Abstract
The systematically difficult task of diagnosing Lyme disease can be simplified by sensitive and specific laboratory tests. The currently recommended two-tier test for serology is highly specific but falls short in sensitivity, especially in the early acute phase. We previously examined serially collected serum samples from Borrelia burgdorferi-infected rhesus macaques and defined a combination of antigens that could be utilized for detection of infection at all phases of disease in humans. The five B. burgdorferi antigens, consisting of OspC, OspA, DbpA, OppA2, and the C6 peptide, were combined into a fluorescent cytometric bead-based assay for the detection of B. burgdorferi antigen-specific IgG antibodies. Samples from Lyme disease patients and controls were used to determine the diagnostic value of this assay. Using this sample set, we found that our five-antigen multiplex IgG assay exhibited higher sensitivity (79.5%) than the enzyme immunoassay (EIA) (76.1%), the two-tier test (61.4%), and the C6 peptide enzyme-linked immunosorbent assay (ELISA) (77.2%) while maintaining specificity over 90%. When detection of IgM was added to the bead-based assay, the sensitivity improved to 91%, but at a cost of reduced specificity (78%). These results indicate that the rational combination of antigens in our multiplex assay may offer an improved serodiagnostic test for Lyme disease.
INTRODUCTION
With an estimated 300,000 new cases each year, Lyme disease is the most common vector-borne disease in North America (http://www.cdc.gov/lyme/stats/humanCases.html) (1). The burden of this disease in many parts of Europe is also staggering (1, 2). Treatment with antibiotics is generally effective, even more so when employed soon after infection (3). Therefore, early and reliable laboratory diagnosis is critical for effective cure of Lyme disease patients.
Patient serum antibodies specific for Borrelia burgdorferi antigens are detected with currently recommended laboratory tests for Lyme disease. While the detection of antigen may be preferred for early diagnosis, this is encumbered by the absence of detectable spirochetes or spirochetal antigen in the bloodstream once the organism has disseminated. Thus, the use of antigen detection from a blood or skin biopsy specimen has not demonstrated favorable sensitivity (4). The two most commonly used tests for diagnosis of Lyme disease in North America are (i) the two-tier test that includes an enzyme-linked immunosorbent assay (ELISA) and confirmatory Western blotting and (ii) the C6 test, where antibodies to a specific peptide within a conserved region of VlsE, the B. burgdorferi antigen, are detected (5–8). Both the two-tier and C6 tests exhibit high specificity and are most sensitive for patients in the disseminated phases of disease (5, 8–10). In addition, the C6 test has been evaluated as an indicator of treatment outcome in the United States (11, 12), with a ≥4-fold decline in antibody titer (up to 6 months after treatment) in a majority of patients. The Western blot procedure is technically demanding, and the two-tier test may require two blood draws if performed by separate laboratories. Improvement is needed for the diagnosis of patients in the early acute phase of disease and for the minority of B. burgdorferi-infected individuals who may not produce detectable anti-C6 antibodies. A simple test that can quantify antibody responses to multiple antigens with a small amount of patient serum would be of significant benefit.
In addition to the C6 peptide, a precedent exists for the use of several other B. burgdorferi proteins that are known to elicit antibody responses in natural infections. Among those which have been incorporated into Western blot-type tests are included outer surface protein C (OspC) (13, 14), fibronectin-binding protein BBK32 (p35) (15), decorin-binding protein A (DbpA) (16), flagellar protein (FlaB), VlsE (16, 17), and outer surface protein A (OspA). The temporal changes in specific antibody titers to each of these antigens postinfection can affect diagnostic accuracy. We therefore performed an assessment of longitudinal responses to multiple antigens following infection of rhesus macaques (18). These data were used as the rationale for combining specific antigens into a multiplex assay. Among the antigens tested, OspC, DbpA, and C6 each induced responses in the majority of infected animals and those responses exhibited varied kinetics. OspA was also included because levels of anti-OspA antibodies have been shown to be elevated among posttreatment Lyme disease syndrome (PTLDS) patients (19) and because we wish to develop an assay that can be used as a diagnostic test at all stages of disease. We also examined the responses to oligopeptide permease A2 (OppA2) in experimental animals, and this protein was found to be a reliable target diagnostic antigen. We hypothesized that a quantitative assay using this combination of antigens could be used to improve the sensitivity of detection of B. burgdorferi infection in patients at all phases of disease. In this report, we describe the construction and optimization of this five-antigen multiplex assay, along with the evaluation of its performance using human serum samples.
MATERIALS AND METHODS
Production of antigens and ELISAs.
OspA, OspC, and DbpA were produced as glutathione S-transferase (GST) fusion proteins by inserting the genes of interest into the pGex 4T-1 vector (GE Healthcare) as described by Embers et al. (18). The C6 peptide was synthesized by the Louisiana State University AgCenter Biotechnology Laboratory. OppA2 was produced as a glutathione S-transferase (GST) fusion protein by Biomatik, Inc. The C6 ELISA (Immunetics, Inc.) was performed according to the manufacturer's instructions.
Production of antigen-coupled cytometric beads.
