|Home | About | Journals | Submit | Contact Us | Français|
An in vitro assay designed to measure the functional activity of vaccine-induced antibody is a necessary component of any vaccine development program. Because traditional efficacy studies of vaccines to prevent neonatal diseases caused by group B Streptococcus (GBS) are unlikely given the effectiveness of current antibiotics and screen-based surveillance practices, the ability to efficiently and effectively measure functional antibody responses may be of particular importance. GBS, like other encapsulated bacterial pathogens, are susceptible to opsonization by specific antibody and complement and subsequent killing by the host's effector cells. The in vitro opsonophagocytosis and killing assay (OPA) mimics this in vivo defense strategy and has been used for decades to measure the functionality of natural and/or vaccine-induced GBS-specific antibody. Here we describe a fluorescence-based OPA (flOPA) that measures the ability of specific antibody to opsonize fixed, fluorescently labeled GBS or antigen-coated fluorescent microspheres for uptake by differentiated HL-60 cells in the presence of complement. Compared to the classical OPA, the flOPA is standardized with respect to effector cells, complement and antigenic targets. The GBS flOPA is also less time-intensive and has the potential to measure antibody to multiple antigens simultaneously. Quantitative functional antibody determinations using the flOPA may serve as a surrogate measure of GBS vaccine effectiveness in lieu of traditional phase 3 efficacy trials.
Several phase 1 and phase 2 clinical trials of candidate group B Streptococcus (GBS) capsular polysaccharide (CPS)-protein conjugate vaccines have either been completed1 or are currently recruiting subjects.2 Successful outcomes of the current, industry-sponsored clinical trails would offer the hope that an effective vaccine against GBS, an opportunistic human pathogen,3 will be licensed in the near future. However, FDA approval of a maternal vaccine to prevent neonatal GBS disease may rely on a surrogate of protection,4 as a traditional phase 3 efficacy trial is not likely to be conducted because of the effectiveness, not only of the antibiotics themselves, but also of the strategies for antepartum administration.5,6 One viable surrogate of the effectiveness of a GBS vaccine is the in vitro opsonophagocytosis and killing assay (OPA) which measures the ability of antibody to opsonize GBS for killing by human effector cells in the presence of complement.7 This is the mechanism by which antibody is thought to provide in vivo protection against infection by this encapsulated pathogen.
The variability in past results of the GBS OPA can be attributed to several factors that could preclude its use as a surrogate of protection. These factors include (1) the source and quality of effector cells, (2) the source and quality of complement, (3) standardized preparation of target GBS and GBS antigens, and (4) the labor- and time-intensiveness of the assay. Recently we showed that differentiated HL-60 (dHL-60) cells can substitute for human peripheral blood leukocytes and that commercially available baby rabbit complement can be used in an OPA to measure the functional activity of vaccine-induced GBS rabbit and human antisera.8 Killing of GBS by human peripheral blood leukocytes in the presence of complement, as performed in the classical GBS OPA, was directly correlated with killing by dHL-60 cells.8 These findings led to the development of a “fluorescent OPA” (flOPA) that uses Alexa Fluor 488-labeled whole, killed GBS cells or antigen-coated fluorescent beads as targets to measure the opsonic activity of antibody by flow cytometry. These reagents allow for standardization of the antigenic target and measurement of functional antibody to several GBS antigens in a single assay. Internalization of either fixed, labeled GBS or antigen-coated fluorescent beads by dHL-60 cells replaces reduction in GBS CFU as the measured output in an assay that can be scaled up and performed in a fraction of the time and with smaller amounts of valuable reagents compared with the traditional OPA.
This study was undertaken to develop a fluorescence-based assay to measure the functional activity of vaccine-induced GBS antibody. Such an assay should be controlled and standardized with respect to the reagents used, and it should also be able to measure the functional activity of antibody directed to different GBS antigens.
