Canine ehrlichiosis is caused by infection with ≥ 1 tick-borne Ehrlichia spp. In the southeastern United States, the following 3 species of Ehrlichia are most commonly implicated: Ehrlichia canis, Ehrlichia chaffeensis, and Ehrlichia ewingii.1-3 Sera from Ehrlichia infected dogs will typically cross-react in immunoassays.4 An IFA that uses E canis antigen as a substrate (ie, the E canis-IFA) is commonly used to screen canine sera for prior exposure to any of the 3 Ehrlichia organisms.5 Thus, seroreactivity to E canis antigens on IFA can indicate exposure to 1 of the listed agents but not necessarily exposure to E canis.
Although not widely available, WB immunoassays that use native E canis antigens have been developed to confirm IFA results.6 However, because infection with ≥ 1 Ehrlichia spp induces cross-reactive antibodies and because unique WB patterns have not been consistently reported, the WB analysis cannot be performed to confirm infection by a specific Ehrlichia spp.1
An ELISAa that detects heartworm (Dirofilaria immitis) antigen, antibody to Borrelia burgdorferi C6 peptide, and E canis antibody is commercially available. The ELISA kit uses synthetic E canis peptides (P30, P30-1) derived from E canis immunodominant epitopes.7
The purpose of the study reported here was to compare results of the E canis-IFA, WB analyses, and commercially available ELISA for titers against Ehrlichia spp in serum samples from dogs for which a diagnosis of ehrlichiosis was suspected by the attending clinician. An approximately equal number of E canis nonreactive serum samples and serum samples with low, medium, and high titers against Ehrlichia spp were chosen from diagnostic accessions submitted to the NCSU-Vector Borne Diagnostic Laboratory for serologic testing. Serum samples were chosen on the basis of IFA titers in an effort to fully represent a range of results among the 4 groups. Serum samples were tested with WB analyses developed independently at NCSU and at a commercial laboratoryb and with the commercially available ELISA.
Materials and Methods
Serum sample selection—Serum samples that were reactive on E canis-IFA were selected from dogs for which a diagnosis of canine ehrlichiosis was suspected by the attending clinician (n = 53). Serum samples were collected over a 12-year period (1988 to 2000) and stored at −70°C. Serum samples underwent an undetermined number of freeze-thaw cycles during the period prior to testing. The commercially available ELISA and the WB analyses were performed by use of the same sample aliquot at a single site. Personnel at the NCSU-Vector Borne Diagnostic Laboratory performed all of the testing reported in our study. Serum samples were obtained from canine blood samples that were usually free of gross evidence of hemolysis or lipemia. Serum samples were chosen on the basis of the recorded titer on the E canis-IFA to include an approximately equal number of low-, medium-and high-titer serum samples. In our study, 19 low-titer serum samples (reciprocal titer, 80 to 160), 17 medium-titer serum samples (reciprocal titer, 320 to 2,560), and 18 high-titer serum samples (reciprocal titer, 5,120 to > 20,480) were used. Nonreactive control sera consisted of 8 clinical accessions that had IFA reciprocal titers of < 20 and 8 serum samples collected from specific pathogen–free dogs that were raised in a tick-free environment. A single serum sample that was reactive on E canis-IFA was from a dog experimentally infected with E canis (14 days after infection).
Serologic testing—An indirect IFA was performed at NCSU by use of E canis (Florida strain) organisms grown in DH82 cells. Starting with an initial serum dilution of 1:20, serial 2-fold dilutions of sera were applied to IFA slides, which were washed and reacted with fluorescein isothiocyanate anti-canine IgG.c Endpoint titers were determined as the last dilution at which brightly stained organisms could be detected with a fluorescence microscope with exciter and barrier filters. Seroreactivity against E canis antigens was defined as a titer of ≥ 80 (cutoff titer).
Western immunoblotting—Western blot analyses were performed by use of assays developed independently at NCSU (ie, NCSU WB analysis) and at a commercial laboratory (ie, commercial WB analysis).b For NCSU WB analysis, E canis (Florida strain) grown in DH82 cells, a continuous canine monocytoid cell line, was purified by sucrose gradient centrifugation and the protein concentration determined. Dilutions of antigen were made in final sample buffer (0.062M Tris-HCl, 5% 2-mercaptoethanol, 10% glycerol, 2% sodium dodecyl sulfate, 0.01% bromphenol blue) at a concentration of 7.5 mg/mL. Antigen was loaded by use of a volume of 20 μL/well and electrophoresed on 12.5% SDS-PAGE mini gels. Proteins were electrotransferred to 0.45-μm nitrocellulose membrane. After blocking with 5% milk in PBS solution or a solution of PBS with 0.05% Tween 20, 0.5% nonfat dry milk, and 1% normal goat serum, proteins were reacted with canine serum samples at a 1:50 dilution and with peroxidase-conjugated goat anti-canine IgG at a 1:400 dilution in 1% milk in PBS solution. Bands were detected with the color reagent 4-chloro-1-napthol. Serum samples from a dog experimentally infected with E canis (Florida strain), with a reciprocal titer of 10,240, were used as a positive control. Prestained color molecular weight standards run with each gel were used to calculate the molecular weight of protein bands.