A Bio-Rad Bio-Plex amine coupling kit (catalog no. 171-406001) was used to couple the antigens to the cytometric beads. The manufacturer's protocol was followed with slight modifications to include both the C6 peptide and the recombinant proteins in the same protocol. Bio-Plex Pro magnetic COOH beads, internally labeled with fluorescent dyes for individual bead identification (Microspheres catalog no. MC10043-01 for OppA2, catalog no. MC10052-01 for OspC, catalog no. MC10026-01 for OspA, catalog no. MC10062-01 for C6, and catalog no. MC10034-01 for DbpA) were vortex mixed for 30 s and then sonicated for 15 s using Bransonic ultrasonic cleaner model 1510R-DTH (Branson Ultrasonics Corporation, Danbury, CT). For bead use at 1× scale, each coupling reaction used 100 μl of beads (1.25 × 106). Monodispersed beads were placed on a DynaMag-Spin magnetic separator (Life Technologies catalog no. 12320D) for 1 min, followed by removal of the supernatant without disturbing the beads. Beads were removed from the magnet and resuspended in 100 μl of the provided wash buffer. Following a 30-s vortex procedure, the beads were placed on the separator for 1 min, followed by gentle aspiration and removal of supernatant. To the beads, 80 μl of bead activation buffer was added, followed by vortex mixing. S-NHS (N-hydroxysulfosuccinimide; Thermo Scientific catalog no. 24510) and EDAC [1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride] (Bio-Rad catalog no. 153-0990) were adjusted to a final concentration of 50 mg/ml in bead activation buffer, and 10 μl of each was added sequentially to the bead suspension. After vortex mixing, the tube was covered with foil and the beads were agitated on a rotator at 180 rpm for 20 min at room temperature. After incubation, 150 μl phosphate-buffered saline (PBS) from the kit was added and beads were vortex mixed for 10 s. The tubes were placed on the magnetic separator for 1 min. Following removal of the supernatant, the beads were washed with 150 μl of PBS. The activated beads were resuspended in 100 μl PBS and vortex mixed for 30 s. For coupling, 12 μg of peptide or protein (without removal of GST) was added to the activated beads. PBS was added to bring the total volume of the reaction to 500 μl, followed by vortex mixing. The reaction tube was covered with foil and agitated on a rotator at 180 rpm for 2 h at room temperature. After incubation, beads were placed on the magnetic separator for 1 min. While beads were on the separator, supernatant was removed and 500 μl of PBS was added, followed by vortex mixing. After bead separation (1 min) and supernatant removal, 250 μl of blocking buffer was added and the tube was then vortex mixed for 15 s. The blocking reaction mixture was covered with foil and agitated on a rotator at 180 rpm for 30 min at room temperature. After incubation, bead separation on the magnetic separator was allowed for 1 min. Following supernatant removal, the beads were resuspended in 500 μl of storage buffer. After 20 s of vortex mixing, magnetic separation was allowed for 1 min. Finally, the supernatant was removed and the beads were resuspended in 150 μl of storage buffer and stored at 4°C in the dark. The bead count was determined using a hemocytometer. The coupled beads were expected to be stable for 10 to 12 months in cold storage.
Detection of specific antibodies and controls.
The magnetic bead cocktail was prepared with assay buffer to achieve 50 μl per well. After optimizing of the ratio was performed, the final bead counts per well were 2,000 for C6, 1,000 for DbpA, 1,000 for OspA, 1,000 for OspC, and 1,000 for OppA2. After the beads were added to a 96-well plate, they were washed twice with 200 μl wash buffer using a BioPlex Pro wash station. Samples were prepared for each plate as follows: per well, 50 μl of primary antibody was diluted in sample diluent. Serum (human or monkey) was diluted 1:200, and the anti-C6 and anti-OspC monoclonal antibodies were diluted 1:1,000 (or as otherwise specified). Anti-OspA was diluted 1:100.
Anti-C6 monoclonal 4B10 was produced by Genemed Synthesis and was diluted to 1.0 mg/ml in PBS. The anti-OspC hybridoma (B5.1) was kindly provided by Robert Gilmore (Centers for Disease Control and Prevention [CDC]) and was used as a dialyzed, concentrated hybridoma supernatant at a concentration of 0.4 mg/ml. The anti-OspA (CB10) was supplied by Jorge Benach and used as a hybridoma supernatant (undiluted).
Following addition of the primary (sample) antibody, plates were sealed with light protecting sealer, shaken at 1,100 rpm for 30 s, and incubated for 1 h at 300 rpm at room temperature. Beads were washed three times with 100 μl wash buffer. The secondary (detection) antibody was diluted 1:1,000 in detection antibody buffer, and 25 μl was added to each well. Antibodies (all from Southern Biotech, Inc.) used for detection included the following: goat anti-rhesus IgG (heavy- and light-chain [H+L])-phycoerythrin (PE) (catalog no. 6200-09); goat anti-mouse IgG (H+L) R-phycoerythrin (RPE), human absorbed (catalog no. 1031-09); goat anti-human IgG-PE (catalog no. 2040-09); and mouse anti-human IgM-PE (catalog no. 9020-09). Plates were sealed with light protecting sealer and shaken at 1,100 rpm for 30 s, followed by 1 h of incubation at 300 rpm at room temperature. Beads were washed 3 times with 100 μl wash buffer. Finally, 125 μl of sheath buffer (Bio-Rad) was added to each well and the plate was sealed with new tape followed by shaking at 1,100 rpm for 30 s. Plates were stored at 4°C and were covered with aluminum foil until analysis. Typically, the assay was performed 1 day and analyzed the next day. Prior to analysis, plates were shaken at 1,100 rpm. Samples were read on a Bio-Plex 200 suspension array system and analyzed using Bio-Plex Manager v6.1 software (Bio-Rad Laboratories).
Animals, infection, and treatment.
Practices in the housing and care of animals conformed to the regulations and standards of the PHS Policy on Humane Care and Use of Laboratory Animals and to the Guide for the Care and Use of Laboratory Animals. The Tulane National Primate Research Center is fully accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care International. The Institutional Animal Care and Use Committee of the Tulane National Primate Research Center approved all animal-related protocols, including the infection and treatment of and sample collection from nonhuman primates. All animal procedures were overseen by veterinarians and their staff. For blood collection, monkeys were anesthetized with ketamine (10 mg/kg of body weight) by intramuscular injection.
In a separate study, 10 male rhesus macaques (Indian origin), 3 to 4 years of age, were fed upon by 10 to 20 Ixodes scapularis nymphs harboring B. burgdorferi strain B31. Infection was confirmed by skin biopsy culture and PCR. At 4 months postinoculation (p.i.), five animals received antibiotic treatment consisting of one 25 mg tablet of doxycycline (Bio-Serv) administered twice a day for 28 consecutive days. This dose corresponded to ≥5 mg/kg/day to ensure that an effective blood level was achieved. Blood was collected at 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24 or 25, 30 or 31, 32 or 33, 34 or 35, 39, 47 or 48, 50, 58, and 69 weeks p.i.
