Exopolysaccharide compositions and use thereof

Abstract

A composition for the in vivo stimulation of cancer immunity, said composition including at least one exopolysaccharide (EPS) characterised in that the at least one EPS is a trimeric sugar comprised predominantly of 1-3 and 1-4 linked glucose and galactose residues.

Description

[0001] Exopolysaccharide Compositions and use thereof

[0002] The present invention relates to compositions containing exopolysaccharides and method of use of said polysaccharides in anti-cancer or anti-tumour applications.

[0003] Although tthhee following description rreeffeerrss exclusively ttoo capsular exopolysaccharides derived from Bifidobacterium, the skilled person will appreciate that the compositions of the present invention can be isolated or derived from other bacteria and is not limited to capsular polysaccharides or those derived from Bifidobacterium.

[0004] Inconsistent patient response to standard of care therapy is one of the great challenges of cancer medicine. It is now thought that host-intrinsic factors underpin resistance mechanisms mediating inconsistent therapeutic response(7). Advancement of multi-omic technologies has heralded the gut microbiome as one such factor regulating cancer progression (2). Seminal studies have demonstrated this phenomenon in both mice(3-5) and humans(6-8), implicating specific species and strains which can be protective against cancer progression and enhance responses to standard of care therapy(3, 9, 10). Although the concept of anti- tumour activity from intestinal bacteria is established, most studies to date focus on immunogenic tumours (e.g., melanoma, lung cancer) (11-14) and gastrointestinal indications (such as colorectal cancer) (9, 15). The influence of gut bacteria on extra-intestinal non-immunogenic indications, like breast cancer, is not well characterised.

[0005] Many species of lactic acid bacteria, such as Bifidobacterium and Lactobacillus, are known to inhibit the progression of solid tumour types in pre-clinical models (9, 16, 17). Several species of Bifidobacterium, including bifidum(15), breve(18, 19), and pseudolongum(10) have been studied and associated with anti-tumour immunity. This pre-clinical data is increasingly supported by human studies, which show that abundances of various Bifidobacterium species positively correlate with patient outcomes across multiple cancer indications (77, 14). Collectively, there is good evidence to support Bifidobacterium-based therapeutic interventions to treat cancer.

[0006] Historically, microbiome-based cancer therapeutics have lacked strong mechanistic data on the microbial active compounds which initiate a beneficial host response. Microbial metabolites are the most described mechanisms (9, 10, 20, 21), although microbial structural compounds are now increasingly understood to drive protective responses (15, 22). A class of compound receiving interest in this context are capsular and secreted exopolysaccharides (EPS). These polymeric sugar chains can vary in size, glycosyl contents, and structural decoration, and are known drive discrete host immune response pathways (23, 24). Although the role of structural polysaccharides are better understood in the context of fungal pathogenesis(23), the cancer protective roles of microbial polysaccharides are now being releveled. A seminal study by et al, (3) demonstrated Bacteroides fragilis polysaccharide A

[0007] (PSA) to mediate anti-tumour immunity response combined with immune checkpoint inhibitors. More recently, Sharma et aL,(22) showed a strain of Lactobacillus plantarum produces surface polysaccharide which mediates anti-tumour immunity through enhanced activity of macrophage iron sequestration pathways, skewing tumour-associated macrophages to anti-tumour CD8+-permissive state.

[0008] Despite these breakthrough studies involved in direct tumour cell death, a role for EPS in stimulating cancer immunity has not previously been described.

[0009] It is therefore an aim of the present invention to provide a composition that addresses the abovementioned problems. It is a further aim of the present invention to provide a composition for the treatment of cancer and / or immunomodulation.

[0010] It is a yet further aim of the invention to provide a composition for in vivo stimulation of cancer immunity.

[0011] In a first aspect of the invention there is provided a composition for the induction of anti-tumour activity and / or the treatment of cancer, said composition including at least one exopolysaccharide (EPS) characterised in that the at least one EPS is a trimeric sugar comprised predominantly of 1-3 and 1-4 linked glucose and galactose residues.

[0012] As such, the EPS provides in vivo stimulation of cancer immunity.

[0013] In a preferred embodiment the EPS is included with a pharmaceutical carrier to provide an anti-tumour therapeutic compound or composition.

[0014] In a preferred embodiment of the invention the EPS is isolated or cleaved from the genus Bifidobacterium.

[0015] Preferably the EPS trimeric sugars have a molecular weight of at or below 500 da.

[0016] Preferably the EPS trimeric sugars have a molecular weight at or below 475 da.

[0017] In one embodiment the EPS is a capsular EPS. Typically the capsular EPS is cleaved and / or isolated from the genus Bifidobacterium. Typically the EPS is derived or isolated from B. pseudocatenulatum. Further typically the EPS is isolated or derived from Bifidobacterium pseudocatenulatum LH663 or Bifidobacterium pseudocatenulatum 210 and / or B. pseudocatenulatum DSM20438.

[0018] Bifidobacterium pseudocatenulatum LH663 and Bifidobacterium pseudocatenulatum 210 refer to the same strain of Bifidobacterium pseudocatenulatum.

[0019] Typically the EPS is derived or isolated from the bacteria of the genus Bifidobacterium can include any one or any combination of Bifidobacterium adolescentis, Bifidobacterium angulatum, Bifidobacterium ammalis, Bifidobacterium bifidum, Bifidobacterium breve, Bifidobacterium catenulatum, Bifidobacterium dentium, Bifidobacterium gallicum, Bifidobacterium infantis, Bifidobacterium kashirvanohense, Bifidobacterium longum, Bifidobacterium pseudocatenulatum, Bifidobacterium pseudoIongum, Bifidobacterium reteuri, Bifidobacterium scardovii.

[0020] Typically the composition induces CD8+T cell-dependent anti-tumour immunity through the activation dendritic cells.

[0021] In one embodiment the binding of EPS increases CD80+CD86+dendritic cell (DC) maturation increases the systemic infiltration of CD8+T cell specific cDC1 cells. Typically EPS-programmed DCs enhances the polarisation and activation of CD8+T cells at systemic sites, further typically increasing the fraction of IFNγ+TNFa+poly-functional CD8+T cells at the tumour.

[0022] In one embodiment the composition promotes dendric cell activation. Typically the composition increases systemic cDC1 infiltration. Further typically leading to robust CD8+T cell-mediated anti-tumour activity. Typically the composition is active and / or provides anti-tumour activity in respect of breast cancer.

[0023] In a second aspect of the invention there is provided a pharmaceutical composition for the treatment of cancer, said composition presenting anti tumour activity and including at least one EPS.

[0024] In one embodiment the EPS comprises substantially 1-3 and 1-4 linked glucose and galactose residues.

[0025] Typically the composition is active to prevent breast cancer and / or treat breast cancer tumours

[0026] In a third aspect of the invention there is provided an immunotherapy agent said agent including a microbial polysaccharide. Typically the polysaccharide is a capsular EPS. Further typically the polysaccharide is a Bifidobacterium EPS.

[0027] In a further aspect of the invention there is provided an anti-tumour therapeutic compound said compound derived and or isolated from a cell surface capsular EPS.

[0028] In a yet further aspect of the invention there is provided a method of preparing a pharmaceutical composition, said composition including at least one EPS derived or isolated from the genus Bifidobacterium.

[0029] In one embodiment the composition is a chemotherapy adjuvant.

