Evaluation of Indirect Fluorescence Antibody Assay for Detection of Bartonella clarridgeiae and Seroprevalence of B. clarridgeiae among Patients with Suspected Cat Scratch Disease

Abstract

The possibility of Bartonella clarridgeiae being a causative agent of cat scratch disease (CSD) was investigated by using indirect fluorescence antibody assays with 288 suspected CSD patients. Immunoglobulin G antibody to noncocultivated B. clarridgeiae was suitable only for detection of B. clarridgeiae antibody. Significant cross-reactivity between Bartonella henselae and B. clarridgeiae was noted, and no CSD case caused by B. clarridgeiae was detected.

The predominant causative pathogen for cat scratch disease (CSD) is Bartonella henselae (1, 2, 6). Since the isolation of Bartonella species from CSD patients is difficult, indirect fluorescence antibody assay (IFA) for detection of antibodies to Bartonella species is commonly used for serological diagnosis and epidemiological studies (5, 14, 19, 23).

Recently, B. clarridgeiae was suggested as an additional causative agent of CSD (9, 11). This prompted us to reanalyze sera of patients clinically suspected of having CSD, and we investigated the prevalence of immunoglobulin G (IgG) and IgM antibodies to B. clarridgeiae among such patients in Japan.

For detection of IgG and IgM antibodies to B. clarridgeiae, two IFA protocols were evaluated. To investigate cross-reactivity between B. henselae and B. clarridgeiae, 1 ml of each sample from 20 patients with serum titers of IgG antibody to B. clarridgeiae of ≥1:128 was absorbed with 2 mg of B. henselae ATCC 49882 or B. clarridgeiae ATCC 51734 and then tested by IFA. For scanning electron microscopy, B. clarridgeiae cocultivated with Vero cells on small pieces of cover glass was incubated with a 1/50 dilution of the tested sera for 90 min at 35°C in 5% CO₂, followed by the protein A-gold labeling (21), and processed and examined as described previously (20). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blot analysis were performed as described previously (10, 22).

Comparison of two IFA protocols. (i) IgG and IgM antibodies to cocultivated B. clarridgeiae.

IgG antibody to B. clarridgeiae was analyzed by IFA using B. clarridgeiae cocultivated with Vero cells for 24 and 96 h at 35°C in 5% CO₂, as described previously (19). B. clarridgeiae cocultivated with Vero cells for 24 h strongly reacted with 20 healthy human sera (1:2,048 to 1:4,056) (Fig. 1A). When B. clarridgeiae was cocultivated for 96 h, the titers increased by about eight times (1:16,384 to 1:32,768) compared to those for cocultivation for 24 h. These results suggested that titers of IgG antibody to cocultivated B. clarridgeiae might depend on cocultivation time, indicating its nonspecific reaction. Alternatively, cocultivation with Vero cells might enhance the expression of some surface antigenic determinants in B. clarridgeiae. Scanning electron microscopy also showed its nonspecific reaction (Fig. 2). Numerous gold particles were labeled on the surface of B. clarridgeiae cocultivated with Vero cells for 72 h, whereas a few gold particles were observed after 3 h of cocultivation. IgM antibody to B. clarridgeiae was not measured because many nonspecific fluorescences were observed around Vero cells in all samples.

FIG. 1.

IFA using B. clarridgeiae with healthy human or CSD patient's serum (1:128 dilution). Cocultivated bacteria with Vero cells for 24 h strongly reacted with healthy human serum (A), whereas bacteria cocultivated for 0.5 h did not react (B) but strongly reacted with the CSD patient's serum (C).

FIG. 2.

Immuno-gold labeling of B. clarridgeiae with healthy human serum. The healthy human serum nonspecifically reacted with bacteria cocultivated with Vero cells for 72 h (A), but only a few gold particles were observed on the surface of bacteria added to Vero cells and incubated for 3 h (B). Bar, 100 nm.

(ii) IgG and IgM antibodies to noncocultivated B. clarridgeiae.

Of the 20 healthy individuals, 1 had a titer of IgG antibody of 1:64 for B. clarridgeiae grown on the rabbit blood agar medium for 7 days at 35°C in 5% CO₂. The other 19 had titers that were <1:64. Nonspecific reaction was not observed in any of the 20 individuals. IgM antibody to B. clarridgeiae, analyzed for 100 healthy controls, was positive in 20 (20%), with titers of 1:20 to 1:160. Thus, IgM for B. clarridgeiae was unsuitable because of false-positive results with healthy controls.

Altogether, IgG antibody to noncocultivated B. clarridgeiae was suitable only for detection of B. clarridgeiae antibody.

The association of B. clarridgeiae with CSD.

Of the total of 100 sera from healthy individuals with no past history of either lymph node swelling or cat scratch or bite, 3 (3.2%) of 94 sera serologically negative for IgG antibody to B. henselae and 3 (50.0%) of 6 sera with titers of IgG antibody to B. henselae of 1:64 to 1:128 were positive for IgG antibody to B. clarridgeiae, with titers of 1:64 to 1:128 (Table 1).

TABLE 1.

