Both authors contributed equally to the work.
Neonatal colonization of mice with Lactobacillus plantarum producing the aeroallergen Bet v 1 biases towards Th1 and T-regulatory responses upon systemic sensitization
Version of Record online: 29 SEP 2010
© 2010 John Wiley & Sons A/S
Volume 66, Issue 3, pages 368–375, March 2011
How to Cite
Schwarzer, M., Repa, A., Daniel, C., Schabussova, I., Hrncir, T., Pot, B., Stepankova, R., Hudcovic, T., Pollak, A., Tlaskalova-Hogenova, H., Wiedermann, U. and Kozakova, H. (2011), Neonatal colonization of mice with Lactobacillus plantarum producing the aeroallergen Bet v 1 biases towards Th1 and T-regulatory responses upon systemic sensitization. Allergy, 66: 368–375. doi: 10.1111/j.1398-9995.2010.02488.x
Edited by: Hans-Uwe Simon
- Issue online: 1 FEB 2011
- Version of Record online: 29 SEP 2010
- Accepted for publication 18 August 2010
- germ-free mice;
- mucosal immunity;
- primary prevention;
- recombinant lactic acid bacteria
To cite this article: Schwarzer M, Repa A, Daniel C, Schabussova I, Hrncir T, Pot B, Stepankova R, Hudcovic T, Pollak A, Tlaskalova-Hogenova H, Wiedermann U, Kozakova H. Neonatal colonization of mice with Lactobacillus plantarum producing the aeroallergen Bet v 1 biases towards Th1 and T-regulatory responses upon systemic sensitization. Allergy 2011; 66: 368–375.
Background: The use of recombinant lactic acid bacteria (LAB) as vehicles for mucosal delivery of recombinant allergens is an attractive concept for antigen-defined allergy prevention/treatment. Interventions with LAB are of increasing interest early in life when immune programming is initiated. Here, we investigated the effect of neonatal colonization with a recombinant LAB producing the major birch pollen allergen Bet v 1 in a murine model of type I allergy.
Methods: We constructed a recombinant Lactobacillus (L.) plantarum NCIMB8826 strain constitutively producing Bet v 1 to be used for natural mother-to-offspring mono-colonization of germ-free BALB/c mice. Allergen-specific immunomodulatory effects of the colonization on humoral and cellular immune responses were investigated prior and after sensitization to Bet v 1.
Results: Mono-colonization with the Bet v 1 producing L. plantarum induced a Th1-biased immune response at the cellular level, evident in IFN-γ production of splenocytes upon stimulation with Bet v 1. After sensitization with Bet v 1 these mice displayed suppressed IL-4 and IL-5 production in spleen and mesenteric lymph node cell cultures as well as decreased allergen-specific antibody responses (IgG1, IgG2a, and IgE) in sera. This suppression was associated with a significant up-regulation of the regulatory marker Foxp3 at the mRNA level in the spleen cells.
Conclusion: Intervention at birth with a live recombinant L. plantarum producing a clinically relevant allergen reduces experimental allergy and might therefore become an effective strategy for early intervention against the onset of allergic diseases.
Allergic diseases have become a substantial public health burden in industrialized countries, and their prevalence has steadily increased over the last decades. At present, allergen-specific immunotherapy (SIT) is the only curative and disease-modifying treatment in the majority of patients (1), which however has some drawbacks: SIT vaccines contain crude extracts from natural sources containing a mixture of allergens and contaminant components, which may pose a risk for de novo sensitization to new allergens (2). In this respect, standardized recombinant allergens could improve the safety profile and efficacy of SIT (3). To keep the allergens at the injection site, traditional SIT utilizes aluminium hydroxide as an adjuvant. Although aluminium adjuvants are widely used with good results in humans, they promote Th2-like responses (4), which might reduce SIT efficacy. Recently, lactic acid bacteria (LAB) have been introduced into the field of SIT (5). We and others have shown that certain LAB strains shift the allergen-specific Th2 response towards a more balanced Th1/Th2 profile (5). In this regard, the use of such nonpathogenic LAB with intrinsic Th1-promoting or immunomodulatory properties as vehicles for mucosal delivery of recombinant allergen is an attractive concept for development of well-tolerated and effective allergy vaccines (6).
