Pneumococcal surface protein A (PspA) is effective at eliciting T cell-mediated responses during invasive pneumococcal disease in adults

Abstract

Humoral immune response is essential for protection against invasive pneumococcal disease and this property is the basis of the polysaccharide-based anti-pneumococcal vaccines. Pneumococcal surface protein A (PspA), a cell-wall-associated surface protein, is a promising component for the next generation of pneumococcal vaccines. This PspA antigen has been shown to stimulate an antibody-based immunity. In the present study, we evaluated the capacity of PspA to stimulate CD4⁺ T cells which are needed for the correct development of a B cell based immune response in humans. Cellular immunity to PspA was evaluated by whole-blood culture with different pneumococcal antigens, followed by flow cytometric detection of activated CD4⁺CD25⁺ T cells. T cell-mediated immune responses to recombinant PspA proteins were assessed in acute-phase and convalescent blood from adults with invasive pneumococcal disease and in blood from healthy subjects. All cases had detectable antibodies against PspA on admission. We found that invasive pneumococcal disease induced transient T cell depletion but adaptive immune responses strengthened markedly during convalescence. The increased production of both interleukin (IL)-10 and interferon (IFN)-γ during convalescence suggests that these cytokines may be involved in modulating antibody-based immunity to pneumococcal disease. We demonstrated that PspA is efficient at eliciting T cell immune responses and antibodies to PspA. This study broadens the applicability of recombinant PspA as potent pneumococcal antigen for vaccination against S. pneumoniae.

Introduction

Streptococcus pneumoniae, an encapsulated bacterium, remains an important cause of morbidity and mortality in humans. The 23-valent capsular polysaccharide (PS) fails to protect children below 2 years of age probably because their immune system is immature and, as a consequence, incapable of responding correctly to carbohydrate antigen. A 7-valent protein–polysaccharide conjugate vaccine has been developed and offers protection against pneumococcal disease caused by vaccine serotypes in young children [1]. This vaccine exploits T cell-dependent properties such as immunogenicity in early infancy, stimulation of high levels of IgG isotype antibodies and enhanced immunological memory response [2]. In adults, the plain PS offers a clinical protection rate about 60% against invasive pneumococcal disease (IPD) and confers serotype-specific protection [3]. However, the challenge of achieving full serotype coverage with PS-conjugate has stimulated attempts to develop protein-based pneumococcal vaccines. These vaccines might provide better protection against IPD, as these proteins are shared by all S. pneumoniae. Such vaccines could be also highly immunogenic and would probably induce antigen-specific memory.

One of the most intensively studied candidate proteins is the pneumococcal surface protein A (PspA), a cell-wall-associated surface protein. PspA is known to play a major role in the pneumococcal virulence; it bonds human lactoferrin and interferes with complement deposition on the bacterial surface [4,5]. PspA is relatively variable at the DNA and protein sequence levels; two major alleles have been identified (family 1 and family 2) and are present on almost all the pneumococcal isolates from patients with IPD [6]. Each PspA family is subdivided into several clades: PspA family 1 contains two clades (1 and 2) and PspA family 2 three clades (3, 4 and 5) [7]. To accommodate this variability, it has been postulated that a combination of three PspA antigens from family 1 (clade 2) and family 2 (clade 3 and clade 4) would elicit protection against the vast majority of S. pneumoniae. It has been hypothesized that the polymorphism of PspA is attributable to immunological selection and confirmed indirectly that this antigen is readily accessible to antibodies on the surface of the intact pneumococci [8]. This is also supported by the observation that, in humans, vaccination with a recombinant PspA derived from strain Rx1 (family 1, clade 2) elicits broadly cross-reactive antibody responses to a wide range of PspA proteins [9]. The pneumococcal surface antigen protein A (PsaA) is structurally conserved [10] and plays a role in adherence to host mucosae [11]. It is another potential vaccine component but, until now, it has not been used as vaccine antigen in humans. Our study focused on these two cell-wall proteins, but other vaccine candidates such as pneumolysin have been identified [12].

Immunity to pneumococcal diseases is assumed to depend mainly on the humoral arm of the immune system. The antibody response to both PspA and PsaA has been studied in children [13–15] and in adults over 50 years of age with pneumococcal disease [16]. During the course of the invasive disease, antibodies to PspA peak at the convalescent phase. CD4⁺ T cell help, through T cell receptor (TCR) recognition, is required for the IgG antibody response to PspA [17,18].

