Acellular Bordetella pertussis vaccine enhances mucosal interleukin-10 production, induces apoptosis of activated Th1 cells and attenuates colitis in Gαi2-deficient mice

Abstract

Mice deficient for the inhibitory G protein subunit α2 (Gαi2^−/−) spontaneously develop a progressive inflammatory bowel disease resembling ulcerative colitis, and have a T helper 1 (Th1)-dominated immune response prior to onset of colitis, which is further augmented after the onset of disease. The present study was performed to investigate whether the Gαi2^−/− mice were able to down-regulate the Th1-dominated inflammatory mucosal immune response and/or induce an anti-inflammatory Th2/T regulatory response and thereby diminish the severity of colitis following treatment with acellular Bordetella pertussis vaccine. The acellular vaccine against B. pertussis, the causative agent of whooping cough, has been demonstrated to induce a Th2-mediated response in both man and mice. We therefore treated Gαi2^−/− mice intraperitoneally with a three-component acellular B. pertussis vaccine. The treated Gαi2^−/− mice showed significantly increased interleukin (IL)-10 production in intestinal tissue, associated with significantly reduced colitis and decreased mortality, compared to untreated Gαi2^−/− mice. The attenuation of colitis in Gαi2^−/− mice was due, at least partly, to the B. pertussis surface antigen filamentous haemagglutinin (FHA), which almost completely inhibited proliferation of CD4⁺ T cells and stimulated apoptosis of activated CD4⁺ T helper 1 cells. In conclusion, the three-component acellular B. pertussis vaccine containing filamentous haemagglutinin increases the production of IL-10 in the intestinal mucosa, induces apoptosis of activated Th1 cells and attenuates colitis in Gαi2^−/− mice.

Introduction

Mice with a targeted deletion of the G-protein αi2 subunit (Gαi2) develop a wasting disease, manifested by diarrhoea and bloody stools, clinically and histopathologically closely resembling ulcerative colitis in humans [1]. The idiopathic inflammatory bowel diseases (IBD), ulcerative colitis (UC) and Crohn's disease (CD), represent a group of intestinal inflammatory diseases of unknown aetiology and pathogenesis. Whereas Crohn's disease has clearly been shown to be a T helper 1 (Th1)-mediated disease [2–5], the mechanisms driving the inflammation in UC is less evident [6,7]. Although the initiating mechanism in different forms of IBD might differ, some of the downstream inflammatory pathways appear to be shared. In both CD and UC there is an enhanced synthesis of proinflammatory cytokines including interleukin (IL)-1β, IL-6 and tumour necrosis factor (TNF)-α and an influx of non-specific cells into the mucosa [8]. Also, most animal models of IBD are mediated by a Th1 cell response [9].

Previous studies in Gαi2^−/− mice have established that mice that have not yet developed colitis have increased levels of Th1 cytokines in the large intestine, although not as high as in animals with established colitis [10,11]. Thus, activation of the intestinal immune system in Gαi2^−/− mice precedes histopathological and clinical signs of inflammation, suggesting that immune abnormalities, including dysregulation of the balance of Th1 and Th2 cytokine responses, play an important role in the induction of Th1 cytokine-mediated colitis in this model.

Bordetella pertussis is a Gram-negative bacterium, causing whooping cough. It colonizes the human respiratory tract and secretes various virulence factors, i.e. filamentous haemagglutinin (FHA) and pertactin (PRN), which mediate adherence to human epithelial cells and monocytes [12,13]. FHA is associated with the outer membrane of B. pertussis and is secreted into culture medium during growth, and serves as a ligand for at least two leucocyte integrins, leucocyte response integrin (LR1) and complement receptor 3 (CR3)(CD11b/CD18) [14–16]. PRN is also associated with the bacterial outer membrane, but its ligand still remains unidentified.

Previously it was shown that mice immunized with an acellular B. pertussis vaccine consisting of detoxified B. pertussis toxin (PTd), FHA and PRN, generated a Th2 response [17], which remained as long as 6 months after immunization [18]. Furthermore, patients treated with a three-component acellular B. pertussis vaccine demonstrated high levels of the Th2 cytokine IL-5, but no detectable levels of the Th1 cytokine interferon (IFN)-γ or IL-2 [19]. In addition, non-obese diabetic (NOD) mice treated with a diphtheria–tetanus toxoid—acellular Pertussis vaccine were significantly protected from developing diabetes [20].

