Probiotic strains with designed health properties

Exploitable results

The fact that probiotics may influence the intestinal physiology through modulation of the endogenous flora or the immune system is presently well recognised. For example, numerous observations support the claim that given LAB strains may act as immunomodulators (i.e. may stimulate the immune system or reduce abnormal situations such as inflammation), a property that seem to vary substantially between species and perhaps between strains belonging to the same species. As for other probiotic traits, there is virtually no understanding of the mechanisms accounting for the observed effect and of the bacterial components involved in the cross-talk between a probiotic and the host. As a consequence it is almost impossible today to propose a rational screening procedure to select new and efficient probiotic strains. As the bacterial cell wall is probably recognized at first by the intestinal immune system, mutants affected in their cell wall structure were constructed in order to verify if modification of the lipoteichoic acids leads to altered immunomodulation capacity. Mutations in the alanine racemase encoding gene (Alr- mutant) or in the D-alanine incorporation operon (Dlt- mutant) were studied in two genetic backgrounds, i.e. Lactobacillus plantarum and Lactococcus lactis. These strains were chosen because they differ substantially in their capacity to survive in the mammal gastro-intestinal tract. The cell wall mutants were first compared in vitro to their wild type counterpart, mainly for their capacity to induce secretion of specific cytokines from human peripheral mononuclear cells (PBMC). This led to the identification of two types of interesting candidates that presented somehow opposite immune properties. Thus strains with opposite cytokine stimulation profile were selected to analyse their in vivo effect in mouse models mimicking human diseases. First, the Dlt- mutant of L. plantarum (defective in D-Ala incorporation in teichoic acids (TA)) turned out to be much more anti-inflammatory than the parent strain, both in vitro and in vivo. Experiments conducted in the mouse TNBS colitis model proved it to be equally protective as a L. lactis secreting mIL-10. The second L. plantarum cell wall mutant selected for its altered immunological properties was the Alr- mutant that produces modified TA and peptidoglycan (PG) as a results of a defect in Alanine racemase. This mutant was used as live vaccine delivery vehicle, first with a potent immunogen, i.e. the C subunit of Tetanus toxin (TTFC). Mice immunisation experiments by the intragastric, intravaginal and rectal routes proved that it behaved as a substantially improved antigen delivery system as compared to the WT strain, especially leading to enhanced mucosal immune responses. The Alr- mutant of L. plantarum and L. lactis behave similarly. This remarkable property was confirmed by the construction and immunological evaluation of an L. plantarum Alr- recombinant strain producing the UreB antigen of H. pylori. From all the candidate anti-Helicobacter strains constructed within the project, only the Alr-/UreB strain was shown to partially protect against a Helicobacter challenge, upon intragastric immunization and in absence of mucosal adjuvants. Quite surprisingly, the L. lactis Dlt- mutant exhibited different properties than the equivalent mutant in L. plantarum, as it turned out to be much more immunogenic than its wild type counterpart. In summary, cell wall mutants have been engineered that exhibit significantly enhanced immune properties as compared to the wild type strains. One strain corresponds to a much more potent anti-inflammatory candidate that induced equivalent protection in experimental colitis as a strain producing a heterologous immune modulator, i.e. IL-10. The three other ones present a remarkably enhanced capacity to deliver antigens at mucosal surfaces, thus representing improved candidates to develop live mucosal vaccines. This work allowed us to highlight the importance of the TA and PG composition in the immune modulation properties of given strains, and to identify host factors involved in their recognition. In addition, the Alr- mutants which are depending on D-Alanine for growth, should behave as biologically contained strains.

