Pneumococcal surface protein A (PspA) is a multi-factorial virulence factor that functions in complement evasion and lactoferrin binding. In this study the authors investigate antibody responses to PspA using a panel of sera collected from infected individuals. The data presented shed new light on the specificity of IgG responses to PspA and its relationship to the production of serotype specific antibodies to surface polysaccharides.
Human antibody responses to pneumococcal surface protein A and capsular polysaccharides during acute and convalescent stages of invasive disease in adult patients
Article first published online: 21 NOV 2013
© 2013 Federation of European Microbiological Societies. Published by John Wiley & Sons Ltd. All rights reserved
Pathogens and Disease
Volume 70, Issue 1, pages 40–50, February 2014
How to Cite
Kolberg, J., Aase, A., Næss, L. M., Aaberge, I. S. and Caugant, D. A. (2014), Human antibody responses to pneumococcal surface protein A and capsular polysaccharides during acute and convalescent stages of invasive disease in adult patients. Pathogens and Disease, 70: 40–50. doi: 10.1111/2049-632X.12106
- Issue published online: 12 FEB 2014
- Article first published online: 21 NOV 2013
- Accepted manuscript online: 22 OCT 2013 02:06PM EST
- Manuscript Accepted: 23 SEP 2013
- Manuscript Revised: 13 SEP 2013
- Manuscript Received: 26 APR 2013
- Streptococcus pneumonia ;
- pneumococcal surface protein A;
- monoclonal antibodies;
- human antibodies;
The IgG antibody responses to pneumococcal surface protein A (PspA) and capsular polysaccharides in acute and convalescent-phase sera from 10 adult patients with invasive pneumococcal disease were analysed. The relatedness between the strains were characterized by capsular serotyping (1, 4, 7F, 9V, 12F and 19F), multilocus sequence typing (MLST) and sequencing of the gene coding for PspA. Immunoblotting with the patient's own infecting strain used as whole cell antigen revealed strong antibody responses to PspA in 4 of 10 patients. Two of these patients showed cross reactivity of PspA antibodies within PspA families 1 and 2 by ELISA measurements with recombinant PspA proteins. Using ELISA, we found increased levels of capsular-specific antibodies during convalescent phase in 9 of 10 patients. All patients, except one, revealed low antibody levels in their acute phase sera. The binding of serum antibodies to live pneumococci using the patient's own infective strain was measured by flow cytometry. The antibodies binding to the live pneumococci corresponded to the serotype-specific polysaccharides by ELISA. Low antibody-binding activities to their infective strain in the acute serum may explain why they were not protected.
The Gram-positive bacterium Streptococcus pneumoniae is an important human pathogen causing a variety of diseases including pneumonia, meningitis, sepsis and otitis media and is a major cause of morbidity and mortality worldwide, particularly among the young and elderly. The pneumococcal polysaccharide capsule, of which there are more than 90 distinct serotypes, is probably the most important virulence factor due to its antiphagocytic properties (Jonsson et al., 1985; Sanchez et al., 2011). Capsular polysaccharides are the main component of current vaccines, and selected polysaccharides are used alone or conjugated to a carrier protein [for review see (Gladstone et al., 2011)]. There is also work on serotype independent vaccines, and the most extensively studied protein candidate is the pneumococcal surface protein A (PspA).
Important functions of PspA are its ability to inhibit complement deposition at the bacterial cell surface and binding of lactoferrin to prevent lactoferrin-mediating killing (Hammerschmidt et al., 1999; Mukerji et al., 2012). PspA has three major sequence domains, a C-terminal choline-binding region that attaches the protein to the cell surface, a proline-rich region and a N-terminal well-exposed domain which is thought to form an antiparallel coiled-structure (Yother & White, 1994; Jedrzejas et al., 2000). The PspA proteins are exceptionally diverse in their α-helical regions resulting in a considerable molecular and antigenic heterogeneity between different pneumococcal isolates (Hollingshead et al., 2000). The α-helical domain includes a clade defining region which is used to divide PspA molecules into three different families of which family 1 comprises clades 1 and 2; family 2, clades 3,4 and 5, and family 3, clade 6 (Hollingshead et al., 2000). Over 98% of PspAs belong to the highly cross-reactive families 1 and 2 (Baril et al., 2004).
Several studies have shown PspA to be highly immunogenic and protective in animal infection models (Roche et al., 2003; Oliveira et al., 2010; Hotomi et al., 2011), whereas data on the response to this protein in severe human infections are relatively scarce (Baril et al., 2004; Linder et al., 2007). Furthermore, in contrast to polysaccharide vaccine studies, few data exist on antibody responses to these structures during pneumococcal infection (van Mens et al., 2011). In this study, we have analysed the IgG antibody responses in detail to PspA and polysaccharides in acute and convalescent-phase sera from 10 adult patients with invasive pneumococcal disease using immunoblotting and ELISA. Unlike other studies, we have used the patient's own isolate as antigen to study the highly variable PspA proteins by immunoblotting. In addition, we have measured the binding of serum antibodies to the patient's own invading strain (live pneumococci) using flow cytometry. By this method, we can quantify the total amount of antibodies that may exert functional activity.