To date, the cellular-based protective immunity to pneumococcal disease remains to be better clarified. In this study, we evaluated the potential of PspA and PsaA antigens to elicit a T cell-mediated immune response in adults with IPD. The cellular immune responses to pneumococcal antigens were compared with those in healthy subjects. We investigated the cellular-based immunity to S. pneumoniae surface proteins by whole-blood culture with recombinant forms of PspA and PsaA, followed by flow cytometric detection of activated CD4⁺ CD25⁺ T cells.

Methods

Study population

The patients were admitted to Croix-Rousse Hospital in Lyon or Bellevue Hospital in Saint Etienne (France) between October 2001 and September 2002. The appropriate Ethics Committee approved the study before any subjects were enrolled, and the study was conducted in accordance with the Declaration of Helsinki. All the subjects gave their written informed consent before entering the study. For this study, cases were defined as patients ≥ 18 years with a invasive pneumococcal disease. Cases were subclassified as culture-confirmed or probable cases. Culture-confirmed cases were defined as patients with positive culture of S. pneumoniae isolated from blood or another normally sterile fluid. Probable cases were defined as follows: (1) clinical signs, chest radiography, clinical outcome compatible with pneumococcal disease; (2) positive urinary pneumococcal soluble antigen testing (Binax Now® Strep. pneumoniae urinary antigen test, Binax Corp., Portland, ME, USA); and (3) antibody concentration to one of the three recombinant PspA antigens risen at least fourfold between the acute and convalescent phases. Immunocompromised patients, including those with HIV infection, were excluded. Three healthy control subjects consulting the out-patient clinics at Bellevue or Croix Rousse Hospitals were closely age-matched with IPD cases.

Microbiological analysis

For each pneumococcal isolate, we determined the PspA family by polymerase chain reaction (PCR), as described previously [6,7].

Blood sampling

Peripheral blood specimens were obtained from the cases on the day on admission (acute samples) and 21 ± 7 days later (convalescent samples). Single blood samples were obtained from the control subjects. Whole blood specimens were processed within 4 h after sampling. The C-reactive protein concentrations from acute and control samples were determined. Culture supernatants were stored at − 80 ° C. Sera were stored at − 80 ° C and shipped on dry ice to University of Alabama at Birmingham (UAB) in the United States for further analysis.

Description of the studied antigens

Table 1 describes the antigens used for this study. Tetanus toxoid (TT, 5 µg/ml; Sanofi Pasteur, Marcy l’Etoile, France) and phytohaemagglutinin (PHA, 10 µg/ml; Sigma, St Quentin Fallavier, France) were used as positive controls for specific and non-specific T cell activation, respectively. RPMI-1640 culture medium (Sigma) was used as negative control.

Table 1.

Description of the antigens used in the study.

Antigen	Abbreviation	Characteristics	Origin
Positive controls
Phytohaemagglutinin	PHA	Non-specific positive control	Sigma, France
Tetanus toxoid	TT	Specific positive control	Sanofi Pasteur
Crude pneumococcal antigens
Strain SJ04	SJ	Capsular serotype 9, family 2 clade 3 PspA	Laboratory of Bacteriology, Hôpital de la Croix Rousse
Strain Rx1	Rx1	Laboratory non-encapsuled strain, family 1, clade 2 PspA	UAB*
Strain JY2141	JY	Laboratory non-encapsuled strain, expressing only the first 115 amino-acids of Rx1 PspA	UAB
Native PspA
Native full length PspA from Rx1	FL-Rx1	Culture supernatant of Rx1 strain	UAB
Full length PspA from JY2141	FL-JY	Negative control for FL-Rx1, culture supernatant of JY strain	UAB
Recombinant antigens
PspA-Rx1MI	PspA1	Truncated PspA family 1, clade 2 (AA: 1–364)	Sanofi Pasteur
PspA-EF3296LL	PspA2	Truncated PspA family 2, clade 3 (AA: 1–478)	Sanofi Pasteur
PspA-EF5668	PspA3	Truncated PspA family 2, clade 4 (AA: 1–369)	Sanofi Pasteur
PsaA	PsaA	Conserved PsaA	Sanofi Pasteur

^*UAB = University of Alabama at Birmingham.

