Serological response to feline coronavirus in the UK domestic cat population

Abstract

Feline coronavirus (FCoV) is a ubiquitous virus of cats, with highly variable seroprevalence, ranging between 4 and 95% worldwide in the domestic cat population. It is the causative agent of feline infectious peritonitis (FIP), a disease which, prior to the recent introduction of effective antiviral drug treatments, was invariably fatal. The recent outbreak of FIP, originating in Cyprus and caused by the novel recombinant FCoV-23, highlighted the need to increase understanding of the serological response to FCoV. This study aimed to characterize the serological response to the virus and to assess cross-neutralization of type II FCoV subtypes as well as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). In serum samples of 1117 UK cats collected between 2020 and 2023, we detected an FCoV-II neutralizing antibody (nAb) seroprevalence of 13.6% and an overall seroprevalence of 59.5%; commonly cited risk factors (sex, age and breed) did not significantly impact the nAb seroprevalence nor nAb titre against FCoV-II. Antibodies against subtypes FIPV-79–1146 and FIPV-DF-2 were found to be cross-reactive and additionally neutralized both isoforms of FCoV-23. In contrast, no cross-neutralization was observed against SARS-CoV-2. Overall, these results highlight that although FCoV infection offers effective cross-protection between FCoV strains and subtypes, this does not extend to betacoronaviruses.

Introduction

Feline coronavirus (FCoV), an Alphacoronavirus in the order Nidovirales [1], is a ubiquitous virus of cats, with a prevalence of 4–95% [2,3] depending on the population tested. Risk factors for infection include the age, sex and breed of cats, as well as whether they are housed in multi-cat households or shelters [4,12].

FCoV is a single-stranded RNA virus classified into two genetic subtypes, FCoV-I and FCoV-II. Within each subtype, there are two distinct pathotypes: feline enteric coronavirus (FECV) and feline infectious peritonitis virus (FIPV). FIPV is the causative agent of feline infectious peritonitis (FIP), which was invariably fatal until the recent introduction of effective antiviral drug treatments [13]. Repurposed human antivirals, GS-441524 (a metabolite of remdesivir) and GC-5734, are now available for the treatment of FCoV infections and FIP [14,16].

The virus is transmitted via the faecal–oral route [17], with the primary source of infection being cat litter trays. Cats become infected with FECV, and, according to the widely accepted internal mutation theory, in a proportion of infected cats, the virus mutates into FIPV, which produces a tropism switch and the subsequent development of FIP [18,20]. Although several mutation sites have been implicated, no single causative mutation has been identified, suggesting FIP is likely due to a combination of mutations working in concert [3,21]. However, direct transmission of FIPV has been shown to occur [22,23], known as the circulating virulent/avirulent theory. This direct transmission has been the cause of previous small epizootic incidents, and this, coupled with the recombination with canine coronavirus, was understood to result in the rapid spread of a recent FCoV outbreak in Cyprus in 2023 [24,27].

FCoV can be effectively cleared by the immune response in infected animals. Following infection, neutralizing antibodies (nAb) are produced, with the majority of nAbs targeting the viral spike (S) protein that is exposed on the surface of the virion and facilitates viral entry into host cells. The FCoV S protein comprises three domains: the large ectodomain, a transmembrane domain and a small endodomain. The large ectodomain contains the S1 and S2 subunits responsible for host-cell receptor recognition [28,29] and membrane fusion [30]. The receptor-binding domain (RBD) is formed by the S1 subunit, which consists of domains 0, A, B, C and D; the S2 subunit forms the stalk of the spike protein, which fuses with the viral envelope. Additionally, a role for domain 0 of S1 in the development of FIPV has been proposed [26,31].

In the UK, although FCoV testing is widespread, there has been no recent large-scale serosurvey of the prevalence of infection, which would determine exposure to the virus and inform the development of preventative measures. However, large-scale serosurveys for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, a betacoronavirus, have been conducted in the UK cat population [32]. Together, these studies highlight that both FCoV and SARS-CoV-2 regularly infect cats [2,3, 32]. Antigenic cross-reactivity between FCoV and other alphacoronaviruses has been observed previously; recent studies also report antigenic cross-reactivity between alpha- and betacoronaviruses, such as between FCoV and SARS-CoV [33], as well as SARS-CoV-2 [34,35].

The recent emergence of the recombinant and highly pathogenic FCoV-23 strain, together with the coronavirus disease 2019 (COVID-19) pandemic, has highlighted the importance of investigating the potential for cross-reactivity between viruses. Thus, this study examined cross-reactivity and cross-neutralization between FCoV and SARS-CoV-2 within the UK domestic cat population.

