Salmonellosis detection and evidence of antibiotic resistance in an urban raccoon population in a highly populated area, Costa Rica

Abstract Wild animals are involved in zoonotic disease transmission cycles. These are generally complex and poorly understood, especially among animals adapted to life in human ecosystems. Raccoons are reservoirs and effective carriers for infectious agents such as Salmonella throughout different environments and contribute to the transference of resistance genes. This study examined the presence of circulating Salmonella sp. in a population of raccoons in a tropical urban environment and evaluated resistance to antibiotics commonly used to treat salmonellosis. A total of 97 raccoons of different ages and sex were included in this study. 49% (38–60 CI) of the faecal samples were positive for Salmonella spp. The study identified 15 circulating serovars with the most prevalent being S. Hartford (7/15), S. Typhimurium (4/15) and S. Bovismorbificans (4/15). These serovars correspond to the serovars detected in humans with clinical symptoms in Costa Rica. 9.5% of the Salmonella strains recovered demonstrated ciprofloxacin resistance, and 7.1% showed resistance to nalidixic acid. This study provides evidence of multiple Salmonella serovars circulating in a population of urban raccoons in Costa Rica. Furthermore, the study confirms the existence of antimicrobial resistance to two antibiotics used to treat human salmonellosis. The findings emphasize the role of the raccoon as a reservoir of Salmonella in the Greater Metropolitan Area of Costa Rica (GAM) and stress the need for active monitoring of the presence and possible spread in antibiotic resistance due to this peri‐domestic carnivore.

Data from such ecosystems are fundamental to understanding the potential pathways and dynamics of zoonotic transmission (Adams & Lindsey, 2010;Conover, 2002).
Salmonella is a pathogenic agent that persists in the environment. This is due to the bacteria's specific resistance or more general crossresistance mechanisms to a variety of deleterious environmental conditions, but it is also a result of constant environmental contamination by the infected hosts, allowing for effective Salmonella expansion into new environments and new vertebrate hosts. This contamination process is associated with the type of vertebrates acting as reservoirs as well as the abiotic conditions, especially in endemic areas (Bondo et al., 2016bb;Hoelzer et al., 2011;Kumar, 2012;Murray, 1991;Spector & Kenyon, 2012).
Furthermore, the presence of antimicrobial resistance (AMR) and the spread of the respective resistance genes represent a serious public health crisis, constraining the effective treatment of affected human patients (O'Neill, 2016). AMR in Costa Rica has a strong economic and public health impact (Ministerio de Salud., 2017), and PAHO has stated that changing to a holistic surveillance of drug resistance helps to fight AMR (Grundmann et al., 2011). Therefore, surveillance should also be extended into wildlife species (Queenan, Häsler, & Rushton, 2016).
The World Health Organization listed Salmonella bacteria as a priority pathogen for promoting research and development (R&D) of new antibiotics (WHO, 2019). Although they are still classified below the high priority group, the increase in quinolone resistance within this group is of serious concern. While wild animals can both harbour and spread resistance genes, they can also serve as sentinels for environmental AMR (Bondo et al., 2016aa;Duncan et al., 2012;Jobbins & Alexander, 2015;Loncaric et al., 2013).
Increasing conflicts due to the presence of raccoons in urban zones of Costa Rica is a recent phenomenon that is attributed to the close proximity of raccoons and human residents inside the most populated areas of Costa Rica (Baldi et al., 2016). The close proximity of humans and raccoons can potentially increase the exchange of pathogens and resistance genes within urban landscapes. This presents a serious public health concern that must be considered and addressed (Bondo et al., 2016aa;Bradley & Altizer, 2007;Conover, 2002). Even more so in Costa Rica, Salmonella epidemiology is limited to reported clinical cases in hospitals (INCIENSA, 2015).
This study determined the prevalence and diversity of endemic Salmonella spp strains and provides evidence of antibiotic resistance in urban raccoons within the Great Metropolitan Area (GAM) of Costa Rica.

| Study area
In the period between 2013 and 2016, we captured a total of 97 urban raccoons (P. lotor) in 15 different geographic localities within the GAM of Costa Rica (Figure 1). The first 20 individuals were captured opportunistically and used in a pathology-based study. The remaining 77 animals were sampled, tagged and released. Permission