Materials and methods
Patients, sera and bacterial isolates
The materials used in this study were collected in 1993/1994 from hospitalized adult patients with invasive pneumococcal disease and have partly been described elsewhere (Kolberg et al., 2000). The capsular serotype of the isolates obtained from blood was determined by the Quellung reaction using specific antisera from Statens Serum Institut, Copenhagen, Denmark. The patients and strains characteristics are summarized in Table 1. Acute sera were obtained on the day of admission to the hospital and early convalescent sera > 2 weeks later. For some of the patients, late convalescent sera were obtained > 4 weeks later. The pneumococcal strain Spn 51 and its PspA knock-out mutant were obtained from Dr Sven Hammerschmidt, Greifswald, Germany (Kolberg et al., 2006). Pooled sera from seven healthy adults (mean age in years: 57, range 36–65) who had not been given any S. pneumoniae vaccine and without known pneumococcal diseases were used as a normal control serum.
|Serotype||Strain||PspAa Family||Clade||ST||Pasient ID||Ageb||Clinical disease||Daysc|
|7F||1679/94||2||3||191||1679||59||Moderately severe pneumonia||3|
|7F||1678/94||2||3||191||1678||42||Severe pneumonia with complications||2|
|9V||1674/94||2||3||162||1674||55||Severe pneumonia with complications||3|
|12F||1675/94||1||2||218||1675||76||Severe pneumonia with complications||1|
PspA sequencing and typing
The central region of the pspA gene was amplified using the oligonucleotide primers LSM13 and SKH2, as described in the study of Heeg et al. (2007). PCR products were purified and sequenced as described under using the primers described by Heeg et al. (2007). Sequence edition was performed using the SeqScape version 2.5 (Applied Biosystems). Sequence alignment was performed using mega version 5.1 and clustal omega (http://www.ebi.ac.uk/Tools/webservices/services/msa/clustalo). PspA typing was performed by using the nucleotide sequences of the isolates in the http://www.ncbi.nlm.nih.gov/blast program to assess their relatedness to reference strains of known clade types. Phylogenetic relationships were illustrated using the Neighbour-Joining method in Clustal Omega.
Multilocus sequence typing (MLST)
Genomic DNA was isolated by boiling of a 1 μL loopful of bacteria in 100 μL Tris-EDTA buffer for 10 min. After centrifugation at 16 000 g for 5 min, the supernatants were stored at −20 °C. MLST was performed as described elsewhere (Enright & Spratt, 1998). Sequencing reactions were performed with an ABI Prism BigDye Terminator cycle sequencing ready reaction kit (ABI, Foster City, CA) according to the manufacturer's recommendation. Sequencing was performed using an ABI 3730 DNA analyzer (Applied Biosystems). The sequence types (STs) were obtained by using the MLST database (http://www.mlst.net).
SDS-PAGE and immunoblotting
Whole bacterial cell lysates were subjected to SDS-PAGE with 4% stacking gel and 7.5% separating gel, electrotransferred to a nitrocellulose membrane which were blocked with Sigma's casein-based blocking buffer over night or for 90 min at room temperature. The human sera were diluted 1 : 1000 in blocking buffer and incubated for 2 h at room temperature. The mAbs were used at different dilutions to give signals similar to that of the patient sera. The nitrocellulose strips were washed five times in PBS with 0.05% Tween®20 and incubated for 1 h with peroxidase-labelled rabbit anti-human IgG or anti-mouse immunoglobulins. These secondary antibodies were from DakoCytomation, Denmark, and used at a dilution of 1 : 1000 for 1 h. After washing, the strips were in 1 min covered with chemiluminescence reagent (Western lightning® Plus-ECL) from PerkinElmer, Inc., Boston, MA, placed between two stiff plastic sheets and then scanned via Kodak Image Station 2000 in the first part of the study and later with Carestream 2200 Pro or 4000R Pro. The major band detected on the strips incubated with PspA-specific mAbs was used as a reference to find the position of the major PspA antigen on strips incubated with human sera. Kodak Imaging (Carestream) Software was used for calculating the intensity of IgG binding to the PspA antigen.
The mAbs used in this study are listed in Table 2.
|143,F-2||IgA||Epitope within aa 1–450 of PspA from TIGR4||This work and (Kolberg et al., 2001)|
|149,B-3||IgG2a||Epitope within aa 1–450 of PspA from TIGR4||This work and (Kolberg et al., 2006)|
|159,D-7a||IgG2a||Epitope within aa 67–236 of PspA from strain Rx1 and aa 1–311 of strain 1675/94||This work and (Kolberg et al., 2003)|
|160,F-4a||IgG1||Epitope within aa 1–67 of PspA from strain Rx1 and aa 1–311 of strain 1675/94||This work and (Kolberg et al., 2003)|
|169,H-6a||IgG1||Epitope within aa 67–236 of PspA from strain Rx1 and aa 1–311 of strain 1675/94||This work and (Kolberg et al., 2003)|
|183,F-6a||IgM||Positive reactions with Spn 51 and nonreactive with the pspA negative mutant. Epitope not within aa 1–303 of Rx1 PspA. Epitope surface exposed on strain Rx1 shown by flow cytometry.||This work and (Kolberg et al., 2003)|
ELISA measurements of antibodies against family 1 and 2 PspA proteins
Recombinant proteins were produced in E. coli by GenScript, Hong Kong, based on the available sequence for amino acid (aa) residues 1–450 of PspA from TIGR4 (family 2) and our sequence analyses for aa 1–311 for PspA from strain 1675/94 (family 1). The received PspA fragments contained a C-terminal His-tag, and the purity was > 75% after a one-step affinity purification by a Ni-Hi Trap column. Analyses showed the expected positive and negative reactions with the mAbs used in this study.