Methods

Samples

Serum and fluid samples were provided by the University of Glasgow Veterinary Diagnostic Service (VDS). These were remnant samples from a mixture of sick and healthy cats submitted by UK veterinary clinics for diagnostic testing, which were heat-inactivated at 56 ℃ for 30 min and stored at 4 ℃ and −80 ℃ until required. Sampling dates ranged from April 2020 to October 2023. The study was granted ethical approval by the University of Glasgow Veterinary Ethics Committee (EA27/20). Sample metadata were collected from the VDS Laboratory Information Management System (LIMS) for data analysis. Each sample was labelled with a unique identifier barcode by the VDS; this was used to match the metadata from the LIMS.

Pseudotype-based virus neutralization assay

To detect the presence of FCoV S protein-specific neutralizing antibodies, a pseudotype-based virus neutralization assay (PVNA) was conducted. This method has been described previously in the context of SARS-CoV-2 [32,36]. In summary, human embryonic kidney 293T (HEK-293T) and feline embryo-A (FEA) [37] cells were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) (1X, Thermo Fisher Scientific, Paisley, UK) supplemented with 10% FBS (Thermo Fisher Scientific, Paisley, UK), 1% l-glutamine (200 mM, Thermo Fisher Scientific, Paisley, UK) and 1% penicillin-streptomycin (Thermo Fisher Scientific, Paisley, UK) (complete DMEM) and kept at 37 ℃ in an atmosphere of 5% CO2 (37℃/5% CO₂) until required.

HEK-293T cells were transfected with codon-optimized FCoV spike plasmids corresponding to FIPV-79–1146 (GenBank: DQ010921) and FIPV-DF-2 (GenBank: JQ408981). The FCoV-23 (long) sequence was a consensus sequence corresponding most closely to GenBank: PQ133182 with the following amino acid mutations: Y89D, T206N, S237F, L253F, H259Y, M272V, I387S, H389Y, V671A, F810S, F885V, S928T, N961K, D984E, H1005N, S1284I, A1295V, S1307G, K1324E and N1444STOP. The FCoV-23 (short) sequence was a consensus sequence corresponding most closely to GenBank: PQ133179 with the following amino acid mutations: D52G, M68V, ΔY53-A60, S191P, L543Q, S699N, D780E, R979Q, Y1043D, S1080I, S1103G, I1127T and N1240STOP. Constructs were synthesized by GenScript (Rijswijk, the Netherlands). Spike plasmids were co-transfected with p8.91 (HIV gag-pol) and pCSFLW (firefly luciferase) in the presence of polyethyleneimine (Polysciences, Warrington, PA, USA). The transfected cells were incubated at 37 °C/5% CO₂ for 48 h, then filtered through a 0.45-µm filter, aliquoted and stored at −80 ℃ until use. To determine the concentration of pseudotype virus to use in assays, pseudotype viruses were diluted serially from neat to 1 in 64 in complete DMEM, plated onto FEA cells and incubated for 48 h at 37 °C/5% CO₂. Post-incubation, luciferase activity was measured using an EnSight luminometer (Revvity, Abingdon, Oxfordshire, UK) following the addition of Steadylite substrate (Revvity, Abingdon, Oxfordshire, UK). The appropriate pseudotype dilutions to use in assays were determined as the highest dilution with luciferase activity greater than or equal to 2×10⁶ counts per second.

Firstly, heat-inactivated samples were screened at a fixed dilution, where sera were diluted 1 in 50 in complete DMEM in duplicate for each pseudotype tested, alongside a known positive control (pooled FCoV positive serum samples) and a ‘no serum’ negative control. Sera were then incubated with viral pseudotype at 37 °C/5% CO₂ for 1 h, plated onto FEA cells and incubated at 37 °C/5% CO₂ for 48 h. After 48 h, results were read to determine luciferase activity using a Perkin Elmer EnSight luminometer, following the addition of Steadylite substrate. Results that showed at least an 80% reduction in luciferase activity when compared with the no serum control were classed as positive; those with over 20–50% inhibition of luciferase activity were classed as equivocal and followed up with titration to eliminate false negative results. The cut-off threshold for positive results was determined using a receiver operating characteristic curve (Fig. S5, available in the online Supplementary Material).

FCoV nAb activity was quantified for those sera that tested positive or equivocal at fixed dilution by serial dilution of samples between 1 in 50 and 1 in 36,450 in complete DMEM, alongside a ‘no serum’ control. As previously, diluted samples were incubated with pseudotypes in triplicate, plated onto FEA cells and incubated for 48 h at 37 °C/5% CO₂. After incubation, luciferase activity was measured using an EnSight luminometer with the addition of Steadylite. Antibody titres were estimated by determining the dilution at which the mean infectivity was reduced to 20% of the value of the ‘no serum’ control.