The indigenous flora of the gastrointestinal tract is required to maintain immune homeostasis within the gut and promotes the maturation of the host’s immune system. Meanwhile, it is assumed that a balanced constitution of the microflora is one of the factors promoting protection from allergy and asthma (7). Several clinical studies indicated that a reduced presence of lactobacilli or bifidobacteria in the early intestinal microbiota of atopic children precedes the occurrence of allergic diseases (8). Along these lines, human trials using specific LAB strains for oral interventions have pointed out a potential for primary prevention of allergic diseases and encouraged further research (9).
Germ-free (GF) mice, born and raised in sterile isolators lacking microbial communities, enable the introduction and investigation of specific single microbial strains to study their interaction with the host’s immune system both within the intestinal tract and in peripheral tissue. Studies employing GF animals have shown that the intestinal microbiota shapes the development, distribution, differentiation and inflammatory profile of immune cells in the gut and peripheral sites within the host (10).
Recently, it has been proposed that many factors affecting the initiation and course of allergic disorders act within a narrow window of opportunity, either pre-, peri- and/or postnatally (11). Thus, preventive strategies intervening at an early developmental stage seem promising to modulate immune responses with a sustained effect.
The aim of our study was to test a strategy of early prophylaxis of type I allergy based on live recombinant LAB producing a specific allergen. We here show that neonatal mono-colonization of GF mice with the Lactobacillus (L.) plantarum NCIMB8826 strain producing the major birch pollen allergen Bet v 1 attenuates the development of birch pollen allergy later in life. The mechanisms involve a shift towards a nonallergic Th1 phenotype accompanied by increased regulatory responses.
Recombinant Lactobacillus plantarum constitutively producing Bet v 1
Lactobacillus plantarum NCIMB8826 was used as the final host to carry a plasmid for constitutive intracellular Bet v 1 production and grown at 37°C in MRS medium (Difco, Becton Dickinson, Franklin Lakes, NJ, USA). Escherichia (E.) coli MC1061 was used as intermediate host for cloning and cultured at 37°C in Luria broth. Antibiotics were used at following concentrations: for E. coli, ampicillin (100 μg/ml) and for L. plantarum, erythromycin (5 μg/ml). Molecular biology techniques, electrotransformation and cloning techniques were previously described (12). Briefly, a polymerase chain reaction (PCR) was performed on the E. coli expression vector pMW175 carrying the Bet v 1-encoding cDNA fragment using the sense OMEC190 and antisense OMEC191 primers as described in (12). Subsequently, the amplified fragment was subcloned into the pZero-2 plasmid (Invitrogen, Groningen, Netherlands), verified by DNA sequencing, the insert digested with the restriction enzymes NcoI–XbaI and subcloned into the NcoI–XbaI digested plasmid pGIT032 (13) carrying the constitutive L. plantarum promoter for lactate dehydrogenase (pldhL) and an erythromycin resistance gene. The resulting construct (pMEC181) carried the Bet v 1-encoding sequence under control of pldhL. Finally, L. plantarum NCIMB8826 was electrotransformed with pMEC181 giving rise to an erythromycin-resistant L. plantarum constitutively producing intracellular Bet v 1 (recLp). To generate a control strain (conLp), L. plantarum NCIMB8826 was electroporated with the empty plasmid pGIT032. In vitro stability was demonstrated after growth over 100 generations as described (12).
