Characterization of Serological Responses to Pertussis

Serum samples.Serum samples were obtained from culture-positive children (79 serum samples from 51 individuals) recruited at Cincinnati Children's Hospital Medical Center from 1999 to 2002 and were characterized in a previous study (25). The characteristics of the culture-positive population are described in Table 1. Additional samples were obtained from some subjects at yearly intervals. Eighty-eight percent of the children had received the appropriate immunization series. One individual had never received pertussis vaccine. Control sera (48 serum samples from 34 donors) were collected at the University of Cincinnati from healthy adult volunteers who did not recall having a long-lasting cough in recent memory. Informed consent and assent were obtained from all subjects and their parents or guardians. This study was approved by the institutional review boards of the Cincinnati Children's Hospital Medical Center and the University of Cincinnati and was conducted according to the experimental guidelines of the U.S. Department of Health and Human Services.

Secondary antibodies.Alkaline phosphatase-conjugated goat antibody against human immunoglobulin G (IgG), IgA, and IgM were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA).

Antigens.We purified PT and FHA from the culture supernatant of B. pertussis BP338 (12, 24, 30). The LOS of BP338 and LPS of B. parapertussis CN8234 were purified by phenol extraction with proteinase K treatment (10, 27). Purified BrkA protein was a gift from Rachel Fernandez (26). PAL was purified from a glutathione S-transferase fusion protein expressed from pGEX-PAL, provided by Nicholas Carbonetti (5). The purities of all antigens were determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. No contaminants were detected in any of the purified preparations.

Production of CatACT and C-FHA.A DNA fragment encoding the first 400 amino acids of the cyaA gene was PCR amplified from BP338 by using the oligonucleotides 5′-GCTAGCATGCAGCAATCGCATCAGGCT-3′ and 5′-AAGCTTTCACACTGCGCCCAGCGAC-3′. The open reading frame of FHA from codon 2373 to the C terminus was amplified by PCR with oligonucleotides 5′-ATATGCTAGCCAGAGCAAATTCTACGGCTCGCG-3′ and 5′-CCCAAGCTTTTTGTTGGTTTCATAGAAGACCCGGTAG-3′ with Pfu-turbo polymerase (Stratagene, La Jolla, CA) and 1 M GC-Melt (Clontech, Palo Alto, CA). The DNA sequences of both fragments were verified.

The fragments were cloned into the NheI and HindIII sites of pET21b (Novagen, Madison, WI), resulting in plasmid pSM021 for CatACT and plasmid pMW10 for C-FHA. Protein was expressed in Rosetta2(DE3)pLysS (Novagen), induced with isopropyl-β-d-thiogalactopyranside (2 mM), and purified from inclusion bodies.

For CatACT, lysis buffer (50 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA) supplemented with lysozyme and 0.1 mM phenylmethylsulfonyl fluoride (PMSF) was added at 4 ml/g bacteria. The bacteria were incubated for 20 min at 25°C and lysed by the addition of deoxycholic acid (4 mg/g bacteria) and sonication. The bacterial pellet was washed and dissolved in denaturing buffer (8 M urea, 50 mM Tris-HCl, pH 8.0, 0.2 mM CaCl₂, 0.1 mM PMSF), which was added at 8 ml/g bacteria. The urea extract was passed over a DEAE column (0.5 ml/min, 4°C), and the flowthrough was dialyzed into 25 mM sodium carbonate buffer at pH 8.0.

C-FHA was solubilized from the inclusion bodies (29) and loaded onto a HiTrap-Ni column fast-performance liquid chromatography system (Amersham Bioscience, Piscataway, NJ). The column was washed with 20 mM sodium phosphate buffer (pH 7.2) with 0.5 M NaCl and eluted with a linear gradient of imidazole (0 to 0.5 M, 20 ml).

ELISAs.ELISAs for quantitation of FHA and PT antibodies (IgG and IgA) were performed by the method of Manclark et al. (23) with Food and Drug Administration pertussis antiserum lot 3 and lot 5 as references.

