Antibody and cytokine responses in kittens during the development of feline infectious peritonitis (FIP)

2.1. Experimental design

Twelve, 8–9-week old specific pathogen-free (SPF) kittens were obtained from a commercial supplier and housed in pathogen-free quarters. The kittens were divided into three groups, one of which was left as controls (Control), one was immunised with recombinant FIPV 79-1146 spike protein (S) and the third was immunised with a recombinant modified spike protein from which three major enhancing epitopes had been deleted (SDAD). The recombinant proteins were obtained after co-transfection of sf9 cells with BacPAC 6 Baculovirus linear DNA and the Baculovirus transfer plasmid pBacPAC 8 containing the FIPV coding sequences. The proteins were administered subcutaneously with 0.1% Carbopol 941 as adjuvant, being given on three separate occasions, 3-weeks apart. Kittens were challenged oronasally with 10⁵CCID₅₀FIPV79-1146 3 weeks after the final immunisation.

Daily health checks were performed from the day of challenge. These included taking rectal temperatures and assessing the kittens for evidence of inappetence, depression, dyspnoea, ascites, jaundice, lymphadenopathy, diarrhoea, ocular or nasal discharge, oral or nasal ulceration, hypersalivation, conjunctivitis, sneezing or neurological signs. Serial blood samples were collected for the assessment of serological responses and monitoring of stimulated peripheral blood mononuclear cells (PBMC) cytokine mRNA levels.

2.2. FCoV virus neutralisation test

Virus neutralisation tests were performed in 96 well flat bottom microtiter™ plates (Costar, USA) using doubling dilutions of serum. Each serum sample was assessed in quadruplet wells. Each sample was mixed with an equal volume of cell growth medium (Eagle's minimum essential medium (MEM) (Sigma) supplemented with 5% (v/v) foetal bovine serum (Sigma), 0.23% (w/v) sodium bicarbonate, 1% (v/v) non-essential amino acids (GibcoBRL), 0.05% (v/v) lactalbumin hydrolysate (GibcoBRL), 2 mM l-glutamine and 50 μg/ml gentamicin) containing 100 CCID50 of FIPV 79-1146, then incubated at 35°C for 30 min. 5×10³ whole feline embryo (WFE) cells (Harbour et al., 1991) were then added and the cultures were incubated for four days at 37°C. Titres were expressed as the reciprocal of the dilution that completely neutralised the FIPV 79–1146.

2.3. Serum anti-FCoV IgG ELISA

Serum FCoV-specific IgG was detected using an ELISA method. Ninety-six well microtiter™ plates (Dynatech Laboratories, USA) were coated overnight at 5°C with 100 μl of sucrose gradient-purified FIPV 79–1146 in carbonate/bicarbonate coating buffer (Voller et al., 1982). The plates were then washed five times in PBS/0.1% (v/v) Tween 20. A doubling dilution series of each serum sample was added and the plates were incubated for 2 h at room temperature, then washed as before. Mouse anti-cat IgG monoclonal antibody (K62.2G4, a gift from Dr K. Haverson, University of Bristol, UK) was added at the optimum dilution in PBS/0.1% (v/v) Tween 20 and the plates incubated for a further 30 min at room temperature. After incubation, washing was repeated and goat anti-mouse/alkaline phosphatase-conjugated monoclonal antibody (Sigma) added at a 1:1000 dilution. The plates were incubated for 30 min at room temperature. After further washing, paranitrophenol phosphate (1 mg/ml) was added in carbonate/bicarbonate coating buffer. The reaction was allowed to progress for 30 min, then stopped by adding 50 μl of 0.02% (v/v) sulphuric acid. Titres were expressed as the reciprocal of the highest dilution of serum to give an optical density (OD) of at least twice that of the standard negative serum.