Analysis of Bet v 1 production
Analysis of Bet v 1 production was performed as previously described (12). Briefly, cell extracts of recLp and conLp were mixed with a protease inhibitor (Complete, Boehringer Mannheim, Germany). For western blot, cell extracts or purified Bet v 1 (Biomay, Vienna, Austria) were separated by gel electrophoresis, blotted onto nitrocellulose and detected with a mouse monoclonal anti-Bet v 1 antibody followed by HRP-conjugated goat anti-mouse antibodies. For protein quantification by ELISA, recLp extracts were compared with a Bet v 1 standard curve and expressed as μg/109 colony forming units (CFU) and percentage of protein content measured using a kit (Bio-Rad, Munich, Germany).
Germ-free BALB/c mice were kept under sterile conditions with a 12- h light–dark cycle at 22°C. Sterile pellet diet (ST1, Bergman, Czech Republic) and water were fed ad libitum. Faecal samples were weekly controlled for microbial contamination (14). Experiments were approved by the local ethics committee.
Colonization and in vivo strain stability
Germ-free animals were colonized with 2 × 108 CFU recLp or conLp by intragastric tubing. Drinking water was supplemented with erythromycin (50 μg/ml) to ensure long-term stability of the recombinant strain in vivo. Colonization and presence of the plasmid pMEC181 were analysed weekly by plating on MRS agar ±5 μg/ml erythromycin and expressed as log CFU/g faeces ± standard error of the mean (SEM). Conformational integrity of the protein was assessed by western blot analyses of 2 × 109 CFU recLp freshly grown from faeces.
Immune responses induced by colonization with recLp (experiment I)
Eight-week-old GF mice were inoculated with recLp and mated 20 days later (Fig. 1A). The offspring (n = 6) mono-colonized with recLp via their mothers (group 1) was killed at 56 days to analyse cellular and humoral immune responses in sera, spleens and small intestines. Age-matched GF animals served as controls (group 2).
Influence of recLP on birch pollen allergy (experiment II)
Mice neonatally mono-colonized with recLp (group 1), conLp (group 2) or noncolonized GF mice (group 3) were intra-peritoneally (i.p.) sensitized at 56 days of age three times in a 10-day interval using 1 μg Bet v 1 adsorbed to 2 mg aluminium hydroxide (Serva, Heidelberg, Germany) in 200 μl saline (Fig. 1B). Mice were killed 7 days after the last immunization. Age-matched untreated GF mice served as controls when appropriate. Experiments were performed in two independent sets of 5–7 mice/group, and results were pooled for analysis.
At the killing process, serum was collected for antibody analyses. Pooled mesenteric lymph nodes (MLN, experiment II only) and spleens were removed for in vitro cytokine assays and real-time RT-PCR (a spleen segment was stored in RNAlater® Solution (Ambion, Austin, TX, USA) at −20°C until RNA isolation). Spleen and MLN cells were prepared as previously described (15). Small intestines were excised, weighed and gut lavages performed for the determination of Bet v 1–specific IgA (12).
Cellular immune responses
Cytokine production upon in vitro Bet v 1 restimulation of spleen or pooled MLN cells was assessed as described in (15). Briefly, 5 × 106 cells/500 μl were incubated in 48-well plates (Corning, Oneonta, NY, USA) ± Bet v 1 for 48 h. Interleukin (IL)-4, IL-5, IL-10 and interferon (IFN)-γ were measured from supernatants using ELISA kits (RnD, Mineapolis, MN, USA) with sensitivities <5 pg/ml (IL-4, IL-5), <12 pg/ml (IL-10) and <2 pg/ml (IFN-γ) and reported in pg/ml after subtraction of baseline levels of nonstimulated cultures. Values below assay sensitivity were considered “not detectable” (n.d.).