[0030] In one embodiment the pharmaceutical composition includes and / or is administered with at least one immune checkpoint inhibitor. In one embodiment the inhibitor is a PD1 inhibitor.

[0031] In one embodiment, the nucleotide sequence of the Bifidobacterium or Bifidobacterium pseudocatenulatum LH663 / Bifidobacterium pseudocatenulatum 210 may have 80, 85, 90, 95 or 99 % identity to the sequence set out in SEQ ID NO: 1 to 5.

[0032] In one embodiment, the nucleotide sequence of the Bifidobacterium or Bifidobacterium pseudocatenulatum LH663 / Bifidobacterium pseudocatenulatum 210 may have 80-99 % identity to any one or any combination of the sequences set out in SEQ ID NOs: 1 to 5.

[0033] In one embodiment, the nucleotide sequence of B. pseudocatenulatum DSM20438 may have 80, 85, 90, 95 or 99 % identity to the sequence set out in SEQ ID NO: 1 to 5.

[0034] Preferably the EPS is obtained from Bifidobacterium pseudocatenulatum 210 and / or B. pseudocatenulatum DSM20438.

[0035] Preferably the EPS is obtained from Bifidobacterium pseudocatenulatum strain with SO- 99 % identity to any one or any combination of the sequences set out in SEQ ID NOs: 1 to 5.

[0036] Specific embodiments of the invention are now described with reference to the following figures wherein

[0037] The present invention concerns the use of the capsular exopolysaccharides (cEPS) isolated from Bifidobacterium for the modulation of the immune system and treatment of cancer. The cEPS is a mixture of low molecular weight (475da) trimeric sugars, comprised predominantly of 1-3 and 1-4 linked glucose and galactose residues. We believe the cEPS we have generated has been cleaved from a larger structure by our isolation protocol, converting a natural product to a smaller form not found in nature.

[0038] In this study, we demonstrate that several species of Bifidobacterium (B. bifidum, B. reteuri, and B. pseudocatenulatun) have therapeutic utility for the treatment of pre- clinical breast cancer. We highlight that administration of a single strain of B. pseudocatenulatum (LH663) / Bifidobacterium pseudocatenulatum 210 can reduce primary tumour burden across models of the major breast cancer subtypes and enhance response to standard of care chemotherapy and immunotherapy. In a novel mechanism, we demonstrate B. pseudocatenulatum 210 capsular EPS mediates systemic anti-tumour immunity from the gut. B. pseudocatenulatum 210 EPS induces CD8+T cell-dependent anti-tumour immunity through the activation dendritic cells. Binding of 210 EPS increases CD80+CD86+DC maturation in vitro and in vivo, preferentially increasing the systemic infiltration of CD8+T cell specific cDC1 cells. 210 EPS-programmed DCs enhance the polarisation and activation of CD8+T cells at systemic sites and, crucially, increase the fraction of IFNγ+TNFa+poly- functional CD8+T cells at primary tumour. We demonstrate that 210-derived EPS can be isolated and used therapeutically independently of any live parental B. pseudocatenulatum cells, highlighting Bifidobacterium EPS aass aa novel cancer immunotherapy compound.

[0039] Methods

[0040] Mice

[0041] C57 BL / 6 mice were purchased and maintained in-house at the Disease Modelling Unit at the University of East Anglia. BALB / C mice were purchased from Charles River and maintained in-house. Animals used throughout were age 8-12 weeks and were randomly mixed between cages prior to experiment onset. All animal experiments were performed in accordance with UK Home Office regulations and the European Legal Framework for the Protection of Animals used for Scientific Purposes (European Directive 86 / 609 / EEC).

[0042] Orthotopic breast tumour growth experiments

[0043] Synergic breast cancer cells were injected in 50μl of a 1:1 mixture of phosphate buffered saline (PBS) and Matrigel (Coming Life Sciences, Coming, USA) into the left inguinal mammary fat pad of age-matched female mice. PyMT-BO1, E0771 and 4T1 cells were each injected 1x105cells and BRPKp110 cells were injected at 5x105cells. The experimental duration for each breast model used is outlined in the associated figures. In situ tumour volumes were measured thrice weekly with digital calipers from the onset of a palpable tumour, using the formula: length x width2x 0.52(26).

[0044] In vivo experimental interventions (orthotopic experiments)

[0045] Animals were orally administered thrice weekly with live Bifidobacterium strains (1x1010CFU / 200μl) or isolated EPS (80μg / 200μl) from the onset of a palpable tumour to endpoint. For chemotherapy experiments, cyclophosphamide (Sigma) was administered by intraperitoneal injection at 100mg / kg on days 10 and 17 (for BRPKp110 experiment) or days 7 and 14 (PyMT-BO1 experiment). For checkpoint immunotherapy experiments, tumour-bearing animals were intraperitoneally administered 20mg / kg anti-PD-1 mAb (Clone J43, BioXCell) or matched isotype control on day 7, 10, 13 (4T1 experiment) or day 10, 13, 16, 19 (BRPKp110 experiment). Cellular depletions were induced through intraperitoneal injection of 400μg anti-CD8-a (clone 2.43, BioXCell) or matched isotype control mAb one day prior to Bifidobacterium administration (day 9), followed by 200μg injections thereafter on days 13, 16, and 19. Depletion was verified by flow cytometry of primary tumour immune cells. For microbiome depletion, BRPKp110 tumour-bearing animals were treated with antibiotics by oral gavage (200 μl in water vehicle) on day 3, 6 and 8 prior to therapeutic intervention. The antibiotic cocktail contained Img / ml Amphotericin B, 25mg / ml Vancomycin, 50mg / ml Neomycin and 50mg / ml Metronidazole (all purchased from Sigma). Animal drinking water was also supplemented with Img / ml Ampicillin (Sigma). For adoptive cell transfer experiments, 1x106bone marrow-derived dendritic cells (BMDCs) were cultured with either 5x106CFU live 210 cells, 80μg of EPS derived from B. pseudocatenulatum 210 or B. Iongum B71, or vehicle control. Treated BMDCs were adoptively transferred by intravenous injection to naive BRPKp110 tumour- bearing animals on days 10, 14, and 18.

[0046] PyMT spontaneous tumour growth experiments

[0047] MMTV-PyMT (B6.FVB-Tg(MMTV-PyVT)634Mul / LellJ) mice were obtained from Jackson Laboratory (stock 022974) on a congenic C57BL / 6 background. Female mice heterozygous for the PyMT transgene were use for tumour growth studies. Mice were palpated for tumours and treated twice weekly with oral administrations of B. pseudocatenulatum 210 from 8 weeks of age, with the first palpable tumours being observed from 12 weeks. Tumour growth was measured using digital calipers and animals were sacrificed once tumours reached 1cm3or at the 24-week experimental endpoint.

[0048] Germ free colonisation experiments

[0049] Germ free mice were and maintained in isolators orally administered 1x1010CFU of B. pseudocatenulatum 210 or vehicle control (PBS). Animals were sacrificed at 6, 24, and 48 hours post administrations and gut contents were isolated from the upper and lower colon, caecum, and small intestine in sterile MSC class II cabinets. Gut contents were isolated at 50mg per sample, per animal, and mechanically homogenised in 1ml of PBS. 100μL of the slurry solution was then cultured on MRS-cysteine agar plates and incubated anaerobically (37°C for 48 hours. Visible colonies were counted, and bacterial load (CFU / g) was calculated for each condition with the following: CFU / g = (no. of colonies x dilution factor) / (volume cultured x weight of faecal sample). Gut contents from PBS treated germ free animals were confirmed to be sterile at each timepoint by the same method.

