Serological Study of SARS-CoV-2 Antibodies in Japanese Cats: Analysis of Risk Factors Among Cat Lifestyles

Little is known about the epidemic status of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in cats living in Japan, and about the inuence of cat lifestyles on the SARS-CoV-2 infection epidemic in cats. We developed protein A/G-based ELISA, which was standardized using positive rabbit antibodies. The measured values of this ELISA were consistent with those of conventional anti-feline IgG-based ELISA. We then collected blood samples from 1,969 cats that visited veterinary clinics in Japan from June to July 2020 and examined for the presence of anti-SARS-CoV-2 antibodies. Using protein A/G-based ELISA, nine cats were found to have SARS-CoV-2 S1-specic IgG, among which four had RBD-specic IgG. Among those nine samples, one showed neutralizing activity. Based on these, we estimated that the prevalence of SARS-CoV-2 neutralizing antibodies in cats living in Japan was 0.05% (1/1,969 samples). This prevalence did not differ much from the prevalence of neutralizing antibodies for SARS-CoV-2 in humans conducted in Japan at that time. Furthermore, we searched for factors associated with the prevalence of neutralizing antibodies in cats using our data and information from ve countries (China, Croatia, France, Germany, and Italy). The prevalence of SARS-CoV-2 in cats was correlated with the rate of keeping indoor-only. Correlations of the seroprevalence of SARS-CoV-2 antibodies in cats with factors associated with cat-rearing styles were analyzed using the Kolmogorov–Smirnov normality test, followed by Spearman’s correlation coecient test. The prevalence of anti-SARS-CoV-2 IgG in cats and humans was compared using Fisher’s exact test. Statistical signicance was set at p < 0.05. All statistical analyses were performed using GraphPad Prism 9.0 (GraphPad Software, Inc., San Diego, CA, USA).

Wuhan, China. Referring to the cumulative COVID-19 cases per 1,000 people in Japan and Wuhan, China, at that time from the World Health Organization database (accessed on May 25, 2020), the levels of SARS-CoV-2 infection in cats were assumed to be similar in these countries. Thus, the assumed seroprevalence in Japan was tentatively set to ca. 3.0% for the sample size calculation. Based on this seroprevalence assumption in Japan, the sample size was calculated to be 1,562 cases (α error = 0.05, β error = 0.80) using goodness-of-t tests. Considering the loss of samples because of a lack of information (i.e., clinical history and cat lifestyle), 1,969 cat blood samples were collected from June 1, 2020, to July 31, 2020, in this study (Table 1).
Samples were collected with owners' consent by veterinarians at 101 animal hospitals in Japan from June 1, 2020, to July 31, 2020. The sampling period was set at 2 months after the outbreak of the virus, when the second wave of infections occurred in Japan (Additional le 1: Supplementary  Figure 1). The correlation of IgG reactivity to the S1 protein, as assessed by protein-A/G based ELISA, with that assessed via conventional ELISA using anti-feline IgG conjugated with horseradish peroxidase (anti-feline-IgG based ELISA) was examined. We examined 34 samples collected during the COVID-19 pandemic period, including 9 positive and 25 negative samples in the protein-A/G-based ELISA, detecting the anti-S1-protein IgG antibody, and 162 negative control samples ( Figure 1). The comparison of the IgG reactivity measured by protein-A/G-based ELISA with that measured by anti-feline-IgGbased ELISA revealed the presence of a strong correlation (r 2 = 0.83 and p < 0.0001).
Seroprevalence of anti-SARS-CoV-2 antibodies in cats in Japan The screening of the 1,969 samples by ELISA to S1 protein identi ed nine samples with IgG reactivity (seroprevalence: 0.46%) (Figure 2a). In addition, ELISA of these nine positive samples to detect the anti-RBD-protein IgG antibody revealed four samples with IgG reactivity (seroprevalence: 0.20%) ( Figure 2b). Comparing IgG reactivity to the S1 protein with that to the RBD protein showed that these four samples had high IgG reactivity to both proteins ( Figure 2c).
Because ELISA can produce false-positive results owing to IgG cross-reactivity among coronaviruses [14], the neutralization activity of these nine positive samples to SARS-CoV-2 in the in vitro setting was determined. One sample neutralized SARS-CoV-2, which was diluted to 1:80, whereas the remaining eight samples (diluted up to 1:20) did not. It was recorded for the sample that neutralized SARS-CoV-2 that the cat had no respiratory or gastrointestinal symptoms or fever in the previous 3 months. It was kept indoors only and had no history of escape from the house (Table 2).
Accordingly, the seroprevalence of SARS-CoV-2 antibodies in cats in Japan was assumed to be 0.05% (1/1,969 samples).

Cat-lifestyles in Japan
The cat-lifestyle in Japan was investigated because the frequency of interaction between cats and people can be determined by the owner. The factors associated with cat-lifestyle included living environment (i.e., housing-indoors-only, free-access-outside, and living-outdoors styles), neutering, and multiple animal holdings (Table 1). In Japan, the housing-indoors-only style was ca. 88%, representing the major living environment for cats. The ratio of neutering was ca. 80%. The rate of housing multiple animals was ca. 40%.