Two pneumococcal crude soluble antigens were prepared from the Rx1 strain (a non-encapsulated laboratory strain belonging to the PspA family 1, clade 2) and from JY2141 (JY), a strain derived from Rx1 which expresses a non-efficient truncated PspA of 15 kDa. Both strains were provided by UAB. A third crude soluble antigen was prepared from strain SJ04 (SJ), a blood culture isolate from an adult hospitalized in Lyon (capsular polysaccharide serogroup 9, PspA family 2, clade 3). Briefly, isolates were cultured in sterile chemically defined medium (CDM; JHR Biosciences, Denver, CO, USA) with 10% fetal calf serum (Sigma) and 0·73 g/l l-cysteine hydrochloride monohydrate (Sigma) at 37°C for 10 h (before autolysis occurred). Cultures were centrifuged for 15 min at 1500 g, the supernatants were collected, and the pellets were washed three times in sterile phosphate-buffered saline (PBS). The bacteria were then sonicated in PBS for 15 min and centrifuged for 15 min at 2500 g, and the supernatants were filtered through 0·22-µm membranes in sterile conditions. The total protein concentrations were determined using the colorimetric detergent compatible (DC) protein assay (Bio-Rad, Marnes-la-Coquette, France).

The antigens were a truncated PspA (amino acids 1–364) from strain Rx1 (family 1, clade 2: PspA1), a truncated PspA (amino acids 1–478) from strain EF3296LL (family 2, clade 3: PspA2), a truncated PspA (amino acids 1–369) from strain EF5668 (family 2, clade 4: PspA3) and a mature non-lipidated PsaA from American Type Culture Collection (ATCC) strain 6326 (all from Sanofi Pasteur). Each truncated protein comprised the α-helical region and a portion of the proline-rich region of the respective full-length PspA, but lacked the repetitive choline-binding site.

Two full-length PspA species (FL-Rx1 (1–589 amino acids) and FL-JY) were prepared at the UAB from strains Rx1 and JY2141, respectively. FL-JY (PspA-) was used as a negative control for FL-Rx1 (PspA+). Briefly, supernatants containing FL-Rx1 and FL-JY proteins were prepared at UAB. They were analysed by dot-blot on nitrocellulose membranes with a monoclonal antibody (XIR278, produced at UAB) to ensure they were, respectively, PspA⁺ (Rx1 strain) and PspA^– (JY strain). Lowry's protein assay was used to determine the PspA concentrations. Samples were sterilized by irradiation (3000 rads) and then frozen immediately before shipping on dry ice to Lyon, France.

Peripheral blood leucocytes and antibody analysis

Polymorphonuclear neutrophils, CD3⁺ T cells, CD3⁺ CD4⁺ T cells, CD3⁺ CD8⁺ T cells, CD19⁺ B cells and CD3^– CD56⁺ (natural killer) cells were counted in acute blood and control blood by using an automatic cell counter (SYSMEX SP-100, Roche Diagnostics, Meylan, France) and a fluorescence activated cell sorter (FacScan) flow cytometer (Becton Dickinson, Pont de Claix, France).

Antibody levels to recombinant PspA1, PspA2, PspA3 and PsaA were measured by enzyme-linked immunosorbent assay (ELISA) as described elsewhere [16]. The concentration of antibody reactive with Rx1 PspA was determined using a serum pool with known antibody titres provided by Sanofi Pasteur [9].

Whole blood analysis and lymphocyte activation

Whole blood (100 µl) was incubated in triplicate in sterile 45 × 8·8 mm propylene tubes (Micronic Systems, Lelystad, the Netherlands) with final antigen concentrations of 5 µg/ml except for PHA, which was tested at the equivalent of 10 µg/ml. On day 1, 800 µl of RPMI-1640 was added to each tube and the samples were incubated for 7 days at 37°C in 5% CO₂ in air. On day 7, culture supernatants were collected and stored at −80°C until cytokine assays.