• Urban raccoons in the Greater Metropolitan Area of
Costa Rica interact closely with human beings, enhancing the potential transmission and dissemination of zoonotic pathogens such as Salmonella.
• We demonstrate a high prevalence of Salmonella strains and a great diversity of Salmonella strains in this raccoon population. Additionally, we show evidence of antibiotic resistance to important antibiotics such as ciprofloxacin and nalidixic acid.
• A clearer understanding of the existing risk of close human-raccoon contact will further awareness among health and wildlife management authorities and the general public about the need to avoid close contact with these animals while improving the management of this species in urban areas. Assessment of the carcase and internal organs was performed according to previously standardized protocols (Woodford op., 2000) including weight, length and tooth eruption measurements. Tissue samples from internal organs were fixed in 10% neutral buffered formalin for histopathological examination.
Rectal swab samples of all 97 raccoons were collected directly from the rectum and placed in sterile bags (Whirl-Pak®). All faecal swabs were stored at ~4°C and transported to the previously mentioned laboratories within 3 hr or less. They were then placed in Stuart medium for bacterial culture. The faecal swabs from 11 animals could not be processed due to a problem in labelling. Location, sex, age and weight were recorded for each captured raccoon (Table 1), and individuals were classified as either juvenile or adult based on weight, size and dentition (Grau, Sanderson, & Rogers, 1970).

| Salmonella spp.
A total of 86 rectal swab samples were cultured and characterized via biochemical tests. We followed the methods recommended by the United States Department of Agriculture (USDA) for the isolation and identification of Salmonella colonies with some modifications (Laboratory Quality Assurance Staff & USDA/FSIS/OPHS, 2016). Briefly, the primary culture was performed by adding 1 g of stool or a rectal swab to 10 ml of sterile buffered peptone water (Oxoid®) which was then homogenized. Samples were incubated at 35°C ± 2°C for 24 hr, and 1 ml and 0.1 ml aliquots were then transferred to 10 ml of tetrathionate broth (HAJNA; Acumedia®) and 10 ml of Rappaport-Vassiliadis broth (Acumedia®), respectively. Tubes were incubated in a water bath at 42°C ± 0.5°C for 24 hr. After incubation, an aliquot from each enrichment broth was streaked onto Xylose Lysine Tergitol 4 Agar (XLT4, Acumedia®) and

| Statistical analysis
Descriptive and comparative statistics were performed by using r (R Development Core Team, www.r-proje ct.org) to establish the prevalence, and chi-square test was used to assess the association between Salmonella positive animals and their sex, gender and weight.

| Geocoding and spatial analysis
A map was generated with the georeferenced points of each sample taken in the three different zones. We assigned a number to the geographic point where the serovar was collected and identified.
Serovar information from 15 strains of Salmonella was identified and geocoded. The map was generated using arcgis 10.1 software (ERSI).

| RE SULTS
A total of 97 raccoons were sampled, but only 86 rectal swab samples could be analysed for Salmonella detection during the 20-month sampling period. Of the total raccoons sampled, 47% (n = 46) were males and 54% (n = 51) females, 15% were juveniles (15)  TA B L E 2 Salmonella serovars detected in the faecal samples of the raccoons indication of disease, and additionally, gross examination during necropsy (n = 20) showed them to be in good health. The body weight had a mean value of 4.87 (CI: 4.50-5.26) kg. Table 1 shows the weight range per sex and age class.  Table 3. There was no statistical evidence of associations between sex, gender or weight parameters or with Salmonella cases for the animals (n = 86) included in these analyses.

| Salmonella
Antimicrobial resistance was detected in 7.1% (3/42) of the isolates against ciprofloxacin. These three isolates presented a MIC of 0.5 μg/ml. A reduction in the susceptibility against nalidixic acid (MIC: 32 μg/ml) was also detected in one isolate 2.3% (1/42). The rest of the samples were susceptible to all antibiotics tested. Two Salmonella isolates obtained from raccoons sampled from two different geographical locations showed resistance to both ciprofloxacin and nalidixic acid. The serovars resistant to ciprofloxacin were identified as S. Typhimurium (1), S. Kentucky (1) and one Salmonella without serovar identification (S. none); resistance to nalidixic acid was detected in the following serovars: S. Typhimurium (1) and S.
none (1). Serovars resistant to both antibiotics were as follows: S.