Patient samples.
Two separate panels of serum samples were obtained from the CDC. The first panel consisted of 32 samples; 12 of the samples were derived from patients at various stages of Lyme disease, including early/acute, early/convalescent, Lyme arthritis, and neurological Lyme disease (disseminated phase), and the other 20 were derived from control subjects and patients with “look-alike” diseases, including fibromyalgia, infectious mononucleosis, multiple sclerosis, rheumatoid arthritis, severe periodontitis, and syphilis. The second panel consisted of 92 samples: 32 samples obtained from Lyme disease patients and 60 control samples. In total, the patient samples (n = 44) consisted of 14 from patients with early disease (acute phase), 14 from patients with early disease (convalescent phase), 8 from patients with Lyme arthritis, 2 from patients with early cardiac Lyme disease, and 6 from patients with neurologic Lyme disease. The acquisition and characterization of patient samples in the CDC biorepository have been described in detail previously (20).
RESULTS
The dynamic linear range of specific antibody detection using the Multiplex assay is much broader than that of standard ELISA.
In order to examine the potential for this assay to be used to quantify and compare antibody titers, we performed standard ELISA and the Multiplex assay using multiple dilutions of an anti-C6 monoclonal antibody. As shown in Fig. 1A, using standard ELISA, the upper detection limit is an unchanged absorbance of ∼3.8 (in the dilution range of 1:800 to 1:4,000) which drops an optical density at 450 nm (OD450) of less than 1.0 with a 50-fold dilution of 1:20,000. In contrast, a linear decline over the entire range of dilution was detected using the Multiplex assay, with a mean fluorescence intensity (MFI) change of 11,000 (from ∼15,000 to 4,000) over the 50-fold dilution. By using a series of 25- to 50-fold dilutions, we compared the linear ranges of the assays using a best-fit (R2) calculation. Here the reciprocal dilution (log2) was plotted against the assay value (OD450 or MFI). As shown in Fig. 1, the R2 for dilutions in the range of 1:800 to 1:40,000 or for the four lowest dilutions (1:8,000 to 1:40,000) demonstrated that the linear range is broader for the Multiplex assay (0.6608 versus 0.8443) and is also a better linear fit for the lower antibody concentrations (0.9512 versus 0.9885). This implies that the use of standard curves with monoclonal antibodies for quantification of responses would be more accurate with the cytometric bead assay than with the ELISA. While more dilutions on the upper and lower ends are needed to determine the actual dynamic range, these data indicate improved precision with the Multiplex test.
FIG 1.
Dilution of anti-C6 monoclonal antibody shows the broader linear range of the Multiplex assay in comparison to standard ELISA. Shown is the dilution of anti-C6 monoclonal antibody (MAb) using ELISA (A) or the cytometric bead Multiplex assay (D) on a linear scale. To determine the R2 value, each result was plotted using the Log base 2 reciprocal dilution. Shown are the values determined using ELISA (B and C) and the multiplex assay (E and F) with dilutions from the series of dilutions and the lower range (1:8,000 to 1:40,000), respectively.
The Multiplex assay more accurately reflects differences in antigen-specific antibody levels.
In order to evaluate how the antibody responses within the range found in serum of infected animals were represented by both methods, we compared the levels of antibodies to three antigens using serially collected serum from two monkeys infected with B. burgdorferi by tick feeding. One animal (IH11) was treated with doxycycline between weeks 16 to 20, and the other animal (IP55) was not treated. Using comparisons of data generated from standard ELISA and the Multiplex assay (Fig. 2), the patterns of antibody reactivity over time were shown to be similar. However, the Multiplex assay was more sensitive to differences in antibody level. For example, comparing anti-OspC responses, a sharp increase in antibody level was detected at week 30 for animal IP55 by the Multiplex assay, but it was observed as a minor increase by standard ELISA. Comparing the anti-DbpA responses, the patterns of antibody response for the two animals looked similar by ELISA but separated sharply after week 30 as shown by the Multiplex assay results. A decline in C6 titer was expected following antibiotic treatment (11), and this was observed for animal IH11. The ELISA indicated a steady decline to below baseline (the average of preimmune serum values) by week 30, whereas the Multiplex assay indicated that levels rose slightly after the beginning of treatment and did not drop below baseline until week 35; a low level of antibody was also detected at week 58 with the Multiplex assay.
FIG 2.
Comparison of IgG antibody detection by ELISA to that by Multiplex assay in longitudinal samples from infected monkeys reveals sensitivity differences. Shown are standard ELISA results versus Multiplex results using antigens OspC (A and D), DbpA (B and E), and C6 (C and F). One animal (IH11) was treated with doxycycline between week 16 and week 20, and the other animal (IP55) was not treated.
The Multiplex assay reveals specific differences in the polyvalent antibody responses during infection.
Results from a time course analysis of antibody responses to multiple antigens by B. burgdorferi-infected monkeys are shown in Fig. 3. Each animal, infected by tick-mediated transmission, showed a different profile of reactivity. For each of the antigens utilized in this assay, the ranges of antibody levels (i.e., the MFIs) for animals that produced positive responses were similar. For example, the levels of anti-DbpA antibodies were much higher than those of anti-C6 or anti-OspA antibodies.
FIG 3.
The Multiplex assay reveals specific differences in the polyvalent antibody responses during infection. Data represent results of time course analysis of antibody responses to multiple antigens by B. burgdorferi-infected monkeys (mk). Antigen specificity was shown by testing monoclonal antibody (MAb) reactivity to C6, OspA, and OspC or to a well-characterized serum sample from a B. burgdorferi needle-inoculated animal (ctrl mk serum). 1° antibody, primary antibody; 2°, secondary antibody.
