A Live Diarrheal Vaccine Imprints a Th2 Cell Bias and Acts as an Anti-Inflammatory Vaccine¹

Enterotoxigenic Escherichia coli (ETEC)³ remains problematic in developing countries, particularly for young children who have not acquired immunity and for travelers to endemic areas (reviewed in Ref.1). Enteric symptoms of E. coli infection are often referred to as “traveler’s diarrhea.” E. coli becomes enterotoxigenic upon acquisition of a plasmid or plasmids containing the heat-stable enterotoxin (2) or the cholera-like exotoxin, which is commonly termed the heat-labile enterotoxin (3, 4). Both toxins are responsible for inducing fluid loss and electrolyte imbalance. Virulence by ETEC is also in part contributed by the acquisition of the plasmid for the pili or fimbriae colonization factor Ags, which enhance the colonization of E. coli in the gastrointestinal tract. The ETEC pili are a heterogenous group of fimbrial adhesins and are responsible for adherence to small intestinal epithelium via their fimbriae or long, hair-like projections extending from the bacterial cell surface to epithelial mannose-containing glycoproteins (5). Previous studies show that oral delivery of purified colonization factor Ags fails to induce significant serum IgG or secretory IgA (S-IgA) Abs (6), and these anti-fimbriae Abs (7, 8) fail to protect human volunteers from pathogenic ETEC challenge (7). Thus, live vaccines may be necessary to retain immunogenicity of colonization factor Ags (9).

We have previously reported that oral immunization with a Salmonella vaccine vector expressing the ETEC fimbriae, colonization factor Ag I (CFA/I), elicits high Ab titers, as evidenced by elevated serum IgG1 and mucosal IgA (10). These elevated Ab titers are supported by a biphasic Th cell response in which Th2 cells are rapidly induced and precede the development of the Th1 cells (10). Such findings contrast to immune responses induced by conventional Salmonella vaccine vectors, which generally stimulate Th1 cell-dominating responses toward both Salmonella and passenger Ags (11, 12, 13). These responses are characterized by elevated IFN-γ-regulated IgG2a production and suboptimal mucosal IgA responses. We have also observed that the Salmonella-CFA/I vaccine fails to elicit proinflammatory cytokines upon infection of macrophages (14). Although the isogenic Salmonella vector strain H647 could elicit elevated levels of TNF-α, IL-1, and IL-6, the Salmonella-CFA/I, H696, vaccine remains, as if it were stealth, and fails to elicit these cytokines despite exhibiting identical infectivity to its isogenic Salmonella strain.

Experimental autoimmune encephalomyelitis (EAE) is an animal model of an inflammatory, demyelinating human disease of the CNS, multiple sclerosis (MS; Refs.15, 16, 17). EAE can be induced either by the immunization against specific myelin Ags, including myelin basic protein (MBP), proteolipid protein (PLP), myelin oligodendrocyte glycoprotein, or by passive transfer of activated myelin-specific CD4⁺ Th1 cells (18, 19, 20). Encephalitogenic T cells secrete Th1-type cytokines, IFN-γ and IL-2, and induce local macrophage and microglial activation, infiltration of inflammatory cells from peripheral lymphoid tissue, and demyelination (21, 22, 23). EAE can be suppressed by adoptive transfer of myelin-specific Th2 cells, which secrete IL-4, IL-10, and IL-13 and inhibit Th1 cell development (24, 25, 26, 27). The protective effect of Th2-type cytokines has been also demonstrated by direct delivery into the CNS using retrovirally transduced T cells (28, 29).

These studies clearly indicate that induction of Th2 cell-like responses is a viable strategy for treatment of diseases dominated by Th1-mediated pathophysiology. Several experimental strategies for induction of Th2 cell-dependent protection in EAE have been reported, including oral administration of myelin Ags for oral tolerance (30, 31), immunization using altered peptide ligands bearing encephalitogenic T cell epitopes (32, 33), and immunization with Ags that selectively induce Th2-type responses (34). Low dose tolerance, believed to act by suppressing the induction of encephalitogenic T cells by Th2 cells, has also been adapted for MBP-specific Th2 cells (35). In addition, high dose oral tolerance, believed to result in clonal deletion/anergy of encephalitogenic T cells, has also been tested (36). It has also been observed that MBP-specific Th2 cell responses suppress PLP-specific EAE, suggesting that the bystander suppression alters the T cell microenvironment to become Th2-type dominated (31). Immunization with an altered peptide ligand for PLP_139–151 induces Th2 cells specific for the immunogen and suppresses EAE upon PLP challenge (33). Other myelin Ag-specific EAE challenges have also been suppressed, supporting a role for bystander suppression in this model. Protection to EAE could also be achieved upon immunization with non-self Ags associated with down-regulation of CD4 coreceptor at the time of the EAE onset and/or EAE-specific immune deviation from Th1- to Th2-type by the change in the cytokine microenvironment in which encephalitogenic T cells developed (34, 37). Collectively, these observations clearly provide a precedence for developing strategies for the treatment of autoimmune diseases using non-self Ags to induce bystander Th2 cell-mediated suppression.