2.4. PBMC isolation and culture

Blood samples were centrifuged at 2000×g for 10 min. Plasma was removed and stored at −20°C prior to assessment of antibody response to FCoV (see above). Buffy coat cells were diluted to 5 ml in sPBS, overlaid onto 5 mls of Ficoll-Paque (Pharmacia Biotech) in 15 ml conical tubes and centrifuged for 20 min at 750×g, room temperature. The cells which collected at the interface were transferred to a fresh tube, washed twice with sPBS, then re-suspended at 5×10⁶ cells per ml in lymphocyte growth medium (RPMI 1640 (Sigma) supplemented with 5% (v/v) foetal bovine serum (Sigma), 1% (v/v) HEPES, 0.1 μg/ml glutamine, 50 μg/ml gentamicin) and Con A at 10 μg/ml. Separated cells consisted of greater than 90% lymphocytes, with over 90% viability as determined by trypan blue exclusion. The cells were cultured in 2 ml aliquots in 12-well tissue culture plates (Costar, USA) for 4 h at 37°C. After culture both the adherent and non-adherent cells were harvested. The non-adherent cells were transferred directly to fresh tubes, washed twice with sPBS, then pelleted. 1 ml of TRI REAGENT (Sigma) was added to each culture well to lyse the adherent cells, pipetted repeatedly, then transferred to the tube containing the corresponding cell pellet. Repeated gentle pipetting ensured complete cell lysis prior to storage at −70°C.

2.5. RT-PCR for the detection of cytokine mRNA

RNA was extracted using 1 ml of TRI REAGENT per 5×10⁶ cells. The cells were lysed by repeated pipetting until the lysate changed from a viscous to a watery consistency. Each lysate was then vortexed for 30 s prior to adding 200 μl of HPLC grade chloroform (Sigma). Each sample was vortexed for a further 30 s, then centrifuged for 15 min at 13 000×g, at 4°C (Hettich Zentrifugen EBA12R). The upper aqueous layer was transferred to a fresh 1.5 ml microcentrifuge tube and an equivalent volume (∼0.5 ml) of cold isopropanol (Sigma) was added. The sample was vortexed, then chilled at −20°C for at least 30 min (preferably greater than 2 h ). After chilling the sample was centrifuged for 15 min at 13 000×g, 4°C. The supernatant was discarded and the RNA pellet washed by adding 1 ml of 75% ethanol (molecular grade – Hayman), vortexed, then centrifuged for ten minutes at 8000×g, 4°C. The supernatant was again discarded and the RNA was dried briefly in air, prior to being re-suspended in diethylpyrocarbonate (DEPC)-treated water (∼20 μl per 5×10⁶ cell equivalents). The amount of RNA extracted varied from between 10–30 μg per 5×10⁶ cells, with A₂₆₀: A₂₈₀ ratio between 1.7–2.0.

The RNA was DNAse treated. Each aliquot of RNA solution was heated to 94°C for 5 min, then chilled quickly on ice. The reaction mixture was adjusted to contain 0.1 μg/μl total RNA, 20 mM Tris-HCl (pH 8.3), 50 mM KCl, 2.5 mM MgCl₂, 0.1 U/μl DNAse I (molecular biology grade, Gibco BRL) and DEPC-treated water to a final volume of 50 μl. The mixture was incubated for 15 min at room temperature. To terminate the reaction 2.5 mM EDTA was added and the mixture incubated at 65°C for 15 min.

To perform reverse transcription the RNA solution (containing 2–5 μg of total RNA) was incubated with 0.5 μg oligodeoxythymidine triphosphate (dT)12-18 primers for 10 min at 70°C, then chilled quickly on ice for at least 1 min. Each sample was adjusted with additional reagents to the following conditions: 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 2.5 mM MgCl₂, 0.5 mM deoxynucleic acid triphosphate (dNTP) mix, 10 mM DTT and 10 U/μl SuperScript II RT reverse transcriptase (GibcoBRL), in a total volume of 20 μl. The samples were incubated at 45°C for 55 min, heat inactivated at 70°C for 15 min, then chilled on ice. Each sample was then treated with 0.1 U/μl of E.coli RNase H and incubated for 20 min at 37°C. The cDNA was either used directly in the PCR reaction, or stored at −20°C, prior to further processing.

