Previous Article
Infection and Immunity, August 1999, p. 4303-4305, Vol. 67, No. 8
0019-9567/99/$04.00+0
LETTERS TO THE EDITOR
Measurement of Human Antibodies to Type III Group B
Streptococcus
 |
LETTER |
We disagree with the conclusions reached by Bhushan et al.
(5) in their study comparing different methods of coating
enzyme-linked immunosorbent assay (ELISA) plates for estimation of
group B Streptococcus type III (GBS III)
polysaccharide-specific antibody concentrations in human serum. We
believe that interpretation of the data as suggested in their paper
could lead to serious errors in the use of serum antibody
concentrations as surrogates of vaccine efficacy. We propose four
alternate conclusions that are supported by the data given by Bhushan
et al. (5).
(i) ELISAs that used free GBS III polysaccharide or
polysaccharide mixed with methylated human serum albumin (mHSA) as a
coating antigen failed to detect all of the antibodies reactive with
type III polysaccharide in a quantitative precipitin assay. Bhushan et al. found that ELISA methods using GBS III polysaccharide alone or type III polysaccharide mixed with mHSA as coating antigens measured substantially lower concentrations of specific immunoglobulin G (IgG) in a reference serum than did methods using polysaccharide covalently conjugated to HSA or to biotin. However, the specific IgG
concentration detected by the conjugated polysaccharide methods in
ELISAs agreed very closely with the concentration of specific antibodies in the reference serum (standard human reference serum III
[SHRS III]) determined by quantitative precipitin analysis with the
purified type III polysaccharide as antigen, while the ELISA methods
using unconjugated polysaccharide underestimated the antibody
concentration by more than 50%. The data provided in the paper are
inadequate to permit the assessment of the relative sensitivities of
the free versus conjugated polysaccharide methods, but experiments in
our laboratory have shown that assays using free polysaccharide as the
coating antigen are less sensitive than those using polysaccharide
conjugated to HSA (III-HSA) (6).
(ii) Antibody measurements in the III-HSA assay are highly
correlated with antibody measurements in the RABA. The radioactive antigen binding assay (RABA) is an assay in which soluble
polysaccharide (unencumbered by attachment to a plastic surface) is
allowed to interact in solution with antibodies; the antibody
concentration is determined by quantifying the amount of polysaccharide
complexed to specific antibodies. The RABA is the gold standard assay
because it is the only immunoassay shown to give results that correlate with neonatal susceptibility to GBS III infection (1-3).
To test the validity of the unconjugated polysaccharide ELISA,
III-specific IgG in 16 serum specimens from healthy adults was measured
by ELISA independently in the laboratory of Bhushan et al. at the Food
and Drug Administration (FDA) and by RABA in the laboratory of one of
us (C.J.B.). Both the FDA and the Baylor laboratories also measured
specific IgG concentrations by the III-HSA ELISA method. Antibody
concentrations determined by the RABA and the III-HSA ELISA
(irrespective of the laboratory in which the III-HSA ELISA was
performed) were in better agreement than those determined by RABA and
the free polysaccharide plus mHSA method (unpublished data). These
results support the previously published excellent correlation
(r = 0.92) between antibody concentrations determined
by the III-HSA ELISA and by the RABA (6). Of particular concern is the overestimation of specific antibodies by the free polysaccharide plus mHSA method in the eight serum samples that contained <1.0 µg of specific antibodies per ml according to RABA. In those eight samples, the range of specific IgG concentrations determined by III-HSA ELISA was <0.1 to 0.5 µg/ml at the FDA and <0.05 to 0.47 µg/ml at Baylor. By contrast, the free polysaccharide plus mHSA method measured >1.0 µg of specific IgG per ml (range, 1.0 to 4.2 µg/ml) in each of the samples. Overestimates by the free
polysaccharide plus mHSA method are of serious concern because Bhushan
et al. propose to use this ELISA to establish the minimum level of
maternal antibody that is protective against neonatal infection.
Overestimation of specific antibody levels would yield a falsely
elevated value for the minimum protective level in such studies.