As expected, significant variation in antibody reactivity was seen among human samples. Multiplex test results from a panel of 12 Lyme disease patients and multiple control patients from look-alike disease categories are shown in Fig. 4. Lyme disease patients typically produced specific antibodies for one or more antigens, but only one Lyme arthritis patient (LA in the bottom panel of Fig. 4) in this panel had significant levels of antibodies against all five antigens. Only one of four early-phase/acute-phase Lyme disease patients (EL/A) tested positive, and this was detected by using anti-DbpA antibodies. Importantly, very little reactivity to control samples was detected; the exception was the weak responses to multiple antigens shown by two rheumatoid arthritis (RA) patients (bottom panel) and to OppA2 by one healthy control (top panel) from a region of nonendemicity (HNE). However, OppA2-GST (produced in house) was the source antigen for this set of experiments. A purer form was later acquired from an industry source and utilized for subsequent analyses (compiled in Tables 1 and 2). The IgG results displayed in the top panel, in summary, showed that we achieved 100% sensitivity for Lyme disease patients in the convalescent phase (EL/C) with Lyme arthritis (LA) or neuroborreliosis (NL). Of the patients in the acute phase (EL/A), 1 of 4 was positive (25%) using the Multiplex assay whereas 0 of 4 (0%) was positive using the 2-tier test.
FIG 4.
Results from human serum samples determined using the five-antigen multiplex assay. The MFI values (y axis) of antibodies to each particular antigen are indicated by colored bars. The patient categories are indicated below the x axis, according to the following key: healthy, from an area of endemicity (HE); healthy, from an area of nonendemicity (HNE); fibromyalgia (FB); infectious mononucleosis (MN); multiple sclerosis (MS); rheumatoid arthritis (RA); severe periodontitis (PD); syphilis (SY); early Lyme/acute phase (EL/A); early Lyme convalescent phase (EL/C); Lyme arthritis (LA); and neuroborreliosis (NL). Each sample indicated on the x axis represents an individual patient.
TABLE 1.
Comparison of the 5-antigen Multiplex assay to the EIA and 2-tier test shows improved sensitivity while maintaining specificitya
| Patient category | 5-antigen Multiplex-IgG assay |
EIAb |
2-tierb,c |
C6 |
|||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No. of subjects with indicated result/total no. of subjects |
% sens | % spec | |||||||||||||||
| OspA | OspC | C6 | DbpA | OppA2 | Total | No. of subjects with indicated result/total no. of subjects | % sens | % spec | No. of subjects with indicated result/total no. of subjects | % sens | % spec | No. of subjects with indicated result/total no. of subjects | % sens | % spec | |||
| Lyme disease | |||||||||||||||||
| Early/acute phase (n = 14) | 0/14 | 1/14 | 3/14 | 3/14 | 5/14 | 7/14 | 50 | 4/12* | 33.3 | 1/14 | 7.1 | 6/14 | 42.8 | ||||
| Early/convalescent phase (n = 14) | 0/14 | 3/14 | 12/14 | 6/14 | 11/14 | 13/14 | 92.9 | 13/14 | 92.9 | 11/14 | 78.6 | 12/14 | 85.7 | ||||
| Arthritis (n = 8) | 3/8 | 4/8 | 8/8 | 8/8 | 8/8 | 8/8 | 100 | 8/8 | 100 | 8/8 | 100 | 8/8 | 100 | ||||
| Neuro (n = 6) | 0/6 | 2/6 | 5/5* | 3/6 | 4/6 | 5/6 | 83.3 | 5/6 | 83.3 | 5/6 | 83.3 | 6/6 | 100 | ||||
| Cardio (n = 2) | 0/2 | 0/2 | 2/2 | 2/2 | 2/2 | 2/2 | 100 | 2/2 | 100 | 2/2 | 100 | 2/2 | 100 | ||||
| Non-Lyme disease | |||||||||||||||||
| Syphilis (n = 6) | 0/6 | 1/6 | 2/5* | 1/6 | 1/6 | 2/6 | 66.7 | 5/6 | 16.7 | 1/6 | 83.3 | 0/5* | 100 | ||||
| Mononucleosis (n = 6) | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 100 | 2/5* | 60 | 0/6 | 100 | 1/6 | 83.3 | ||||
| Periodontitis (n = 6) | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 100 | 0/6 | 100 | 0/6 | 100 | 0/5* | 100 | ||||
| Multiple sclerosis (n = 6) | 0/6 | 1/6 | 0/6 | 0/6 | 0/6 | 1/6 | 83.3 | 1/6 | 83.3 | 0/6 | 100 | 0/6 | 100 | ||||
| Fibromyalgia (n = 6) | 0/6 | 0/6 | 0/6 | 0/6 | 1/6 | 1/6 | 83.3 | 0/6 | 100 | 0/6 | 100 | 0/6 | 100 | ||||
| Rheumatoid arthritis (n = 6) | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 100 | 2/6 | 66.7 | 1/6 | 83.3 | 0/6 | 100 | ||||
| Healthy endemic (n = 12) | 0/12 | 0/12 | 0/12 | 1/12 | 1/12 | 1/12 | 91.7 | 3/11* | 72.7 | 0/12 | 100 | 0/12 | 100 | ||||
| Healthy non-end stage (n = 12) | 0/12 | 0/12 | 0/11* | 0/12 | 0/12 | 0/12 | 100 | 1/11* | 90.9 | 0/12 | 100 | 1/12 | 91.6 | ||||
| All Lyme disease | 3/44 | 10/44 | 30/44 | 22/44 | 30/44 | 35/44 | 79.5 | 32/42 | 76.1 | 27/44 | 61.4 | 34/44 | 77.2 | ||||
| All non-Lyme disease | 0/60 | 2/60 | 2/60 | 2/60 | 3/60 | 5/60 | 91.7 | 14/57 | 75.4 | 2/60 | 96.7 | 2/58 | 96.6 | ||||
sens, sensitivity; spec, specificity; *, equivocal results excluded.
Data are from a Vidas Lyme IgM and IgG polyvalent assay by bioMérieux, Inc.
Data are from IgM and IgG immunoblotting assays by MarDx Diagnostics, Inc.
TABLE 2.