Lymphocyte activations were studied as described previously [19]. Briefly, erythrocytes were lysed with NH₄Cl (155 mM), KHCO₃ (10 mM) and ethylenediamine tetraacetic acid (EDTA) (0·1 mM), and leucocytes were recovered by centrifugation. Leukocytes were then triply stained with fluorescein isothiocyanate-conjugated anti-CD25 (Dako, Trappes, France), phycoerythrin (PE)-conjugated anti-CD4 (Dako) and cyanin 5-PE-conjugated anti-CD3 (Immunotech, Marseille, France) for 15 min at 4°C in the dark. The stained cells were washed and resuspended in PBS supplemented with 0·1% bovine serum albumin (BSA) and EDTA (5 mM), then analysed in a flow cytometer to detect CD25-positive T cells. Irrelevant fluorescein isothiocyanate (FITC)-conjugated IgG1 (Dako) was used as a negative control to confirm specific CD25 detection.

Cytokine assays

Interferon (IFN)-γ, a potent proinflammatory cytokine, interleukin (IL)-10 (produced mainly by Th2 type lymphocytes) and IL-4 (a major factor in B cell activation and differentiation) were measured in 7-day culture supernatants with commercial ELISA kits (R&D Systems, Lille, France) with detection limits of 10 pg/ml (IFN-γ and IL-4) and 40 pg/ml (IL-10).

Statistical analysis

Antibody concentrations were log-transformed to approximate a normal distribution. The results are reported as the anti-logarithm of the log₁₀ geometric mean antibody concentrations (GMC), together with the mean and the 95% confidence intervals for values for each group. Student's paired t-test was used to compare antibody concentrations between acute and convalescent sera from cases. Wilcoxon's test (one-sided) was used to compare flow cytometry results between acute and convalescent blood from cases, and the Mann–Whitney test (two-sided) was used to compare acute and convalescent blood from cases with blood from healthy controls. Owing to a large number of missing results and samples with values below the detection limit, a permutation test (two-sided) was used to compare cytokine concentrations in acute and convalescent blood culture supernatants, and to compare culture supernatants from acute phases in cases and from controls. sas® software (version 8 for Windows, SAS Institute Inc., Cary, NC, USA) or StatXact® software (version 5 for Windows, Cytel Software Corp., Cambridge, MA, USA) was used as appropriate, and figures were drawn with Sigmaplot® software (version 8·0 for Windows, SPSS Science, Chicago, IL, USA).

Results

Clinical and microbiological findings

Sixteen cases were included in this study: 12 cases were culture-confirmed and four were probable cases. There were 11 men and five women, with a median age of 65·5 years [25–75 interquartiles (IQ): 60–71]. The median age of the 44 control subjects (33 men, 11 women) was 65·5 years (IQ: 57–72·5). The median C-reactive protein concentration was 174·5 mg/l (IQ: 121–338) in acute samples and 7·0 mg/l (IQ: 3·6–12·5) in controls. All patients recovered on appropriate anti-microbial chemotherapy. Ten S. pneumoniae isolates were recovered by blood culture, one by cerebrospinal fluid culture and one by pleural fluid culture. Eight isolates could be subcultured and shipped to UAB for PspA family classification: three expressed a family 1 PspA and the others expressed family 2 PspAs. Eight isolates were examined for their capsular serogroup: two were serogroup 1, and one each was serogroup 3, 4, 8 and 9; one strain was non-typable. Nine isolates were susceptible to penicillin, two were intermediate-resistant [minimum inhibitory concentration (MIC): 0·12–1 mg/l] and one was resistant (MIC: > 1 mg/l).

Cellular and cytokine responses

As shown in Table 2, peripheral blood leucocyte and polymorphonuclear neutrophil counts were higher and mean lymphocyte counts were lower in acute samples than in the controls. CD3⁺ CD8⁺ T cell depletion was marked in acute samples and CD3⁺ CD4⁺ T cell depletion was moderate. All the controls had normal results. All the cell phenotypes studied differed significantly between the cases and controls, except for CD19⁺ B cells (normal in both groups). Figure 1 shows the percentage of cultured CD4⁺ CD25⁺ T cells reacting with each antigen (see definitions in Table 1), as detected by flow cytometry. CD4⁺ CD25⁺ T cells always represented < 2% of total peripheral blood mononuclear cells (PBMC) after 7 days of culture with RPMI-1640 or FL-JY. Percentages of stimulated CD4⁺ CD25⁺ T cells were lower in PHA- and TT-stimulated acute samples than in convalescent and control samples; there were no significant differences between convalescent and control blood. The percentage of CD4⁺ CD25⁺ T cells stimulated with all pneumococcal antigens, except for FL-JY (PspA^– control), was always higher in convalescent blood than in acute blood; this difference was always statistically significant except for PspA2, which was borderline significant. This was also the case for the TT and PHA antigens. On one hand, cellular responses to Rx1, FL-Rx1 and PsaA antigens were significantly stronger in convalescent samples than in control samples. On the other hand, responses triggered by recombinant PspA antigens were stronger in controls than in convalescent blood; this was statistically significant in the case of PspA1. As shown in Fig. 2, IL-10 concentrations in cultures stimulated with Rx1 antigen and FL-Rx1 were higher than corresponding IFN-γ concentrations. IL-4 was never detected.