Immunofluorescence

Immunofluorescence assays (IFA) were carried out by the VDS, a commercial diagnostic laboratory at the University of Glasgow, as part of routine screening prior to this study. Data on IFA results were provided as part of the metadata and were analysed in relation to this study’s findings.

Briefly, IFAs were completed according to a previously described protocol by Addie and Jarrett [38], the assay detecting all antibodies against FCoV. Samples with titres of 10 or below were determined to be seronegative, whereas those with titres of 20 or over were seropositive. Titres of over 1,280 were classed as >1,280.

Immunoblotting assay

To investigate the antibody response against FCoV, purified virus stocks were prepared from an FCoV-II virus (FCoV-VDS) provided by the University of Glasgow VDS and an FCoV-I virus (FIPV-Black (TN406HP) [29] kindly provided by Herman F. Egberink (Utrecht University)). Briefly, Crandell–Rees feline kidney (CRFK) cells, maintained in complete DMEM, were infected with each virus at a multiplicity of infection of 0.1 for an hour at 37 °C/5% CO₂, and then the cells were aspirated and overlaid with complete DMEM and incubated for 30 h. Virus supernatant was then collected and purified. Purified virus was prepared as follows: virus supernatants were pelleted through a 20% sucrose cushion by ultracentrifugation (28,000 g for 2 h at 4 ℃). The resulting pellets were resuspended in PBS, and the final virus stocks were stored at −80 ℃ until required.

To determine the concentration of virus to use in immunoblot serology assays, purified virus stocks were diluted 1 in 2, 1 in 5 and 1 in 10 in 2X Laemmli buffer (Merck) and run on a 4–12% Invitrogen NuPAGE Bis-Tris-MES gel by electrophoresis to visualize viral proteins. The dilution that produced clearly defined bands for the three FCoV viral proteins (spike, nucleoprotein and membrane protein) was then used to make nitrocellulose membranes for the serology assay, as detailed below.

Nitrocellulose membranes for immunoblotting were prepared by separating purified FCoV virus via electrophoresis using either a 4–12% Invitrogen NuPAGE Bis-Tris-MES or a 3–8% Invitrogen NuPAGE Tris-Acetate pre-cast gel (Thermo Fisher Scientific, Paisley, UK), which was transferred by blotting onto nitrocellulose membranes. Following transfer, membranes were blocked by incubation overnight at 4 ℃ in the presence of milk-PBS/Tween (1X PBS with 0.05% Tween 20).

To screen samples, 2-mm strips were cut from FCoV nitrocellulose membranes, and membranes were probed with sera at a 1 in 10 dilution in 1% milk in PBS/Tween, along with 300 mM EDTA. Sera were incubated overnight at 4 ℃ with continuous shaking. Membranes were washed with PBS/Tween and then incubated with 1.5 mg ml⁻¹ biotinylated anti-cat IgG (Vector Laboratories, Peterborough, UK) in 1% milk in PBS/Tween and subsequently 1 mg ml⁻¹ streptavidin peroxidase (Sigma-Aldrich, Burlington, MA, USA) in 1% milk in PBS/Tween. Biotinylated anti-cat IgG and streptavidin peroxidase were both incubated for 2 h at room temperature with continuous shaking, and strips were washed in PBS/Tween after each step. Membrane strips were developed by the addition of HRP chromogenic substrate and placing them on a shaker until distinct bands appeared. The HRP chromogenic substrate was prepared by adding 5 ml of Solution A (25 ml PBS (1X, Sigma-Aldrich, Burlington, MA, USA) and 0.03 g 4-chloro-naphthol (97%, Sigma-Aldrich, Burlington, MA, USA) to 10 ml of Solution B (10 ml methanol (99.8%) and 15 µl hydrogen peroxide (30 wt% in H₂O, Sigma-Aldrich, Burlington, MA, USA). FCoV positive results were defined as those that displayed three virus-specific bands: one in the 43–50 kDa region and two in the 25–35 kDa region.

Data analysis

Data were cleaned by removing duplicate samples; where more than one sample per animal was tested, the earliest sample was selected for analysis, and where results did not match, the positive result was selected. Individual cats were categorized as positive if their serum tested positive for at least one pseudotype on PVNA (neutralization titre ≥50). For comparisons with SARS-CoV-2 serology results, data were gathered from previously published material [32]; the highest titre elicited against both viruses by each sample was taken and used for analysis.