Inclusion of IgM testing improves sensitivity but reduces specificity
| Patient category | 5-antigen Multiplex-IgM assay |
IgG + IgM test |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| No. of subjects with indicated result/total no. of subjects |
% sens | % spec | % sens | % spec | ||||||
| OspA | OspC | C6 | DbpA | OppA2 | Total | |||||
| Lyme disease | ||||||||||
| Early/acute phase (n = 14) | 0/14 | 3/14 | 4/14 | 4/14 | 1/14 | 7/14 | 50 | 78.5 | ||
| Early/convalescent phase (n = 14) | 1/14 | 6/14 | 4/14 | 5/14 | 4/14 | 9/14 | 64.3 | 100 | ||
| Arthritis (n = 8) | 1/8 | 1/8 | 1/8 | 0/8 | 0/8 | 2/8 | 25 | 100 | ||
| Neurological disease (n = 6) | 0/6 | 4/6 | 1/6 | 1/6 | 1/6 | 4/6 | 66.7 | 83.3 | ||
| Cardiac disease (n = 2) | 0/2 | 1/2 | 1/2 | 2/2 | 1/2 | 2/2 | 100 | 100 | ||
| Non-Lyme disease | ||||||||||
| Syphilis (n = 6) | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 100 | 66.7 | ||
| Mononucleosis (n = 6) | 0/6 | 2/6 | 1/6 | 2/6 | 0/6 | 2/6 | 66.7 | 66.7 | ||
| Periodontitis (n = 6) | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 0/6 | 100 | 100 | ||
| Multiple sclerosis (n = 6) | 0/6 | 2/6 | 2/6 | 0/6 | 0/6 | 2/6 | 66.7 | 66.7 | ||
| Fibromyalgia (n = 6) | 0/6 | 2/6 | 2/6 | 2/6 | 0/6 | 2/6 | 66.7 | 66.7 | ||
| Rheumatoid arthritis (n = 6) | 1/6 | 1/6 | 0/6 | 0/6 | 0/6 | 2/6 | 66.7 | 66.7 | ||
| Healthy, from area of endemicity (n = 12) | 1/12 | 1/12 | 2/12 | 2/12 | 0/12 | 2/12 | 83.3 | 75 | ||
| Healthy, from area of nonendemicity (n = 12) | 0/12 | 0/12 | 2/12 | 0/12 | 0/12 | 2/12 | 83.3 | 83.3 | ||
| All Lyme patients | 2/44 | 15/44 | 11/44 | 12/44 | 7/44 | 24/44 | 54.5 | 90.9 | ||
| All non-Lyme patients | 2/60 | 8/60 | 9/60 | 6/60 | 0/60 | 12/60 | 80 | 78.3 | ||
Comparison of IgG results from the 5-antigen Multiplex assay to those from the EIA and 2-tier test shows that sensitivity was improved while specificity was maintained.
Following the analysis of the 32-sample CDC panel (panel I), we acquired an additional panel of 92 samples; 32 were obtained from Lyme disease patients and 60 from controls (panel II). Each sample was tested “blind,” the results were sent to the CDC, and then the patient sample information was shared. This information included (i) the disease identity, (ii) the enzyme immunoassay (EIA) test results, and (3) Western blot results (combined to give 2-tier result). The C6 ELISA was performed by our laboratory, using a kit supplied by Immunetics, Inc. Before the identities of the sample categories were obtained, we could not establish the mean and standard deviation values of the negative-control samples in order to establish cutoff values for positivity. Once we had tested the samples, the healthy controls that were included on each plate were used to establish the cutoff as the mean + 3× the standard deviation. When obtained, equivocal results were excluded.
The combined results obtained for panels I and II are shown in Table 1. After obtaining the disease identity for each sample, we determined that single-antigen positivity with respect to C6, OppA2, or DbpA was specific, whereas single responses to OspC were not because the only case where OspC alone was detected did not represent a Lyme patient. However, for the purposes of reporting the data, we have listed any sample with a single positive result within the five-antigen test as positive. Among the five antigens, the Lyme disease patients most frequently had specific antibodies to C6 and OppA2. Responses to DbpA were also common, but only one patient was singly positive for DbpA (Fig. 4, top panel [EL/A]), whereas multiple patients were singly positive for C6 or OppA2. Responses to OspA were confined to Lyme arthritis patients, and, surprisingly, only 4 of 28 early-phase Lyme disease patients possessed detectable anti-OspC antibodies. Finally, the only patients who showed antibodies to all 5 antigens were a few of the Lyme arthritis patients.
Several of the samples that produced “false-positive” results by the Multiplex assay were scored as positive by EIA and/or Western blot analysis. For example, one of our false positives was from a healthy patient in an area of endemicity. The EIA was positive, and 4 bands were present on the Western blot (data not shown). Since 5 bands were required to meet the threshold, it was scored negative by the 2-tier test, although the possibility that this person had been exposed to B. burgdorferi cannot be ruled out. One of our false-positive samples was from a fibromyalgia patient who was dual-antigen positive by our test and also by Western blot analysis (not shown). We therefore included the specificity achieved against all of the controls who would have had the lowest likelihood of exposure (healthy patients from areas of nonendemicity) in our analysis and found it to be 100%. For the EIA, the two-tier test, and C6, the specificities against HNE patients were 90.9%, 100%, and 91.6%, respectively. The comparisons of the levels of sensitivity (true positive rate) and specificity (true negative rate) obtained using IgG results only are shown in Table 1.
Comparison of diagnostic test results from early-phase Lyme disease patients shows the highest level of sensitivity using the 5-antigen Multiplex assay.
