Carbapenem-resistant K. pneumoniae

The clinical CRK isolate used in our study was obtained from Soonchunghyang University Hospital Pathogenic Resource Bank and cultured at the Probiotics Microbiome Commercialization Center [34]. The CRK isolate was grown in MacConkey broth (BD Difco, USA) for 24 h under aerobic conditions at 37 °C. The culture was then streaked onto MacConkey agar (BD Difco, USA) plates supplemented with 10 mg/mL imipenem (Sigma-Aldrich, USA) and incubated for 18 h at 37 °C. Next, colonies were picked, transferred to MacConkey broth, and incubated at 37 °C for 18 h. Subsequently, bacterial growth was checked using a spectrophotometer (DR 1900, Hatch, USA), and the optical density (OD) was measured at 600 nm. Following this, pathogens (OD: 1.0, 2 × 109 CFU/mL) were stored using sterile glycerol at − 80 °C until further use.

Whole genome analysis of microbiome therapeutic PMC101

The PMC101 strain was isolated from a vaginal sample of a healthy woman and subjected to whole-genome sequencing analysis. Genomic DNA from the strains was extracted using the QIAamp DNA Mini kit (Qiagen, Germany) and analyzed by CJ Bioscience (Gyeonggi-do, Korea) based on PacBio sequencing as previously reported [35]. PacBio sequencing data were assembled using the PacBio SMRT Analysis software (v.2.3.0) with the HGAP2 protocol (Pacific Biosciences, USA). The resulting contigs from the PacBio sequencing data were circularized using Circulator 1.4.0 (Sanger Institute, UK). Protein coding sequences (CDSs) were predicted using Prodigal 2.6.2 [36] and classified based on Clusters of Orthologous Groups (COG) (EggNOG; http://eggnogdb.embl.de). For taxonomic purposes, the average nucleotide identity was calculated using the orthologous average nucleotide identity (OrthoANI) method [37].

Establishment of a cell bank, food-grade media optimization, and scale-up production for microbiome therapeutic PMC101

After identifying the PMC101 strain, the Master Cell Bank (MCB) preparation was conducted. The MCB for microbiome therapeutic PMC101 was prepared using a cleaned and decontaminated biosafety cabinet, following ICH Q5D-compliant Good Manufacturing Practice (GMP) guidelines [38]. Subsequently, the bacterial strain was cultured and transferred into cryovials containing 60% glycerol, which had been autoclaved beforehand aseptically following the previous study [39]. Following this, the cryovials were subjected to shaking to mix the bacterial culture and glycerol for 5 min and stored at − 80 °C for future use. Working Cell Banks (WCBs) were subsequently prepared from the MCB under defined culture conditions. Quality control tests were performed to confirm that the MCB and the WCB are genetically identical and to ensure the WCB is free of contaminants. The prepared WCB was also stored at − 80 °C following the earlier conditions.

Moreover, to discover an effective food-grade medium (FGM) suitable for the large-scale production of PMC101, seven different types of FGMs were prepared based on the commercial MRS medium composition, as described in our previous study [40]. FGMs compositions were yeast-peptone (MB cell, Korea), Tween 80 (Sigma-Aldrich, USA), d-glucose (Sigma-Aldrich, USA), magnesium sulfate (Sigma-Aldrich, USA), sodium acetate 3 M (Biosesang, Korea), sodium acetate (Sigma-Aldrich, USA), Dipotassium phosphate (Sigma-Aldrich, USA), and trisodium citrate (MB cell, Korea). After preparation, the pH values of FGMs were adjusted to 6.4 using NaOH 10 M (Sigma Aldrich, USA). Initial pHs of FGM 1, FGM 2, FGM 3, FGM 4, FGM 5, FGM 6, and FGM 7 were 6.4, 6.3, 6.2, 6.5, 6.1, 5.8, and 6.0, respectively. Next, all FGMs were sterilized at 120 °C for 15 min. The PMC101 strain was then inoculated into the prepared FGMs and incubated at 37 °C for 18 h to reach the endpoint of growth. All cultures were initiated with the same OD600 value (0.043 ± 0.01) at the start (0 h). Subsequently, the growth of PMC101 and pH were checked, and the FGM with the highest bacterial growth rate was selected for the forthcoming experiments. Accordingly, PMC101 was cultured in 5 L of fresh, selected FGM using a lab-scale fermenter system (FMT-STS07, Fermentec, South Korea), achieving an OD600 of 1.0 under aerobic conditions. Cells were then collected via centrifugation at 4000 RPM for 10 min at room temperature and adjusted to an approximate concentration of 1 × 109 CFU/mouse/200 μL for the in vivo experiment.