For ELISAs for quantitation of protein antigens, wells of 96-well microtiter plates were coated with 100 μl of CatACT or C-FHA at a concentration of 1 μg/ml in 50 mM carbonate buffer, pH 9.6 (coating buffer), for 16 h at room temperature. Phosphate-buffered saline (PBS; pH 7.2) was used as the buffer for antigen coating for the PAL ELISA. After three washes with PBS containing 0.05% Tween 20, twofold serial dilutions of sera in incubation buffer (PBS with 10% nonfat dry milk and 0.05% Tween 20) were added at 100 μl/well. The plates were incubated at room temperature for 2 h. Secondary antibody (1:5,000) in 100 μl of incubation buffer was added, and the plate was incubated for 2 h. The wells were washed, and 200 μl of substrate (FAST-pNPP kit; Sigma, St. Louis, MO) was added. After 30 min at room temperature, 25 μl of 12% NaOH was added to the wells and the optical density at 405 nm (OD₄₀₅) was measured. Antibody values were expressed as the reciprocal serum dilution (titer) that gave an absorbance of the mean plus 2 standard deviations (SDs) of the values for the negative control wells for anti-CatACT antibodies or the reciprocal serum dilution that gave 0.5 as absorbance of the mean for anti-C-FHA and PAL antibodies.

The BrkA ELISA was performed by the procedure developed by the laboratory of R. Fernandez. In brief, the wells were coated with 100 μl of BrkA (20 μg/ml) in coating buffer with 5 mM MgCl₂ at 37°C for 1 h. After four washes with Dulbecco's modified PBS (DPBS), the wells were blocked with 150 μl of DPBS with 1% bovine serum albumin (BSA) at 37°C for 1 h. One hundred microliters of diluted serum (1:100) in DPBS with 1% BSA was added to the appropriate wells. After incubation at 37°C for 1 h, the wells were washed and incubated with 100 μl of secondary antibody (1:5,000) in DPBS with 1% BSA, incubated at 37°C for 1 h, and developed as described above. The antibody content was expressed as the OD₄₀₅.

For the LOS and LPS ELISAs, the wells of 96-well microtiter plates were coated with 100 μl of LOS or LPS (20 or 10 μg/ml) in coating buffer at 37°C for 2 h. The wells were washed, 300 μl of incubation buffer was added for 30 min at 37°C, and the wells were washed again. Serum diluted 25-fold in incubation buffer was added at 100 μl/well. The subsequent steps were performed as described for the CatACT ELISA, except that 1 h of incubation for sera and secondary antibodies was used. The antibody content was expressed as the OD₄₀₅.

Statistical analysis.Statistical differences between the two study groups were examined by Student's t test with SYSTAT11 (SYSTAT Software, Point Richmond, CA). Receiver operating characteristic (ROC) analysis was performed with Win Episcope2 (http://www.clive.ed.ac.uk/winepiscope/ ).

Antibody responses to vaccine antigens.PT and FHA are vaccine antigens and are also used in conventional ELISAs for the serodiagnosis of pertussis (8, 9). Consistent with previous results, the IgG and IgA antibody responses to PT were significantly higher in the culture-positive group than in the control group at both the time of diagnosis and 1 year later (Fig. 1). Mean PT IgM levels did not differ between the culture-positive group and the control group (data not shown).

• Open in new tab
• Download powerpoint

FIG. 1.

Distribution of responses to vaccine antigens in the culture-positive individual group and the control group. (A) PT IgG; (B) PT IgA; (C) FHA IgG; (D) FHA IgA. Visit 1, antibody values for serum obtained at the time of diagnosis; visit 2, antibody values for serum obtained approximately 1 year after diagnosis; visit 3, antibody values for serum obtained approximately 2 years after diagnosis. Box-whisker plots represent boxes for medians with 25th and 75th percentiles and whiskers for 10th and 90th percentiles. Closed circles represent outliers. The numbers in each box chart represent the probability values obtained by Student's t test.

