Waning and aging of cellular immunity to Bordetella pertussis

Abstract

While it is clear that the maintenance of Bordetella pertussis-specific immunity evoked both after vaccination and infection is insufficient, it is unknown at which pace waning occurs and which threshold levels of sustained functional memory B and T cells are required to provide long-term protection. Longevity of human cellular immunity to B. pertussis has been studied less extensively than serology, but is suggested to be key for the observed differences between the duration of protection induced by acellular vaccination and whole cell vaccination or infection. The induction and maintenance of levels of protective memory B and T cells may alter with age, associated with changes of the immune system throughout life and with accumulating exposures to circulating B. pertussis or vaccine doses. This is relevant since pertussis affects all age groups. This review summarizes current knowledge on the waning patterns of human cellular immune responses to B. pertussis as addressed in diverse vaccination and infection settings and in various age groups. Knowledge on the effectiveness and flaws in human B. pertussis-specific cellular immunity ultimately will advance the improvement of pertussis vaccination strategies.

Waning and aging features of human B- and T-cell responses specific for Bordetella pertussis are discussed, and this knowledge may be instrumental in the development of improved vaccines and vaccination strategies for pertussis.

Graphical Abstract Figure.

Waning and aging features of human B- and T-cell responses specific for Bordetella pertussis are discussed, and this knowledge may be instrumental in the development of improved vaccines and vaccination strategies for pertussis.

INTRODUCTION

Immunity to Bordetella pertussis is typified by an ambiguity. On the one hand, shortly after exposure to live B. pertussis or pertussis vaccine, high levels of specific antibodies and memory B and T cells are raised that can provide effective immune protection from disease. On the other hand, the acquired protective immunological memory is relatively short-lived; this is seen not only after natural infection (Wendelboe et al. 2005) and immunization with whole cell pertussis vaccines (wP) but also, and even more rapidly so as was recently discovered, after use of current acellular pertussis vaccines (aP) (Klein et al. 2012; Witt et al. 2013; Gambhir et al. 2015). Following an exposure to any pathogen or vaccine, cells from the specific immune system will pass through a clonal expansion phase and subsequently a contraction and maintenance phase, as they differentiate into various subsets of memory cells. Ideally, individuals maintain lifelong immunity after natural infection and vaccination as is the case for some viruses such as measles (Bouche, Ertl and Muller 2002). While clearly this is not the case for B. pertussis, it is currently unknown at which pace waning of immune mechanisms to B. pertussis is observed after different priming conditions and which threshold of sustained functional immune memory is required to provide protection to (ex-) patients and vaccinees from infection and disease at a subsequent exposure.

In the absence of a known correlate of protection for pertussis, such a threshold is difficult to define. The general view is that antibodies can prevent the attachment of B. pertussis to cells of the upper and lower respiratory tract; hence, antibodies with adhesin specificity and opsonizing or bactericidal effector function may provide protection. In addition, cell-mediated immunity (CMI) of the proper CD4⁺ T helper cell type is also implied, either by its own effector mechanism or by helping the antibody response (Plotkin and Gilbert 2012; Fedele, Cassone and Ausiello 2015). Many studies have described the waning of human B. pertussis-specific antibody levels after infection and vaccination (Giuliano et al. 1998; Versteegh et al. 2005; Hallander et al. 2009; Dalby et al. 2010; May et al. 2012; Fry et al. 2013), indicating the absence of durable effector mechanisms in the humoral compartment. This raises the question as to whether there is maintenance of cellular memory to B. pertussis that can provide rapid replenishment of the humoral compartment or mediate cellular immunity, involving both B-cell populations and CD4⁺ T helper cell types, but these have been studied less extensively. B- and T-cell components of the specific immune response need each other for the development of effective immunological memory. At various key stages in the specific B-cell response, specific CD4⁺ T cells provide cognate help to specific B cells, which is a prerequisite for the formation of germinal centers in which B-cell memory develops (Fig. 1A) (Slifka and Ahmed 1998; McHeyzer-Williams et al. 2012). There are examples that indeed inadequate CD4⁺ T-cell responses may be a limiting factor in developing B-cell responses (Lazuardi et al. 2005). On the other hand, specific memory B (B_mem) cells are the prime antigen presenting cells in the recall of memory CD4⁺ T-cell responses (de Wit et al. 2015). During these cellular interactions, activated B and MHC class II restricted CD4⁺ T cells exchange antigen-specificity, differentiation, proliferation and survival signals (Fig. 1B). So at various time points in the immune response mutual B–T cell cross-talk and licensing is required for the optimal development and maintenance of functional B as well as T-cell memory to B. pertussis. It is however unknown whether B. pertussis-specific B- and T-cell mechanisms equally wane, or which of these arms in the pertussis immune response may become limiting first.

