Previous Article | Next Article ![]()
Clinical and Diagnostic Laboratory Immunology, January 2005, p. 165-170, Vol. 12, No. 1
1071-412X/05/$08.00+0 doi:10.1128/CDLI.12.1.165-170.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Department of Bacteriology, National Center for Microbiology, Instituto de Salud Carlos III,1 Department of Pediatrics, Lactants and Immunodeficiencies Section, 12 de Octubre University Hospital, Madrid, Spain,2 Vaccines Section, Eijkman-Winkler Institute for Microbiology, Infectious Diseases and Inflammation, Utrecht University Hospital, Utrecht, The Netherlands3
Received 20 July 2004/ Returned for modification 9 September 2004/ Accepted 9 November 2004
|
|
|---|
|
|
|---|
Efficacy trial data of pneumococcal vaccination for human immunodeficiency virus (HIV)-infected patients from prospective studies are not feasible. This drawback highlights the need for further immunogenicity studies to assess the impact of vaccination in immunocompromised patients. Enzyme-linked immunosorbent assay (ELISA) has been widely used to establish correlates of protection in the evaluation of sera of vaccinees (19). However, establishing effectiveness in preventing pneumococcal disease based on serological criteria is hindered by the absence of an ideal serological correlate of protection in vitro. The 3rd International Symposium on Pneumococci and Pneumococcal Diseases (Anchorage, Alaska, 2002) recommended that opsonophagocytic assays (OPAs) should be used to supplement ELISA antibody concentration measurements (11). Clinical and mouse protection data for the selected serotypes for which this assay works indicate that the opsonophagocytosis assay provides the best basis for protection (12, 24). Measuring OPA titers has revealed low functional activity of conjugate vaccine-induced antibodies in bone marrow transplant recipients (18) and different responses to conjugate and plain polysaccharide vaccines in children and adults with sickle cell disease (25).
The present study aimed to investigate PCV7 immunogenicity in pediatric patients with perinatal HIV infection by assessing antibody response measured by ELISA, the functional capacity of the antibody response measured by OPA, and the antibody response after one and two doses of vaccine. We focused on the analysis of OPA responders within the framework of ELISA consensus threshold antibody levels based on efficacy trials in healthy individuals and on the serotype-specific differences reported here. Our study involved a cohort of 56 HIV-infected children and adolescents attending the Pediatric Clinic Unit of 12 de Octubre Hospital (Madrid, Spain).
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. Demographic characteristics, pneumococcal vaccination, and surrogate markersa
|
Specific anticapsular antibody determination. ELISA was performed as previously described (4, 21) with minor modifications to determine the concentration of specific anticapsular IgG to pneumococcal serotypes 6B, 14, and 23F. Briefly, serum reference standard assay, quality-control specimens, and test specimens were absorbed with a 1:50 dilution of 10 µg/ml of cell wall polysaccharide (Statens Seruminstitut, Copenhagen, Denmark). ELISA microtiter plates (Nunc Maxisorp, PGC Scientifics Corp., Gaithersburg, Md.) were coated with 0.5 µg of purified polysaccharide antigen (American Type Culture Collection, Manassas, Va.). Every plate included paired pre- and postvaccination samples, our own laboratory reference standard serum with calculated concentrations of IgG to a specific polysaccharide (based on cross-standardization with the 89-SF serum reference-assay standard, kindly provided by Carl E. Frasch, Centers for Biologics, Rockville, Md.), four quality controls, and a negative serum specimen that contained no IgG to the individual polysaccharide studied.
Pneumococcal strains. The pneumococcal serotypes selected for the OPA included the least immunogenic serotypes (6B and 23F) and serotype 14 (20). These three serotypes are thought to be the most prevalent in invasive pneumococcal disease among HIV-infected children (13). Pneumococcal strains were obtained from our collection (Reference Laboratory of Pneumococci, Centro Nacional de Microbiología, Madrid, Spain) and serotyped as previously described (8).
