Prevalence of Bartonella henselae and Bartonella clarridgeiae in an Urban Indonesian Cat Population

Abstract

We studied evidence of Bartonella henselae and Bartonella clarridgeiae infection in 54 cats living in Jakarta, Indonesia. By using an indirect immunofluorescence assay, we found immunoglobulin G antibody to B. henselae in 40 of 74 cats (54%). The blood of 14 feral cats was cultured on rabbit blood agar plates for 28 days. Bartonella-like colonies were identified as B. henselae or B. clarridgeiae by using restriction fragment length polymorphism analysis and direct sequencing of the PCR amplicons. Of the cats sampled in the study, 6 of 14 (43%; all feral) were culture positive for B. henselae; 3 of 14 (21%; 2 feral and 1 pet) culture positive for B. clarridgeiae. This is the first report that documents B. henselae and B. clarridgeiae infections in Indonesian cats.

The genus Bartonella consists of 11 validated species, of which 2 are associated with cats and 4 have been shown to cause human disease. Bartonella bacilliformis, the type strain, is the etiologic agent of Carrion’s disease and is thought to be transmitted by sand flies of the genus Lutzomyia (14). Bartonella elizabethae has been isolated only once (8), and its vector and reservoir are unknown. Bartonella quintana, the etiologic agent of trench fever (34), is thought to be transmitted by the human body louse (Pediculus humanus corporis) (18, 31), and any potential natural reservoirs, other than humans, have not been demonstrated. Bartonella henselae, the etiologic agent of cat scratch disease (CSD), has been identified as a cause of bacillary angiomatosis in immunocompromised persons (22, 27). The domestic cat is the reservoir and vector for human B. henselae disease, and cat fleas (Ctenocephalides felis) may serve as a putative arthropod vector (6, 17, 18, 32). Bartonella clarridgeiae, a recent addition to the genus (21), has been isolated several times from pet and feral cats (12, 13, 19) and is suspected of having the same feline host as B. henselae. Although an association of B. clarridgeiae with human cases has been reported twice (19, 23), the role of this organism in causing human disease is unclear.

The prevalence of B. henselae immunoglobulin G (IgG) antibody in pet and feral cats from the United States, Canada, Japan, Portugal, Denmark, Austria, Switzerland, Egypt, and southern Africa has been shown to vary from 0 to 74%, depending upon geographic location (3, 9, 15, 16). Pet and impounded cats from the United States, The Netherlands, France, and Australia were often determined to be asymptomatic, but 25 to 66% were found to be B. henselae culture positive (1, 2, 5, 10, 12, 17, 20, 30, 36).

The objectives of this study were to estimate both the prevalence of B. henselae IgG antibody and Bartonella species bacteremia in a sample of the cat population of Jakarta, Indonesia.

MATERIALS AND METHODS

Bacterial strains.

The following Bartonella type strains used in this study were obtained from the American Type Culture Collection (ATCC; Rockville, Md.): B. bacilliformis KC584 (ATCC 35686), B. clarridgeiae Houston-2 (ATCC 51734), B. elizabethae F9251 (ATCC 49927), B. henselae Houston-1 (ATCC 49882), B. quintana OK90-268 (Fuller strain), Bartonella vinsonii Baker (ATCC VR-152), and Bartonella vinsonii berkoffii 93-CO1 (ATCC 51672). Bartonella grahamii V2 and Bartonella doshiae R18 were kindly provided by Richard Birtles.

Blood and serum collection.

Between October 1995 and October 1996, EDTA-treated whole blood and serum samples were collected from 74 cats (both feral and pet) residing in areas proximal to the United States Navy Medical Research Unit Number 2 (NAMRU-2) and from Center for Infectious Diseases Research at the National Institutes of Health Research and Development (P3M) facilities in Jakarta (West Java), Indonesia (6°10′ S/106°50′ E). Samples were sent to the Centers for Disease Control and Prevention (Atlanta, Ga.) for culture and serological testing. Feral cats were trapped and their ages were determined, based upon the level of erosion of permanent teeth. Pet cats were enrolled through a local veterinary clinic.

Microbiology.

Blood samples were directly plated on commercially available rabbit blood-heart infusion agar (Becton Dickinson Microbiology Systems, Cockeysville, Md.), followed by incubation at 32°C in a humidified CO₂-enriched environment (27, 35), and kept for 28 days. Cultures identified as having colony morphology consistent with Bartonella species were harvested from the plates by using sterile Dacron-tipped swabs and 2 ml of brain heart infusion broth (Becton Dickinson Microbiology Systems) and stored at −70°C.

Organisms were identified by using Gram stain, oxidase and catalase tests, and substrate utilization as measured by RapID ANAII diagnostic panels (Innovative Diagnostics Systems, Norcross, Ga.).

Controls.

Experimental controls included the purified genomic DNA of the established Bartonella species. Controls also included blood from bacteremic cats naturally infected with B. henselae and blood from nonbacteremic cats studied in our lab. The specificities of the amplified PCR products were confirmed by restriction endonuclease fragment length polymorphism (RFLP) and direct dideoxy sequencing.