This report demonstrates that acellular B. pertussis vaccine significantly attenuates colitis in an animal model of IBD. Treatment of Gαi2^−/− mice with a three-component acellular B. pertussis vaccine, consisting of PTd, FHA and PRN adsorbed in alum, resulted in significantly increased levels of IL-10 accompanied by less mortality and attenuation of colitis. Studies both in vitro and in vivo demonstrated that this attenuated colitis was due, at least partly, to FHA-induced apoptosis of activated CD4⁺ Th1 cells.

Materials and methods

Mice

Gαi2^−/− mice [1] were bred and kept at the Department of Experimental Biomedicine, Göteborg University. Homozygous Gαi2 mutant males on a 129SvEv × C57BL/6 background were bred with heterozygous females, and the offspring were genotyped by polymerase chain reaction (PCR) analysis. One hundred percent of the Gαi2^−/− mice on this background develop colitis, and have to be sacrificed due to severe disease between 15 and 25 weeks of age. Wild-type mice, continuously derived by breeding heterozygotes from the littermates above, do not develop colitis and were used as controls where indicated to verify that any effects of the pertussis vaccine is universal and not due to the G protein deficiency.

Ovalbumin (OVA) T cell receptor (TCR)-transgenic mice, clone DO11·10, on a BALB/C background, recognizing the 323–339 peptide fragment of OVA were used at 10–14 weeks of age for the generation of Th1 and Th2 cell lines. The animal facility is kept pathogen-free using microisolator cages and sterile workbenches, and mice are monitored routinely by health screening according to FELASA recommendations. All animal experiments were approved by the local ethics committee in Göteborg.

Treatment of Gαi2^−/− mice with B. pertussis vaccine

Gαi2^−/− mice were treated either with a three-component acellular B. pertussis vaccine, consisting of 25 µg formaldehyde—glutaraldehyde treated pertussis toxoid (PTd), 25 µg formaldehyde-treated filamentous haemagglutinin (FHA) and 8 µg formaldehyde-treated pertactin (PRN) per human dose (0·5 ml) (SmithKline Beecham Biologicals, Rixenart, Belgium), or a one-component acellular vaccine consisting merely of 40 µg hydrogen peroxide-treated pertussis toxoid per human dose (Statens Serum Institute, Copenhagen, Denmark). Both vaccines also contain 0·5 mg aluminium hydroxide as an adjuvant. The mice were given 0·2 human doses (0·1 ml) (equal to 5 µg PTd, 5 µg FHA and 1·6 µg PRN) of the vaccine intraperitoneally (i.p.) at 6, 10 and 14 weeks of age, and were killed at 19 weeks of age. Additionally, Gαi2^−/− mice were treated with 10 µg ammonium sulphate precipitated FHA (Institut Mériux, France) i.p. twice a week for 15 weeks, starting at 4 weeks of age, without the co-administration of any adjuvant. In experiments analysing the immunoregulatory role of FHA and PRN, Gαi2^−/− mice were treated i.p. with 5 µg/ml FHA or PRN (Institut Mériux, France). The lipopolysaccharide (LPS) content was < 7 ng/mg protein in purified FHA and PRN according to the Limulus assay. Both preparations were free from other B. pertussis proteins as judged by analysis on sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). Untreated Gαi2^−/− mice and wild-type mice were used as controls, as indicated in the Results section.

Measurement of cytokine production from intestinal tissue

After careful removal of Peyer's patches and fecal contents, the small and large intestines were opened lengthwise and cut into 50-mg pieces. The tissue pieces were washed extensively in phosphate buffered saline (PBS) and poured over nylon net to retrieve the tissue. The intestinal pieces were incubated in 24-well plates at 37°C, 5% CO₂ in 500 µl of Iscove's medium (Seromed, Berlin, Germany) containing 10% heat-inactivated fetal calf serum (FCS), 5 × 10⁻⁵ M 2-mercaptoethanol, 1 mMl-glutamine and 50 µg/ml gentamycin (hereafter referred to as Iscove's total medium). After 24 h of culture, the supernatants were collected and stored at −70°C until analysis of cytokine content. The spontaneous production of IL-10 and IFN-γ was determined by PharMingen OptEIA^TMSet cytokine kits (BD Biosciences, San Jose, CA, USA), according to the manufacturer's recommendations.

