Long-term antibody responses following human ... - Wiley Online Library

Blackwell Science, LtdOxford, UKCEIClinical and Experimental Immunology0009-9104Blackwell Publishing Ltd, 2002 130 Original Article S. A. Cawthraw et al.

Clin Exp Immunol 2002; 130:101–106

Long-term antibody responses following human infection with Campylobacter jejuni S. A. CAWTHRAW*, R. A. FELDMAN†, A. R. SAYERS* & D. G. NEWELL* *Veterinary Laboratories Agency (Weybridge), Addlestone, and †Queen Mary’s College, University of London, London, UK

(Accepted for publication 16 July 2002)

SUMMARY Epidemiological evidence suggests prior infection of humans by Campylobacter jejuni leads to protection against disease following further exposure. It is known that infections elicit strong antibody responses following the onset of disease and that antibody levels are elevated in putatively immune populations. To determine if systemic and mucosal antibodies induced by a confirmed infection remain at elevated levels for prolonged periods, repeat serum, saliva and urine samples were taken from campylobacter patients from 1 week and up to a year postinfection. Antibodies were monitored by ELISAs using three different antigen preparations: acid-glycine extracts (AE) of C. jejuni strain 81116 and an aflagellate mutant (R2), and a whole-cell R2 sonicate, and by Western blotting. Levels of serum IgG antibodies against 81116AE and R2 sonicate, but not R2AE, remained significantly raised over time when compared to a comparison population. Serum anti-sonicate IgA antibody levels were initially significantly raised but decreased over time to levels similar to the comparison group. There were no significant differences in levels of salivary IgA against the AEs. Anti-sonicate salivary IgA and IgG levels were initially significantly higher than in the comparison group. Both declined over time but the IgG levels remained significantly higher. Significant correlations were seen between serum IgG levels and age and duration of illness. Serum antibodies against flagellin, 40 kDa and 29 kDa antigens were still detectable in most patients up to a year postinfection, as were salivary antibodies to flagellin, the major outer-membrane protein and a 40 kDa antigen. Keywords

antibodies antigens

Campylobacter jejuni

INTRODUCTION Campylobacter jejuni is a major cause of acute bacterial enteritis in children and adults in industrialized countries. In non-industrialized countries, however, although colonization in preschool children is common, disease is rare in individuals over the age of 2 years [1] and this absence of disease may be associated with raised antibody levels [2]. This has led to the suggestion that repeated challenges with campylobacters induce antibody responses, which protect a child or adult from disease, though not necessarily from colonization [3]. Experimental evidence from human volunteer [4] and non-human primate [5] studies indicate that prior challenge with campylobacter can induce protective immunity. The efficacy and duration of this acquired protective immunity is unclear, but it is possible that such protection could confound the results of case–control studies to identify risk Antibodies to C. jejuni

Correspondence: Professor D. G. Newell, Veterinary Laboratories Agency, New Haw, Addlestone, Surrey KT15 3NB, UK. E-mail: [email protected] © 2002 Blackwell Publishing Ltd

saliva

serum

factors and monitor possible intervention strategies for disease resulting from food-borne infection [6]. These epidemiological problems are further complicated by evidence suggesting that the majority of campylobacter infections go unreported [7]. Thus, there is a recognized need for a method to measure the immune status of individuals involved in epidemiological studies [6]. Ideally, such a method would utilize clinical samples obtained non-invasively. Given the enteric nature of campylobacter infections, it is likely that mucosal antibodies play a key role in such immunity. Although specific faecal, urinary and mammary antibody responses have been reported, particularly of the immunoglobulin A (IgA) isotype [4,8,9], salivary anti-campylobacter antibodies have not yet been investigated. Moreover, most research to date appears to have focused on the characterization of antibodies induced during the immediate convalescent phase following campylobacteriosis. The specificity of longer-lasting, and potentially protective, antibodies induced by infection, of either serum or mucosal origin, is unknown. Recently the specificity of serum antibodies in poultry abattoir workers, occupationally exposed to