Isolate identification.

DNA was extracted from the harvested bacterial material by using the QIAamp blood kit (Qiagen, Inc., Chatsworth, Calif.) in accordance with the manufacturer’s recommendations.

Two oligonucleotides homologous to the citrate synthase (gltA) gene sequences of B. henselae Houston-1 (GenBank accession no. L38987) were used as primers, BhCS781.p (5’-GGGGACCAGCTCATGGTGG-3’) and BhCS1137.n (5-AATGCAAAAAGAACAGTAAAC-3’), resulting in a 379-bp product (24).

All PCRs were performed with a PTC200 DNA-Engine (MJ Research, Watertown, Mass.). The PCR amplification was performed with 10 μl of sample in a 100-μl reaction mixture containing 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl₂, 0.001% gelatin, 0.1% Brij 35, 200 μM (each) deoxynucleotide triphosphate (dATP, dCTP, dGTP, and dTTP), 0.5 μM (each) primer BhCS781.p and BhCS1137.n, and 0.2 U of thermostable Ampli-Taq DNA polymerase (Perkin-Elmer Cetus, Norwalk, Conn.). Reaction conditions have been described previously in detail (24).

Twelve microliters of each PCR-amplified gltA product was used for gltA RFLP analysis. A panel of three restriction endonucleases was used as described in the manufacturer’s specifications in a 20-μl final volume: Hinfl, MseI, and TaqI (Promega, Madison, Wis.) (24). The digests were analyzed by electrophoresis on a 2% agarose gel in 1× Tris-borate-EDTA buffer. Gels were stained with ethidium bromide and visualized by UV fluorescence (29). RFLP patterns were compared to the RFLP patterns from all recognized Bartonella species.

The specificities of the amplified products were confirmed by direct sequencing. The primers BhCS781.p and BhCS1137.n were used to sequence the PCR products. All PCR products were sequenced in both directions with the Prism dye terminator kit (Applied Biosystems Incorporated, Foster City, Calif.) by using an ABI-Prism model 377 autosequencer (Applied Biosystems Incorporated). PCR product sequences were compared to gltA sequences available in GenBank (release 101) by using the FASTA algorithm implemented in the Wisconsin Sequence Analysis Package (Genetics Computer Group) (11).

Serologic testing.

Indirect immunofluorescence antibody testing (IFA) of all cat serum specimens was performed as previously described by using B. henselae (Houston-1) and B. quintana (OK90-268) whole cells cocultivated with E6 Vero cells (7, 28). An IFA result having a dilution end point titer of ≤64 was considered negative. Vero cells were cultured by the Centers for Disease Control and Prevention cell culture core facilities.

Data analysis.

All statistical tests based on 2 × 2 contingency tables were done by using Fisher’s exact test. Group comparisons of geometric mean titer (GMT) values was done by taking the log₁₀ transformation of each subject’s titer value and comparing group mean log₁₀ values via a t test of independent samples. The log₁₀ transformation was used to normalize the data. Because the assumption of equal population variances was violated, the P values associated with the Welch-Satterthwaite adjustment were used for interpretation. All analyses were conducted by using SPSS (release 7.5) (25).

RESULTS

Of the cats included in this study, 53 of 74 (72%) were under 1 year of age and 21 (28%) were judged to be older than 1 year of age; 57 of 74 (77%) were feral and 17 (23%) were pets; and 42 of 74 (57%) were female and 10 (14%) were male. For 22 (30%), gender was not ascertained.

Of the 74 cats tested for B. henselae IgG antibodies, 44 were positive by IFA (Table 1). Overall, there was a statistically significant association (P = 0.047) between B. henselae IgG antibody status and ownership status, with pet cats testing B. henselae antibody positive at a higher prevalence (82%) than feral cats (52%). However, when the data were split by age group, <1 year versus ≥1 year, the statistically significant association between B. henselae IgG antibody status and ownership status held true only for cats younger than 1 year of age, with 100% (9 of 9) of the pet cats testing antibody positive but only 45% (20 of 44) of the feral cats testing antibody positive (P = 0.003). Thus, generalizing that pet cats test positive more often than feral cats appears to hold true only when cats are less than 1 year of age. In fact, in the sample of cats greater than or equal to 1 year of age, the opposite appeared to be true, with a higher percentage of feral cats testing B. henselae positive (77%) than pet cats (63%). However, the difference was not statistically significant.

TABLE 1.

