ABSTRACT
We evaluated the utility of Western blot (WB) bands of Bartonella henselae in detecting anti-B. henselae immunoglobulin M (IgM) for serodiagnosis of cat scratch disease (CSD). IgM band patterns were examined using sera from 92 patients clinically suspected of having CSD and from 130 healthy individuals. Positive WB bands were observed in 49 (53.5%) of the 92 patient sera. Three bands at 8 to 10, 31 to 35, and 70 kDa were regarded as relevant for B. henselae because all of the positive sera yielded at least one of the three bands, and none of the healthy control sera showed reactivity to any of them. In contrast, the positive rate of the patient sera by conventional indirect fluorescence antibody assay (IFA) for B. henselae IgM was 28.3% (26/92) among the patients. These finding suggest that the IgM-WB assay, although cumbersome to perform, can be used for confirmatory diagnosis of CSD with no false positivity in the control sera. Purification of proteins in the specific bands may contribute to the development of an IgM enzyme-linked immunosorbent assay (IgM-ELISA) with improved specificity and sensitivity.
INTRODUCTION
Cat scratch disease (CSD), an infection caused by the fastidious Gram-negative bacillus Bartonella henselae, is a worldwide zoonosis associated with various clinical manifestations (1–3). Although typical clinical features at onset are fever and subacute regional lymphadenopathy with pain, the infection may manifest atypically with systematic symptoms such as encephalopathy, osteolytic lesions, pneumonia, Parinaud's oculoglandular syndrome, hepatic or splenic granuloma, endocarditis, and neuroretinitis (1–3). Annual incidence is estimated to be 22,000 cases in the United States (4) and 10,000 cases in Japan (5).
Cats are the main reservoir of B. henselae, and people become infected incidentally from cat scratches or bites. A role of the cat flea as a vector for human transmission has also been reported (3).
The laboratory diagnosis of CSD mainly depends on serological analysis because it is difficult to isolate B. henselae from patients. The indirect fluorescent antibody assay (IFA) for anti-B. henselae immunoglobulin G (IgG) is regarded as a gold standard for detecting B. henselae antibodies with high specificity. However, cross-reacting antibodies against Coxiella burnetii and Brucella spp. may compromise IFA specificity (6, 7). Besides, its sensitivity varies widely in different reports (6, 8, 9). Although the detection of IgM antibodies is crucial for the diagnosis of active B. henselae infection, the sensitivity of IFA for anti-B. henselae IgM (IgM-IFA) so far reported is very low and is lower than that of IgG-IFA (8, 10). For example, Maurin et al. (10) reported the sensitivity of 52.9% for IgG-IFA assay, but only2.9% for IgM-IFA by an evaluation employing 68 CSD patients.
The detection of anti-B. henselae IgM by various enzyme-linked immunosorbent assays (IgM-ELISA) has been proposed as an alternative to IgM-IFA (9, 11, 12). However, it is necessary to increase the sensitivity and specificity of the assay to improve its clinical utility by identifying B. henselae proteins (epitopes) that react specifically to IgM antibodies in infected patients. Therefore, it is of great importance to elucidate B. henselae antigens that exhibit reactivity to IgM from patients with suspected CSD. To explore such proteins, we developed a Western blot (WB) assay for CSD-specific IgM (IgM-WB) by use of whole-cell proteins of B. henselae. From this analysis, three of the WB bands showed very specific reactivities to the patient sera, and thus, the utility of the IgM-WB assay for a confirmatory serodiagnosis of CSD was evaluated in comparison with that of conventional IgM/IgG-IFA.
MATERIALS AND METHODS
WB analysis.Whole-cell lysate suspensions of B. henselae ATCC 49882 were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and WB analysis. These procedures are the same as those reported previously but with minor modifications (13). Briefly, the B. henselae suspension was separated by SDS-PAGE using a 5-to-20%-gradient polyacrylamide gel (ATTO, Japan). To detect human IgM, polyclonal rabbit anti-human IgM conjugated with horseradish peroxidase (HRP) (Dako) was used as secondary antibody. Super Signal West Dura substrate (Thermo Fisher Scientific) was used as the chemiluminescent substrate for the Western blotting.
Immunofluorescence assay.The in-house antigen slides for use in the IgM-IFA were prepared using B. henselae ATCC 49882 in a solution of approximately 108 cells/ml. For the IgG-IFA, the slides were prepared using B. henselae ATCC 49822, which was cocultivated with Vero cells. The details of the procedure are described elsewhere (8). The antibody titer was expressed as the highest dilution of serum that yielded a positive staining result. Sera with IgG titers of ≥1:256 or IgM titers of ≥1:20 were regarded as positive, whereas sera with IgG titers of <1:256 or IgM titers of <1:20 were regarded as negative (14). The IFA tests were performed by one well-trained technician for consistency. The Western blot assay was performed by another experienced technician. Both technicians were blinded to the clinical and laboratory findings of the specimens.