Primers were designed to the feline housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (G3PDH) after sequencing it. The cytokine primer sequences used initially were those designed by Rottman et al. (1995), based on published sequences (IL-2 – Cozzi et al., 1993; IL-6 – Bradley et al., 1993; Ohashi et al., 1993). A number of primers were redesigned based on new published data (IL-4 forward (F) and reverse (R); IL-10 (R); IFNγ (R); shown on Table 1 (IL-4 – Schijns et al., 1995b; IFNγ – Argyle et al., 1995; Schijns et al., 1995a). At the start of the study the nucleic acid sequence for feline IL-12 was not available. To synthesise primers for the more highly conserved IL-12 p35 subunit, consensus sequences for man, macaque, mouse, pig and cow were determined using the Genbank database. Primer sequences for IL-12 (F) and (R) are shown on Table 1. Potential sites for PCR primers were identified and consensus sequences aligned using computer assistance (MacVector™ 5.0 – Kodak). Primers were made by Genosys.

The components for the first round of the PCR reaction were combined together into a master mix adjusted to give reaction conditions: 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 2.0 mM MgCl₂, 200 μM dNTP mix, 0.1 U/μl Taq DNA polymerase (AmpliTaq Gold TM - Perkin Elmer), 2 μl (5% [v/v]) of the cDNA solution and 0.3 μM primers. DEPC-treated water was added to make a final volume of 20 μl. In order to study large numbers of samples, the PCR was performed using a 96-well polycarbonate plate format, on a temperature cycler with a heated lid (Touchdown – Hybaid). The optimised PCR protocol for detecting message for IL-2, IL-4, IL-10, IL-12 and G3PDH consisted of an initial incubation of 94°C (10 min), followed by 30 cycles of melting 94°C (1 min), annealing 58°C (1.5 min) and extension 72°C (1 min), then a single extension of 72°C (10 min). For IL-6 and IFNγ the optimised protocol consisted of an initial incubation of 94°C (10 min), followed by 35 cycles of 94°C (1 min), 50°C (1.5 min) and 72°C (1 min), then a single extension of 72°C (10 min).

Controls were incorporated with each set of PCR reactions. Positive controls included amplifying cDNA generated from kit supplied RNA, using the kit supplied PCR primers, and later, using experimental cDNA samples known to give strong positive results. Negative controls were established for each cytokine and G3PDH estimation, and included `no RT', `no primer' and `no cDNA' controls.

2.6. Detection of RT-PCR products

Once it had been demonstrated that the probes detected only single and specific bands, the PCR products were applied directly to nylon membranes (Hybond™-N Amersham), prior to detection using the enhanced chemiluminescence detection system (ECL™, Amersham Life Science). The PCR products were denatured using a thermal cycler (Touchdown – Hybaid) at 100°C for 10 min, then chilled quickly on ice. Three microlitres of the denatured PCR product were then applied directly to the nylon membrane and air dried, then fixed by UV cross linking for 45–60 s on a UV transilluminator (UVP, Cambridge, UK).

Oligonucleotide probes were designed using computer assistance (MacVector™ 5.0 – Kodak) according to published sequences for IL-2, IL-4, IL-6, IL-10, and IFN-γ, and the feline G3PDH sequence (Helps, unpublished), and on consensus sequences from man, macaque, mouse, pig and cow for IL-12 (Table 2 ). Probes were made by Genosys.