Figure 1.

Multiple cognate interactions between B cells and CD4⁺ T cells specifc for B. pertussis. (A) Development of primary and recall B- and T-cell responses: The presence of B. pertussis (or antigens) trapped in draining lymph nodes is sensed by naïve B cells with IgM B cell receptors (BCR) with low affinity for antigen. Naïve B cells become activated, take up antigen via their BCR into lysosomal compartments, and process and present antigenic peptides in the context of MHC class II molecules. Dendritic cells (DC) also sense the pathogen, become activated and start presenting antigenic peptides in the context of MHC class II molecules while migrating to the draining lymph node, where they activate naïve CD4⁺ T cells with T-cell receptors (TCR) specific for the presented MHC class II-peptide complexes. CD4⁺ T-cell proliferate and differentiate into different functional subsets (Th1, Th2, Th17, Treg, T_FH). Surface expression of CXCR5 enables activated B and T_FH cells to comigrate to the CXCL13 rich B–T cell borders of the draining lymph node. Visualized are four stages (1–4) in the development of adaptive immune response in lymphoid organs in which reciprocal interactions between activated B cells and CD4⁺ T_FH cells with the same antigen specificity determine the clonal burst, differentiation and maintenance of both memory B and CD4⁺ T cell subsets (based on current literature and adapted from Tangye and Tarlinton 2009; Yoshida et al. 2010; Crotty 2011; McHeyzer-Williams et al. 2012). Stage 1: B cells that have productive and long-lasting interaction with specific CD4⁺ T_FH cells are licensed to start a germinal center (GC) reaction in which B cells undergo activation-induced cytidine deaminase (AID)-mediated BCR class switching and affinity maturation through somatic hypermutation. Activated B cells lacking cognate CD4⁺ T cell help can clonally expand and proceed into non-GC short-lived plasma cell development that may involve class switching to IgG. Inset: molecular interactions between activated B and T_FH cells in stage 1 are detailed in Fig. 1B and are representative for interactions at subsequent stages. Stage 2: GC B cells with sufficiently adapted BCR receive survival signals from CD4⁺ T_FH cells and leave the GC as either long-lived plasma cells or memory B cells. Stage 3: low amounts of recall antigen are sensed by memory B cells, taken up and processed and presented to memory CD4⁺ T_FH cells that localize close to B-cell follicles. In the secondary response, memory B cells are the prime antigen presenting cells for CD4⁺ T cells, which localize close to the B-cell follicles. Productive interaction with memory T_FH cells promotes memory B-cell expansion and 2nd plasma cell generation. Also a 2nd GC reaction can be started. Stage 4: 2nd GC B cells with further refined BCRs receive survival signals from 2nd CD4⁺ T_FH cells and can leave the GC. Whether 2^nd GC reactions generate both long-lived plasma cells and memory B cells like primary GC reactions is ill defined. IgM, IgG, IgA, IgE: Immunoglobulin M, -G, -A, -E. (B) [Inset:] Receptors and signals involved in cognate interaction between activated B and CD4⁺ T_FH cells sharing specificity for a B. pertussis antigen (also representative for other functional subsets of specific CD4⁺ T cells). Cognate interaction is promoted through CXCR5-mediated colocalization in the lymph node based on chemokine attraction. Reciprocal licensing occurs through MHC class II restricted recognition of CD4⁺ T_FH cells of cognate B. pertussis-specific peptides processed and presented by B. pertussis-specific B cells that have taken up antigen via their BCR. Efficiency of MHC class II presentation of peptide fragments is enhanced by BCR affinity for antigen. Further modifying interactions between specific B cells and CD4⁺ T_FH cells occur via accessory molecules, such as CD40-CD40L, OX40L-OX40, ICOS-ICOSL and (not shown) SLAM. Cells also exchange differentiation, proliferation and survival signals via cytokine production and expression of cytokine receptors. Recently, it was found that CD4⁺ T_FH cells besides predominantly expressing IL-21 can also secrete IL-4 and IFN-γ, known to regulate B-cell responses (Reinhardt, Liang and Locksley 2009).

Figure 1.