Patients and samples.WB analysis was performed on sera from 92 randomly selected patients clinically suspected of having CSD due to lymphadenopathy and/or fever of unknown origin together with a known history of close contact with a cat or dog. These specimens were obtained when a serological diagnosis for CSD using IFA was ordered to our clinical laboratory at Yamaguchi University. Of the 92 serum samples, 26 were positive (titer ≥1:20), and 66 were negative by IgM-IFA. Among the 26 IgM-positive cases, 22 were positive with IgG titers of ≥1:256, and 4 were negative with IgG titers of 1:64 to 1:128. Among the 66 IgM-negative cases, 33 were positive by IgG-IFA, and 33 were negative with titers of either 1:64 to 1:128 (15 cases) or <1:64 (18 cases).
The serum samples of 130 healthy individuals with no history of cat scratches/bites were examined as controls, and negative results for both IgG-IFA (titer: <1:256) and IgM-IFA (titer: <1:20) were obtained in all samples.
Statistical analysis.The difference in sensitivity between IgM-WB analysis and IgM-IFA was tested by chi-square test. A P value of <0.05 was considered to indicate statistical significance.
RESULTS
WB analyses.B. henselae WB bands were analyzed for their reactivity with anti-B. henselae IgM presumed to be present in sera from the patients. The sera from the healthy individuals were also analyzed as a control.
Of the 92 patient sera, 49 sera showed a variable degree of reactivity to bands that spread out over the entire range of the strip. The rate of positivity of the IgM-WB bands in patients was 53.3% (49/92). In contrast, the sera from healthy individuals were all negative except for three sera that yielded a single band at 16 kDa. Thus, the specificity of the IgM-WB assay was 97.7% (127/130). Typical patterns of anti-IgM reactive bands seen in patients are shown in Fig. 1. A total of 8 bands were detected as a whole at 8 to 10, 13 to 15, 31 to 35, 42 to 45, 60, 70, 130, and ≥254 kDa (Table 1). Although there was no band common to all patient sera, the three highest rates of positivity were observed for the bands located at molecular weights of 8 to 10 kDa, 31 to 35 kDa, and 70 kDa, with respective rates of 57.1%, 44.9%, and 30.6%. Besides, we found that all of the patients with positive bands shared at least one of the 3 major bands: i.e., the probability of the patient sera possessing at least one of the 3 bands was calculated as 100% (Table 2). Therefore, we regarded these three bands as relevant antigens representing a major immunodominant determinant of B. henselae.
Typical patterns of IgM Western blot analysis using Bartonella henselae whole-cell antigens. Western blot analysis was performed by use of B. henselae whole-cell antigens to examine the reactivities of anti-B. henselae IgM from the sera of 7 representative patients suspected of having cat scratch disease (P1 to P7) and 3 healthy control individuals (C1 to C3). Molecular size markers (kDa) are indicated in the left margin.
Patterns of IgM-reactive bands observed among the 49 subjects positive for cat scratch disease
Frequencies of anti-B. henselae IgM-Western blot bands among the 49 positive subjects
Comparison between IgM-WB analysis and IgM/IgG-IFA.The test results of the IgM-WB analysis and IgM/IgG-IFA examined in the 92 patients with suspected CSD can be compared in Table 3. The IgM-IFA assay results column represents comparison of results between IgM-IFA and IgM-WB. The positive rate in the patients of 53.3% (49/92) by the IgM-WB method was significantly higher than that of 28.3% (26/92) by IgM-IFA (P < 0.0018). It is notable that 26 IgM-IFA-positive sera were all positive by IgM-WB, whereas among the 66 IgM-IFA-negative sera, 23 of 66 (34.8%) were found to be positive by IgM-WB analysis. Although not shown in the table, the healthy control sera were all negative (100%) by IgM-IFA.
Result of method comparison between IgM-WB and IgM/IgG-IFA
The comparison between IgG-IFA and the IgM-WB assay is shown in the IgG-IFA assay results column Table 3. It is noted that 37 (67.3%) of the 55 IgG-IFA-positive sera were also positive for IgM-WB, whereas 12 (32.4%) of the 37 IgG-IFA-negative sera were found to be positive by IgM-WB. This finding indicates the ability of the IgM-WB assay to detect cases at an early stage of CSD.
DISCUSSION
There have been many reports on WB analyses that use antigens prepared from B. henselae observing many bands that react to IgG from the sera of CSD patients (15–18). However, the diagnostic utility of these bands is questionable because many of the bands were also observed among healthy controls. In contrast, there are very few reports on the utility of WB analysis for the detection of IgM in CSD patient sera as described below (15, 16). The results are mixed, and thus the utility of WB-IgM assay for the diagnosis of CSD remains to be elucidated.