[0050] Bacterial culture

[0051] All Bifidobacterium strains used for this study were isolated previously by the laboratory of Prof. Lindsay Hall. The strains were cultured at 37°C in MRS broth with L-cysteine (50mg / L) (Sigma) in an anaerobic chamber (Don Whitley Scientific, Bingley, UK). Strains were cultured for one week and then preserved by lyophilisation in the exponential phase of growth. Lyophilised bacteria were equally distributed across individual dosage vials which were stored at -80°C. To ensure accurate dosing, at least three vials from each lyophilisation batch were enumerated by counting CPUs on MRS agar plates across serial PBS dilutions. The mean CPU was calculated for each batch and vials were resuspended in PBS (to 1x1010CFU / 200μl / animal) immediately prior to experimental administrations.

[0052] For acid-killed experiments, peracetic acid pre-treatment of bacteria was performed as previously described(27). Briefly, lyophilised bacteria were reconstituted in 10ml of sterile PBS at a concentration of ~ 1x1010CFU / ml. Peracetic acid (Sigma) was added to a final concentration of 0.4% and bacteria were incubated at RT for 1 hour. The bacteria were washed three times in sterile PBS and then resuspended to the appropriate final concentration for animal administrations. Bacterial killing was confirmed by culturing 100μL of acid-killed bacteria on MRS agar under aerobic conditions (outlined in section 2.3.1.) and validating the absence of bacterial growth compared with a non-acid treated positive control.

[0053] Cell culture All tumour cell lines were cultured in high glucose DMEM (Thermofisher) supplemented with 10% foetal bovine serum (FBS) (Hyclone, Thermo fisher) and 100 units / ml penicillin / streptomycin (Thermofisher). Cells were seeded onto flasks coated with 0.1% porcine gelatin (Sigma) and incubated at 37°C and 5% C02.

[0054] BMDCs were generated from flushed bone marrow from tibias and femurs of C57 BL / 6 mice. Isolated bone marrow was treated with red blood cell lysis buffer for 5 minutes, washed twice in PBS, and cultured overnight in RP MI 1640 medium supplemented with 10% FBS, 1% penicillin / streptomycin (Gibco), 20ng / ml IL-4 (R&D) and lOng / ml GM-CSF (R&D). On the second day, supernatant with non- adherent cells was removed and culture media replaced. Cell culture medium was replaced again on the fifth day, and semi-adherent BMDCs were collected on the seventh day for experimental use.

[0055] Lung histology

[0056] Harvested organs were incubated overnight in 4% paraformaldehyde (PFA) at 4°C and processed with the Leica Tissue Processor ASP-300-S (Leica Biosystems, Milton Kenes, UK). The tissues were incubated in formalin, dehydrated through increasing concentrations of ethanol (from 70% to 100%), washed in three changes of xylene (Sigma-Aldrich) and then embedded in paraffin (Sigma- Aldrich). Paraffin blocks were sectioned at 6μm using a rotary microtome (Leica Biosystems, RM2235), mounted onto positively charged glass slides (Thermofisher), and incubated overnight at 37°C. Prior to histological staining, FFPE tissue sections were deparaffinised in xylene (Sigma) and rehydrated through sequentially decreasing concentrations of ethanol (100%-70%), then into water. H&E staining was performed using a Leica ST5020 tissue multi-stainer (Leica Biosystems, Nussloch, Germany) and sections were then mounted with coverslips with Neo- Mount™ (Sigma). Images were captured using the Olympus BX60 microscope (Olympus, Southend-on-Sea, UK) with a microscope camera Jenoptik C10 and ProgRes CapturePro software v.2.10.

[0057] Flow cytometry

[0058] Tumours and lungs were excised and mechanically homogenised using scalpels. Tumours were digested in 0.2% collagenase IV (Thermofisher) and lungs in 0.2% collagenase I (Thermofisher), with 0.01% hyaluronidase (Sigma) and 0.01% DNase I (in HBBS) at 37°C under agitation for 60 minutes (tumours) or 30 minutes (lungs). Tissues were then passed through a 70μm filter (Thermofisher) and washed in PBS before further staining. Spleens and lymph nodes were mechanically dissociated through 70μm strainers. Lymph nodes were immediately processed for staining and tumours, lungs, and spleens were resuspended red blood cell lysis buffer (Thermofisher) for 5 minutes. Blood was collected by cardiac puncture in EDTA and washed twice in red blood cell lysis prior to cell staining. Cell from all tissues were washed in PBS and 1x106cells per sample were then resuspended in FACS buffer (2% FBS in PBS) prior to further staining with extracellular antibodies.

[0059] For intracellular cytokine analysis, cells were resuspended in 200μl RPMI supplemented with 10% FBS, 50μM 2-Mercaptoethanol, 50ng / ml Phorbol 12- Myristate 13-Acetate (PMA), 750ng / ml lonomycin and 10μg / ml Brefeldin-A (all purchased from Sigma) and incubated in a 96-well U-bottom plate (Sigma) at 37°C, 5% CO2for 4 hours. Cells were blocked with Fc-receptor blocking reagent (Thermofisher) and incubated in relevant extracellular antibody (Table X) and fixable Live / Dead Red (Invitrogen) solutions for 30 minutes (at 4°C in the dark). Cells were washed twice in FACS buffer, fixed with 4% PF A for 30 minutes, and then resuspended in FACS buffer prior to analysis. Where intracellular staining was required, cells were treated using the eBioscience™ Foxp3 / Transcription Factor Staining Buffer Set, as per the manufacturer’s instructions, then stained with intracellular antibodies.

[0060] Data collection was conducted on the BD LSR Fortessa cell analyser and analysed using Flowjo software (BD). All samples were initially gated using FSC-A vs. FSC- H to identify single cells (Singlets), which were then gated for Live / Dead Red negative. Major immune populations were identified using the marker profiles outlined in Table X.

[0061] Serum isolation

[0062] Blood was collected by cardiac puncture immediately following animal sacrifice by rising level of CO2. Blood was allowed to coagulate for 30 minutes and then centrifuged at 12000 g for 15 minutes at 4°C. Serum was removed from the separated pellet of coagulated blood and stored at -80°C.

[0063] Untargeted metabolomics

[0064] The untargeted metabolomics assay was performed by Biocrates (Innsbruck, Austria). The commercially available MxP® Quant 500 kit from Biocrates was used for the quantification of endogenous metabolites of various biochemical classes. Lipids and hexoses were measured by flow injection analysis-tandem mass spectrometry (FIA-MS / MS) using a SCIEX API 5500 QTRAP® (AB SCIEX, Darmstadt, Germany) instrument with an electrospray ionization (ESI) source, and small molecules were measured by liquid chromatography-tandem mass spectrometry (LC MS / MS), also using a SCIEX API 5500 QTRAP@ (AB SCIEX, Darmstadt, Germany) instrument. The experimental metabolomics measurement technique is described in detail by patents EP1897014B1 and EP1875401B1. Briefly, a 96-well based sample preparation device was used to quantitatively analyse the metabolite profile in the samples. This device consists of inserts that have been impregnated with internal standards, and a predefined sample amount was added to the inserts. Next, a phenyl isothiocyanate (PITC) solution was added to derivatise some of the analytes (e.g., amino acids), and after the derivatization was completed, the target analytes were extracted with an organic solvent, followed by a dilution step. The obtained extracts were then analysed by FIA-MS / MS and CC-MS / MS methods using multiple reaction monitoring (MRM) to detect the analytes. Concentrations were calculated using appropriate mass spectrometry software (Sciex Analyst®) and data were imported into biocrates' MetIDQTM software for further analysis.

