DECOLONIZATION OF ENTEROBACTERIA, SUCH AS KLEBSIELLA PNEUMONIAE, FROM THE GUT USING STRAINS OF KLEBSIELLA OXYTOCA

Abstract

The present invention relates to probiotic bacteria of the species Klebsiella oxytoca that are used for a decolonization of multidrug resistant (MDR) Enterobacteria, such as Klebsiella pneumoniae (K. pneumoniae), from the gut of a subject. The decolonization can both be therapeutic, i.e. after colonization of the gut by the multiresistant pathogen(s), or as a preventive measure before a re-colonization of the gut, as required after antibiotic treatment.

Claims

1. A method for preventing and/or treating antibiotic resistant Enterobacteriaceae in the microbiome of an animal wherein said method comprises administering to said animal a bacterium selected from Klebsiella oxytoca and species related to Klebsiella oxytoca outcompetes said Enterobacteriaceae for use of beta-glucosidic sugars.

2. The method according to claim 1, wherein said prevention and/or treatment is through decolonization of said Enterobacteriaceae.

3. The method according to claim 1, wherein said species related to Klebsiella oxytoca is selected from the group consisting of K michiganensis, K. grimontii, K. aerogenes, and Klebsiella that have a carbohydrate utilization pattern substantially similar to K. oxytoca strain MK01.

4. The method according to claim 1, which reestablishes colonization resistance after antibiotic dysbiosis in the microbiome of an animal.

5. The method according to claim 1, wherein said microbiome is located in the gut of said animal.

6. The method according to claim 1, wherein said bacterium that is administered is a Klebsiella strain isolated from a human.

7. The method according to claim 1, wherein said bacterium does not exhibit additional antibiotic-resistance in addition to a naturally occurring ampicillin/amoxicillin-resistance phenotype.

8. The method according to claim 1, wherein said prevention and/or treatment of said antibiotic resistant Enterobacteriaceae is in the context of bacterial infection, bloodstream infection, allogeneic hematopoietic stem cell transplantation, and bacterial invasion into the liver, spleen and/or mesenteric lymph nodes (MLN) in said animal.

9. The method according to claim 1, which utilizes fecal microbiota transplant (FMT).

10. The method according to claim 1, wherein said administration is in a dosage of between 10.sup.12 and 10.sup.7 CFU of said bacterium.

11. The method according to claim 1, wherein said administration is in combination with at least one bacterial species selected from the group consisting of Clostridium spec., Bacteroides spec., Laclobacillus spec., Enterococcus spec., Acutalibacter spec., Bfidobacterium spec., Muribaculum spec., Flavonfractor spec., Akkermansia spec., Turicimonas spec., and Blautia spec., Firmicutes spec.

12-13. (canceled)

14. The method according to claim 1, wherein the animal is a human.

15. The method according to claim 1, wherein the Enterobacteriaceae is Klebsiella pneumoniae.

16. The method according to claim 1, wherein the bacterium that is administered outcompetes said Enterobacteriaceae for use of sucrose and/or cellobiose.

17. The method according to claim 9, wherein said administration is via sterile saline, a suppository or capsule, or as a probiotic.

18. The method according to claim 17, wherein the capsule is stomach-acid resistant.

19. The method according to claim 11, wherein said use is in combination with at least one bacterial species selected from the group consisting of Clostridium innocuum, Clostridium clostridioforme, Bacteroides caecimuris, Lactobacillus reuteri, Enterococcus faecalis, Acutalibacter muris, Bifidobacterium animalis subsp. animalis, Muribaculum intestinale, Flavonifractor plautii, Akkermansia muciniphila, Turicimonas muris, Blautia coccoides, Firmicutes spec., Blautia coccoides YL58, Enterococcus faecalis KB1 and Clostridium clostridioforme YL32.

20. The method according to claim 19, wherein said use is in combination with at least one bacterial species selected from Blautia coccoides YL58, Enterococcusfaecalis KB1 and Clostridium clostridioforme YL32.

Description

[0045] The present invention will now be described further in the following examples with reference to the accompanying Figures, nevertheless, without being limited thereto. For the purposes of the present invention, all references as cited herein are incorporated by reference in their entireties.

[0046] FIG. 1 shows a diagram of recovered CFUs of K. pneumoniae MDR1 in differentially treated SPF-H mice cecal content. Mean and SEM of two independent experiments with n=4-5 mice per group. p values indicated represent a non-parametric Kruskal-Wallis test *p<0.05. **p<0.01. ***p<0.001. ****p<0.0001.

[0047] FIG. 2 shows a diagram of the colonization kinetics of CFUs of K. pneumoniae MDR1 in differentially treated SPF-H mice. Mean and SEM of three independent experiments with n=4-6 mice per group.