We compared the results obtained from early-phase Lyme disease patients in the acute phase using the 5-antigen multiplex assay with those obtained with the EIA, 2-tier test, and C6 ELISA and found that the 5-antigen multiplex assay was more sensitive than the others (Table 1). Specifically, the detection in the early acute phase was 50%, showing a distinct improvement over the EIA (33.3%) and the 2-tier test (7.1%); the C6 test was a close second with 42.8% sensitivity. Our multiplex assay also showed high sensitivity (92.9%) in the early convalescent stage. Combining these results, the early-phase sensitivity of the multiplex assay was 20/28, or 71.4%, which is a moderate improvement over the sensitivities obtained for the C6 test (18/28 [64.3%]) and EIA (17/26 [65.4%]) but a significant improvement over that obtained for the 2-tier test (42.9%). Importantly, we saw between 83% and 100% sensitivity for all other Lyme patient categories. The improved sensitivity came with a moderate decrease in specificity compared to levels seen with the 2-tier (96.7%) and C6 (96.6%) tests, which remained over 90% (91.7%).
The multiantigen assay IgM responses showed greater sensitivity for early-phase Lyme patients but lower specificity.
In order to analyze the IgM responses using our Multiplex assay, we needed to establish a cutoff value. In general, we found the background levels of IgM against all of the antigens to be high. A sample of the results is shown in Fig. S1 in the supplemental material. For each plate, we used the values from control patients to determine the baseline. However, the wide range and variation made the establishment of cutoff values using the few available samples likely less accurate. The IgM results on their own demonstrated lower sensitivity and specificity than the IgG results (Table 2). As expected, the values for IgM specificity were higher in the early phases of infection (50% for acute-phase patients; 64% for convalescent-phase patients) than in the late stage (25% for Lyme arthritis patients). By combining the results from the IgG and IgM Multiplex assays (Table 2), we could improve the detection of early-phase Lyme disease and increase the overall sensitivity to 90.9%. However, the specificity would decline to 78%.
DISCUSSION
The goal of developing a reliable diagnostic test that can detect B. burgdorferi sensu lato infection in patients at all phases of disease (21) is being pursued with multiple methods and technologies. From the perspective of the use of disease-related biomarkers, approaches involving metabolomic and immunologic profiles have shown promise (22–24). In particular, a liquid chromatography-mass spectrometry-based analysis of metabolites was shown to correctly classify early-phase Lyme disease patients and healthy controls with a sensitivity of 88%. Given that the test is not specific to B. burgdorferi, a remarkable specificity of 95% was achieved in tests against “look-alike diseases” such as fibromyalgia and rheumatoid arthritis (22). However, patients with late-stage disease were not tested and the potential for the levels of biomarkers included in the profile to wane over time is a concern. Early detection of infection has also seen improvement with PCR-based and antigen detection methods in blood (25, 26). This relies on B. burgdorferi or its antigens being present in the blood, a condition which is confined to a small window of time in early infection.
Antibody-based detection methods have also seen improvements. Here, test performance has been and continues to be weak the early stage of infection, but the overall sensitivity of newer tests compared to that of two-tier testing has improved. The C6 ELISA, which utilizes a peptide derived from a conserved region (C6) of the VlsE protein (27), has been shown to perform as well as or better than the two-tier test (5, 10), depending on the disease phase. In addition, it has been shown to have potential as an indicator of treatment outcome (11, 12). Peptides derived from another B. burgdorferi sensu lato protein (OppA2) have also been shown to detect specific antibodies in Lyme disease patients (28). Most recently, multiplex assays for the detection of infection in dogs (29) and in humans (30) have shown strong potential for this diagnostic platform. Specifically, a combination of 10 antigens was used to test serum antibodies from Lyme disease patients at different phases (30). This test showed improved sensitivity over the two-tier test (87.5% versus 67.5%) in the early phase but was not compared to the C6 test or tested against samples from patients with multiple look-alike diseases for specificity.
Here we report the results from preliminary testing of a five-antigen multiplex assay for diagnosis of Lyme disease. These results were compared to those obtained for the EIA, the two-tier test, and the C6 ELISA. We saw improved sensitivity in all categories of patients using the Multiplex test compared to the other tests. The improvement over the C6 test was only moderate, but it was more evident in the comparisons to the two-tier test and EIA for sensitivity and specificity, respectively. Comparison of the multiplex assay to the C6 test specifically showed improvement for early-stage detection, while the improved sensitivity in later stages was also met with reduced specificity. In several cases, the C6 ELISA result was positive and a low titer of anti-C6 antibody was detected by the multiplex assay; however, the level was below the cutoff, so it was scored as negative by the multiplex assay. Establishment of an index for positivity should improve multiplex test performance in this regard. The antigens most frequently targeted by patient antibodies were C6, OppA2, and DbpA, with single-positive detection obtained using C6 and OppA2. The use of OspA and OspC added little to the sensitivity of the assay. We expected that few patients would have anti-OspA antibodies, as these have been shown to be restricted to patients with Lyme arthritis and/or PTLDS (19, 31). However, we elected to keep this antigen as part of the assay (i) in order to cover all patient categories and stages and (ii) because nonspecific responses are infrequent. We did, however, expect a higher proportion of patients to possess anti-OspC antibodies that would be detectable by this assay. The OspC antigen is a full-length molecule derived from B. burgdorferi strain B31. This strain has been classified as having ospC genotype 1 (type A) (32, 33). Given the sequence diversity of this gene, seroreactivity can vary in accordance with the infecting strain's ospC type (34). By inclusion of a recombinant OspC protein from another genotype, such as type K, we may be able to improve the sensitivity (34). Inclusion of other diagnostic antigens that have shown strong potential may also improve the assay (35).
In order to establish a more solid index of positivity, high numbers of healthy negative-control patients are needed. For minimizing plate-to-plate variation, standard curves determined using monoclonal antibodies against C6 and OppA2 are required and under development. This will also facilitate quantification of antibody titers, which may be of tremendous value in measuring treatment response. These two antigens appear to be essential for test performance. OppA2 incidentally showed 100% specificity with respect to IgM antibody detection (Table 2). By reducing the number of antigens used for IgM testing, in addition to the establishment of proper cutoff values, we can improve upon this platform test. In summary, we have demonstrated that a multiplex bead-based assay that is confined to a few select antigens can provide specific and sensitive serodiagnosis of Lyme disease.
Supplementary Material
ACKNOWLEDGMENTS
This research was supported by the National Institute of Allergy and Infectious Diseases, grants 1R21AI100166-01 and 4R33AI100166-03.
