Qualitative and Quantitative Analysis of the Local and Systemic Antibody Response in Mice and Humans with Helicobacter Immunity and Infection

Abstract

Immunization can prevent or cure an otherwise chronic gastric Helicobacter infection in several different animal models. The goal of the present study was to compare the titers and specificities of local and systemic antibody responses generated by Helicobacter infection and immunization. Protective immunization results in levels of specific gastric antibody significantly lower than induced by infection. However, antibodies from protectively immunized mice preferentially recognize immunodominant proteins of 10–22 and 30 kDa. Immunoblot analysis of infected mice and humans demonstrated that the serum IgA, but not serum IgG, binding profiles yield an accurate profile of the antigenic specificity of the host's gastric IgA. Therefore, serum IgA may be useful in evaluating the immunodominant antigens at the gastric mucosa of infected persons and possibly in determining the immunogenicity of orally applied Helicobacter vaccines.

Colonization of the gastric antrum by the gastric pathogen Helicobacter pylori is accompanied by a serum and a local antibody response [1–3]. Although a Helicobacter infection endures for the life of the host, oral immunization can prevent or cure chronic Helicobacter infections in several animal models of Helicobacter infection [4]. Thus, whereas the immune response induced by infection is not adequate to eradicate the organism, immunization can prevent and/or clear the infection. Our goal in this study was to make a systematic quantitative and qualitative comparison of the humoral immune response to Helicobacter organisms induced via infection or protective immunization and to identify any antigens that might be preferentially recognized by protected animals. Defining these differences may help elucidate the mechanism(s) by which protective Helicobacter immunity is achieved and possibly define markers for protective immunity.

Materials and Methods

Patients

The study group consisted of 53 patients (26 men and 27 women, 23–79 years old) with dyspepsia or peptic ulcer disease referred for gastrointestinal endoscopy. Of these, 37 were H. pylori—positive and 16 were negative by culture.

Mice

Six-week-old C57BL/6 mice (Taconic Farms, Germantown, NY) were housed in microisolator cages and fed autoclaved laboratory chow and water ad libitum.

Bacteria

Helicobacter felis was isolated and characterized as previously described [5]. H. pylori was isolated from a patient presenting with peptic ulcers and characterized on the basis of colony and bacterial morphology, Gram's stain, and the production of urease, catalase, and oxidase. Bacterial cultures were maintained as previously described [5, 6].

Helicobacter lysates

Bacterial lysates for immunization, ELISA, and immunoblotting were prepared as previously described [7]. Briefly, Helicobacter organisms were harvested from liquid culture by centrifugation and resuspended in 2 mL of PBS per 200 mL of bacterial culture. Bacteria were lysed by sonication (Sonics and Materials, Danbury, CT), and any remaining whole organisms were removed by centrifugation and filtration through a 0.22-μM—pore filter. The concentration of protein was determined by the Lowry assay [8].

Immunization and challenge of mice

All immunizations, inoculations, and challenges were delivered by gastric intubation using flexible tubing on the end of an 18-g needle. Chronic infection was established with 10⁸H. felis on 2 consecutive days. Challenge of immunized mice was with 5 × 10⁶H. felis 7 days after the final immunization. Oral immunizations consisted of four weekly doses of 2 mg of H. felis lysate—soluble antigen combined with 10 μg of cholera toxin (List Biological Laboratories, Campbell, CA).

Diagnosis of infection

Patients and mice were considered infected if their biopsy specimens were positive by the urease test or if organisms were detected in histologic sections. For urease activity, two biopsy specimens from each patient or mouse were placed in 0.5 mL of Stuart's urease test broth [9]. Colonization by Helicobacter was confirmed by a change in broth color from orange to red within 24 h. Urease tests were simultaneously performed on naive mice as controls for bacterial flora that might generate false positives. For histologic evaluation, tissue was surgically removed from the greater curvature of the stomach, from the duodenum to the gastric cardia, of each mouse. Human biopsies were taken from the gastric antrum during endoscopy. Tissues were fixed in formalin and processed for histologic examination. Paraffin-embedded sections were mounted on slides and silver-stained by the Steiner method to facilitate identification of Helicobacter. Patients were also considered positive if H. pylori could be cultured from gastric biopsies. Biopsies were homogenized in 200 μL of Brucella broth, and the homogenate was spread on Columbia agar plates containing 7% horse blood. Cultures were grown and identified as previously described [5, 6].