Table 2.

Peripheral blood leucocyte, polymorphonuclear neutrophil and lymphocyte mean counts in acute blood samples from cases with invasive pneumococcal disease (IPD) and in control subjects.

Mean (range) (µl)	Subjects with IPD (n = 16) acute blood	Control group (n = 44)	Mann–Whitney testa (two-sided)
Peripheral blood leucocytes	11242 (3410–25230)	6453 (3010–17000)	**
Polymorphonuclear neutrophils	9672 (2540–21150)	4007 (1210–13260)	**
Lymphocytes	1112 (330–1850)	1832 (600–3620)	**
CD3+ T cells	735 (228–1494)	1381 (323–2716)	**
CD3+ CD4+ T cells	560 (172–1062)	915 (220–1738)	**
CD3+ CD8+ T cells	224 (40–481)	438 (92–1048)	**
CD19+ B cells	148 (30–388)	168 (12–561)
CD3– CD56+ cells	101 (40–233)	198 (3–814)	*

^aP < 0·05 and P< 0·01.

^*P < 0·05

^**P < 0·01.

^bNormal values for adults (µl) (used by G. Cozon at the Laboratory of Immunology in Lyon): peripheral blood leucocytes 4000–10000; polymorphonuclear neutrophils 200–7500; lymphocytes 1000–4000; CD3⁺ T cells 790–2460; CD3⁺ CD4⁺T cells 455–1575; CD3⁺ CD8+ T cells 240–1050; CD19⁺ B cells 60–540; CD3^– CD56⁺ cells 50–700.

Fig. 1.

(a,b) Percentages of cultured CD4⁺ CD25⁺ T cells reacting with the different antigens as detected by flow cytometry. Box plots graphs: the boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box furthest from zero indicates the 75th percentile. Whiskers above and below the box indicate the 95th and 5th percentiles. The outlining dots represent the values out of the 95th and 5th percentiles. Horizontal bar is defined at 2% of CD25⁺ CD4⁺ T cells in all the graphs. Statistical results: *P < 0·05 and **P < 0·01.

Fig. 2.

Concentration of cytokines [interleukin (IL)-10 and interferon (IFN)-γ] from the supernatants using Rx1 and FL-Rx1 antigens: acute blood from cases (closed circle), convalescent blood from cases (open triangle), blood from controls (closed square).The dots at the bottom of each graph are below the detection limit. Statistical results: *P < 0·05 and **P < 0·01.

Humoral responses

As shown in Table 3 and Fig. 3, all cases had detectable antibodies against both PspA and PsaA on admission. The geometric mean antibody concentrations to PspA and PsaA in the convalescent sera of cases were, respectively, 6·6 and 5·2 times higher than in the corresponding acute sera. Antibody concentrations in these acute samples were similar to those already reported in healthy subjects in our previous study [16]. Pneumococcal strains from eight patients were sent to UAB; three were PspA family 1 (one clade 1 and two clade 2) and five were PspA family 2 (two clade 3, one clade 4 and two with non-determined clade). Seven paired serum samples from these eight patients were available and changes in antibody concentrations were more pronounced against the recombinant PspA antigen corresponding to the PspA strain isolated from the same patient (at least a fourfold increase between acute and convalescent sera, except for one patient presenting only a twofold increase). However, for these eight patients, no difference was seen in cellular or cytokines responses according to the PspA family of the pneumococcal strain and the corresponding recombinant antigen (data not shown).

Table 3.

Antibody concentrations (µg/ml) against three recombinant PspA proteins and against a recombinant pneumococcal surface protein A (PspA) in subjects with invasive pneumococcal disease (IPD).