Humoral immune responses
Allergen-specific serum IgG1, IgG2a and IgA levels were determined by ELISA as described in (15). Briefly, 96-well microtiter plates (Nunc, Roskilde, Denmark) coated with Bet v 1 were applied with sera at appropriate dilutions (experiment I: 1/10 for IgG1, IgG2a, IgE and IgA; experiment II: 1/10000 for IgG1, 1/100 for IgG2a and 1/10 for IgE, 1/10 for IgA in both serum and intestinal lavages). Rat anti-mouse IgG1, IgG2a and IgA antibodies (1/500, Pharmingen, San Diego, CA, USA) and peroxidase-conjugated mouse anti-rat IgG antibodies (1/2000; Jackson, Immuno Labs., West Grove, PA, USA) were used for the detection. As it was shown that allergen-specific IgG interferes with allergen-specific IgE detection (16), total IgE was additionally measured using an ELISA kit (Bethyl, Montgomery, TX, USA) Antibody levels were reported as optical density or absolute units (total IgE). Bet v 1–specific IgA in intestinal lavages was reported as optical density/gram.
RNA was isolated from spleens of recLp colonized or GF mice sensitized to Bet v 1 (groups 1 and 3, experiment II) using a RNeasy® Minikit (Quiagen, Valencia, CA, USA). After DNase treatment, RNA integrity and purity was determined by gel electrophoresis and photometry (260/280 nm). Reverse transcription into cDNA was performed using oligo(dT)15 primers (ImProm-IITM Reverse Trancription System, Promega, Madison, WI, USA). Universal Probe Library (Roche, Mannheim, Germany) probes were used for the quantification of TGF-β1 (UPL#15), IL-10 (UPL#13), Foxp3 (UPL#13) and β-actin (UPL#101). Gene expression was determined using FastStart TaqMan® Probe Master Mix (Roche) at 95°C for 10 min, followed by 45 cycles of 15 s at 95°C and 30 s at 60°C. Relative quantification was performed using GenEx software (MultiD Analyses AB, Göteborg, Sweden).
Statistical analysis was performed with one-way analysis of variance (anova) followed by Student’s t-test for comparison between groups. Data were expressed as mean ± SEM.
Recombinant L. plantarum constitutively produces Bet v 1in vitro and after colonization
Bet v 1 detected in cell extracts of recLp, but not the control strain is shown in Fig. 2A. 109 CFU recLp contained 0.8 ± 0.3 μg of Bet v 1 corresponding to 0.5–0.7% of total soluble cellular protein. The in vivo presence of recLp and pMEC181 in recLp in experiment II was stable, as demonstrated by equal bacterial counts seen after plating in the presence and absence of antibiotics (Fig. 2B) and preserved Bet v 1 production (Fig. 2B, insert).
Mono-colonization with recombinant L. plantarum induces allergen-specific cellular Th1 but no humoral response in naïve mice (experiment I)
Spleen cells from mice mono-colonized with recLp displayed high IFN-γ production upon Bet v 1 stimulation in vitro, while Th2 cytokines (IL-4 and IL-5) were undetectable (Table 1). IL-10 production did not differ from noncolonized GF mice. At the humoral level, no induction of allergen-specific antibodies (IgG1, IgG2a, IgA or IgE) or changes in total IgE levels were detectable in sera of mono-colonized mice compared with GF controls. Furthermore, no significant production of allergen-specific specific IgA was detectable in gut lavages.
|Colonization||Cytokines (pg/ml)||Serum||Gut lavage (OD)|
|recLp||419.3 ± 48.8*||n.d.||n.d.||69.1 ± 5.5||0.34 ± 0.06||0.16 ± 0.01||0.50 ± 0.11||0.06 ± 0.01||222.8 ± 35.4||0.10 ± 0.0|
|None||5.7 ± 4.9||n.d.||n.d.||52.3 ± 10.6||0.29 ± 0.04||0.19 ± 0.03||0.44 ± 0.08||0.06 ± 0.02||224.4 ± 28.6||0.09 ± 0.0|
Mono-colonization with recombinant L. plantarum suppresses allergen-specific antibody responses in Bet v 1–sensitized mice (experiment II)
Allergen-specific IgG1 and IgG2a responses induced by sensitization with Bet v 1 were significantly suppressed by mono-colonization with recLp, but not with conLp or in naive GF mice (Fig. 3E, F). Allergen-specific IgE was reduced, and total IgE levels were significantly suppressed (Fig. 3H, G). No significant changes occurred in allergen-specific IgA in sera (group 1, recLp: 0.32 ± 0.05 OD, group 2, conLp: 0.25 ± 0.03 OD, group 3, GF sens: 0.27 ± 0.03 OD, group 4, GF naive: 0.02 ± 0.01 OD, experiment II) or intestinal lavages (group 1, recLp: 0.09 ± 0.02 OD, group 2, conLp: 0.11 ± 0.03 OD, group 3, GF sens: 0.13 ± 0.02 OD, group 4, GF naive: 0.05 ± 0.02 OD, experiment II).