 (Continued)

Pertussis affects all age groups (Hewlett and Edwards 2005). Since the immune system undergoes changes throughout life, ranging from immatureness in infants to immunosenescence in elderly (van Gent et al. 2009; Fulop, Larbi and Pawelec 2013), age-related differences in the induction and waning of B. pertussis-specific immune responses are conceivable. Aspects of aging in T-cell immunity are for example the decrease of the total amount of T lymphocytes with age (Schindowski et al. 2002), the reversal of the CD4/CD8 ratio, the decrease of naïve T cell and increase of terminally differentiated T cells frequencies (Pera et al. 2015) and a larger role for T-cell regulation, as the ratio effector cells/regulatory cells declines (van der Geest et al. 2014). Aging is also proven to affect B cells. The frequency of effector (memory) B cells increases while the pool of naïve B cells diminishes and the diversity and affinity of antibodies are more limited (Colonna-Romano et al. 2006; Aw, Silva and Palmer 2007; Morbach et al. 2010).

This review will summarize current knowledge on the waning of human cellular immunity to B. pertussis after various priming conditions and the observed effects of age, in an effort to understand why pertussis immunity is, in general, relatively short-lived. Ultimately, knowledge on cellular key players responsible for the relatively rapid loss of immunity, especially after aP priming, will advance efforts to improve pertussis vaccines and vaccination strategies.

Waning patterns in B. pertussis-specific B-cell responses

The specific B-cell response begins when naïve B cells recognize antigen with their clonally expressed B-cell receptor, become activated and start proliferating. While a part of these activated specific B cells directly develops into short-lived antibody secreting plasma cells, another part moves, with the help of so-called follicular helper T cells (T_FH) (Crotty 2011), into germinal centers. Subsequently, cells emerge either as long-lived antibody secreting plasma cells that home to the bone marrow or as B_mem cells that recirculate in the blood as they head to secondary lymphoid organs (Fig. 1A). Upon recall with antigen part of these B_mem cells will, aided by memory T_FH cells, differentiate quickly into antibody secreting plasma cells, providing an initial rapid boost of the antibody response. Another part will reenter a germinal center reaction developing a second generation of B_mem cells (Manz et al. 2005; Tangye and Tarlinton 2009; Yoshida et al. 2010). Much of this mechanistic knowledge comes from animal models, but in order to better understand immunity to vaccine preventable infectious diseases human cellular B-cell responses against various pathogens have started to gain more attention (Crotty et al. 2003; Bauer and Jilg 2006; Buisman et al. 2009; Pinna et al. 2009; Alam et al. 2011; Kakoulidou et al. 2013; Ndungu et al. 2013).

Assays to enumerate B_mem cells to B. pertussis based on their in vitro differentiation into antibody-secreting cells (ASC) and detection in ELISpot were applied by Buisman et al. (2009), to first describe that specific long-term B_mem cells could be detected in vaccinated children whose antibody levels had already waned (Hendrikx et al. 2011a). Such B_mem cells may have a protective role, provided they can propagate a booster response rapidly enough to outpace pathogenesis of B. pertussis (Pichichero 2009). In the mouse, a direct protective role was shown for B. pertussis-specific B_mem cells in the absence of antibodies (Leef et al. 2000; Mahon, Brady and Mills 2000). In another study by Hendrikx et al., a second aP booster vaccinaton in Dutch wP-primed children 9 years of age heightened B_mem cell responses. The follow-up after one year showed a decline of these B_mem cell levels compared to +1 month after booster but still an enhancement compared to pre-booster levels, indicating that waning of B_mem cell levels occurs gradually (Hendrikx et al. 2011c). Whether waning of pertussis-specific B_mem cell levels after a second aP booster is different in aP-primed children is currently under investigation (A. Buisman, pers. comm). In an Italian study comparing two aP vaccines in 104 children, still >80% in both groups presented a positive B_mem ASC response 5 years after aP priming (Carollo et al. 2014). Buisman's group also compared B-cell memory in 6-year-old children either wP or aP primed in infancy and having received an aP booster at 4 years of age. For wP-primed children, levels of filamentous hemagglutinin (FHA)- and pertactin (Prn)-specific B_mem cells were higher at both 28 days and 2 years post-booster compared to pre-booster levels and there were no significant differences between the levels of 28 days vs 2 years post-booster. In contrast, aP-primed children showed a 2–5-fold decrease of pertussis toxin (Ptx)-, FHA- and Prn-specific levels at 2 years compared to 28 days post-booster, suggesting that aP-primed children may experience faster B_mem cell waning (Schure et al. 2013).

