Isabella Pali-Schöll Department of Pathophysiology Medical University of Vienna Währinger Gürtel 18-20 1090 Vienna Austria
Background: Aluminium (ALUM) is used as experimental and clinical adjuvant for parenteral vaccine formulation. It is also contained in anti-acid drugs like sucralfate (SUC). These anti-acids have been shown to cause sensitization to food proteins via elevation of the gastric pH. The aim of this study was to assess the oral adjuvant properties of ALUM, alone or contained in SUC, in a BALB/c mouse model.
Methods: Mice were fed SUC plus ovalbumin (OVA) and compared with groups where ALUM or proton pump inhibitors (PPI) were applied as adjuvants. The humoral and cellular immune responses were assessed on antigen-specific antibody and cytokine levels. The in vivo relevance was investigated in skin tests.
Results: The highest OVA-specific immunoglobulin G1 (IgG1) and IgE antibody levels were found in mice fed with OVA/SUC, followed by OVA/ALUM-treated animals, indicating a T helper 2 (Th2) shift in both groups. Antibody levels in other groups revealed lower (OVA/PPI-group) or baseline levels (control groups). Positive skin tests confirmed an allergic response in anti-acid or adjuvant-treated animals.
Conclusions: Our data show for the first time that ALUM acts as a Th2-adjuvant via the oral route. This suggests that orally applied SUC leads to an enhanced risk for food allergy, not only by inhibiting peptic digestion but also by acting as a Th2-adjuvant by its ALUM content.
The rate of food allergy in western populations has been constantly rising during the last decades. At present, about 3.5–4% of adults and about 6% of children are affected in the United States (1–3). Although no coherent explanation of this development has been found yet, there is an agreement that the risk of developing food allergies depends on genetic predisposition, increased exposure to allergens (4, 5) and several environmental factors (6). The remarkable boost of allergies in western countries has led to several theories about this phenomenon. Many of them address the potential link of modern lifestyle to an altered immune response (7). Our own studies introduced anti-acid drugs as responsible factors for sensitization in mouse models (8–10) and more importantly also in humans (8, 11). The underlying mechanism is based on elevation of the gastric pH and therefore blocking of the gastric digestion. Some of these anti-acid drugs, e.g. sucralfate (SUC), contain aluminium (ALUM), which is commonly used as a parenteral T helper 2 (Th2)-adjuvant (12, 13).
Alum is a ubiquitous element in western regions. It is present in drinking water – especially in urban areas – as well as in food such as soy-based milk products, baking powder, frozen products and many more (14). While plants are getting contaminated by growing on ALUM-rich farmland, many food additives such as preservatives, colouring agents, leavening agents and anticaking agents are intentionally mixed with ALUM (15). Especially, food consumed in the United States has been shown to contain high concentrations of ALUM compared with other countries (15, 16). Furthermore, ALUM is used in water purification, sugar refining and brewing. Many drugs use ALUM either as an additive (antacids, analgesics, antidiabetic drugs, etc.) or are contaminated with ALUM (17). Depending on these variables, ingestion of ALUM in humans can vary greatly between countries from 2 to 160 mg/person/day (15).
Food allergens have several characteristics in common like stability or persistence to digestion. The remaining proteins or fragments may be better recognized by the immune system (18). This process may even be augmented by ALUM via the oral route, based on the fact that it causes potentiation of the immune response by adsorption of proteins on the parenteral route (19).
Here, we investigate for the first time the impact of ALUM intake on the process of sensitization via the oral route and the impact of ALUM as a component of the anti-acid drug SUC.
Material and methods
Measurement of pH changes in simulated gastric fluid after addition of adjuvants
Simulated gastric fluid (SGF) contained 0.2% of sodium chloride, 0.32% of pepsin and 0.7% of hydrochloric acid; pH of the final solution was 1.2. After the addition of 200 μl SUC (1 g of SUC/5 ml suspension; Ulcogant, Merck, Vienna, Austria), 200 μl ALUM (Serva, Heidelberg, Germany) or 200 μl phosphate buffer saline (PBS; negative control) the pH was determined at different time points from 0 to 7200 s (pH meter; Metrohm, Zofingen, Switzerland). Additionally, different amounts of test substances (400, 600, 800 or 1000 μl) were added to investigate dose-dependent effects.