Mono-colonization with recombinant L. plantarum suppresses allergen-specific Th2 responses in a model of type I allergy (experiment II)
In spleen cell cultures, IL-4 and IL-5 production (Fig. 3A, B) was significantly suppressed and IFN-γ (Fig. 3C) levels were increased in mice colonized with recLp, but not the control strain (conLp) compared with GF animals sensitized to the allergen (GF sens). IL-10 levels did not differ statistically between the groups (Fig. 3D).
Also in pooled MLN cells, suppressed IL-5 and enhanced IFN-γ production was noted in mice mono-colonized with recLp and sensitized to the allergen. IL-4 production was not detectable. Interleukin-10 seemed reduced in mice colonized with recLp or conLp compared with GF-sensitized mice (GF sens) (Table 2).
|Experimental set-up||Cytokines (pg/ml)|
|recLp (group 1) + sens||36.0 ± 0.2||n.d.||16.4 ± 3.4||14.1 ± 3.6|
|conLp (group 2) + sens||17.8 ± 0.7||n.d.||53.1 ± 22.5||17.2 ± 4.0|
|None (group 3) + sens||13.1 ± 4.9||n.d.||58.5 ± 20.4||30.7 ± 4.1|
|None (group 4)||9.5 ± 5.6||n.d.||5.4 ± 3.8||18.9 ± 1.6|
Regulatory markers are increased in spleen cells of mice mono-colonized with recombinant L. plantarum (experiment II)
A significant up-regulation of mRNA of the regulatory marker Foxp3 in spleen cells of mice mono-colonized with recLp was detectable compared with GF-sensitized controls. (Fig. 4). IL-10 and TGF-β mRNA was enhanced, though without statistical significance.
The hygiene hypothesis (16) has brought up new perspectives for the prevention of allergy preparing the ground for application of LAB for modulation of the intestinal microflora to influence allergy development (7). Furthermore, LAB have been demonstrated to act as potent mucosal adjuvants and/or antigen-delivery systems (6). In the light of this, recombinant LAB strains genetically engineered to produce and deliver allergens to mucosal surfaces to induce tolerance have emerged as new tools for mucosal intervention against type I allergy (5, 12, 17–24). We previously evaluated two LAB strains, L. plantarum and L. lactis, for their potential to modulate allergic immune responses in a model of birch pollen allergy in conventional mice, showing a strong Th1 inducing capacity and induction of counter-regulatory immune responses when admixed with Bet v 1 (25). Aiming at standardized antigen delivery to the mucosal surfaces by the bacteria, we constructed a recombinant LAB producing Bet v 1. Accordingly, prophylactic intranasal application of the recombinant strains succeeded in suppressing systemic and local (within the airways) allergen-specific Th2 responses in adult mice (12).