Because oral immunization with the Salmonella-CFA/I vaccine induces Th2-type responses, we questioned whether it could provide protection against EAE. The data clearly show that oral vaccination with the Salmonella-CFA/I before EAE induction results in a Th2-type bias in peripheral lymphoid tissues and provides protection against EAE development.

Female SJL/J mice 6–8 wk old were obtained from The Jackson Laboratory. All mice were maintained at Montana State University (MSU) Animal Resources Center under pathogen-free conditions in individually ventilated cages under HEPA-filtered barrier conditions and were fed sterile food and water ad libitum. The mice were free of bacterial and viral pathogens, as determined by Ab screening and histopathologic analysis of major organs and tissues. All experiments adhered to the “Guide for the Care and Use of Laboratory Animals,” prepared by the Committee on Care and Use of Laboratory Animals of the Institute of Laboratory Animal Resources, National Research Council (National Institutes of Health Publication No. 86-23, revised 1985) and were approved by MSU’s Institutional Animal Care and Use Committee.

The ΔaroA Salmonella enterica serovar Typhimurium-CFA/I vector vaccine, strain H696, expressed functional CFA/I fimbriae on the vector’s cell surface (9). This phenotype was maintained by a plasmid bearing a functional asd gene to complement the lethal chromosomal Δasd mutation in the parent Salmonella strain allowing stabilized CFA/I fimbrial expression in the absence of antibiotic selection (10). For each experiment, groups of five female SJL/J mice (five per group), pretreated with an oral 50% saturated sodium bicarbonate solution, received a single oral dose of ∼5 × 10⁹ CFU of the Salmonella-CFA/I construct or its isogenic control strain H647, which lacked the CFA/I operon (10). Control mice received PBS only.

The encephalitognic PLP peptide (PLP_139–151) (HSLGKWLGHPDKF) was synthesized by Global Peptide Services, and HPLC-purified to >90%. One or 4 wk after oral immunization with the Salmonella vaccines or PBS, mice were given s.c. challenges with 200 μl of 100 μg of PLP_139–151 emulsified in a modified Freund’s adjuvant containing 1.5 mg of dead Mycobacterium tuberculosis strain H37RA (Difco Laboratories) per milliliter of IFA. The mice also received i.p. 200 ng of Bordetella pertussis toxin (PT; List Biological Laboratories) on days 0 and 2 relative to the day of challenge. Mice were monitored daily for clinical signs, and clinical scores were assigned as follows (38): 0, normal; 1, a limp tail; 2, hind limb weakness; 3, hind limb paresis; 4, quadriplegia; 5, death.

For histological evaluation of tissue pathology, spinal cords were removed, fixed with neutral-buffered formalin, routinely processed and embedded into paraffin, and sectioned at 5 μm. Cross (transverse) sections from the spinal cord lumbar region were stained with H&E for pathological changes and inflammatory cell infiltration. Adjacent sections were stained with Luxol Fast Blue (LFB; Ref.39) and examined for loss of myelin. Pathological manifestations were scored separately for cell infiltrates and demyelination. Each H&E section was scored from 0 to 4: 0, normal; 1, cell infiltrate into the meninges; 2, one to four small focal perivascular infiltrates; 3, five or more small focal perivascular infiltrates and/or one or more large infiltrates invading the parenchyma; 4, extensive cell infiltrates involving 20% or more of the white matter (38). In each LFB-stained section, myelin was also scored from 0 to 4: 0, normal appearance; 1, one small focal area of demyelination; 2, two to three small focal areas of demyelination; 3, one to two large areas of demyelination; 4, extensive demyelination involving 20% or more of the white matter (38).

