Molecular survey of Bartonella henselae and Bartonella clarridgeiae in pet cats across Japan by species-specific nested-PCR

SUMMARY

Cats are known to be the main reservoir for Bartonella henselae and Bartonella clarridgeiae, which are the agents of ‘cat-scratch disease’ in humans. In the present study, we investigated the prevalence of the two Bartonella species on 1754 cat bloods collected from all prefectures in Japan during 2007–2008 by a nested-polymerase chain reaction (PCR) targeting the 16S–23S rRNA internal transcribed spacer region. Overall, Bartonella DNA was detected in 4·6% (80/1754) of the cats examined. The nested-PCR showed that 48·8% (39/80) of the positive cats were infected with B. henselae mono-infection, 33·8% (27/80) with B. clarridgeiae mono-infection and 17·5% (14/80) were infected with both species. The prevalence (5·9%; 65/1103) of Bartonella infection in the western part of Japan was significantly higher than that (2·3%; 15/651) of eastern Japan (P < 0·001). Statistical analysis of the cats examined suggested a significant association between Bartonella infection and FeLV infection (OR = 1·9; 95% CI = 1·1–3·4), but not with FIV infection (OR = 1·6; 95% CI = 1·0–2·6).

INTRODUCTION

Bartonella bacteria are haemotropic Gram-negative organisms that persistently infect a wide variety of mammalian erythrocytes [1] and several Bartonella species have been recognized to be pathogenic for humans [2]. Bartonella henselae is known to be the agent of ‘cat-scratch disease’ (CSD) and B. clarridgeiae has also been suspected as causing a few cases of CSD in humans [3]. In addition, Bartonella koehlerae has been associated with culture-negative endocarditis in human [4].

Cats are recognized to be the reservoir for B. henselae, Bartonella clarridgeiae and B. Koehlerae, which are transmitted by cat fleas between cats [5–7]. In contrast to humans, cats infected with the Bartonella species do not usually develop any symptoms but present relapsing bacteraemia for months or years [8–10]. Epidemiological studies have been conducted in cats in many countries and the prevalence of Bartonella varies from 24·0% (65/271) to 39·5% (81/205) in USA [11, 12], 8·1% (8/99) to 16·5% (72/436) in France [13, 14], 13·0% (13/100) in Germany [15], 61·3% (19/31) in the Philippines [16], 27·6% (76/275) in Thailand [17] and 19·1% (25/131) in Taiwan [18]. In these studies, B. henselae tends to be the more dominant species than B. clarridgeiae in many countries. In Japan, the prevalence of Bartonella infection in cats from 10 prefectures was reported as 7·2% (50/690) and B. henselae was the predominant species among the isolates identified in 2000 [19]. On the other hand, B. koehlerae has never been isolated from any cats not only in Japan, but also in other Asian countries to date. Therefore, surveillance of B. henselae and B. clarridgeiae infection in cats in Japan seems to be more crucial for preventing cat-associated bartonellosis rather than that of B. koehlerae.

Isolation of Bartonella bacteria is the most important method for detection of current status of bacteraemic cats. However, primary culture of Bartonella requires long incubation periods since the bacteria are fastidious and the growth is considerably slow [20]. Several conventional and real-time polymerase chain reactions (PCRs) have been applied as alternative diagnostic tools for the isolation of Bartonella bacteria from clinical specimens [21]. On the other hand, identification of Bartonella species by DNA sequencing with PCR amplicons takes a long time and at considerable cost [22,23]. Consequently, it is necessary to develop an easy and rapid identification method for Bartonella instead of DNA sequencing at the clinical site.

The objectives of the present study are to investigate the prevalence of CSD-associated Bartonella species by a nested-PCR capable of discriminating between B. henselae and B. clarridgeiae and to evaluate correlation between FIV or FeLV and Bartonella infection in pet cats from all prefectures in Japan.