Protocol of sensitization of mice
BALB/c mice (female: 6–8 weeks) were obtained from the Institute for Laboratory Animal Science and Genetics (Medical University of Vienna) and treated according to European Community rules of animal care with the permission of the Austrian Ministry of Science (GZ 66.009/0108-C/GT/2007). On days 0, 7, 14, 21, 28, 35, 42 and 49, BALB/c mice (n = 5 per group) were immunized intragastrally as shown in Table 1. This immunization scheme was performed according to our established food allergy model (20). Blood samples were drawn by tail-bleeding on days 0 (preimmune serum), 14, 28, 42, 56, 70 and 84.
*1 g sucralfate/5 ml suspension; Ulcogant, Merck, Vienna, Austria.
†Serva, Heidelberg, Germany.
‡Nexium 40 mg, AstraZeneca, Wedel, Germany; 64 μg esomeprazole/20 μl 0.9% NaCl 24 h (i.v.) and 1 h (i.v. and i.m.) prior to oral immunization.
0.2 mg OVA/100 μL PBS plus 50 μl sucralfate*
100 μl PBS plus 50 μl sucralfate*
0.2 mg OVA/100 μl PBS plus 18 μl 2% alum† and 32 μl PBS
0.2 mg OVA/100 μl PBS plus 50 μl PBS
150 μl PBS
0.2 mg OVA/100 μl PBS plus 50 μl PBS, pretreated with a PPI‡
150 μl PBS, pretreated with a PPI‡
Antibody detection in serum by enzyme-linked immunosorbent assay
Microtiter plates (Maxisorp, Nunc, Roskilde, Denmark) were coated with ovalbumin (OVA; Sigma, Steinheim, Germany) or hazelnut extract as control allergen (20 μg/well in 100 μl 50 mM NaHCO3, pH 9.6) over night at 4°C. Standard serial dilution of the respective purified mouse antibody [immunoglobulin E (IgE), IgG1 or IgG2a 0.1 mg/ml isotype control; BD Biosciences, Franklin Lakes, NJ, USA] served as standard. Each well was blocked with 200 μl TBST/1% bovine serum albumin (BSA) for 2 h at 4°C. Subsequently, serum samples were diluted 1 : 100 for IgG1 and IgG2a, or 1 : 10 for IgE (in TBST/0.1% BSA) and incubated over night at 4°C. The detection antibody (r-anti-mIgG; r-anti-mIgG2a; r-anti-mIgE; BD Biosciences) was diluted 1 : 500 in TBST/0.1%BSA and incubated for 2 h at room temperature (RT). The secondary antibody (POX-anti-rat; GE Healthcare, Vienna, Austria) was diluted 1 : 1000 in TBST/0.1%BSA and was incubated for 1 h at RT. ABTS 0.1% and H2O2 1% in citric acid buffer were used for developing. Readout was performed in a microtitre plate reader (Spectra Max Plus, Molecular Devices, CA, USA) at wavelengths of 405 and 490 nm.
Skin tests of mice
On the day of killing (day 84), Evans blue (100 μl of 5 mg/ml NaCl 0.9%; Merck, Darmstadt, Germany) was injected into the tail vein of mice. Subsequently, 30 μl of OVA (50 μg/ml in PBS), 30 μl of codfish extract (50 μg/ml in PBS) as irrelevant control allergen, mast cell degranulation compound 48/80 (20 μg/ml in PBS; Sigma) as positive control or PBS as negative control were administered intradermally into shaved abdominal skin. After 20 min, mice were killed and skinned. A positive response can be seen as a blue colour reaction on the inside of the abdominal skin because of vascular leakage.