Testing was blinded until final analysis; at this stage, metadata were provided by the VDS and were used to categorize samples into the following:

• Sex groups: male, female and cats of unknown sex

• Breed groups: pedigree cats (British Shorthair, Ragdoll, Maine Coon etc.), non-pedigree cats (domestic shorthair, domestic longhair etc.) and cats of unknown breed

• Age groups: kitten (0–1 years), junior (1–2 years), prime (3–6 years), mature (7–10 years), senior (11–14 years), super senior (15+ years) and cats of unknown age

Analysis and preparation of figures were carried out in Microsoft Excel (version 2408), R (version 4.5.1) and BioRender. Differences between nAb seroprevalence and categorical data were assessed using chi-squared tests and Fisher’s exact tests. Differences in nAb titres were assessed using a Mann–Whitney U test and a Kruskal–Wallis test. Where pairwise comparisons were required between multiple groups, a Mann–Whitney U test was conducted, and the Bonferroni correction was applied for multiple comparisons. Regression and correlation were assessed using linear regression models (Pearson).

Results

Cohort description and sampling

FCoV nAb seroprevalence was measured in residual cat sera (n=1,117), collected from domestic UK pet and shelter cats between April 2020 and October 2023, which were previously tested for SARS-CoV-2 [32]. Samples were provided by the University of Glasgow’s Veterinary Diagnostic Service; these were submitted from practices throughout the UK, with sizeable clusters discernible around London, Glasgow, Blackpool and Cambridgeshire as shown in Fig. 1(a).

Fig. 1. Distribution of the screened pet cat population within the UK. Sera from cats submitted to veterinary practices around the UK were collected between April 2020 and October 2023 (n=1117) (a). The size of circles over locations is proportional to the number of samples submitted at that postcode. The circle appearing in the middle of the English Channel indicates samples collected from the island of Jersey. (b) Sampled cats were characterized based on three common risk factors, sex, breed and age, using metadata.

The sample population comprised 56.3% males (n=629), 38.2% females (n=427) and 5.5% with no recorded sex (n=61). Cats were classified into six age groups based on the American Association of Feline Practitioners’ guidelines [39,40], namely kitten (0–1 years), junior (1–2 years), prime (3–6 years), mature (7–10 years), senior (11–14 years), super senior (15+years) and those with no recorded age. The sample population was biassed towards younger age groups, with approximately half (46.6%, n=520) being kittens. Older age groups comprised ~13.2% (n=148) of the population, with 10.7% (n=120) and 2.5% (n=28) belonging to senior and super senior groups, respectively. With respect to breed, 33.2% (n=371) of the population were pedigree cats. British Shorthair, Maine Coon and Ragdoll were the most common breeds, while non-pedigree cats made up 55.9% (n=625) of the population, and 0.9% (n=121) had no recorded breed. Maine Coon and Ragdoll cats were overrepresented at 6.3% and 5.9% in the tested population (Fig. 1b), compared to the UK statistics of 2% and 3%, respectively [41]. However, the overall pedigree-to-non-pedigree ratio in this study aligned with current UK statistics [8,41].

FCoV nAb seroprevalence was similar across FCoV-II isolates

To assess FCoV-II nAb seroprevalence across our cohort, we produced HIV(FCoV) pseudotype, bearing the Spike protein from two FCoV-II strains, namely the US isolate, FIPV-79–1146 and the vaccine strain, FIPV-DF2, and carrying a firefly luciferase marker gene. Serum samples were screened by PVNA to investigate differences in nAb seroprevalence between strains FIPV-79-1146 and FIPV-DF2. These assays revealed no significant variation in nAb seroprevalence for the two strains, with an overall nAb seroprevalence of 13.6% (n=152); individual nAb seroprevalence by strain was 11.5% and 12.4% for FIPV-79-1146 and FIPV-DF2, respectively (χ² (1)=0.43, P=0.51, chi-squared).

A subset of samples (n=663) had already been screened for FCoV-II antibodies by IFA during routine diagnostic testing, using a method described previously [38], revealing 395 (59.5%) FCoV seropositive individuals (Fig. S1). This subset of samples contained 88 FCoV nAb seropositive individuals, of which only 56 tested positive by IFA (Fig. S2).

Risk factors showed no impact on FCoV nAb seroprevalence

The impact of risk factors such as sex, age and breed was evaluated (Fig. 2). Although the sample population was biassed towards younger males, none of the assessed risk factors had a significant impact on FCoV nAb seroprevalence. While the number of seropositive kittens was significantly higher (OR=2.31 (95% CI: 1.022–6.12), P=0.038, Fisher’s exact) compared to the junior age group, there was no further evidence for statistically significant differences in nAb seroprevalence between age groups; grouping age groups to compare younger (Kitten, Junior, Prime) and older (Mature, Senior, Super Senior) cats did not show significant differences in FCoV nAb seroprevalence either (χ² (1)=0.0026, P=0.96, chi-squared). Similarly, comparing groups of pedigree and non-pedigree cats showed that breed had no significant impact on FCoV nAb seropositivity (χ² (1)=0.013, P=0.91, chi-squared). However, comparisons of nAb seroprevalence between sex groups approached significance (χ² (1)=2.77, P=0.096, chi-squared).