[0048] FIG. 3 shows a diagram of the colonization kinetics of CFUs of K. pneumoniae MDR2 in differentially treated SPF-H mice. Mean and SEM of one independent experiments with n=5 mice per group.

[0049] FIG. 4 shows a diagram of the clearance rates of K. pneumoniae MDR1 in precolonized and control animals. P values indicated represent a log-rank test ****p<0.0001 (control/E. coli 103 vs K. oxytoca MK01), ####p<0.0001 (control/E. coli 103 vs K. oxytoca MK02).

[0050] FIG. 5 shows a diagram of the clearance rates of K. pneumoniae MDR1 according to the examples, below. Mean of two independent experiments with n=5 mice per group. P values indicated represent a log-rank test **p<0.01 (control vs. K. oxytoca MK01), #p<0.05 (control vs K. oxytoca MK02).

[0051] FIG. 6 shows that the protective effect of commensal K. oxytoca is widespread among different K. oxytoca strains. Mean of one independent experiment with n=3-5 mice per group. SUS=susceptible, INT=intermediate, and RES=protected donors.

[0052] FIG. 7 shows the clearance rates of another K. pneumoniae strain (MDR2 ST258) in K. oxytoca precolonized and control miceMean and SEM of one out of two independent experiments with n=5 mice per group. p value indicated represent a log-rank test **p<0.01.

[0053] FIG. 8 shows that K. oxytoca and K. pneumoniae compete for specific carbon sources with a broader substrate range found in protective K. oxytoca isolates. (A) Amount of carbon source with detected growth for different K. oxytoca isolates, K. pneumoniae MDR1 and E. coli 103. Positive growth was defined as OD >0.2 after 24 h of incubation. Strains highlighted with * were used to generate the Venn diagram displayed in (B). (C-D) CFUs of K. pneumoniae MDR1 after 24 h of aerobic cocultivation with selected K. oxytoca or E. coli strains in minimal medium supplemented with Sucrose and Cellobiose as sole carbon sources. Values represent one out of two independent experiments with n=3 replicates per condition.

[0054] FIG. 9 shows that K. oxytoca requires cooperating bacteria from the phylum Firmicutes to reestablish long-term colonization resistance in germfree mice. (A) Resulting CFUs of K. pneumoniae MDR1 after anaerobic co-cultivation with selected Oligo-MM.sup.12 bacteria (V=Verrucomicrobia, B=Bacteroidetes, A=Actinobacteria and F=Firmicutes) and K.oxytoca MK01 after 24 h. Mean and SEM of three independent experiments with n=2-4 samples per group. p values indicated represent a non-parametric Kruskal-Wallis test *p<0.05, ****p<0.0001. (B) CFUs of K. pneumoniae MDR1 in native Oligo-MM.sup.12 and differentially precolonized GF animals at day 21. Mice received selected Firmicutes from the Oligo-MM.sup.12 microbiota two weeks before challenge with K. pneumoniae MDR1. At day 7 mice were additionally precolonized with MK01. CFUs were assessed until day 28 after K. pneumoniae MDR1 colonization. Table below indicates present bacteria in each group. Mean and SEM of one experiments with n=2-5 samples per group.

[0055] FIG. 10 shows fecal colonization kinetics of multi-drug resistant Enterobacteriaceae P. mirabilis (A), E. cloacae (B), and E. coli (C) in MK01 pretreated and untreated control mice over time. Pooled data of two independent experiments with n=6-7 mice per group are displayed. Round circles and straight line: control; triangles and dashed line: MK01.

EXAMPLES

[0056] Methods

[0057] Isolation of Commensal Klebsiella Strains from Human Stool Samples

[0058] Stool samples were homogenized in 1 ml PBS and plated in serial dilutions on Simmons Citrate Agar plates supplemented with inositol, tryptophan and bile salts (SCITB) (Simmons Citrate Agar; Cheng V C C, Yam W C, Tsang L L, et al. Epidemiology of Klebsiella oxytoca-associated diarrhea detected by Simmons citrate agar supplemented with inositol, tryptophan, and bile salts. J Clin Microbiol. 2012; 50(5):1571-1579. doi:10.1128/JCM.00163-12). Yellow, inositol fermenting colonies were picked and plated on LB-amp plates to check for ampicillin resistance. Positive colonies were screened with species specific PCRs for the pehX gene of K. oxytoca (Kovtunovych G, Lytvynenko T, Negrutska V, Lar O, Brisse S, Kozyrovska N. Identification of Klebsiella oxytoca using a specific PCR assay targeting the polygalacturonase pehX gene. Res Microbiol. 2003; 154(8):587-592. doi:10.1016/SO923-2508(03)00148-7) or the internal transcribed spacer (ITS) of the rrn operon for K. pneumoniae (Liu Y, Liu C, Zheng W, et al. PCR detection of Klebsiella pneumoniae in infant formula based on 16S-23S internal transcribed spacer. Int J Food Microbiol. 2008; 125(3):230-235. doi:10.1016/j.ijfoodmicro.2008.03.005).