Immunoblot patterns of the NCSU WB analysis were categorized as positive, nonreactive, or indeterminate. Sera in the positive category had immunoblot patterns of antigenic proteins that were indicative of E canis exposure with detection of the 27- to 30-kd multigene family of proteins as well as detection of 0 to 4 other proteins that are also detected routinely by IFA in serum samples from dogs experimentally infected with E canis . Sera from dogs that reacted with ≥ 1 E canis antigens but did not recognize the 27- to 30-kd, genus-specific, multigene family of proteins on NCSU WB analysis were classified as indeterminate. Nonreactive sera were defined as not reacting with any E canis antigens on NCSU WB analysis.
For commercial WB analysis, an E canis isolated was used that was cultured in canine 030.9 cells. An E canis–enriched membrane fraction was prepared by differential centrifugation and solubilized in 1% SDS. Standard methods were used for SDS-PAGE and electrophoretic transfer of these culture-derived E canis antigens to nitrocellulose. Synthetic-peptide–based E canis antigens were synthesized, conjugated to bovine serum albumin, and applied to the nitrocellulose by airbrushing. Immunoblots were blocked with nonspecific proteins; then, strips were cut from blot sheets, incubated with serum-plasma samples at a 1:50 dilution for 2 hours, washed, incubated with peroxidase-conjugated anti-canine lgGe at a 1:3,000 dilution for 1 hour, washed, and peroxidase signal-developed for 15 minutes with a commercial substrate system.f
Immunoblot patterns of the commercial WB analysis were categorized as positive, nonreactive, or indeterminate. Sera in the positive category recognized the 27- to 30-kd proteins of the culture-derived E canis antigen and 1 or both of the synthetic peptide antigens. Sera that reacted with only the culture-derived 27- to 30-kd antigens or 1 of the synthetic peptide antigens on the commercial WB analysis were classified as indeterminate. Nonreactive sera were defined as not reacting with the culture-derived or synthetic E canis antigens.
Commercially available ELISA—Detection of E canis antibody was performed by use of the commercially available ELISA.a The ELISA provides for simultaneous detection of canine heartworm antigen, E canis antibodies, and antibodies to B burgdorferi C6 peptide in canine serum, plasma, or whole blood. The ELISA format uses the proprietary assay device of the manufacturer, which provides reversible chromatographic flow of sample and automatic sequential flow of wash and enzyme substrate. Synthetic peptides that duplicate immunodominant regions of E canis surface proteins were made by use of an automated peptide synthesizer.8,g The synthetic peptides were conjugated to bovine serum albumin and to horseradish peroxidase by use of standard methods.9 The bovine serum albumin-peptide conjugate was deposited onto a porous polyethylene flow matrix. The horseradish peroxidase-peptide conjugate was added to a conjugate diluent containing nonspecific proteins and detergent. Two drops of test sample were mixed with 5 drops of conjugate and applied to the flow matrix. Ehrlichia canis antibody (if present) in the sample binds to the synthetic peptide-horseradish peroxidase conjugate and to the synthetic peptide-bovine serum albumin conjugate. The deposited synthetic peptide-bovine serum albumin conjugate was then exposed to wash and substrate reagents in the course of the assay. Results of the commercially available ELISA were considered to be positive if blue color developed in the area of deposition of the synthetic peptide-bovine serum albumin conjugate.
Results
The WB analyses and commercially available ELISA serum sample results were arranged on the basis of IFA-titer groups (ie, nonreactive, low, medium, or high; Table 1). Of the 16 serum samples in the nonreactive group (ie, titer on E canis-IFA of ≤ 20), all were nonreactive on both WB analyses and the commercially available ELISA. Likewise, of the 17 serum samples in the medium-titer group (ie, E canis-IFA reciprocal titer of 320 to 2,560) and the 18 serum samples in the high-titer group (ie, E canis-IFA reciprocal titer of 5,120 to > 20,480), all had positive results on both WB analyses and the commercially available ELISA. Results for both WB analyses were positive or nonreactive for the 51 serum samples in these 3 groups; no indeterminate results were found on NCSU or commercial WB analyses.