(iii) The III-HSA ELISA measures antibodies that cross-react with
PN-14 polysaccharide but that also bind to GBS III polysaccharide. Bhushan et al. suggest that conjugation of the type III polysaccharide to a protein alters the antigenic specificity of the polysaccharide. Because the type III ELISA methods using conjugated polysaccharide also
detected antibodies cross-reactive with the structurally related type
14 pneumococcal (PN-14) polysaccharide, the investigators conclude that
the conjugated polysaccharide ELISAs are less specific than assays
using free polysaccharide. Antibody cross-reactions between GBS III and
PN-14 polysaccharides have been recognized for many years (4,
6-8). We have reported that in some subjects immunization with
the purified type III polysaccharide evokes antibodies that cross-react
with PN-14 polysaccharide (9). Of importance is that these
are truly cross-reacting antibody populations: the GBS III-HSA ELISA
measured no increase in specific antibodies in the postimmunization
sera of four patients who responded to vaccination with PN-14
polysaccharide (6). Thus, the III-HSA ELISA detects
antibodies that bind to type III polysaccharide and cross-react with
PN-14, but it does not detect all antibodies that recognize PN-14. The
lower sensitivity of the free polysaccharide method accounts for the
inability of this method to detect lower-affinity antibodies such as
those that cross-react with PN-14 polysaccharide.
(iv) The III-HSA ELISA, but not the free polysaccharide ELISA,
has been shown to correlate directly with opsonic activity of serum
against GBS III. In addition to the problem of assay sensitivity,
it is critical to determine which assay system gives results that most
closely reflect the concentration of functionally active and protective
antibodies. Specific IgG concentrations measured by the III-HSA method
have been shown to correlate with opsonophagocytic killing activity of
the antisera to GBS III in sera from subjects immunized either with
type III polysaccharide or with polysaccharide conjugated to tetanus
toxoid (10). No such relationship with functional activity
has been demonstrated for antibody detection by the free polysaccharide method.
On the basis of these data and the results presented in the article by
Bhushan et al. (5), neither the free polysaccharide method
nor the polysaccharide plus mHSA ELISA method can be considered a valid
technique for measurement of naturally acquired GBS III polysaccharide-specific IgG in human serum.
| |
REFERENCES |
| 1.
|
Baker, C. J., and D. L. Kasper.
1976.
Correlation of maternal antibody deficiency with susceptibility to neonatal group B streptococcal infection.
N. Engl. J. Med.
294:753-756[Abstract].
|
| 2.
|
Baker, C. J.,
D. L. Kasper,
I. B. Tager,
A. Paredes,
S. Alpert,
W. M. McCormack, and D. Goroff.
1977.
Quantitative determination of antibody to capsular polysaccharide in infection with type III strains of group B Streptococcus.
J. Clin. Investig.
59:810-818.
|
| 3.
|
Baker, C. J.,
M. S. Edwards, and D. L. Kasper.
1981.
Role of antibody to native type III polysaccharide of group B Streptococcus in infant infection.
Pediatrics
68:544-549[Abstract/Free Full Text].
|
| 4.
|
Baker, C. J.,
D. L. Kasper,
M. S. Edwards, and G. Schiffman.
1980.
Influence of preimmunization antibody levels on the specificity of the immune response to related polysaccharide antigens.
N. Engl. J. Med.
303:173-178[Abstract].
|
| 5.
|
Bhushan, R.,
B. F. Anthony, and C. E. Frasch.
1998.
Estimation of group B Streptococcus type III polysaccharide-specific antibody concentrations in human sera is antigen dependent.
Infect. Immun.
66:5848-5853[Abstract/Free Full Text].
|
| 6.
|
Guttormsen, H.-K.,
C. J. Baker,
M. S. Edwards,
L. C. Paoletti, and D. L. Kasper.
1996.
Quantitative determination of antibodies to type III group B streptococcal polysaccharide.
J. Infect. Dis.
173:142-150[Medline].
|
| 7.
|
Jennings, H. J.,
C. Lugowski, and D. L. Kasper.
1981.
