Cats shedding pathogenic Leptospira spp.—An underestimated zoonotic risk?

Shedding of DNA of pathogenic Leptospira spp. has been documented in naturally infected cats in several countries, but urinary shedding of infectious Leptospira spp. has only recently been proven. The climate in Southern Chile is temperate rainy with high annual precipitations which represents ideal preconditions for survival of Leptospira spp., especially during spring and summer. The aims of this study were to investigate shedding of pathogenic Leptospira spp. in outdoor cats in Southern Chile, to perform molecular characterization of isolates growing in culture, and to assess potential risk factors associated with shedding. Urine samples of 231 outdoor cats from rural and urban areas in southern Chile were collected. Urine samples were investigated for pathogenic Leptospira spp. by 4 techniques: qPCR targeting the lipL32 gene, immunomagnetic separation (IMS)-coupled qPCR (IMS-qPCR), direct culture and IMS-coupled culture. Positive urine cultures were additionally confirmed by PCR. Multilocus sequence typing (MLST) was used to molecularly characterize isolates obtained from positive cultures. Overall, 36 urine samples (15.6%, 95% confidence interval (CI) 11.4–20.9) showed positive results. Eighteen (7.8%, 95% CI 4.9–12.1), 30 (13%, 95% CI 9.2–18), 3 (1.3%, 0.3–3.9) and 4 cats (1.7%; 95% CI 0.5–4.5) were positive in qPCR, IMS-qPCR, conventional culture, and IMS-coupled culture, respectively. MLST results of 7 culture-positive cats revealed sequences that could be assigned to sequence type 17 (6 cats) and sequence type 27 (1 cat) corresponding to L. interrogans (Pathogenic Leptospira Subgroup 1). Shedding of pathogenic Leptospira spp. by cats might be an underestimated source of infection for other species including humans. The present study is the first one reporting growth of leptospires from feline urine in culture in naturally infected cats in South-America and characterisation of culture-derived isolates. So far, very few cases of successful attempts to culture leptospires from naturally infected cats are described worldwide.

Introduction shed pathogenic Leptospira spp., while the reported antibody prevalence in cats in rural areas (25.2%) [18] is among the highest reported worldwide. Therefore, the role of cats as a reservoir for pathogenic Leptospira spp. could be very important.
Thus, to provide scientific knowledge for a better understanding of the role and relative importance of cats as a source of pathogenic Leptospira spp. infection, the aims of the study were: (1) to determine the proportion of cats shedding pathogenic Leptospira spp. in Southern Chile, (2) to determine the molecular profile of the cultured isolates, and (3) to identify possible risk factors regarding Leptospira spp. shedding.

Ethical note
Procedures were approved by the Universidad Austral de Chile Animal Care Committee (No. 269/2016). Informed written owner consent was obtained for all cats included in the study.

Study design
The sampling size was statistically estimated by assuming an expected prevalence of urinary shedding of pathogenic Leptospira of 4% with a 95% confidence interval (CI) of the estimation with a 5% precision, indicating that at least 236 cats had to be included. Between October 2016 and March 2017, cats of rural and urban origin from the Los Ríos and Los Lagos region in Southern Chile were included, trying to cover a geographic area in which a high antibody prevalence in cats had been reported previously [18].

Cats
Included cats were patients of the small animal hospital of the Universidad Austral de Chile in Valdivia presented for health issues or for elective procedures (e.g. neutering, vaccination), client-owned cats participating in castration projects offered by the city of Valdivia or other communities, and cats presented by their owners specifically for study enrolement. Cats were offered one free vaccination and deworming as compensation for participating, if this was clinically justifiable. Cats were excluded from the study if they had received antimicrobial treatment within the last four weeks before sampling due to a decreased likelihood for detection of Leptospira spp. even if they were infected, or if they were living exclusively indoors or had only access to a balcony because of a low risk for infection.

Sampling
Urine samples were collected via cystocentesis using a 21 G needle attached to 5 ml syringe. Cystocentesis was performed with palpation of the urinary bladder or ultrasound-guided. At the time of sampling, a physical examination was performed in each cat. A questionnaire with information about the cat and potential risk factors was answered by the owner. It included questions about signalment, living conditions, consumption of raw meat, drinking out of puddles, contact with rodents or eating rodents, and contact with other cats, dogs, or livestock and vaccination status.