The two proteins were used for the separate analyses of IgG antibodies against PspA of family 1 and 2. Nunc MaxiSorp Immuno plates (Roskilde, Denmark) were coated over night at 4 °C with 2.0 μg protein/ml in PBS with 0.02% sodium azide. Plates were washed with PBS with 0.05% Tween®20 in AquaMax 4000 Microplate Washer (Molecular Devices, CA), blocked with 10% skim milk in PBS for 1 h at 37 °C. The blocking buffer was removed without washing. Dilutions of serum samples in 100 μL 5% skim milk were set up in triplicates. The plates were incubated for 2 h at 37 °C, washed and then incubated for 90 min at 37 °C with peroxidize-conjugated polyclonal goat anti-human IgG (Sigma) diluted 1 : 5000. The plates were washed followed by addition of the substrate p-nitrophenyl phosphate (1 mg mL−1) in 10% diethanolamine buffer. The second convalescent serum from patient 1675 was used as a standard, with a dilution of 1 : 1000 and was assigned an antibody titre of 1 U mL−1.
ELISA measurement of IgG antibodies against capsular polysaccharides
The procedure is based on the WHO published method (www.vaccine.uab.edu and Goldblatt et al., 2011). Capsular polysaccharides of types 1 and 11A were from American Type Culture Collection, the others capsular and cell wall polysaccharides were obtained from Statens Serum Institut, Copenhagen, Denmark. For serotypes 11A and 12F, the IgG antibody concentrations were not determined in the new human pneumococcal reference serum, 007sp (Goldblatt et al., 2011), and for these two types, we therefore used reference serum 89-SF (Quataert et al., 2004). Briefly, the plates were coated with 2.5 μg mL−1 of capsular polysaccharides. The sera were diluted in absorption solution containing 5.0 μg mL−1 of cell wall polysaccharides and 5.0 μg mL−1 of 22F capsular polysaccharides followed by incubation for 20 h at 18 °C. After washing, the plates were incubated for 2 h at room temperature with secondary antibodies as described previously. The plates were then washed, followed by addition of the substrate.
Binding of antibodies in patient sera to their own infecting strains using flow cytometry
Antibodies against the infecting strain were measured by incubating the patient's sera with their own isolate (Kolberg et al., 2011). By this method, we can quantify the antibodies binding to live bacteria; these are the antibodies that may exert functional activity. The pneumococci were grown over night on Colombia agar in 5% CO2. Colonies were harvested and dispersed in Hanks balanced salt solution supplemented with 2% BSA (HBSS-BSA) to a density of 0.7 at 650 nm. The WHO pneumococcal reference serum 007sp was used to make a standard curve for the serotypes 1, 4, 7F, 9V and 19F, whereas the reference serum 89-SF was used as a standard for the serotypes 11A and 12F. Twofold dilutions of patients' sera or reference sera (50 μL) were incubated with 5 μL of bacteria at 37 °C for 45 min followed by two washing steps. Next, bound antibodies were detected with a phycoerythrin (PE)-conjugated anti-human IgG (Sigma) diluted 1 : 100 for 30 min followed by one wash step. Finally, the bacteria were resuspended in HBSS-BSA containing 1 μg mL−1 Hoechst 34580. The fluorescence (PE) intensity of bound antibodies was analysed using flow cytometry (Attune®, Applied Biosystems) gating on combined forward scatter and the Hoechst 34580 fluorescence. The geometric mean fluorescence of patient's serum was interpolated on the corresponding pneumococcal reference serum (standard curve) to calculate the specific antibody IgG concentration in arbitrary units per mL (AU per mL).
Patients and strain characteristics
The infecting strains from 10 adult patients with invasive pneumococcal disease (uncomplicated pneumonia to pneumonia with meningitis) covering seven serotypes are listed in Table 1.
Capsular, PspA and MLST types of the isolates
Three of the strains were of capsular type 7F, two were 9V and the five other isolates were of types 1, 4, 11A, 12F and 19F, respectively (Table 1). The classification of the isolates into PspA clades was performed by comparison of the nucleotide sequences with those of the PspA typed reference strains given in (Donati et al., 2010). The identity of the matches was 98–100%. Among the 10 isolates, PspA family 2 was expressed in eight of which seven were clade 3 and one clade 4, the two other strains belonged to family 1 (Table 1). The classification of isolate 1675/94 to clade 2 was supported by its reactivity with mAb 159,D-7 which detects an epitope within amino acid (aa) residues 67–236 of the clade 2 reference strain Rx1 (Kolberg et al., 2003). A dendrogram based on the PspA sequences was constructed (Fig. 1). The two 9V strains showed identical nucleotide sequences in the N-terminal region (1200–1500 nucleotides) and were both ST 62. The three 7F strains revealed identical nucleotide sequences and were ST 191. In the PspA N-terminal region showed 1532 nucleotides of strain 1672/94 (11A), 99% identity with the invasive isolate AP200 (11A) from Italy (Camilli et al., 2011) with both strains being ST 62. Similarly, the analysed N-terminal region (aa 1–326) of strain 1680/94 (serotype 4) was 100% identical with the Norwegian isolate TIGR4 (Tettelin et al., 2001). Both strains were ST type 205 and were immunoblot positive with mAbs 143,F-2 and 149,B-3. The epitopes for the two mAbs have been shown by flow cytometry to be surface exposed on strain 7/87 (Kolberg et al., 2001) which after genome sequencing has been renamed to TIGR4.