	Cases (n = 15)c
GMCa (95% CI)b (µg/ml)	Acute sera	Convalescent sera	Ratio convalescent/ acute samples
Family 1, clade 2 PspA	3·2 (1·8–5·8)	21·2 (11·1–40·6)	6·6**d
Family 2, clade 3 PspA	4·2 (2·1–8·3)	15·7 (5·7–43·2)	3·7*
Family 2, clade 4 PspA	5·3 (2·7–10·1)	12·1 (5·2–28·0)	2·3**
PsaA	11·3 (4·8–27·0)	58·6 (24·7–138·8)	5·2**

^aGMC: geometric mean antibody concentration.

^bCI: confidence intervals.

^cSera missing for one patient.

^dPaired t-test:

^*P < 0·05 and

^**P < 0·01.

Fig. 3.

Antibody concentrations against the recombinant pneumococcal surface protein A (PspA) and PsaA antigens in acute and convalescent sera from the patients with pneumococcal disease. Box plot graphs: the boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box furthest from zero indicates the 75th percentile. Whiskers above and below the box indicate the 95th and 5th percentiles. The outlining dots represent the values out of the 95th and 5th percentiles. Statistical results: *P < 0·05 and **P < 0·01.

Discussion

In a PspA-based vaccine context, it is important to study scale and orientation of T cell responses because an efficient cooperation between B and T cells is necessary for induction of antibody and memory responses against this protein antigen. However, very few studies have examined T cell responses to S. pneumoniae in humans. Arva et al. [20,21] showed that whole killed S. pneumoniae cells induced strong in vitro proliferation and cytokine production (IFN-γ and IL-10) by PBMC from healthy donors. However, little is known about specific CD4⁺ T cell responses and the profile of cytokine response to pneumococcal cell-wall proteins during the course of IPD in adults. In this study, we focused on acquired immunity to non-capsular pneumococcal antigens and we characterized cellular activation, cytokine production and antibody responses to crude pneumococcal antigens and to two cell-wall proteins (PspA and PsaA).

We first confirmed that during the acute phase of pneumococcal infection there is a transient systemic T cell depletion, and that cellular and antibody responses strengthened markedly during convalescence after successful antibiotic treatment. The relative immunodeficiency during the acute phase of pneumococcal infection, relative to age-matched healthy controls, was characterized by a deficit of systemic T cells and CD3^– CD56⁺ lymphocytes, and by weak levels of responses to PHA, TT and pneumococcal antigens (as shown in Fig. 1). Meanwhile, CD19⁺ B cell numbers remained within the normal range. The lymphopenia might be either a cause or a consequence of the acute infection. On one hand, moderate lymphopenia is not uncommon in people aged 65 years and older [22], and may favour infection. On the other hand, the T cell migration and sequestration in infected tissues (and in peripheral lymphoid organs) might also contribute to peripheral lymphopenia and is in favour of an early role of CD4⁺ T cells in protective responses to pneumococcal infections, as demonstrated in mice models [23,24]. Indeed, Kadioglu and colleagues showed that a subpopulation of CD4 cells migrates specifically towards the toxin pneumolysin. This migrated cell population expresses significantly increased levels of CD25 and this expression was linked to activated state of T cells in response to the pneumococcal toxin pneumolysin [23]. Moreover, previous studies have shown that the lymphopenia affects mainly activated T lymphocytes with a type-1 cytokine profile, suggesting that they might be involved in vivo in the immune response to S. pneumoniae [25]. As expected, the lymphopenia resolved during convalescence (not shown) with clinical improvement, and responses to mitogens and specific antigens increased accordingly. The link between T cell recovery and immune and clinical improvements argues in favour of a central role of these CD4⁺ T cells to control an effective immune response against S. pneumoniae.