[0065] Metabolomics analysis to generate PCoA plots, heatmaps and differential metabolite comparisons were conducted using the MetaboAnalyst (V5.0) platform. The data was first normalised by sum, log10 transformed, and auto scaled prior to univariate analyses, as described previously (28). Differential metabolite levels were compared by two-sample t test with statistical FDR threshold value < 0.05. Metabolite abundances were then input into the MetaboAnalayst platform to allow comparisons of perturbed metabolic pathways, as previously described(29).

[0066] Mesoscale discovery (MSD) cytokine analsis

[0067] Tissue samples were weighed into a MPBio Lysing Matrix E bead beating tube (MPBio) with 1ml of homogenisation buffer (150 mmol / L NaCl, 20 mmol / L Tris, 1 mmol / L EDTA, 1 mmol / L EGTA, 1% Triton X-100, pH 7.5+ cOmpleteTM protease inhibitor (Roche)). An MPBio Fast Prep bead beater (MPBio) was used to homogenise the tissues at a speed 4.0 for 40 seconds, followed by speed 6.0 for 40 seconds. Samples were centrifuged at 12,000 x g for 12 minutes (4°C) and then stored at — 80°C until analysed. Samples were run on a custom Mesoscale Discovery U-PLEX Mouse Kit (MSD) according to the manufacturer’s instructions. Plate was read using an MSD QuickPlex SQ 120 imager (MSD, Rockville, MD, USA). Shotgun metagenomics

[0068] Caecal DNA was extracted using the MPBio FastDNA™ SPIN Kit for Soil (MPBio) following the manufacturers protocol, with an extension of the bead beating time to three minutes. The recovered DNA was validated using a Qubit® 2.0 fluorometer (Invitrogen) prior to library preparation.

[0069] Sequencing was performed as previously described by McKee (2021) (30) and Alikhan (2022) (37). A modified Illumina Nextera low input tagmentation approach was used. A master mixture containing 9μL of TD Tagment DNA Buffer, 0.09μL TDE1, Tagment DNA Enzyme and 4.01 μL PCR grade water was loaded at 3 μL / reaction to a chilled 96 well plate. Genomic DNA was normalised to 0.5ng / μL with lOmM Tris-HCl. 2μL of normalised DNA (Ing total) was pipette mixed with the 5μL of the tagmentation mix and heated to 55 °C for 10 minutes in a PCR block. A PCR master mix (4ul kapa2G buffer, 0.4 μL dNTP’s, 0.08 μL Polymerase and 4.52 μL PCR grade water) from the Kap2G Robust PCR kit (Merck Life Science) was added at 9pL per samples. 2μL of each P7 and P5 of Nextera XT Index Kit v2 index primers (Illumina) were added to each well. Then, the 7μL of Tagmentation mix was added and mixed. The PCR was run with 72°C for 3 min, 95°C for 1 min, 14 cycles of 95°C for 10s, 55°C for 20s and 72°C for 3 minutes. Following the PCR reaction, the libraries were quantified using the Quant-iT dsDNA Assay Kit, high sensitivity kit and run on a FLUOstar Optima plate reader. Libraries were pooled in equal quantities. The final pool was double- SPRI size selected between 0.5 and 0.7X bead volumes using KAPA Pure Beads (Roche, Burgess Hill, UK). The final pool was quantified on a Qubit 3.0 instrument and rruunn oonn a D5000 ScreenTape (Agilent) using the Agilent Tapestation 4200 to calculate the final library pool molarity. The pool was run at a final concentration of 1.5pM on an Illumina Nextseq500 instrument using a Mid Output Flowcell (NSQ® 500 Mid Output KT v2(300 CYS) following the Illumina recommended denaturation and loading recommendations which included a 1% PhiX spike in (PhiX Control v3 Illumina Catalog FC-110-3001). Data was uploaded to Basespace (www.basespace.illumina.com) where the raw data was converted to FASTQ files for each sample.

[0070] Primary analysis was conducted using the MicrobiomeAnalyst (V2.0) platform(32). Default filtering settings were used to remove low count and low variance and ensure robust results. Total sum scaling was applied and downstream analysis at the phyla and species level was conducted using the platform in-built tools. Analysis included microbiome a diversity (Shannon measure with t-test), 0 diversity plotted as PCoA (Bray— Curtis distance method), and univariate species abundance comparisons using t test (adjusted cut-off < 0.05).

[0071] PyMT tumour microbiome analysis

[0072] Mouse mammary tumours from the indicated anatomical regions were dissected in sterile conditions in tissue culture hoods with autoclaved tools, reagents, and protective equipment. Tumour samples were manually homogenised in 1ml of PBS, with 100μL of the resulting homogenate spread on microbial culture plates. For aerobic culture, homogenate was spread on Columbia blood agar (CBA) (Oxoid) + 5% horse blood; Man Rogosa Sharpe (MRS) agar (Oxoid); brain heart infusion (BHI) agar (Oxoid), and incubated aerobically at 37°C for five days. For anaerobic culture, homogenate was spread on MRS agar (+ 50mg / L-cysteine), BHI agar (+ 50mg / L-cysteine), Peptone Yeast Extract Glucose Starch (PYGS) agar (Thermo) and cultured anaerobically at 37°C for three days. Dissection tools were dipped in sterile PBS and plated on under the same conditions as an environmental control, whilst mouse skin swabs were plated as a positive control to confirm culture conditions could support microbial growth.

[0073] Bifidobacterium pseudocatenulatum qPCR Bacterial load was of B. pseudocatenulatum 210 was estimated by qPCR using a species-specific primer (GroEL gene) designed previously by Junick and Blaut (2012)(33). The reactions were performed in duplicate in 12.5μl of LightCycler® 480 SYBR Green I Master (Roche Diagnostics, cat. 0470751600), 2.5 μl of forward and reverse lOpM primers, 7.3μl of RNase-free water (Qiagen), and 0.2μl of template DNA. Standard curves were generated by serial dilutions of 100 ng / μL of monoculture- extracted DNA to reach the lowest concentration of 0.001 ng / μL. Samples were run on the LightCycler® 480 system (Roche) with the programme as follows: 5 minutes incubation at 95 °C, followed by 45 cycles with 15 seconds at 94° C, 15 seconds at 64° C and 15 seconds at 72°C. The melting curve analysis followed with 5 seconds at 95°C, 5 minutes at 65°C and continuous temperature increase to 97°C. Samples were finally cooled to 40°C for 30 seconds before completion. Data was analysed with the LightCycler® 480 Software (v.1.5) (Roche).