In this study, we compared the performance of the IgM-WB assay using a chemiluminescent substrate with that of IgM/IgG-IFA by analyzing sera from 92 patients with clinically suspected CSD and sera from 130 healthy individuals. Unexpectedly, we observed IgM-WB-positive bands at a frequency (sensitivity) of 53.3% (49/92), whereas the specificity of the healthy control sera was 97.7% (127/130). In contrast, the sensitivity of IgM-IFA was 28.3% (26/92), although the specificity was 100% (130/130). These findings clearly point to the significantly better utility of the IgM-WB assay than the conventional IgM-IFA for the diagnosis of CSD.
In the 49 CSD patients, 8 positive IgM-WB bands were observed widely scattered between 8 and 254 kDa, with one band above 254 kDa. No single band was common to all, but all 49 sera with positive bands shared at least one of the three major bands at 8 to 10, 31 to 35, and 70 kDa. However, the positive IgM-WB bands observed in 3 of the 130 healthy sera were all located at 16 kDa. Therefore, we interpret these findings to indicate that the 3 bands shared by the sera of IgM-WB-positive patients would be specific for use in the diagnosis of CSD. The superior sensitivity of our IgM-WB assay to the IgM-IFA is obvious because 23 of 66 IgM-IFA-negative sera were found to be positive by the IgM-WB assay. Besides, all 26 IgM-IFA-positive sera were positive by our IgM-WB assay, a finding that validates its superior performance.
To our knowledge, only two reports have investigated the utility of IgM-WB analysis for the diagnosis of CSD (15, 16). McGill et al. (15) reported that they could not identify any specific bands among CSD patients, whereas Litwin et al. (16) reported the usefulness of their IgM-WB assay using B. henselae outer membrane protein. By use of 42 CSD patient sera, they compared the assay performance of their IgM-WB assay with that of IgM-IFA and IgM-ELISA. IgM-WB-positive bands were found in 12 sera (28.6%) at a common location of 8 kDa. They also showed a significant correlation between reactivity to the 8-kDa protein and positive results with IFA and ELISA, and thus concluded that the 8-kDa band could be used for the detection of IgM antibodies.
In the present study, we identified two new bands at 31 to 35 and 70 kDa in addition to the one at 8 kDa. There is a possibility of nonspecific antibodies in the CSD patient sera cross-reacting to the two bands. However, the frequencies of patient sera reacting to the bands at 31 to 35 and 70 kDa were as high as 44.9% and 30.6%, respectively. Therefore, we think it rational that they are not nonspecific reactions but relevant to the CSD patients.
The higher sensitivity achieved in our IgM-WB assay than that by Litwin et al. may be attributed to the attainment of these two additional bands, which was made possible by our use of highly sensitive chemiluminescent substrate as a coloring agent. Because no false-positive reactions were observed for the three bands in healthy sera, we believe the antigens in these bands can be used for the development of an IgM-ELISA with higher sensitivity and specificity than the present methods.
In interpreting the diagnostic utility of the IgM-WB assay in contrast to the IgG-IFA, we think the following facts are quite notable: i) 37 (67.3%) of the 55 IgG-IFA-positive sera (with titers of ≥1:256) were also positive by IgM-WB assay, and ii) 12 (32.4%) of the 37 IgG-IFA-negative sera (titers of ≤1:128) were found positive by our IgM-WB assay. Furthermore, although not described in the results, when analysis was limited to the 19 IgG-IFA-negative sera with titers between 1:64 and 1:128 (borderline negative), 10 (52.6%) were positive by the IgM-WB assay.
These differences in the detection rate between the IgM-WB assay and the IgG-IFA can be explained by the earlier but shorter detection period of IgM than IgG in the clinical course of CSD; the timing of sampling for testing thus makes a difference. Therefore, these facts reemphasize the importance of simultaneous testing for both antibodies for the serological diagnosis of CSD.
In summary, we showed that the IgM-WB assay could be used for the serological diagnosis of CSD with unprecedented high sensitivity and specificity. We also identified three IgM-WB bands at 8 to 10, 31 to 35, and 70 kDa as relevant to sera from CSD patients. The characterization of B. henselae antigens in these bands will contribute to the development of an ELISA with high accuracy for the serological diagnosis of CSD.
ACKNOWLEDGMENTS
This work was supported in part by two grants from the Japan Society for Promotion of Sciences (JSPS), one to Hidehiro Tsuneoka (no. 17K09013: 2017-19) and the other to Ken-ichiro Otsuyama (no. 15K15226: 2015-17).
We are grateful to the patients and doctors for allowing us to analyze the blood specimens of the patients with suspected CSD.
FOOTNOTES
- Received 17 August 2017.
- Returned for modification 7 September 2017.
- Accepted 19 October 2017.
- Accepted manuscript posted online 1 November 2017.
- Copyright © 2017 American Society for Microbiology.