- Construction and evaluation of anti-inflammatory strains In this part of the work emphasis was on engineering different Lactobacillus strains to secrete murine interleukin-10. Five strains, namely Lb. casei LMG6904, Lb. casei BL23, Lb. plantarum NCIMB 8826, Lb. plantarum LMG 9211 and Lb. rhamnosus LMG 10770 could be transformed and secreted mIL10 . In all cases, the expression module consisted of the constitutive lactococcal P1 promoter and the gene usp-45 secretion leader sequence fused in frame to the encoding mature mIL10 and was located on the high copy number plasmid, pT1mIL10, which specifies resistance to erythromycin. Initial tests with intragastric administration of recombinant Lb. plantarum NCIMB 8826 in the DSS model for chronic colitis set up in Balb/c and C57Bl6 mice did not result in a clear improvement of the anti-inflammatory properties. In collaboration with IPL we have introduced the trinitrobenzene sulfonic acid (TNBS)-induced colitis model for evaluation in our animal facility. Interestingly, while the recombinant Lc. lactis strain ssequence ecreting mIL-10 protected against colitis and was thus largely superior to its wild type counterpart, this was not observed in the case of Lb. plantarum. This result can best be explained by the fact that the protective effect of IL-10 delivery is very dose dependent. Optimisation of the treatment regimen with Lb. plantarum/mIL-10 is thus necessary before making any firm conclusion. The strains Lb. casei BL 23, Lb. plantarum NCIMB 8826 and Lb. rhamnosus LMG 10770 were further analyzed in greater molecular detail. Using PCR-based detection of the pT1mIL10 plasmid we established that all three strains stably maintained the plasmid after storage at -70°C in MRS medium containing 20% glycerol. The plasmid copy number was essentially the same in all three strains and comparable to that observed in Lactococcus lactis. All three strains secreted biologically active IL-10, albeit in lower amounts than L.lactis. - In vivo analysis of the recombinant strains With the aim of following the fate and transit of a bacterial inoculum trough the intestine we constructed a plasmid that expresses the Green Fluorescent Protein (GFP). The GFP gene was cloned under control of the lactococcal P1 promoter. Commercially available variants of GFP, which should give improved fluorescence, could not be successfully used in L. lactis. We then constructed a synthetic GFP gene (GFPma), in which the codons were adapted to the preferred A/T rich codon usage of lactobacilli. Using fluorescence microscopy we could now clearly detect individual bacterial cells from overnight cultures of L. lactis MG1363, transformed with plasmid pP1GFPma. With support from the EU officer, it was decided to extend the collaboration with IPL to the micro array analysis of the colon samples of mice receiving anti-inflammatory LAB strains prior to TNBS colitis induction. Experiments are still in progress. - Construction of a biologically contained strain The original aim was the elimination of antibiotic selection markers from the recombinant strains by targeting the expression units to non-essential genomic regions of the lactobacilli (in collaboration with UNISI). In the course of the project it was decided to adapt a different strategy that had already been initiated within another framework at VIB. Rather than merely removing the antibiotic resistance marker, we had set out to construct a GMO that incorporated biological containment features. The target strain was L. lactis MG1363, secreting human IL-10. Our approach was to exchange the chromosomal thymidylate synthase gene, thyA, for the expression cassette encoding hIL-10. The rationale for choosing the thyA gene derives from the observation that starvation for thymine in a thyA mutant is bactericidal rather than bacteriostatic (as commonly observed for most other auxotrophies). We cloned and sequenced the thyA gene from L. lactis MG1363 together with its upstream and downstream flanking regions. We further constructed conditional-replicative plasmids carrying a series of hIL10 expression cassettes flanked by 1 kb of the regions upstream and downstream of the MG1363 thyA gene. These were used to obtain L. lactis strains in which the thyA gene was replaced by the hIL10 gene following double homologous crossover. A representative strain, designated Thy12, contains the usp45-hIL10 sequence downstream of the MG1363 thyA promoter and secretes hIL-10. Survival of the strain is strictly dependent on the availability of thymine or thymidine. Accumulation of the genetically modified strain in the environment is very unlikely, as rapid death occurs after thymidine starvation.