[0074] Exopolysaccharide isolation and purification

[0075] EPS isolation and purification were performed based on the protocol by Ruas- Madiedo (2021) (34). A 20ml bacterial suspension in MRS broth was cultured overnight at 37°C under anaerobic conditions. 200μl of the liquid culture were plated on MRS + L-cysteine (50mg / L), MRS + L-cysteine (50mg / L) + fructose, or MRS + L-cysteine (50mg / L) + arabinose agar plates and incubated for 96 hours anaerobically at 37°C. The bacterial lawn was harvested by adding 1ml of MilliQ water to the plate and scraping with a spreader. This step was repeated as required. One volume of 2M NaOH was added to the harvested bacterial biomass and the resulting solution was stirred gently for 16 h at 150rpm Next, the bacterial biomass solution was centrifuged for 25 minutes at 9200 rpm at 4°C and the resulting supernatant was collected. EPS was precipitated by adding two volumes of ice-cold absolute ethanol to the supernatant and storing at 4°C for at least 48 hours. The precipitated mixture was centrifuged for 25 minutes at 9200 rpm at 4°C. The supernatant was discarded, the precipitate was dissolved in 10ml MilliQ water and transferred to a pre-soaked Spectra / Por® Dialysis Membrane (Spectrum Laboratories, Inc., USA) with a molecular weight cut off of 8,000 or 10,000 Da. The dissolved precipitate was dialysed against MilliQ water with a daily water change for at least 48 hours. The dialysis product was transferred to sterile empty petti dishes (5ml / petri dish), coveted with parafilm which was then poked, and lyophilized overnight using Alpha 1-4; CHRIST LOC-lm (Christ, Germany) lyophilizer. The lyophilized crude EPS (cEPS) was harvested using a 10μl sterile plastic loop, transferred to a sterile cryotube and stored at 4°C.

[0076] 20mg of crude EPS were dissolved in 4ml of Buffet I (50 mM Tris-HCl pH 7.5, 10 mM MgSO4*7H2O), after which 1000x stock of DNase I (dissolved in Buffet I) was added to a final concentration of 5.5μg / ml. The solution was gently stirred on a shaker at 37°C for 6 hours. Then, stock of 100x Ptonase E dissolved in Buffet II (50 mM Tris-HCl pH 7.5, 2% EDTA pH 7-8) was added to a final concentration of 50μg / ml. The solution was gently stirred on a shaker for 18 hours at 37°C. As a next step, 60% TCA were added to a final concentration of 12%. The solution was transferred to 2ml Eppendotf tubes and incubated on a shaker at 21.5°C for 30 minutes at 350rpm, after which it was centrifuged at 13000 g for 25 minutes at 4°C. The supernatant was collected and adjusted to a pH of 5 with 10M NaOH. The solution was transferred to a pre-soaked Spectra / Por® Dialysis Membrane (Spectrum Laboratories, Inc., USA) with a molecular weight cut off of 10,000 Da and dialysed against MilliQ H2O at 4°C for 48 hours with a daily water exchange. The dialysed sample was then lyophilised and transferred to a sterile cryotube and stored at 4°C.

[0077] Statistical analysis Statistical analyses were performed using GraphPad Prism 9 software. Unless otherwise stated, Kolgorov-Smimov tests were performed to confirm normality of data and Student’s t-test (unpaired, two-tailed, at 95% confidence interval) were using to generate P-values. Where multiple t-tests were performed, a false discovery rate (FDR) of q<0.05 was used. Significant observations are represented according to the following annotation: ****P < 0.001, ***P < 0.001, **P < 0.01, *P < 0.05. In some figures, raw P-values which did not meet statistical significance are presented next to the corresponding data. Where no P-value is presented, statistical analyses did not identify a significant observation. Full details of specific statistical tests performed corresponding to each dataset can be found the respective figure legends. Quantifications show mean values ± SEM unless otherwise indicated.

[0078] Results

[0079] A Bifidobacterium cocktail treatment reduces tumour burden in breast cancer models

[0080] To explore the impact of Bifidobacterium in breast cancer progression, we first conducted shotgun metagenomics of faecal material from breast cancer patients to assess the prevalence and diversity of Bifidobacterium species. Analysis of the microbiome pre- and post-tumour resection surgery demonstrated that Bifidobacterium was consistently in the top 10 most abundant genus in human patients (Figure 1A). Characterisation of the Bifidobacterium populations demonstrated a relatively high level of species diversity, with Bifidobacterium longum, Bifidobacterium bifidum and Bifidobacterium adolescentis among the most highly represented (Figure IB).

[0081] Given the diversity of Bifidobacterium species observed in human patients, we devised a Bifidobacterium cocktail (Bifcocktail) to test in pre-clinical tumour models, comprised of four distinct strains: B. longim subsp. longim NCIMB 8809, B. bifidum LH80, B. pseudocatenulatum 210, and B. choerinum LH506. Oral administration of the Bifcocktail to animals bearing BRPKp110 luminal A-like breast tumours (42, 43) led to a significant reduction in BRPKp110 tumour volume (Figure 1C and reduced early dissemination of GFP+tumour cells to the lungs (Figure 1D-E). To dissect the microbial mechanisms driving tumour response, we deconstructed the Bifcocktail, testing the combined treatment against each constituent strain individually in the luminal B-like PyMT-BO1 breast tumour model (44). Whilst the Bifcocktail was ineffective in this model, three of the constituent strains (LH80, 210, and LH506) induced an anti-tumour response (Figure IF). Comparable tumour reductions were observed after administration of the effective single strains in the BRPKp110 model (Figure 1G), albeit with slightly less potent efficacy with strain LH506. The activity of CD8+T cells is consistently associated with the anti- tumour activity of Bifidobacterium (4, 10, 15) and is a vital pathway mediating patient outcomes. Only strain 210 treatment induced CD8+polarisation to an effector- memory subtype in PyMT-BO1 tumours (Figure 1H), suggesting a CDS^T cell- dependent mechanism driving 210 efficacy, and alternative mechanisms mediating response to LH80 and LH506.

[0082] B. pseudocatenulatum 210 exhibits consistent an3-tumour ac3vity across Breast Cancer models and enhances standard-of-care therapy

[0083] We focused on B. pseudocatenulatum 210 due to its strong and consistent anti-tumour effects. When administered as a monotherapy, 210 induced consistent anti-tumour efficacy across several pre-clinical orthotopic models representing luminal A (BRPKp110), luminal B (PyMT-BO1), and triple negative (4T1) breast cancer — significantly reducing tumour burden (Figure 2A). In addition, histological analyses showed that 210 also reduced 4T1 metastatic spread to the lungs (Figure 2B-C), demonstrating its systemic anti-tumour efficacy. In the autochthonous PyMT model, 210 monotherapy not only reduced tumour burden, but also delayed tumour formation, suggesting its potential for longer-term anti-tumour activity and prophylactic utility (Figure 2D). Furthermore, administration of 210 enhanced the efficacy of standard-of-care treatments, including cyclophosphamide chemotherapy in luminal BRPKp110 tumours and anti-PD-1 immunotherapy in triple negative 4T1 tumours (Figure 2E- F).

