INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Infection with Bordetella pertussis leads to temporary immunity and protection against subsequent disease (20, 28). Protective immunity can also be induced by immunization with whole-cell and acellular vaccines (6, 11, 15, 26). The mechanism by which natural infection or specific immunization induces protection against pertussis in children is still unclear.

Research into protection has traditionally concentrated on the role of humoral immunity, and studies of humans have shown that natural infection or immunization with whole-cell or acellular vaccines induces pertussis-specific antibodies (12, 13, 14, 18, 22, 32); however, large clinical trials have not provided evidence of a direct correlation between serum antibody titers and disease protection (33, 34, 35).

Although circulating antibodies may play a role in toxin neutralization and prevent bacterial attachment to respiratory epithelial cells, evidence that other immunological mechanisms, such as cell-mediated immunity (CMI), may be required for sufficient and long-lasting protection against B. pertussis is emerging (3, 5, 9, 10, 39). In mice infected with B. pertussis, CMI may help eradicate bacteria from the respiratory tract, probably by activating type 1 CD4⁺ T-helper cells (24). The induction of CMI responses to B. pertussis vaccine constituents has also been clearly demonstrated for humans (2, 4, 31, 37).

Little is known about the capacity of different pertussis vaccines to elicit CMI in infants or whether these responses correlate with antibody titers against B. pertussis antigens and/or disease protection. Even less is known about the duration of immunity and about differences in the types of protection induced by immunization and natural infection. The long-term evaluation of antibody profiles and CMI may therefore contribute to our understanding of the immunological characteristics of pertussis-specific protection and provide important clues as to the best vaccination schedule.

The aim of this study was to compare pertussis-specific humoral and cellular immunity in children 5 years after a primary vaccination with a combined diphtheria, tetanus, tricomponent acellular pertussis, and hepatitis B vaccine (DTaP-HBV; InfanrixHepB; SmithKline Beecham) with immunity after natural infection.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Study design and study population. This was a multicenter, single-blind study (i.e., all of the laboratory examinations were carried out without technicians knowing the group to which the children belonged). Previously healthy male and female children aged 5 to 6 years at the time of enrollment, free of health problems as established by their medical histories and a clinical examination before entering the study, and who had received DTaP-HBV at 3, 5, and 11 months of life or had acquired pertussis at the age of 3 to 11 months were considered eligible for inclusion. Pertussis was defined as an illness with 21 or more days of paroxysmal cough and evidence of B. pertussis infection in cultures or diagnostic serologic tests (6, 38).

Vaccine. The per-dose composition of DTaP-HBV was 25 limes flocculation units of diphtheria toxoid, 10 limes flocculation units of tetanus toxoid, 25 µg of pertussis toxin (PT), 25 µg of filamentous hemagglutinin (FHA), 8 µg of pertactin (PRN), and 10 µg of hepatitis B surface antigen. The vaccine was manufactured by SmithKline Beecham Biologicals (Rixensart, Belgium). Three lots (lots 16705, 16707, and 16708) were used, and the vaccine was injected intramuscularly into the thigh.

Pertussis antigens. The B. pertussis soluble antigens used to evaluate the humoral and cell-mediated immune responses were kindly supplied by SmithKline Beecham. In agreement with the methods used by Zepp et al. (39), PT was detoxified by treatment with glutaraldehyde followed by formalin. FHA and PRN were treated with formalin in order to inactivate any residual traces of PT.

Patient enrollment and evaluation. The children were enrolled between December 1999 and January 2000 at two public health clinics operated by the Italian National Health Service: the Institute of Infectious Diseases, University of Palermo, and the Institute of Pediatrics and Neonatology, University of Bologna. The serum and heparinized venous blood samples were obtained at these local health units and then sent to the central laboratories for evaluation (Department of Hygiene and Microbiology, University of Palermo, and Pediatric Department I, University of Milan). The 1-ml serum samples were used to measure antibody titers against PT, FHA, and PRN; the 7-ml heparinized venous blood samples were used to assess CMI in the lymphoproliferation and cytokine assays.

Humoral immunity. Serum antibodies were measured at the Department of Hygiene and Microbiology, University of Palermo. IgG antibody titers against PT, FHA, and PRN were determined using a standardized enzyme-linked immunosorbent assay (ELISA) according to previously described procedures (5, 16, 35). The reference-line method was used to calculate the ELISA units with standardized software (Unitcalc [1992]; Biosys INOVA, Stockholm, Sweden), and the results were expressed in ELISA units per milliliter. For each ELISA, the estimated minimum detection level and the positive serologic response to each pertussis antigen were considered as previously described (5, 16). Humoral immunity was evaluated in each study population.