Mario T. Philipp is an inventor and an owner of U.S. patent no. 6,475,492 and its foreign counterparts, which are licensed to Immunetics, Inc.
Footnotes
Supplemental material for this article may be found at http://dx.doi.org/10.1128/CVI.00685-15.
REFERENCES
- 1.Mead PS. 2015. Epidemiology of Lyme disease. Infect Dis Clin North Am 29:187–210. doi: 10.1016/j.idc.2015.02.010. [DOI] [PubMed] [Google Scholar]
- 2.van den Wijngaard CC, Hofhuis A, Harms MG, Haagsma JA, Wong A, de Wit GA, Havelaar AH, Lugnér AK, Suijkerbuijk AWM, van Pelt W. 16 June 2015. The burden of Lyme borreliosis expressed in disability-adjusted life years. doi: 10.1093/eurpub/ckv091. [DOI] [PubMed] [Google Scholar]
- 3.Asch ES, Bujak DI, Weiss M, Peterson MG, Weinstein A. 1994. Lyme disease: an infectious and postinfectious syndrome. J Rheumatol 21:454–461. [PubMed] [Google Scholar]
- 4.Aguero-Rosenfeld ME, Wang G, Schwartz I, Wormser GP. 2005. Diagnosis of Lyme borreliosis. Clin Microbiol Rev 18:484–509. doi: 10.1128/CMR.18.3.484-509.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bacon RM, Biggerstaff BJ, Schriefer ME, Gilmore RD Jr, Philipp MT, Steere AC, Wormser GP, Marques AR, Johnson BJ. 2003. Serodiagnosis of Lyme disease by kinetic enzyme-linked immunosorbent assay using recombinant VlsE1 or peptide antigens of Borrelia burgdorferi compared with 2-tiered testing using whole-cell lysates. J Infect Dis 187:1187–1199. doi: 10.1086/374395. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mogilyansky E, Loa CC, Adelson ME, Mordechai E, Tilton RC. 2004. Comparison of Western immunoblotting and the C6 Lyme antibody test for laboratory detection of Lyme disease. Clin Diagn Lab Immunol 11:924–929. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Nyman D, Willen L, Jansson C, Carlsson SA, Granlund H, Wahlberg P. 2006. VlsE C6 peptide and IgG ELISA antibody analysis for clinical diagnosis of Lyme borreliosis in an endemic area. Clin Microbiol Infect 12:496–497. doi: 10.1111/j.1469-0691.2006.01374.x. [DOI] [PubMed] [Google Scholar]
- 8.Steere AC, McHugh G, Damle N, Sikand VK. 2008. Prospective study of serologic tests for Lyme disease. Clin Infect Dis 47:188–195. doi: 10.1086/589242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Embers ME, Jacobs MB, Johnson BJ, Philipp MT. 2007. Dominant epitopes of the C6 diagnostic peptide of Borrelia burgdorferi are largely inaccessible to antibody on the parent VlsE molecule. Clin Vaccine Immunol 14:931–936. doi: 10.1128/CVI.00075-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wormser GP, Schriefer M, Aguero-Rosenfeld ME, Levin A, Steere AC, Nadelman RB, Nowakowski J, Marques A, Johnson BJB, Dumler JS. 2013. Single-tier testing with the C6 peptide ELISA kit compared with two-tier testing for Lyme disease. Diagn Microbiol Infect Dis 75:9–15. doi: 10.1016/j.diagmicrobio.2012.09.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Philipp MT, Bowers LC, Fawcett PT, Jacobs MB, Liang FT, Marques AR, Mitchell PD, Purcell JE, Ratterree MS, Straubinger RK. 2001. Antibody response to IR6, a conserved immunodominant region of the VlsE lipoprotein, wanes rapidly after antibiotic treatment of Borrelia burgdorferi infection in experimental animals and in humans. J Infect Dis 184:870–878. doi: 10.1086/323392. [DOI] [PubMed] [Google Scholar]
- 12.Philipp MT, Wormser GP, Marques AR, Bittker S, Martin DS, Nowakowski J, Dally LG. 2005. A decline in C6 antibody titer occurs in successfully treated patients with culture-confirmed early localized or early disseminated Lyme borreliosis. Clin Diagn Lab Immunol 12:1069–1074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Magnarelli LA, Fikrig E, Padula SJ, Anderson JF, Flavell RA. 1996. Use of recombinant antigens of Borrelia burgdorferi in serologic tests for diagnosis of Lyme borreliosis. J Clin Microbiol 34:237–240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wilske B, Fingerle V, Herzer P, Hofmann A, Lehnert G, Peters H, Pfister HW, Preac-Mursic V, Soutschek E, Weber K. 1993. Recombinant immunoblot in the serodiagnosis of Lyme borreliosis. Comparison with indirect immunofluorescence and enzyme-linked immunosorbent assay. Med Microbiol Immunol 182:255–270. [DOI] [PubMed] [Google Scholar]
- 15.Magnarelli LA, Levy SA, Ijdo JW, Wu C, Padula SJ, Fikrig E. 2001. Reactivity of dog sera to whole-cell or recombinant antigens of Borrelia burgdorferi by ELISA and immunoblot analysis. J Med Microbiol 50:889–895. doi: 10.1099/0022-1317-50-10-889. [DOI] [PubMed] [Google Scholar]
- 16.Goettner G, Schulte-Spechtel U, Hillermann R, Liegl G, Wilske B, Fingerle V. 2005. Improvement of Lyme borreliosis serodiagnosis by a newly developed recombinant immunoglobulin G (IgG) and IgM line immunoblot assay and addition of VlsE and DbpA homologues. J Clin Microbiol 43:3602–3609. doi: 10.1128/JCM.43.8.3602-3609.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Branda JA, Aguero-Rosenfeld ME, Ferraro MJ, Johnson BJ, Wormser GP, Steere AC. 2010. 2-tiered antibody testing for early and late Lyme disease using only an immunoglobulin G blot with the addition of a VlsE band as the second-tier test. Clin Infect Dis 50:20–26. doi: 10.1086/648674. [DOI] [PubMed] [Google Scholar]