Opsonophagocytosis activity titers. Flow cytometric OPA for each of the three serotypes was performed largely as previously described but with minor modifications (10). Three strains of serotypes 6B, 14, and 23F were grown thrice consecutively to log phase in Todd-Hewitt medium supplemented with 1% of decomplemented human pooled serum to ensure a high level of encapsulation. This step was critical for measuring specific anticapsular antibodies (17). Subsequently, strains were heat inactivated and fluorescein isothiocyanate labeled. Human pooled serum was depleted of IgG by protein G affinity chromatography and used as a complement source. In optimizing experiments, the addition of complement did not significantly change OPA titers relative to those of corresponding non-heat-inactivated strains. This ensured a high degree of encapsulation of the strains used in the assay.
Human polymorphonuclear leukocytes (PMNs) were isolated from the blood of healthy, Fc
receptor-typed volunteers (9) by Ficoll-Histopaque gradient centrifugation. Sera from vaccinees were heated (30 min at 56°C) to inactivate the complement before it was assayed. Samples of 2.5 x 106 pneumococci and three serial fivefold dilutions starting with a 1:10-diluted serum sample were added to each of the 96 wells of a microtiter plate (ratio of PMN to bacteria = 1:20). After 30 min of opsonization at 37°C on a microtiter plate shaker, 1.25 x 105 freshly isolated PMNs were added per well and phagocytosis was carried out for 30 min at 37°C with vigorous shaking. After three washing steps with ice-cold bovine serum albumin-Hanks' balanced salt solution, cells were transferred to fluorescence-activated cell sorter tubes, immediately fixed with paraformaldehyde (2%) in phosphate-buffered saline, and analyzed in a flow cytometer (FACSCalibur; Becton Dickinson). The percentage of fluorescein isothiocyanate-positive PMNs was used as a measure of the phagocytic activity of the serum. Results are expressed as the log10 of the phagocytosis titer, which is the reciprocal of the serum dilution yielding 25% fluorescein isothiocyanate-positive PMNs. A positive control serum was included on each plate to control the reproducibility of the assay.
Statistical analysis.
All patients were included in the analysis, and data were analyzed with the SPSS 8.0 program. Descriptive statistics (mean, standard deviation, and median) were calculated for all samples. Student's t test for paired samples was used to compare means of variables (prevaccination with postdose 2 and postdose 1 with postdose 2). ELISA and OPA data were correlated using the Pearson's correlation coefficient (r). Chi-square (
2) contingency tests were used to analyze differences in frequencies of responders and nonresponders for pairs of response criteria. All statistics were regarded as significant for probability values of
0.05.
|
|
|---|
OPA titers.
Mean OPA titers are summarized in Table 2. Significant differences were found between prevaccination and postdose 2 OPA titers (6B, t = 4.05, P
0.000, n = 41; 14, t = 2.669, P = 0.011, n = 42; 23F, t = 2.034, P = 0.048, n = 42). However, no significant differences between the first and second doses were found in any serotype (6B, t = 1.825, P > 0.05, n = 37; 14, t = 1.957, P > 0.05, n = 37; 23F, t = 1.229, P > 0.05, n = 38).
|
View this table: [in a new window] |
TABLE 2. Mean antibody responses measured by OPA and ELISA against serotypes 6B, 14 and 23F
|
Correlation between OPA and ELISA.
Correlation differed significantly between serotypes and times studied, being higher and significant for each serotype only at 3 months after the administration of dose 2. Considering serotypes 6B, 14, and 23F, postdose 2 r values were 0.754 (P
0.000), 0.414 (P = 0.006) and 0.593 (P
0.000), postdose 1 r values were 0.673 (P
0.000), 0.072 (P > 0.05) and 0.420 (P = 0.002) and, prevaccination r values were 0.216 (P > 0.05), 0.246 (P > 0.05), and 0.544 (P
0.000), respectively. The distribution of antibody concentrations and OPA titers after one and two doses of PCV7 and prior to vaccination are shown in Fig. 1.