To verify the possibility of using the S-layer protein of Lactobacillus crispatus M247 as a fusion partner for heterologous gene expression in Gram-positive bacteria, the antiviral polypeptide cyanovirin-N (CV-N) was expressed as fusion with the S-layer in Streptococcus gordonii. For this purpose we included in the genetic fusion the signal sequence and transcription/translational signals of the S-layer gene. Recombinant CV-N/S-layer proteins included different portions of the S-layer fusion partner, and were expressed in the supernatant of recombinant S. gordonii. To construct CV-N/S-layer fusion, the CV-N gene was cloned in frame into the S-layer gene in plasmid pSMB366. Different genetic fusion were constructed, using unique restriction sites within the S-layer for cloning CV-N. The CV-N gene was PCR-amplified from plasmid pelB/pET26b(+) by using two different set of primers and cloned in pSMB366. This plasmid is a derivative of the previously developed vector pSMB47, and contains the S-layer gene with its original promoter and terminator sequence. The integration of this vector into the chromosome of S. gordonii was based on the homologous recombination into the tet(M) gene of the broad range transposon Tn5253 of the recipient strain. Two recombinant plasmids were obtained, depending on the restriction sites used for cloning, which were named pSMB392 and pSMB393. Both plasmids included the S-layer signal sequence responsible for the export of the protein at the cell surface, and the S-layer transcription/translational signals. These plasmids were used to transform S. gordonii GP201, and transformants were selected. The recombinant S. gordonii strains were subjected to cell fractionation, and culture supernatants were concentrated by filtration. Expression of CV-N/S-layer fusions in recombinant S. gordonii was assayed by Western blot analysis, using both anti-CV-N rabbit serum diluted 1:1000, and anti-S-layer rabbit serum diluted 1:3000. Two recombinant S. gordonii strains were constructed, expressing CV-N as a fusion with different portion of S-layer protein. In Western blot analysis reactive bands were detected in GP1405 and GP1406 culture supernatant, using both anti-CV-N and anti-S-layer rabbit sera. The estimated molecular weights of the bands corresponding to mature proteins produced by GP1405 and GP1406 are 21 and 46 kDa, respectively, according to those deduced from the amino acid sequence (Fig. 1B). Only negligible traces of the recombinant proteins were found in surface associated protein fraction. Both the two fusion proteins, which included different portion of the S-layer, were expressed efficiently in S. gordonii. Therefore, shortening the fusion partner did not affect significantly the expression of recombinant proteins. These results demonstrated that the S-layer of L.crispatus M247 could be used as a fusion partner for heterologous gene expression. This strategy may be widely applicable for expressing heterologous proteins in commensal lactic acid bacteria as a possible approach for local delivery of bioactive molecules.