To identify infiltrating lymphocytes, immunohistochemistry was performed on cryosections of spinal cord from SJL/J mice between days 11 and 12 after PLP_139–151 challenge from each immunization group. Mice were euthanized, and spinal cords were removed via saline injection into the spinal column. Lumbar regions of cords were embedded in OCT cryoembedding media (Sakura Finetek) and snap-frozen with a dry ice/2-methyl butane slurry (−90°C). Lumbar regions of cords were transversely sectioned at −16°C in a cryostat and mounted on Plus Charge (Erie Scientific). Frozen sections were air-dried overnight at room temperature (RT) and fixed the next day in 75% acetone/25% absolute ethanol mixture for 5 min at RT and rinsed immediately in three changes of Dulbecco’s PBS (DPBS). Appropriate rinsing was done between all immunohistochemistry steps using a rinse buffer (DPBS with 0.2% goat serum and 0.05% Tween 20). Normal spleens were used as positive controls for CD4, CD8, and SK208 staining; for Mac-1⁺ (CD11b) macrophage staining, a Salmonella-infected spleen was used. Endogenous peroxidase was blocked 10 min with Peroxidase Block (no. S2001; DakoCytomation) followed by an endogenous biotin block (according to manufacturer’s instructions, Avidin/Biotin Blocking kit, Vector Laboratories) and a normal serum block (10% goat and 2.5% mouse sera in rinse buffer) for 30 min at RT. The primary Abs were diluted in normal serum block, and all were incubated for 30 min at RT. Biotinylated rat anti-mouse Abs (BD Pharmingen) specific for CD4 (clone GK1.5, IgG2a) and Mac-1 (CD11b, clone M1–70, IgG2b) were used at 1.0 μg/ml. Isotype-matched biotinylated Abs were used as negative controls. To stain for neutrophils, an indirect staining procedure was used in which rat anti-mouse neutrophil (SK208; compliments of Dr. M. Jutila, Montana State University; Refs.40 and 41) mAb was applied to sections as an undiluted hybridoma supernatant for 30 min at RT, followed by the secondary biotinylated F(ab′)₂ of goat anti-rat IgG adsorbed to mouse at 2.0 μg/ml for 30 min. Rat IgG (10 μg/ml; Jackson ImmunoResearch Laboratories) was used as a negative control. Biotinylated Abs were detected with 1.0 μg/ml Streptavidin-HRP (BioSource International/TAGO) in rinse buffer for 20 min at RT. Following a DPBS/0.05% Tween 20 buffer rinse, AEC⁺ chromogen (DakoCytomation) was applied and color developed using microscopic monitoring. Color reaction was halted with DPBS, followed by a water rinse, a light hematoxylin counterstain, and coverslipping with an aqueous mounting media.

CFA/I-specific endpoint titers from dilution of immune sera or fecal extracts were measured by an ELISA, as previously described, using purified CFA/I fimbriae Ag (10) as a coating Ag. Specific reactivity to CFA/1 fimbriae was determined using HRP conjugates of goat anti-mouse IgG-, IgG1-, IgG2a-, and IgA-specific Abs (1 μg/ml; Southern Biotechnology Associates), and ABTS (Moss) enzyme substrate was used to develop the color reaction. The absorbance was measured at 415 nm on a Kinetics Reader model EL312 (Bio-Tek Instruments). Endpoint titers were expressed as the reciprocal dilution of the last sample dilution, giving an absorbance ≥0.1 OD units above the OD₄₁₅ of negative controls after a 1-h incubation.

Lymphocytes from the various tissues (spleens, Peyer’s patches (PP), cervical lymph nodes (CLN), and spinal cords) from the different immunization groups (PBS, H647, and H696) following PLP_139–151 challenge were cultured at 5 × 10⁶/ml in medium alone or in the presence of OVA (10 μg/ml), CFA/1 fimbriae (10 μg/ml), or PLP_139–151 peptide (30 μg/ml) in a total volume of 1.5 ml in a 24-well tissue culture plate. Lymphocytes were cultured for 60 h at 37°C. The culture supernatants were collected by centrifugation and saved at −80°C until assayed. IFN-γ, IL-4, IL-10, and IL-13 were measured on duplicate sets of samples by capture ELISA, as previously described (42). For the IL-13 ELISA, 2.0 μg/ml rat anti-mouse IL-13 mAb (clone 38213) was used as the capture Ab, and 0.2 μg/ml biotinylated goat anti-mouse IL-13 Ab was used as the detecting Ab (R&D Systems). The color reaction was developed using goat anti-biotin HRP conjugated Ab (Vector Laboratories) and ABTS (Moss), as previously described (42). Cytokine concentrations were extrapolated from standard curves generated by recombinant murine cytokines IFN-γ, IL-4, IL-10 (R&D Systems), and IL-13 (PeproTech).