Antibody responses against PspA analysed by immunoblotting
The sera from each of the patients were analysed against whole cell bacterial cell lysates of their own infecting strain (Fig. 2). The problem of identifying the positions, that is, mobility in gel, of the highly variable PspA proteins on the immunoblots was solved by the use of our panel of PspA specific mAbs. The similarities between the two serotype 9V strains as given above were strengthened by the same immunoblot reaction pattern against mAbs 143,B-3 and 160,F-4. The same was found for the three serotype 7F isolates that also were recognized by mAb 143,B-3. The identifications were strengthened by several bands by the mAbs with the same mobility as bands detected with the patients' sera. All strains revealed bands in the 100 kDa region, except for the 9V strain 1674/94 which showed mol wt < 50 kDa (Fig. 2c). Somewhat surprisingly, the other 9V strain (1673/94) of the same PspA and ST (162) showed upon immunoblotting PspA bands in the high mol wt region.
Antibody responses to PspA were not detected by immunoblotting in the sera from two patients with uncomplicated pneumonia (Fig. 2a). The sera from three patients, one with meningitis, showed similar levels of PspA antibodies both in acute and convalescent-phase sera, whereas the sera from patient 1673 revealed a slight increase (1.7) in the index between convalescent and acute phase sera. (Fig. 2b). A normal serum pool from seven healthy adults who had not been vaccinated and without known pneumococcal disease was tested against lysates of the strains 1675/94 (family 1, clade 2) and the TIGR4 like strain 1680/94 (family 2, clade 3). This serum pool showed immunoreactive bands similar to that of the acute phase sera from the patients from whom these strains were isolated (Fig. 2b and c), indicating that antibodies to PspA were present at the time of admission to the hospital. An index between convalescent and acute phase sera lower than 2 was therefore regarded as negative. The sera from four patients revealed an increase in the staining intensity of the PspA bands for convalescent to acute phase sera ranging from 2.9 to 7.4 (Table 3).
|Patient ID||Serotype of isolated strain||C-1/A||C-2/A|
Antibody responses against PspA analysed by ELISA
The amino terminal parts of family 1, clade 2, (strain 1675/94) and family 2, clade 3, (TIGR4 identical to 1680/94) PspA were used as coating antigens, respectively. The acute phase sera from some patients revealed slightly higher and for others slightly lower values than the normal serum pool from healthy adults (Fig. 3). When the seven sera included in the normal serum pool were analysed individually against clade 3 PspA, great variations were found. The range was 25–133 U mL−1, and the median was 94. Paired sera from patients with increases in PspA levels lower than 2 were judged as nonresponders. Most of the strains isolated in this study belonged to PspA clade 3 and sera from patients infected with these strains showed the strongest antibody responses against this antigen (Fig. 3b). Four of the 10 patients were found to have raised antibodies against clade 3 PspA. Three of these paired sera revealed also positive staining indexes by immunoblotting (Table 3). The forth ELISA positive serum against PspA/TIGR4, from patient 1680, infected with a strain similar to TIGR4, was PspA negative by immunoblotting. Analyses of the 10 paired sera against the other coating antigen (clade 2 PspA/1675/94) showed only a strong increase in antibody levels in the convalescent sera from patient 1675 (infected with a PspA clade 2 strain), whereas the paired sera from patient 1674 revealed low antibody levels compared with the normal serum pool (Fig 3a). The sera from these two patients showed positive response indexes by immunoblotting (Table 3). The PspA antibody level in the convalescent serum from patient 1681 was slightly more than twice of the normal serum pool, but this patient was regarded as nonresponder by this technique because no increase was seen from the acute phase serum. Here, it should be noted that this patient with an uncomplicated pneumonia was admitted to the hospital on the same day as she felt ill (Table 1). These baseline levels could be interpreted as antibodies from previous exposures to S. pneumoniae. No PspA-specific antibodies in the sera from this patient (1681) were found by immunoblotting (Fig. 2a). Patient 1672 did not respond to the PspA antigen as measured by immoblotting and the same was found against the two recombinant proteins used as antigens in ELISA. Increased levels of antibodies against PspA were detected by immunoblotting in the convalescent sera from patient 1678, but no response was found by the ELISA measurements. Here it should be noted that full length PspA in the patient's own strain was used as antigen by the first mentioned method, whereas a recombinant fragment from another strain was used as coating antigen in ELISA. However, PspA from both strains were family 2, clade 3 proteins.