Isolation of splenocytes and detection of cytokines
Mice were killed and the spleens were removed. Spleen cell suspensions of individual spleens were prepared immediately by cutting, mincing and filtering through 70 μm nylon meshes (BD Biosciences, Schwechat, Austria). Cells were resuspended in RPMI medium (Gibco Invitrogen, Lofer, Austria), supplemented with 10% foetal calf serum, 1%l-glutamine and 1% penicillin/streptomycin. Mononuclear cells were isolated by density separation (Lympholyte-M; Cedarlane, Hornby, ON, Canada) according to manufacturer’s instructions. Cells were plated (4 × 105 cells/well) in sterile round-bottom 96-well tissue culture plates (Costar, New York, USA). For stimulation, OVA (400 μg/ml), codfish extract (400 μg/ml), Concanavalin A (5 μg/ml; Sigma) as positive control or medium were added. The cells were cultured for 72 h at 37°C and 5% CO2. Supernatants were harvested and stored at −20°C until further use for cytokine determination. Measurement of interleukin-4 (IL-4), IL-5 and interferon-γ (IFN-γ) was performed by enzyme-linked immunosorbent assay (ELISA) with anti-mouse cytokine-antibodies and standards (Bender MedSystems, Vienna, Austria) in supernatants of stimulated splenocytes (diluted 1 : 2), according to manufacturer’s instructions.
Beta-hexosaminidase secretion assay from rat basophile leukaemia cells
Rat basophile leukaemia (RBL)-2H3 cells (generously provided by Dr Arnulf Hartl, Paracelsus Medizinische Privatuniversitaet Salzburg, Austria) were cultured in RPMI medium supplemented with 10% foetal calf serum, 4 mM l-glutamine, 2 mM sodium pyruvate, 10 mM HEPES, 1% penicillin/streptomycin and 100 μM 2-mercaptoethanol at 37°C and 5% CO2. Rat basophile leukaemia cells (4 × 104 cells/well in 100 μl medium) were seeded in a 96-well round-bottom plate (96 MicroWell, Nunclon, Nunc, Roskilde, Denmark) and incubated overnight at 37°C and 5% CO2. Cells were passively sensitized with pooled sera from mice (diluted 1 : 7) for 2 h at 37°C and 5% CO2. Afterwards, cells were washed twice with Tyrode’s buffer (137 mM NaCl, 2.68 mM KCl, 1.4 mM CaCl2, 1.7 mM MgCl2, 5 mM d-glucose, 10 mM HEPES, 0.42 mM NaH2 PO4·H2O, 12 mM NaHCO3) at pH 7.2, supplemented with 0.1% BSA (Sigma). Ovalbumin (0.1 μg in 100 μl /well), codfish extract (0.1 μg in 100 μl /well) as irrelevant control allergen or ionomycin (0.4 μg in 100 μl/well; Sigma) as positive control were diluted in Tyrode’s buffer and added to the wells for 30 min at 37°C and 5% CO2. As a negative control RBL-cells were incubated with Tyrode’s buffer without any allergens to calculate spontaneous release. Cell lysis was performed by adding 10 μl of 10% Triton-X (Sigma) in Tyrode’s buffer for 5 min for total release of β-hexosaminidase (100% release). Of the supernatants, 50 μl was mixed with 50 μl assay solution (5 ml 0.1 M citric acid, pH 4.5 with NaOH plus 80 μl 4-methyl-umbelliferyl-N-acetyl-beta-d-glucosaminide; Sigma) and incubated for 1 h at 37°C. The reaction was terminated by adding 100 μl glycine buffer (200 mM glycine, 200 mM NaCl, pH 10.7). Finally, fluorescence was read at 360/465 nm in a fluorometer (Cytofluor 2350, Millipore, Vienna, Austria) and the results were calculated as percentage of total release.
Statistical comparison of antibody and cytokine levels between groups was performed using a one-way anova with the Bonferroni correction for multiple comparisons using the software, spss (version 16.0 for Windows; SPSS Inc., Chicago, IL, USA). Differences were considered statistically significant at P-values < 0.05.
Sucralfate raises pH of simulated gastric fluid
The pH of SGF (3 ml) was determined at defined time points (0–7200 s) after adding 200 μl SUC, 200 μl ALUM or 200 μl PBS (negative control). While SUC caused a marked pH elevation up to 3.4, no pH changes could be observed after the addition of ALUM or PBS (Fig. 1). Additionally, SUC showed a dose-dependent effect, whereas amounts of ALUM up to 1000 μl/3 ml SGF did not change the pH (data not shown).