Results of WB analyses and the commercially available ELISAa for serum samples with various indirect IFA titers for detection of Ehrlichia canis.
| E canis-IFA reciprocal titers | No. of serum samples | No. of positive results | ||||
|---|---|---|---|---|---|---|
| NCSU WB | COM WB | NCSU WB and COM WB | ELISA | |||
| Nonreactive (>20) | 16 | 0 | 0 | 0 | 0 | 0 |
| Low (1:80 to 1:160) | 19 | 10 | 11 | 7 | 6 | 4 |
| Medium (320 to 2,560) | 17 | 17 | 17 | 17 | 17 | 17 |
| High (5,120 to >20,480) | 18 | 18 | 18 | 18 | 18 | 18 |
E canis-IFA = An IFA that uses E canisantigen as a substrate. NCSU WB = WB analysis was performed by use of assays developed independently at NCSU. COM WB = WB analysis developed at a commercial laboratory.
A variety of results were found on WB analyses and the commercially available ELISA for the 19 serum samples in the low-titer group (E canis-IFA reciprocal titer of 80 to 160; Table 2). Of the 19 serum samples in the low-titer group, 10 had positive results by use of the NCSU WB analysis, 11 had positive results by use of the commercial WB analysis, and 6 had positive results by use of the commercially available ELISA. Seven of the 19 serum samples had positive results on both WB analyses, and 4 of the 19 serum samples had positive results on both WB analyses and the commercially available ELISA. Ten of the 19 serum samples had indeterminate results in 1 or both WB analyses. Nine indeterminate results were found by use of the NCSU WB analysis and 2 by use of the commercial WB analysis. The serum sample from the single experimentally infected dog was in the low-titer group (ie, titer on E canis-IFA of 160) and had positive results on both WB analyses but was nonreactive on the commercially available ELISA.
Number of the commercially available ELISAa positive and nonreactive results and results of WB analyses for serum samples with low E canis-IFA titers.
| No. of low-titer serum samples* | WB analyses | Commercially available ELISA | ||
|---|---|---|---|---|
| NCSU | COM | Pos | Neg | |
| 7 | Pos | Pos | 4 | 3 |
| 1 | Pos | Ind | 0 | 1 |
| 2 | Pos | Neg | 0 | 2 |
| 4 | Ind | Pos | 2 | 2 |
| 1 | Ind | Ind | 0 | 1 |
| 4 | Ind | Neg | 0 | 4 |
| 0 | Neg | Neg | 0 | 0 |
E canis-IFA reciprocal titer of 80 to 160.Pos = Positive result. Neg = Negative (ie, nonreactive) result. Ind = Indeterminate result.
Discussion
Sera from dogs infected with different Ehrlichia spp can have several possible reaction patterns when evaluated by use of native E canis antigens on WB analyses, which will vary depending upon the Ehrlichia spp.1,6 Sera from E canis –, E ewingii –, and E chaffeensis –infect-ed dogs all react with similar higher molecular weight proteins ranging from 34- to 110-kd but have several possible reactions to the 27- to 30-kd proteins. On the basis of our criteria, sera must react with the 27- to 30kd proteins to confirm prior exposure to E canis by WB analysis.10 Sera from E canis –infected dogs react, sera from E ewingii –infected dogs do not react, and sera from E chaffeensis –infected dogs have variable reaction to the 27- to 30-kd proteins.6,11 The lack of consistent reaction of E chaffeensis –positive serum samples to the 27- to 30-kd proteins on WB analysis for E canis has been observed for canine and human serum samples and may be related to antigenic differences among E chaffeensis strains.6,11,12 Sera that react to ≥ 1 E canis proteins in the range of 34-to 110-kd but do not react to the 27- to 30-kd proteins were classified as indeterminate on NCSU WB analysis and nonreactive on commercial WB analysis. Therefore, WB analyses results by use of E canis antigens would be reported as positive for serum samples from E canis –infected dogs; as indeterminate (by NCSU WB analysis) or nonreactive (by the commercial WB analysis) for serum samples from E ewingii –infected dogs; and as positive, indeterminate (NCSU and the commercial WB analyses), or nonreactive (the commercial WB analysis) for E chaffeensis –infected dogs. The large proportion of serum samples (10/19) with indeterminate WB analysis results in the low-titer group, compared with serum samples in the nonreactive, medium-, and high-titer groups where no indeterminate results were found for 51 serum samples, provides evidence that a high proportion of these sera was from dogs infected with an Ehrlichia spp other than E canis.