[0084] CD8+T cells drive the an3-tumour response to B. pseudocatenulatum 210

[0085] Next, we wanted to characterise the immune response driving tumour inhibition of 210. No significant changes were observed in the gross infiltration of major lymphoid populations in the tumour microenvironment, suggesting immune effects were more likely caused by differential polarisation and activation of immune populations. In agreement with our initial data in PyMT-BO1 tumours, 210 administration also increased BRPKp110 CD8+T cell polarisation towards an effector-memory subtype (Figure 3A), albeit not to a statistically significant level in the BRPKp110 model. CD8+cells in the spleen and tumour-draining lymph node (tdLN), were more polarised to the central-memory subtype, a state associated with long-lived antigenic memory (Figure 3B). Additionally, blood circulating CD8+T cells showed enhanced activation, marked by increased expression of the CD44 activation marker (Figure 3C). Cytokine analysis demonstrated that 210 induced increases in TNFa and (non- statistically significantly) IFNy in the primary tumour (Figure 3D) and were mirrored by a significant increase in serum IFNy levels (Figure 3E). BRPKp110 tumour infiltrating CD8+cells in both BRPKp110 and 4T1 tumour models (Figure 3F-G), secreted higher levels of TNFa and IFNy following 210 administration suggesting their involvement in gross increases observed for these cytokines. Enhanced cytolytic activity of CD8+T cells from 210-treated animals was confirmed by increased granzyme B and CD107a expression (Figure 3H-I). Concurrently, tumour infiltrating T helper cells were not more polarised or activated, aside from a small increase in IL-4 production restricted to the BRPKp110 model. Likewise, BRPKp110 tumour-infiltrating NK cells were not more abundant and did not produce any higher levels of inflammatory cytokines (including IFNy and TNFa), suggesting CD8+T cells may be the sole cytotoxic effector population mediating 210 anti-tumour response. Notably, depletion of CD8+T cells in vivo in BRPKp110 tumour-bearing animals abolished the anti-tumour efficacy of 210, validating a CD8+-dependent mechanism of action (Figure 3J).

[0086] B. pseudocatenulatum 210 enhances DC activation

[0087] Alongside the CD8+T cell response, we investigated the activity of innate immune cells given they orchestrate T cell populations. Within BRPKp110 and PyMT-BO1 primary tumours, we observed a significant reduction in CD206+tumour- associated macrophages (TAMs). These CD206+TAMs, historically characterised as ‘M2-like’ and pro-tumourigenic (45), were replaced with, MHCII+ TAMs (‘Ml- like’) indicative of a shift towards a more anti- tumourigenic, CD8+-privileged tumour microenvironment (TME). Although the inflammatory state of the TME is vital, we also showed systemic activation of CD8+T cells (in the spleen, tdLN, and blood), which may reflect upstream pathway activation away from the localised TME. Systemic analysis of circulating innate cells highlighted a near-significant increase in the circulating pool of DCs (Figure 4A).

[0088] DC infiltration and activation are pivotal for priming and activation of CD8+T cells due to their role as antigen presenting cells (46, 47). Systemic analysis indicated that 210 induced increases in circulating DCs (Figure 4A); particularly in the CD8+-specific cDC1 population, while no significant changes were observed in the T helper-specific cDC2 population (Figure 4B). Furthermore, circulating DCs and cDC1 cells primed by 210 displayed enhanced expression of CCR7, the chemokine receptor associated with pro-inflammatory migration to tdLN and the mounting of an anti-tumour immune response(46) (Figure 4C-D). This suggests that 210 may preferentially stimulate CD8+T cell responses via cDC1, pointing to a potential mechanistic cascade for its anti-tumour efficacy. Within the tdLN, a key site for CD8+T cell programming, we detected an increased infiltration of cDC1 cells (Figure 4E) alongside elevated expression of DC CCR7 (Figure 4F). Additionally, DCs in the tdLN of 210-treated animals showed elevated expression of CD80+CD86+maturation markers, approaching statistical significance (Figure 4G). Within the primary tumour, we observed a markedly increased infiltration of DCs and eDC subsets (Figure 4H), alongside a concurrent increase in DC maturation (Figure 41). Collectively, these data suggest that 210 enhances the systemic migration and maturation of DCs, potentially boosting their ability to sample and cross- present tumour-associated antigens and effectively stimulate naive CD8+T cells towards an effector phenotype. To further assess whether 210- induced DC activation was directly responsible for the enhanced anti-tumourigenic efficacy and increased CD8+activity, we co- cultured BMDCs with 210 cells and adoptively transferred the treated BMDCs to naive animals bearing BRPKp110 tumours (Figure 4J). Notably, 210-conditioned BMDC treatment significantly reduced BRPKp110 tumour burden and enhanced CD8+cell activity within the TME, as evidenced by increased production of inflammatory IFNy cytokine by CD8+cells. This confirms that 210-activated DCs are capable of robustly stimulating anti-tumour CD8+T cell immunity (Figure 4K).

[0089] B. pseudocatenulatum 210 anti-tumour effects are mediated by EPS

[0090] Having defined a DC and CD8+T cell-based mechanism underpinning 210 anti- tumour efficacy we next wanted to identify the functional compound(s) driving these host-microbe interactions. While immunogenic metabolite secretion is a commonly described mechanism in bacterial-mediated cancer immunity (48) untargeted metabolomics analysis of the serum from BRPKp110 and PyMT-BO1 tumour-bearing animals did not identify any significant changes to the global metabolome profile, or to the abundance of individual circulating metabolites following 210 administration (Figure 5A-B). Furthermore, differential expression analysis of serum metabolites across both models showed no consistent patterns of change (Figure 5C, suggesting that circulating metabolites were unlikely to be responsible for the observed 210-induced anti-tumour effects. To gain further clarity on the microbial dynamics and activity of 210, we conducted colonisation studies in germ free monocolonised mice, using standard oral administration at 1x1010 CFU. As expected, no viable bacteria were detected in vehicle control mice, whilst 210 monocolonised mice only had minimal viable bacterial colonies, localised primarily in the colon and caecum. Interestingly, live 210 cells were not able to establish prolonged colonisation and were largely cleared from the gut within 24 hours post-administration. This suggested a transient interaction between 210 and the host. Further validation using B. pseudocatenulatum specific PCR analysis of wild-type animals, confirmed that 210 cell load was most prominent in the caecum and colon, with 210 signature detectable up to 24 hours post- administration (Figure 32B). Despite the clearance of live bacteria, 210-specific signatures (GroEL gene), were detectable in faecal pellets. 210 gene signatures returned to approximate baseline levels up to 48 hours post oral administration, indicating that dead bacterial cells or bacterial fragments may persist longer in the gut environment. This persistence suggests that even non-viable bacterial components may play a role in modulating the immune response. Next, we sought to determine whether 210 could translocate to or colonise tumour tissues. Selective qPCR of BRPKp110 primary tumours and culturomic analysis of MMTV-PyMT primary tumours failed to detect any 210 cells, suggesting the absence of a direct tumour microbiome-based mechanism. To further investigate if 210’s anti-tumour effect was dependent on other gut commensal bacteria, we pre-treated BRPKp110 tumour bearing animals with a broad-spectrum antibiotic cocktail to deplete the native microbiota. Notably, antibiotic treatment did not rescue tumour growth inhibition in 210-treated animals, suggesting 210 exerts it effects independently of the commensal microbiome and directly induces anti-tumour immunity. The absence of involvement from other commensals, combined with the poor colonisation and viability of 210, and lack of 210-induced metabolic changes, suggested a mechanism driven by bioactive structural components rather than active microbial processes. To test whether viability of 210 cells was necessary for therapeutic efficacy, we administered peracetic acid-killed 210 cells, a method shown by Moor et al (27) to effectively kill microbes whilst preserving their surface structures. Acid killed 210 was able to inhibit BRPKp110 tumour growth in vivo to a similar extent as live 210 cells (Figure 6A), confirming that 210’s anti-tumour activity is not dependent on metabolite production or live cell interactions. Given that Bifidobacterium species and strains are known to produce immunogenic EPS on their cell surfaces, which have been shown to mediate specific immune responses in lymphoid populations (49) and DCs (23), we hypothesized that 210 EPS might be the key driver of its anti-tumour effects. To evaluate this, we administered isolated preparations of 210 EPS to BRPKp110 tumour bearing animals. Strikingly, 210 EPS treatment significantly reduced tumour progression, mirroring efficacy of live 210 cells (Figure 6B). Furthermore, 210 EPS replicated the observed increases in intratumoural CD8+T cell cytokine expression and effector-memory differentiation (Figure 6C), suggesting that EPS alone can trigger the same immunological responses as live 210. Within the TME, 210 EPS also reduced the infiltration of CD206+TAMs (Figure 6D). Consistent with live 210 treatment, 210 EPS elevated systemic IFNy levels (Figure 6E), and DC and cDC1 infiltration to levels even higher than those observed with live 210 (Figure 6F-G). Accordingly, 210 EPS did not induce significant changes in infiltration or activity of other lymphoid effector populations, such as T helper or NK cells in the TME. To validate that the 210 EPS-specific interactions with DCs were causative in the mechanism, we performed a BMDC adoptive transfer experiment. BMDCs conditioned with 210 EPS, but not vehicle control or EPS from a B. longum strain (B71) shown to be ineffective in tumour reduction, were able to induce significant anti- tumour activity in naive animals (Figure 6H). Moreover, 210 EPS-conditioned BMDCs stimulated robust CD8+T cell activation (Figure 61) within the tumour. These data demonstrate that B. pseudocatenulatum 210 EPS is the key bioactive compound driving the DC-based immune response, leading to CD8+T cell- mediated tumour suppression.