Clinical evaluation of colitis

A blinded clinical evaluation of the colitis score of Gαi2-deficient mice was quantified by assessing the manifestation of hunch back, stool consistency, weight loss and thickening of the colonic wall: Grade 1, no clinical signs of colitis; Grade 2, no visible hunch back, slight diarrhoea, no or only minor (< 2%) weight loss and minor thickening of the colonic wall; Grade 3, no visible hunch back, diarrhoea, moderate weight loss (< 5%) and thickening of colonic wall; Grade 4, visible hunch back, diarrhoea, substantial weight loss (5–10%) and thickening of the entire colonic wall; and Grade 5, severe hunch back, diarrhoea, extensive weight loss (> 10%) together with major thickening of the entire colonic wall. Gαi2^−/− mice that died due to severe colitis during the experiment were considered equivalent to the colitis score 5.

Histological evaluation of colitis

Tissue samples, 5–7 mm, from the small and large intestinal mucosa were fixed in 10% formalin, dehydrated in gradients of alcohol, embedded in paraffin, cut at 4 µm and stained with haematoxylin-eosin in a conventional manner [21]. Evaluation of the intensity of the inflammation was carried out according to previously established criteria [22]. Briefly, the colonic tissue was graded by the following scale: Grade 1, normal mucosa; Grade 2, enhanced glands with intraepithelial granulocytes, enhancement of lymphocytes and/or eosinophils in the stroma; Grade 3, goblet cell depletion, reduced mucin production in some glands, marked increase of inflammatory cells in the stroma; Grade 4, marked gland and mucosal atrophy, crypt abscesses and massive increase of acute inflammatory cells and follicle formation in deeper cell layers; and Grade 5, ulcerations with gland and mucosal atrophy, crypt abscesses, extensive stromal inflammation and deep follicles.

Lymphocyte stimulation

Spleen cells were isolated by passing the tissue through a nylon net, erythrocytes were lysed with ammonium chloride and the single cell suspension was washed in 1% FCS/PBS. Cell cultures were performed in Iscove's total medium. The total spleen cell population was added to flat-bottomed 96-well plates at 2 × 10⁵ cells/well in a total volume of 200 µl. The cells were cultured in the presence of α-CD3 [10% supernatant from the α-CD3 monoclonal antibody (mAb)-producing cell line 452C11] together with the B. pertussis surface antigens FHA (5 µg/ml) or PRN (5 µg/ml) for 72 h with the addition of 1 µCi [³H]TdR (Roche, Palo Alto, CA, USA) per well for the final 6 h of culture to determine proliferative responses. The incorporation of [³H]-thymidine was measured in a β-scintillation counter. Additionally, 100 µl of the supernatants were collected after 24 h of culture and stored at −70°C until analysis of cytokine content. The spontaneous production of IL-2 was determined by Pharmingen OptEIA^TMSet cytokine kits (BD Biosciences, San Jose, CA, USA), according to the manufacturer's recommendations.

Serum IgG subclass enzyme-linked immunosorbent assay (ELISA)

Mice were bled at 19 weeks of age. Maxisorp microtitre plates (Nunc, Roskilde, Denmark) were coated with purified goat anti-mouse IgG antibodies (Southern Biotechnology) at 5 µg/ml in PBS at 4°C overnight. Coated plates were then washed in PBS and blocked with 1% BSA/PBS for 1 h at 37°C. Sera were added in 1 : 1000 dilutions, and serial threefold dilutions in 1% BSA/PBS were performed. Following incubation at 4°C overnight and washing in PBS/0·05% Tween 20, bound antibodies were demonstrated with alkaline phosphatase conjugated antimouse IgG1 or IgG2a (Southern Biotechnology) and visualized with p-nitrophenyl phosphate tablets (p-NPP) (Sigma) 1 mg/ml in ethanolamine buffer pH 9·8. The absorbance at 405 nm was determined in a Labsystems Multiscan ELISA reader. The total immunoglobulin levels in serum were calculated from the plotted standard curve of serial dilutions of a mouse reference serum (ICN, Aurora, USA) with known IgG subclass concentrations.

Analysis of apoptotic and necrotic cells

Gαi2^−/− spleen cells were cultured for 40 h in the presence of 1·25, 2·5 or 5 µg/ml FHA together with α-CD3 (10% 452C11 supernatant). Freshly isolated or cultured Gαi2^−/− spleen cells, at 0·5–5 × 10⁵/100 µl in 1% FCS/PBS, were incubated with fluorescein isothiocyanate (FITC) or phycoerythrin (PE)-conjugated mAbs (BD Biosciences) for 30 min at 4°C. The cells were then incubated either in a buffer containing 1% annexin-V FITC using TACS^TM annexin-V FITC Apoptosis Detection Kit (R&D Systems, Minneapolis, MN, USA) or in 10% propidium iodide for 15 min in the dark and thereafter diluted in binding buffer according to the manufacturer's recommendations. Analysis was performed on 10 000 live lymphocytes per sample as defined by forward- and side-scatter on a FACScan (BD Biosciences).