Capsular-specific antibodies in acute and convalescent sera
The results for the measurements of capsular-specific serum IgG antibodies by the WHO recommended ELISA method using the WHO reference standards for weight-based units are shown in Fig. 4a. The sera from each patient were only tested against the capsular polysaccharides of the corresponding infecting strain. Statistical analyses using paired t-test revealed no significant increase in anti-capsular IgG antibodies from acute to 1st convalescent serum on all 10 patients. However, one patient (1673) had extremely high antibody levels in the acute phase serum compared with all the other patients. When this patient was excluded from the data analyses, there was a significant increase in anticapsular antibodies levels from acute to first convalescent serum (P < 0.01).
The absolute antibody concentrations in the pooled normal serum from healthy individuals were higher for 11A, 12F and 19F than the other serotypes examined. Among the seven serotypes tested, the highest antibody concentrations in the first convalescent sera (> 12 μg mL−1) were found for serotypes 11A and 12F. Five of the patients listed in Table 1 had uncomplicated pneumonia. One of the patients in this group (1680, serotype 4 strain isolated) was the only one in this study that did not show a significant antibody response to serotype-specific polysaccharides. Three other patients in the group with uncomplicated pneumonia [1681 (serotype 1), 1682 (7F) and 1672 (11A)] showed substantial increases (6–11 times) of the IgG values from acute to convalescent sera. The fifth patient in the group with uncomplicated pneumonia [1673 (9V)] revealed a very high IgG serum concentration (> 15 μg mL−1) at the admission to the hospital followed by a decline in first and second convalescent serum samples. For the one patient diagnosed with moderate severe pneumonia (1679), there was an eight times increase in the values from acute to convalescent-phase sera, but the absolute level in the latter serum was relatively low (< 2 μg mL−1) compared with those from the other patients in this study. Four patients were diagnosed with severe pneumonia with complications; one of these patients had meningitis (1677) (19F) with a doubling of serotype-specific antibodies from acute to convalescent phase. The sera from patient 1678 revealed an eight times increase in the values from acute to convalescent-phase sera, but the absolute level in the latter sera were < 2 μg mL−1. The two other patients (1674 and 1675) with severe pneumonia showed IgG values in convalescent-phase sera in the range from 5 to 15 μg mL−1 (4–8 times increases).
Two of the three patients infected with 7F strains belonging to the same clone revealed relative weak IgG anti-capsular specific responses, whereas a third patient showed a strong responses (Fig. 4a).
Comparison of antibody responses to capsular polysaccharides and PspA
Both immunoblot and ELISA measurements revealed that two patients (1672 and 1681) were nonresponders to the PspA antigens, but they showed strong antibody responses to the serotype polysaccharides of the infecting strains. The sera from four patients which revealed immunoblot positive IgG anti-PspA responses (Table 3, Fig. 2c) contained also capsular type specific antibodies. Patient 1680 (infecting strain undistinguishable from TIGR4) had not raised antibodies to PspA as analysed by immunoblotting or capsular polysaccharides. However, antibodies to PspA were detected when a recombinant N-terminal fragment of PspA/TIGR4 was used coating antigen in ELISA (Fig. 3a).
Antibody responses measured by binding to the patient's own infecting strains
The IgG antibody levels in the acute- and convalescent-phase sera were measured by flow cytometry against the isolated infecting pneumococcus and the results are shown in Fig. 4b. By this method, we measure antibodies against the serotype-specific polysaccharides, but also against other surface exposed polysaccharide and protein antigens. The three 7F strains belonged to the same clone. Thus, the sera from these three patients were measured against one of the 7F strains (1678/94). Similarly, one of the two 9V strains was used for the assessment of antibodies in sera from patients with 9V isolates (1674/94). The antibody levels in the normal serum pool were found to be at the same level as those in the different acute phase sera. However, the reaction pattern for capsular type-specific antibodies in these paired sera (Fig. 4a) was similar to that found for antibody binding to surface exposed structures on both polysaccharides and proteins (Fig. 4b). For all patients, there was a substantial increase in antibody levels from acute to convalescent phase for all patients, except patient 1680 from whom the serotype 4 strain 1680/94 (same clone as TIGR4) was isolated. This patient was the only one which did not show any capsular type-specific antibody response (Fig. 4a). Patient 1673 showed a very high level of serotype-specific IgG antibodies in the acute phase serum, but lower values in early and late convalescent-phase sera in ELISA. In contrast, the flow cytometry–binding analyses showed a strong increase from acute to convalescent phase.