Fig. 2. Commonly reported risk factors to FCoV infection and FIP development showed no impact on nAb seroprevalence to FCoV. FCoV nAb seroprevalence was measured by PVNA (n=1,117); positive results were those that had a titre over 50. Results were analysed by risk factors, showing that (a) there was no significant variation in nAb seroprevalence between sex groups (n=1,056). (b) There was no significant impact of breed group on seropositivity (n=976). (c) There were significantly more seropositive kittens when compared to junior cats, but no differences were found between other age groups (n=1,116). Significance is denoted as follows: *P<0.05, ns, no significance.

FCoV nAb titres showed cross-reactivity between FCoV-II strains FIPV-79-1146 and FIPV-DF2

To further characterize the antibody response to FCoV, samples that tested positive at a 1 in 50 dilution were titrated to compare the strength of the nAb response against both strains. Overall, results showed a broad correlation (r (150)=0.69, 95% CI: 0.60–0.77, P<0.001, Pearson) between the nAb titres elicited against FIPV79-1146 and FIPV-DF2, with no significant differences when comparing titres (W=11086, 95% CI: −16.00–6.99, P=0.54, Mann–Whitney U test) (Fig. 3). Although some cats showed nAb against only one strain [8.5% (13/152) FIPV-DF2 positive only, 15.8% (24/152) FIPV-79–1146 only], the majority neutralized both [75.7% (115/152)], suggesting that FCoV-II strains can reliably induce cross-neutralizing antibody responses against other FCoV-II strains. An amino acid alignment comparing the S1/S2 and RBD region of FIPV-DF2 and FIPV-79–1146 showed that of the seven amino acid differences between the Spike proteins, three were localized to the RBD region (V457A, Y531H and H582P from FIPV-79-1146 to FIPV-DF2, Figs S3 and S4) and may have contributed to the strain-specific neutralization detected in a minority of cats.

Fig. 3. FCoV nAb titres elicited against FIPV-79–1146 and FIPV-DF-2 showed close correlation. FCoV nAb titres were measured by PVNA for two FCoV-II strains, FIPV-79-1146 and FIPV-DF-2 (n=1,117). Positive samples (n=152) were those that presented with a titre over 50. The graph shows FCoV nAb titres for both strains screened, with dotted lines depicting the cutoff for positive results. Background colours are assigned for the groups of samples: samples positive for both strains tested in light purple, positive only for FIPV-79–1146 in pink, positive only for FIPV-DF-2 in dark purple and negative for both in peach. Samples were also tested for FCoV-II by immunofluorescence at the University of Glasgow VDS are coloured pink. A linear regression line was fitted (in red) to investigate the correlation of the nAb titres between the two strains tested (r=0.69, Pearson).

Risk factors had a limited impact on FCoV nAb titres

The effects of sex, breed and age on the FCoV nAb titres were examined. Cats were termed FIPV-79–1146- or FIPV-DF2-dominant based on the strain against which they elicited a stronger response, and the impact of the risk factors on nAb titres was determined (Fig. 4). There was no statistically significant impact of sex or breed on the FCoV nAb titre [W=1905.50 (95% CI: −20.99–20.00), P=0.89, and W=1905.50 (95% CI: −23.00–17.00), P=0.60, respectively, Mann–Whitney U test], suggesting that these factors do not impact the nAb response elicited following FCoV infection. However, while age, similar to sex and breed, showed no significant impact on nAb titres (χ² (5)=4.63, P=0.46, Kruskal–Wallis, pairwise comparisons), cats in the super senior age group tended to have higher titres compared to cats in other age groups (P>0.05, pairwise Mann–Whitney U test).

Fig. 4. Commonly reported risk factors to FCoV had no impact on FCoV nAb titres. FCoV nAb seroprevalence was measured by PVNA (n=1,117); positive results were those that had a titre over 50. Results were analysed by risk factors, showing that (a) there was no significant variation in nAb titres between sex groups (n=145; P=0.89, Mann–Whitney U test). (b) There was no significant impact of breed group on nAb titre (n=152; P=0.6, Mann–Whitney U test). (c) Age had no significant impact on FCoV nAb titres (n=151; P=0.46, Kruskal–Wallis). Significance between groups was calculated by tests outlined above, where *P<0.05, and ns, no significance (boxes represent the 25–75th percentile, while whiskers represent the 5–95th percentile).