Generation of Th1 and Th2 cell lines

Splenic CD4⁺ T cells from OVA-transgenic mice were purified with Dynabead CD4 Positive Isolation Kit (Dynal Biotech, Oslo, Norway). Red blood cells (RBC) lysed syngeneic splenic feeder cells (1 × 10⁶/ml) were depleted of CD4⁺ and CD8⁺ T cells using Dynabead CD4 and CD8 Positive Isolation Kit and the non-selected cells were collected and irradiated at 2200 rad. CD4⁺ spleen T cells (3 × 10⁵/ml) were cultured for 5 days together with syngeneic splenic feeder cells (1 × 10⁶/ml) in the presence of rIFN-γ (ΒD Biosciences) and α-IL-4 (a kind gift from Professor Nils Lycke, Göteborg University), or rIL-4 (BD Biosciences) and α-IFN-γ (a kind gift from Professor Nils Lycke, Göteborg University) (all at 10 µg/ml) to obtain Th1 and Th2 cells, respectively. Cells were cultured for an additional 2 days in 96-well plates precoated with α-CD3 (10% 452C11 supernatant), in the presence of 10 µg/ml rIL-2 (BD Biosciences) and with or without the addition of 5 µg/ml FHA.

Statistical analysis

The Mann—Whitney U-test for unmatched data was used for analysis of significance. P-values ≤ 0·05 were regarded as significant.

Results

Decreased mortality and attenuated colitis in Gαi2^−/− mice treated with three-component B. pertussis vaccine

The colitis that develops spontaneously in Gαi2^−/− mice is progressive and results in death of the animal. The attenuation of colitis in Gαi2^−/− mice treated with a three-component B. pertussis vaccine was investigated. Indeed, Gαi2^−/− mice treated with the three-component B. pertussis vaccine showed diminished mortality as compared to untreated Gαi2^−/− mice. All the treated Gαi2^−/− mice (n = 12) were alive at 19 weeks of age, whereas 25% of the untreated Gαi2^−/− mice (n = 12) had died due to severe colitis at the same time-point (Fig. 1a). The clinical evaluation of the colitis score clearly demonstrated a significantly decreased colitis (P = 0·05) in the animals treated with B. pertussis vaccine at 19 weeks of age (Fig. 1b). More than 65% of the untreated Gαi2^−/− mice achieved a colitis score of 4 or 5, whereas only 33% of the mice in the B. pertussis vaccine-treated group had a colitis score of 4 and no mice in this group manifested a colitis equivalent to 5. Due to mortality, three of the untreated Gαi2^−/− animals were not graded histopathologically, although it could be anticipated that these animals would have assessed colitis score 5, as their death was due to severe colitis. The blinded histopathological evaluation of the colitis showed decreased intestinal inflammation in the B. pertussis-treated Gαi2^−/− mice (P = 0·05 including dead animals, P = 0·15 excluding dead mice) compared to untreated controls (Fig. 1c). Treated mice showed preserved mucin, normal glands and only minor inflammation in the stroma compared to untreated Gαi2^−/− mice, where the majority of the mice exhibited severe inflammation, atrophic glands, crypt abscesses and disappearance of mucous. Furthermore, the histopathological evaluation revealed that only 25% of the treated mice showed colitis score as 4 compared to the untreated group, where 67% of the mice had this, or a higher, degree of inflammation (Fig. 1c), which correlates with the data obtained by clinical evaluation of colitis. In conclusion, treatment of Gαi2^−/− mice with a three-component B. pertussis vaccine significantly attenuates colonic inflammation and reduces mortality.

Fig. 1.