101

102

S. A. Cawthraw et al.

campylobacters for periods of less than, or more than, 3 months, has been described [10]. Several C. jejuni antigens, including flagellin and a protein antigen of 40 kDa were identified as potentially initiating longer-term antibody responses. However, the role of these antigens in the induction of persistent mucosal or serum antibodies, following an infection with C. jejuni which resulted in diarrhoea, is unknown. The aim of this study was to detect and characterize, by enzyme-linked immunosorbent assay (ELISA) and Western blotting, those anti-C. jejuni antibodies persisting in the serum, saliva and urine of individuals with a diarrhoeal illness and a positive stool culture for Campylobacter. Questionnaires were completed by each patient to establish whether factors such as age, duration and severity of illness, travel history and antibiotic treatment influenced the detectable antibody responses. In addition, using selected control groups, attempts were made to determine criteria for the establishment of population-based assays of protective immunity.

MATERIALS AND METHODS Clinical samples One hundred sequential patients, who consulted general practioners in the London area during 1995–96 and who had campylobacter-culture-positive stools, were identified. Eighty-four of these patients (age range 4–69 years) were recruited into the survey. Each patient (or, in the case of minors, their parents) filled in a questionnaire requesting information on age, gender, duration of illness, whether hospitalized or not, any antibiotic therapy prescribed, self-administration of antidiarrhoeals and recent travel history. All procedures for the survey and samples were approved by the Ethics Research Committee of East London and the City Health Authority. Each patient recruited was visited by a trained nurse on up to four occasions. The nurse collected serum, saliva and urine samples. Whole saliva was collected by oral swabs (Omni-SAL saliva collection device, Saliva Diagnostic Systems, Singapore). The first clinical samples were taken between 1 week and 2 months postinfection (mean = 40 days), with three follow-up saliva and urine samples taken in most cases approximately every 1–2 months thereafter. A second serum sample was taken at the time of the last saliva/urine sampling (mean 218 days postinfection). Fiftynine of the 84 patients provided a complete set of clinical samples, with the rest being sampled on three or less occasions. Samples were stored frozen at −20°C until assayed. Serum (n = 39) and saliva (n = 43) samples were also collected from unrelated populations of individuals (comparison groups) (blood donors and laboratory and office staff of the Veterinary Laboratories Agency) with no known history of campylobacter infection and no recent diarrhoeal illness. Enzyme-linked immunosorbent assays (ELISA) Immunoglobulin G and IgA anti-campylobacter antibodies in serum and saliva were measured by ELISA. Previous investigations [10] have showed that the acid-glycine extractable antigens (AE) [11] of C. jejuni strain 81116 provides a suitable capture antigen for such studies. The use of a single strain ensures that antibodies directed largely against common campylobacter antigens are detected. A considerable proportion of this material is flagellin, which is highly antigenic. To determine the relative contribution of antibodies directed against flagellin, an AE of R2, an

aflagellate (flaA– flaB–) site-directed mutant of C. jejuni strain 81116 [12] was also used as the capture antigen. ELISAs were performed as described previously [10]. Briefly, the capture antigen was coupled to the ELISA plate at 2 µg/ml. Human sera was used at dilutions of 1 in 800 and saliva samples were used at dilutions of 1 in 4 regardless of isotype detected. Specific IgG and IgA antibodies were detected using horseradish peroxidase (HRP)labelled rabbit antihuman IgG or IgA (1 in 2000 dilution) (Dako, Glostrup, Denmark). A sonicate of C. jejuni 81116 R2 (5 µg/ml) was also used to determine the contribution of non-acid extractable and nonflagellin antigens. For these ELISAs, sera were used at dilutions of 1 in 3700 and 1 in 80, and saliva samples at dilutions of 1 in 4 and 15 in 100, for IgG and IgA antibody detection, respectively. Specific antibodies were assayed in the first urine samples from 25 of the patients. The urine was used either neat or at a fivefold concentration (10× concentrate diluted 1 in 2 in phosphate buffered saline) in the assay. The urine samples were concentrated using a centrifugal concentrator with a 10-kDa cut-off (Amicon, UK).