Oral treatment of mice with sucralfate results in a specific Th2-response
Female BALB/c mice were treated either with OVA/SUC, OVA/ALUM, OVA/proton pump inhibitor (PPI) or OVA/PBS (Table 1). Control groups were applied either PBS/SUC or PBS/PPI or PBS alone. Mice treated with OVA/SUC revealed the highest OVA-specific IgE response (mean 113.6 ng/ml, SEM 17.3). Also, OVA/ALUM feedings induced a significant increase in OVA-specific IgE antibodies (mean 64.7 ng/ml, SEM 16.4). A moderate increase in OVA-specific IgE antibodies (mean 39.7 ng/ml, SEM 6.3) was found in the OVA/PPI-treated group, while the OVA/PBS-treated group showed low OVA-specific IgE antibody levels (mean 21.9 ng/ml, SEM 10.5). No OVA-specific IgE antibodies were observed in any of the control groups (Fig. 2A). A significant increase of OVA-specific IgG1 antibodies (mean 32.7 μg/ml, SEM 4.4 and mean 32.2 μg/ml, SEM 8.2, respectively) was found in mice treated with OVA/SUC and OVA/ALUM. OVA-specific IgG1 antibodies increased moderately (mean 14.2 μg/ml, SEM 3.9) in OVA/PPI-treated mice, while low OVA-specific IgG1 antibody levels (mean 9.6 μg/ml, SEM 5.9) were measured in the OVA/PBS-treated group. No OVA-specific IgG1 antibodies were observed in any of the control groups (Fig. 2B). No significantly elevated OVA-specific IgG2a antibodies could be seen in any of the groups (Fig. 2C). The presence of biologically relevant OVA-specific IgE in OVA/SUC and OVA/ALUM-treated groups on day 84 was further shown in an RBL-assay, using pooled mouse sera (Fig. 3). Control groups did not show OVA-specific IgE. No reactions were found to the control allergen codfish/hazelnut extract either in ELISA or in RBL-assay (data not shown).
Levels of IL-5 and IFN-γ are increased in splenocytes of OVA/SUC-treated mice
A significant elevation of IL-5 and IFN-γ in supernatants of splenocytes was observed after stimulation with OVA in OVA/SUC-treated mice (Fig. 4). Results of IL-4 analysis revealed no significant elevation in any group (data not shown). Cytokines remained at base level after stimulation with codfish extract, which was used as control allergen (data not shown).
Intradermal skin tests show a positive reaction after OVA stimulation in SUC-treated mice
The in vivo effect of the Th2-response was investigated by intradermal skin tests. Positive skin test reactions to OVA were seen in mice treated with OVA/SUC (5/5) (Fig. 5A) or OVA/ALUM (4/5) (Fig. 5C) as well as in the OVA/PPI-treated group (4/5) (Fig. 5F). Treatment with OVA/PBS led to few (2/5) (Fig. 5D), whereas PBS/SUC (Fig. 5B), PBS (Fig. 5E) or PBS/PPI (Fig. 5G) showed no positive skin reaction after OVA stimulation.
We have shown previously that anti-acid treatment is an important factor for the induction of food-specific IgE in adult humans (11) and also in mouse models (8–10). This effect is most probably caused by the increased gastric pH. As pepsinogen cannot be converted into its active form pepsin at increased pH levels, protein digestion is reduced or even blocked. These studies, including human data, therefore suggest that acid-suppressing drugs are responsible for an enhanced risk of food allergy development.
Among anti-acid drugs, SUC is a common medication for the treatment of gastro-oesophageal reflux disease in humans. It is a complex of sucrose octasulphate and ALUM and provides mucosal protection by creating an adherent complex with the protein exudates of the oesophageal/gastric mucosa. It also has pepsin-, proton- and bile acid-binding capacities that enhance tissue resistance (21–23). The proton-binding effect of SUC can be seen in our study as marked pH elevation in vitro. These data are consistent with observations of other groups (24). An elevation of the pH of SGF, after addition of SUC (Fig. 1), was shown in vitro up to 3.4. Proton pump inhibitors were used as a substance to mimic the anti-acidic effects of SUC. To compare both substances, we measured the gastric pH in BALB/c mice in vivo after application of PPIs using a pH microelectrode (20) and could show a pH elevation up to 5.3. We have shown that a reliable blockage of gastric digestion in vitro can already be seen starting from a pH of 2.75 (25). This means that although the effect of SUC and PPIs on the gastric pH is not the same (pH 3.4 vs 5.3), we suggest that both substances have a comparable blocking effect on the gastric digestion.