CMI. CMI was evaluated at Pediatric Department I, University of Milan. The peripheral blood mononuclear cell (PBMC) fraction was obtained within 24 h of delivery by means of centrifugation on a density gradient (Becton Dickinson, Rutherford, N.J.) and washed twice in phosphate-buffered saline (Organon Teknika Corp., Durham, N.C.); the number of viable cells was determined using the trypan blue exclusion test (8). The PBMCs were suspended at 3 × 10⁶/ml in RPMI 1640 medium (Gibco Laboratories, Grand Island, N.Y.) containing 0.5% penicillin and 1% glutamine (8). For both proliferation assays and cytokine production, 0.1 ml of the lymphocyte suspension per well was added in triplicate to 96-well flat-bottom microtiter tissue culture plates (Costar Corp., Cambridge, Mass.) (8). A total of 0.1 ml of complete medium and pooled human plasma (5%) was added to each well. PBMC proliferation and cytokine secretion were evaluated in all of the children with sufficient cell recovery from each blood donation, without any a posteriori selection.

Statistical analysis. The data from all of the antibody titer, proliferation, and cytokine determination experiments were recorded in a computerized database. The descriptive analyses were made using the SPSS (Chicago, Ill.) statistical package. The differences between proportions were assessed by means of the ² test with Yates correction or Fisher's exact test; the differences between mean values were assessed using the Mann-Whitney U test. Each difference with a P value of <0.05 (two tailed) was considered significant.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Serum antibody determinations. Table 1 shows the percentages of children with a positive humoral immune response and their geometric mean titers (GMT). A positive response to PRN was significantly more common in the DTaP-HBV recipients than in the pertussis group, but there was no significant difference in the proportions of seroresponders to PT or FHA. Significantly higher anti-PRN GMT were observed in the DTaP-HBV recipients than in the pertussis group, but the GMT of the antibodies against PT and FHA were not significantly different between the two groups.

Proliferative response. Table 2 shows the lymphoproliferative responses to the three soluble B. pertussis antigens by percentage of responders and magnitude of lymphoproliferation. However, due to the small quantity of cells recovered from the blood of some children, only 48 subjects were tested (see Materials and Methods). The proportions of proliferation-positive responses to at least one antigen were 23.5% in vaccine recipients and 21.4% in the pertussis group. No significant differences were observed in the proportions of proliferation-positive responders to each antigen, and there was no significant difference between the SIs of the two groups. The magnitude of the lymphoproliferation response to PRN was significantly higher in the DTaP-HBV recipients than in the pertussis group, whereas no significant difference was observed in the lymphoproliferation responses to PT or FHA.

Cytokine profile. Table 3 shows the cumulative results of IFN-, IL-2, IL-4, and IL-5 production induced in the B. pertussis antigen-stimulated PBMC cultures. For this assay, due to the small quantity of cells recovered from the blood of some children, only 40 subjects were tested (see Materials and Methods). Analysis of PBMC cytokine production against B. pertussis antigens revealed that responses were still detectable in a minority of the children in both groups. The PBMCs of all of the study subjects showed low rates of production of IFN-, IL-2, IL-4, and IL-5 in response to each antigen. All the cytokines were highly produced in PHA-stimulated PBMCs of the study population. Regardless of the stimulatory antigen, IFN- secretion was significantly higher in the pertussis group. In contrast, production of IL-2 and IL-5 was higher, but not significantly so, in the vaccine recipients. Little or no IL-4 was produced by the antigen-stimulated cells from both groups.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This study compared the long-term persistence of humoral and cellular immunity to pertussis antigens in children who had received three doses of DTaP-HBV at the ages of 3, 5, and 11 months with that observed in children who had suffered from pertussis during their first year of life. After 5 years, the residual immunity due to the administration of the combined DTaP-HBV was quite similar to that induced by natural infection. In both cases, only a small proportion of the children had significant serum IgG concentrations specific to all three B. pertussis antigens, and T-cell responses persisted in a minority of subjects. Notwithstanding the low levels of significance, which are partly a reflection of the relatively small number of samples, the cytokine data show clearly that the infection induces a type 1 response (IFN- and no IL-4 or IL-5) but that DTaP-HBV induces a type 2 skewed response (IL-5 and little or no IFN-).

To the best of our knowledge, this is one of the few studies evaluating humoral and CMI responses to B. pertussis antigens a long time after vaccination in comparison with those remaining after natural infection. Our results also allow an assessment of immune responses in infants vaccinated at 3, 5, and 11 months of age, the schedule commonly used in Italy and other European countries (21).