Detection of pathogenic Leptospira spp. by qPCR and culture
To increase the overall diagnostic sensitivity for the detection of pathogenic Leptospira spp. in urine, four diagnostic tools were used in each sample: lipL32 quantitative qPCR, immunomagnetic-separation (IMS)-coupled with lipL32 qPCR (IMS-qPCR), culture for Leptospira spp., and IMS-coupled with culture (Fig 1).
Quantitative real-time lipL32 polymerase chain reaction. For each sample, at least 1 mL was centrifuged at 6,500 rpm for 15 minutes. The pellet was resuspended in 1 mL of 1X phosphate buffered saline (PBS) [137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 , 1.4 mM KH 2 PO 4 (pH 7)] and then transferred to a 1.5 mL microcentrifuge tube and re-centrifuged at 11,000 rpm for 5 minutes. Finally, the supernatant was discarded and the pellet was resuspended in 1mL of 1X PBS [26]. After this first cleaning and purification step, 100 μL were separated for a pretreatment step of immune-separation (see below) before proceeding with DNA extraction and direct qPCR. The rest of the urine was subjected to a DNA extraction-purification protocol using the High Pure PCR Template Preparation kit (Roche), following the manufacturer instructions.
The DNA templates obtained were analyzed in a qPCR system (Roche LightCycler 2.0), using a TaqMan probe and targeting the lipL32 gene which is specific only for pathogenic Leptospira spp. [26]. The amplification mixture for each sample included 0.7 μM primers, 0.15 μM probe, 10 μL Master Mix TaqMan universal (Roche) and 5 μL DNA template, in a total volume of 20 μL. Samples were amplified with the following program: Initial denaturation at 95˚C for 2 min, followed by 40 cycles of denaturation for 5 s at 95˚C for 5 and annealing/elongation for 30 s at 58˚C [26,27]. The PCR system included a negative and positive control in order to survey the proficiency of the reaction as well as DNA extraction-negative and -positive controls. The positive control for both, DNA extraction and PCR protocols, corresponded to a pure culture of pathogenic Leptospira spp. in two dilutions with a known concentration of leptospires (10 4 /mL and 10 2 /mL), where the two dilutions were used. Amplification efficiency (E) was calculated from the slope of the standard curve in each run using the following formula (E = 10 −1/slope ).
Immunomagnetic-separation (IMS)-coupled lipL32 quantitative polymerase chain reaction. The urine sample aliquot was pre-treated with an IMS method published previously [28], followed by a DNA extraction purification and molecular confirmation qPCR. The IMS-qPCR consisted of 4 steps (in silico peptide production; polyclonal antibody production; coating magnetic beads with polyclonal antibodies, and IMS-qPCR Leptospira spp. detection). The magnetic separation of pathogenic Leptospira spp. from urine samples and the subsequent washing steps were carried out using the automated BeadRetrieverTM System (Invitrogen Life Technologies, Grand Island, NY). Pathogenic Leptospira spp. were selectively concentrated from 0.5 mL processed samples obtained from the coated magnetic beads, as described above. The final product was suspended in 0.5 mL PBS for DNA extraction purification followed by qPCR for pathogenic Leptospira spp. as described above.
Direct culture. Two dilutions of urine (1:10 and 1:100) with 5 ml of liquid Ellinghausen-McCullough-Johnson-Harris (EMJH) medium supplemeted with 5-fluorouracil (0.2 mg/ml) were prepared immediately after urine sampling. Samples were cultured at 29˚C for 13 weeks according to previous recommendations [29]. In this liquid medium, growth was defined as turbidity that becomes apparent at an estimated bacterial density of 1x10 8 bacteria/mL. Each sample was observed for evidence of growth using dark-field microscopy for observation of viable, motile Leptospira spp. bacteria. DNA of positive isolations was used for molecular characterization using the above described extraction protocol.
Immunomagnetic separation coupled to culture. After immunomagnetic separation as described above, 100 μL of the final product were transferred into 5 ml EMJH medium and cultured at 29˚C for 13 weeks. Each sample was observed for evidence of colony growth, as defined above. DNA of positive cultures was used for molecular characterization.
Estimation of pathogenic Leptospira spp. shedding load. Pathogenic Leptospira spp. cell numbers (genome equivalents) were estimated from either of the detection methods. Estimation was based on the concentration of Leptospira spp. DNA, obtained from positive samples, that was measured in a Nanoquant spectrophotometer (TECAN group, Männedorf, Schweiz) adjusted for a 10 6 dilution. With the reference of the molecular weight of the genome of Leptospira interrogans serovar Hardjo-prajitno (GenBank accession number EU357983.1) a standard curve for estimation of pathogenic Leptospira spp. numbers in urine using a Roche 2.0 real-time PCR, according to the following equation was established [30,28]: Genome equivalent ¼ DNA concentration ðng=mlÞ � ð6:022 � 10 23 mol À 1 Þ ð4:659 � 10 6 base pairsÞ � ð1 � 10 9 ng=g � 660g=molÞ ðaccesion number EU357983:1Þ ðbase massÞ Molecular characterization by multilocus sequence typing (MLST). The "Multi Locus Sequence Typing" (MLST) method was used to molecularly characterize strains obtained from positive cultures. MLST was performed according to a protocol described by Thaipadungpanit et al. (2007) with the following housekeeping genes: glmU, pntA, sucA, tpiA, pfkB, mreA, and caiB [31]. Briefly, PCR amplification was carried out using Leptospira spp. DNA obtained from positive urine cultures, where the initial denaturation step was 94˚C for 5 min, followed by 35 cycles of 94˚C for 30 sec, annealing/elongation was at 50˚C for 60 sec for all genes, extension was at 72˚C for 50 sec, and then the final extension was at 72˚C for 7 min. The PCR products were purified using the Geneaid™ "Gene/DNA Genetic Extraction" kit, and sequencing was done in both directions with primers initially used for PCR amplification. Sequencing was performed using the sequencer kit "BigDye Terminator v. 3.1 cycle sequencing kit (ABI)" and the automated DNA sequencer "ABI Prism 3130xl Genetic". The sequences were analyzed using the Chrommas and Bioedits programs, and these were derived from the international database for free use (https://pubmlst.org/leptospira/) to obtain the allelic profile and to assign the sequence type (ST).