Labelling and detection of the oligonucleotide probes was performed using ECL 3′-oligolabelling and detection systems. All procedures were performed following the manufacturer's recommendations. Oligonucleotide probes were labelled with fluorescein-11-dUTP to give a final probe concentration of 0.6 μM. Blots were pre-hybridised for two to eight hours at 42°C. Labelled probe was added at 5 ng/ml of hybridisation buffer (5×SSC, with 0.1% (w/v) kit-supplied hybridisation buffer component, 0.02% (w/v) SDS, and 20-fold dilution of kit-supplied liquid block). Hybridisation was performed for 2 h at 42°C. Membranes were washed as directed and unbound probe removed by incubating twice in an excess of stringency buffer (1×SSC with 0.1% (w/v) SDS) for 15 min at 42°C. Membranes were rinsed, blocked and conjugated, excess conjugate was removed and the signal generated as per the manufacturer's recommendations. The luminescent image was detected on autoradiography film (Hyperfilm™-ECL- Amersham), the film being exposed for 5 min prior to development (GBX developer and replenisher – Kodak) and fixation (GBX fixer and replenisher – Kodak).

2.7. Semi-quantification of RT-PCR product

To allow comparison between samples it was necessary to relate the amount of starting material to the detected product. To accomplish this the protocol was standardised and measurements were made at as many steps as possible. PBMCs were incubated at fixed concentrations and the concentration of the extracted RNA was determined by spectrophotometry (Ultrospec II – LKB Biochrom); one unit of optical density at 260 nm being equivalent to approximately 40 mg of single stranded RNA (Ausubel et al., 1995).Total RNA isolated (mg)=40×optical density×dilution×volume (ml)

To control for variation in the RT-PCR, G3PDH, a `housekeeper' gene involved in glycolysis, was selected as an internal control and its mRNA assessed at the same time as that for the cytokines.

Following amplification, RT-PCR products were dotted onto membranes in duplicate. To control for variation in the labelling procedure all samples for a particular cytokine were analysed simultaneously. The autoradiograph images were quantified by densitometry using a Model GS-700 Imaging Densitomiter (Bio-Rad Laboratories, UK) and molecular analysis software version 2.1 (Macintosh/Bio-Rad Laboratories, UK). The calculated mean relative optical density (OD) corresponds to the volume of the dot multiplied by its density, after background density had been subtracted. Duplicate dot values were averaged. The G3PDH signals were used to control for variation in the quantity of RNA between samples and the efficiency of the RT-PCR.

Using cDNA known to give strong signals, dilution series were performed to determine the relationship between product and template. The relative densities were then plotted and linear regression correlation performed.

2.8. Analysis of results

Results were expressed as the log10 of the cytokine mRNA/G3PDH mRNA ratio. Each cDNA sample represented a single time point for an individual cat. Where cats were considered in groups, these values were averaged and the standard deviation calculated. When graphed the log₁₀ of the cytokine mRNA/G3PDH mRNA ratio was adjusted to zero at the start of the experiment, and all later values expressed as changes relative to that time point. Statistical analysis was performed using ANOVA with Bonferroni's post test. The relationship between OD and the concentration of the PCR product was analysed using Pearson's correlation coefficient.

3.1. Clinical progress

The proteins were well tolerated, and no abnormal signs developed in any of the kittens during the period prior to virus challenge. In the week after the challenge all of the kittens developed a transient pyrexia (defined as >39°C). The temperatures then returned to normal, before pyrexia reappeared as clinical disease became apparent (Fig. 1 ).

Within 4 weeks of the challenge all of the kittens developed clinical signs typical of FIP (Fig. 2 ); characterised by ascites and dyspnoea, with occasional other signs including sneezing, hypersalivation and diarrhoea. None of the groups showed accelerated development of disease. Kittens were euthanased when they became anorexic and depressed. All were subject to gross post mortem examination and showed typical lesions consisting of granulomata in the visceral organs; many also had peritoneal or pleural exudates.

3.2. Serological responses

All of the kittens were seronegative for FCoV at the start of the study. No neutralising antibodies were detected before challenge. Three weeks following challenge, all of the kittens had developed low to moderate neutralising antibody titres, which rose further until euthanasia (Fig. 3 ). There was no indication of an anamnestic response in the immunised groups compared to the unimmunised control group.