[0059] In Vivo K. pneumoniae Colonization

[0060] If required, mice received ampicillin (0.5 g/L) for six consecutive days starting three days before colonization with K. pneumoniae. Inoculi were prepared by culturing bacteria overnight at 37 C. in LB broth with 50 g/ml ampicillin and 25 g/ml cloramphenicol. Subsequently, the culture was diluted 1:25 in fresh medium, and subcultured for 4 hours at 37 C. in LB broth. Bacteria were resuspended in 10 ml phosphate-buffered saline (PBS) and required amounts were calculated. Mice were orally inoculated with 10.sup.8 CFU of K. pneumoniae diluted in 200 l PBS. Weight of the mice was monitored, and feces were collected at different time points (1, 3, 6, 9, 14, 21, 28 and 42 days) after colonization for measuring pathogen burden and 16S rRNA sequencing or mice were sacrificed at day 6 to evaluate pathogen burden in the intestinal organs.

[0061] In Vivo Commensal Colonization

[0062] Ampicillin pretreatment and preparation of inoculi was performed as described for K. pneumoniae. Mice were orally inoculated with 10.sup.8 CFU of commensal bacteria (K oxytoca or E. coli) diluted in 200 l PBS. Weight of the mice was monitored, and feces were collected at different time points (1, 3, 6, 9, 14 days) after colonization for measuring bacterial numbers and 16S rRNA sequencing.

[0063] In vivo MDR Enterobacteriaceae colonization (P. mirabilis, E. cloacae, E. coli) SPF-mice pretreated with 10.sup.8 CFUs of K. oxytoca MK01 or untreated control mice reveived ampicillin (0.5 g/L) for six consecutive days starting three days before colonization with multi-drug resistant Enterobacteriaceae (P. mirabilis, E. cloacae, E. co/i). Inoculi were prepared by culturing bacteria overnight at 37 C. in LB broth with 50 g/ml ampicillin. Subsequently, the culture was diluted 1:25 in fresh medium, and subcultured for 4 hours at 37 C. in LB broth. Bacteria were resuspended in 10 ml phosphate-buffered saline (PBS) and required amounts were calculated. Mice were orally inoculated with 10.sup.8 CFU of multi-drug resistant Enterobacteriaceae (P. mirabilis, E. cloacae, E. coli) diluted in 200 l PBS. Weight of the mice was monitored, and feces were collected at different time points (1, 3, 6, 9, 14, 21, 28 and 42 days) after colonization for measuring pathogen burden of mice. See also FIG. 10.

[0064] Mice precolonized with the commensal MK01 isolate showed significantly reduced CFUs of multi-drug resistant bacteria over time compared to control mice. Complete clearance was observed in portions of MK01 pretreated mice for all three multi-drug resistant bacteria but not in any control mice. The efficacy of reduction varied between 100-fold and more than 1000-fold differences at different time points for each pathogen.

[0065] Quantification of Fecal K. pneumoniae Colonization

[0066] Fresh fecal samples were collected, and weight was recorded. Subsequently, fecal samples were diluted in 1 ml PBS and homogenized by bead-beating with 1 mm zirconia/silica beads for two times 25 seconds using a Mini-Beadbeater-96 (BioSpec). To determine CFUs, serial dilutions of homogenized samples were plated on LB plates with 50 g/ml ampicillin and 25 g/ml cloramphenicol. Plates were cultured at 37 C. for 1 day before counting. CFUs of K. pneumoniae were calculated after normalization to the weight of feces. Quantification of K. oxytoca and E. coli was performed accordingly and samples were plated on LB plates with 50 g/ml ampicillin.

[0067] In Vitro Assays

[0068] Mice were sacrificed and cecal content was isolated, weighted and diluted in a 1:1 ratio with PBS or BHI and homogenized for two times 25 seconds using a Mini-BeadBeater-96 (BioSpec). Samples were either prepared aerobically or anaerobically in an anaerobic chamber. Bacteria were grown in appropriate media and normalized to 10.sup.6 CFUs. Tubes were inoculated with 10 l K. pneumoniae (OD 0.2) and cultivated at 37 C. under aerobic or anaerobic conditions for 24 hours. 25 l of each sample was serial-diluted in 96 well plates and plated on selective agar plates with ampicillin and chloramphenicol to recover viable amounts of K. pneumoniae.

[0069] Antibiotics Susceptibility Testing

[0070] For assessment of minimal inhibitory concentrations, the commercial ETEST kit from the vendor Biomerieux was used. Antibiotic susceptibility testing against four lead antibiotics classes was performed according to manufacturer's instructions.