It is not known how long the immunity provided by pertussis vaccines and natural infection persists. Data from both old and recent studies indicate that, in immunocompetent subjects, neither childhood immunization nor infection provides long-term immunity against B. pertussis (7, 17), which is why older children and adults are major reservoirs of infection (1). He et al. have shown that Finnish children become susceptible to clinical pertussis after school entry, thus suggesting that immunity persists for no longer than 5 years after the last immunization at 2 years of age (19). In another study, no specific cell responses to PT and FHA were observed in one of three adults who had received a dose of bivalent vaccine (PT and FHA) 6 years before being tested (17), which suggests that the CMI induced by pertussis immunization starts to decrease and may even become undetectable over 6 years.

In our study, the proportion of children with a significant pertussis-specific humoral and cellular response 5 years after vaccination or natural infection was relatively low. Both DTaP-HBV and natural infection provided similar immune responses and long-term declines. Although caution must be used in interpreting these data because no correlate of protection has been firmly established, our results suggest that most vaccinated or previously infected children should receive a booster in order to maintain long-term pertussis-specific humoral and CMI response.

Like Zepp et al. (39), we used detoxified PT for in vitro stimulation because native PT is known to display mitogenic activities on lymphocytes (22, 25). Consequently, detection of specific responses using native antigens may be hampered by nonspecific T-cell proliferation due to the toxin. The observation of Zepp et al., who demonstrated a PT-specific antibody and T-cell response following vaccination with glutaraldehyde- and formalin-detoxified PT (39), supports the observations that, in our study, the weak antibody and CMI responses to PT appear to be a consequence of the limited duration of immunity and that they do not seem to be a reflection of the effect of glutaraldehyde and formalin treatment of this antigen.

In our population, the humoral and proliferative responses to PRN seemed to be greater and more frequent after immunization than after natural infection, possibly because of antigenic differences between the PRN of the bacteria causing infections and that used for in vitro stimulation. The PRN used for immunizations had been treated with formaldehyde, which may have destroyed some of the antigenic epitopes recognized by the T cells induced during natural infection (10). Similar data were also obtained by He et al., who suggested that the selective pressure caused by an immunization program lasting more than 40 years might have changed the PRN of the B. pertussis bacteria existing in the childhood population (17).

A comparison of antibody levels and CMI responses suggests that antibodies to the three pertussis components may persist at detectable levels for longer than we could detect significant CMI responses. These data are in agreement with one previous report (37) but not with others (5, 36, 39). There are a number of possible explanations for this finding. Technical reasons meant that only a relatively small amount of PBMCs could be used, and this may have decreased the sensitivity of the assay. We chose a CMI response (SI) cutoff point of 4, which has also been used in other recent studies (5, 37). A lower cutoff point would increase the sensitivity but decrease the specificity of the proliferation assay. However, as previously observed (37, 39), it is possible that CMI shows an immediate strong stimulation after infection or a vaccine booster because of an effective priming of a memory response to relevant B. pertussis antigens.

Since CMI is induced and regulated by cytokines, we also studied the profiles of the major cytokines elicited by in vitro PBMC stimulation with B. pertussis antigens. It has been shown that recovery from pertussis and the immunization of children with whole-cell pertussis vaccines selectively induces a type 1 immune response (23, 29) but that acellular pertussis vaccines induce T cells that secrete both type 1 and type 2 cytokines (4, 30). In our study, a preferential type 1 cytokine response with IFN- secretion was observed in our pertussis group whereas a type 2 skewed response (IL-5 and little or no IFN-) was observed in the vaccinated children; however, the quantitative differences in the amounts of cytokines produced by DTaP-HBV and natural infection were minimal.

Our findings also demonstrate that cytokine synthesis does not persist in a high proportion of children 5 years after natural infection or pertussis immunization. These data are in contrast with the results of Ryan et al. (30). This may be due to natural boosting caused by subclinical infections in their population, regardless of the use of pertussis vaccines.

In our study, proliferation and cytokine production appeared to be low only in response to pertussis antigen whereas they seemed to be high when PBMC responses against PHA were tested. This difference helps to allay any worries about the assay system and provides a convincing explanation for the declining responses observed in a high proportion of vaccinated and previously infected children.

In conclusion, our study shows that long-term pertussis-specific humoral and CMI responses after DTaP-HBV administration are very low and comparable with those observed after natural infection. These findings suggest the need for careful consideration of antipertussis immunization policies, with not only the vaccination of infants but also booster administrations in adolescents and adults being considered in order to maintain adequate levels of a pertussis-specific immune response.