of the VS-FCV strains and not field or vaccine strains): E→D at position 294, N→S at 1055, and T at 1314 (various amino acids in field strains). There were seven VS-FCV-specific amino acid residues in the capsid gene, including E→K at 398, V→T at 430, T→V at 438, A→K at 448, D→E at 452, R→K or D at 581, and S→D at 592, and no consistent changes in ORF-3. All seven changes in the capsid occurred generally in the same region from 398 to 592. Interestingly, protein structure was predicted to differ in the capsid with one extra glycosylation site in the VS-FCVs compared with field strains.
Discussion
Within 6 years, hundreds of cases of a frequently fatal, previously undescribed disease in cats were identified in at least six epidemic clusters. In addition to the alarming possible impact of this emerging infectious disease in cats and cat owners, this problem warrants epidemiological and immunopathological investigation because of the immunopathological and clinical similarities of some cases to emerging hemorrhagic fevers in humans (Pearks Wilkerson et al 2004). Cats infected with VS-FCV may have systemic vascular compromise and hemorrhagic-fever like signs in part due to viral invasion of epithelium and endothelium, coupled with host cytokine responses. Sequencing results appear to implicate changes in the capsid gene in the emerging phenotype.
VS-FCV infection is distinctive in its clinical severity, tropism for epithelial and endothelial cells, multisystemic attack, induction of systemic vascular compromise, and rate of involvement of visceral organs including lungs, pancreas, and liver with frequent signs of fever, edema, multiple-organ failure, hemorrhage, shock and death. Published pathological investigation of cats infected with the VS-FCV variants documented subcutaneous edema, ulceration, and segmental to full-thickness epithelial necrosis of the stratum basale, stratum spinosum and follicles (Pesavento et al 2004). Many affected cats had pulmonary edema and liver or pancreatic compromise although VS-FCV was not consistently recovered by culture or PCR from livers of affected cats, in contrast to PCR results from a similar syndrome in lagomorphs caused by rabbit hemorrhagic disease (RHD) virus (Ohlinger et al 1990). Immunohistochemical staining with a monoclonal antibody to FCV and transmission electron microscopy documented VS-FCV antigen within affected endothelial and epithelial cells, including within the skin and lungs (Pesavento et al 2004).
The mechanisms for lesion development are not known but appear to be at least partly immune-mediated. This is supported by the reported increased disease severity in older cats (Hurley et al 2004), similar to that in RHD, where young rabbits experience self-limiting diseases while older infected rabbits have almost 100% mortality (Mutze et al 1998). The three most altered cytokines in affected skin tissue from cats with VS-FCV infection were MIP-1α, IL-10, and TNF-α. MIP-1α is secreted by numerous cell types, chemoattractant for macrophages and monocytes, pyrogenic and a potentiator of IFN-γ production (Davatelis et al 1988). IL-10 is secreted by TH2 and macrophages, although it feeds back and inhibits further macrophage cytokine release. In the skin, IL-10 stimulates mast cells and IgA-producing B-cells, and upregulates MHC-II expression. TNF-α is a TH1 cytokine and may be very important in the pathogenesis of VS-FCV by virtue of its ability to increase vascular permeability, stimulate acute phase responses from liver, and induce complement activation, fever, and shock. Affected areas were compared to control tissue for the same cat with a paired t-test, similar to previous studies on vitiligo, an idiopathic, possibly immune-mediated local skin disorder in humans (Grimes et al 2004). This allowed our results to demonstrate a local cytokine effect which is not necessarily systemically active. Causation cannot be directly inferred from this association of increased cytokines with VS-FCV lesions, but overall, the cytokine and pathology findings suggest some contribution of immunopathogenic sequelae following direct viral invasion of endothelium and epithelium; in severe, generalized cases this could lead ultimately to systemic vascular compromise, microthrombus formation, disseminated intravascular coagulation, and death. This contrasts with the pathogenesis of other diseases associated with systemic vascular compromise, such as inflammatory vasculitis in Rocky Mountain spotted fever (Yamada et al 1978), immune complex vasculitis in feline infectious peritonitis (Pedersen 1995), or direct viral tissue cytotoxicity leading to edema in rabbit hemorrhagic disease (Marcato et al 1991, Ramiro-Ibanez et al 1999).
The underlying reasons for the emergence of VS-FCV are not known. One possible hypothesis is the emergence of a novel virulent VS-FCV genotype. If a novel emerging FCV genotype was responsible for the highly virulent phenotype, then not only should all isolates from within an epidemic be closely related genetically but consistent genetic changes should be detected from VS-FCV strains from among various outbreaks. Caliciviruses are non-enveloped, positive-sense, single-stranded RNA viruses, and include FCV, rabbit hemorrhagic disease virus (RHDV), and vesicular exanthema of swine virus. Like other members of the family Caliciviridae, FCV is prone to high mutation rates and minimal repair, a mechanism that has been implicated in the previous emergence of RHDV (Ohlinger et al 1990). Antigenic change and the emergence of FCV quasispecies have been reported previously during persistent infection with FCV (Pedersen and Hawkins 1995, Radford et al 1998). In the Los Angeles feline outbreak, all VS-FCV strains obtained from case cats clustered very closely together and contained a characteristic 3 nt deletion (Hurley et al 2004). However, our data show by sequencing of the capsid gene as well as the whole genome that VS-FCVs are not all members of a single clade; rather these mutant viruses are emerging from several different lineages inter-mixed with other field strain FCVs. Similar results have been obtained by researchers studying the RHDV, where no specific genetic viral determinants have been related to the emerging rabbit disease.
A second plausible mechanism for the emerging disease is that several different mutant viruses may share common pathogenic determinants when they interact with feline cells: particularly, if much of the disease pathogenesis is immune-mediated, the initial virus–host interaction could initiate common immune sequelae even if the original viral genotype differed. Caliciviruses have a unique single structural capsid protein which functions in RNA and host cell attachment. The virus has numerous arch-like capsomeres, each of which is a capsid protein dimer (Prasad et al 1994). Numerous mutations were detected in VS-FCV capsid genotypes in this study, all occurring generally in the same region from nt 398 to 592 and corresponding to an extra predicted glycosylation site in the amino acids. It is intriguing to speculate that novel receptor or other host–virus interactions could occur as a result of mutated capsid protein structure, particularly targeting epithelial or endothelial cells. Additional research to investigate possible altered host receptor–VS-FCV capsid interactions is warranted. The skin's endarterial circulation and unique local immune cell populations (particularly dendritic cells capable of secreting TNF-α) are additional intriguing targets for further evaluation of VS-FCV–host interactions.
In conclusion, genomic instability, particularly in RNA viruses, coupled with various ineffectual or counterproductive host responses can periodically result in novel emerging infections, rarely with the scope of morbidity and mortality represented by emerging VS-FCVs. Detailed comprehensive studies of epidemiology, pathogenesis, and molecular characterization of such emerging pathogens will be key to understanding and managing outbreaks in all host species.
Acknowledgments
We thank Niki Drazenovich and Mike Bannasch for technical assistance and veterinarians and shelter personnel involved with sample and data collection. Funding was provided by Maddie's® Fund, the Winn Feline Foundation, and the UC Davis Centers for Companion Animal Health and Vectorborne Diseases.