MATERIALS AND METHODS

Cat blood and DNA extraction

Between 2007 and 2008, blood samples were collected from 1754 cats in animal clinics from all (47) prefectures of Japan. In total 14–41 samples were collected from each prefecture. Western and eastern parts of Japan were divided at a longitude of approximately 137° east. The samples were transferred into commercial blood collection tubes and stored at 4 °C or −20 °C, until the DNA extraction.

Before collecting the samples, the physical condition of cats was checked and the demographic information, including age and sex were obtained by clinical veterinarians. In addition, FeLV and FIV infections in the cats were examined using the Snap Combo FeLV Ag/FIV Ab test kit (Idexx Laboratories, ME, USA). The genomic DNA of Bartonella was extracted from 200 µl of each blood sample by using QIAamp DNA Blood Mini Kit (QIAGEN, Hilden, Germany) and kept at −20 °C until further examination.

Detection of B. henselae and B. clarridgeiae DNAs by nested-PCR

A nested-PCR was applied to detect Bartonella DNAs by modifying the PCR targeting a part of the 16S-23S rRNA internal transcribed spacer (ITS) region of Bartonella species. The genus-specific primers (URBarto1 and URBarto2) reported in the previous study [24] were used for the first PCR. The primer sets used for the second PCR were newly designed in silico to be specific to B. henselae (URBhen-f and URBhen-r) or B. clarridgeiae (URBcla-f and URBcla-r), respectively (Table 1). B. henselae strains of Houston-I^T (Accession No. L35101), CAL-1 (Accession No. AF369527) and San Ant 2 (Accession No. AF369529) and B. clarridgeiae strains of Houston-II^T (Accession No. AF312497) and Kyoto19-2 (Accession No. AB674239) were used for design of the species-specific primers. The DNA sequences obtained from GenBank database were analyzed by using GENETYX-win software, version 9 (Genetyx Corp., Tokyo, Japan).

Table 1.

Primer information for the first and second PCRs used in this study

PCR	Target species	Primer	Sequence (5′→3′)	Reference
First	Bartonella species	URBarto1	CTTCGTTTCTCTTTCTTCA	[24]
		URBarto2	CTTCTCTTCACAATTTCAAT
Second	B. henselae	URBhen-f	TTGCTTCTAAAAAGCTTATCAA	This study
		URBhen-r	CAAAAGAGGGATTACAAAATC
Second	B. clarridgeiae	URBcla-f	ATGCTAAAAGTTGCTATATTGG	This study
		URBcla-r	CCTCACACTAAAATATAAAAAAC

First and second steps of the nested-PCRs were carried out in a 20 µl volume of mixture as follows: 1 µl of a 10 µM solution of each primer, 10 µl of 2 × Go Taq DNA polymerase master mix (Promega, Madison, WI, USA), 7 µl of nuclease-free distilled water and 1 µl of a sample DNA (1–10 ng/μl) for the first PCR or an amplicon of the first PCR for the second PCR. The first PCR was performed in the same way as the previous report [24]. Conditions for the second PCR were as follows: 3 min at 95 °C followed by 40 cycles of 30 s at 95 °C, 30 s at 50 °C, 1 min at 72 °C and a final extension of 3 min at 68 °C. Amplicons of the PCRs were separated by electrophoresis on 3% agarose gels and visualized by staining with ethidium bromide. Any sample showing the product size of 700–722 bp by the first PCR was considered to be positive for genus Bartonella and used for the following second PCR. B. henselae or B. clarridgeiae-specific primers for the second PCR were predicted to amplify products of 254 bp or 283 bp to 285 bp, respectively. Positive controls were the genomic DNAs from B. henselae Houston-I^T and B. clarridgeiae Houston-II^T and negative control was nuclease-free distilled water in the first PCR.

Statistical analysis

Chi-square test was applied to analyze the association between Bartonella infection and demographic factors. A P < 0·05 was considered as statistically significant in this study. In relation to the association between Bartonella and FeLV or FIV infections, the odds ratio was calculated with a 95% confidence interval.

