Abstract
Canine leptospirosis, which is caused by infection with pathogenic Leptospira species, occurs worldwide, but information regarding the causative Leptospira serotypes and genotypes and their effects on virulence in dogs remains limited. Monitoring acute leptospirosis in dogs as sentinels can also aid in estimating the risk of human leptospirosis, particularly when the disease is rare, as it currently is in Japan. Among 283 clinically suspected cases of leptospirosis diagnosed from August 2007 to March 2011 in Japan, 83 cases were laboratory diagnosed as leptospirosis by blood culture, a rise in antibody titres in paired sera using a microscopic agglutination test (MAT) and/or DNA detection using flaB-nested PCR. The infected dogs comprised hunting dogs (31 dogs) and companion animals (50 dogs) and two unknown; 63.4 % of the infected dogs were males. The mortality rate was 53.2 %. A rise of at least fourfold in MAT titre was detected in 30 dogs whose paired serum samples were obtained, and the predominant reactive serogroup was Hebdomadis (53.3 %), followed by Australis (16.7 %) and Autumnalis (16.7 %). Leptospira interrogans was isolated from 45 dogs of the following serogroups: Australis (16), Autumnalis (six), Canicola (one), Hebdomadis (21) and Icterohaemorrhagiae (one). All of these serogroups caused lethal infections (57.1–100 %). Genetic heterogeneity was demonstrated in serogroups Australis, Autumnalis and Hebdomadis by multilocus sequence typing (MLST) and/or RFLP analysis based on PFGE. In serogroup Hebdomadis, each genotype determined by MLST had a unique mortality rate in the infected dogs. Although classic canine leptospirosis is associated with serovars Canicola and Icterohaemorrhagiae, serogroup Hebdomadis has become the predominant serogroup causing high mortality in Japan. This study suggests that the virulence of members of serogroup Hebdomadis in dogs may be associated with the genotypes in this serogroup.
-
The GenBank/EMBL/DDBJ accession numbers for the flaB sequences of canine Leptospira interrogans isolates determined in this study are AB700138–AB700210.
-
A supplementary figure is available with the online version of this paper.
Introduction
Leptospirosis is a worldwide zoonosis caused by infection with pathogenic Leptospira species, affecting almost all mammals (Bharti et al., 2003; Faine et al., 1999). Canine leptospirosis is widespread throughout the world, and dogs serve as both incidental hosts for various leptospiral serovar strains and maintenance hosts for serovar Canicola (Sykes et al., 2011). Monitoring acute infection of leptospirosis in dogs as sentinels can also aid in estimation of the risk to humans in specific areas (Martins et al., 2012). In accidental infections, dogs exhibit acute or subacute hepatic and renal failure, which historically is associated with serovars Canicola and Icterohaemorrhagiae (Goldstein, 2010). However, the widespread use of bivalent vaccines comprising these serovars and the increased contact between dogs and wildlife reservoirs in expanding suburban environments are likely to result in changes of the prevalent Leptospira serovars or the emergence of new serovars in the USA and Europe (Ellis, 2010; Gautam et al., 2010; Stokes et al., 2007). Although there are numerous reports on serological investigation of canine leptospirosis, characterization of Leptospira isolates causing clinical disease in dogs is limited (Suepaul et al., 2010; Sykes et al., 2011). No recent information on leptospiral species and members of serogroups (serovars) causing clinical disease in Japan has been reported.
In human leptospirosis, it was considered previously that distinct clinical syndromes were associated with specific serogroups, especially serogroup Icterohaemorrhagiae, but this has now been refuted (Levett, 2001). In contrast, it has been suggested recently that a specific clone in Leptospira interrogans serovar Copenhageni was related to severe pulmonary haemorrhage syndrome (Gouveia et al., 2008). A specific clone [a genotype defined by multilocus sequence typing (MLST)] of L. interrogans serovar Autumnalis was associated with outbreaks of leptospirosis in Thailand (Thaipadungpanit et al., 2007). A serological study indicated that members of serogroup Pomona caused more severe renal disease and were associated with a worse outcome in dogs than disease caused by serogroup Autumnalis or Grippotyphosa (Goldstein et al., 2006); however, the differences in the virulence of different Leptospira genotypes in dogs remain unknown.