Sample collection for antibody measurement

Serum and gastric juice (neutralized with an equal volume of Tris-HCl [pH 7.4]) were obtained from fasting patients at the time of endoscopy. Mouse serum was obtained 1 day before necropsy by tail bleed. Mouse saliva and gastric and intestinal mucus were collected using polywicks (Polyfiltronics, Rockland, MA) [10]. Antibodies were extracted from wicks by vortexing each in 300 μL of protease inhibitor solution for 30 s [10]. Wicks were then placed in 500-μL microfuge tubes with a hole bored through the bottom, placed inside 1.5-mL microfuge tubes, and centrifuged at 12,000 g to extract the sample. The extracted volume was combined with the original 300 μL of protease inhibitor solution [10]. Intestinal washes were obtained by rinsing the small intestine with 2 mL of cold protease inhibitor solution. Gastric washes were collected by rinsing the stomach with 300 μL of cold protease inhibitor solution. All samples were stored at −20°C until assayed.

ELISA and immunoblotting

ELISAs were performed as previously described [5–7] using serial log₁₀ dilutions of each sample. Titers were defined as the highest log₁₀ dilution generating a 450 nm reading of 0.05 optical density units above background. Immunoblotting was performed using Helicobacter whole cell lysates resolved on 12.5% polyacrylamide gels electrotransferred to Immobilon P membranes (Millipore, Bedford, MA) and blocked with 5% skim milk. Membranes were washed three times in phosphate buffer for 5 min between each subsequent step. Individual lanes were incubated with 1.5 mL of sample diluted in 5% skim milk for 1 h and then incubated with isotype-specific, alkaline phosphatase—conjugated goat antibody (see ELISA) diluted 1:500 in 5% skim milk for an additional hour. Membranes were developed with 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium substrate tablets.

Statistical analysis

The presence or absence of experimental infection among groups of mice was evaluated by χ² analysis. Data representing antibody titers are expressed as means ± SE of the mean, and comparisons between experimental groups were evaluated by the two-tailed Student's t test.

Results

Gastric IgA levels are significantly lower in immunized mice than in infected mice

Serum, intestinal, gastric, and salivary H. felis—specific antibody titers were determined for two separate experiments, in which mice were either infected with H. felis, orally immunized, or orally immunized and challenged with H. felis. Both infection and immunization generated significant anti-Helicobacter titers in intestinal secretions (1.7 ± 0.5–2.7 ± 0.4) and serum (3.2 ± 0.3-3.9 ± 0.1). However, the most notable observation was that oral immunization generated prechallenge gastric IgA titers (0.7 ± 0.3–1.3 ± 0.6) significantly lower than those induced by a chronic infection (1.9 ± 0.6-2.8 ± 0.3), as measured in either gastric washes (P <.05) or gastric mucus collected with absorbent wicks (P <.002). Challenge of immunized animals resulted in a significant rise of H. felis—specific gastric IgA, compared with prechallenge titers (P <.01) measured in gastric washes. These titers were not significantly different from those in mice chronically infected with H. felis. Despite the low prechallenge titers in immunized mice, a separate group receiving the same immunization were protected: 88% (7/8) of immunized mice were H. felis—negative 6 weeks after challenge, as determined by urease negativity and the absence of bacteria in silver-stained histologic sections. In contrast, despite their ability to mount a robust gastric antibody response, 100% (5/5) of naive mice became infected with H. felis when given the same infectious dose.

Gastric IgA from infected and protectively immunized mice differ with respect to antigenic specificity

Immunoblots were performed to compare the specificity of the antibodies between the 2 groups. Figure 1 shows that representative samples of gastric washes from immunized/challenged mice had an immunoblot profile distinct from that of infected mice. Although many common antigens were recognized by both groups, gastric IgA from mice that had been immunized with whole cell lysate prior to challenge recognized a distinct set of dominant antigens of two molecular weight ranges: a group of lower molecular weight proteins in the range of 10–22 kDa, and an antigen of ∼30 kDa. Conversely, the gastric IgA from infected mice specifically recognized several antigens between 30 and 66 kDa that were not immunodominant in immunized/challenged animals. The immunoblot profiles shown in figure 1 were representative of their respective groups (n = 8/group)

Figure 1

Detection of specific H. felis antigens by gastric IgA from mice immunized with H. felis lysates plus cholera toxin and then challenged with live H. felis (lanes a–d) or mice infected with H. felis for 6 weeks (lanes e–h); 20 μg of whole cell sonicate antigen per lane was used for blotting.