[0091] B. pseudocatenulatum EPS is highly variable and functions in a strain- dependent manner

[0092] To gain insight into the uniqueness of 210 EPS relative to other strains and species of Bifidobacterium, we undertook comparative analyses of the putative EPS enzymatic clusters within the genomes of the strains used in the study. Single components of the enzymatic clusters within the 210 genome showed minor homology to some enzymes within the B. bifidum LH80, B. cboerinum LH506 and B. Ionium subsp. Ionium NCIMB 8809 genomes, including glycosyl transferases and ABC-transporters. However, the majority of the predicted 210 EPS cluster components were not homologous to those identified in the other Bifidobacterium species, suggesting the resulting EPS structures from these strains may be highly variable from one another and species specific. Comparatively, the architecture of the EPS clusters putatively encoded by B. Ionium 8809 and B. Ionium B71, whose EPS were both shown to be ineffective in reducing tumour burden, displays near complete homology across most cluster components, suggesting their resulting EPS structures may be similar. Comparison of the 210 EPS cluster architecture with that of other strains of B. pseudocatenulatum showed high levels of homology between 210 and the B. pseudocatenulatum DSM20438 type strain, with lower levels of homology to individuals genes encoded by the strain B. pseudocatenulatum LH14 (Figure 7A-B). As expected, none of the B. pseudocatenulatum strains encoded genes with significant homology to those identified in the B. Ionium B71 strain previously used as a negative control. Given the similarity of the putative EPS genes encoded by 210 and DSM20438, and the large differences to those of LH14, we administered each of these strains to BRPKp110 tumour-bearing animals to ascertain whether these differences observed in the architecture of EPS clusters would be reflected functionally in (EPS-driven) tumour reduction. Indeed, administration of DSM20438 resulted in significant tumour inhibition comparable with that of 210 (Figure 7C), whilst strain LH14 was ineffective in reducing tumour progression (Figure 7D). AdministraPon of B. longum B71 was also ineffective in reducing tumour burden (Figure 7E). Immunological analyses of DSM20438- treated animals showed further functional consistency between DSM20438 and 210, as primary tumour CD8+T cells were more activated (with higher IFNg expression) and polarised to an effector-memory subtype. Comparatively, animals administered with LH14 did not display the characteristic increase in primary tumour CD8+cytokines or effector polarisation. DSM20438-treated tumours also displayed highly elevated infiltration of DCs and eDC subsets, alongside a dramatic increase in DC maturation. The DC infiltrate increase was further reflected in the tdLN, showing a systemic DC-increase following DSM20438 administration. In vitro analysis of BMDCs exposed to purified EPS from 210, DSM20438 and B. longum B71 further mirrored the functional differences observed in vivo. Although each EPS sample was able to stimulate the BMDC maturation in a dose dependent manner, reaching maximal maturation at super-physiological doses, the EPS from 210 and DSM20438 demonstrated a much greater ability to enhance BMDC MHCII expression and CD80+CD86+maturation at a more physiologically relevant dose (10μg) than the B71 EPS (Figure 7F). This demonstrates that the B. pseudocatenulatum EPS isolated from 210 and DSM20438 are functionally similar in vitro whilst having greater immunostimulatory potential at lower doses than the ineffective B. longum B71 EPS.

[0093] Put together, these data suggest that B. pseudocatenulatum 210 and DSM20438 induce the same anti-tumour immunological mechanism through the production of similar EPS structures. Our data also suggest that the EPS produced by different Bifidobacterium species is highly strain dependent. We observed functional similarities between B. pseudocatenulatum 210 and B. pseudocatenulatum DSM20438 in vivo and their respective EPS isolates in vitro, with corresponding functional differences compared to strain of B. pseudocatenulatum LH14 and other species of Bifidobacterium (B. gm g71). The similarities and differences between different strains and EPS functionality was closely reflected by the diversity of EPS genomic clusters among these strains. These data reinforce an important and unique role for B. pseudocatenulatum EPS in mediating anti-tumour immune responses.

[0094] Discussion

[0095] Therapeutically harnessing gut microbes to treat chronic disease is a fundamental aim in microbiome research. Whilst numerous studies have highlighted the potential of microbiota- based interventions, the success of translating these findings to extra-intestinal disease therapies remains limited. A significant barrier to this translation is our incomplete understanding of the specific microbial-derived functional compounds that drive therapeutic responses. Without identifying these key compounds, predicating how they may be influenced by variable patient- specific conditions becomes difficult. In this study, we demonstrate the broad potential of various species of Bifidobacterium to treat breast cancer.

[0096] We provide strong evidence that B. pseudocatenulatum 210 inhibits breast tumour progression and enhances the efficacy of standard-of-care therapies through presentation of cell surface capsular EPS. This is the first in vivo study demonstrating Bifidobacterium EPS as an anti-tumour compound, advancing the concept of microbial polysaccharides as novel immunotherapy agents.

[0097] The link between Bifidobacterium and cancer has been an active area of research for the past decade. Sivan et al, (4) were the first to show that a cocktail of four Bifidobacterium species eennhhaanncceedd i immmmuunnee checkpoint inhibitors (ICIs) responsiveness to melanoma, a mechanism dependent on DCs-and live bacteria. While their study implicated secreted metabolites, it left open questions about the specific microbial components driving the response. Subsequent studies have demonstrated the anti-tumour potential of single Bifidobacterium strains (70, 15, 19), including in colorectal cancer and melanoma models. However, few studies have explored the efficacy of Bifidobacterium in breast cancer or identified the microbe- derived active compounds mediating protective effects.