In this study, we performed laboratory diagnoses, including culture, antibody detection using the microscopic agglutination test (MAT), and DNA detection by flaB-nested PCR, of clinical specimens collected from dogs clinically suspected of leptospirosis in ten prefectures in Japan. We characterized leptospiral isolates by serological and molecular methods such as MLST and RFLP analysis based on PFGE, and examined the association of Leptospira serogroups and genotypes with the outcome of infected dogs.
Methods
Clinical sample collection.
Clinical samples of blood, serum and urine were collected from dogs that exhibited: (i) at least two of the four symptoms of leptospirosis – fever, vomiting, hyperaemia and haemorrhage of the mucous membranes, and jaundice; and/or (ii) acute renal involvement (abnormal values in creatinine and/or blood urea nitrogen) of unknown origin; and/or (iii) acute hepatic involvement (abnormal values in alanine aminotransferase, aspartate aminotransferase and/or alkaline phosphatase) of unknown origin, in ten prefectures of Japan (Ibaraki, Chiba, Mie, Fukuoka, Saga, Nagasaki, Kumamoto, Miyazaki, Kagoshima and Okinawa) from August 2007 to March 2011.
Isolation of leptospires from dogs clinically suspected of having leptospirosis.
To isolate leptospires, two drops of blood were inoculated into 4 ml liquid Korthof’s medium containing 10 % rabbit serum and liquid Ellinghausen–McCullough–Johnson–Harris (EMJH) medium (Faine et al., 1999), and cultivated at 30 °C for 3 months. The cultures in which no leptospires were observed under dark-field microscope during the 3-month incubation period were deemed negative.
Antibody detection from dogs clinically suspected of having leptospirosis.
A MAT was performed to detect anti-Leptospira antibodies in patient serum samples (Faine et al., 1999) using a battery of previously described reference strains (Koizumi et al., 2008). Briefly, 25 µl of twofold serially diluted serum samples was incubated with the same volume of leptospiral cultures for 3 h at 30 °C. The end point was determined by the presence of ~50 % free, unagglutinated leptospires compared with the control suspension (Levett, 2001). At least a fourfold increase in antibody titre between acute and convalescent serum samples was judged as positive. These reference strains were cultivated in liquid modified Korthof’s medium as described above.
Identification of the serogroups of dog isolates.
The serogroups of the isolates were identified by MAT using a panel of anti-Leptospira rabbit sera for serovars Australis, Autumnalis, Bataviae, Canicola, Castellonis, Cynopteri, Djasiman, Grippotyphosa, Hardjo, Hebdomadis, Icterohaemorrhagiae, Javanica, Mini, Pomona, Pyrogenes, Sejroe, Sarmin and Tarassovi.
DNA detection from dogs clinically suspected of having leptospirosis.
DNA was extracted from blood or sediment from plasma, serum or urine after centrifugation (16 000 g, 10 min) using a DNeasy Blood and Tissue kit (Qiagen). Extracted DNA was subjected to nested PCR to detect the leptospiral flaB gene, as described previously (Koizumi et al., 2008), with slight modifications. The reaction volume of the first PCR mixture, which contained 5 µl DNA template, was 50 µl. Next, 1 µl first PCR product was added to 19 µl second PCR mixture. DNA sequencing of the amplicons was performed using a BigDye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems) with the second PCR primers (L-flaB-F2; 5′-TGTGCACAAGACGATGAAAGC-3′, L-flaB-R2; 5′-AACATTGCCGTACCACTCTG-3′).
flaB sequencing.
DNA was extracted from the Leptospira isolates, and leptospiral flaB was amplified using the first PCR primers (L-flaB-F1; 5′-CTCACCGTTCTCTAAAGTTCAAC-3′, L-flaB-R1; 5′-TGAATTCGGTTTCATATTTGCC-3′), after which the nucleotide sequences of the amplicons were determined as described above.
MLST.