Conformational aspects critical to the immunospecificity of the type III group B streptococcal polysaccharide.
Biochemistry
20:4511-4518[Medline].
|
| 8.
|
Jennings, H. J.,
K.-G. Rosell, and D. L. Kasper.
1980.
Structural determination and serology of the native polysaccharide antigen of type-III group B Streptococcus.
Can. J. Biochem.
58:112-120[Medline].
|
| 9.
|
Kasper, D. L.,
C. J. Baker,
R. S. Baltimore,
J. H. Crabb,
G. Schiffman, and H. J. Jennings.
1979.
Immunodeterminant specificity of human immunity to type III group B Streptococcus.
J. Exp. Med.
149:327-339[Abstract/Free Full Text].
|
| 10.
|
Kasper, D. L.,
L. C. Paoletti,
M. R. Wessels,
H.-K. Guttormsen,
V. J. Carey,
H. J. Jennings, and C. J. Baker.
1996.
Immune response to type III group B streptococcal polysaccharide-tetanus toxoid conjugate vaccine.
J. Clin. Investig.
98:2308-2314[Medline].
|
| | | | |
Dennis L. Kasper
Michael R. Wessels
Hilde-Kari Guttormsen
Lawrence C. Paoletti
Channing Laboratory
Department of Medicine
Brigham and Women's Hospital
Department of Microbiology and Molecular Genetics
Harvard Medical School
Boston, Massachusetts
|
| | | | |
Morven S. Edwards
Carol J. Baker
Department of Pediatrics, Microbiology and Immunology
Baylor College of Medicine
Houston, Texas
|
| |
AUTHORS' REPLY |
Methods for estimation of human antibodies to the type III
polysaccharide (PS) of group B streptococcus (GBS) should specifically measure antibodies to the type III PS. Studies by Bhushan et al. (4) were initiated to compare different ELISA antigens for their sensitivity and specificity for measurement of immunoglobulin G
(IgG) anti-GBS PS antibodies. We found the enzyme-linked immunosorbent assay (ELISA) described by Guttormsen et al. (7) to be
sensitive but it lacked the needed specificity for estimation of GBS
type III PS antibodies as shown in Fig. 1A of Guttormsen et al.
(7) and Fig. 5 of Bhushan et al. (4), where
binding to the pneumococcal type 14 (PN-14) PS was evident. The lack of
specificity was overcome by working out conditions for attachment of
the free GBS type III PS to the ELISA plate.
While there is no perfect method for estimation of these antibodies, in
our hands, free GBS type III PS bound directly to the plate or mixed
with methylated human serum albumin (mHSA) measured most specifically
GBS type III antibodies. The sensitivity was sufficient to measure at
least 0.05 µg of IgG antibody per ml, similar to that reported by
Guttormsen et al. for their assay (7).
We have addressed the four alternative conclusions proposed by Kasper
and coworkers. We disagree with their conclusions proposed in response
to our paper (4).
(i) The question is about antibody specificity, not use of quantitative
precipitation per se. We have compared antibody concentrations estimated using four different PS preparations in reference sera, sera
from immunized and from nonimmunized adults. We found that apparent
assay sensitivity was dependent on the serum used. When reference serum
19, a pool of sera from adults immunized with a GBS tetravalent PS
vaccine containing low levels of PN-14 antibody, was used at an
antibody concentration of 40 to 80 ng/ml, antibodies bound equally to
both free and conjugated GBS type III antigen preparations
(4). When the SHRS-III reference serum, a serum pool
prepared from five individuals receiving GBS type III-tetanus toxoid
conjugate, was used at the same antibody concentration, marked
differences were seen between antibody binding to conjugated PS
preparations and to free PS or PS mixed with mHSA. Antibody concentrations calculated in micrograms per milliliter for two GBS
hyperimmune immune globulin intravenous (IGIV) preparations (004 and
006) were also comparable with all four antigen preparations (4,
6). Thus, when the postvaccination serum contained mostly GBS
type III-specific PS antibodies, there was little difference in
antibody binding among the four antigen preparations, but when the
serum also contained antibodies to the PN-14 PS, then marked differences were seen in measured antibody concentrations among the
different antigen preparations. Analyzing SHRS-III, when PS alone or PS
mixed with mHSA were used as ELISA coating antigens, the GBS type III
antibody concentration was less than half (33 µg/ml) of that
estimated when PS conjugated to biotin or HSA (81 to 83 µg/ml) was
used as coating antigen. The large differences in antibody
concentrations estimated in SHRS-III are because SHRS-III also contains
41.0 µg of PN-14 PS antibody per ml.