It is extremely important to maintain the integrity of target GBS antigens used in the flOPA. Target antigens used in this study included ethanol-fixed whole GBS type III strain M781 and recombinant GBS alpha-like protein (rAlp3) coupled to fluorescent beads. A type III CPS-specific inhibition ELISA confirmed that ethanol fixation of GBS did not alter this carbohydrate structure. There was no substantial difference in the ability of ethanol-fixed or live GBS to inhibit binding of type III CPS-specific antibody to microtiter plates coated with type III CPS (data not shown). rAlp3 coupled to fluorescent microspheres inhibited binding of specific antibody to rAlp3-coated microtiter plates in a dose-dependent manner (data not shown). These data demonstrate that neither the methods used to prepare labeled, fixed GBS whole cells nor the conjugation procedure used to couple GBS proteins to fluorescent microspheres altered the integrity of the target antigens.
Internalization of target GBS or antigen-coated beads by dHL-60 cells was revealed by microscopy. Uptake of whole, ethanol-fixed GBS type III cells by dHL-60 cells in the presence of type III CPS-specific antibody and complement was visualized by differential interference and confocal microscopy (Fig. 1). Additionally, internalization by dHL-60 cells of rAlp3-coated fluorescent beads in the presence of specific antibody and complement was shown with use of confocal microscopy and further visualized by three-dimensional rendering of the confocal images (Fig. 2). These analyses strongly suggest that both whole GBS and protein-coated fluorescent microspheres are indeed internalized by dHL-60 cells and not simply associated with the dHL-60 cell surface.
A high signal-to-background ratio is desirable in any quantitative assay; however, in the absence of antibody, nonspecific opsonization by complement can result in high background. Lowering the amount of added complement from 10% to 5% reduced the complement-mediated internalization of labeled GBS, with additional reduction when the amount was further reduced to 2% (Fig. 3). Opsonization of GBS type III in the presence of rAlp3 antibody with 1% added complement gave the lowest level (<5%) of non-specific uptake while retaining a substantial uptake percentage. Based on these results, all subsequent assays were performed with 1% baby rabbit complement.
Titration of standard rabbit reference serum to GBS type III (SRRS III) in the presence of 1% complement showed that >50% uptake of fluorescently labeled GBS was achieved with low background at a serum dilutions of 1:1,000 and 1:2,000. Uptake was less than 14% in the absence of antibody, complement, both antibody and complement, or with non-immune rabbit serum (Fig. 4). These data show that 1% rabbit complement is (1) sufficient for optimal GBS uptake in this assay, and (2) provides an optimal signal-to-background ratio.
Rabbit antibody to rAlp3 cross-reacts with the Rib protein found on GBS type III strain M781.9 This antibody opsonized rAlp3-conjugated fluorescent beads and also labeled GBS type III cells presumably via the cross-reactive Rib protein (Fig. 5). rAlp3-conjugated fluorescent beads were not opsonized and thus were not internalized by dHL-60 cells in the presence of rabbit antibody to recombinant beta C protein, or pre-vaccination rabbit serum. These data show functional activity of antibody on fixed and labeled GBS of both protein- and CPS-specific epitopes and demonstrates that the functional activity of antibody to more than one antigen can be measured using the flOPA.
The functional activity of sera from six rabbits with known amounts of type III CPS-specific IgG given two doses of GBS type III-TT vaccine was determined using the flOPA and fluorescently labeled GBS type III strain M781. The ability of the sera to opsonize GBS did not always correlate with the amount of type III CPS-specific IgG in the assay (Fig. 6). Over 50% uptake of labeled GBS was measured with 0.11 μg IgG/ml from rabbit 4209, whereas only 0.0007 μg IgG/ml from rabbit 4185 was required to achieve the same level of uptake. Although serum from rabbit 4169 that responded poorly to vaccination (0.38 μg/ml of specific IgG) resulted in low uptake (<50%), and serum from the best responder (rabbit 4172; 27.47 μg/ml of specific IgG) tested opsonized GBS for uptake at a concentration of 0.1 μg/ml, a correlation between specific IgG and uptake could not be derived due to the low number of samples tested. However, these data exemplify the importance of measuring the activity of antibody, as the specific amount of antibody alone does not necessarily reflect functional immunity.