The ANOVA followed by a posthoc Tukey test was applied to show differences in clinical scores in vaccinated vs PBS-dosed mice. The Student t test was used to evaluate differences between variations in Ab titers and cytokine production levels.

To test the hypothesis that bystander suppression by oral vaccination of Th2 cell-promoting Salmonella-CFA/I vaccine could alter the course of EAE development, three groups of SJL/J mice were orally vaccinated with PBS, the isogenic empty Salmonella vaccine vector (H647), or Salmonella-CFA/I (H696). One or 4 wk after oral immunization, each group of mice was challenged with PLP_139–151 and PT. To determine whether mice were effectively immunized by Salmonella-CFA/I, CFA/I-specific copro-IgA and serum IgG Ab endpoint titers at week 2 were measured by ELISA. CFA/I-specific Ab titers (Fig. 1, A and B) were similar to endpoint titers obtained in BALB/c mice as was the dominance of IgG1 subclass responses when compared with IgG2a. The unvaccinated control group showed the expected disease course with a mean day onset of disease of 8.8 ± 1.2 days (Table I) with disease peaking between days 12 and 20 after PLP_139–151 challenge (Fig. 1,A). Mice orally vaccinated with the Salmonella vector showed similar disease kinetics (Fig. 1,C) with a mean day onset of 7.83 ± 1.37 days (Table I). However, while the Salmonella-CFA/I vaccinated mice did not show a significant delay in disease onset (Fig. 1,C and Table I), the overall clinical outcome was significantly milder in these mice (p < 0.001). These data suggest that at the peak of the Th2 cell dominance, previously observed for vaccination with strain H696 (10), the disease course was clearly altered, with these mice recovering by 21 days after EAE induction. Unvaccinated and Salmonella control groups showed disease persistence throughout the observation period.

To determine whether Salmonella-CFA/I was capable of providing protection during the IFN-γ-dominated phase, a second set of experiments was performed in which oral immunization occurred 4 wk before EAE induction (Fig. 1,D). We have previously shown that at this time point, the Th2 cell response is reduced, while the Th1 cell response becomes elevated (10). To test this, SJL/J mice were orally vaccinated with PBS, H647, or H696 strains and challenged (immunized) with PLP_139–151 on day 28 and evaluated for clinical disease for 33 days (Fig. 1,D). As with the 1-wk vaccination studies, the unvaccinated control group developed a similar, but more severe course (Table I). Both the PBS- and H647-vaccinated groups showed a similar mean day onset of clinical disease (Table I and Fig. 1 D). In contrast, the Salmonella-CFA/I-vaccinated groups showed a significantly delayed mean day onset of 10.5 ± 1.5 days when compared with PBS- or H647-dosed mice (p < 0.001). Although the Salmonella-CFA/I-vaccinated mice showed higher clinical scores than mice vaccinated 1 wk before challenge, all the vaccinated mice resolved their EAE by 24 days after challenge. The H647-vaccinated mice, while showing reduced disease compared with the PBS-dosed group, had elevated clinical scores when compared with the H647-dosed mice in the 1 wk postimmunization challenge. Thus, the protective effect of the Salmonella-CFA/I strain persists throughout the Th2 and Th1 cell phases of vaccine-specific immune responses.

Typically, EAE is characterized by CNS inflammatory cell infiltration and loss of myelin. To determine the extent of demyelination and inflammatory cell infiltration among the 1-wk vaccination groups, sections of spinal cord were stained with LFB and H&E. A striking difference in the extent of demyelination was observed among the different treatment groups when compared with normal SJL/J mouse spinal cord (Fig. 2, upper panel). H696-immunized mice showed no demyelination or only mild demyelination (Fig. 2, upper panel). Spinal cord sections from PBS- and H647-immunized mice showed demyelination with the H696-vaccinated mice, revealing the most mild myelin loss compared with PBS- and H647-vaccinated mice (Table II; p < 0.001).