Ethics approval and experimental design for assessing PMC101 efficacy in mice models with CRK infection

Nine-week-old specific-pathogen-free female BALB/c mice were purchased from Doo Yeol Biotech, Korea. The mice were acclimated for 1 week prior to the experiments. They were housed individually in cages within a temperature-controlled environment (23 °C ± 2 °C) under a 12 h light–dark cycle and with a relative humidity range of 30–70%. To disrupt the gut microbiota, mice were administered vancomycin solution (Sigma-Aldrich, USA) at 150 mg/kg (200 µL per mouse) on days − 2 and − 1. Cyclophosphamide (Sigma-Aldrich, USA) at 450 mg/kg (200 µL per mouse) was also administered 3 days prior to infection to induce neutropenia [34]. Three days after neutropenia, mice were infected with 200 µL of CRK suspension (2 × 109 CFU/mouse) orally as a single dose on days 0, 2, and 7. CRK bacterial colonies obtained from the 18 h culture were suspended in a sterile saline solution (0.9% sodium chloride) to achieve the desired concentration of CRK in each mouse model. Additionally, a 200 µL solution of 0.2 M NaHCO3 (Sigma-Aldrich, USA) was administered after infection to increase the severity of infection by neutralizing stomach acidity [34]. Following this, the antibiotic-induced dysbiosis mice were randomly divided into two groups: an infection group and a treatment group, with free access to food and water. The treatment group received PMC101 culture, whereas the infection group only received sterile saline through oral gavage under the same conditions. The prophylactic dosage in mice models was determined according to previous studies [37, 41]. All animal experiments were conducted in a biosafety level 2 room (LML 20-591) at the Probiotic Microbiome Center (PMC) animal facility, Soonchunhyang University, in accordance with the World Health Organization’s guidelines. The facility is registered with the Ministry of Food and Drug Safety (Registration No. 657). The animal experiments of this study were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Soonchunhyang University (IACUC Number SCH23-0003).

Efficacy of microbiome therapeutics PMC101 against CRK infection in antibiotic-induced gut microbiome dysbiosis mouse model

The effectiveness of microbiome therapeutics PMC101 was evaluated against CRK infection in an antibiotic-induced dysbiosis mouse model (n = 6 mice per group). The mice in the probiotic treatment group received 200 µL of PMC101 via oral gavage (1 × 109 CFU/mouse) four times a week on days − 3, − 2, − 1, 0, 2, 6, 7, and 9, respectively. The remaining mice’s survival rate, illness severity score, and body weight were evaluated until the end of the experimental period. Additionally, fecal samples were collected from each cage, resuspended in 1 mL of phosphate-buffered saline (PBS), and homogenized for 60 s. The total aerobic bacterial burden and CRK were then quantified using nutrient agar (consisting of 5.0 g/L peptone, 3.0 g/L beef extract, and 15.0 g/L agar, BD Difco, USA) and MacConkey agar containing imipenem, respectively.

Investigating the efficacy and mechanistic insights of microbiome therapeutic PMC101 in preventing CRK translocation in mice

The prophylactic mechanism of PMC101 in preventing CRK translocation was evaluated using an antibiotic-induced dysbiosis mouse model (n = 12 per group). The mice in the PMC101 treatment group received 200 µL of PMC101 via oral gavage (1 × 109 CFU/mouse) for 6 days starting three days before CRK inoculation. At the end of the experiment, mice from each group were sacrificed, and blood and organ samples, such as the lungs, intestines, and kidneys, were freshly collected. Cytokine analysis was performed using blood samples, and CRK in the lungs and kidneys was quantified by utilizing MacConkey agar containing imipenem. Histopathological and tight junction gene expression analyses were conducted on the intestinal epithelium. Additionally, fecal samples were collected during the experimental period to assess the impact of PMC101 on SCFA production. Moreover, metagenomics analysis was conducted using the collected fecal samples to assess the effect of PMC101 on gut microbiome composition.