The commercially available ELISA includes synthetic peptide reagents that duplicate amino acid sequences found in specific E canis proteins and is designed to react with sera from dogs infected with E canis. Homologous proteins expressed in strains of E chaffeensis contain the same amino acid sequence8,11 and induce antibodies in sera that are cross-reactive in the commercially available ELISA. However, not all sera from E chaffeensis–infected dogs recognize this epitope, suggesting that it is not expressed in all strains of E chaffeensis or that it is differentially expressed during the infection cycle.
In addition to the commercially available ELISA used in our study, several other ELISAs have been developed for detection of E canis antibody. These include an ELISA13,h that uses cultured whole organism as the detection antigen and an ELISA that uses the recombinant major antigenic protein 2 of E canis as the detection antigen.14 In a study comparing the performance of these 3 ELISAs with that of the E canis-IFA, the sensitivity of the 3 ELISAs ranged from 71% to 86%, which was improved to a range of 91% to 100% when only serum samples with IFA titers of > 1:320 were considered.15
In our study, the correlation of IFA, WB analyses, and the commercially available ELISA results for the 16 nonreactive, 17 medium-titer, and 18 high-titer serum samples was excellent. Each of the serum samples in the nonreactive group was nonreactive on all assays, and each of the medium- and high-titer serum samples had positive results on all assays. No indeterminate WB analyses results were found among these 3 groups. Results of IFA, WB analysis, and the commercially available ELISA correlated poorly for serum samples in the low-titer group. Only 4 of 19 low-titer serum samples had positive results in each of the WB analyses and the commercially available ELISA. These discordant results may reflect enhanced IFA sensitivity, poor IFA specificity, or exposure of dogs with low IFA titers to Ehrlichia spp other than E canis. Sera from dogs reactive to E canis antigen as measured by IFA have been found to be infected with Ehrlichia from several related spp, including E canis, E chaffeensis, and E ewingii.6,10,16 The overall less intense reaction observed in sera from dogs infected with an Ehrlichia spp other than E canis to the heterologous E canis antigen has been observed by use of IFA and WB analyses.2,6 This would result in a measurable but low IFA titer that is reduced in comparison to that of E canis –infected dogs and could explain the clustering of unconfirmed IFA results in the low-titer group.
An advantage of the commercially available ELISA is that it supplies a rapid result (approx 8 minutes) and can be used in a clinical setting. An advantage of the IFA is that results are reported in a quantitative fashion. A high IFA titer (≥ 320) would be indicative of an E canis infection. A low titer on the E canis-IFA (≤ 160) may result from a weak response to an E canis infection but may also result from an Ehrlichia infection that is not caused by E canis. The lack of specificity on IFA for E canis antibody detection, presumably reflecting potential cross-reactivity to other Ehrlichia spp, resulted in poor correlation for low-titer serum samples in our study.
In our study, serum samples with IFA reciprocal titers of ≥ 320 had uniformly positive results on both WB analyses and the commercially available ELISA. The correlation between specific IFA titers and results of the WB analyses and the commercially available ELISA is accurate for the IFA as performed in the NCSU-Vector Borne Diagnostic Laboratory. The E canis-IFA has not been standardized among veterinary diagnostic laboratories. There can be variation in IFA procedure, variation in the quality and quantity of E canis antigen used to prepare slides for IFA, and variation in the technical capacity to discern positive results and nonreactive results by fluorescence microscopy that yield substantial differences in reported IFA titer values among laboratories.1,17 A study using sequential serum samples from dogs experimentally infected with E canis, E chaffeensis, and E ewingii would be required to further define the relative sensitivity and specificity of the E canis-IFA in comparison to WB analyses and the commercially available ELISA results.
| IFA | Immunofluorescence assay |
| WB | Western blot |
| NCSU | North Carolina State University |
SNAP 3Dx, IDEXX Laboratories Inc, Westbrook, Me.
IDEXX Laboratories Inc, Westbrook, Me.
Cappel, ICN Pharmaceuticals Inc, Costa Mesa, Calif.
“Jake” isolate, North Carolina State University, Raleigh, NC.
Jackson Immunodiagnostics, West Grove, Pa.
BioRad, Richmond, Calif.
Perkin Elmer, Applied Biosystems Division, Foster City, Calif.