The IgG and IgA antibody responses to FHA were also significantly higher in the culture-positive group than in the control group at both the time of diagnosis and 1 year later (Fig. 1). High values of IgG antibodies to FHA were seen in the control group. This was thought to be due to proteins produced by other bacterial species that cross-react with FHA (2).

Different cutoff values have been proposed for serodiagnosis by the use of PT IgG and FHA IgG. Using the mean plus 2 SDs of the values for the control group, we calculated a value of 48 ELISA units (EU)/ml to be the cutoff value for seropositive responses to PT, giving a specificity of approximately 95% (Table 2). This value is more than double the traditional definition of 20 EU/ml (4). Under the traditional cutoff definition, approximately 13% of the control group would be considered seropositive, reducing the specificity of the test to 87%. In contrast, Baughman et al. (4) proposed a cutoff value of 94, and this value would reduce the sensitivity of the PT IgG assay to 73%. The sensitivity of the FHA IgG test was only 35.3% by the use of our calculated cutoff value of 163 EU/ml (Table 2). This value also falls between the conventional value, 20 EU/ml, and a much higher value, 229 EU/ml, suggested by Baughman et al. (4). The conventional cutoff value reduced the specificity of FHA IgG to 39.6%, while the higher value reduced the sensitivity to 15.7%.

Antibody responses to nonvaccine antigens.The responses to five nonvaccine antigens were also characterized (Fig. 2). The mean IgG responses to CatACT (Fig. 2A) and C-FHA (Fig. 2C) of the culture-positive individuals were higher than those of the control group at the time of diagnosis and 1 year later. However, the IgA responses to CatACT (Fig. 2B) and C-FHA (Fig. 2D) were not different between the two groups. The sensitivity and specificity of CatACT IgG as a diagnostic tool were 62.8% and 91.7%, respectively (Table 2), while the sensitivity and specificity of C-FHA IgG as a diagnostic tool were 39.2% and 95.4%, respectively (Table 2). Mean CatACT IgM levels did not differ between the culture-positive group and the control group (data not shown).

• Open in new tab
• Download powerpoint

FIG. 2.

Distribution of responses to nonvaccine antigens in the culture-positive individual group and the control group. (A) CatACT IgG; (B) CatACT IgA; (C) C-FHA IgG; (D) C-FHA IgA; (E) LOS IgG; (F) LOS IgA; (G) PAL IgG; (H) PAL IgA. Visit 1, antibody values for serum obtained at the time of diagnosis; visit 2, antibody values for serum obtained approximately 1 year after diagnosis; visit 3, antibody values for serum obtained approximately 2 years after diagnosis. Box-whisker plots represent boxes for medians with 25th and 75th percentiles and whiskers for 10th and 90th percentiles. Closed circles represent outliers. The numbers in each box chart represent the probability values obtained by Student's t test.

The IgG response to LOS was not different between the culture-positive group and the control group (Fig. 2E), and interestingly, the LOS IgG titers did not decrease with time. In contrast, the LOS IgA titers did differ between the two groups. The sensitivity of LOS IgA as a diagnostic tool was relatively low, 29.4% (Table 2).

Two proteins, BrkA and PAL, did not appear to be very immunogenic. The titers of IgG antibody against BrkA were measured in only six culture-positive individuals and six control individuals (data not shown). Most of the individuals had no detectable response to BrkA, and there was no mean difference between the culture-positive individuals and the control individuals (P = 0.226). Similarly, most of the subjects did not have antibodies to PAL, and the responses for the culture-positive individuals did not differ from the responses of the control group (Fig. 2G and H).