In this study, we have analysed isolates from a small number of adult, hospitalized patients with invasive pneumococcal disease in Norway in 1993/1994. The patients' IgG antibody responses to two important, but chemically different virulence factors (serogroup-specific polysaccharides and the PspA protein) were analysed. PspA proteins among different pneumococcal strains show antigenic diversity and are also variable with regard to molecular sizes (Waltman et al., 1990; McDaniel et al., 1994). Immunoblotting for measuring antibody responses is therefore often performed with available N-terminal PspA fragments from family 1 and 2, but this may be insufficient to cover all protein variants. It is therefore important to note that we measured the patients' antibody responses against whole cell bacterial lysates of their own isolates. The positions of the PspA proteins on the blots were identified by using a panel of mAbs made against Norwegian pneumococcal strains. The PspA antibody levels by immunoblotting for four of the patients increased ranging from 2.9 to 7.4 from acute to convalescent phase. Here, it should be noted that antibodies detected by the immunoblotting technique are limited to linear epitopes and would not identify conformational dependent epitopes. Analyses of the sera by ELISA revealed in addition increased antibody levels from acute to convalescent phase for a patient that was found as a nonresponder to PspA by immunoblotting. However, full-length PspA was antigen by immunoblotting, whereas N-terminal PspA fragments were used by ELISA. A serum pool from nonvaccinated, healthy adults with no history of known pneumococcal disease was found to contain PspA antibodies both to family 1 and 2 by immunoblotting. Nonsymptomatic pneumococcal carriage or undiagnosed less severe infections may explain the presence of these antibodies. PspA antibodies in healthy, adult individuals have also been reported by others (Virolainen et al., 2000; Linder et al., 2007; Simell et al., 2008).
Increased levels of antibodies to PspA were detected by immunoblotting and ELISA in convalescent sera in 5 of 10 adult patients. Our findings seem to be in contrast to the analyses in (Linder et al., 2007) who found that nearly all of 19 adult patients with pneumococcal bacteraemia had increased levels of IgG antibodies against PspA during convalescence. However, a direct comparison between these studies is difficult because nonresponders in our study were defined as increases by < 100% and in the other study < 25%.
In agreement with other studies, we found that PspA family 2 (8 of 10) dominated among isolates from adult patients with invasive pneumococcal disease (Hollingshead et al., 2006; Rolo et al., 2009; Qian et al., 2012). The PspA in all three 7F isolates were clade 3 and the two 9V strains were also of the same PspA type. However, several larger studies have shown that PspA clades are independent of capsular serotypes (Heeg et al., 2007; Rolo et al., 2009; Qian et al., 2012).
Although the PspA is a highly variable protein which is independent of capsular serotypes, three 7F isolates of the same ST (191) showed 100% identity in their N-terminal PspA nucleotide sequences. These strains belong to the same clone. ST type 191 is according to the MLST database, associated with serotype 7F. Serotype 7F isolates of PspA family 2 and ST 191 have been found to be associated with increased virulence (Sadowy et al., 2006). Identical N-terminal PspA sequences were found for the two 9V isolates. The reason for the variation in the mol wts for PspA is not known (Waltman et al., 1990). There might be differences outside the analysed PspA region (aa 1–508) of the two 9V strains because they showed different mobility patterns upon immunoblotting. They should most likely be regarded as subclones from the same mother clone. It thus seems to be of value to include mol wt analyses by immunoblotting as a marker in the studies of the PspA relatedness between different pneumococcal strains.
We found that two patients out of 10 had raised cross-reacting IgG anti-PspA antibodies in their convalescent phase sera against family 1 and 2 proteins as coating antigen in ELISA. A wider repertoire of antibodies against PspA epitopes could probably be expected because these adult patients might have been exposed to multiple pneumococcal infections. Our data with a limited number of patients are somewhat in contrast to the findings reported by (Hollingshead et al., 2006). By using an ELISA technique with rabbit antibodies made against one of three N-terminal recombinant fragments from different clades, they found that two-thirds of the clinical isolates from adults over 50 years of age reacted with more than one of the serum pools used. One explanation for these apparently different findings could be that we measured cross-reacting antibodies after natural infections whereas they used rabbit antibodies to show similarities between bacterial cell lysates. It might also be that the two recombinant proteins used in our study (amino acid sequence identity 28.1%) contain less cross-reacting epitopes than the PspA antigens that were used by Hollingshead et al. in their rabbit immunizations. This could also be the reason why Baril et al. found a very high degree of anti-PspA cross-reacting antibodies in convalescent sera from 14 adult patients with invasive disease in ELISA (Baril et al., 2004), in contrast to our findings.
It is generally accepted that capsular type-specific antibodies play an important role in protection against invasive pneumococcal disease and that a certain threshold of antibody concentration is needed for protection. By using the WHO reference ELISA, a threshold of 0.2 μg mL−1 against strain-specific polysaccharide antigen has been derived after immunogenicity and vaccine efficacy trials in infants and children with a seven valent conjugate pneumococcal vaccine (Westerink et al., 2012). The protective antibody levels of anti-capsular polysaccharide IgG for adults have not been established (Musher et al., 2011). It is likely that the situation in this age group is more complex as adults might have had several pneumococcal infections and therefore also raised antibodies to other important, possible protective proteins including PspA. Furthermore, multiple infections may result in antibodies directed against several types of highly variable proteins like PspA.