The flOPA offers several improvements over the traditional OPA: (1) assay components such as complement and effector cells are standardized; (2) functional antibody responses to a specific GBS antigen (surface-expressed proteins or CPS) can be measured by coupling target antigens to fluorescent beads; (3) functional antibody to any of the nine GBS serotypes can be prepared to measure whole-cell responses to naturally occurring GBS infections; and (4) the time required to perform the assay is reduced by as much as 75%. Six different batches of dHL-60 cells were used to generate the data presented in Figures 4 and and6.6. The percent uptake of labeled GBS M781 cells by dHL-60 cells in the presence of 1% baby rabbit complement (no antibody) averaged 9.5 (range 6.2 to 11.5) with a standard deviation of 1.9 (95% confidence limits: 7.5 and 11.5). The observed minimal variation from day-to-day experiments with different batches of dHL-60 cells demonstrates the robustness of using this cell line in the flOPA.
Perhaps most importantly, the flOPA can be multiplexed to measure opsonic antibody to several antigens simultaneously. Indeed, purified type III CPS covalently coupled to amine-modified R590 microspheres (a generous gift of Michael Hickey, Flow Applications, Inc., Okawville, IL) used in lieu of labeled GBS M781 cells were effective targets in the flOPA (data not shown), suggesting that reagents to measure the antibody response to any GBS CPS can be prepared and used in a multiplex assay as demonstrated for pneumococcal CPSs.10 The major drawback to the flOPA include the need for a costly flow cytometer and skilled personnel to operate and maintain it.
The effectiveness of vaccines is commonly determined by large phase 3 efficacy trials in settings that allow direct comparison to controls. Developing a maternal GBS vaccine to prevent neonatal GBS disease cannot ethically include a traditional efficacy trial, as antibiotics are effective to reduce incidence of at-risk deliveries and also to treat those who have contracted GBS disease. Fortunately, the in vitro OPA is a viable correlate of GBS vaccine effectiveness that can circumvent the need for a traditional efficacy trial for licensure of a maternal GBS vaccine. If a standardized, high-throughput, well-controlled OPA based on the one described in this report correlates positively with a quantitative ELISA, then direct antibody measurement may be an appropriate correlate of vaccine efficacy. Until such a relationship is confirmed, one can determine the association between the functional activity of either vaccine-induced or naturally occurring GBS-specific antibody and GBS disease. The data presented herein provide the framework necessary to establish a high-throughput, multiplex OPA to effectively measure functionally active antibody against one or more GBS vaccine antigens. Such an assay could be used as a surrogate of protection in the absence of phase 3 efficacy trials.
GBS type III strain M781 was grown in 200 ml of Todd-Hewitt broth to an optical density at 650 nm of 0.3 (mid-exponential growth). GBS cells were pelleted by centrifugation and fixed by suspending cells in 70% ethanol for 15 min at room temperature. Fixed GBS were then washed by centrifugation thrice with 10 mM phosphate-buffered saline (PBS, pH = 7.4; Gibco, Grand Island, NY) and vortexed extensively to break any chains. Fixation was confirmed by lack of growth on blood agar plates.
Alexa Fluor 488 carboxylic acid succinimidyl ester mixed isomer dye (1 mg; Invitrogen, Molecular Probes, Eugene, OR) was suspended in 50 μl of dimethylformamide and combined with 20 mg of ethanol-fixed GBS resuspended in 1 ml of 0.2 M NaHCO3. The labeling reaction was incubated at room temperature in the dark with mixing for 1 h. Fluorescently labeled GBS were repeatedly washed with 10 mM PBS to remove excess dye. Fluorescent labeling of GBS was verified by microscopy (Zeiss Axioskop 2 plus, Carl Zeiss Ltd., Maple Grove, MN). GBS were diluted in 10 mM PBS (pH 7.4) to a concentration of 3.6 × 107 cells/ml, aliquoted, and stored at -80°C.