- 18.Embers ME, Hasenkampf NR, Jacobs MB, Philipp MT. 2012. Dynamic longitudinal antibody responses during Borrelia burgdorferi infection and antibiotic treatment of rhesus macaques. Clin Vaccine Immunol 19:1218–1226. doi: 10.1128/CVI.00228-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Chandra A, Wormser GP, Marques AR, Latov N, Alaedini A. 2011. Anti-Borrelia burgdorferi antibody profile in post-Lyme disease syndrome. Clin Vaccine Immunol 18:767–771. doi: 10.1128/CVI.00002-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Molins CR, Sexton C, Young JW, Ashton LV, Pappert R, Beard CB, Schriefer ME. 2014. Collection and characterization of samples for establishment of a serum repository for Lyme disease diagnostic test development and evaluation. J Clin Microbiol 52:3755–3762. doi: 10.1128/JCM.01409-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Marques AR. 2015. Laboratory diagnosis of Lyme disease: advances and challenges. Infect Dis Clin North Am 29:295–307. doi: 10.1016/j.idc.2015.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Molins CR, Ashton LV, Wormser GP, Hess AM, Delorey MJ, Mahapatra S, Schriefer ME, Belisle JT. 2015. Development of a metabolic biosignature for detection of early Lyme disease. Clin Infect Dis 60:1767–1775. doi: 10.1093/cid/civ185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Schmidt C, Plate A, Angele B, Pfister H-W, Wick M, Koedel U, Rupprecht TA. 2011. A prospective study on the role of CXCL13 in Lyme neuroborreliosis. Neurology 76:1051–1058. doi: 10.1212/WNL.0b013e318211c39a. [DOI] [PubMed] [Google Scholar]
- 24.Soloski MJ, Crowder LA, Lahey LJ, Wagner CA, Robinson WH, Aucott JN. 2014. Serum inflammatory mediators as markers of human Lyme disease activity. PLoS One 9:e93243. doi: 10.1371/journal.pone.0093243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Cheung CSF, Anderson KW, Benitez KYV, Soloski MJ, Aucott JN, Phinney KW, Turko IV. 4 November 2015. Quantification of Borrelia burgdorferi membrane proteins in human serum: a new concept for detection of bacterial infection. Anal Chem doi: 10.1021/acs.analchem.5b02803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Eshoo MW, Crowder CC, Rebman AW, Rounds MA, Matthews HE, Picuri JM, Soloski MJ, Ecker DJ, Schutzer SE, Aucott JN. 2012. Direct molecular detection and genotyping of Borrelia burgdorferi from whole blood of patients with early Lyme disease. PLoS One 7:e36825. doi: 10.1371/journal.pone.0036825. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Liang FT, Steere AC, Marques AR, Johnson BJ, Miller JN, Philipp MT. 1999. Sensitive and specific serodiagnosis of Lyme disease by enzyme-linked immunosorbent assay with a peptide based on an immunodominant conserved region of Borrelia burgdorferi vlsE. J Clin Microbiol 37:3990–3996. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Signorino G, Arnaboldi PM, Petzke MM, Dattwyler RJ. 2014. Identification of OppA2 linear epitopes as serodiagnostic markers for Lyme disease. Clin Vaccine Immunol 21:704–711. doi: 10.1128/CVI.00792-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wagner B, Freer H, Rollins A, Erb HN. 2011. A fluorescent bead-based multiplex assay for the simultaneous detection of antibodies to B. burgdorferi outer surface proteins in canine serum. Vet Immunol Immunopathol 140:190–198. doi: 10.1016/j.vetimm.2010.12.003. [DOI] [PubMed] [Google Scholar]
- 30.Lahey LJ, Panas MW, Mao R, Delanoy M, Flanagan J, Binder SR, Rebman AW, Montoya JG, Soloski MJ, Steere AC, Dattwyler RJ, Arnaboldi PM, Aucott JN, Robinson WH. 7 October 2015. Development of a multi-antigen panel for improved detection of Borrelia burgdorferi infection in early Lyme disease. J Clin Microbiol doi: 10.1128/JCM.02111-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Akin E, McHugh GL, Flavell RA, Fikrig E, Steere AC. 1999. The immunoglobulin (IgG) antibody response to OspA and OspB correlates with severe and prolonged Lyme arthritis and the IgG response to P35 correlates with mild and brief arthritis. Infect Immun 67:173–181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Alghaferi MY, Anderson JM, Park J, Auwaerter PG, Aucott JN, Norris DE, Dumler JS. 2005. Borrelia burgdorferi ospC heterogeneity among human and murine isolates from a defined region of northern Maryland and southern Pennsylvania: lack of correlation with invasive and noninvasive genotypes. J Clin Microbiol 43:1879–1884. doi: 10.1128/JCM.43.4.1879-1884.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Bunikis J, Garpmo U, Tsao J, Berglund J, Fish D, Barbour AG. 2004. Sequence typing reveals extensive strain diversity of the Lyme borreliosis agents Borrelia burgdorferi in North America and Borrelia afzelii in Europe. Microbiology 150:1741–1755. doi: 10.1099/mic.0.26944-0. [DOI] [PubMed] [Google Scholar]
- 34.Ivanova L, Christova I, Neves V, Aroso M, Meirelles L, Brisson D, Gomes-Solecki M. 2009. Comprehensive seroprofiling of sixteen B. burgdorferi OspC: implications for Lyme disease diagnostics design. Clin Immunol 132:393–400. doi: 10.1016/j.clim.2009.05.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Weiner ZP, Crew RM, Brandt KS, Ullmann AJ, Schriefer ME, Molins CR, Gilmore RD. 2015. Evaluation of selected Borrelia burgdorferi lp54 plasmid-encoded gene products expressed during mammalian infection as antigens to improve serodiagnostic testing for early Lyme disease. Clin Vaccine Immunol 22:1176–1186. doi: 10.1128/CVI.00399-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.