Information collected from the in vitro and in vivo tests during the DEPROHEALTH project served as a basis to establish an interactive polyphasic database, implemented on BioNumerics software(Applied Maths BVBA, Belgium) which allows to: - determine ranges for each parameter measured, - identify links between parameters measured (useful in revealing possible mechanisms of activity) and - facilitate future comparison of probiotic strains with - the potential to predict relevant functionalities. Possible application: The comparison of the results obtained after screening for some critical parameters to a database of well-known reference strains (supported by literature data or clinical information) may be indicative for potential functionality. This results for the producer in: - An early detection of less performing strains, saving money through the use of relatively simple in vitro or in vivo techniques and helping to focus clinical trials. - Early orientation during the preparation of the scientific dossier of a probiotic strain; again saving time and costs. - Clear referencing of the position of their own strains in the range covered by other probiotic strains (reference strain identity can be kept anonymous at any time). For the consumer this means: - A probiotic strain with better performance. - To be assured to consume a strain with a scientific dossier, build on a sound screening approach. - Less contradictory information and more robust claims for specific applications. Limitations: - The value of this database is linked to the use of standardised laboratory methods. Using these established methodologies it is possible to investigate new or existing strains. However, although the intra-laboratory reproducibility was evaluated, the inter-laboratory reproducibility of the methods used has not sufficiently been studied at this stage. This can limit the current portability of the database as part of a wider laboratory network. - For some of the currently used techniques, a final validation by a specific clinical trial is desirable. This trial should ideally compare a placebo to 2 strains with opposite properties (according to the screening protocol used). The outcome of the clinical trial should confirm the predicted efficiency. Strains successfully used in the past for a particular application can be compared with strains that have not been used successfully before. Unless clearly explainable, the results of the 'a posteriori' screening test should not be in disagreement with former clinical findings. Dissemination and use potential: - Despite the uncertainty of the current portability of the database as part of a wider laboratory network, implementation in a service activity at one single institute (for industry, consumer organisations, etc...), is surely feasible and can be implemented immediately. - When inter-laboratory reproducibility permits, the use of specialised software such as BNServer, directly linked to the database, allows external laboratories to link to a central database, which will automatically perform identification and reporting of ranges for any particular test. The software exists (several applications known) and can be configured to allow registered, password protected and encrypted access and, based on a log system can prepare for invoicing. - The current 'probiotic' functionality can be extended to include safety parameters, technological (fermentation, organoleptic) or even clinical parameters. - The databasing potential, when demonstrated for 'probiotics' can be extended to other food microorgansims, including food pathogens and possibly linked to existing European initiatives (e.g. food safety issues: Workpackage 4 in Med-Vet-Net, the EnterNet (http://www.hpa.org.uk/hpa/inter/enter-net_ menu.htm), the SalmGene (http://www.salmgene.net/) and ListerNet (http://www.invs.sante.fr/publications/2004/listernet).) Current status The database has been created. An initial set of data has been loaded (commercial and scientific reference strains). A possible partner, willing to implement the BNServer part has been identified. Financing is needed for: - setting up the proper login and safety scripts, - testing inter-laboratory reproducibility, - extending the database with strains and experimental protocols to cover a wider potential of probiotic characteristics (lab partners have been identified for safety parameters and for technological parameters). Current status The database has been created. An initial set of data has been loaded (commercial and scientific reference strains). A possible partner, willing to implement the BNServer part has been identified. As the result of an EOI (Standardisation of screening- and evaluation tools for probiotic safety and functionality assessement, ProbioStandard) a large number of interested laboratories have been identfied. (Community) financing is needed for: - testing inter-laboratory reproducibility; - extending the database with strains and experimental protocols to cover a wider potential of probiotic characteristics (lab partners have been identified for safety parameters and for technological parameters); - buying and setting up the software with proper login and safety scripts. The comparison of the results obtained after screening for some critical parameters to a database of well-known reference strains (supported by literature data or clinical information) may be indicative for potential functionality. This results for the producer in: - An early detection of less performing strains, saving money through the use of relatively simple in vitro or in vivo techniques and helping to focus clinical trials. - Early orientation during the preparation of the scientific dossier of a probiotic strain; again saving time and costs. - Clear referencing of the position of new strains in the range covered by other probiotic strains (reference strain identity can be kept anonymous at any time, allowing industry to estimate chances of success upon commercialisation. For the consumer this means: - a probiotic strain with better performance, - to be assured to use a strain with a scientific dossier, build on a sound screening approach - more robust claims for strains that are targeted to a specific application. (Community) financing is needed for: - testing inter-laboratory reproducibility (minimum set of high quality laboratories with long time relevant experience; many potential labs identified through former EOI ProbioStandard). - extending the database with strains and experimental protocols to cover a wider potential of probiotic characteristics (lab partners have been identified for safety parameters and for technological parameters). - buying and setting up the software with proper login and safety scripts (SME has been identified).