Analysis of factors associated with cat-lifestyles with seroprevalence in cats in six countries
The ratios of factors of cat-lifestyle also varied among ve countries (China, Croatia, France, Germany, and Italy) [21][22][23][24][25][26][27]. Estimating the important factor of cat-lifestyle, we statistically analyzed the correlation of the factors of cat-lifestyle with seroprevalence in cats among these -countries (Table  3). Comparing the factors of cat-lifestyle, the ratio of housing-indoors-only style and that of neutering in Japan were the highest among them. On the other hand, the ratio of housing multiple animals in Japan was lower than in Germany and Italy and higher than in China. The factors included not only housing indoor only, free access outside, and living outdoors in the living environment but also neutering and multiple animal holdings. The ratio of housing-indoors-only alone was negatively correlated with viral seroprevalence (r = -0.900, p = 0.042); other factors were not. Considering the epidemic situation of COVID-19 in humans, the seroprevalence in cats was positively correlated with the ratio of living outdoors (r = 0.996, p = 0.042); other factors were not.

Discussion
In the present study, to analyze the epidemic situation of SARS-CoV-2 in cats, we collected blood samples from 1,969 cats that visited veterinary clinics in Japan in June and July 2020 to determine the presence of SARS-CoV-2-speci c antibodies. Of these samples, one showed neutralizing activity (seroprevalence: 0.05%, 1/1,969 samples), which was not signi cantly different from the concurrent human seropositivity at that time [15]. This suggests that the prevalence of SARS-CoV-2 infection in cats may be related to the COVID-19 epidemic status in humans. We next surveyed the cat lifestyle from owners of the 1,969 cats by questionnaire. We added information from ve countries (China, Croatia, France, Germany, and Italy) where the prevalence of neutralizing antibodies had already been reported, and searched for factors associated with the prevalence of neutralizing antibodies in cats. The percentage of cats living indoors-only and outdoors correlated with the prevalence of SARS-CoV-2 neutralizing antibodies in cats.
We performed ELISAs for S1 and RBD of SARS-CoV-2 followed by a neutralization test. Although sero-epidemiological studies targeting the nucleocapsid of SARS-CoV-2 have been conducted [29,30], the nucleocapsid-protein-speci c IgG of cats infected with SARS-CoV-2 cross-reacted with feline coronavirus (FCoV) and other coronaviruses [30,31]. By contrast, S1 has a unique structure in each coronavirus [32], and feline S1-speci c IgG infected derived from FCoV, Feline infectious peritonitis virus, or SARS-CoV-2 barely cross-reacted with other coronaviruses [14]. RBD-speci c IgG of SARS-CoV-2 also barely cross-reacted with RBD of feline coronaviruses [20]. In addition, recent analysis of 15 different antibody titers in the blood of COVID-19 patients revealed in humans that measurement of IgG against spike proteins, especially S1 and RBD, correlates with the presence of neutralizing antibodies and is an excellent indicator of past infection [33]. Because identi cation of the presence of neutralizing antibodies requires special laboratories equipped with biosafety level 3 facilities and involves the risk of analyzing multiple samples, screening for the presence of S1-and RBD-speci c IgG and then con rming the presence of neutralizing antibodies in IgG-bearing samples are currently considered to be an appropriate criterion standard.
ELISA to determine the presence of feline anti-SARS-CoV-2 IgG in previous reports used anti-feline IgG antibodies [13,16,17,19,20]. However, as of September 30, 2021, no feline anti-SARS-CoV-2 IgG antibodies are commercially available. Because of the general unavailability of controls, the results of each ELISA cannot be compared across publications. An ELISA system using Protein A/G has recently been used to investigate infectious diseases across animal species [34,35]. Because SARS-CoV-2 is spreading worldwide, it is important to compare the results across species. The detection system developed in the present study can generalize antibody titers because rabbit anti-SARS-CoV-2 antibodies are commercially available. Future research using this analysis system is warranted.
Similarities in seroprevalence between humans and cats have been observed in Japan. This suggests that the opportunity for SARS-CoV-2 infection in cats is in uenced by the extent of COVID-19 prevalence in humans. Cats kept by COVID-19 patients or people with a history of COVID-19 have a high probability of being infected with SARS-CoV-2 [19,36]. The prevalence obtained in the present study is supported by these reports.
Changes in our lifestyle, represented by social distancing and staying at home, have reduced the spread of COVID-19 [37,38]. The behavior of cats is largely determined by their owners. Therefore, we suspected that cat lifestyle is involved in the prevalence of SARS-CoV-2 infection in cats. Because of the small number of positive samples in this study, we examined antibody prevalence and cat lifestyles in countries that had reported the prevalence of SARS-CoV-2 neutralizing antibodies in cats. We consider that whether cats were kept indoors or outdoors may have played a role in the epidemic. Cats kept outdoors may not only form their own feline social networks, but may also be casually petted by people passing. As we have found similarly, previous reports indicate that cats are often asymptomatic even when they are affected by SARS-CoV-2 [8]. Therefore, infected outdoor cats without clinical symptoms can transmit SARS-CoV-2 to cats and to humans by contact and droplet transmission [39]. For cats housed indoors, the public maintains a physical distance, whereas cats housed outdoors are not considered to maintain an appropriate distance from the public. Therefore, housing cats indoors is considered one of the most effective preventive measures to control the spread of SARS-CoV-2 between cats and humans.
There are several limitations to the present study. First, due to ethical constraints, we were unable to interview owners about their history of COVID-19.
Therefore, it remains unclear to what extent transmission from cats carrying neutralizing antibodies to SARS-CoV-2 to humans occurs. In addition, because of the limited number of positive samples we obtained, we compared our results with previous reports. Therefore, we expected that there would be variations in the survey methods, and we were not able to analyze directly the relationship between cat housing methods and the spread of SARS-CoV-2 in cats. It is hoped that future research will ll in the gaps in our analysis.
As of July 2020, 0.05% of cats in Japan had a history of infection with SARS-CoV-2, which was similar to the epidemic situation to humans. Because cats are animals that live near humans, the infection status of SARS-CoV-2 is considered to be linked to that in humans. Therefore, follow-up of the epidemic status of SARS-CoV-2 in cats is warranted. Factors that have been implicated in the prevalence of SARS-CoV-2 in cats are whether the cats are housed indoors or outdoors. The epidemic of SARS-CoV-2 infection in cats may be controlled by the efforts of cat owners.