Decreased mortality and attenuated colitis in inhibitory G-protein α2 subunit (Gαi2^−/−) mice treated with three-component Bordetella pertussis vaccine. Gαi2^−/− mice were given 5 µg B. pertussis three-component vaccine at the age of 6, 10 and 14 weeks of age, and untreated Gαi2^−/− mice were used as controls. (a) The numbers of mice that died due to severe colitis up to 19 weeks of age were recorded. Results are pooled from two independent experiments with a total of 12 mice in each group. (b) Blinded clinical evaluation of colitis in the large intestine from three-component B. pertussis vaccine treated and untreated Gαi2^−/− mice at 19 weeks of age. The clinical colitis was assessed as declared in Materials and methods, rating the inflammation score from 1 to 5. Results shown are from two independent experiments with a total of 12 mice in each group at the start of the experiment. Each dot represents one mouse and the median for each group is denoted. (c) Blinded histopathological evaluation of colitis in the large intestine from three-component B. pertussis vaccine treated and untreated Gαi2^−/− mice at 19 weeks of age. The histopathological intensity of colonic inflammation was analysed according to previous established criteria, rating the inflammation score from 1 to 5. Results shown are from two independent experiments with a total of 12 mice in each group at the start of the experiment, although showing histological score of only nine animals in the untreated group as three mice died (indicated in the figure by a cross) due to severe colitis during the study, and were not included in the histological evaluation. Each dot represents one mouse and the median for each group is denoted.

Significantly increased mucosal IL-10 levels in Gαi2^−/− mice treated with three-component B. pertussis vaccine

Gαi2^−/− mice treated with three-component B. pertussis vaccine were also investigated for the ability to induce a Th2/Tr1 response. Interestingly, Gαi2^−/− mice treated with a three-component acellular B. pertussis vaccine at 6, 10 and 14 weeks of age showed significantly increased spontaneous production of the Th2/Tr1 cytokine IL-10 in both the small (P = 0·02) and large (P = 0·03) intestine at 19 weeks of age, compared to untreated Gαi2^−/− mice (Fig. 2). In contrast, the spontaneous production of the Th1 cytokine IFN-γ was similar in the two groups. Thus, three-component B. pertussis vaccine treatment of Gαi2^−/− mice significantly increased the spontaneous production of IL-10 in both the small and large intestinal mucosa.

Fig. 2.

Increased spontaneous production of interleukin (IL-10) but not interferon (IFN)-γ in small and large intestinal mucosa of Bordetella pertussis vaccine-treated inhibitory G-protein α2 subunit (Gαi2^−/−) mice at 19 weeks of age. Mice were given three doses of 5 µg B. pertussis three-component vaccine at the age of 6, 10 and 14 weeks of age, and were killed at 19 weeks of age; 50 mg intestinal tissue was cultured for 24 h, and the spontaneous production of IL-10 and IFN-γ was determined by enzyme-linked immunosorbent assay (ELISA). Results are pooled from two separate experiments and shown as mean ± s.d. from 12 mice in each group.

Decreased serum IgG2a/IgG1 ratio in Gαi2^−/− mice treated with three-component B. pertussis vaccine

We also wanted to determine whether treatment with the three-component B. pertussis vaccine, apart from the above shown local protective effects, could down-regulate the systemic Th1 response recorded previously in Gαi2^−/− mice prior to colitis [11]. The total serum IgG1 antibodies, a subclass mediated by the Th2 cytokine IL-4, and total IgG2a subclass antibodies, a subclass mediated by the Th1-cytokine IFN-γ, were measured by ELISA. When Gαi2^−/− mice or wild-type mice were treated with three-component B. pertussis vaccine, a significantly (P = 0·004 and P = 0·005, respectively) decreased serum IgG2a/IgG1 ratio was recorded, indicating a systemic switch towards a Th2 response (Fig. 3). Thus, treatment of Gαi2^−/− mice with the three-component B. pertussis vaccine induces systemic Th2 response.

Fig. 3.

Significantly decreased serum IgG2a/IgG1 ratio from inhibitory G-protein α2 subunit (Gαi2^−/−) mice treated with three-component Bordetella pertussis vaccine. Gαi2^−/− mice were given 5 µg B. pertussis three-component vaccine at the age of 6, 10 and 14 weeks of age and untreated Gαi2^−/− mice were used as controls, and the serum IgG subclasses were determined by enzyme-linked immunosorbent assay (ELISA) analysis at 19 weeks of age. Results are shown as the ratio of serum IgG2a : IgG1 and are based on two separate experiments and shown as mean ± s.d. from 12 B. pertussis treated mice and nine untreated mice. Absolute values for IgG1 were 2768 ± 1156 µg/ml in B. pertussis-treated Gαi2^−/− mice compared to 1391 ± 1980 µg/ml in untreated Gαi2^−/− mice. In wild-type mice IgG1 levels were 7816 ± 2314 µg/ml and 1013 ± 869 µg/ml in B. pertussis and untreated mice, respectively. For IgG2a the absolute values were 8068 ± 2785 µg/ml in B. pertussis-treated and 8523 ± 3738 µg/ml in untreated Gαi2^−/− mice, 9764 ± 3433 µg/ml in B. pertussis treated and 3711 ± 1611 µg/ml in untreated wild-type mice.