Correlation between titres elicited against novel FCoV-23 and other FCoV-II strains

Following the emergence of FCoV-23, the causative agent of a significant FCoV outbreak in Cyprus [24,27], a subset of samples (15/1,117) was evaluated by PVNA against FCoV-23 pseudotypes to detect neutralization and identify cross-neutralization with this recently-emerged strain. It has recently been shown that the loss of domain 0 in the FCoV Spike protein enhances viral entry and fusogenicity [26]. Therefore, to determine if domain 0 plays a role in viral neutralization, we compared serological responses against the two isotypes of the novel FCoV-23 strain, namely the FCoV-23 (short) and FCoV-23 (long) (Fig. 5). A similar magnitude of neutralizing activity and strong positive correlation was observed between the titres elicited against FCoV-23 (short) and FCoV-23 (long) [r (14)=0.99, 95% CI: 0.99–0.99, P<0.001, Pearson], consistent with strong cross-neutralization between the two forms of FCoV-23. Furthermore, when comparing neutralizing antibody responses against both FCoV-23 strains with FIPV-79-1146 and FIPV-DF2, a strong correlation was observed between all strains (r (14)=0.74–0.99, P<0.001, Pearson), indicating cross-reactivity of neutralizing responses against FCoV-II strains, including FCoV-23.

Fig. 5. FCoV nAb titres cross-react between various FCoV-II strains. FCoV nAb titres were measured by PVNA against several FCoV-II strains, including the novel FCoV-23 isotypes FCoV-23 (long) and FCoV-23 (short) (n=16). Positive samples presented with a titre over 50. The graph shows FCoV nAb titres for both strains screened, with dotted lines depicting the cutoff for positive results. FCoV nAb titres showed strong correlation when comparing the two FCoV-23 isotypes, FCoV-23 (long) and FCoV-23 (short) (a) with each other (r=0.99, Pearson), (b) with FIPV-79–1146 (r=0.75 and r=0.74 respectively, Pearson) (c) and with FIPV-DF-2 (r=0.98 and r=0.98 respectively, Pearson). Background colours are assigned for the groups of samples: samples positive for both strains tested in light pink, positive only for the strain on the X-axis in dark peach, positive only for the strain on the Y-axis in light orange and negative for both in light peach. A linear regression line was fitted (in red) to investigate the correlation of the nAb titres between the two strains tested.

FCoV neutralizing antibodies do not cross-neutralize SARS-CoV-2

Serum samples screened for FCoV were part of a larger dataset that had been tested previously for SARS-CoV-2 nAb seroprevalence [32], facilitating the comparison of nAb titres against FCoV and SARS-CoV-2 for evidence of cross-reactivity (Fig. 6). Comparison of titres elicited against both FCoV-II strains and SARS-CoV-2 showed no correlation [r (69)=−0.087, 95% CI: −0.31–0.15, P=0.47, Pearson], suggesting no evidence for cross-reactivity between these two coronaviruses.

Fig. 6. SARS-CoV-2 seropositive samples do not cross-neutralize FCoV on PVNA. FCoV seropositive samples titrated against both FCoV and SARS-CoV-2 were investigated (n=71). The percentage neutralization of the two viruses is shown on the graph; as two strains were tested for FCoV, the results for the strain that presented with the highest titre were plotted. Cut-off for positive samples is at a point of 90% and 80% reduction in neutralization activity when compared to the no serum control, for SARS-CoV-2 and FCoV, respectively; cut-off points are indicated by dashed lines. Samples are classified into four groups distinguished by different background colours: those seropositive for both viruses in pink, those seropositive for only FCoV in light purple, those seropositive for only SARS-CoV-2 in peach and those negative for both in dark purple. There was no correlation in the percentage neutralization of the two viruses (r=−0.08, Pearson).

To confirm the absence of cross-reactivity between nAbs elicited against FCoV and SARS-CoV-2, a subset of samples that tested positive for SARS-CoV-2 were tested by immunoblot analysis for antibodies recognizing FCoV-I and FCoV-II (Fig. 7). No reactivity was observed against either FCoV-I or FCoV-II. These results confirm the absence of cross-reactivity between FCoV and SARS-CoV-2 linear and conformational epitopes and indicate that, despite co-circulating in the cat population, infections with these two viruses can be differentiated using serological tests.