[0071] Statistical Analysis

[0072] Experimental results were analyzed for statistical significance using GraphPad Prism v8.2 (GraphPad Software Inc.). Differences were analyzed by Student's t test and one-way ANOVA. P values indicated were calculated by a non-parametric Mann-Whitney U test or Kruskal-Wallis test comparison of totals between groups (Segata et al. 2011). OTUs with Kruskal-Wallis test <0.05 were considered for analysis. P values lower than 0.05 were considered as significant: *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

[0073] Mouse Model for Experiments

[0074] Pretreatment with ampicillin destroys natural colonization resistance in SPF-H mice and establishes a long term colonization of the pathogen

[0075] As a basis for the experiments, a recently published gut colonization model for K. pneumoniae in mice (Sorbara M T, Dubin K, Littmann E R, et al. Inhibiting antibiotic-resistant Enterobacteriaceae by microbiota-mediated intracellular acidification. J Exp Med. 2018; 216(1):84-98. doi:10.1084/jem.20181639) was used. First, it was verified that this animal model was also suitable to model long term colonization of MDR K. pneumoniae in SPF.

[0076] In-house SPF mice with in-house gut flora (SPF-H) were treated for 6 consecutive days with ampicillin and colonized on day 3 of treatment with 10.sup.8 CFUs K. pneumoniae MDR1, a clinical isolate belonging to ST395 previously described.

[0077] Ampicillin treated mice showed a significant higher pathogen burden especially at the beginning with CFUs reaching up to 109 CFUs/ml compared to 103 to 104 CFUs/ml in control mice. In addition, these mice were unable to clear K. pneumoniae MDR1 even after 6 weeks of colonization suggesting a stable long-term integration of K. pneumoniae MDR1 in the microbiome of the amp-treated mice.

[0078] Weight loss or any sign of infection was absent in both groups, indicating that K. pneumoniae is an asymptomatic colonizer in healthy mice. Next, changes in the microbiota during the six weeks of K. pneumoniae MDR1 colonization were assessed. Regarding species richness, no changes could be observed in the microbiota of control mice, whereas richness of the amp-treated animals was almost reduced to zero after 1 day post colonization (p.c.). After the end of amp treatment, species richness recovered over time almost reaching levels similar to control animals after six weeks. Nevertheless, species evenness in these animals was still lower after 6 weeks of colonization, indicating that some long-term effects due to the antibiotic treatment were still detectable.

[0079] Therefore, the mouse colonization model established a stable long-term colonization in healthy mice without body weight loss or other signs of infection, making it a useful model to study the impact of specific commensal regarding their potential to prevent or decolonize MDR K. pneumoniae from the dysbiotic mouse gut and to re-establish homeostatic microbiome conditions.

[0080] The Oligo-Mouse-Microbiota (Oligo-MM.sup.12) is a community of 12 mouse intestinal bacteria to be used for microbiome research in gnotobiotic mice, and consists of Clostridium innocuum, Bacteroides caecimuris, Lactobacillus reuteri, Enterococcus faecalis, Acutalibacter muris, Bifzdobacterium animalis subsp. animalis, Muribaculum intestinale, Flavonifractor plautii, Clostridium clostridioforme, Akkermansia muciniphila, Turicimonas muris, and Blautia coccoides (Garzetti D, Brugiroux S, Bunk B, et al. High-Quality Whole-Genome Sequences of the Oligo-Mouse-Microbiota Bacterial Community. Genome Announc. 2017;5(42):e00758-17. Published 2017 Oct 19. doi:10.1128/genome A.00758-17).

[0081] The Ampicillin Mouse Model Induces Reproducible and Robust Colonization with Different Klebsiella pneumoniae Strains

[0082] In order to test the suitability of the mouse model with other relevant Klebsiella strains, the above SPF-H mice were colonized with another MDR strain from the highly relevant ST258 (K. pneumoniae MDR2) which is a major spread in the US This strain achieved similar CFUs as the initially tested K. pneumoniae MDR1. None of the mice cleared any of the strains tested even after six weeks of colonization indicating that the ampicillin mouse model works for Klebsiella strains in general. Similarly to what was observed with K. pneumoniae MDR1, mice did not lose any weight for the other strains tested indicating that immunocompetent mice did not suffer from infection with those strains.

[0083] Screening of Human Stool Samples for Commensal Klebsiella StrainsHigh Prevalence of K. Oxytoca in Protected Stool Samples

[0084] Different age groups of healthy volunteers were screened for commensal Klebsiella strains. In total, 32 Klebsiella strains could be isolated from 29 healthy individuals, 16 species from 13 healthy adult volunteers and 16 species from 15 healthy children.

[0085] In order to analyze spread of antibiotic resistance in the Klebsiella isolates, genomes were sequenced using Illumina sequencing and analyzed regarding resistance genes to predict the resistance phenotype.