Comparative anti-infective activity of Lactobacillus fermentum strains in a murine CRK colonization model

A murine colonization model was established to evaluate the anti-infective efficacy of two different Lactobacillus fermentum strains (L.f 1 and L.f 2) compared to PMC101 in vivo. Mice were randomly divided into infection and treatment groups (n = 6 per group). From day − 3 to day − 1, mice received a daily oral gavage of a broad-spectrum antibiotic cocktail consisting of metronidazole (150 mg/kg, Sigma-Aldrich, USA), kanamycin (150 mg/kg, Sigma-Aldrich, USA), and vancomycin (150 mg/kg) in a total volume of 200 µL per mouse. On day − 3, neutropenia was induced via a single intraperitoneal injection of cyclophosphamide at 450 mg/kg (200 µL per mouse). On day 0, after oral administration of bicarbonate to mice, they were challenged with 200 µL of carbapenem-resistant Klebsiella (CRK) suspension containing 1 × 10⁹ CFU/mouse. Prior to CRK infection, mice were treated with 9 × 109 CFU/mL of L. fermentum strains, including PMC101, in 150 mL of drinking water twice a week for 2 weeks. On day 7 post-infection, 5-fluorouracil was administered intraperitoneally at a dose of 400 mg/kg (200 µL per mouse). Stool samples were collected during the study to quantify the levels of colonization. For CFU determination, stool samples were serially diluted, and 10 µL of each dilution was plated on MacConkey agar supplemented with 10 mg/mL imipenem. Plates were incubated overnight at 37 °C. Bacterial load was calculated and expressed as log10 CFU/mL.

Histopathological assessment

Moreover, to evaluate the efficacy of PMC101 on the intestinal epithelium of the antibiotic-induced dysbiosis mouse model, the intestinal samples from each cage were sliced into three segments corresponding to the duodenum, jejunum, and ileum on day 7. The extracted organs were fixed in neutral buffered formalin, and Hematoxylin and eosin (H&E) staining was conducted at the Korea Pathology Technical Center (KPNT, Cheong-si, Korea). Following this, samples were processed for 13 h using the STP120 Spin Tissue Processor (Myr, Spain), and tissue sections were cut into approximately 4 μm slices before being mounted on slides. The slides were then visualized using an Easyscan Pro6 slide scanner (Motic, Hong Kong). Histopathology was scored on H&E-stained ileum sections, and pathological changes and mucosal inflammation were measured.

Cytokine analysis of blood serum

Blood samples (1 mL) were collected from sacrificed mice’s hearts on day 7 using a syringe (Korea Vaccine, Korea). Subsequently, blood samples were transferred to the lab immediately after collection and kept at room temperature for 20 min to facilitate clotting. After clotting, the blood samples were centrifuged at 4000 RPM for 10 min to separate the blood serum. Then, the blood serum samples were transferred to a fresh Eppendorf tube and stored at − 80 °C for further analysis. Before starting the cytokine analysis, frozen serum samples were thawed, and cytokine levels were evaluated using IL-6, IL-10, TNF-α, and IFN-γ LEGEND MAX Mouse ELISA Kit (BioLegend, USA), according to the manufacturer’s instructions.

Analysis of tight junction protein expression using western blot

Following previous studies, the integrity of the mice’s intestinal epithelium was evaluated during PMC101 treatment on day 7 using a western blot assay [36,37,38]. For protein extraction, the intestinal tissue was collected freshly and homogenized in 1× RIPA lysis buffer (Abcam, UK) and Protease Inhibitor Cocktail (Cell Signaling Technology, USA). The mixture was then centrifuged at 14,000 RPM at 4 °C for 15 min. Extracted proteins were then quantified, transferred into polyvinylidene difluoride membranes, and incubated with primary antibodies, including ZO-1 (Protein Tech, USA) and β-actin (Cell Signaling Technology, USA), followed by incubation with horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology, USA). Membranes were then visualized using a ChemiDoc XRS+ system (Bio-Rad, USA), and densitometric measurements of the obtained bands were subsequently conducted with ImageJ software (version 1.8.0).