One-component B. pertussis vaccine does not attenuate colitis in Gαi2^−/− mice

Having shown that Gαi2^−/− mice treated with a three-component B. pertussis vaccine containing B. pertussis PTd, FHA and PRN showed less severe inflammation as compared to untreated mice, the Gαi2^−/− mice were treated with a one-component vaccine containing only B. pertussis toxoid apart from alum as adjuvant. The one-component acellular B. pertussis vaccine neither decreased the severity of colonic inflammation, as judged by clinical or histopathological evaluation, nor diminished the mortality of the Gαi2^−/− mice (not shown). In accordance, no shift in the spontaneous production of IL-10 or IFN-γ in the intestinal mucosa or in the serum IgG2a/IgG1 ratio was recorded in the one component B. pertussis vaccine-treated group (not shown). Thus, the one-component B. pertussis vaccine containing B. pertussis toxoid only does not attenuate the colitis in Gαi2^−/− mice.

B. pertussis surface antigen FHA suppresses lymphocyte proliferation

As treatment with the three-component B. pertussis vaccine attenuated colitis in Gαi2^−/− mice, but the one-component B. pertussis vaccine did not, we further wanted to elucidate the immunoregulatory properties of the B. pertussis surface antigens filamentous haemagglutinin (FHA) and pertactin (PRN). Therefore, splenic T cells from Gαi2^−/− as well as wild-type mice were stimulated with α-CD3 in the presence of FHA or PRN. The α-CD3-induced proliferation was significantly suppressed (P = 0·02) when splenic T cells were cultured with 5 µg/ml FHA, whereas the presence of 5 µg/ml PRN had no effect (Fig. 4a). Furthermore, there was neither an inhibitory nor an additive effect by culturing cells with FHA and PRN in combination compared with FHA alone. In contrast to the inhibited T cell proliferation, co-culture of T cells with FHA, but not PRN, resulted in increased levels of IL-2 (Fig. 4b). In conclusion, stimulation of splenic T cells with α-CD3 in the presence of FHA leads to significantly decreased proliferation in combination with increased levels of IL-2.

Fig. 4.

Decreased T lymphocyte proliferation and increased levels of interleukin (IL)-2 in the presence of filamentous haemagglutinin (FHA). (a) Splenocytes from inhibitory G-protein αi2 subunit (Gαi2^−/−) or wild-type mice were cultured in the presence of α-CD3 together with the Bordetella pertussis surface antigens FHA (5 µg/ml) or pertactin (PRN) (5 µg/ml) for 72 h with the addition of 1 µCi [³H]TdR for the final 6 h of culture to determine proliferative responses. (b) Splenocytes from Gαi2^−/− or wild-type mice were cultured in the presence of α-CD3 together with the B. pertussis surface antigens FHA (5 µg/ml) or PRN (5 µg/ml) and the supernatants were collected after 24 h of culture and the spontaneous production of IL-2 was determined by enzyme-linked immunosorbent assay (ELISA). Results are shown as mean ± s.d. of individual mice from one representative experiment of two giving similar results, both with four mice in each group.

In order to investigate further whether FHA was the colitis attenuating component of the three-component vaccine, FHA was given i.p. twice a week for 15 weeks to Gαi2^−/− mice without the co-administration of any adjuvant. This treatment did not, however, abrogate the colitis development in the Gαi2^−/− mice (not shown). Taken together, FHA possesses immunoregulatory properties but do not attenuate colitis in Gαi2^−/− mice on its own.

FHA induces apoptosis of activated Th1 lymphocytes

The suppressed proliferation in combination with increased IL-2 levels in the supernatant indicated that IL-2 was produced but not consumed in these cultures. This suggested either an increased susceptibility to apoptosis [23] or an increased necrosis of T lymphocytes in the presence of FHA. To investigate whether FHA induced apoptosis, splenic T cells from Gαi2^−/− as well as wild-type mice were stimulated with α-CD3 in the presence of 1·25, 2·5 or 5 µg/ml FHA. The frequency of annexin-V⁺ apoptotic cells increased in a dose-dependent manner (not shown).