Western blotting The antigenic specificities of serum and salivary antibodies were Western blotted against the total protein profile of C. jejuni strain 81116, to detect primarily antigens with conserved linear epitopes. The methods used for Western blotting have been described previously [10]. Briefly the total protein profile of C. jejuni strain 81116 was produced by SDS-PAGE on a 10–25% (w/v) gradient gel. The separated proteins were electroblotted onto supported nitrocellulose (Electran, BDH). The blot was blocked with 3% (w/v) dried skimmed milk and cut into strips. Each strip was incubated in sera (diluted 1 in 50) or undiluted saliva overnight at 4 °C. The strips were then incubated in horseradish peroxidase (HRP)labelled rabbit antihuman IgG or IgA (1 in 2000 dilution) (Dako, Glostrup, Denmark) for 1 h at room temperature. Bound IgG and IgA antibodies were visualized using 3-amino-9-ethyl carbazole as the chromogen. Statistical analysis Statistical analyses were performed on the ELISA results to compare any differences in the antibody levels between samples from the patient and comparison groups. Logarithmic transformations were performed on the data prior to analysis by Student’s t-test. To determine whether the trends over time (i.e. the time postinfection that the samples were taken) were significant, analyses of covariance were undertaken. Analyses of variance were also performed to determine any effects of age ( £30, 31–50, >50), gender, hospitalization, medical treatment (antibiotics or antidiarrhoeals), prior travel and duration of illness ( £7 days or >7 days) on the ELISA results.

RESULTS For the first sera collected from all patients, the mean optical densities (O.D.) for IgG antibodies against the AE of R2 and the AE of 81116 were 0·19 ± 0·06 and 0·45 ± 0·39, respectively (Fig. 1). As the only difference between these strains was the deletion of the flaA and flaB genes, these results indicated that a large proportion of antibodies were directed against flagellin. Comparison of the mean O.D.s from the first and second sera collected indicated a significant decrease (P < 0·001) over time of antibodies detectable

© 2002 Blackwell Publishing Ltd, Clinical and Experimental Immunology, 130:101–106

Antibodies to C. jejuni

103

1.25 a

e 1.00 ELISA O.D.450

c 0.4

0.75 c 0.50

d/f

a/c b

Fig. 1. ELISA results (O.D.450) indicating levels of serum IgG and salivary IgA antibodies detectable against glycine/acid extracts (AE) of C. jejuni strains 81116 and its aflagellate mutant R2. Samples were taken from patients shortly after a diagnosed infection (first) and up to a year later (last) and from comparison groups. Significant differences between the first or last samples and the corresponding comparison group are indicated as follows (P-values): α < 0·001, β < 0·001, χ = 0·005. , 81116 AE; , R2 AE.

with the 81116 AE (0·25 ± 0·20) but not the R2 AE (0·17 ± 0·05) supporting the role of the antiflagellin antibodies in the convalescent immune response. The mean ODs of specific serum IgG against the R2 and 81116 AEs for the comparison group were 0·15 ± 0·06 and 0·17 ± 0·09, respectively. Both these O.D.s were significantly lower than the first patient sera for both AEs (P < 0·001). However, with the second sera only the O.D.s using the 81116AE were significantly lower than those of the comparison group ( P = 0·005). The mean specific salivary IgA antibody levels detected with 81116AE were also slightly higher than with R2AE (Fig. 1) in both sets of patient serum samples tested. However, there was no detectable fall in antibody levels over time. Moreover, no statistically significant differences were detected in the IgA antibody levels of saliva samples from the patient and comparison groups. These studies, using the acid extract antigens, indicate the importance of surface antigens, especially flagellins. To investigate the importance of other, potentially cellular, antigens a sonicate preparation derived from the aflagellate mutant was used. For the first and second sera collected from the patients, the mean O.D. for IgG antibodies was 0·62 ± 0·4 and 0·38 ± 0·24, respectively (Fig. 2) which was a significant decrease over time (P < 0·001). However, several paired sera showed very little change and in 6/59 cases the O.D. of the second sera was increased by more than 10%. The mean O.D. of the sera from the comparison group (0·24 ± 0·10) was significantly lower (P < 0·001) than the mean O.D. of both the first and second sera from the patient group. The mean O.D. for circulating IgA antibodies directed against whole-cell R2 sonicate for the first and second sera from the