RESULTS

Prevalence of Bartonella DNA in pet cats

The first PCR detected Bartonella DNA in 4·6% (80/1754) of the cats examined. The following second PCR demonstrated that the DNAs of B. henselae, B. clarridgeiae and both species accounted for 48·8% (39/80), 33·8% (27/80) and 17·5% (14/80) of the DNA-positive cats, respectively.

The Bartonella DNA by regions was detected in 2·3% (15/651) of cats from 7 of 18 prefectures in the eastern part of Japan (Fig. 1). In contrast, the positivity rate was found to be 5·9% (65/1103) in cats from 20 of 29 prefectures in the western part of Japan. The prevalence of Bartonella DNA in the western part was significantly higher than that observed in the eastern part of Japan (P < 0·001; Table 2).

Fig. 1.

Geographical distribution of the Bartonella-infected cats. Blue and red colors indicate the areas of eastern and western parts of Japan, respectively. The areas are segregated by a vertical dashed line located at a longitude of approximately 137° east. Asterisks represent the prefectures where Bartonella-infected cats were found.

Table 2.

Bartonella DNA prevalence in cats in Japan by respective demographic factors

Demographic factor		No. examined	No. positive for Bartonella DNA (%)	P value
Area of Japan	East	651	15 (2·3)	<0·001
	West	1103	65 (5·9)
Sex	Male	931	43 (4·6)	0·840
	Female	815	36 (4·4)
	Unknown	8	1 (12·5)	–
Age	1 year >	218	14 (6·4)	0·153
	1 year ⩽	1523	65 (4·3)
	No record	13	1 (7·7)	–

The prevalence of Bartonella DNA by gender was 4·6% (43/931) in male cats and 4·4% (36/815) in female cats. Only one of eight cats was positive for Bartonella, where gender was unknown. Prevalence by age was 6·4% (14/218) in cats under 1 year of age and 4·3% (65/1523) in cats over 1 year of age. Only one of 13 cats was positive for Bartonella where age was unknown. No significant difference was observed in the prevalence when sex and age of the cats were examined and compared (Table 2).

Table 3 summarizes the prevalence of Bartonella DNA in relation to FeLV or FIV infections of cats. Prevalence of Bartonella DNA in cats that were positive or negative for FeLV infection were 7·7% (16/209) and 4·1% (64/1545), respectively; in cats with FIV infection positive or negative the prevalence was 6·4% (26/407) and 4·0% (54/1347), respectively. Statistical analysis demonstrated a significant association between Bartonella infection and FeLV infection (OR = 1·9; 95% CI = 1·1–3·4) but not with FIV infection (OR = 1·6; 95% CI = 1·0–2·6).

Table 3.

Correlation between FeLV or FIV infections and Bartonella infection in cats

Feline immunosuppressive viral diseases		No. examined	No. positive for Bartonella (%)	OR (95% CI)
FeLV infection	Positive	209	16 (7·7)	1·9 (1·1–3·4)
	Negative	1545	64 (4·1)
FIV infection	Positive	407	26 (6·4)	1·6 (1·0–2·6)
	Negative	1347	54 (4·0)

OR, odds ratio; CI, confidence interval.

DISCUSSION

The nested-PCR used in this study could differentiate B. henselae and B. clarridgeiae infections in cat blood. Bacterial culture with blood samples and PCR and DNA sequencing with the DNA of the isolates are considered as gold-standard diagnosis methods for Bartonella infection in cats and the identification of Bartonella species. In addition, the indirect immunofluorescence assay with serum samples is also used for diagnosis of Bartonella infection. However, primary culture of Bartonella bacteria requires a long period of time, which can extend to 4 weeks because of the fastidious property of the organism [20]. Furthermore, serological tests are not adequate in understanding the present status (vs. historic legacy) of infection in cats [25]. The nested-PCR applied in this study can be a useful tool for estimating active infection in cats without culture of Bartonella bacteria. Furthermore, the nested-PCR can differentiate B. henselae and B. clarridgeiae infection in cats. Consequently, the nested-PCR may contribute to the rapid detection and identification of Bartonella in cats at the clinical site.