![]() View larger version (13K): [in a new window] |
FIG. 1. Distribution of antibody concentrations (ELISA) and opsonophagocytic assay (OPA) titers prior to vaccination and after one and two doses of 7-valent protein-conjugated vaccine for serotype 6B (A), serotype 14 (B), and serotype 23F (C).
|
4 for serotype 6B, <25 to
25 for serotype 14, and <6 to
6 for serotype 23F. We used both criteria to maximize the possibility of detecting any potential benefit associated with vaccination. Responders versus nonresponders. Scatter plots representing the increases in OPA and ELISA after dose 2 for each serotype are shown in Fig. 2. Following the first response criterion, the quadrant analysis revealed that responders were observed in 43.9% (18 of 41, OPA-ELISA mean and standard deviation of 21.3 ± 2.5 and 17.4 ± 3.35, respectively), 28.5% (12 of 42, OPA/ELISA mean of 134.3 ± 5.37 and 159.05 ± 3.61, respectively), and 38.0% (16 of 42, OPA/ELISA mean of 16.12 ± 2.48 and 13.29 ± 3.03, respectively) of cases for serotypes 6B, 14, and 23F, respectively. Nonresponders occurred at frequencies of 31.7% (13 of 41), 33.3% (14 of 42), and 40.4% (17 of 42) for serotypes 6B, 14, and 23F, respectively. OPA-only responders had frequencies of 12.1% (5 of 41), 7.1% (3 of 42), and 7.1% (3 of 42) for serotypes 6B, 14, and 23F, respectively, while those of ELISA-only responders were 12.1% (5 of 41), 30.9% (13 of 42), and 14.2% (6 of 42), respectively.
![]() View larger version (12K): [in a new window] |
FIG. 2. Distribution of serotype-specific vaccine responders as a function of OPA increase and ELISA increase after dose 2. (A) Serotype 6B; (B) serotype 14; (C) serotype 23F. Quadrant lines were drawn at the twofold increase thresholds for the ELISA and opsonophagocytic assays.
|
2 = 17.7 and P
0.000,
2 = 4.2 and P = 0.04, and
2 = 10.1 and P = 0.001 for serotypes 6B, 14, and 23F, respectively). Correlation of vaccination response with demographic features and clinical variables. No correlation was found between response to vaccination and demographic features or clinical characteristics such as immune and clinical category before highly active antiretroviral therapy and CD4 cell counts.
|
|
|---|
Antibody concentrations in prevaccination serum samples were higher in our pediatric AIDS patients than those reported for healthy U.S. children (2, 22) and corresponding OPA titers were significantly low, except for serotype 14. Poor correlations between OPA titers and IgG concentrations prior to vaccination suggested that a higher level of nonspecific antibodies in the pediatric AIDS patients may account for this difference. An effect of prevaccination with the 23-valent polysaccharide vaccine might lead to higher concentrations, but we have ruled out this explanation in this case since the previous immunization occurred between 11 and 84 months earlier and specific antibodies are known to be lost rapidly after pneumococcal vaccination in HIV-infected patients (16). Total IgG concentrations in AIDS patients higher than in immunocompetent adults were previously reported (7).
Recent developments in the field have included an absorption step with an irrelevant polysaccharide (22F) for the evaluation of antibody concentrations that may influence correlations (5). Feikin et al. reported that IgG ELISAs for serotypes 6B and 23F were more specific for serotype-specific pneumococcal capsular antibodies with 22F polysaccharide absorption by reducing nonspecific antibody binding (the exception being serotype 14). However, correlations between ELISA IgG concentrations and opsonophagocytic titers were not significantly improved and continued to be poorly correlated in the case of prevaccination sera, both with and without 22F absorption (7). In previous work, nonfunctional antibodies failed to be completely removed since a similar reduction in prevaccination and postvaccination antibody levels was observed in control paired samples with a significant OPA rate (data not shown). Moreover, assignments of concentrations from the standard reference serum (89S-F) are calculated without 22F absorption. Taking these matters into account, we decided not to include 22F absorption in these ELISA determinations.