Comparisons

Last saliva

1st saliva

Comparisons

Last saliva

1st saliva

Last sera

0.00

Comparisons

0.2

b

0.25

Last sera

b

1st sera

f

b a/c

Comparisons

ELISA O.D.450

d e

0.6

0.0

a

1st sera

0.8

Fig. 2. ELISA results (O.D.450) indicating levels of serum and salivary IgG and IgA antibodies detectable against a whole-cell sonicate of C. jejuni strain R2. Significant differences between the first or last samples and corresponding comparison group are indicated as follows (P-values): α–δ < 0·001, ε = 0·029, φ = 0·002. , IgG; , IgA.

patient group were 0·31 ± 0·20 and 0·23 ± 0·12, respectively (Fig. 2). Four of 59 patients showed a greater than 10% increase in circulating IgA antibodies over time. The mean O.D. for serum IgA antibodies in the comparison group (0·21 ± 0·09) was significantly lower than that of the first sera (P < 0·001) but not significantly different to the second sera ( P > 0·05) from the patient group. Salivary IgG and IgA antibodies, directed against the R2 sonicate, were investigated (Fig. 2). The mean O.D. for specific salivary IgG from the first saliva sample decreased over time (0·60 ± 0·42–0·42 ± 0·16). The corresponding mean O.D.s for specific salivary IgA were 0·74 ± 0·44 and 0·54 ± 0·28, respectively. In contrast, the mean O.D.s of the saliva samples from the comparison group were 0·53 ± 0·34 and 0·34 ± 0·10 for IgA and IgG, respectively. Statistical analyses indicated that the specific salivary IgG levels in comparison group were significantly lower than those in both the first and last saliva samples (P < 0·001 and p = 0·002, respectively) from the patient group. Specific salivary IgA levels were also significantly lower in the comparison group compared to the first saliva sample (P = 0·029) but not to the last saliva samples (P = 0·069) from the patient group. As observed previously with the serum antibodies, not all saliva samples showed a decrease in specific antibodies over time. The O.D. levels, for specific salivary IgG and IgA antibodies, increased (by greater than 10%) in eight and 13 of 59 patients, respectively. Nevertheless, the decrease in the mean O.D. over time was significant for IgG antibodies (P = 0·001) but not quite for the IgA antibodies (P = 0·05). Specific antibody levels in the urine samples from patients were found to be very low. Even after a fivefold concentration, only 11 of the 25 samples tested (taken 18–44 days postinfection) gave O.D.s for specific IgA of greater than 0·1. On the basis of


104


these preliminary experiments no further work was done using urine. The analysis of variance for the detection of the effects of age, gender, hospitalization, medical treatment, recent travel and duration of illness on the O.D.s of the first and last serum and saliva samples tested against all 81116 and R2 antigens is shown in Table 1. There was a consistently significant association between higher O.D.s obtained from the serum IgG antibody assays and increasing patient age. This trend was not observed with salivary IgA antibodies. Notably there was also a significant association between serum IgG antibodies directed against the cellular antigens and both duration of illness and the need for hospitalization. Western blotting, of the first serum samples from each patient, for IgG antibodies detected a large number of antigens in C. jejuni strain 81116 and identified variable responses between patients (Fig. 3). In the 82 initial serum samples tested the antigens most commonly, and most strongly, recognized were flagellin (95%), a 40 kDa antigen (83%) and a 32 kDa antigen (80%). A 66 kDa antigen (66%) and a 29 kDa antigen (66%) were also recognized frequently by these sera. A number of other antigens, particularly in the 25–35 kDa region, were recognized occasionally, although the responses to these were usually weak. The major outer membrane protein (MOMP) was recognized, albeit weakly, by only 15 of 82 (18%) samples. Western blots of IgG antibodies were also performed using the 61 follow-up serum samples, taken 4–12 months postinfection. These blots showed that the responses to many of the antigens were weaker by the time of the second serum sample but that antibodies were still present directed against the flagellin (97%), the 40 kDa (77%) and the 29 kDa antigens (69%). Sera from 40 patients were also Western blotted for IgA antibodies. Detectable serum IgA responses were mainly directed against the 62 kDa flagellin protein, although anti-40 kDa antigen antibody responses were seen in 15 of 40 (37·5%) sera.