Because demyelination is caused by the inflammatory cells, the degree of inflammatory cell infiltration in the CNS (Fig. 2, lower panel) was evaluated in spinal cord sections stained with H&E (Fig. 2, lower panel). The H696-vaccinated group showed little or no inflammation similar to that observed in normal mice. By contrast, PBS- and H647-vaccinated mice had clear inflammatory infiltrates in the meninges and CNS parenchyma regions. The degree of cellular infiltration in the spinal cords is summarized in Table II. The H696-vaccinated group scored 0.4, which represented either no infiltration or minimal meningeal infiltration. PBS- and H647-vaccinated groups had scores in the 2.5–2.7 range, indicating between one to four small perivascular infiltrations and more than four small perivascular/or large parenchymal invasions. Thus, protection from clinical disease in H696-vaccinated mice was clearly associated with reduction in inflammatory CNS infiltrates.

To determine the general composition of the inflammatory infiltrates, immunohistochemistry was performed to identify cell types typically present in EAE tissue (Fig. 3). Infiltrates in spinal cord sections from PBS-vaccinated mice consisted of CD4⁺ T cells, Mac-1⁺ cells, and SK208⁺ neutrophils. By contrast, spinal cords from control vector-vaccinated mice showed staining for a small number of Mac-1⁺ cells and only a few neutrophils. Finally, CD4⁺ T cells and SK208⁺ neutrophils were absent from infiltrates in Salmonella-CFA/I-vaccinated mice; only a few Mac-1⁺ cells were detected. These data suggested that protection was due to suppressed development of myelin-specific, pathogenic Th1 cells and/or prevented migration of pathogenic T cells to the CNS.

The participation of Th1/Th2-type cytokines in the development of organ-specific autoimmune disease has been well-documented (43, 44). To determine whether the protective effects of Salmonella-CFA/I vaccination are Ag-specific, cytokine responses were measured ex vivo in lymphocytes isolated from spleens, PP, CLN, or spinal cords from vaccinated and unvaccinated mice, and pulsed in vitro with OVA, CFA/I fimbriae, or PLP_139–151 (Fig. 4). Unstimulated lymphocytes were included as controls. As we previously observed (10), cytokine responses were clearly Th2 cell-dominant in H696-vaccinated mice, significantly exceeding IFN-γ production (Fig. 4). Not having previously measured IL-13, we observed considerable enhancement in H696-vaccinated mice, but not in the PBS-dosed mice. Some splenic and CLN IL-13 was induced in mice orally vaccinated with the Salmonella vector, but not to the levels observed with H696-vaccinated mice. Upon restimulation with PLP_139–151 peptides, the IFN-γ levels in H696-vaccinated mice were significantly reduced in all tissues examined when compared with unvaccinated controls or H647-vaccinated mice. The major production of PLP_139–151-specific IFN-γ was primarily observed with CLN from PBS- and H647-vaccinated mice, suggesting that this was the site of T cell activation. Conversely, the PBS- and H647-vaccinated mice showed minimal to no Th2-type cytokine production in any of the tissues examined. These data suggest that the Salmonella-CFA/I vaccine imparted a Th2-type bias and, in an Ag-independent fashion, biased the development of PLP_139–151 CD4⁺ T cells to becoming Th2- type. Thus, oral immunization with Salmonella-CFA/I vaccine induced PLP_139–151-specific immune deviation for ultimate protection against EAE.

Previous work shows that Salmonella-CFA/I vaccine fails to elicit proinflammatory cytokines from infected macrophages, unlike its isogenic strain which, in a dose-dependent fashion, induces the expected elevations in TNF-α, IL-1α, IL-1β, and IL-6 (14). It is apparent from this study that the presence of the CFA/I fimbriae impedes the expected proinflammatory response. This lack of an inflammatory response is not attributed to increased production of IL-10 or IL-12p40 because these did not significantly differ from those levels induced by the Salmonella vector strain (14). Further evidence that the CFA/I fimbriae alters Salmonella’s behavior is supported by in vivo studies in which CFA/I fimbria-specific Th2 cells were dominated by IL-4, IL-5, IL-6, and IL-10 and accompanied by elevated serum IgG1 and mucosal IgA Ab titers (10). We hypothesized that this property of the Salmonella-CFA/I vaccine could be exploited to develop an anti-inflammatory vaccine.