Moreover, the Gαi2^−/− mice were treated with 5 µg FHA i.p. This resulted in increased apoptosis of the CD4⁺ spleen T lymphocytes as compared to untreated mice (Fig. 5a). In addition, 66% of the CD44^hi splenocytes from mice treated with FHA in vivo was apoptotic as compared to 33% in untreated mice (Fig. 5c). However, FHA was not toxic, as treatment with FHA did not result in increased necrosis of either CD4⁺ T cells or CD44^hi cells, as determined by PI staining (Fig. 5b, d). In contrast, there was no increased apoptosis among the CD44^lo CD4⁺splenocytes from the FHA-treated compared to untreated mice (33% and 30%, respectively) (not shown). The increased apoptosis of activated CD4⁺ T cells was confirmed by the finding that the frequency of peripheral blood CD4⁺ T lymphocytes expressing CD44 was decreased by 10% in Gαi2^−/− mice treated in vivo with the three-component B. pertussis vaccine, compared to untreated Gαi2^−/− mice (not shown).

Fig. 5.

Filamentous haemagglutinin (FHA) induces increased apoptosis of activated CD4⁺ T cells. Inhibitory G-protein αi2 subunit deficient (Gαi2^−/−) mice were given 5 µg FHA i.p. once and untreated Gαi2^−/− mice were used as controls. Mice were killed 48 h after treatment and spleen cells were isolated and double-stained with either annexin-V fluorescein isothiocyanate (FITC) and CD4 phycoerythrin (PE) for detection of apoptotic CD4⁺ T cells (a), PI and CD4 PE for detection of necrotic CD4⁺ T cells (b), annexin-V FITC and CD44 PE for detection of apoptotic CD44^hi cells (c) or PI and CD44 PE for detection of necrotic CD44^hi cells (d).

To investigate whether there was a difference between Th1 and Th2 cells in the susceptibility to apoptosis induced by FHA, short-term CD4⁺ Th1 and Th2 cell lines were generated from spleen cells from OVA-transgenic mice. Indeed, Th1 cells showed a more than four times as high frequency of apoptotic cells as the Th2 cell population in the presence of FHA. Although the frequency of apoptotic cells was not different in the two cell types in the absence of FHA, Th1 cells exposed to FHA showed 74% increased relative frequency of apoptotic cells compared to untreated Th1 cells, whereas Th2 cells cultured with FHA showed a 16% increased relative frequency of apoptotic cells as compared to untreated Th2 cells (Table 1). Thus, FHA induces apoptosis of activated T lymphocytes both in vivo and in vitro, and Th1 cells are much more sensitive to FHA-induced apoptosis.

Table 1. Filamentous haemagglutinin (FHA) induces apoptosis preferentially of Th1 CD4⁺ T cells.

Cell type	FHA added	Frequency of annexin-V+ apoptotic cells (%)	Relative increase of apoptotic cells (%)
Th1a	–	14·8
Th1	5 µg/ml	25·8	74
	Th2	–	12·4
Th2	5 µg/ml	14·4	16

^aCD4⁺ T cells isolated from ovalbumin (OVA)-transgenic mice were cultured for 5 days together with feeder cells in the presence of OVA p323, interferon (IFN)-γ and α-interleukin 4 (IL-4) or α-IFN-γ to obtain the Th1 and Th2 cell lines, respectively. Thereafter, cells were washed and 5 µg/ml filamentous haemagglutinin (FHA), 10 µg/ml rIL-2 and 10%α-CD3 was added to the cultures for two consecutive days. No FHA was added to controls, as indicated in the table. Results shown are from one representative experiment of two, showing similar results. Individual cell lines were generated for the two separate experiments.

Discussion

This study shows that treatment of Gαi2^−/− mice with the acellular three-component B. pertussis vaccine attenuates colitis, as shown by reduced mortality together with decreased clinical as well as histopathological score. The attenuated colitis in Gαi2^−/− mice was probably mediated, at least partly, by a significantly increased level of IL-10 in the intestinal mucosa and increased apoptosis of Th1 cells.

IL-10, produced by both Th2 and Tr1 cells [24], is crucial in regulating the intestinal mucosal immune response as Tr1 clones, producing high levels of IL-10, have been shown to suppress the proliferation of CD4⁺ T cells and prevent colitis in a murine model of IBD [24]. Systemic IL-10 administration prior to and after initiation of colitis has been successful in animal models [25–27], but has been ineffective at reversing established inflammation in patients [28–30]. Recent results suggest that IL-10 is regulating the immunosuppressive effect of TGF-β by up-regulating the TGF-βRII on activated CD4⁺ T cells, and making them susceptible to TGF-β-mediated suppression [31].