Western blots were also performed using the first and the last saliva samples. Compared to the circulating antibody responses, the salivary responses were visually weaker. The only salivary IgG antibodies detected were directed against flagellin. However, in the first samples specific IgA salivary antibodies were directed against a number of antigens, notably of molecular masses 116, 97, 66, 62, 47, 43–45, 40, 32 and 29 kDa. The most commonly and strongly recognized of these was flagellin (62 kDa), recognized by 20 of 24 (83%) patients. Other antigens recognized by the majority of these 24 samples tested included the 66 kDa (67%), the 40 kDa (50%) and the 29 kDa (71%) antigens. The 43–45 kDa major outer-membrane protein was also widely recognized by antibodies in these sera (62%), although this was usually a weak response. The antibody responses to most of the antigens observed using the follow-up salivary samples were visually weaker. Nevertheless, in the 55 final saliva samples tested IgA antibodies were still detectable directed against the flagellin (60%), MOMP (62%), 47 kDa (50%), 40 kDa (56%) and 29 kDa (24%) antigens.

DISCUSSION In this study we have investigated the kinetics, isotypes and specificity of serum and mucosal antibodies in patients following naturally acquired campylobacter infections. The results confirm [13] that acute campylobacter infections induce specific serum IgG and IgA in the convalescent phase (1 week to 2 months). In general these antibodies declined over time but the IgG antibodies, in particular, remained detectable in sera collected up to one year postinfection. Circulating antibody levels in some patients even increased over time. It is possible that the immune responses of these patients were boosted by subsequent exposure to campylobacters, but no patient follow-up was undertaken to clarify this.

Table 1. The P-values following analysis of variance testing the effects of questionnaire responses on ELISA results for serum IgG and saliva IgA antibodies directed against C. jejuni sonicate and acid-extract antigens

Se IgG anti-R2 son Sa IgA anti-R2 son Sa IgG anti R2 son Se IgA anti R2son Se IgG anti 81116 AE Se IgG anti R2 AE Sa IgA anti 81116 AE Sa IgA anti R2 AE

Sample

Gender

Age

Duration

Hospital

Antibiotic prescribed

Use of antidiarrhoeals

Recent travel

first last first last first last first last first last first last first last first last

0·679 0·425 0·361 0·079 0·761 0·642 0·019 0·428 0·506 0·081 0·676 0·045 0·829 0·825 0·174 0·963

0·177 0·019 0·068 0·654 0·058 0·699 0·014 0·064 0·005 0·012 0·002 0·011 0·089 0·475 0·262 0·356

0·846 0·019 0·144 0·548 0·795 0·842 0·873 0·137 0·635 0·577 0·748 0·610 0·344 0·416 0·873 0·435

0·067 0·001 0·269 0·637 0·156 0·486 0·256 0·223 0·143 0·223 0·127 0·313 0·445 0·400 0·808 0·219

0·700 0·281 0·671 0·931 0·045 0·941 0·967 0·859 0·929 0·828 0·542 0·780 0·752 0·766 0·039 0·929

0·160 0·123 0·564 0·689 0·966 0·614 0·029 0·162 0·836 0·618 0·410 0·072 0·070 0·540 0·453 0·127

0·840 0·431 0·431 0·756 0·465 0·828 0·901 0·946 0·454 0·821 0·306 0·346 0·322 0·891 0·087 0·711

Sa = salivary; Se = serum; son = sonicate; AE = acid extract. P-values in italics = approaching significance (0·1 > P > 0·05). P-values in bold = significant at P £ 0·05.