Prevalence of B. henselae IgG antibodies, as determined by IFA, in cats from Jakarta, Indonesia, by ownership status, age, and gender

Age (yr) and gender	No. of cats positive/total no. tested (%)
	Pet	Feral	Total
<1
Male		5/6 (83)	5/6 (83)
Female	9/11 (82)	12/16 (75)	21/27 (78)
Undetermined		3/22 (14)	3/22 (14)
Subtotal	9/11 (82)	20/44 (45)	29/55 (53)
≥1
Male	1/2 (50)	1/2 (50)	2/4 (50)
Female	4/6 (67)	9/11 (82)	13/17 (76)
Subtotal	5/8 (63)	10/13 (77)	15/21 (71)
Grand total	14/19 (74)	30/57 (53)	44/76 (58)

The B. henselae antibody titer values for all cats ranged from 31 to 2,048 with a GMT of 95.11. The B. henselae GMT for pet cats (45.33) was lower than the GMT for feral cats (119.25). The comparison of normalized mean GMT values for the pet and feral cats, after a log₁₀ transformation, was statistically significant with t(51) = 3.55 at a P of 0.001. Although cats of <1 year of age had a higher B. henselae IgG GMT than cats of ≥1 year of age, this difference held true statistically for feral cats only [with t(38) = 3.00 at P of 0.005], not pet cats [t(9) = −1.397 at P of 0.197)].

Six of 14 cats (43%; all feral) were culture positive for B. henselae and 3 of 14 (21%) were culture positive for B. clarridgeiae (2 feral and 1 pet) (Table 2). Of the cats that were culture positive for B. henselae, all had IgG antibodies to B. henselae. No cats were found to be doubly infected. Organisms resembling Bartonella species were isolated from 9 of 14 blood cultures (64%), in most cases after 7 to 14 days for B. henselae (range, 7 to 19 days) and 12 to 20 days for B. clarridgeiae (range, 12 to 28 days). Organisms were identified as being similar in enzymatic profile to Bartonella species with a RapID ANAII panel score of 000671. Catalase and oxidase tests, Gram stain, growth requirements and characteristics, and colony morphology were also consistent with Bartonella species identification. The identities of the cultured organisms were confirmed to the species level with both RFLP analysis and dideoxynucleotide sequencing of the PCR amplicons. No differences were observed between the sequences obtained in this study and those found in GenBank (release 101.0) in which all B. henselae sequences and B. clarridgeiae sequences were identical to previously released sequences (accession no. L38987 and U84386, respectively). The cats that were culture positive for B. clarridgeiae were found to have negative titers (<64) to B. henselae antigen.

TABLE 2.

Identification of bacteremia by Bartonella isolate and gender of cat from Jakarta, Indonesia

Gender	No. positive for isolate/total no. positive (%)
	B. henselae	B. clarridgeiae
Male	2/6 (33)	2/6 (33)
Female	2/6 (33)	1/6 (17)
Undetermined	2/2 (100)	0/2 (0)
Total	6/14 (43)	3/14 (21)

DISCUSSION

Cats are the zoonotic reservoir for B. henselae (17, 18, 26, 32). We demonstrated an overall B. henselae IgG antibody prevalence of 54% in our sample of cats living in Jakarta, Indonesia. The prevalence of B. henselae IgG antibody was highest among pet cats (74%), although this group had overall lower titers (GMT = 45). The prevalence observed among feral cats (53%) was lower than that among pet cats, although the titers were highest among this group (GMT = 119). The B. henselae IgG antibody prevalence data do not mirror previous observations that cats of >1 year of age have higher prevalences than cats of ≤1 year of age; however, the titers for these two groups do mirror these observations (GMTs, 117 versus 70). Feral cats of >1 year of age had much higher titers generally than pet cats of the same age group (GMTs, 154 versus 34), which is supported by previous prevalence findings. Other authors have indicated that feral cats tend to have higher prevalences of B. henselae IgG antibodies than pet cats do (4, 5, 17). The B. henselae IgG antibody prevalence in the United States has been observed to range from 41 to 85% (17, 36). The only IgG prevalence data presently available for cats of Asian countries was obtained from a large study conducted in Japan, where a 15% seroprevalence was observed (33), although a large bacteremia prevalence study was conducted in Australia (2).

The inference from our study is that susceptible human populations with cat exposure are at risk for B. henselae infection. B. clarridgeiae has been implicated in only two human cases of disease in North America. One patient manifested mild CSD-like symptoms (19), while a more seriously infected CSD patient had severe debilitating illness lasting 3 months that required hospitalization (23). The role of B. clarridgeiae in human disease remains to be fully elucidated. The organism has been isolated from cats in Europe (12, 13); however, no human cases have been identified to date. Previously identified high-risk groups for B. henselae infection include young children and immunocompromised persons, particularly AIDS patients who are exposed to young cats and kittens (17). Zangwill and colleagues (36) demonstrated that 39 of 48 cats associated with CSD patients (81%) had antibodies to B. henselae. Those authors also found that people who owned at least one kitten of ≤1 year old were 15 times more likely to contract CSD than those who owned older cats. Individuals who had been scratched or bitten by a kitten were 27 times more likely to become infected, and people who had at least one kitten with fleas were 29 times more likely to become infected than people whose animals were free of fleas.

The majority of cats in our study were 1 year of age or younger. The high prevalence of B. henselae antibody is consistent with previous observations in young cats (5, 17).

Despite the limitations resulting from the small sample size of our study, the data indicate that B. henselae infection in pet and feral cats is common in the urban center of Jakarta, Indonesia. This is the first report to document a distribution of B. clarridgeiae beyond North America and Europe, suggesting that, like B. henselae, B. clarridgeiae may also have a cosmopolitan distribution among Felis domesticus.