To investigate B. pertussis-specific B- and T-cell responses after a (natural) pertussis infection, our group set up an observational, cross-sectional study (the SKI study), comprising ∼300 (ex-) symptomatic pertussis patients of various ages and whose blood samples were collected between 2008 and 2012 at a known time after their laboratory confirmed diagnosis, ranging between <1 month and >10 years post-diagnosis (Han et al. 2013; van Twillert et al. 2014). In a subset of the SKI study, we measured pertussis-specific B_mem cell responses in 174 (ex-) patients. Peak levels reduced 2–3-fold within 9 months after antigen encounter (van Twillert et al. 2014). Besides our study, one other study evaluated B_mem cell responses induced by live B. pertussis in humans. In the randomized phase I clinical trial of the live-attenuated B. pertussis vaccine BPZE1, the seven subjects who exhibited nasopharyngeal colonization accumulated strong Ptx-, FHA- and/or Prn-specific B_mem cell responses between day 0 and 28, demonstrating the immunogenicity of BPZE1 in humans. At follow-up, 5–6 months after vaccination, these responses had declined. Despite suboptimal vaccine dosage, some subjects sustained elevated antigen-specific B_mem cell levels as compared to day 0, while others had responses that had declined to undetectable levels (Jahnmatz et al. 2014a).

Summarizing, we conclude that dynamics of B. pertussis-specific B_mem cells show an expansion phase soon after exposure to B. pertussis antigens, followed by a relatively rapid decay within several months and a maintenance phase in which levels can be detectable for years. The fact that peripheral B_mem cell peak levels wane after the acute phase of antigen exposure is to be expected due to normal contraction and homing to secondary lymphoid organs. The question remains whether the measured maintenance levels of specific B_mem cells are high enough for a quick recall reaction.

The induction and maintenance of pertussis-specific B_mem cells throughout life

Most studies that were performed to investigate B_mem cell responses to B. pertussis have so far concentrated on vaccination effects on frequencies of B cells specific for vaccine antigens Ptx, FHA and Prn in children and adolescents (Hendrikx et al. 2011a,b,c; Schure et al. 2013; Carollo et al. 2014; Jahnmatz et al. 2014b). The Buisman group showed that the levels of specific B_mem cells, measured in the maintenance phase, increased with age in groups of 3-, 4-, 6- and 9-year-old wP vaccine-primed children (Hendrikx et al. 2011a). In the SKI study that contains (ex-) pertussis cases of all ages, our group equally found an increase of B. pertussis antigen-specific B_mem cell levels with age, however, in the expansion phase (van Twillert et al. 2014). Here, higher levels of Ptx-, FHA- and Prn-specific B_mem cells were found in adults and (pre-) elderly compared to under-fours and schoolchildren. In subsequent phases after diagnosis, B_mem cell numbers declined for all groups to maintenance levels with no significant differences related to age. A re-analysis of these findings is summarized in Fig. 2A, showing waning of early induced levels of Ptx-specific B_mem cells within 3 months after exposure across age groups, but significantly stronger early peak levels at older age. The accumulating number of encounters with antigen with age, be it through vaccination or by circulating B. pertussis, was proposed to be involved in the higher B_mem cell levels found in older age groups (van Twillert et al. 2014). However, other age-related factors in B-cell immunity cannot be excluded. The gradual increase of specific B_mem cell levels after subsequent booster doses has earlier been found within the first 3 months after vaccination with diphtheria toxoid (Nanan et al. 2001). In summary, from these few studies it appears that B. pertussis-specific B_mem cell frequencies increase with age.

Figure 2.