Antibodies to C. jejuni

Fig. 3. Western blot showing patient antibody responses to C. jejuni strain 81116. Strips 1–5: IgG responses in the first serum samples from five patients; (6) rabbit antimajor outer-membrane protein; (7) IgA response in the first saliva sample from a patient. F indicates flagellin; M indicates the major outer-membrane protein.

To define partly those C. jejuni antigens, which initiated the longer-lived antibody responses, several ELISA antigen preparations were used. The acid extractable material of C. jejuni comprises a number of peripherally associated surface protein antigens including flagella and the PEB 1–4 proteins [14]. By comparison of the data from ELISAs using acid extracts from strain 81116 and its aflagellate mutant, R2, the contribution of antiflagellin antibodies to the overall antibody response was determined. To assay antibodies directed against other, potentially cellular, antigens a whole bacterial sonicate of R2 was used. The results suggested that flagellin was an important antigen in the detection of specific antibodies during the convalescent and postconvalescent phases, but that cellular antigens were also important for the longer-lived responses. This was especially evident for the IgG systemic, and possibly mucosal, responses. The role of systemic antibodies in anti-campylobacter immunity is unknown. C. jejuni is generally considered an enteroinvasive pathogen. However, bacteraemia is documented infrequently, suggesting circulating antibodies may play a role in limiting invasion but not colonization. Because C. jejuni can shed surface material, which may be taken up and presented as antigen at the intestinal mucosal surface, intestinal lumen colonization alone may induce a systemic antibody response. Nevertheless, the presence of serum IgG anti-campylobacter sonicate antibodies, but not anti-acid extract antibodies, in the second serum sample, directly and significantly, correlated with duration of infection and hospitalization, i.e. more serious illness. This may therefore be a reflection of intestinal tissue invasion and translocation by some campylobacter strains. Thus, such long-lasting circulating

105

antibodies may be indicators of both protective immunity and strain virulence. Mucosal anti-campylobacter antibodies have been previously detected in a variety of host samples including milk and faeces [9,15]. In our study both specific salivary IgA and IgG antibodies were detected during convalescence but the salivary IgG antibodies were detectably longer-lasting. Although IgG antibodies are not actively secreted at mucosal surfaces, they can be extruded into saliva from the circulation via the gingival crevices and in Helicobacter pylori infections such antibodies have proved a better indicator of chronic infection than mucosal IgA antibodies [16]. Whether long-lasting serum or salivary antibodies could be used to detect potential immune protection, to inform epidemiological investigations, is debatable. Statistically significant differences were detected between specific IgG and IgA antibodies in the sera of patients, taken up to 2 months postinfection, and the individuals with no recent history of diarrhoeal disease used for comparative purposes. Furthermore, these significant differences remained detectable in the patients’ sera collected 4–12 months postinfection. However, a major problem recognized in this study was the identification of an appropriate control population for such comparative purposes. Unfortunately, the collection of matched samples from uninfected controls was not feasible. The use of a negative control, such as serum from an agammaglobulinaemic patient, may have been useful but may not have reflected the populations potentially under epidemiological investigation. It was also evident that there was considerable variability in the data within the patient and comparison groups. This variability may indicate the prior exposure to campylobacters of some of the individuals in the comparison group and, possibly, the poor immune response of some individuals within the patient group. Such variation is reflected in the observed large standard deviation bars in Figs 1 and 2. Nevertheless, given the sample sizes and the differences between the means of each group, some statistically significant differences were confirmed between the patient and the comparison groups. Such statistically significant differences were largely restricted to longer-lasting serum antibodies of the IgG isotype and to detect these the incorporation of flagellin into the capture antigen was clearly important. Ideally, human sero-epidemiological studies would use clinical material obtained non-invasively. The results from this study indicate that saliva may be useful in non-invasive tests to detect anti-campylobacter antibody responses, particularly of the IgG isotype. However, the efficacy of such assays seems to be very dependent on the antigen used. Interestingly in this case the flagellin antigens were apparently not essential. By Western blotting a number of C. jejuni antigens were identified that elicited salivary and serum antibody responses, which persisted for 4, or more, months postinfection and were, potentially, markers of protective immunity. These antigens included flagellin, the MOMP and a 40 kDa antigen. Both the flagellin and MOMP have been identified previously as immunogens during infection [3] and are potential virulence factors. The immunogenicity of the 40 kDa antigen has been reported previously [10]; however, this antigen has yet to be fully characterized. The antigenicity of flagellin has been investigated in some detail [17,18]. Both conserved- and serotype-specific epitopes are expressed [19]. The Western blots presented here suggest that at least some, presumably conserved, flagellin epitopes are presented, presumably at the intestinal surface, to induce mucosal