In a second part, we also studied cellular activation and cytokine production in response to non-capsular pneumococcal antigens in vitro. It was evaluated by whole-blood culture with antigen, followed by flow cytometric detection of activated CD4⁺ CD25⁺ T cells. This method can be used to examine cellular immunity with small amounts of whole blood (5–10 ml), and avoids both mononuclear cell isolation and the use of radionuclides. In addition, activated lymphocyte subpopulations can be distinguished without purification steps. Activated T cells observed after culture with antigen express CD25 and proliferate in response to IL-2. These activated CD4⁺ CD25⁺ T cells are different from CD4⁺ CD25⁺ immunoregulatory T cells (T_reg) [26] observed in circulating blood. As a matter of fact, those CD4⁺ CD25⁺ T cells are absent from control cultures with RPMI-1640 medium, overexpress CD4 [27] and increase, in percentage terms, from day 4 to day 7 culture (G. Cozon, personal communication). We showed that CD4⁺ T cells observed after culture with TT (specific activation) or with pneumococcal antigens present an activated state as they express CD25 and proliferate in response to IL-2. Our study does not allow us to conclude if one or the other pneumococcal antigens is more effective at inducing a strong immune activation; no statistically significant differences were observed, due perhaps to the small number of analyses or because of the known cross-reactivity between antibodies to PspA [9]. However, specific cellular responses to some pneumococcal antigens were stronger in convalescent patients than in healthy controls (Rx1, FL-Rx1 and PsaA) and specific antibody titres also rose in all the patients during the convalescence phase, suggesting that a specific and effective immunity can be induced by challenge with the appropriate protein antigen. Although this specific immune response was elicited by the IPD, its protective role remains to be clearly demonstrated.

In healthy adults, it has been observed that natural immunity to the pneumococcus after natural exposure resulted in anti-capsular and weak anti-protein antibody responses [28]. In our study, we also found that T cells and cytokine profiles to the studied pneumococcal antigens were present in healthy adults who have had a long-life exposure to pneumococcal colonization of their upper respiratory tract. Indeed, in mice models, Malley et al. [29] suggested that immunity to pneumococcal colonization can be acquired after intranasal exposure to pneumococci in the absence of antibody but this protection required the presence of CD4⁺ T cells at the time of the challenge. CD4⁺ T cells may thus be implicated in the development of protection against pneumococcal disease but CD4⁺ T cell-dependent protection in humans may use different mechanisms that involve or not antibody production.

We found that the degree of CD4⁺ T cell activation by Rx1 (a crude pneumococcal antigen) and FL-Rx1 (the native protein) was similar. A similar profile of activation was also observed with the three recombinant PspA antigens, but the percentages of activated cells were generally lower. The recombinant PspA antigens elicited stronger responses in controls than in convalescent cases; this may be due to the fact that an acute pneumococcal infection results in the preferential stimulation of responses against epitopes not present in the recombinant PspA. However, no activation was shown with the PspA negative antigen (from strain FL-JY), whereas an activation was observed for all the PspA-containing antigens (crude pneumococcal antigens, native PspA, recombinant PspA antigens), in agreement with the existence of a specific response.

Another marker of cellular activation is the type of cytokines produced. Indeed, cytokines required to induce antibody production against polysaccharide antigens and protein antigens may be different, because only protein antigens can recruit cognate CD4⁺ T cell through TCR recognition of peptide–major histocompatibility complex (MHC) class II complexes on the surface of antigen-presenting cells (APC). In our study, the fact that a similar production of IL-10 (produced mainly by Th2 type lymphocytes) and IFN-γ (a potent proinflammatory cytokine) is detected during convalescence and in control subjects stimulated by pneumococcal antigens, and that this production was higher than that observed during the acute phase of infection, suggests that these cytokines may be involved in modulating systemic antibody responses. The fact that proinflammatory cytokines are induced in early innate response to S. pneumoniae has been shown to influence subsequent antibody production against protein–antigen. For example, tumour necrosis factor (TNF)-α strongly stimulates the primary anti-PspA response and the development of PspA-specific memory in response to intact S. pneumonia (strain R36A) [30]. This is in accordance with other reports, which show that T cell with type 1 cytokine profile are important for help with induction of the humoral response against S. pneumonia [25]. On the other hand, whereas type 2 cytokines are known as inducers of humoral responses, Khan et al. suggested that IL-10 may have a suppressive effect on the humoral response to S. pneumonia [30]. The cytokines environment requires further exploration of the induction of protein-specific antibodies, particularly in the context of a new vaccine.

In conclusion, this study shows clearly that adults with acute pneumococcal infection have a peripheral immunodeficiency specific to pneumococcal antigens. In addition, we demonstrated that this immunodeficiency is temporary. Cellular immunity normalizes during convalescence, and a specific cellular and antibody response to some cell-wall pneumococcal antigens (particularly PspA) is elicited, showing their relative accessibility to host immune system. These results suggest that these antigens induce both cellular and antibody responses, but their roles in protection remain to be further documented. If protective efficacy is demonstrated, these antigens could constitute a third-generation of vaccine candidates.