Fig. 7. Immunoblot analysis for cross-reactivity between FCoV and SARS-CoV-2. Samples reacted similarly on both FCoV-I and FCoV-II immunoblots. Serum samples were blotted on membranes prepared with purified FCoV strains FCoV-VDS (a) and FIPV-Black (TN406HP) (b), respectively; membrane strips were incubated with individual serum samples to detect the presence of antibodies recognizing FCoV viral proteins. Three groups of sera tested were FCoV non-neutralizing positive (lane 1), FCoV neutralizing positive (lanes 2–3) and SARS-CoV-2 neutralizing positive (lanes 4–5) samples; neutralization status was previously determined by PVNA. SARS-CoV-2 seropositive samples showed no evidence for cross-reactivity with FCoV viral proteins on either FCoV-I or FCoV-II immunoblots, as shown by the lack of bands in lanes 4–5. The expected molecular weight of FCoV viral proteins is depicted as S (spike protein), N (nucleoprotein), M (membrane protein) and E (envelope protein).

Discussion

To gain a deeper understanding of the serological response against FCoV infection in UK domestic cats, we conducted a large-scale serosurvey using PVNA. The FCoV nAb seroprevalence was 13.6% in the samples that were tested. Considering that the UK pet cat population was last estimated to be ~12.5 million [42], these results suggest that at least 1.7 million pet cats have been exposed to FCoV. These findings are in line with a previous study that reported an FCoV nAb seroprevalence of 4% in single-cat and 62% in multi-cat households in Türkiye [2]. While nAb seroprevalence provides an indication of previous infection, it is important to acknowledge that persistent FCoV infection occurs in a subset of animals, and, therefore, the true seroprevalence might be lower than estimated in this study. However, as persistent shedding only occurs in ~13% of infections [43], these individuals would not significantly impact population-level seroprevalence results.

FCoV seroprevalence has been estimated using ELISA and IFA and reported values worldwide are reported to range between 7 and 91% [3,28]. A recent meta-analysis estimated seroprevalence to be 32.4% globally [29]. In this study, FCoV IFA testing was conducted on approximately half of the cat samples that had been tested for neutralizing antibodies, resulting in an overall FCoV seroprevalence of 59.5% (n=395/663), which aligns with FCoV seroprevalence reported by other UK studies (22–88%) [3]. The differences in FCoV seroprevalence and FCoV nAb seroprevalence, estimated by IFA and PVNA tests, respectively, are consistent with the two assays measuring antibodies against different antigens (the PVNA detects antibodies targeting epitopes on the Spike protein, while the IFA detects antibodies binding all viral proteins produced during infection). These differences highlight the importance of assay selection when interpreting seroepidemiological data. In this study, Spike-specific nAbs were detected in only a fraction of FCoV-infected animals, and more cats displayed non-neutralizing antibodies that were detected by IFA. Nevertheless, a subset of FCoV nAb reactive individuals displayed a lack of IFA seropositivity; this discrepancy can be explained through the difference in sensitivity of the two tests, the PVNA being more sensitive when compared to the IFA. Accordingly, some individuals will display antibody titres on the PVNA assay, while showing no reactivity on IFA (Fig. S2).

Recent studies reported that cats experimentally infected with FCoV-I did not develop antibodies against FCoV-II strains [34,35]. In this study, we observed that the majority of samples that elicited a neutralizing antibody response to one FCoV-II strain also displayed neutralization when tested against the second FCoV-II strain (Fig. 3), with a positive correlation [r (150)=0.69, 95% CI: 0.60–0.77, P<0.001, Pearson] between the antibody responses elicited against strains FIPV-79-1146 and FIPV-DF-2. Considering that cats in this study were sampled in the UK, where FCoV-I is understood to be the predominant circulating genetic subtype, the cats tested were most likely to have been infected by an FCoV-I strain. Thus, the results presented here not only provided evidence for cross-reactivity and cross-neutralization between FCoV-II strains but also between FCoV genetic subtypes, which was further supported by the results of immunoblotting assays (Fig. 7). Nevertheless, since FCoV-II strains have originated from multiple recombination events between FCoV-I and canine coronaviruses, different FCoV-II strains could have variable spike protein sequences; this could explain the differences between our findings compared to those of previous studies, suggesting that while there is cross-reactivity between some strains, this may be strain-specific. Indeed, we observed a small subset of seropositive samples that recognized only one of the strains tested, as shown in Fig. 3. A key molecular difference between the two FCoV-II strains used in this study (namely FIPV-FD2 and FIPV-79–1146) lies within their RBD, where three amino acid residues differ (Figs S3 and S4). These sequence differences might contribute to subtle structural alterations within the RBD, potentially influencing antigenicity and susceptibility to neutralization. This finding highlights the importance of mapping specific epitopes involved in cross-neutralization across FCoV strains.