[0086] As expected, all of the isolates were resistant against ampicillin, confirming the occurrence of natural resistance against the antibiotics class of penicillins in Klebsiella (i.e. the naturally occurring ampicillin/amoxicillin-resistance phenotype). All isolated K. oxytoca strains did not exhibit any further resistance genes, whereas all other commensal K. pneumoniae strains showed at least one additional resistance with some of the strains belonging to ST395 showing even multi-drug resistance. Microbiological resistance testing verified the bioinformatics prediction. Therefore, K. oxytoca exhibited far less wide-spread antibiotic resistance as commensal K. pneumoniae strains. This further supports the concept of the present invention to use K oxytoca strains to fight MDR K. pneumoniae, since transfer of resistance apparently is less common but metabolic and ecologic behavior is similar.

[0087] Addition of K. Oxytoca to Amp Treated Cecal Content Efficiently Reduces CFUs of Recovered K. pneumoniae In Vitro

[0088] In vitro, SPF-H mice were treated with ampicillin for six consecutive days to diminish their natural colonization resistance. One day before sacrifice, mice were treated with K. oxytoca strain isolated from donor MK1903 (K. oxytoca MK01), MK1901 (K. oxytoca MK02) or left untreated as a control. Pre-colonization was assessed the next day by plating the isolated cecal content on selective agar plates. Cecal content was spiked with 10.sup.6 CFUs of K. pneumoniae MDR1 and incubated for 24 h under anaerobic conditions before content was plated on LB-chloramp plates to recover viable amounts of K. pneumoniae MDR1. All animals were successfully colonized with K. oxytoca.

[0089] Ampicillin treatment eradicates most microbes from the SPF-H flora resulting in a significant increase of pH value in all treated groups compared to the untreated control group, which was not affected by the pre-colonization with K. oxytoca. Strikingly, recovered CFUs of K. pneumoniae MDR1 were significantly reduced upon pre-colonization with K. oxytoca (FIG. 1). A strong anti-correlation between the level of pre-colonization with K. oxytoca and the resulting CFUs of MDR K. pneumoniae with a coefficient of determination R.sup.2=0.59 for K. oxytoca MK01 and R.sup.2=0.68 for K. oxytoca MK02 was observed. K. oxytoca therefore has the potential to reduce MDR K. pneumoniae.

[0090] Pretreatment of Antibiotic-Treated Mice with Probiotic K. oxytoca Facilitates Clearance of MDR K. pneumoniae from the Feces

[0091] Next, it was tested whether K, oxytoca could also help to clear MDR K. pneumoniae in antibiotic disturbed microbiota situations in vivo. SPF-H mice were precolonized with K. oxytoca MK01 or K. oxytoca MK02 for two weeks before colonization with 10.sup.8 CFUs of K. pneumoniae MDR1. Mice received ampicillin for 6 consecutive days during each intervention three days before until three days after colonization. Feces were taken after different time points of precolonization and colonization in order to assess levels of K. oxytoca and MDR K. pneumoniae. None of the mice lost body weight upon colonization with K. oxytoca or MDR K. pneumoniae. Colonization kinetics revealed that both groups colonized with K. oxytoca strains exhibit significantly reduced levels of K. pneumoniae MDR1 (FIG. 2).

[0092] K. oxytoca pretreated animals could clear MDR K. pneumoniae faster and up to 80% ( 13/16 animals for K. oxytoca MK01 and 12/15 animals for K. oxytoca MK02) cleared the colonization after 28 days, whereas control animals could only clear the pathogen in 21.4% of the cases ( 3/14 animals). This experiment demonstrates that K. oxytoca efficiently reduces fecal CFUs of MDR K. pneumoniae and supports clearance from the microbiome of colonized animals.

[0093] K. oxytoca Facilitates Clearance of Another K. pneumoniae Strain from the Feces

[0094] To investigate if the effect of K. oxytoca would also lead to the same level of protection against another MDR K. pneumoniae strains, ampicillin treated mice were colonized with K. oxytoca A or left uncolonized as a control. After two weeks, mice were colonized with K. pneumoniae MDR2, and fecal colonization was assessed after different time points of colonization. K. oxytoca MK01 efficiently reduced fecal CFUs of infected mice already by day 1 by a factor of 10,000 (FIG. 3). After 6 days, precolonized mice completely cleared K. pneumoniae MDR2 from the feces, resulting in a clearance rate of 100%, which was even more efficient than initially observed for K. pneumoniae MDR1 with rates above 80%. These experiments revealed that human K. oxytoca strains exhibit a specific effect which is not observed with other commensal related enterobacteria, such as E. coli, and showed a broad-spectrum activity against different MDR K. pneumoniae strains.