Titers of IgG antibodies to B. henselae and B. clarridgeiae in sera of 100 healthy controls and 288 patients with suspected CSD

Group	Titer of IgG antibody to B. henselaea	No. of samples tested	No. of sera with titer of IgG antibody to B. clarridgeiaeb of:						No. (%) positive
			<64	64	128	256	512	1,024
Healthy controls	<64	94	91	2	1	0	0	0	3 (3.2)
	64-128	6	3	2	1	0	0	0	3 (50.0)
Patients	<64	145	141	4	0	0	0	0	4 (2.8)
	64-128	59	45	9	5	0	0	0	14 (23.7)
	≥256	84	39	30	12	2	0	1c	45 (53.6)

^aB. henselae cocultivated with Vero cells.

^bB. clarridgeiae not cocultivated with Vero cells.

^cTiter of IgG antibody to B. henselae was 1:4,096.

Sera from 288 patients (221 children and 67 adults) clinically suspected of having CSD because of either lymphadenopathy or fever of unknown origin and a previous history of cat scratch or contact were reevaluated for IgG antibody to B. clarridgeiae. Paired acute-phase and convalescent-phase sera were obtained from 35 patients. They were divided into three groups by titers of IgG antibody to B. henselae. Group 1 included 145 samples with titers of <1:64, group 2 included 59 patients whose sera had titers of 1:64 to 1:128, and group 3 included 84 samples with titers of ≥1:256. In group 1, 4 (2.8%) were positive for B. clarridgeiae with a titer of 1:64, and 14 (23.7%) of group 2 and 45 (53.6%) of group 3 were positive for IgG antibody to B. clarridgeiae with titers of 1:64 or more (Table 1). One patient's serum had high titers of 1:1,024 for IgG antibody to B. clarridgeiae and 1:4,096 for IgG antibody to B. henselae. Thus, the positive rate for B. clarridgeiae IgG increased significantly in proportion to the increase of the titer of B. henselae IgG among suspected CSD patients (P < 0.0001 for group 1 versus group 2; P = 0.02 for group 2 versus group 3). None of paired sera from 35 patients showed a fourfold rise in titers of IgG antibody to B. clarridgeiae. There was no difference in the positive rate for IgG for B. clarridgeiae between the control group and group 1 (P = 1.000).

Serological cross-reaction.

When the sera were absorbed with B. henselae, titers of IgG for both B. henselae and B. clarridgeiae were significantly reduced or disappeared (Table 2). When the sera were absorbed with B. clarridgeiae, titers of IgG antibody to B. henselae did not change, whereas titers of IgG antibody to B. clarridgeiae were not detected. Western blotting with a CSD patient's serum showed that it reacted predominantly with a 58-kDa protein of B. henselae and 58-, 37-, and 32-kDa proteins of B. clarridgeiae (Fig. 3B). The serum absorbed with B. henselae resulted in diminished reaction, and there was no band reacting with either strain (Fig. 3C). On the contrary, the serum absorbed with B. clarridgeiae still reacted against a 73-kDa protein of B. henselae (Fig. 3D). Although this protein band was thin, its intensity was the same as that of unabsorbed serum (Fig. 3B and D).

TABLE 2.

Titers of IgG antibodies to both B. henselae and B. clarridgeiae in sera of patients with CSD before and after absorption with B. henselae or B. clarridgeiae

Patient no.	Titer of IgG
	Before absorption		After absorption with B. henselae		After absorption with B. clarridgeiae
	B. henselae	B. clarridgeiae	B. henselae	B. clarridgeiae	B. henselae	B. clarridgeiae
1	128	128	<64	<64	128	<64
2	128	128	<64	<64	128	<64
3	128	128	<64	<64	128	<64
4	128	128	<64	<64	128	<64
5	128	128	<64	<64	128	<64
6	512	128	<64	<64	512	<64
7	512	128	<64	<64	512	<64
8	512	128	<64	<64	512	<64
9	512	128	<64	<64	512	<64
10	1,024	128	<64	<64	1,024	<64
11	1,024	128	<64	<64	1,024	<64
12	1,024	128	<64	<64	1,024	<64
13	1,024	128	<64	<64	1,024	<64
14	2,048	128	64	<64	2,048	<64
15	2,048	128	128	<64	2,048	<64
16	2,048	128	512	<64	2,048	<64
17	2,048	256	256	<64	2,048	<64
18	4,096	256	256	<64	4,096	<64
19	4,096	128	1,024	<64	4,096	<64
20	4,096	1,024	1,024	<64	4,096	<64

FIG. 3.

Cross-reaction between B. henselae and B. clarridgeiae in CSD patient's serum. (A) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of whole-cell protein of B. henselae (lane 1) and B. clarridgeiae (lane 2) incubated on blood agar. (B to D) Immunoblot analyses by using the serum from a CSD patient (B), the serum absorbed with B. henselae cells (C), and the serum absorbed with B. clarridgeiae cells (D). Arrowheads show that the 73-kDa band still reacted with the serum after absorption with B. clarridgeiae cells.

In view of the facts that cats are infected with both B. henselae and B. clarridgeiae simultaneously (4, 8, 12), B. clarridgeiae could be a causative agent of CSD (9, 11), and there was cross-reaction between B. henselae and B. clarridgeiae, as shown between B. henselae and Bartonella quintana (5, 7, 13, 15, 17, 18), serological diagnosis of CSD should be carried out with caution (3), or a specific marker, such as B. clarridgeiae flagellin (16) as antigen, is recommended to overcome the cross-reaction. With respect to association of B. clarridgeiae with CSD, our study failed to demonstrate a case of CSD caused by B. clarridgeiae. Further accumulation of CSD patients and improvement of the serological method would shed more light on the role of B. clarridgeiae in the pathogenesis of CSD.