(ii) Kasper and coworkers correlate antibody concentrations determined
by GBS III-HSA conjugate ELISA with a radioactive antigen binding assay
(RABA). The GBS PSs induce IgG, IgA, and IgM antibodies (1, 2,
7). Our purpose was to specifically measure those antibodies that
will potentially cross the placenta to protect the neonate. We
therefore made no comparisons with RABA. Guttormsen et al.
(7) showed that GBS III RABA did not measure antibody concentrations <1 µg/ml because of the limitation in preparing labeled PS of sufficiently high specific activity. RABA also does not
distinguish among immunoglobulin isotypes or subclasses (3, 7) and measures only precipitable antibody. Protection against GBS disease is mediated by opsonsophagocytosis and is the functional correlate of protection.
(iii) Since antibodies to PN-14 PS are not protective against GBS
disease (8), we think an assay that specifically measures antibody to GBS type III PS, and is not sensitive to binding of PN-14
PS antibodies, would be appropriate to determine maternal anti-PS
antibody concentrations necessary for protection against GBS disease.
Earlier studies done by Kasper and coworkers (8) showed that
rabbit antiserum to PN-14 did not react with GBS type III native PS in
agar gel diffusion but reacted strongly with partially desialyated GBS
core PS. They immunized a group of adult volunteers with GBS type III
native PS, GBS core PS, or polyvalent pneumococcal PS vaccine. Sera
from immunized volunteers receiving the GBS type III vaccine were
highly opsonic, while only 1 of 12 serum samples from volunteers
receiving GBS core PS or pneumococcal vaccine had an opsonic response
to GBS type III. They also demonstrated that development of opsonic
antibody and natural immunity to GBS was correlated with antibody to
the native GBS type III PS rather than antibody to the core PS. Thus,
if antibody to PN-14 does not protect against GBS disease, why use an
assay that is also sensitive to binding of antibodies to the PN-14 PS?
We are concerned that GBS III-HSA ELISA will overestimate antibody
concentrations if the serum contains PN-14 PS antibody and will
underestimate GBS III-specific antibodies because the change in
conformation of the PS (after conjugation) that results in a
PN-14-reactive epitope replaces an alternative GBS type III-specific epitope.
We have also compared four different GBS Ia PS preparations as coating
antigens in ELISA to determine GBS type Ia PS-specific antibodies (free
PS, PS mixed with mHSA, PS conjugated to biotin, or PS conjugated to
HSA) (5). These PS preparations were used to evaluate GBS Ia
antibody in sera from recipients of a GBS tetravalent PS (Ia, Ib, II,
or III) vaccine, in sera from women receiving GBS type Ia-conjugated PS
vaccine, and in sera from nonimmunized healthy women of childbearing
age. Unlike the GBS type III ELISA, GBS Ia PS antibody concentrations
estimated by all four PS antigen preparations were similar
(5).
(iv) IgG concentrations measured by using free PS or PS mixed with mHSA
correlate well with opsonic activity. The correlation coefficients of
ELISA titers versus opsonic titers were 0.90 by using PS mixed with
mHSA as the coating antigen and 0.60 by using PS conjugated to HSA for
sera from nonimmunized women. We have also looked at GBS type III
postvaccination sera from adults, and the correlation coefficients for
antibody concentrations versus opsonic titer for PS mixed with mHSA and
for PS conjugated to HSA were 0.63 and 0.52, respectively
(5). Thus, an ELISA using free GBS type III PS or PS mixed
with mHSA correlates better with functional activity.