| D ISCUSS I ON
Although raccoons are viewed as a natural reservoir for Salmonella and Salmonella has been previously reported in raccoons in temperate zones, this study establishes, for the first time, the presence of Salmonella bacteria and identifies antibiotic resistance to quinolones in raccoon populations in a tropical urban zone in Costa Rica (lat: 9.7489°N).
Previous reports from varying temperate zone urban, rural and natural environments have provided contradictory results as to the prevalence of Salmonella serovars in raccoons (Compton et al., 2008;Jardine et al., 2011;Lee et al., 2011;Rainwater et al., 2017;Very et al., 2016). In tropical urban settings, we found S. Hartford to have the highest prevalence followed by S. Typhimurium and S.
This contrasts with our findings, where S. Hartford was identified at 18.3% (2/11), 20% (5/25) and 0%, respectively, within the three study areas. The reported prevalence S. Typhimurium showed that this serovar is among the more consistently recorded serovars in raccoons in rural or natural temperate areas. We found S. Typhimurium (16.6%) and S. Bovismorbificans (9.5%) to be the second most prevalent serovars in urban areas in Costa Rica. This differs from previous reports in temperate zones (Bondo et al., 2016a(Bondo et al., , 2016bCompton et al., 2008;Jardine et al., 2011;Lee et al., 2011;Very et al., 2016;White et al., 1975) and might be associated with the type of urban organization and type of climate, which are very different from the tropical climate and urban organization in Costa Rica.
Our findings show that the serovars S. Bovismorbificans and S. Typhimurium circulate among the studied raccoon population.
S. Bovismorbificans is a serovar that has been previously associated with outbreaks in humans, especially in developed countries (Blaylock et al., 2015;Gilsdorf et al., 2005;Hendriksen et al., 2011;Mattila et al., 1998). It can colonize and persist in different species of animals and may contaminate meat products for human consumption (Nógrády, Imre, Kostyák, Tóth, & Nagy, 2010). In Costa Rica, S. raccoons were similar to those recovered from clinically ill patients in Pennsylvania, USA. We did not perform a PFGE analysis of our raccoon samples; therefore, it is not possible to establish how many serovars are potentially involved in an exchange or overlap of serovars (Very et al., 2016). Certainly, in our study there is a significant similarity in the number of serovars isolated in raccoons and those reported in humans in the study area (47% national cases; INCIENSA, Antimicrobial resistance in raccoon Salmonella isolates from our study yielded prevalences similar to other studies Lee et al., 2011;Very et al., 2016). However, only Lee et al., 2011 previously reported resistance to nalidixic acid without resistance to ciprofloxacin in Salmonella from raccoons. Nalidixic acid resistance was detected previously in S. Infantis (2) and S. Typhimurium (1) serovars in raccoons. Both antibiotics are used for salmonellosis and other infections in humans (Pickering, 2012).
While raccoons can be infected with Salmonella AMR bacteria from different sources (Bondo et al., 2016a;Martinez, 2009;Radhouani et al., 2014;Rosenblatt-Farrell, 2009), it appears likely that, in the urban environment, sources are human food, waste and domestic animals (pets) (Baldi et al., 2016). Molina et al., 2016 showed high levels of contamination in animal food and by-products dedicated to the manufacture of these foods in Costa Rica with different Salmonella strains as well as antibiotic resistance genes (Molina et al., 2016 Nevertheless, plasmid-mediated quinolone resistance in Salmonella is an alternative resistance mechanism that employs an enzyme to shield DNA gyrase and type IV topoisomerase from quinolone inhibition (Al-Gallas et al., 2013;van Hoek et al., 2011).
Most importantly, this shows that a level of AMR is already present in some of the Salmonella strains that are circulating in the raccoon population in Costa Rica. This poses the risk of AMR gene transfer as a mechanism of resistance dispersal by different means (physical and biological forces) (Allen et al., 2011;Okeke & Edelman, 2001). Our findings suggest that raccoons could play a role in the dissemination of antibiotic resistance genes in urban areas in Costa Rica.

| CON CLUS ION
Our study provides evidence of the presence of various Salmonella serovars and related antibiotic resistance in urban raccoons in Costa Rica and highlights the need to monitor raccoons as potential sentinel candidates for Salmonella surveillance and environmental antibiotic resistance (Martinez, 2009;Rosenblatt-Farrell, 2009).

ACK N OWLED G EM ENTS
We would like to thank Dr. PK Robbins-Walzer for editing manuscript, and Mario Romero and Martha Piche for the field work assistance. Funding for this study was provided by a grant from the

CO N FLI C T O F I NTE R E S T
The authors have no conflict of interests.