Two patients (1678 and 1679) infected with serotype 7F strains revealed antibody concentrations at this level in their acute phase sera, whereas the other eight patients showed much higher values. Thus, a serum level for serotype-specific IgG antibodies which is assumed to be protective in children does not seem to be valid in the adult population based on the results in our limited study. Another explanation might be that the acute serum does not reflect the antibody level when the infection was initiated, but rather shows the escalating antibody response induced by the infection. Furthermore, only one patient (strain 1680/94 same clone as TIGR4) did not show a significant increase from acute to convalescent-phase sera of capsular type specific antibodies. The absolute convalescent-phase antibody levels for the other patients varied from relative low to high, but one should have expected these levels to be adequate for protection. However, antibody concentrations measured by ELISA do not always be equal to their functional properties. A prerequisite for the antibodies to exert effector functions such as opsonophagocytosis are their capacity to bind surface structures on live bacteria. Reference strains are often used in binding assays. In this study, we used the patient's own infecting strains for flow cytometry assessment of IgG antibody binding to surface exposed epitopes, that is,carbohydrate as well as protein structures. Like ELISA with serotype-specific polysaccharides as coating antigens, the sera from patient 1680 was the only one which did not show a substantial increase in antibody binding levels by the flow cytometry measurements. This raises the question of the protective capacities of the infection induced antibodies in this group of patients including both young and old adults. It is likely that the immune response (antibody production) induced at the initial stage of infection, progresses too slowly to deal with the acute pneumococcal infection, similar to what is observed in meningococcal sepsis (Erlich & Congeni, 2012). The outcome of the pneumococcal infections will probably also depend on other host-specific immunological factors (Calbo & Garau, 2011).
Three of the four patients who by immunoblotting were found to have increased levels of IgG anti-PspA antibodies in their convalescent sera also revealed strong responses to serotype-specific polysaccharides. There are few reports on the measurements of immune responses to the two above-mentioned structures. In adult patients with invasive disease, low levels of antibodies against capsular polysaccharides, rather than antibodies directed against PspA, may predispose for disease (Zysk et al., 2003). They used a 23 valent pneumococcal polysaccharide vaccine as coating antigen in their ELISA, but absorption with pneumococcal cell wall and 22F polysaccharide were not performed. For the determination of PspA antibodies, they used ELISA with recombinant protein fragments and found that PspA antibody titres did not change significantly after invasive pneumococcal infection, whereas in our studies, we found for some patients substantial increases in IgG binding to immunoblots with the patient's own isolates as antigen.
In summary, by immunoblotting and ELISA measurements, an IgG antibody response to PspA was found for 5 of 10 adult patients with invasive pneumococcal disease. This might be an underestimated number because full-length PspA was not available for the ELISA measurements. A prerequisite for antibodies to exert functional properties is their binding to the invading bacteria. Binding of IgG antibodies measured by flow cytometry to the patient's own isolate revealed a close relation to the patterns seen by ELISA measurements of capsular type-specific antibodies. This could indicate that the majority of antibodies detected in the flow cytometry assay represent serotype-specific antibodies and that the polysaccharides are more expressed on live pneumococci than PspA. However, the in vitro cultivation of the pneumococcal strains might have influenced the surface expression of polysaccharides, that is, the opacity. It is known that there is a great variation between strains to form opaque variants (Serrano et al., 2006). Increased expression of polysaccharides for some strains would probably not result in the close relationship in antibody levels by the two methods. Therefore, the PspA antigens might be less important for the development of protective antibodies than capsular polysaccharides. Low levels of serotype-specific antibodies in the patients acute phase sera, may explain why they were not protected.
We thank Dr Sven Hammerschmidt, Ernst Moritz Arndt University of Greifswald, Greifswald, Germany, for the gift of the pneumococcal strain Spn 51 and its PspA knock-out mutant. Karin Bolstad, Elisabeth Fritzsønn, Anne-Cathrine Kristoffersen and Jan Oksnes provided excellent technical assistance.
- 2004) Characterization of antibodies to PspA and PsaA in adults over 50 years of age with invasive pneumococcal disease. Vaccine 23: 789–793. , , , , , & (
- 2011) Factors affecting the development of systemic inflammatory response syndrome in pneumococcal infections. Curr Opin Infect Dis 24: 241–247. & (
- 2011) Complete genome sequence of a serotype 11A, ST62 Streptococcus pneumoniae invasive isolate. BMC Microbiol 11: 25. , , , , , , , , , , , & (
- 2010) Structure and dynamics of the pan-genome of Streptococcus pneumoniae and closely related species. Genome Biol 11: R107. , , et al. (
- 1998) A multilocus sequence typing scheme for Streptococcus pneumoniae: identification of clones associated with serious invasive disease. Microbiology 144(Pt 11): 3049–3060. & (