Early and steady-state Ptx-specific B- and T-cell responses in (ex-) pertussis cases and impact of age. B cell (A) and T cell (B) responsiveness to Ptx in symptomatic (ex-) pertussis cases are two dimensionally plotted in relation to time elapsed after last exposure to antigen or diagnosis (Y-axes) and age at blood donation (X-axes). Responses are subgrouped in recently exposed (<3 months after date of diagnosis or after date of last vaccination if vaccination took place after clinical infection) versus retrospectively exposed (≥3 months) cases (Y-axis) and in younger (<35 years) versus older (≥35 years) age at blood sampling. Peripheral blood was obtained after informed consent from (ex-)pertussis cases at various time points after their laboratory confirmed diagnosis of symptomatic B. pertussis infection as part of the observational SKI-study (NVI-243, ABR NL16334.040.070). (A) Thawed PBMC from 179 (ex-)pertussis cases were cultured for 5 days in culture medium supplemented with an optimized cocktail of polyclonal and specific stimuli (3 mg/ml CpG, 10 ng/ml IL-10 and 10 ng/ml Ptx) to differentiate B_mem cells into ASC. Ptx-specific ASC were determined by using an ELISpot assay. Spots from at least two countable cell dilutions were used and expressed as geometric mean ASC per 10⁵ plated cells. Negative control wells not coated with Ptx but incubated with cells were used to calculate background. These background spot numbers were subtracted from the antigen-specific ASC. In 32 of tested pertussis cases (18%), B. pertussis infection was not the last exposure to Ptx, due to later vaccination. Strength of Ptx-specific B_mem cell responses was classified and indicated per case (white = no response ACS/10⁵ PBMC <2, green = medium response ACS/10⁵ PBMC 2–70, and red = strong response ACS/10⁵ PBMC > 70). The percentages in the quadrants indicate the fraction of strong Ptx-specific ACS responses of the studied subgroup. Waning of strong B_mem cell responsiveness in time was seen across age groups, being highest in older cases. In the early phase after exposure, strong responses were seen more frequently in younger versus older cases (Fisher's exact test, P = 0.0121). (B) Fresh PBMC from 62 (ex-) pertussis cases were stimulated with 1 μg/ml heat-inactivated Ptx protein for 7 days and Ptx-specific proliferation was assed using [³H]Thymidine incorporation the last 18 hours. Stimulation Index (S.I. = geomean CPM peptide/geomean CPM medium) was calculated. Strength of Ptx-specific T-cell responsiveness was classified and indicated per case (white = no response S.I. < 2, green = medium response S.I. 2–10, and red = strong response S.I. > 10). The percentages in the quadrants indicate the fraction of cases with strong Ptx lymphoproliferative responses of the studied subgroup. Waning of strong Ptx-specific lymphoproliferation in time was seen across age groups, but was highest in the older cases (Fisher's exact test, P = 0.0391).

Dynamics of B. pertussis-specific T-cell responses

In the course of infection or vaccination, dendritic cells will process endocytosed whole bacteria or antigen material and present protein fragments in the context of MHC class II molecules to naïve CD4⁺ T cells (Fig. 1A). These may clonally respond if their unique T-cell receptor recognizes a particular expressed MHC class II peptide specificity. Depending on the quality of their TCR signal and the innate cytokine and costimulatory environment, CD4⁺ T cells will then expand and differentiate into several of at least five T helper types (Th) of CD4⁺ effector cells, the Th1, Th2, Th17, Treg (regulatory T cell) or T_FH lineages, characterized by the transcription factors they express and the cytokines they secrete (Christie and Zhu 2014). After contraction of most of the peak response, part of the CD4⁺ T cell lineages will be maintained as central memory T cells. Quickly after sensing the presence of recall antigen and depending on their population size and imprinted memory potential, central memory CD4⁺ T cells will proliferate and differentiate into effector memory T cells. In these CD4⁺ T-cell recall reactions, antigen-specific B_mem cells are the main antigen presenting cells (Ise et al. 2014) and producers of cytokines (Shen and Fillatreau 2015), and hence strongly impact the outcome (Fig. 1A and B).

Although features of CD4⁺ T_FH cells may be very relevant in immunity to B. pertussis, they have not been studied in pertussis models so far. Instead most attention has gone to the analysis of other functional T-cell lineages, although these may also partially play a role in T_FH cell responses (Crotty 2011; McHeyzer-Williams et al. 2012; Hale and Ahmed 2015). Th1 and Th17 cells mediate the recruitment and activation of neutrophils and macrophages, which, with the help of opsonizing antibodies produced by plasma B cells, can take up and kill pathogens, as shown for B. pertussis (Higgs et al. 2012). Experimental models in mice and baboons have indicated the importance of pertussis-specific Th1 and Th17 type CD4⁺ T cells and redundancy of Th2 type CD4⁺ T cells in protection against pertussis (Mills et al. 1993; Redhead et al. 1993; Barbic et al. 1997; Mahon et al. 1997; Ross et al. 2013; Warfel, Zimmerman and Merkel 2014). Natural pertussis infection and wP vaccination in humans are known to induce Th1 and Th17 responses while aP vaccination yields a mixed Th1/Th2 response (Ausiello et al. 1997b, 1999; Ryan et al. 1997, 1998; Esposito et al. 2001; Mascart et al. 2003, 2007; Schure et al. 2012b). The discrepancies between these lineages of CD4⁺ T-cell responses to B. pertussis now seem essential for the effectiveness of the immune response, since currently used aP vaccines, although efficacious in the short term, protect less long (Klein et al. 2012; Witt et al. 2013; Gambhir et al. 2015). Collectively, these observations emphasize the importance of studies investigating both functionality and longevity of human pertussis-specific T-cell responses.