We also examined how demographic factors, namely sex, age and breed, shaped serological responses to FCoV. These factors are commonly cited as risk factors associated with FCoV infection and the development of FIP. Sexually intact younger male cats [4,6] are reported to show an increased predisposition to FIP, suggesting that both neutering status and sex are important risk factors. However, the results of the present study found no evidence of differences in either nAb seroprevalence or the magnitude of titres elicited against FCoV between male and female cats, similar to previous studies by Pedersen et al. [7] and Yin et al. [44], who reported no evidence that sex or neutering status predisposed animals to FCoV infection. Differences in cat populations, housing conditions and geographical location are all factors likely to influence studies such as these and could explain differences observed across studies. Additionally, the study setup, for example, whether the study was conducted in a shelter, real world, laboratory [7] or a veterinary setting [5,6, 8, 44], might influence findings.

Kittens showed higher nAb seroprevalence against FCoV compared to the junior age group (Fig. 2c); however, this trend was not reflected in the magnitude of nAb titres (Fig. 4c). These findings align with previously published studies that observed FCoV seroprevalence decreasing with age [9], with 80% of infections occurring in cats under the age of 2 and 50% in kittens under 7 months of age [7]. Additionally, more recent studies reported that FCoV RNA in samples was associated with kittens being under the age of 1 year [10,11]. These observations can be explained by an immature immune system in kittens and differences in housing conditions, as kittens are more likely to be housed in multi-cat housing at breeders' premises or shelters [12], which increases the risk of viral transmission.

Many studies have reported that pedigree cats are more likely to be infected with FCoV and are more susceptible to developing FIP [12,45] compared to non-pedigree cats, with risk factors such as housing, hygiene and potential for host genetic factors influencing susceptibility. In this study, breed did not appear to be a risk factor as there was no difference in nAb seroprevalence or titres against FCoV between pedigree and non-pedigree cats. It is important to interpret these findings with caution, since our sample set likely reflects a population biassed towards sick animals, since animals with clinical signs are more likely to be sampled by veterinary practices. This factor, as well as the overrepresentation of younger male cats, could have introduced an element of bias to the study. Therefore, assessing these factors in a larger, randomly sampled population and including factors such as housing, which this study was unable to do, would be required to assess the true impact of these factors on FCoV infection and FIP.

To evaluate the serological responses elicited to the novel FCoV-23 isotypes (long and short) and compare these with the results against other older FCoV-II strains, PVNAs were carried out using viral pseudotypes bearing either the FCoV-23 (long) or the FCoV-23 (short) Spike proteins. There was a very strong correlation [r (14)=0.99, 95% CI: 0.99–0.99, P<0.001, Pearson] in nAb titres elicited against the two FCoV-23 isotypes (Fig. 5a), suggesting that although the loss of domain 0 has previously been implicated in enhancement of viral entry and viral fusogenicity [26], it is not a major target for neutralizing antibodies. Furthermore, neutralization of the two FCoV-23 pseudotypes was strongly correlated with that of FIPV-79-1146 and FIPV-DF-2 (Fig. 5b and c). This supported the hypothesis that cross-reactivity is present between FCoV-II strains, regardless of geographical location and whether strains are novel or historic, while also indicating that UK cats are likely to have nAb against FCoV-23 infection given the strong cross-neutralization between FCoV strains. Overall, these results, along with the recent FCoV-23 outbreak, highlight the importance of monitoring not just FCoV-I, but all FCoV strains and infections, as due to the increase in animal movement, such as through international adoptions, the geographical distribution of strains might be changing.

Finally, since cats are regularly exposed to FCoVs and SARS-CoV-2, the potential for cross-reactivity between the two viruses was investigated. There was no correlation between the nAb titres elicited against the two viruses in the tested animals, indicating a lack of cross-reactivity (Fig. 6). This was further confirmed by immunoblot analysis (Fig. 7), which showed a lack of reactivity between both FCoV-I and FCoV-II viral proteins and SARS-CoV-2 seropositive samples. This lack of cross-reaction suggests that single antigen pan-coronavirus vaccine candidates [35] might not be effective in improving current prevention strategies against FCoV, as they are unlikely to offer cross-protection between different coronavirus genera, such as Alpha- and Betacoronaviruses.

Abbreviations

COVID-19

coronavirus disease 2019

CRFK

Crandell-Rees feline

DMEM

Dulbecco’s Modified Eagle’s Medium

FCoV

feline coronavirus

FEA

feline embryo-A cells

FECV

feline enteric coronavirus

FIP

feline infectious peritonitis

FIPV

feline infectious peritonitis virus

HEK-293T

human embryonic kidney 293T

IFA

immunofluorescence assays

LIMS

Laboratory Information Management System

nAb

neutralizing antibody

PVNA

pseudotype-based virus neutralization assay

RBD

receptor-binding domain

S

spike protein

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

VDS

Veterinary Diagnostic Service