Data analysis
The overall prevalence of urinary shedding (as detected by any of the 4 techniques) of pathogenic Leptospira spp. with their 95% confidence was calculated. Mann-Whitney test was used for comparison of Leptospira spp. loads between direct qPCR and IMS-qPCR. Statistical analysis to assess risk factors was performed by ordinary and mixed-logistic regression models using the location of the cat as a random slope effect to account for potential environmental heterogeneity.
The strategy for building the model consisted in obtaining unconditional models for initial screening of variables. Variables associated with the outcome (P<0.25) were eligible for inclusion in the conditional model that was built using a forward approach and a P value of 0.05 was used for interpreting the outcomes. Potential interactions were tested based on their biological significance. The best model was assessed by using the Bayesian information criterion (BIC) index as a measure of goodness-of-fit. Two analysis were performed: a) one for cats positive with any of the 4 techniques, combined as "Leptospira-shedding" cats, and b) one for "viable Leptospira-shedding" cats (culture-and IMS-coupled culture-positive cats). Analysis was done with lme4 package [32] of R (R Development Core Team (2008)) [33]. A power post hoc analysis was performed using WebPower package [34] for a evaluation of variables that were statistically not significant in the model.

Shedding of pathogenic Leptospira spp.
In total, 231 cats from 11 locations at 7 municipal districts were included into the study ( Fig  2). Overall, 30 cats were PCR-positive (Table 1), and 7 cats were culture-positive.
All 18 cats that were positive on qPCR were also positive on IMS-qPCR (Table 2). IMS-qPCR identified 12 additional cats as shedders of pathogenic Leptospira spp. Seven cats were positive in culture, 3 cats in direct culture and 4 other cats in IMS-coupled culture. None of them was positive in both culture methods. One of the cats that was positive in IMS-coupled culture was also positive in conventional qPCR and on IMS-qPCR. The other 6 culture-positive cats were PCR-negative. Individual data on signalment and health status are illustrated in Table 2. Data of geographic distribution of positive cats are combined in Table 3.
Multi locus sequence typing of isolates obtained of 7 positive cultures revealed that sequences of 6 cats were identical to available sequences of L. interrogans sequence type (ST) 17, and the sequence of one cat was identical to available sequences of L. interrogans ST 27 (Table 4).