To test this hypothesis, EAE-susceptible SJL/J mice were orally vaccinated with Salmonella-CFA/I vaccine, and the clinical and histopathological disease was compared with mice receiving either no vaccine (PBS) or its isogenic construct (lacking the CFA/I operon). We first questioned whether SJL mice would make an appropriate immune response to the vaccine, as previously shown for BALB/c (10) and B6 mice (40). The data clearly indicated that SJL mice are vaccine responders generating a robust anti-CFA/I Ab response that is similar in magnitude and kinetics to those exhibited by similarly vaccinated BALB/c and B6 mice with the similar biases in serum IgG1 and IgG2a. In addition, the studies clearly showed that Salmonella-CFA/I-vaccination significantly reduces clinical and histopathological EAE in PLP_139–151-immunized mice. This was evidenced by a shorter duration of acute disease and lower peak clinical scores and recovery. By contrast, vaccinated mice and mice receiving control vaccination showed poor disease recovery throughout the 33-day observation period. Although the severity of EAE observed in mice receiving Salmonella-CFA/I vaccine 4 wk before PLP_139–151 immunization was worse than that observed in mice with EAE induced 1-wk postvaccination, all vaccinated mice in both vaccination regimens recovered completely. It was not surprising that the Salmonella-CFA/I-vaccinated mice in the 1-wk regimen showed less severe clinical disease, because the Th2 cell response is dominant during the early stages of Salmonella inflection, and CFA/I fimbria-specific Th1 cells generally do not peak until 3–4 wk postinfection (10). We also noted a difference in the magnitude of EAE in PBS-dosed mice, which may be related to the differences in age of the challenged mice. Such differences may be attributed to age as others have previously suggested that EAE is age-dependent (45, 46).

Surprisingly, mice vaccinated with the Salmonella vector strain H647 also showed reduced EAE in both dosing regimens, which was a consistent finding in four of four experiments conducted independently. In addition, this occurred in the presence of an intact IFN-γ response and in the absence of a clear Th2 cell deviation. An explanation for this finding may be that proinflammatory responses induced by Salmonella infection alone may be competing with PLP_139–151-induced inflammation. In all four experiments, this was consistently observed. Similar observations have been made for EAE protection (47) and treatment of relapsing-remitting MS in humans (48) following vaccination with bacillus Calmett-Guérin. Bacillus Calmett-Guérin, a Th1-type promoting bacterium, caused the formation of granulomas in spleen and liver of infected mice, and the investigators demonstrated that myelin oligodendrocyte glycoprotein-specific Th1 cells were recruited into the liver granulomas rather than to the CNS (47). This suggests that vaccination with Th1-inducing microbial agents may redirect a Th1 cell response away from an Ag-targeted site of inflammation. Additional studies are warranted to directly address this possibility in the Salmonella-CFA/I vaccination model.

In addition to reduced clinical disease, mice receiving oral Salmonella vaccine also exhibited clear evidence of reduced tissue pathology. Salmonella-CFA/I-vaccinated mice showed minimal demyelination and inflammatory cell infiltration during peak EAE, and the composition of inflammatory infiltrates differed significantly among the three experimental groups. CD4⁺ T cells and Mac-1⁺ cells and some neutrophils dominated the spinal cord infiltrates in PBS-vaccinated mice, while infiltrates in Salmonella vector-vaccinated mice consisted almost exclusively of Mac-1⁺ cells (not neutrophils). Thus, a reduction in T cells may account for the observed reduced pathology in Salmonella-vector vaccinated mice. The lack of T cells in spinal cord infiltrates limited the ability to study T cell cytokines in the CNS of Salmonella-CFA/I-vaccinated mice. This evidence supports the notion that oral vaccination with Salmonella-CFA/I vaccine can reduce peripheral or distal inflammation in a tissue other than the mucosal compartment. This ability to reduce inflammation in the CNS occurred in an Ag-independent fashion because there was no primary structural homology between PLP_139–151 and CFA/I fimbria.