106


IgA responses detectable in saliva. Surprisingly, these salivary antiflagellin IgA antibodies did not appear to be detectable using ELISA assays based on flagellin-containing acid extracts. One explanation for this anomaly is that antibodies to flagellin epitopes common to members of the Campylobacter and Helicobacter generi [20] may confound the results. An alternative explanation is the presence of cryptic linear epitopes in acid extracted campylobacter flagellin [21]. The nature of the antigenicity of the MOMP protein during infection is debatable. Previous reports suggest that this protein is an important systemic antigen during human infection [22]. However, in experiments using hyperimmune rabbit sera this protein appears to largely express conformational epitopes [23]. This is largely supported, in the current study, by the relative infrequency of human serum antibodies directed against the MOMP detectable by Western blotting. Interestingly, such antibodies were more frequently detected in the mucosal samples suggesting that some MOMP linear epitopes are presented at the intestinal surface. Such observations suggest a compartmentalization of the systemic and mucosal responses during campylobacter infection. In conclusion, the results of this study indicate that some serum and salivary IgG antibody responses, developed by individuals during, and subsequent to, infection by C. jejuni, can still be detectable up to 1 year postinfection. Long-term specific IgA antibody responses could not be detected. The kinetics of these IgG responses in convalescent patients, and the presence of similar responses in putatively protected individuals in endemically exposed environments [10], are compatible with protective immunity. Several potential marker antigens of long-term immunity have been identified, including flagellin, the MOMP and a 40 kDa antigen. Tests utilizing these antigens may be suitable for assessing the immune status of individuals, although the protective nature of the antibodies detected, as yet, remains unknown.

5

6 7

8

9

10

11

12

13

14

15

16

ACKNOWLEDGEMENTS This work was supported by the Department of Health, United Kingdom, grant number DH 186.

17

18

REFERENCES 1 Tauxe RV. Epidemiology of Campylobacter jejuni infections in the United States and other industrialized nations. In: Nachamkin I, Blaser MJ, Tompkins LS, eds. Campylobacter jejuni: current status and future trends. Washington, DC: American Society for Microbiology, 1992:9– 19. 2 Blaser MJ, Black RE, Duncan DJ, Amer J. Campylobacter jejunispecific serum antibodies are elevated in healthy Bangladeshi children. J Clin Microbiol 1985; 21:164–7. 3 Newell DG, Nachamkin I. Immune responses directed against Campylobacter jejuni. In: Nachamkin I, Blaser MJ, Tompkins LS, eds. Campylobacter jejuni: current status and future trends. Washington, DC: American Society for Microbiology, 1992:201–6. 4 Black RE, Perlman D, Clements ML, Levine MM, Blaser MJ. Human Volunteer Studies with Campylobacter jejuni. In: Nachamkin I, Blaser MJ, Tompkins LS, eds. Campylobacter jejuni: current status and future