[0095] In addition, the question was addressed whether K. oxytoca could also prevent colonization of MDR K. pneumoniae in gnotobiotic and GF mice which are naturally susceptible to colonization. Oligo-MM.sup.12 mice which harbor 12 specific bacterial strains and no other gamma-proteobacteria as well as GF mice, which are devoid of any bacteria in the gut, were colonized with K. oxytoca MK01, K. oxytoca MK02 or E. coli 103 two weeks prior colonization with MDR K. pneumoniae, and fecal colonization was monitored over 28 days. Similar effects could be observed as initially seen in ampicillin treated SPF-H mice.

[0096] Both K. oxytoca strains efficiently reduced fecal CFUs of MDR K. pneumoniae and achieved total clearance after 6 (K. oxytoca MK02) or 9 days (K. oxytoca MK01) post colonization. In contrast, control group and E. coli 103 colonized mice exhibited similar high amount of K. pneumoniae during the whole course of colonization without any reduction or clearance (FIG. 4). This experiment highlighted that K. oxytoca is also capable to protect gnotobiotic mice with a defined microbiome composition from colonization with MDR K. pneumoniae indicating that the protection might occur via a direct competitive mechanism between the related Klebsiella species.

[0097] To further test that K. oxytoca actively reduces K. pneumoniae, GF animals were colonized with K. oxytoca MK01 and E. coli 103 prior to colonization with K. pneumoniae or mice were left untreated. Body weight loss and fecal burden was assessed over 4 weeks. All colonized mice gained up to 20% of weight during the course of experiment indicating that colonization with other enterobacteria supported the proper development of the animals. In terms of fecal colonization with MDR K. pneumoniae significantly reduced CFUs in the K. oxytoca MK01 treated animals were found, and to a lower extend also in the E. coli 103 colonized mice. This experiment shows that K. oxytoca directly inhibits the growth of MDR K. pneumoniae, at least in the beginning of the colonization.

[0098] Probiotic K. oxytoca Protects Susceptible Animals from Luminal and Tissue Invasion and Systemic Dissemination of the Pathogen

[0099] To examine the protective effect further, the characterization of the colonization dynamics was extended. To test if observed differences in the fecal colonization are also visible in the intestinal organs, SPF-H mice were precolonized with K. oxytoca MK01, K. oxytoca MK02 or left uncolonized as a control prior to colonization with K. pneumoniae MDR1. Luminal and tissue invasion of K. pneumoniae MDR1 was assessed in the gastrointestinal organs including small intestine, cecum and colon as well as in the liver, and lymphatic organs including spleen and mesenteric lymph nodes (MLN) on day 6 p.c. A strong reduction in the gastrointestinal organ content as well as in the tissues in both precolonized groups compared to the control groups was observed. In addition, K. oxytoca colonized groups did not have any bacteria in the liver, spleen or MLN in contrast to the control animals, suggesting a systemic spread of the pathogen in control mice, elevating the risk for blood stream infections. This experiment demonstrated a broad-spectrum activity of K. oxytoca in all gastrointestinal organs including reduction of pathogenic bacteria in the lumen and tissue and preventing the systemic spread into the liver and lymphatic tissues such as spleen and MLN.

[0100] The Adaptive Immune-System is Dispensable for the Observed Phenotype Supporting the Hypothesis of a Direct Competitive Effect

[0101] To exclude the hypothesis that adaptive immunity plays a major role in the protective phenotype, the same organ burden experiment using Rag2.sup./ SPF-H animals was performed, which are devoid of any B- and T-cell populations. The data were highly similar with the phenotype initially overserved in WT SPF-H mice. Similarly, CFUs in the gastrointestinal organs including cecum, colon and small intestine were significantly reduced in the lumen and the tissue in K. oxytoca A precolonized animals. In addition, bacterial burden in the liver and spleen was significantly reduced in precolonized mice indicating that K. oxytoca MK01 could prevent the systemic spread of the pathogen without a sufficient adaptive immune response. This supports a direct mechanism of the commensal itself rather than an immune-mediated protection.

[0102] K. oxytoca Helps Anaerobic SCFA Producers to Recover Faster after Antibiotic Dysbiosis Thereby Reestablishing Colonization Resistance.

[0103] In view of the results regarding an immune-mediated mechanism it was analyzed how K. oxytoca colonization may changes the microbiome composition in the mice. To do so, cecal samples from the organ burden (day 6 p.c.) were sequenced. In addition, cecal SCFA levels of these mice were determined. A strong clustering regarding the beta-diversity in precolonized animal on day 6 p.c, in the cecal content was observed. In terms of alpha diversity K. oxytoca treated animals displayed significantly higher species richness and evenness, and a highly diverse microbiome was observed in K. oxytoca precolonized animals in comparison to control animals which showed a dysbiotic species composition mainly consisting of Klebsiella in the phylum of Proteobacteria, and Enterococcaceae in the phylum of Firmicutes. Surprisingly, the microbiome of K.oxytoca treated animals had almost recovered full species diversity as found in untreated SPF-H mice with regard to species present and species richness. Furthermore, significantly elevated levels of butyrate and propionate in SPF-H WT and Rag2.sup./ were found in the mice, further supporting that K. oxytoca helps to reestablish colonization resistance by facilitating the recovery of anaerobic SCFA producers. Based on these findings, SCFA concentrations can also serve as a biomarker to predict disease outcome after infection with certain Enterobacteriaceae, like K. pneumoniae.