Because no primary structural homology could be identified between the PLP_139–151 and CFA/I fimbrial subunit, we hypothesized that the mechanism for the observed efficacy against EAE involved immune deviation. The presence of CFA/I fimbriae alters the host recognition of the Salmonella vaccine, and it is evident that a potent Th2 cell response was induced despite the proinflammatory challenge produced by the EAE regimen. In this context, the development of inflammatory, pathogenic PLP_139–151-specific Th cells was diminished, indicated by the reduced PLP_139–151-specific IFN-γ-producing T cells in Salmonella-CFA/I-vaccinated mice. Instead, PLP_139–151-specific Th2 cells were induced and sustained for at least 4 wk. In contrast, the PBS- and Salmonella vector-immunized mice showed elevated levels of IFN-γ and minimal to no production of Th2-type cytokines. Thus, it appears that the Salmonella-CFA/I vaccine altered the microenvironment, possibly in the draining lymph nodes of the head and neck, to prevent the development of encephalitogenic Th1 cells.

We were also surprised to find the quite robust IL-13 responses by Ag-restimulated lymphocytes because we had not previously evaluated this Th2-type cytokine. IL-13 is a signature Th2-type cytokine that exhibits functions similar to those observed for IL-4. It was evident in this study that while IL-13 was induced against the CFA/I fimbriae, the majority was induced by PLP_139–151, especially in the CLN. It has been reported that IL-13 can suppress EAE, using transfection methods (49) or adoptive transfer of Th2 cells (26), and it has been suggested that IL-13 suppresses EAE by inhibiting the production of proinflammatory cytokines, such as IL-1β and TNF-α, from macrophages/monocytes (50, 51) or NO by glial cells (52). Although IL-13 and IL-4 share a receptor, the ability of IL-13 to suppress EAE appears distinct from that of IL-4 (53). In our studies, the concomitant induction of IL-4 and IL-13 by oral vaccination with a live vaccine suggests a clear advantage for our Salmonella-CFA/I vaccination strategy to prevent inflammatory autoimmune diseases.

In addition to the stimulation of IL-4 and IL-13, the Salmonella-CFA/I vaccine also stimulates production of IL-10 (10), which has a well-documented beneficial impact on EAE (25, 26, 54, 55, 56). IL-10 delays disease onset and reduces clinical disease when directly administered at the time of disease induction (54) and in transgenic mice overexpressing IL-10 (25, 55). Moreover, targeted disruption of IL-10 results in loss of EAE suppression (25, 55). IL-10 acts during the induction, rather than in the effector stage of disease, because it fails to suppress EAE induced by adoptive transfer of myelin-specific cells (57). Administration of IL-10 may actually be detrimental in the adoptive transfer model (57). In our study, IL-10 was already present at the time of disease induction and may have contributed to deviation of PLP_139–151-specific CD4⁺ T cells toward a Th2 cell phenotype, either independently or in conjunction with IL-4 and IL-13.

Data from other infection models suggest that preconditioning the host with a Th2-inducing organism may bias and alter the development of encephalitogenic T cells. Infection with the Schistosoma mansoni worm initially produces a Th1 cell response, which then shifts to Th2-type response during egg production, implying that the eggs are responsible for inducing Th2-type responses (58). However, the mechanism for protection by S. mansoni differs between a live infection or exposure to eggs alone. S. mansoni infection results in the down-regulation of IFN-γ and IL-12p40, but leaves IL-10 and IL-4 unchanged, suggesting mechanisms other than Th2 cell activation in EAE protection and possibly involving macrophages (59). By contrast, immunization with S. mansoni eggs results in EAE protection that is accompanied by the down-regulation of IFN-α and increases production of IL-4, IL-10 (59, 60), and TGF-β (60). Consequently, adapting the host to a Th2 cell-dominated environment by helminth infection is currently being considered as a strategy for the treatment of MS (61).

In conclusion, this study has shown that oral administration of an anti-inflammatory vaccine, Salmonella-CFA/I, is capable of preventing the development of the inflammatory murine disease, EAE, in an Ag-independent fashion. Expression of the ETEC fimbriae is necessary because mice receiving the “naked” Salmonella vector showed persistent clinical disease throughout the observation period. Future studies will address the ability of this vaccination strategy to induce long-term changes in cytokine environment and whether it can treat ongoing disease in a treatment paradigm.

We thank Nancy Kommers for her assistance in preparing this manuscript and Dr. Mark A. Jutila, Veterinary Molecular Biology, Montana State University, for kindly providing the SK208 mAb.

The authors have no financial conflict of interest.