19

20

21

22 23

trends. Washington, DC: American Society for Microbiology, 1992:207–15. Russell RG. Campylobacter jejuni colitis and immunity in primates: Epidemiology of natural infection. In: Nachamkin I, Blaser MJ, Tompkins LS, eds. Campylobacter jejuni: current status and future trends. Washington, DC: American Society for Microbiology, 1992:148–57. Advisory Committee on the Microbiological Safety of Foods. Interim report on campylobacter. HMSO, London 1992. Tompkins DS, Hudson MJ, Smith H et al. A study of intestinal infectious disease in England. Microbiol Findings Cases Controls Communicable Dis Public Health 1999; 2:108. Russell RG, Sarmiento JI, Fox J, Panigrahi P. Evidence of reinfection with multiple strains of Campylobacter jejuni and Campylobacter coli in Macaca nemestrina housed under hyperendemic conditions. Infect Immun 1990; 58:2149. Lane EM, Batchelor RA, Bourgeois AL, Burr DM, Olson JG. Urine and faecal IgA response during naturally acquired infection with Campylobacter jejuni. Lancet 1987; i:1141. Cawthraw SA, Lind L, Kaijser B, Newell DG. Antibodies, directed towards Campylobacter jejuni antigens, in sera from poultry abattoir workers. Clin Exp Immunol 2000; 122:55–60. McCoy EC, Doyle D, Burda K, Corbeil LB, Winter AJ. Superficial antigens of Campylobacter (Vibrio) fetus: characterization of an antiphagocytic component. Infect Immun 1975; 11:517–25. Wassenaar TM, Bleumink-Pluym NMC, van der Zeijst BAM. Inactivation of Campylobacter jejuni flagellin genes by homologous recombination demonstrates that flaA but not flaB is required for invasion. EMBO J 1991; 10:2055–206. Blaser MJ, Duncan DJ. Human serum antibody response to Campylobacter jejuni infection as measured in an enzyme-linked immunosorbent assay. Infect Immun 1984; 44:292. Pei Z, Ellison RT, Blaser MJ. Identification, purification and characterization of major antigenic proteins of Campylobacter jejuni. J Biol Chem 1991; 266:16363–9. Ruiz-Palacios GM, Calva JJ, Pickering LK et al. Protection of breastfed infants against Campylobacter diarrhoea by antibodies in human milk. J Pediatrics 1990; 116:707–13. Luzza F, Maletta M, Imeneo M, Doldo P, Marasco R, Biancone L, Pallone F. Salivary specific IgG is a sensitive indicator of the humoral response to H. pylori. FEMS Immunol Med Microbiol 1995; 10:281–3. Nuitjen PJ, van der Zeijst BAM, Newell DG. Localisation of immunogenic regions of the flagellin proteins of Campylobacter jejuni 81116. Infect Immun 1991; 59:1100–5. Nachamkin I, Yang X. Immune response to campylobacter flagellin. In: Nachamkin I, Blaser MJ, Tompkins LS, eds. Campylobacter jejuni: current status and future trends. Washington, DC: American Society for Microbiology, 1992:216–22. Newell DG. Monoclonal antibodies directed against the flagella of Campylobacter jejuni: cross-reacting and serotypic specificity and potential use in diagnosis. J Hyg (Lond) 1986; 96:377. Lee A, Logan SM, Trust TJ. Demonstration of a flagellar antigen shared by a diverse group of spiral-shaped bacteria that colonize intestinal mucus. Infect Immun 1987; 55:828. Newell DG. Monoclonal antibodies directed against the flagella of Campylobacter jejuni: production, characterisation and lack of effect on the colonisation of infant mice. J Hyg (Lond) 1986; 96:131. Blaser MJ, Hopkins JA, Vasil ML. Campylobacter jejuni outer membrane proteins are antigenic for humans. Infect Immun 1984; 43:986. Newell DG, McBride H, Pearson AD. The identification of outer membrane proteins and flagella of Campylobacter jejuni. J Gen Microbiol 1984; 130:1201.