Conjugation of recombinant Alp3 (rAlp3) protein to fluorescent microspheres was achieved by combining 2 mg of purified rAlp3,9 with 1 ml of FluoSpheres aldehyde-sulfate microspheres (505/515, 1.0 μm; Invitrogen, Molecular Probes, Eugene, OR) in 1.5 ml of 50 mM phosphate buffer (pH = 6.5). Sodium cyanoborohydride (15 mg; Matreya, Pleasant Gap, PA) was added to the mixture, and the reaction was incubated in the dark at room temperature overnight on an orbital rotator.
rAlp3-conjugated microspheres were washed four times in 50 mM PBS (pH = 7.4) to remove residual sodium cyanoborohydride and unconjugated protein. The microspheres were then stored in the dark at 4°C. Microspheres were diluted in MEM to the desired concentration prior to the assay.
Pre-vaccination and immune sera from rabbits that received two doses of III-TT vaccine were used.11 The levels of type III CPS-specific IgG induced from the vaccine ranged from 0.38 to 27.47 μg/ml. Standard rabbit reference sera specific to GBS type III CPS contained 0.93 mg IgG/ml.
The ability of rabbit antisera to opsonize fluorescently labeled GBS M781 or rAlp3-coupled microspheres for killing in the presence of complement was evaluated using an in vitro OPA as described previously,8 with minor exceptions.
Briefly, 25 μl of heat-inactivated rabbit serum (56°C, 30 min) was combined with 25 μl modified Eagle'ss medium, 25 μl of baby rabbit complement diluted to yield a final dilution of 1% in the reaction mixture (Cedarlane, Burlington, NC), 150 μl dHL-60 cells (4.5 × 106 cells), and 25 μl of either labeled M781 (~9 × 105 bacteria) or rAlp3-conjugated microspheres (~1.5 × 106 microspheres). Control reactions excluded complement and/or antibody, all components except the target, all components except for the effector cells, or used pre-immunization serum. Reaction mixtures were incubated at 37°C for 1 h with end-over-end mixing.
Cytometric analyses were used to evaluate uptake of the target by effector cells. Samples were assayed on a Cell Lab Quanta cytometry system (Beckman Coulter, Fullerton, CA), and results were reported as the percentage of uptake above background levels observed in the negative controls.
Uptake of the target by effector cells was visually confirmed using an LSM 5 PASCAL confocal microscope (Carl Zeiss Ltd., Maple Grove, MN). OPA setup and incubation was performed as described above. After incubation, dHL-60s were fixed in 2% paraformaldehyde for 1 h before staining with Alexa Fluor 555 Wheat Germ Agglutinin Conjugate (Invitrogen Molecular Probes, Eugene, OR) to delineate cell boundaries. Stain (1 μg of dye in 750 μl of cold MEM) was added to each 250 μl reaction mixture and allowed to sit on ice for 15 min. Cells were pelleted by centrifugation (2,000 xg for 1 min) and washed once in fresh MEM prior to visualization. Image slices were taken to confirm that GBS/microspheres were internalized and not externally associated with dHL-60 cells. Three-dimensional rendering of confocal images was performed at the Harvard NeuroDiscovery Center with use of Imaris Suite 5.5 software (Bitplane Inc., Saint Paul, MN).
We are indebted to Keith Crawford at BWH for access to the flow cytometer, to Mark Kazmierczak at the Channing Laboratory for helpful advice with confocal microscopy, and to Michael Hickey, Flow Applications, Inc., for technical help and reagents. We also thank Lai Ding at the Harvard NeuroDiscovery Center for performing the 3-D rendering of confocal images and Paul Guttry for review of this manuscript. This research was supported by NIH-NIAID grant AI-060603.