In the present study, Bartonella DNA was detected from 4·6% (80/1754) of the cats collected from all the prefectures in Japan using PCR. Maruyama et al. [19] showed that the prevalence of Bartonella by culture was 7·2% (50/690) in pet cats derived from 10 prefectures in Japan. It has been reported that the detection of Bartonella DNA by PCR is more sensitive than culture when performed on clinical specimens [26]. However, the positivity rate of the previous study by culture was higher than that of the present study by PCR, suggesting that the difference may be due to epidemiological biases, such as the number of prefectures examined and the study date reflecting a change in hygienic status of rearing environment. We investigated the Bartonella prevalence from cat samples collected from all prefectures in Japan, whereas only 10 out of 47 prefectures had been examined in the previous study [19]. Thus, the data of the present study may reflect the accurate status of Bartonella infection in domestic cats throughout Japan.

Our data showed that Bartonella prevalence in cats in the western part of Japan was significantly higher than that in the eastern part. According to previous studies in Asian countries excepting Japan, the prevalence of Bartonella infection in cats ranged from 19·1% in Taiwan [18] to 64·3% in Indonesia [27]. Bartonella prevalence in Asian cats varies by country, suggesting that the prevalence in tropical and sub-tropical countries tend to be higher than those of the countries with temperate climates, such as Japan. It has also been reported that cats raised in countries with high temperatures and humid climates tend to exhibit a higher seroprevalence of B. henselae than those cats, which dwell in cold and dry climates [28]. High temperature and humid climate is thought to be suitable for growth of ectoparasites such as cat fleas, which is the vector for B. henselae and B. clarridgeiae [5, 26]. Maruyama et al. [29] reported that a significantly higher seroprevalence of B. henselae was observed in cats with flea infestations than that in flea-free cats. Bartonella prevalence in cats in Kagoshima and Okinawa Prefectures located in the most south-western part of Japan was significantly higher than the other areas in Japan and flea infestation rates in the cats have also been reported to be 46·0% in Kagoshima Prefecture and 58·0% in Okinawa Prefecture. Of which, Okinawa Prefecture belong geographically to the sub-tropical region [19]. In the present study, although we could not investigate flea infestation in cats examined, the higher prevalence of this infection in the western part of Japan may possibly be attributed to the status of flea infestation in the cat population.

The percentages of PCR positive cats for B. henselae, B. clarridgeiae and both species were found to be 48·8% (39/80), 33·8% (27/80) and 17·5% (14/80), respectively. In contrast, it has been reported that B. henselae is much more prevalent (88·0%: 44/50) in cats followed by B. clarridgeiae (10·0%: 5/50), while co-infection with both species was detected only on one cat (2·0%: 1/50) [19]. The previous report suggested that [3, 30] isolation of B. clarridgeiae from cat blood appears to be more difficult than that of B. henselae. Then, the data of the present study based on PCR detection may reflect the accurate infection status of the two Bartonella species in pet cats in Japan.

Though no significant association was observed between prevalence and sex or age of the cats examined, cats <1 year-old tended to show higher prevalence than those aged over 1 year-old. Previous reports in the USA and the Netherlands also indicated that cats under 1 year of age showed a higher prevalence of B. henselae by culture and serology than the older cats [11, 31]. These facts suggest that juvenile cats rather than the adults may be more important reservoirs of CSD for humans. In fact, an association between owning a kitten was demonstrated in CSD patients in Connecticut, USA when compared with healthy persons [32].

In the present study, Bartonella infection was significantly associated with FeLV infection but not with FIV infection in the cats examined. The present data may support the hypothesis that susceptibility of Bartonella infection is possibly increased in cats with FeLV infection, regardless of whether infection is latent or progressive [33]. In contrast, it has also been suggested that Bartonella infection in cats was not affected by FIV [29, 34] nor by FeLV infections [35]. Therefore, the association between feline Bartonella bacteria and the immunosuppressive viruses remains controversial. Additional epidemiological studies should be conducted to assess the cause-and-effect relationship in other countries.