- 2012) Importance of circulating antibodies in protection against meningococcal disease. Hum Vaccin Immunother 8: 1029–1035. & (
- 2011) Continued control of pneumococcal disease in the UK - the impact of vaccination. J Med Microbiol 60: 1–8. , , & (
- 2011) Establishment of a new human pneumococcal standard reference serum, 007sp. Clin Vaccine Immunol 18: 1728–1736. , , et al. (
- 1999) Identification of pneumococcal surface protein A as a lactoferrin-binding protein of Streptococcus pneumoniae. Infect Immun 67: 1683–1687. , , & (
- 2007) Genetic diversity of pneumococcal surface protein A of Streptococcus pneumoniae meningitis in German children. Vaccine 25: 1030–1035. , , , & (
- 2000) Diversity of PspA: mosaic genes and evidence for past recombination in Streptococcus pneumoniae. Infect Immun 68: 5889–5900. , & (
- 2006) Pneumococcal surface protein A (PspA) family distribution among clinical isolates from adults over 50 years of age collected in seven countries. J Med Microbiol 55: 215–221. , , , , & (
- 2011) Protection of pneumococcal infection by maternal intranasal immunization with pneumococcal surface protein A. Adv Otorhinolaryngol 72: 121–125. , , , & (
- 2000) Production and characterization of the functional fragment of pneumococcal surface protein A. Arch Biochem Biophys 373: 116–125. , , , , & (
- 1985) Phagocytosis and killing of common bacterial pathogens of the lung by human alveolar macrophages. J Infect Dis 152: 4–13. , , , & (
- 2000) Streptococcus pneumoniae heat shock protein 70 does not induce human antibody responses during infection. FEMS Immunol Med Microbiol 29: 289–294. , , , , , & (
- 2001) Epitope analyses of pneumococcal surface protein A: a combination of two monoclonal antibodies detects 94% of clinical isolates. FEMS Immunol Med Microbiol 31: 175–180. , , & (
- 2003) Epitope mapping of pneumococcal surface protein A of strain Rx1 using monoclonal antibodies and molecular structure modelling. FEMS Immunol Med Microbiol 39: 265–273. , , , & (
- 2006) Streptococcus pneumoniae enolase is important for plasminogen binding despite low abundance of enolase protein on the bacterial cell surface. Microbiology 152: 1307–1317. , , , , , , & (
- 2011) Polyreactivity of monoclonal antibodies made against human erythrocyte membranes with various pathogenic bacteria. Hybridoma (Larchmt) 30: 1–9. , , & (
- 2007) Human antibody response towards the pneumococcal surface proteins PspA and PspC during invasive pneumococcal infection. Vaccine 25: 341–345. , , , & (
- 1994) Localization of protection-eliciting epitopes on PspA of Streptococcus pneumoniae between amino acid residues 192 and 260. Microb Pathog 17: 323–337. , , & (
- 2012) Pneumococcal surface protein a inhibits complement deposition on the pneumococcal surface by competing with the binding of C-reactive protein to cell-surface phosphocholine. J Immunol 189: 5327–5335. , , , , , , , & (
- 2011) Antibody persistence ten years after first and second doses of 23-valent pneumococcal polysaccharide vaccine, and immunogenicity and safety of second and third doses in older adults. Hum Vaccin 7: 919–928. , , , , , , , & (
- 2010) Combination of pneumococcal surface protein A (PspA) with whole cell pertussis vaccine increases protection against pneumococcal challenge in mice. PLoS ONE 5: e10863. , , et al. (
- 2012) Diversity of pneumococcal surface protein A (PspA) and relation to sequence typing in Streptococcus pneumoniae causing invasive disease in Chinese children. Eur J Clin Microbiol Infect Dis 31: 217–223. , , et al. (
- 2004) Assignment of weight-based antibody units for 13 serotypes to a human antipneumococcal standard reference serum, lot 89-S(f). Clin Diagn Lab Immunol 11: 1064–1069. , , , , , & (
- 2003) Regions of PspA/EF3296 best able to elicit protection against Streptococcus pneumoniae in a murine infection model. Infect Immun 71: 1033–1041. , , & (
- 2009) Diversity of pneumococcal surface protein A (PspA) among prevalent clones in Spain. BMC Microbiol 9: 80. , , , & (
- 2006) Multilocus sequence types, serotypes, and variants of the surface antigen PspA in Streptococcus pneumoniae isolates from meningitis patients in Poland. Clin Vaccine Immunol 13: 139–144. , , , & (
- 2011) Changes in capsular serotype alter the surface exposure of pneumococcal adhesins and impact virulence. PLoS ONE 6: e26587. , , , , , & (
- 2006) Heterogeneity of pneumococcal phase variants in invasive human infections. BMC Microbiol 6: 67. , & (
- 2008) Effects of ageing and gender on naturally acquired antibodies to pneumococcal capsular polysaccharides and virulence-associated proteins. Clin Vaccine Immunol 15: 1391–1397. , , , & (
- 2001) Complete genome sequence of a virulent isolate of Streptococcus pneumoniae. Science 293: 498–506. , , et al. (
- 2011) Longitudinal analysis of pneumococcal antibodies during community-acquired pneumonia reveals a much higher involvement of Streptococcus pneumoniae than estimated by conventional methods alone. Clin Vaccine Immunol 18: 796–801. , , , , , , & (
- 2000) Human antibodies to pneumococcal surface protein A in health and disease. Pediatr Infect Dis J 19: 134–138. , , , , & (
- 1990) Variation in the molecular weight of PspA (pneumococcal surface protein A) among Streptococcus pneumoniae. Microb Pathog 8: 61–69. , , & (
- 2012) Immune Responses to pneumococcal vaccines in children and adults: rationale for age-specific vaccination. Aging Dis 3: 51–67. , & (
- 1994) Novel surface attachment mechanism of the Streptococcus pneumoniae protein PspA. J Bacteriol 176: 2976–2985. & (
- 2003) Immune response to capsular polysaccharide and surface proteins of Streptococcus pneumoniae in patients with invasive pneumococcal disease. J Infect Dis 187: 330–333. , , , , , & (