Evidence is accumulating that programming of the immune system, particularly with respect to tolerance induction, starts already at birth and is under close control of the maternal immune system (11), suggesting this period important for processes of immuno-programming. To take advantage of this early “window of opportunity” (11) for modulating the onset of type I allergy, we mono-colonized GF mice from birth with a Bet v 1 producing L. plantarum by natural colonization via their mothers. To be used for neonatal colonization, our previously published recombinant L. plantarum strain (12) seemed unsuitable, as it carries a plasmid with an inducible promoter that is only highly active in vitro, but not under colonizing conditions. We therefore recloned the bet v 1 gene under the control of a strong promoter leading to constitutive expression of the allergen. After demonstrating stable colonization and excluding a loss of the plasmid in vivo, we first analysed the effects of neonatal colonization on the immune system in the absence of allergic sensitization. At the cellular level, the recombinant strain induced a nonallergic Th1 response with significant IFN-γ but absent IL-4 and IL-5 production in spleen cells upon allergen challenge in vitro. These findings differ from Hazebrouck et al. (26) who found that colonization of adult GF mice with a Lactobacillus casei secreting beta-lactoglobulin leads to both Th1 and Th2 cytokine production, which may suggest an advantage of early neonatal intervention over application of recombinant LAB in adulthood. At the humoral level, we did not detect the induction of allergen-specific serum antibodies, which is similar to findings in adult mice colonized with recombinant LAB (26). A report by Dahlman et al. (27) showing that colonization with recombinant Gram-negative bacteria–induced allergen-specific IgE underlines the suitability of Gram-positive LAB as delivery vector for interventional strategies against allergy (25). However, significant strain-specific differences are even the case within these Gram-positive probiotics, as diverse effects on allergic sensitization have been shown after ingestion (28) or sublingual application (29) of a panel of probiotic LAB strains, which argues for careful strain selection. When neonatally colonized mice were subsequently sensitized with the birch pollen allergen Bet v 1, we found suppressed allergen-specific Th2 responses (IL-4/IL-5) and enhanced levels of IFN-γ in spleen and mesenteric lymph nodes. In contrast to our previous studies in conventional mice using repeated oral applications of recombinant LAB that did not succeed in a suppression of Th2 responses (30), the early and long-term intervention used here thus seems to be more efficient.
Notably, the wild-type L. plantarum itself did not exert any suppressive effects on the allergic immune response. Therefore, the effects elicited by our vector system depended on the expression of the specific allergen and indicates that the bacteria represent an adjuvant system without immunosuppressive effects in the absence of expression of a specific allergen, as shown in previous studies with this strain (12, 25). This aspect seems particularly important for interventions at early developmental stages.
The mechanisms of immunomodulation/-regulation exerted by different recombinant LAB strains are complex. A shift towards Th1 (5) but also induction of regulatory cells (5, 22) were described. As we found both a shift towards a Th1 cytokine pattern and the suppression of antibody responses and Th2 cytokine responses, we also examined regulatory markers. According to previous findings in adult mice tolerized against Bet v 1 (31), we found increased mRNA expression of Foxp3 in splenocytes of neonatally colonized and sensitized mice. This finding suggests that the induction of counter-regulatory cellular Th1 responses was paralleled by development of regulatory T cells after colonization from birth and bacterial persistence in the gut.
Taken together, mono-colonization with recombinant L. plantarum specifically induced immunomodulation of allergic immune responses, supporting the concept of early immuno-imprinting using recombinant LAB. Our results therefore promote the use of recombinant lactic acid bacteria for early allergen-directed prevention of type I allergy.
This new research area of immuno-imprinting would, however, definitely benefit from examination of the predictive value of mouse models, e.g. by analysing correlations between preclinical and clinical data. Moreover, aiming at treating neonates with a positive family history of atopy, suitable treatment protocols including environmental containment strategies and a broader range of protection to a panel of allergens are desirable.
We thank Mrs. Jarkovská, Smolová, Grimová and Ságnerová for technical assistance. Supported by grants: IAA500200710 (AS CR), 303/09/0449 (Czech Science Foundation), 2B06155 (MEYS) and Institutional Research Concept AVOZ50200510.
Conflict of interest
All authors have contributed to the work; the authors have no conflict of interest.