No serum anti-FCoV IgG antibody was detected after the first immunisation, but all the kittens developed low levels of antibody after the second dose of protein had been administered. These titres altered little by the third immunisation, but rose following challenge and continued to increase until euthanasia (Fig. 4 ). There was no indication of an anamnestic response in the immunised groups compared to the unimmunised control group.

3.3. Optimisation of PCR methodology

The PCR protocol was optimised to ensure that, for each cytokine assessed, the prodct generated a single band on an agarose gel and ran at the expected molecular weight (Fig. 5 ). Thirty to thirty-five PCR cycles were shown to be within the linear phase of amplification, as determined by assessing replicate tubes run for multiples of five cycles. Titration of the cDNA template prior to amplification revealed a linear relationship between the amount of template added and the intensity of the resulting autoradiograph image. The intensity of the product was shown to be proportional to the logarithm of the template concentration (Fig. 6 ).

The cytokine assays were found not to be compromised by DNA contamination. This was confirmed by running`no RT'-PCR controls and by removing the DNAse step. The `no RT'-PCR controls failed to generate product bands and removing the DNAse step had no effect on the product. For the cytokines the RT-PCR reactions produced single bands of the expected molecular weight, (albeit rather weak in the case of IL-12), while the `no RT'-PCR reactions failed to generate visible product. The G3PDH `no-RT'-PCR reaction produced a very faint band of the same molecular weight as that seen for the corresponding RT-PCR. Southern blotting and ECL labelling confirmed the specificity of the bands in the RT-PCR reactions and revealed no bands in the cytokine `no RT'-PCR controls. It also confirmed the specificity of the G3PDH band in the `no RT'-PCR control. The expression of this band did not appear to be due to contaminating genomic DNA as it was unaffected by DNAse treatment. Since it was shown that DNAse treatment failed to alter the product generated, the step was not incorporated in later studies.

3.4. Changes in cytokine mRNA levels

Due to the kittens' small size it was not possible to collect sufficient blood to assess PBMC cytokine mRNA at the start of the experiment. The first cytokine assessment therefore occurred 5-weeks into the trial, that is, 1-week after the second immunisation, 5-weeks before viral challenge. No cDNA was generated from the RNA extracted from blood taken on the day of challenge from kittens 2 and 10. A number of individual PCR reactions also failed to generate product. These included the 1-week post challenge reactions for IL-2 (kittens 7 and 9), and IL-12 (kittens 1, 2 and 3), and the day of euthanasia sample for IL-6 (kitten 9).

After PCR amplification and denaturation of the PCR products, all samples pertaining to an individual cytokine were dotted onto a single membrane and labelled en masse by ECL. The results are graphed in Fig. 7 . Between 5-weeks before challenge and on the day of the challenge, the expression of IL-2, IL-6 and IL-10 mRNA increased slightly. This trend was also seen with IL-4 mRNA, while the expression of IL-12 mRNA was slightly reduced. From the day of challenge to euthanasia mRNA levels for IL-2, IL-4, IL-10 and IL-12 were significantly reduced. This downward trend was also followed by IFNγ, but with this cytokine the fall was preceded by a small rise 1-week after challenge, most obviously in the Control group. Overall, the IL-6 mRNA levels remained unchanged as the disease progressed.

While the cytokine mRNA levels showed quite marked variation between individual kittens, there was little evidence that the immunisations had any effect. Statistically there was no significant difference between the three treatment groups; all showed suppression of IL-2, IL-4, IL-10, IL-12 and IFNγ mRNA levels after challenge. When looking at the IL-6 levels in individual kittens, two basic patterns emerged; the majority showed elevated IL-6 expression, at least initially, while in three of the kittens (Nos. 4, 6 and 12) there was no elevation in IL-6 (Fig. 8 ). Interestingly, these were three of the earliest kittens to show evidence of clinical disease.