Pioneers researching human B. pertussis-specific T-cell responses in vaccine trials in the 1990s began with classical CMI studies. Typically, ³H Thymidine incorporation and cytokine production in culture supernatants of PBMC were measured after in vitro stimulation of PBMC with conventional B. pertussis antigens Ptx, FHA and Prn for several days, thus analyzing both the proliferative capacity and type of the T-cell response at the bulk level (Tomoda, Ogura and Kurashige 1991; Zepp et al. 1996; Ausiello et al. 1997a,b, 1998, 1999; Cassone et al. 1997; He et al. 1998). Based on these classical assays, CMI was proposed to better correlate with protection than pertussis-specific antibody levels, considering its greater magnitude and longer duration (Zepp et al. 1996; Cassone et al. 1997; Giuliano et al. 1998; Ausiello et al. 2000; Meyer et al. 2007).

CMI responses diminishing in time were shown in some studies, such as in Finnish 10–12 year olds, 8–10 years after primary wP vaccination (Tran Minh et al. 1998) and in Italian 4-year-old children, 42 months after primary aP vaccination (Ausiello et al. 1999). On the other hand, limited waning of CMI responses was seen in a Finnish follow-up cohort after an aP booster in adolescents. Here pertussis-specific CMI persisted above the pre-booster levels measured 5 years earlier (Edelman et al. 2007). In another study in adolescents and adults, B. pertussis-specific CMI responses were still elevated one year after aP vaccination whereas pertussis-specific IgG levels had declined considerably (Meyer et al. 2007). A German adolescent cohort showed detectable CMI responses 4–10 years after a pre-school booster dose in both aP- and wP-primed groups (Rieber et al. 2008). In younger children, Zepp et al. (1996) showed CMI responses in aP-vaccinated children that had even increased during the post-primary to pre-booster (at 15–19 months of age) period.

Few studies researched waning of CMI responses in clinically B. pertussis-infected cases. Ausiello et al. (2000) found that clinically infected non-vaccinated children, 8–59 (median 28) months after diagnosis had lower Ptx specific CMI responses compared to children 4–6 years after primary (aP) vaccination. The ex-pertussis cases had not been tested for CMI responses shortly after diagnosis; therefore, no interpretations could be made regarding the waning rate of these infection-related CMI responses. These findings however do suggest that infection and vaccination history differently affect the level of CMI maintenance.

Although classical CMI assessment is very sensitive, a second generation of B. pertussis-specific T-cell assays was developed to study T-cell responses at the single cell level. In IFN-γ ELISpot assays, frequencies of pertussis-specific T cells are determined after in vitro stimulation of PBMC, eventually combined with measuring cytokine levels in the supernatants of these cultures. The persistence of pertussis-specific T-cell responses has thus been studied in several vaccination studies. Dirix et al. (2009) found that the majority of 13 months old Belgian children showed B. pertussis-specific T-cell responses, i.e. secretions of IFN-γ and/or IL-13, 9 months after the last primary vaccine dose. Three years after vaccination in infancy Dutch children had still detectable specific T-cell responses, showing higher frequencies of IFN-γ producing cells and higher cytokine levels in aP- versus wP-primed children (Schure et al. 2012b). When these cohorts were boosted by a pre-school aP booster vaccinaton at the age of 4 and followed for 2 years afterward, aP-primed children had inferior maintenance of IFN-γ producing cells and in vitro produced IL-17 levels as compared to wP-primed children (Schure et al. 2013). In another Dutch study, pertussis-specific IFN-γ producing cell frequencies and T-cell cytokine concentrations had waned hardly in wP-primed 9-year-old children 1 year after an additional aP booster vaccination compared to 1 month post-booster, which the authors related to age (Schure et al. 2012a).