Assessment of the impact of microbiome therapeutic PMC101 treatment on SCFA levels

The levels of acetic, propionic, butyric, and valeric acids in the gut content of mice were determined using gas chromatography-tandem mass spectrometry (GC–MS/MS, GCMS-TQ8040, Shimadzu, Japan) according to previous studies [42, 43]. Briefly, the fresh fecal samples were homogenized in 5 mL of PBS-distilled water (1 g/sample) and then centrifuged at 4000 RPM for 5 min. Afterward, 20 mg of gut particulate matter was placed into an Eppendorf tube and then extracted with 30 µL of 0.1 M NaOH and 430 µL of methanol. The mixture was then incubated at − 20 °C for 20 min to facilitate protein precipitation. After centrifugation at 4000 RPM for 10 min at 4 °C, the collected supernatant was dried at 37 °C. Following this, methoxyamine hydrochloride (MeOX, Sigma, USA) was dissolved in pyridine (Samchun, Korea) to a concentration of 20 mg/mL, and the dried mixture was added to this solution prior to the incubation (60 °C for 90 min). Subsequently, 60 µL of N-tert-Butyldimethylsilyl-N-methyltrifluoroacetamide (MTBSTFA; Sigma, USA) was added to the mixture, which was then incubated for an additional 30 min at 60 °C. Finally, the mixture was centrifuged at 4000 RPM for 10 min at 4 °C, and the supernatant was collected and mixed with 140 µL of pyridine to prepare the final test solution for detection by GC–MS. The quantification of SCFA was assessed by measuring the peak areas for propionate, acetic, butyric, and valeric acids. The mixture of acetic acid (TCI, Japan), propionic acid (TCI, Japan), butyric acid (TCI, Japan), and valeric acid (Sigma, USA) was prepared at a concentration of 1000 ppm and used under the conditions mentioned above.

Evaluating the effect of microbiome therapeutic PMC101 on the human intestinal microbial ecosystem

To investigate the effect of microbiome therapeutic PMC101 on intestinal eubiosis in greater depth, a human intestinal microbial ecosystem simulator (SHIME, ProDigest, Belgium) was used, inoculated according to the manufacturer’s instructions and a previous study [44]. The system consisted of 10 double-jacked vessels that mimic the in vivo gut microbial ecosystem (stomach, small intestine, ascending, transverse, and descending colon) by controlling environmental conditions such as pH, temperature, inoculum, and residence time [30]. In brief, fresh fecal samples of four healthy children volunteers without a history of antibiotic treatment for 6 months before this study were collected and homogenized with 100 mL of sterilized phosphate buffer (0.1 mol/L, pH 7.0) (8.8 g/L K2HPO4, 6.8 g/L KH2PO4, 0.1 g/L sodium thioglycolate, and 15 mg/L sodium thionite) (ProDigest, Ghent, Belgium) for 10 min using a Bag Mixer 400 (Interscience, Louvain-La-Neuve, Belgium). Next, the samples were inoculated into the colon vessel, both control and treatment groups, where the microbial inoculum was stabilized for 2 weeks in a carbohydrate-based medium and allowed to adapt to the specific environmental conditions of the colon vessels. Subsequently, 50 mL of PMC101 culture (1 × 109 CFU/mL) was added to the small intestine vessel on day 15 in one set. Probiotics were not used in another set as a control. Subsequently, samples were collected from the three colon vessels (ascending, transverse, and descending colon vessels) on days 0, 1, 5, and 7. The collected samples were then utilized to quantify the normal flora, beneficial bacteria, and CRK using nutrient agar, MRS agar, and MacConkey agar, respectively. Additionally, the samples were analyzed through metagenomics to assess the effect of PMC101 treatment on gut microbiota composition. The Pediatrics Department of Soonchunhyang University Cheonan Hospital collected fresh human fecal samples. The experimental method was reviewed and approved by the Ethics Committee of the Pediatrics Department at Soonchunhyang University Cheonan Hospital IRB (No. SCHCA 2019-12-016), and informed consent was obtained from all subjects and their parents.