[0104] K. oxytoca has the Therapeutic Potential to Cure Already Established MDR K. Pneumoniae Colonization

[0105] As it was shown that K. oxytoca has probiotic potential by efficiently blocking colonization and permanent establishment of MDR K. pneumoniae in the microbiome of antibiotic-treated and gnotobiotic animals, additional tests were performed, whether K. oxytoca could also lower or replace MDR K. pneumoniae after an already established colonization. To do so, mice were treated with ampicillin and infected with K. pneumoniae MDR1. After three days of colonization, mice were subsequently colonized with K. oxytoca MK01 or K. oxytoca MK02, and fecal burden of K. pneumoniae MDR1 was assessed after different time points. Mice treated with K. oxytoca MK01 or K. oxytoca MK02 cleared the colonization with K. pneumoniae MDR1 significantly faster when compared with the control mice (FIG. 5).

[0106] After 42 days both strains K. oxytoca A or K. oxytoca B were reaching similar levels as high as 80% ( 8/10) clearance, whereas only one animal (11.11%) of the control group could clear the colonization (FIG. 5). In addition, to verify if the mice also cleared K. oxytoca from the microbiome or if the bacterium was still detectable, feces were plated on amp plates. In the K. oxytoca MK01 group, 5 out of 10 mice (50%) cleared the bacterium, and 3/10 (30%) cleared K. oxytoca MK02. This observation indicated that in most of the cases K. pneumoniae MDR1 was successfully replaced by K. oxytoca.

[0107] Taken together, this experiment supports the potential of K. oxytoca to act as both a potential probiotic strain and having therapeutic potential to cure MDR carriers who are already colonized with MDR K. pneumoniae strains.

[0108] Protective Effect of Commensal K. oxytoca is Strain-Specific but Observed Against Another MDR K. pneumoniae Strain from the ST258

[0109] To address the question whether the protective effect is limited to a specific strain of commensal K. oxytoca or is shared between multiple commensal isolates or related enterobacteria in general, SPF-H were precolonized with different commensal K. oxytoca strains or a commensal ampicillin-resistant E. coli strain isolated from a healthy volunteer (strain EC103). The K. oxytoca strain panel included next to MK01 additional isolates from protected donors (strain MK02, MKO4, MIK15, MR03), intermediate donors (strain MK06, MK08, MK16, MK17) and the only isolate from a susceptible donor (MR01), as well as the type strain (ATCC13182) (FIG. 6). Importantly, no differences were observed in mice in the colonization levels between K. oxytoca strains and E. coli. Strikingly, mice precolonized with all human K. oxytoca isolates except for MR01 promoted decolonization of K. pneumoniae MDR1 similar compared to MK01 (FIG. 6). In contrast, neither K. oxytoca MR01 nor E. co/i EC103 efficiently promoted K. pneumoniae MDR1 decolonization.

[0110] To investigate if the protective effects of different K. oxytoca strains are also observed against other K. pneumoniae strains, mice were precolonized with either K. oxytoca strains (MK01, MR01, ATCC13182) and E. coli (EC103) and then colonized with K. pneumoniae strain MDR2 (ST258) (FIG. 7). In line with the previous observations MK01 and ATCC efficiently reduced fecal colonization of MDR2 already by day 1 by a factor of 10,000, whereas no decolonization could be observed for MR01 and E. coli. Strikingly, MK01-precolonized mice completely cleared K. pneumoniae MDR2 within 9 days from the feces. Taken together, these experiments revealed that all tested commensal K. oxytoca strains except for MR01 promoted efficient clearance of K. pneumoniae strains from at least two different sequence types.

[0111] K. oxytoca and K. pneumoniae Compete in the Gut Lumen for Specific Carbon Sources with a Broader Substrate Range Found in Protective K. oxytoca Isolates

[0112] The inventors tested whether both bacteria K. oxytoca and K. pneumoniae occupy similar spatial locations in the gut. FISH analysis of the gut lumen demonstrated that both Klebsiella species appeared to be distributed within the lumen, suggesting that they share similar spatial niches and therefore compete for nutrients. To test in vitro, whether both Klebsiella strains coexist or compete for the same substrates in the mouse gut, cecal content of germfree mice was spiked with different K. oxytoca strains or E. coli as control and K. pneumoniae and incubated for 24 h under anaerobic conditions. Cocultivation of all commensal K. oxytoca strains except for MR01 and EC103 significantly reduced CFUs of K. pneumoniae MDR1 in vitro by a factor of 100-1000 fold. Protective K. oxytoca isolates were able to utilize more different carbon sources compared to K. pneumoniae and the non-protective strains (FIG. 8). For instance, MK01 could utilize 100 carbon sources whereas MR01 could only utilize in total 55 different carbon sources. In comparison to K. pneumoniae, which could use 56 different carbon sources, MK01 but not MR01 covered with an overlap of 55 carbon sources almost the complete carbon source utilization pattern of K. pneumoniae except for lactulose.