Immunocomb, Biogal, Kibbutz Gal'ed, Israel.
References
- 1↑
Breitschwerdt EB, Hegarty AC, Hancock SL. Sequential evaluation of dogs naturally infected with Ehrlichia canis, Ehrlichia chaffeensis, Ehrlichia equi, Ehrlichia ewingii, or Bartonella vinsonii. J Clin Microbiol 1998; 36: 2645–2651.
- 2
Dawson JE, Ewing SA. Susceptibility of dogs to infection with Ehrlichia chaffeensis, causative agent of human ehrlichiosis. Am J Vet Res 1992; 53: 1322–1327.
- 3
Dawson JE, Biggie KL, Warner CK, et al.Polymerase chain reaction evidence of Ehrlichia chaffeensis, an etiologic agent of human ehrlichiosis, in dogs from southeast Virginia. Am J Vet Res 1996; 57: 1175–1179.
- 4↑
Goldman EE, Breitschwerdt EB, Grindem CB, et al.Granulocytic ehrlichiosis in dogs from North Carolina and Virginia. J Vet Intern Med 1998; 12: 61–70.
- 5↑
Neer TM. Ehrlichiosis. In: Green CE, ed. Infectious diseases of the dog and cat. 2nd ed.Philadelphia: WB Saunders Co, 1998; 139–149.
- 6↑
Rikihisa Y, Ewing SA, Fox JC. Western immunoblot analysis of Ehrlichia chaffeensis, E. canis, or E. ewingii infections in dogs and humans. J Clin Microbiol 1994; 32: 2107–2112.
- 7
O'Connor TP, Esty KJ, MacHenry P, et al.Performance evaluation of Ehrlichia canis and Borrelia burgdorferi peptides in a new Dirofilaria immitis combination assay. In: Seward RL, ed. Recent advances in heartworm disease: symposium '01. Batavia, Ill: American Heartworm Society, 2002; 77–84.
- 8
Ohashi N, Unver A, Zhi N, et al.Cloning and characterization of multigenes encoding the immunodominant 30-kilodalton major outer membrane proteins of Ehrlichia canis and application of the recombinant protein for serodiagnosis. J Clin Microbiol 1998; 36: 2671–2680.
- 9
Hashida SM, Imagawa M, Inoue S, et al.More useful maleimide compounds for the conjugation of Fab' to horseradish peroxidase through thiol groups in the hinge. J Appl Biochem 1984; 6: 56–63.
- 10↑
Suksawat J, Hegarty BC, Breitschwerdt EB. Seroprevalence of Ehrlichia canis, Ehrlichia equi, and Ehrlichia risticii in sick dogs from North Carolina and Virginia. J Vet Intern Med 2000; 14: 50–55.
- 11
Unver A, Rikihisa Y, Ohashi N, et al.Western and dot blotting analysis of Ehrlichia chaffeensis indirect fluorescent-antibody assay-positive and -negative human sera by using native and recombinant E. chaffeensis and E. canis antigens. J Clin Microbiol 1999; 37: 3888–3895.
- 12
Brouqui P, Lecam C, Olson J, et al.Serologic diagnosis of human monocytic ehrlichiosis by immunoblot analysis. Clin Diagn Lab Immunol 1994; 1: 645–649.
- 13↑
Waner T, Strenger C, Keysary A. Comparison of a clinic-based ELISA test kit with the immunofluorescence test for the assay of Ehrlichia canis antibodies in dogs. J Vet Diagn Invest 2000; 12: 240–244.
- 14↑
Alleman AR, McSherry LJ, Barbet AF, et al.Recombinant major antigenic protein 2 of Ehrlichia canis: a potential diagnostic tool. J Clin Microbiol 2001; 39: 2494–2499.
- 15↑
Harrus S, Alleman AR, Bark H, et al.Comparison of three enzyme-linked immunosorbent assays with the indirect immunofluorescent antibody test for the diagnosis of canine infection with Ehrlichia canis. Vet Microbiol 2002; 86: 361–368.
- 16
Rikihisa YA, Ewing SA, Fox JC, et al.Analyses of Ehrlichia canis and a canine granulocytic Ehrlichia infection. J Clin Microbiol 1992; 30: 143–148.
- 17
Seaman RL, Kania SA, Hegarty BC, et al.Comparison of results of serologic testing and a polymerase chain reaction assay to determine the prevalence of stray dogs in eastern Tennessee seropositive to Ehrlichia canis. Am J Vet Res 2004; 65: 1200–1203.