In general, correlations between antibody concentration and opsonophagocytic titer (including all values) in these 56 AIDS patients were poorer than those reported in another study of 12 healthy adults (5) but greater than in a study of 60 renal transplant recipients (15) and another of 33 HIV-infected adults (7). In this regard, the 6B serotype correlation coefficients were more similar to those previously reported in healthy individuals. This suggests that pediatric AIDS patients were better able to produce functional antibodies than HIV-infected adults. The reason for this difference is unclear but might be related to a better response after immunologic restoration with highly active antiretroviral therapy in children than in adults.
Sera from AIDS patients may have higher nonfunctional antibody levels than that from members of the healthy population. This was illustrated when the factors of antibody concentration versus OPA titer were examined by a quadrant analysis of responders (Fig. 2). Ideally, there should be no data points in the upper left and lower right quadrants. As expected, a small number of data points fell in the upper-left quadrant (delimited by values of 2 on the ordinate and abscissa) and a greater percentage, mostly of serotype 14, were distributed in the lower-right quadrant. Our data indicate that although serotype 14 continued to have the highest absolute OPA titers and IgG concentrations, it also had a smallest percentage of responders by the first criterion and, in general, the poorest correlation between OPA titers and IgG concentrations, as was also reported in the study of renal transplant recipients. It is worth noting that the correlation for all serotypes was much greater after the second dose than it was after the first dose, indicating an improvement in the functionality of the antibodies despite no change in the titers for either OPA or IgG. Therefore, booster doses may have a greater effect on the quality of the immune response than on the quantity of the bulk of antipolysaccharide antibodies. Moreover, this observation suggests a relationship between these correlation coefficients and the ability to detect immune responses, since repeating the PCV7 dose improved correlations between OPA and ELISA and made the ELISA values more relevant in terms of protection.
It has been reported that AIDS patients with a CD4 cell count of <500 cells/mm3 have an impaired IgG response to pneumococcal antigens (14, 23). In our study, only one patient had a CD4 cell count of <500, precluding any meaningful analysis of links between the distribution of antibody response and this clinical parameter. No correlation was found between response to vaccination and immune category or clinical category before highly active antiretroviral therapy.
We defined a responder as a patient exhibiting at least a twofold increase in OPA titer and antibody concentration (criterion a). However, the main limitation of these approaches is that clinically relevant cutoff levels (protective efficacy) must be known because the protected or unprotected status of a vaccinee of interest must be determined by criteria independent of the analysis variables. In the World Health Organization workshop on correlates of protection (Anchorage, Alaska, 2002), current scientific opinion based on existing efficacy trial data in children and adults considered the induction of 0.2 µg of postprimary IgG antibody per ml to be a protective threshold against invasive pneumococcal disease, with a single threshold for all serotypes included in the vaccine. As a threshold antibody level should be a more conservative estimate of protection, the geometric mean antibody concentration in PCV7 healthy vaccinees (0.5 µg/ml could be a reasonable threshold value) (2). Subsequently, since opsonophagocytic activity represents qualitative and functional antibody responses, we have considered an additional criterion of response: an OPA titer equal to or greater than the value obtained from the regression curve for each serotype for an IgG concentration of 0.5 µg/ml. Thus, our study reports that a concentrations of 0.5 µg/ml of IgG represented threshold OPA titers of 4, 25, and 6, yielding 46.3, 50.0, and 50.0% OPA responders for serotypes 6B, 14, and 23F, respectively.
We conclude that although significant increases in antibody response have been reported, overall response rates in pediatric HIV infection were generally suboptimal because no booster response was observed after the second dose of PCV7. The present study underlines the need to define how much antibody is needed for protection from different serotypes since analysis of frequency of responders, IgG concentrations, and OPA titers revealed significant differences between serotypes. It also highlights the importance of demonstrating vaccine immunogenicity with suitable endpoints in immunocompromised patients and the difficulties in interpreting serological data.
Financial support was provided by Instituto de Salud Carlos III (grant MPY-1115/03) and Sociedad Española de Microbiología Clínica y Enfermedades Infecciosas (SEIMC). This research was carried out on behalf of the Spanish Network Pneumococcus Study Group (G03/103).
|
|
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2010 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»