Risk factor analyses
For unconditional model (Table 5), health status (P<0.001), eating rodents (P = 0.08), reproductive status (P = 0.11), previous vaccination (P = 0.11), and age >1 year (P = 0.12) were variables primary associated with leptospiruria, and used for building a conditional model for cats that were PCR-and/or culture positive.
Conditional logistic regression after evaluating potential interactions and confounders is illustrated in Table 6 and included four risk factors. Two of them (being sick and being vaccinated) were significantly associated with shedding of pathogenic Leptospira spp. while the other two (age >1year and eating rodents) were not. The random intercept was statistically significant and accounted for 22.9% of the extra variability due to the place where the cat lived. The power estimation for this model was 90% which is considered high enough.   In addition, another model was used to evaluate potential risk factors associated with shedding of viable pathogenic Leptospira spp. (only cats positive on direct culture or IMS-coupled culture). For the unconditional models, only 3 variables were potential candidates: age >1 year, the reproductive status of the cat, and eating rodents were significantly associated with Leptospira spp.-positive urine cultures. Although a mixed model was also assessed, comparison of the goodness-of-the-fit between both indicated that the mixed-model did not fit better. Therefore, ordinary logistic regression model results are shown. The conditional logistic regression model (Table 7) included 2 variables (age >1 year and eating rodents) that were significantly associated with shedding of viable pathogenic Leptospira spp.

Discussion
Overall, 15.6% of cats included in the study were identified as shedders of pathogenic Leptospira spp‥ This is considerably higher than the shedding rates from 0.8% to 4.9% in Thailand, Spain, Germany, Canada, and Malaysia [12,[14][15][16]20]. One possible explanation might be the diagnostic techniques used in the present study as the application of an additional Leptospira spp. concentration technique (IMS-qPCR) increased the proportion of Leptospira spp.-shedding cats identified by PCR from 7.8% to 13.0% (12 additional cats). With direct culture and IMS-coupled culture another 6 shedding cats were identified. Other possible explanations could be related to the humid climatic conditions in the area where the study was performed with very high annual precipitations and the types of pathogenic Leptospira spp. circulating in the study area. In the present study, serovar identification was not performed, but the culturegrown leptospires could be characterized as L. interrogans ST 17 and ST 27 (pathogenic Subgroup 1) based on MLST. Other studies using whole-genome sequencing provided evidence that strains of pathogenic leptospires differ in genetic features and virulence factors for Table 3