Metagenomics analysis of the gut microbiome

To explore the impact of PMC101 on gut microbiota composition, metagenomics analysis was conducted using the stool samples collected from the second animal model and human intestinal microbial ecosystem simulator samples. Briefly, genomic DNAs were extracted from these samples using a QIAamp DNA Fast Stool Mini Kit (Qiagen, Germany) following the manufacturer’s protocol. DNA was then quantified with a Qubit Fluorometer (Thermo Fisher Scientific, UK) using a dsDNA HS Assay Kit (Invitrogen, USA). The quality of extracted DNA was also analyzed using a ChemiDoc (Bio-Rad, USA) after subjecting it to 0.8% agarose gel electrophoresis. Following this, the amplicon PCR was performed to amplify the V4 hypervariable region of the 16S rRNA gene using standard Illumina 16S primers (forward primer: TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG CCT ACG GGN-GGC WGC AG, reverse primer: GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GG-ACT ACH VGG GTA TCT AAT CC) following the optimized conditions [45]. Agencourt AMPure XP (Beckman Coulter, UK) beads were used after each PCR step to purify the PCR products, which had been quantified using a Qubit dsDNA HS Assay Kit. Next, metagenomic libraries were prepared using a Nextera XT DNA Library Prep Kit (Illumina, USA). Later, the samples were normalized, pooled, and mixed with PhiX Control v3 (Illumina, USA) before being sequenced on the iSeq™ 100 platform (Illumina, USA), following the standard Illumina protocol previously used by our team [13, 45]. Finally, the sequencing data were analyzed, and the entire microbial compositions of the samples were exhibited according to a previously described protocol [13]. The Fast Length Adjustment of Short Reads (FLASH) software (version 1.2.11) was used to merge pairs of reads from the initial DNA sections [46]. The Quantitative Insights into Microbial Ecology (QIIME) software (version 1.9.1) was employed for sequence analysis [47]. The richness and diversity of the bacterial populations within the samples were assessed using α indices, i.e., ACE and Shannon [48, 49].

Monitoring the antimicrobial properties of microbiome therapeutic PMC101

A time-kill assay was conducted against the CRK pathogen using the cell-free culture supernatant (CFCS) of PMC101, as described in a previous study with some modifications [50]. To serve this purpose, the PMC101 strain was cultured in MRS broth (BD Difco, USA) and FGM broth overnight, and then CFCS was prepared. In this assay, the initial amount of CRK was set at 1 × 108 CFU/mL, and the pH was maintained at 6.0 ± 0.01. Following this, 300 μL of CRK suspension was added to 15 mL of CFCS derived from MRS broth and FGM, respectively, and incubated at 37 °C. Control groups included CRK pathogen suspended in MRS broth and FGM without CFCS inoculum. At predefined intervals, aliquots were withdrawn from each tube, serially diluted, and plated onto MacConkey agar to determine the viable CFU of CRK. The assay was executed in triplicate, and the mean log CFU/mL of total surviving cells was recorded. Additionally, the pH level of each aliquot’s interval was measured during the experiment using an electronic pH meter (Mettler Toledo, Korea).

Safety evaluation of microbiome therapeutic PMC101

To assess the safety properties of PMC101, hemolytic activity and antimicrobial susceptibility (E-test) were evaluated using a previously established protocol based on data from the official websites of the Korean Ministry of Food and Drug Safety (MFDS) [51,52,53]. Briefly, the hemolytic activity of PMC101 was evaluated on blood agar plates (KisanBio, South Korea) under anaerobic conditions at 37 °C for 48 h. Lacticaseibacillus rhamnosus KCTC 5033 (Korean Collection for Type Culture, Daejeon, South Korea) and S. aureus (NCCP 14780) (National Culture Collection for Pathogens, Cheongju, South Korea) served as negative and positive controls [54]. Antimicrobial susceptibility was assessed using the E-test, following the guidelines of the Clinical and Laboratory Standards Institute (CLSI), and minimum inhibitory concentration (MIC) values were compared to the European Food Safety Authority (EFSA) standards [55]. All experiments adhered to EFSA safety assessment guidelines.