Comparison of test results.All possible combinations of sensitivity and specificity that can be achieved by changing the test's cutoff value can be summarized by using the area under the curve (AUC) of a ROC curve (20). An assay with an AUC of 50% represents an assay that has no discriminating ability, while an AUC of 100% means that the assay can provide perfect discrimination. The test for PT IgG had the largest AUC (Fig. 3) and was the best diagnostic test identified in this study. CatACT IgG, FHA IgA, C-FHA IgG, FHA IgG, LOS IgA, and PT IgA were less sensitive but were still useful for the diagnosis of pertussis (Fig. 3A). In contrast, for CatACT IgA, C-FHA IgA, LOS IgG, PAL IgG, and PAL IgA (Fig. 3B), the AUCs were about 50% and their 95% lower confidence limits were less than 50%, suggesting that tests for these antibodies have little ability to discriminate infected individuals from healthy controls as stand-alone diagnostic tests.

• Open in new tab
• Download powerpoint

FIG. 3.

ROC analysis of tests. (A) ROC curves of seven assays with predictive value (PT IgG, CatACT IgG, FHA IgA, C-FHA IgG, FHA IgG, LOS IgA, and PT IgA); (B) ROC curves of assays that lack predictive value (CatACT IgA, PAL IgG, LOS IgG, PAL IgA, and C-FHA IgA).

The definition of a positive diagnostic test result as an ELISA value greater than the mean plus 2 SDs of the values for the control group resulted in greater than 90% specificity for all assays (Table 2). The PT IgG-based assay had the highest sensitivity (92.2%), followed by the CatACT IgG-based assay, with 62.8% sensitivity. The CatACT IgA, C-FHA IgA, LOS IgG, and PAL IgA and IgG antibodies had very low sensitivities (nearly 10% or less).

Diagnosis based on multiple tests.Combinations of different ELISAs are often used to improve sensitivity. When the results of all 14 tests are considered, the mean number of positive results for the culture-positive individuals was 3.7 (median and mode = 4), with a range from 0 to 9. Only one person was negative by all 14 tests. The combination of PT IgG and FHA IgG detection is widely used; however, all of the FHA IgG-positive patients were also positive for PT IgG, and this combination did not improve the test performance (Table 3). A combination of three assays with nonvaccine antigens (CatACT IgG, C-FHA IgG, and LOS IgA) detected 41 of 51 (80%) culture-positive individuals, and the specificity was 83%.

When the results of all 14 tests are considered for the control group, the mean number of positive results was 0.7 (median and mode = 0), with a range from 0 to 4. Ten serum samples in the control group were positive by two or more tests.

Duration of antibody responses.We also examined the status of individuals a year after diagnosis. By the PT IgG assay, 92.0% of the culture-positive individuals (23 of 25) were seropositive at the time of diagnosis and 64.0% were positive (16 of 25) after 1 year. For the CatACT IgG assay, 68.0% (17 of 25) of the culture-positive individuals were seropositive at the time of diagnosis and 44.0% (11 of 25) were positive after 1 year. Similarly, 48.0% (12 of 15) of the culture-positive individuals were seropositive by the FHA IgA assay at the time of diagnosis and 28.0% were seropositive (7 of 25) after 1 year.

Immune response to B. parapertussis LPS.The antibody responses to the LPS of B. parapertussis were also measured. Both groups had high levels of LPS IgG, and the distribution of the two groups overlapped (Fig. 4). In the test for IgA LPS, eight individuals in the culture-positive groups and one control individual were positive for the LPS of B. parapertussis test under the cutoff definition of the mean plus 2 SDs of the values for the controls. No correlation was detected between IgA against LOS of B. pertussis and IgA against LPS of B. parapertussis.

• Open in new tab
• Download powerpoint

FIG. 4.

Distribution of responses to LPS of B. parapertussis in the culture-positive individual group and the control group. (A) LPS IgG; (B) LPS IgA. Visit 1, antibody values for serum obtained at the time of diagnosis; visit 2, antibody values for serum obtained approximately 1 year after diagnosis; visit 3, antibody values for serum obtained approximately 2 years after diagnosis. Box-whisker plots represent boxes for medians with 25th and 75th percentiles and whiskers for 10th and 90th percentiles. Closed circles represent outliers. The numbers in each box chart represent the probability values obtained by Student's t test.