[0113] Similar patterns were observed for the other protective strains indicating that protective K. oxytoca strains have a broader substrate range covering the substrate range of Kpneumoniae.

[0114] Protective K. oxytoca Strains Outcompete K. pneumoniae for Non-Aromatic Beta-Glycosides Utilization

[0115] Focussing on sugars that could be utilized by K. pneumoniae MDR1 and protective strains, 13 differences between the panel of protective (n=4) and non-protective (n=2) strains in the utilization capacity of various monosaccharides (Tagatose, Psicose, beta-Methyl-D-Glucoside), disaccharides (Cellobiose, Gentiobiose, Palatinose, Sucrose), oligosaccharides (Stachyose), aryl-beta-Glycosides (Arbutin, Salicin) and sugar alcohols (Dulcitol, Inositol, D-Arabitol) were identified (FIG. 8). Interestingly, 6 out of 13 candidates from different groups shared the structural similarity having a beta-glucosidic bond (beta-Methyl-D-Glucoside, Cellobiose, Gentiobiose, Sucrose, Arbutin and Salicin) indicating that beta-glucoside utilization contributes to interspecies competition of Klebsiella strains. Phenotypic screening revealed that all protective K. oxytoca strains (n=22) but not MR01 and EC103 could utilize aryl-beta glucosides (Arbutin and Salicin) and other beta-glucosidic sugars (Cellobiose and Sucrose) as a sole carbon source in minimal medium. In order to test whether these K. oxytoca strains could indeed outcompete K. pneumoniae for beta-Glucosides utilization, both species were co-cultured in a 1:1 ratio in minimal medium supplemented with Sucrose, Cellobiose, Salicin or Arbutin as sole carbon sources for 24 h under anaerobic conditions. Selective plating revealed that growth of K. pneumoniae in Sucrose and Cellobiose supplemented minimal medium was reduced 1,000 to 10,000-fold in co-cultures with all K. oxytoca strains as tested except for MR01 or as a control with EC103 (FIG. 8). Notably, no inhibition was observed in minimal medium supplemented with Salicin and Arbutin. Thus, K. oxytoca effectively outcompetes K. pneumoniae in vitro for non-arylic beta-glucosides like sucrose and cellobiose but not for aromatic beta-glucosides like arbutin and salicin or sugar alcohols like dulcitol or inositol.

[0116] In Order to Maintain Long-Term Full Colonization Resistance in Germfree Mice, K. oxytoca Requires Additional Bacteria

[0117] To characterize whether K. oxytoca directly outcompetes K. pneumoniae in the gut environment, germfree (GF) mice were precolonized with protective K. oxytoca MK01, or as controls with the non-protective MR01 and EC103 and then challenged with K. pneumoniae MDR1 (FIG. 9). At the early timepoint of colonization, K. oxytoca MK01 but not MR01 reduced K. pneumoniae MDR1 colonization compared to GF mice with 1000-fold reduction of K. pneumoniae CFUs at day 3 p.c.. Yet, this reduction slowly decreased but remained stable until 42 days with still significantly reduced CFUs detected in K. oxytoca MK01-precolonized mice in comparison to control mice. Of note, 25% of K. oxytoca MK01-precolonized GF mice cleared K. pneumoniae MDR1 compared to 80-100% in K. oxytoca MK01-precolonized SPF-H mice, suggesting that direct metabolic competition is a major-driver for protection in the initial phase of luminal expansion but cooperating bacteria are required to achieve long-term decolonization.

[0118] K. oxytoca MK01 efficiently reduced K. pneumoniae MDR colonization resulting in gut decolonization already within 9 days. In contrast, E. coli EC103 and MR01 colonized Oligo-MM.sup.12 mice exhibited similarly high colonization levels of K. pneumoniae MDR1 as control mice.

[0119] Colonization of germfree mice with various different minimal consortia of Firmicutes bacteria from the Oligo-MM.sup.12 microbiota revealed that a minimum of three additional bacteria namely Blautia coccoides YL58, Enterococcus faecalis KB1 and [Clostridium]clostridioforme YL32 were able together with K. oxytoca to achieve long-term clearance of K. pneumoniae MDR1 (FIG. 9). This demonstrates that specific members of the phylum Firmicutes cooperate with K. oxytoca to promote decolonization of K. pneumoniae from the gut, with the cooperating bacteria presumably further restricting the availability of alternative carbon sources.