Sampling location (Municipality) All sampled cats n PCR-positive cats n (%) Culture-positive cats n (%) All positive cats n (%)
Valdivia 186   persistence outside the host, during entry into the host, as well as for colonization and persistence within the host organism [35,36], suggesting that not all pathogenic Leptospira spp. are equally virulent [37]. Therefore, it cannot be excluded that the identified strains in the present study might have a higher virulence or be better adapted to invade the species cat than strains in other geographic regions. There is only one study with a much higher proportion (68%) of shedding cats in Taiwan [11]. However, this study was performed after a typhoon with an associated outbreak of leptospirosis in humans [38] and included a high proportion of stray cats (68%), whereas in the present study most of the cats were privately owned well-fed cats, and only 31% of them had been observed by their owners eating rodents. In addition, in the study from Taiwan, a different PCR technique (not a quantitative PCR) was applied and not all primers used in that study were specific for pathogenic Leptospira spp. While PCR can detect DNA of viable and killed bacteria, culture is only positive when a sufficient number of viable Leptospira spp. is present. Recently, a study from Malaysia demonstrated cultivation of pathogenic Leptospira spp. from renal tissue in 3/82 clinically healthy shelter cats and from both, renal tissue and urine, in 1/82 of these cats [20]. In addition to that study, there is one case report of a cat with a positive Leptospira spp. culture originating from the same geographic area as the present study [10]. The present study also yielded positive Leptospira spp. culture results from feline urine. Feline urine with its higher osmolality compared to dogs and humans [39] is a hostile environment for bacterial growth and therefore could prohibit the shedding of viable Leptospira spp. in many cats. This might explain the low proportion of culture-positive cats in the present study. Low pH is another physicochemical property of the urine contributing to the defense system against bacteria [40] and Leptospira spp. do not survive well in acidic urine but remain viable in alkaline urine [41]. However, in the present study, the mean urine pH was identical in culture-positive and in culture-negative cats.
Usage of magnetic separation enhances the analytical specificity and sensitivity of the subsequent detection method, given that it helps to remove the presence of many inhibitory substances or other organisms [42]. Still, only 7 cats (3.0%) were culture-positive (3 direct culture, 4 IMS-coupled culture) compared to 30 (13.0%) PCR-positive cats. However, detection of positive cultures nevertheless demonstrates that feline urine can be a source of possible contamination of the environment with viable Leptospira spp. One possible explanation for the discrepancy in the results of direct and IMS-coupled cultures is the hypothesis that antibodies used in IMS could have a neutralizing effect for the growth of Leptospira spp. in culture.
In the present study, only 1 of the 7 culture-positive cats was also PCR-positive (qPCR and IMS-qPCR). One possible reason for the negative PCR in the culture-positive cats could be the presence of PCR inhibitors. In biologic fluids, such as urine, presence of PCR inhibitors has been reported to cause false negative PCR results [43]. However, positive controls in the present study showed hardly or no inhibition. Interestingly, a similar discrepancy was seen in the only other study in cats reporting the isolation of Leptospira spp. from feline kidneys and urine [20]. In that study, all 4 culture-positive cats had negative urine PCR results. This indicates that there might be substances/chemicals present in the PCR processing but not in the samples that are cultured that could be inhibitory for certain steps in the PCR. Other reasons might also be the level of detection on qPCR and possibly the selection of primers that do not bind to all DNA targets of Leptospira spp. However, molecular characterization of isolates later on revealed sequence types that should have been recognized by PCR.
In the present study, serovar identification was not performed. Nevertheless, it was possible to molecularly characterize isolates via MLST, and identified sequences were assigned to ST 17 and ST 27, which corresponds to L. interrogans (Pathogenic Leptospira Subgroup 1) [44]. Although the Taxonomic Subcommittee does not yet accept MLST for serovar identification, the information reported from the MLST database used in the analysis of the isolated strain in this study, reports the ST 17 as Leptospira interrogans serogroup Icterohaemorrhagiae serovar Copenhageni, and the ST 27 as L. interrogans serogroup Autumnalis serovar Autumnalis. Previously characterized Leptospira spp. isolates from cats´kidney or urine had high similarity to Leptospira interrogans serovar Pomona (1 cat from this area of Chile) [10] or to Leptospira interrogans serovar Bataviae (4 cats from Malaysia) [20]. Whereas, Leptospira interrogans Bataviae is reported to be the most common isolated serovar in humans, dogs and rats in Malaysia [45][46][47], L. interrogans serovar Autumnalis is among the most commonly reported serovars in dogs and cats in Chile [18,48,49], and antibodies against L. interrogans Icterohaemorrhagiae have previously been reported in rodents in this area [50].
Four factors (age (�1 year/>1year), health status, eating rodents, and vaccination status) were present in the final conditional model for all Leptospira spp. shedding cats (cats positive in any of the 4 techniques), and 2 factors (age (�1 year/>1year), eating rodents) for the "viable Leptospira spp. shedding" cats (culture-or IMS-coupled culture-positive cats).