More recently, third-generation flow cytometry (FACS)-based assays have further enhanced the possibilities of single cell analysis of in vitro restimulated pertussis-specific T cells: these assays combine valuable information on class, proliferation, cytokine production and memory phenotype (differentiation stage). Investigators using these methods to interrogate B. pertussis-specific T-cell responses have mostly looked at vaccination effects (aP versus wP), but have not directly addressed waning by comparing early and late time points in the response (Carollo et al. 2012; Schure et al. 2012b; Smits et al. 2013; de Rond et al. 2015). Compared to children aP primed at infancy, wP-primed children displayed equally preserved proliferative capacity but more frequent cytokine responses by specific effector memory T cells in steady state around 10 years of age, even though the time elapsed since the pre-school aP booster dose given to both groups was higher in the wP-primed versus the aP-primed group, suggesting less pronounced waning in the wP-primed group (Smits et al. 2013). Direct ex vivo analysis of antigen-specific human CD4⁺ T cells requires the development of innovative tools such as HLA class II peptide tetramers. Recently, we showed proof of principle for human tetramer analysis based on an immunodominant Prn-specific epitope (Han et al. 2015); however, this method has no high-throughput potential to study waning of T-cell immunity to B. pertussis in large cohorts due to HLA polymorphism. Altogether, studies have shown that vaccination and infection history may differently impact the induction and maintenance of functional pertussis-specific memory T-cell responses. Similar to B_mem cell responses, memory CD4⁺ T-cell responses are sustained for years, however, seemingly with less overt waning.

The impact of age on B. pertussis-specific T-cell responses

Direct investigations of the effect of age on specific T-cell immunity to B. pertussis are scarce, though some studies have focused on distinct age groups while studying differences between responses after aP and wP vaccination or infection. The infant's immune system has been shown to be mature enough to develop Th1 (and Th2 responses) upon vaccination or infection, shown for 0- to 2-month-old neonates, even when born preterm (Mascart et al. 2003, 2007; Knuf et al. 2008; Vermeulen et al. 2010; White et al. 2010). FACS analysis showed higher maintenance levels of Th1 (IFN-γ and TNF-α) cytokine producing Ptx- and Prn-specific CD4⁺ T cells of the effector memory (CCR7⁻CD45RA⁻) phenotype in 9 year olds compared to 4 year olds (Schure et al. 2012a). This was in line with higher pertussis-specific IFN-γ producing T-cell frequencies and cytokine levels found in the Dutch 9 year olds, measured in ELISpot and luminex technology, respectively. Such accumulating T-cell levels with age were assumed to be caused by asymptomic infections, consistent with studies in Italian children that emphasized the impact of high circulation of B. pertussis on CMI (Ausiello et al. 1998, 1999).

Our group investigated B. pertussis epitope-specific CD4⁺ T-cell responses in symptomatic laboratory-confirmed (ex-) pertussis patients in two age groups in the SKI study. Remarkably, multi-epitope specificity long-term after infection was maintained in the younger but was lost in the older age group (Han et al. 2013). When re-analysed at the level of lymphoproliferation to whole proteins (Ptx, Prn), strong responsiveness was generally lost with time but most prominently in the higher age group (shown for Ptx in Fig. 2B). In a more recent analysis, children from the SKI study mounted a significantly higher number of Ptx-specific IFN-γ producing cells soon after infection and retained higher pertussis-specific cytokine maintenance levels compared to adults (van Twillert et al., manuscript in preparation).

Hence, maintenance of CD4⁺ T-cell responsiveness to B. pertussis appears to depend on age. In children steady-state T-cell levels, like those of B_mem cells, may show an increase with age through natural boosters. In contrast, in older age groups CD4⁺ T-cell responsiveness in the maintenance phase seems to decay with age (Fig. 2B), indicating that immunosenescence may come into play.

Conclusion and future perspective

While shortly after pertussis vaccination or natural infection protective pertussis immunity is in place, pertussis still occurs in vaccinated populations, indicating that the maintenance of protective immunological memory is incomplete. Awareness for waning of pertussis immunity has recently risen due to shortcomings of the aP vaccine. B. pertussis-specific B- and T-cell responses both persist longer than humoral immunity (Zepp et al. 1996; Cassone et al. 1997; Giuliano et al. 1998; Ausiello et al. 2000; Meyer et al. 2007; Hendrikx et al. 2011a; Schure et al. 2013). Only recently, more insight in the expansion and waning patterns of these pertussis-specific cellular immune arms is starting to accumulate. Reviewing the state of the art, the B- and T-cell data from (ex-) pertussis cases in the SKI study (illustrated in Fig. 2A and B, respectively) seem representative to some extent for patterns of B- and T-cell responses to B. pertussis found in various other studies. As expected, more intense responses are found shortly after an encounter with B. pertussis antigens, and waning of cellular levels can already be measured within several months. Cumulatively, the studies under review however indicated that both peak and maintenance levels could be influenced by vaccination type or infection history as well as by age. Furthermore, pertussis-specific memory B- and T-cell responses may follow different dynamics.