MLST of the isolates was performed according to Thaipadungpanit et al. (2007). Sequence types (STs) were assigned through the MLST database (). The new STs obtained in this study (ST118 and ST119) have been deposited in this database. The concatenated sequences were aligned in mega4 (Tamura et al., 2007) using clustal w, and phylogenetic distances were calculated in mega4 using the neighbour-joining method.
PFGE.
RFLP analysis of the genomes of Leptospira isolates was conducted by PFGE using the restriction enzyme NotI, as described previously (Koizumi et al., 2009).
Statistical methods.
Each serogroup of Leptospira isolates, Australis (n = 14), Autumnalis (n = 6) and Hebdomadis (n = 21), was tested for serogroup-specific correlation with mortality rate. For the Hebdomadis serogroup, except in one case of mixed infection with ST37 and ST119, each ST, ST37 (n = 8), ST118 (n = 5) and ST119 (n = 6), was tested for genotype-specific correlation with mortality rates. Associations with dead/alive data were tested using a 2×3 Fisher’s exact test.
Results
Features of laboratory-confirmed leptospirosis in dogs
Of the 283 clinically suspected cases, 83 were laboratory diagnosed as leptospirosis by blood culture, antibody detection (at least a fourfold increase in MAT titre in acute and convalescent serum samples) and/or DNA detection, and leptospirosis was detected in dogs from all prefectures surveyed. Blood cultures were performed for 243 animals, and leptospires were isolated from 45 animals. Paired serum samples were obtained from 58 animals, of which a significant increase in antibody titre was detected in 30 serum samples. Blood and/or urine samples for DNA detection were collected from 236 animals and subjected to flaB-nested PCR. Leptospiral flaB was detected in 44 animals. There were 14, three and nine dogs that were positive for culture and DNA detection, culture and antibody detection, and DNA and antibody detections, respectively, and five dogs were positive for all three tests. Culture, MAT using paired serum samples and DNA detection were performed in 76, 33 and 71 of the 83 positive animals, respectively, and positive results were obtained in 45, 30 and 44 animals, respectively. Thus, the sensitivity of these tests was 59.2 % for culture, 90.9 % for antibody detection and 62.0 % for DNA detection.
The ages of the infected animals ranged from 5 months to 13 years (median 5 years), and 63.4 % were males. The infected dogs consisted of hunting dogs (31 dogs) and companion animals (50 dogs) (two were unknown). All the animals routinely went outside. Of the 82 positive animals with a recorded vaccination history, 24 (29.3 %) were vaccinated. We obtained information on the outcome of 79 of the 83 laboratory-diagnosed animals 3 months after the first visit. The mortality rate among them was 53.2 %. The clinical manifestations of the laboratory-confirmed dogs at the initial visit were fever (13.3 %), vomiting (71.1 %), hyperaemia and haemorrhage of the mucous membranes (41.0 %) and jaundice (78.3 %).
Detection of anti-Leptospira antibodies and leptospiral flaB in dogs clinically suspected of having leptospirosis
Thirty dogs were serologically confirmed to have leptospirosis using paired serum samples, although convalescent sera was obtained only from 58 of 271 dogs whose acute serum samples were collected. The predominant reactive serogroup was Hebdomadis (53.3 %), followed by Australis (16.7 %) and Autumnalis (16.7 %) (Table 1). Although definitive diagnosis was made using paired serum samples, it was assumed that a high MAT titre (≥800) in a single sample to a non-vaccinal serovar, accompanied by clinical signs of leptospirosis, is highly suggestive of active infection (Goldstein, 2010). There were 29 dogs whose single serum sample showed a reciprocal MAT titre ≥800 (Table 1). Combined with the results from paired serum samples, the predominant reactive serogroups were Hebdomadis (57.6 %) and Australis (20.3 %), which were detected in eight of the ten prefectures and five of the ten prefectures, respectively (Table 1). In contrast, antibodies against members of serogroups Canicola and Icterohaemorrhagiae, which are generally associated with canine leptospirosis, were detected in only 6.8 and 0 % of the samples tested, respectively. Leptospiral flaB was detected in 44 animals: 23 in blood samples, 16 in urine samples and five in both blood and urine samples. Based on flaB sequences, all the Leptospira species were L. interrogans (data not shown).