Aluminium, also contained in SUC, is widely used as an experimental and clinical adjuvant in vaccine formulations. It is known that ALUM-linked antigens lead to an increased probability of sensitization and induction of Th2 cell-type responses (26). The potentiation of the immune response to antigens is caused by adsorption to the adjuvant. The major mechanisms responsible for the adsorption of antigens by ALUM-containing adjuvants are electrostatic attraction, hydrophobic attraction and ligand exchange (27). One of the mechanisms, which has been proposed to explain how ALUM-containing adjuvants potentiate the immune response is that they convert the soluble antigen into a particulate form and thereby stabilize the conformation (19). However, to the best of our knowledge, it has not been investigated so far if ALUM acts as an adjuvant via the oral route.
Therefore, in this study, we examined whether ALUM can act as a Th2-adjuvant via the oral route and whether this may therefore be an additional mechanism for the sensitization potential of SUC. We fed BALB/c mice with SUC in combination with the well-characterized model allergen OVA, which has been shown to withstand incubation in SGF for up to 60 min (20, 28). This reduces the effect of blocked gastric digestion and emphasizes the adjuvant effect of ALUM. The OVA/ALUM control group was fed the same amount of ALUM via the oral route.
Compared with the low immune response of groups treated without adjuvants, mice fed with OVA plus adjuvants (SUC, ALUM) revealed a highly significant increase of OVA-specific IgE and IgG1 antibodies, suggesting a clear Th2-shift of the humoral immune response. The highest antibody levels were found in the OVA/SUC group, indicating that both mechanisms, blocking of gastric digestion and acting as a Th2-adjuvant, play a role with SUC. In the performed IgE ELISA, much higher concentrated OVA-specific IgG1 may compete against OVA-specific IgE antibodies for epitopes. Currently, there is no method available that would satisfactorily solve this problem. However, as the IgE levels in our groups of interest were significantly elevated even under these competitive circumstances, we believe that this may even strengthen the here presented data.
The biological relevance and high specificity of induced OVA-specific IgE in this study were shown in an RBL-assay. Crosslinking of antigen-specific IgE-antibodies and therefore release of β-hexosaminidase by activated RBL-cells was detectable after stimulation with OVA in OVA/SUC and OVA/ALUM-treated mice. Therefore, RBL-results confirm the Th2-type antibody induction shown in ELISA. On the cytokine level, an enhanced immune response with respect to Th1 and Th2 cytokines could be seen in the OVA/SUC-treated group. Compared with control groups, significantly elevated IL-5 and IFN-γ levels were measured in OVA/SUC-treated mice, while cytokines of all other groups remained at baseline levels.
Although the amount of ALUM in OVA/ALUM-treated mice was exactly adjusted to the amount of ALUM as contained in SUC received by the OVA/SUC group, the latter showed an even higher immune response on the humoral and cellular level. This phenomenon can be explained by the additional effect of SUC on the gastric pH: while ALUM did not raise the pH of SGF in vitro at all, SUC caused an elevation from 1.2 up to 3.4. We tested this remarkable effect of the pH elevation in an isolated form in OVA/PPI-treated mice. This group allowed us to distinguish clearly the effect of the gastric pH shift from the adjuvant effect caused by ALUM. While OVA/PPI-treated mice revealed only the effect of the gastric pH shift and OVA/ALUM-treated mice only the adjuvant effect of ALUM, both effects can be seen in combination in the OVA/SUC group, which revealed the highest immune responses of all groups. These results are further underlined by the positive skin test reactions.
Our data show for the first time that ALUM acts as a Th2-adjuvant via the oral route. This suggests that orally applied SUC leads to an enhanced risk for food allergy, not only by inhibiting peptic digestion but also by acting as a Th2-adjuvant by its ALUM content. Considering the high ingestion of ALUM in western countries via food, water and drugs, this Th2-adjuvant effect may be a partial explanation for the increasing prevalence of food allergy.
We thank Magdolna Vermes, Silke Gruber, Annika Jensen and Julian Kornprobst for their excellent technical assistance, as well as Ulrich Omasits and Matthias J. Brunner for valuable help with statistical analysis. This study was supported by the Hertha-Firnberg stipend T283-B13, SFB F01808-13 of the FWF and granted by the (Foerderungsstipendium) Medical University of Vienna.