A possible mechanism for the increased levels of IL-10 in the Gαi2^−/− mice treated with three-component B. pertussis vaccine is that FHA, present in the three-component B. pertussis vaccine, induces apoptosis of activated Th1 cells, resulting in a relatively increased level of Th2/Tr1 cells producing IL-10. In addition, the presence of apoptotic cells can have a direct immunoregulatory effect in that it induces the secretion of IL-10 and decreases the secretion of TNF-α, IL-1 and IL-12 from activated monocytes [32]. A general shift in the Th1/Th2 balance, rather than an isolated effect on IL-10-producing cells, is supported by the decreased serum IgG2a/IgG1 ratio in Gαi2-deficient as well as in wild-type mice treated with the three-component B. pertussis vaccine, as switching to IgG1 is dependent on IL-4. Nevertheless, it cannot be ruled out that the increased apoptosis of activated Th1 cells and the increased mucosal production of IL-10 are two separate mechanisms of action of FHA, resulting in the strong beneficial effects in this model of IBD.

The fact that treatment with the three-component B. pertussis vaccine attenuates colitis, whereas the one-component B. pertussis vaccine does not, suggests that either of the B. pertussis antigens, PRN or FHA, present in the three-component vaccine, and not unaccompanied B. pertussis toxoid is attenuating the colitis.

FHA has been found previously to inhibit T cell proliferation [33] and induce apoptosis in human monocyte-like cells and bronchial epithelial cells, mediated possibly through the TNF-α receptor pathway [34]. However, this study is, to the best of our knowledge, the first to describe the strong apoptosis-inducing effects of FHA on activated Th1 lymphocytes. Whether this is due to the general higher susceptibility to apoptosis in Th1 cells [35] or a specific action of FHA on this cell population is at present unknown.

FHA has been shown to bind to epithelial cells and macrophages via at least two different leucocyte integrins, leucocyte response integrin (LF1) and complement receptor 3 (CR3) [14–16]. Interestingly, CR3 on T lymphocytes is up-regulated when the cells are activated, and binding of the CR3 receptor inhibits T cell proliferation as well as IL-2 release [36]. It is thus possible that the engagement of the CR3 receptor on activated T cells − possibly also by other ligands than FHA − induces apoptosis of the activated T lymphocytes. Most interestingly, in the light of our results, expression of CR3 has been found to facilitate T cell homing to sites of inflammation [37]. Thus, a possible mechanism for the protective effect of FHA in the B. pertussis vaccine preparation is that FHA induces apoptosis of activated T cells bound to home to sites of intestinal inflammation, leading to fewer activated T cells and less production of proinflammatory cytokines in the intestinal mucosa.

In addition, recent studies have demonstrated that FHA has Th2-promoting effects also on macrophages in supplement to the inhibitory effect on Th1 lymphocytes described by us in the present study, in that it suppresses IL-12 and stimulates the production of IL-6 and IL-10 by macrophages in an IL-10-dependent manner [38]. Furthermore, the same group reported that FHA inhibits IL-12 and stimulates IL-10 production by dendritic cells and that these dendritic cells direct naive T cells into T regulatory type 1 (Tr1) cells [39]. These mechanisms most probably act in concert to help B. pertussis to escape immune evasion by cell-mediated immunity, and they may prove very useful as a therapeutic strategy in several different Th1-dominated chronic disorders. However, as no colitis attenuation was seen when pure FHA was administered to Gαi2^−/− mice, it suggests that a constituent or the combination of several constituents present in the three-component vaccine is needed for FHA to exert its effects in vivo. It is possible that if FHA was administered in the presence of formaldehyde, glutaraldehyde and/or alum, a gel-type depot component that improves the delivery of antigen, as is the case in the three component B. pertussis vaccine, it would be more efficient in exerting its immunomodulatory effects in vivo.

In conclusion, this study shows that a B. pertussis vaccine attenuates colitis in the Gαi2^−/− mice and that this is associated with increased apoptosis of Th1 cells and increased production of IL-10 locally in the mucosa. The acellular three-component B. pertussis vaccine used in this study has been routinely administered to infants without reported side effects, and has been shown to induce a Th2 immune response in patients [19]. It is therefore a promising candidate drug for treatment of, for example, Crohn's disease where a Th1 mediated pathology is clear, but also for those ulcerative colitis patients where the disease has been demonstrated to be mediated by predominantly Th1 cells.