None of the samples tested was specific for serogroup Icterohaemorrhagiae.
Molecular and serological characterization of Leptospira isolates
Leptospires were isolated from 45 animals, and 73 positive cultures were obtained from cultures using Korthof’s and/or EMJH medium in seven prefectures (Table 2). The serogroups of the isolates were identified as Australis (16), Autumnalis (six), Canicola (one), Hebdomadis (21) and Icterohaemorrhagiae (one) by reactivity with the reference antisera (Table 2). Based on flaB sequences, the species of all the isolates were L. interrogans. MLST analysis revealed that all serogroup Australis isolates belonged to ST37, whereas serogroups Autumnalis and Hebdomadis isolates comprised three and four STs including new STs, ST118 and/or ST119, respectively (Fig. 1, Table 3). ST37 was detected in serogroups Australis, Autumnalis, Canicola and Hebdomadis, and ST118 was detected in serogroups Autumnalis and Hebdomadis. Although only one ST (ST37) was detected in serogroup Australis by MLST, PFGE using the restriction enzyme NotI demonstrated genetic heterogeneity in the serogroup Australis (Supplementary Fig. S1 available in JMM Online). Genetic heterogeneity was also observed in each ST of the serogroups Autumnalis and Hebdomadis (data not shown).
Phylogenetic tree based on the concatenated sequences of Leptospira isolates. The sequences were aligned in mega4 using clustal w, and phylogenetic distances were calculated in mega4 using the neighbour-joining method. Numbers on the nodes indicate the bootstrap support after 1000 replicates. Each ST is based on partial nucleotide sequences of seven housekeeping genes and assigned by the MLST database (). Bar, 0.0005 nucleotide substitutions per site.
Influence of the infecting serogroup and ST on the outcomes of infected dogs
Members of all five serogroups of Leptospira isolates caused lethal infections in dogs (57.1 % in Hebdomadis, 78.6 % in Australis and 100 % in Autumnalis, Canicola and Icterohaemorrhagiae) (Table 3), and there was no statistically significant difference in the mortality rate among the serogroups (P = 0.11). In the Hebdomadis serogroup isolates, the mortality rates varied among the STs detected: 0 % in ST103, 28.6 % in ST119, 77.8 % in ST37 and 80 .0% in ST118 (Table 3), but there was no statistically significant difference in the mortality rate among the STs (P = 0.063).
Discussion
Canine leptospirosis occurs worldwide and there are numerous reports on serological investigations on canine leptospirosis, but information on the causative Leptospira serotypes and genotypes and their effects on virulence in dogs remain limited. In human leptospirosis, although an association between clinical severity and specific leptospiral serogroups has been refuted (Levett, 2001), it has been suggested that a specific clone or genotype can be associated with enhanced virulence (Gouveia et al., 2008; Thaipadungpanit et al., 2007). In canine leptospirosis, a serological study indicated differences in the outcomes of dogs infected with leptospires belonging to different serogroups (Goldstein et al., 2006). However, the differences in virulence of Leptospira genotypes among infected dogs remain unknown. In this study, all five serogroups of Leptospira isolates caused mortalities in dogs, and there was no statistically significant difference in the mortality rate among the serogroups (Table 3). Conversely, the nearly significant differences in mortality rates among STs of serogroup Hebdomadis defined by MLST (P = 0.063) suggested that the virulence of Hebdomadis in dogs may be associated with the genotypes in this serogroup. This statistical value may be attributable to the small sample numbers in this study, and a larger sample size may be needed to verify the correlation between Leptospira genotypes and virulence. Reduced virulence in a specific genotype has never been reported, and a new insight into leptospiral virulence could be obtained through a comparative genomic analysis of strains of the serogroup Hebdomadis (Gulig et al., 2010; Nash et al., 2010).