Serum IgA reflects the antigenic specificity of gastric IgA in H. pylori—infected patients and mice experimentally infected with H. felis

Intestinal, serum, and gastric antibodies from the infected mice were compared by immunoblot. Although the specificities of serum and gastric IgA matched almost completely, intestinal antibodies recognized few of the same proteins (figure 2A). IgG was not present in the intestinal washes, and only several bands were visualized by gastric IgG (data not shown). Serum IgG, however, recognized many bands that were not recognized by either gastric IgG or IgA (data not shown). Therefore, serum anti-H. felis IgA appears to be a surrogate marker for the specificity of the gastric IgA response of infected mice. Similar observations were made for human samples. The antigenic specificities of human serum IgA correlated well with gastric IgA, whereas human serum IgG recognized many bands not recognized by the antibodies present in the stomach (figure 2B). Therefore, serum IgA may be a good indicator in both mice and humans of the specificity of the gastric antibodies against Helicobacter infections. The samples shown are representative of their respective species.

Figure 2

Comparisons of specificities of local and systemic antibodies from Helicobacter-infected hosts. A, Serum (lane a), gastric (lane b), and intestinal (lane c) IgA from representative mouse 6 weeks after inoculation with H. felis. B, Gastric IgA (lane d), serum IgA (lane e), gastric IgG (lane f), and serum IgG (lane g) from representative H. pylori—positive patient; 20 μg of whole cell sonicate from either H. felis or H. pylori was used as antigen per each lane for blotting.

Discussion

The present study was designed to examine the local and systemic antibody responses that accompany immunization and infection and to identify any potentially important differences. Our data reveal several new aspects of the gastric humoral response to Helicobacter infection.

First, unlike serum and intestinal antibody, prechallenge gastric IgA titers of immunized mice are significantly lower than those of infected animals. The presence of gastric IgA prior to challenge is attributable to local production, because exhaustive washing of the mucosa to remove ingested saliva did not diminish the amount of IgA collected from the mucus with absorbent wicks. Although there is a significant rise in specific gastric IgA accompanied by challenge of protectively immunized mice, the levels induced are no higher than those induced during natural infection.

Second, although there are many similarities, the specificity of local IgA generated in response to oral immunization is qualitatively different from IgA induced by natural infection. The antigens used for immunization consisted of filtered H. felis sonicates. Thus, although the total antigenic make-up of our immunogen is similar to the antigens potentially seen by the host during a natural infection, these studies demonstrate that orogastric immunization induces the production of local antibodies with antigenic specificities different from those induced by infection. Therefore, the specificity rather than the magnitude or even preexistence of gastric antibodies may be an important element in protective immunity.

Finally, in infected mice and humans, the antigenic specificity of serum IgA, but not serum IgG, reflects the specificity of gastric IgA. Although the presence of serum IgG antibodies is a good marker for current or previous Helicobacter infection, the specificity of serum IgG does not appear to be a good indicator of the specificity of gastric antibody. Serum IgG exhibited a broad spectrum of antigenic specificities not observed for either gastric IgG or IgA. In support of this, it has been shown that gastric antibodies from infected patients do not bind to immunodominant antigens at either 66 or 31 kDa [1], although many H. pylori—infected patients have a serum IgG response against the urease subunits [11]. Therefore, serum IgA specificities may be useful to evaluate immunodominant antigens at the gastric mucosa and, possibly, to determine the immunogenicity of orally applied Helicobacter vaccines.

Presented in part: Digestive Disease Week; 96th Annual Meeting of the American Gastroenterological Association, May 1996, San Francisco.

Informed consent was obtained from all patients or parents of patients whose information is contained within this report. All animal handling and experimentation was reviewed by and performed according to the guidelines of the Institutional Animal Use and Care Committee of Case Western Reserve University. The Case Western Reserve University animal facility is fully accredited by the American Association for Accreditation of Animal Care. This study was approved by the Investigational Review Board of University Hospitals of Cleveland.

Grant support: NIH (DK-46461, AI-40701).