Methods
Collection of samples with clinical history and data pertaining to cat-rearing styles According to a literature review on May 25, 2020, the seroprevalence in cats was assumed based on the lower limit of a 99.9% con dence interval (CI).
Samples were collected with owners' consent by veterinarians at 101 animal hospitals in Japan from June 1, 2020, to July 31, 2020. The sampling period was set at 2 months after the outbreak of the virus, when the second wave of infections occurred in Japan. Veterinarians completed a questionnaire on the clinical history of the animals within 3 months of blood collection. Cat-rearing-style information, such as living environment, neutering, and multiple animal holdings, was surveyed from cat owners visiting animal hospitals. Living-environment information was categorized into three types: hosing-indoors-only, free-access-outside, and living-outdoors (i.e., stray cats and feral cats) styles. Moreover, as negative controls, we also used cat blood samples, which were collected from January 1, 2015, to March 31, 2015, for the Azabu University bioresource banking project. This study was carried out in compliance with the ARRIVE guidelines and, was approved by the Animal Ethics Committee of Azabu University (No. 210407-7).
Data collection of seroprevalence and cat-rearing style in countries by reviewing the relevant literature and cumulative cases of COVID-19 per 1,000 people The data pertaining to the seroprevalence of SARS-CoV-2 antibodies in cats, which were obtained based on a neutralizing antibody assay, were collected by a systematic literature search of PubMed databases using various combinations of the keywords "epidemiology", "prevalence", "surveillance", "survey", "SARS-CoV-2", "feline", "ELISA", "IgG", "neutralizing", and "cat" published between May 1, 2020, and April 31, 2021. The standard protocol and checklists of the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) were followed [41]. The data pertaining to the cat-rearing styles were collected based on a literature search of PubMed databases using various combinations of the keywords "lifestyle", "rearing", "environment", "surveillance", and "cat", and the name of each country published between January 1, 2015, and April 31, 2021. The inclusion criteria were de ned as using over 50 samples in the surveillance and targeted toward the rate of multiple animal holdings, neutering rate, and living environment. Of note, the investigation period could not be restricted to 2020 because of a lack of information. The epidemic situation of COVID-19 in humans in each country was based on the reported cumulative COVID-19 cases per 1,000 people from data from the World Health Organization database (https://covid19.who.int/).

Statistical analysis
The CI of seroprevalence in each report was calculated based on the total number of participants and the positive rates in each report. When multiple reports were available in a country, the mean values of the seroprevalence and those of lower and upper limit values of CIs were calculated.
The correlation of IgG reactivity to the S1 protein, as assessed by protein-A/G-based ELISA, with that assessed via conventional ELISA using anti-feline IgG conjugated with horseradish peroxidase (anti-feline-IgG-based ELISA) was analyzed using Pearson's correlation coe cient test.
Correlations of the seroprevalence of SARS-CoV-2 antibodies in cats with factors associated with cat-rearing styles were analyzed using the Kolmogorov-Smirnov normality test, followed by Spearman's correlation coe cient test. The prevalence of anti-SARS-CoV-2 IgG in cats and humans was compared using Fisher's exact test. Statistical signi cance was set at p < 0.05. All statistical analyses were performed using GraphPad Prism 9.0 (GraphPad Software, Inc., San Diego, CA, USA All sampling procedures complied with national and Japanese regulations, and the animal ethics committee approved this study of Azabu University, Japan (approval number 210407-7). The study was carried out in compliance with the ARRIVE guidelines.

Consent for publication
Not application.

Availability of data and materials
All data generated or analyzed during this study are included in this published article and its supplementary information les.

Competing interests
To the best of our knowledge, the named authors have no con icts of interest, nancial or otherwise.