Classically, L. interrogans serovars Canicola and Icterohaemorrhagiae have been associated with clinical leptospirosis in dogs worldwide. However, the incidence of infection caused by these two serovars has decreased, and other serovars have become prevalent in the USA and Europe, probably due to the widespread use of bivalent vaccines of Canicola and Icterohaemorrhagiae, as well as increased contact between dogs and wildlife reservoirs caused by expanding suburban environments (Ellis, 2010; Gautam et al., 2010; Stokes et al., 2007). In the present study, the most common Leptospira serogroup responsible for acute disease in Japan was L. interrogans Hebdomadis, followed by Australis and Autumnalis (Tables 1 and 2). Although there is no information on the proportion of each serovar, one review article described that serovars Australis, Canicola, Copenhageni and Icterohaemorrhagiae have been isolated from dogs in Japan (Yanagawa, 1992). Bivalent vaccines have been used in Japan since the 1970s, and members of serogroups Autumnalis and Hebdomadis were isolated from mice captured in Miyazaki Prefecture, which was one of the investigation sites in this study (Koizumi et al., 2008). Therefore, it is strongly suggested that the cause of the emergence of new serogroups (Autumnalis and Hebdomadis) in Japan is the same as that of the emergence of new serovars in other foreign countries. In contrast to this study, Iwamoto et al. (2009) indicated recently that antibodies against Icterohaemorrhagiae were commonly detected in healthy unvaccinated dogs throughout Japan. Subclinical disease is common in dogs, and differences in clinical outcomes among leptospiral serogroups have been suggested (Goldstein et al., 2006). Therefore, the serogroup Icterohaemorrhagiae strains present in Japan may cause subclinical infection more often than the serogroups detected in this study. Furthermore, cross-reactions and even paradoxical reactions are often observed in leptospiral infections (Levett, 2001), which may also explain the discrepancy between the previous work by Iwamoto et al. (2009) and the results of this study.
MLST has been widely employed for bacterial genotyping (Maiden, 2006) including Leptospira species (Ahmed et al., 2011; Thaipadungpanit et al., 2007). MLST is a simple molecular technique and can generate reliable, reproducible and easy-to-interpret results that are widely exchangeable compared with the results of PFGE. In the present study, MLST detected more STs than flaB sequencing (data not shown), but the same STs, i.e. ST37 and ST118, were found in more than one serogroup (Table 3). Thus, MLST is suitable for intra-serogroup genotyping but not for serogroup identification, at least for Japanese isolates. Moreover, PFGE using the restriction enzyme NotI demonstrated genetic heterogeneity in the same ST in the same serogroup (ST37 in serogroup Australis; Fig. S1), indicating that MLST has less discriminatory power than PFGE in Leptospira genome analysis.
The present study indicates that canine leptospirosis occurs in various areas of Japan. Both hunting and companion dogs were infected with leptospires, suggesting that infection with leptospires occurred in the living environments of humans, in addition to specific environments such as forests. Acute canine leptospirosis occurs in prefectures such as Ibaraki and Fukuoka where human leptospirosis has not been reported. As fewer than 50 human leptospirosis cases are reported annually, our results strongly suggest that human leptospirosis is underreported in this country. Dogs can also be asymptomatic carriers of leptospires after recovering from acute infection. Subclinical disease is common in both experimental infections (Greenlee et al., 2004, 2005) and natural infections (Stokes et al., 2007). Both dogs and humans were infected with members of the same serogroups in some areas where this study was conducted (Koizumi et al., 2008). Thus, there is a possibility that dogs serve as source of infection for humans. To confirm this possibility, further studies, including the identification of asymptomatic carriers and characterization of their isolates, need to be conducted.
Acknowledgements
We are grateful to the veterinary clinics and veterinary medical associations in Ibaraki, Chiba, Mie, Fukuoka, Saga, Nagasaki, Kumamoto, Miyazaki, Kagoshima and Okinawa for sample collection. We thank M. Suzuki, M. Nakamura, K. Taira, M. Hamasaki, S. Yahiro, K. Matsumoto and H. Osako for their assistance. This work was supported by a Health Sciences Research Grant-in-Aid for Emerging and Re-emerging Infectious Diseases from the Ministry of Health, Labour, and Welfare of Japan (H18-Shinkou-8 and H21-Shinkou-Ippan-004).