Result description: The rotaviral antigen VP8* was expressed in probiotic lactobacilli, such as L. casei and L. plantarum, as well as in Lactococcus lactis alanine racemase mutant (alr-). The VP8* fragment was expressed from different promoters and on different cell compartments (cytoplasmatic, cell-wall anchored and secreted). All of them were tested in a mouse model using cholera toxin subunit B (CTB) as an adjuvant. Female Balb/c mice were given intragastrically 1x109c.f.u. of each of recombinant L. lactis and L. casei strains producing VP8* and a significant local immune response was achieved. IgA antibodies against VP8* were detected in stools, intestinal fluids and cultured Peyer's patches at higher levels in mice given recombinant bacteria secreting VP8*. Protection against rotavirus infection in the adult mouse model was evaluated using the murine EMcN and the simian SA11 rotavirus strains as the challenging viruses. Reduction in viral antigens shed in stools suggests a partial homotypic protection. Antibodies against the VP8* fragment of the rotavirus spike protein VP4 have been shown to block virus entry into the enterocytes and to protect against rotavirus gastroenteritis. Following this rationale, anti-VP8* single chain antibodies (scFv) have been isolated in Escherichia coli and cloned in L. casei, this constitutes a complementary approach with therapeutic implications. Also active anti-NSP4 scFv clones have been isolated, being NSP4 the first viral enterotoxin described. In vivo trails with these synthetic antibodies could be performed in the future. Potential use and expected benefits: Rotavirus is the most important viral agent causing gastroenteritis worldwide and kills 440,000 children each year in developing countries. Therefore, the development of a cost-effective vaccine vehicle, easy to administrate and resistant to harsh manipulation/transport conditions, together with appropriate nutrition and vaccination programmes could notably help to paliate the effects of these infections. The development of lactic acid bacteria vehicles to deliver rotaviral antigens could fit this profile. Furthermore, probiotic strains of Lactobacillus casei with demonstrated anti-rotavirus activity, as well as recombinant strains expressing anti-rotaviral scFv, could reinforce the protecting effect of the recombinant vaccine. Dissemination: In addition to scientific publications, the existence of this project has been communicated through two personal interviews with local journalists, submission of the project profile to various databases and the corporative web pages of the University of Valencia and Instituto de Agroquímica y Tecnología de Alimentos (CSIC). Additionally, the project structure and our results have been communicated to the local Food Products Fair Interaliment (2001). Key innovative features of the results: The results obtained have several relevant innovative aspects related to the location and expression of heterologous proteins in lactic acid bacteria and to the immune response in mice. Difficulties were found to achieve a significant expression of rotavirus capsid protein in lactic acid bacteria, followed by little success in the first immunisation assays in mice. It was evidenced that previously published successful models might not be "the rule". However, this forced the team to test a number of expression vectors, bacterial hosts and inoculation protocols, adjuvants and even mice strains, until the system was put to work. Also, this work reported the isolation and cloning of anti-viral scFv in lactic acid bacteria. Current status and use of results: At present, the last protection experiments are performed in mice, however, in vivo tests with the cloned antibodies or other likely therapeutic molecules are yet to be done. After publication of the results, it would be the time to seek further financing sources and a potential industrial partner interested in the development of further research leading to particular applications in humans and possibly in farm animals. Our research work has proved that an oral immunisation method in mice can be developed for rotaviral antigens that initially offered serious difficulties. Despite the fact that other antigens secreted by lactic acid bacteria have been proven to elicit systemic and mucosal immune response, the underlying mechanisms are totally unknown. Particularly, we would seek collaboration with immunologists to determine the causes of the null mucosal immune response to the VP8* peptide, either purified or expressed in lactic acid bacteria. Since differences in the behaviour and response in different animal species could be expected, further research and networking would also allow to adapt and test the system in farm or experimental animals other than mice, before considering human trails. Existing "clean" recombination methods can be implemented to yield antibiotic-free and other environmentally safe strains (i.e., self-contained). The methods developed and knowledge obtained will facilitate the development of live innocuous bacterial vectors to deliver other antigens (as mucosal vaccines) or therapeutic molecules for many other applications. They seem to be particularly suited for the prevention of gastrointestinal disorders or mucosal infections.

Real-time quantitative PCR is a sensitive method to assess H. pylori load in infected mice, but there is a need to standardise the preparation of DNA from gastric specimens. In the study, mouse stomachs were homogenised by: - complete disruption using a blender (Ultra Turrax) or, - by vortexing with glass beads (1mm diameter). Each procedure was followed by DNA purification by one of three different protocols - two commercially available kits - Qiagen DNA Tissue kit and Qiagen Stool Kit or a phenol-chloroform extraction method. Homogenisation with glass beads followed by the Tissue kit was found to be most suitable protocol combining high extraction and detection efficiency of 16S rDNA. Analyses of PCR inhibition showed a strong correlation with DNA concentrations and indicated that inhibition was probably due to substances co-purified with DNA rather than excess of mouse non-target DNA. To overcome PCR inhibition in samples produced by vortexing and the Qiagen Tissue kit, a 10-fold dilution should be applied before the 16S PCR assay. The procedure was validated by quantifying H. pylori DNA in infected mice and allowed accurate detection of H. pylori DNA even in the case of successfully immunised mice. When infection rates are to be determined, it is important to correct the 16S H. pylori DNA quantities with the host DNA quantities determined as total DNA or as mouse GAPDH DNA. Use of this method should allow reproducible, sensitive and accurate detection of H pylori in mouse stomachs.

Searching for OpenAIRE data...

Exploitable results

Share this page

Download