Adult stage (>1 year) was significantly associated with leptospiruria (odds ratio 6.8, CI 90% 1.08- 40,4). Previous studies already reported older age as a risk factor for Leptospira spp. infection [15,51]. A study in Thailand identified age >4 years as the only risk factor for infection (urine qPCR-positive status and/or serum antibodies in MAT) in a multivariate analysis [15]. A possible explanation for this is longer exposure to infectious environmental sources and therefore a higher risk of being infected. It is not known how long cats remain carriers after infection or if and under which circumstances they can eliminate the infection. Risk factors for Leptospira spp. infection have been investigated in several studies. Contact of cats with livestock animals was significantly associated with presence of Leptospira spp. antibodies in a study performed in Chile [18] but such an association between shedding and contact with livestock animals was not identified in the present study.
The proportion of sick cats was significantly higher in Leptospira spp. shedding cats (36,1%) than in non-shedding cats (5,1%). Clinical cases of leptospirosis in cats are rarely described. In one case series with 3 cats and 2 further reports of single cases, cats presented with acute onset of polydipsia and polyuria, lethargy, anorexia, and hematuria, and had various degrees of kidney disease or increased liver enzymes [8][9][10]. In a more recent case report, a cat presented with a 2-months history of hematuria, had no abnormalities in serum chemistry but a marked leukocytosis [10]. One leptospiruric cat (positive in qPCR and IMS-qPCR) of the present study also suffered from acute renal failure (history of acute lethargy and anorexia, severe renal azotemia), and 2 presented with hematuria (1 cat positive in qPCR, IMS-qPCR and IMS-coupled culture, 1 cat positive in IMS-qPCR). These clinical signs could indeed be associated with leptospirosis and represent acute infection. In the other 10 sick cats with different organs affected, a causal relationship of infection with Leptospira spp. is not likely. However, the higher proportion of sick cats shedding pathogenic Leptospira spp. in the present study suggests that Leptospira spp. could be one of the etiological factors in the development of these clinical problems, and that diagnostic testing Leptospira spp. infection should also be considered in cats with relevant clinical signs. Alternatively, it could be discussed that sick cats might be immunosuppressed and that an impaired immune status increases the risk of leptospiral infection and subsequent shedding. This, however, has not been reported so far. On the other hand, sick cats shedding pathogenic Leptospira spp. could represent chronically infected individuals. It would have been interesting to further monitor these cats to see if they remained chronic shedders.
Eating rodents was found to be associated with an increased risk of shedding viable pathogenic leptospires. In general, rodents are a natural reservoir pathogenic Leptospira spp., and rats are reservoir hosts of the Icterohaemorrhagiae and Copenhageni serovars [52]. Therefore, prey-predator transmission between cats and rodents is possible by hunting and is believed to be the main source of infection in cats [12,53]. In the contrary, a negative association between cat ownership and the presence of Leptospira antibodies in people in an urban population in North America was demonstrated, and cats appear to have a protective role by decreasing contact of people with rodents that act as reservoir hosts [54].
An unforeseen result was the statistically significant association of an increased risk of shedding pathogenic Leptospira spp. and previous vaccination against feline panleukopenia, feline herpes-and calicivirus infection and/or rabies. The reason for this association is unknown. Immunological host-pathogen processes that could affect shedding of leptospires by cats in any direction are conceivable but specific information is missing. Weather there is an interference in susceptibility of vaccinated cats to act as shedders of Leptospira spp. is an interesting aspect that should be further investigated.
Results of the present study show that cats can shed viable pathogenic Leptospira spp. and, therefore, could act as transmitters to humans and livestock. Results of PCR and culture with and without separation and concentration, taken together, showed that: i) the bacterial load in feline urine is relatively low; ii) in 90% of feline shedders it was not possible to culture the organism suggesting that viability of the bacteria in feline urine is negatively affected by the hostile environment, e.g. the relatively high osmolality of feline urine; and iii) separation and concentration followed by qPCR significantly increases the sensitivity of detection of leptospiral DNA in feline urine.
Apart from the possible negative effect on bacterial viability of the hostile environment that is provided by feline urine and the stage of disease, there is another potential cause of negative culture result. There is a known difficulty to isolate Leptospira spp. from infected hosts and get them growing in vitro. Despite the presence of viable leptospires in clinical samples of an infected host, it is in many cases very difficult to isolate the bacteria in media due to the in vivo phenotype of the host-derived bacteria and their lack of capacity to adapt to in vitro conditions. Based on this general problem with cultural isolation of Leptospira spp. from infected humans and animals, it can be concluded that, despite this difficulty, viable Leptospira spp. were detected in the urine of 7 cats, which demonstrates the zoonotic potential of this infection in cats.

Conclusion
In this region of Southern Chile, 15.6% of cats were leptospiruric and 3.0% of cats were Leptospira spp. culture-positive. The present study is one of the first reporting growth of Leptospira spp. from feline urine in culture and describing genetic characteristics of culture-derived isolates. Even though feline urine appears to be a hostile environment for leptospires, shedding cats present a potential risk for zoonotic transmission.