In conclusion, conjugate vaccines are currently under investigation for
protection of women and neonates against GBS disease (9).
There is always a potential problem when antibody responses induced by
a conjugate vaccine (for example GBS type III conjugated to tetanus
toxoid) are measured by using antigens similar to the vaccine (GBS
III-HSA). We found that the GBS III-HSA ELISA overestimated the
SHRS-III response because it also measured antibodies to a PN-14 PS
reactive epitope created following conjugation. Since it is unlikely
that classical randomized efficacy trials will be possible in the
United States because of the highly successful implementation of the
Centers for Disease Control and Prevention prophylaxis guidelines
(10), it is important to use antibody assays that
specifically measure IgG anti-GBS PS antibodies.
| |
REFERENCES |
| 1.
|
Anthony, B. F., and N. F. Concepcion.
1989.
Opsonic activity of human IgG and IgM antibody for type III group B streptococci.
Pediatr. Res.
58:3663-3670.
|
| 2.
|
Anthony, B. F.,
N. F. Concepcion,
C. A. Wass, and D. C. Heiner.
1984.
Immunoglobulin G and M composition of naturally occurring antibody to type III group B streptococci.
Infect. Immun.
46:98-104[Abstract/Free Full Text].
|
| 3.
|
Basham, L. E.,
V. Pavliak,
X. Li,
A. Hawwari,
K. L. Kotloff,
R. Edelman, and A. Fattom.
1996.
A simple, quantitative, reproducible avidin-biotin ELISA for the evaluation of group B streptococcus type-specific antibodies in humans.
Vaccine
14:439-445[Medline].
|
| 4.
|
Bhushan, R.,
B. F. Anthony, and C. E. Frasch.
1998.
Estimation of group B streptococcus type III polysaccharide-specific antibody concentrations in human sera is antigen dependent.
Infect. Immun.
66:5848-5853.
|
| 5.
|
Bhushan, R., and C. E. Frasch.
1998.
Estimation of GBS type Ia- and V-specific antibodies with different polysaccharide antigen preparations, abstr. G-92, p. 309.
In
Program and abstracts of the 38th Interscience Conference on Antimicrobial Agents and Chemotherapy. American Society for Microbiology, Washington, D.C.
|
| 6.
|
Fischer, G. W.,
V. G. Hemming,
K. W. Hunter, Jr.,
H. Gloser,
H. Bachmayer,
C. E. Von Pilar,
T. Helting,
L. E. Weisman,
S. R. Wilson, and P. A. Baron.
1986.
Intravenous immunoglobulin in the treatment of neonatal sepsis: therapeutic strategies and laboratory studies.
Pediatr. Infect. Dis.
5:S171-S175[Medline].
|
| 7.
|
Guttormsen, H. K.,
C. J. Baker,
M. S. Edwards,
L. C. Paoletti, and D. L. Kasper.
1996.
Quantitative determination of antibodies to type III group B streptococcal polysaccharide.
J. Infect. Dis.
173:142-150.
|
| 8.
|
Kasper, D. L.,
C. J. Baker,
R. S. Baltimore,
J. H. Crabb,
G. Schiffman, and H. J. Jennings.
1979.
Immunodeterminant specificity of human immunity to type III group B streptococcus.
J. Exp. Med.
149:327-339.
|
| 9.
|
Kasper, D. L.,
L. C. Paoletti,
M. R. Wessels,
H. K. Guttormsen,
V. J. Carey,
H. J. Jennings, and C. J. Baker.
1996.
Immune response to type III group B streptococcal polysaccharide-tetanus toxoid conjugate vaccine.
J. Clin. Invest.
98:2308-2314.
|
| 10.
|
Schuchat, A.
1999.
Group B streptococcus.
Lancet
353:51-56[Medline].
|
| | | | |
Reva Bhushan
Carl E. Frasch
Division of Bacterial Products
Center for Biologics Evaluation and Research
Bethesda, Maryland
|
Infection and Immunity, August 1999, p. 4303-4305, Vol. 67, No. 8
0019-9567/99/$04.00+0
Top
Letter
References