Remarkably, the peak levels of the Ptx-specific B_mem cells are higher in older recently infected individuals compared to the young, but then contract to similar maintenance levels (Fig. 2A). Yet the Ptx-IgG levels measured in serum of the older patient group were not higher than in the younger group (van Twillert et al. 2014), suggesting that mounting a larger specific B_mem cell response does not always need to correlate with stronger humoral responses by antibody secreting plasma cells. It is known that elderly have a reduced effectiveness in their B-cell compartment. The ability to generate high-affinity antibodies and IgG1 and IgG3 responses is reduced in elderly, as was shown for influenza A hemagglutinin antibody responses (Khurana et al. 2012; Frasca et al. 2013). Since immune complexes provide negative feedback on B_mem cell proliferation (Nimmerjahn and Ravetch 2006), a failure to produce effective antibody responses with age may explain higher peak levels of B_mem cells. In addition, accumulating exposures to specific antigen in life may generally lead to a more efficient B_mem cell burst and maintenance in sequential recall responses (McHeyzer-Williams et al. 2012). In the human population repeat encounters with pertussis antigen tend to occur, either by vaccine boosters or natural exposures to the pathogen, but the phenomenon of accumulating pertussis B-cell memory needs further investigation (Hendrikx et al. 2011a; van Twillert et al. 2014).

In contrast to the B-cell dynamics, incidence of strong Ptx-specific T-cell responsiveness was not higher in the older recently infected cases than in younger ones (Fig. 2B). However, long term after infection strong Ptx-specific lymphoproliferation was reduced in older cases (Fig. 2B). End-stage (CCR7⁻CD27⁻) differentiated (exhausted) memory CD4⁺ T cells have been associated with decreased proliferative capacity, compared to early-stage (CCR7⁺CD27⁺) differentiated memory CD4⁺ T cells (Fritsch et al. 2005) and impairment of human memory CD4⁺ T-cell responsiveness to stimulation has been associated with higher age (Schindowski et al. 2002; Mahnke et al. 2011). Therefore, reduced Ptx-specific CD4⁺ T-cell responsiveness with age may reflect immunosenescence.

Although the B. pertussis-specific memory responses remain poorly resolved, we suggest that the B_mem cell compartment has a tendency to inflate in older subjects (Fig. 2A) while T-cell responsiveness tends to diminish with age (Fig. 2B). Does this suggest that failing pertussis immunity is a matter of T-cell immunity? This seems to be too early to conclude since, besides magnitude, qualitative features of specific memory B_mem and T-cell responses may equate with protective immunity. Affinity of produced antibodies after B cell recall rather than absolute antibody levels or B_mem cell frequencies has been explored as a biomarker of the B-cell response. So far, studies have shown higher avidity of Ptx-specific IgG after natural infection compared to aP vaccination (Barkoff et al. 2012), whereas higher avidity of Ptx- and Prn-specific IgG was found in aP-primed children compared with wP-primed children (Hendrikx et al. 2009). Also, the balance between T helper cell types (T_FH, Th1, Th17 versus Th2 and Treg) and the memory phenotype of the T-cell response are important parameters for effectiveness.

Future research on cellular immunity to B. pertussis should aim to answer which threshold levels and qualitative features of specific B_mem and T cells are needed for sustained protection. Considering that B. pertussis displays immune evasion and modulation strategies (de Gouw et al. 2011; Henderson et al. 2012; Fedele, Bianco and Ausiello 2013; Jongerius et al. 2015), we postulate that a relatively high maintenance level and high quality for both B- and T-cell memory is required. State of the art technology should be used to advance this line of research. Affinity and avidity assays can reveal quality of antibodies in sera or in vitro produced by cultured B_mem cells, flow cytometry-based assays can assess the differentiation stage of memory T cells and high-throughput multiplex platforms can be used to determine the quality and breadth of antibody responses as well as the Th cytokine profile produced by memory T cells, all features implied in the shortcomings of aP vaccine-induced immune protection. This will lead to more knowledge on the protective potential of B and T cells and their individual roles in the waning problem of pertussis immunity, imperative to improve current pertussis vaccines and vaccination strategies.

FUNDING

The authors are funded by research grants of the Dutch Ministry of Health.

Conflict of interest. None declared.