[0098] In this study, we observed that a Bifidobacterium (Bif)-cocktail induced anti-tumour efficacy in luminal A (BRPKp110) but not luminal B (PyMT-BO1) model. Further dissection of the Bif- cocktail revealed that three of its four component strains (B. bifidum LH80, B. choerinum LH506, and B. fiseudocatenulatum 210) independently inhibited tumour progression, whilst B. longum NCIMB 8809 did not. Intriguingly, this suggested that B. longum NCIMB 8809 may have inhibited or diluted the efficacy of the other strains. This highlights a challenge in using microbial consortia or complex communities of live bacteria - interactions within these mixtures may obscure or counteract the therapeutic potential of individual strains. Our findings underscore the importance of tailored microbiome therapies using carefully selected strains to optimise therapeutic pathways. Focusing on the individually effective single strains, analysis of PyMT-BO1 intratumoural CD8+cells showed B. fiseudocatenulatum 210 induced effector-memory polarisation, whilst B. bifidum LH80 and B. choerinum LH506 did not. This finding suggests that although multiple Bifidobacterium species and strains may have therapeutic potential against (breast) cancer, the underlying molecular mechanisms mediating responses are likely distinct. Given these differences, future therapeutic strategies could benefit from using tailored combinations of microbes to synergistically target multiple therapeutic pathways to enhance patient responses across different cancer subtypes. Our investigation into B. pseudocatenulatum 210 revealed potent anti-tumour effects across multiple breast cancer models, including long-term efficacy in the spontaneous PyMT model and enhanced responses to standard therapies, such as chemotherapy and ICI treatment in luminal and triple-negative breast cancers, respectively. These results are the first to demonstrate the in vivo therapeutic potential of B. pseudocatenulatum in cancer. Mechanistically 210’s efficacy was shown to be dependent on CD8+T cells, with 210 administration promoting the maturation of cDC1 DCs and stimulating effector-memory polarisation of CD8+T cells. Importantly, these effects were observed not only within the tumour, but also at systemic sites such as tdLN, spleen, and blood, pointing to a systemic activation of CD8+-driven anti-tumour immunity. Further analysis of the innate immune compartment showed that 210 treatment increased the number of pro- inflammatory macrophages and cDC1 DCs in the blood and tdLN, supporting CD8+T cell activation and anti-tumour responses. The observed increases in systemic DC maturation and activation, alongside the reduced infiltration of CD206+TAMs in the TME, suggest 210 induces asymmetric shift towards a pro- inflammatory, anti-tumour immune profile. Our adoptive transfer experiments with 210-conditioned BMDCs confirmed the causative role of 210- activated DCs in driving CD8+T cell-mediated tumour suppression, providing further evidence for the mechanistic basis of 210’s efficacy.

[0099] Despite the compelling evidence of B. pseudocatenulatum 210’s therapeutic potential, our data showed that 210 did not robustly colonise the gastrointestinal tract of germ-free mice, nor did it induce significant changes to the metabolome. Rather, acid killed 210 maintained anti-tumour efficacy suggesting structural components, rather than live bacterial activity or metabolic output, mediate responses. This led us to hypothesise that 210’s EPS, rather than its metabolites, was the key effector molecule. EPS are known immunogenic components of many gut bacteria, including Bifidobacterium, and have been shown to modulate immune responses through interactions with DCs and other immune cells. In our study, isolated EPS from B. pseudocatenulatum 210 replicated the anti-tumour effects of live 210, including the stimulation of CD8+T cell activity, DC maturation, and TAM repolarisation. Adoptive transfer of BMDCs conditioned with 210-derived EPS further confirmed its role as the critical driver of anti-tumour immunity. Our findings align with emerging evidence of microbial polysaccharides as potent immune modulators in cancer therapy. A recent study demonstrated that soluble 0- glucan polysaccharide could enhance anti-tumour responses by activating Dectin- 1 receptors on immune cells, thus potentiating CD8+T cell activity (50). Similarly, our study demonstrates that 210’s EPS stimulates cDC1 maturation and primes DCs to activate CD8+T cells, providing new mechanistic insight into the anti- tumour activity of microbial polysaccharides. Analysis of the architecture of putative EPS genomic clusters of revealed high levels of genomic variability between individual strains and species, but importantly revealed that the presence of highly homologous EPS-synthesising clusters correlated with functionally similar responses both in vivo and in vitro. This was demonstrated by in vivo inhibition of tumour progression and enhanced anti-tumour immune responses by B. fiseudocatenulatum 210 and B. pseudocatenulatum DSM 20438, which correlated with both strains harbouring highly homologous EPS clusters. This contrasted with the absence of functional responses in vivo from genetically distinct B. pseudocatenulatum LH14 and B. longum B71. Likewise, EPS isolated from 210 and DSM20438 was more immunostimulatory to BMDCs at lower doses in vitro compared with B71 EPS. These results highlight the potential for EPS-based therapies to serve as stand- alone immunotherapeutics or as adjuvants to enhance the efficacy of existing treatments like I Cis. Genetic analyses highlight the high levels of diversity in EPS production across Bifidobacterium, with such diversity correlated to functional differences in the resulting EPS. Thus, genomic analysis of EPS enzymes could be used as a biomarker or screening tool to identify new types of EPS with utility across a wide range of inflammatory contexts.

[0100] The clinical relevance of these findings is underscored by the growing body of evidence linking Bifidobacterium species to favourable positive cancer outcomes (72, 51). Of specific interest, is the correlation of B. pseudocatenulatum, alongside other beneficial species like Roseburia spp. and Akkermansia muciniphila, with enhanced overall response rates and progression-free survival in several cohorts of melanoma patients receiving I Cis (77). By demonstrating that B. pseudocatenulatum EPS drives CD8+T cell-mediated anti-tumour immunity in breast cancer, we provide a strong foundation for future clinical translation of Bifidobacterium-based interventions, particularly in high-risk populations where disparities in cancer outcomes persist.

[0101] In conclusion, this study identifies B. pseudocatenulatum 210 EPS as a novel anti- tumour therapeutic compound. Our findings represent a significant advancement in understanding the functional role of microbial polysaccharides in cancer immunotherapy and provide a compelling case for further exploration of microbial EPS in therapeutic contexts. Future work should focus on translating these findings to human clinical trials, with an emphasis on personalising microbiota- targeted therapies to improve patient outcomes.

[0102] 

[0103] Bifidobacterium comparative genomics analyses

[0104] Bifidobacterium DNA extraction and WGS

[0105] Bacterial pellets were resuspended into MPBio Lysing Matrix E bead beating tubes (MPBio) in sodium phosphate buffer and DNA was extracted using the MPBio FastDNA™ SPIN Kit for Soil (MPBio) following the manufacturers protocol. The recovered DNA was validated using a Qubit® 2.0 fluorometer (Invitrogen).

[0106] Comparative genomics sequencing and bioinformatic analysis

[0107] Genomes of pseudocatenulatum strains were sequenced as previously described(7). The sequencing reads were pre-processed with fastp v0.22(2) with default settings and assembled using Unicycler v0.4.9(3) in “conservative” mode, with the minimum contig length set to 100Obp. After pre-processing, Unicycler v0.4.9 was used to generate a hybrid assembly using both short and long reads, with the minimum contig length of 1000bp. Completeness and contamination of new assemblies were estimated at >99.5% and <0.5% at the family level, respectively, using CheckM vl.2.0(4). All genomes were annotated with Prokka vl.14.6(5).

[0108] 2. Materials and methods

[0109] 2.1. Mouse models

[0110] C57 BL / 6 mice were purchased and maintained in-house at the Disease Modelling Unit at the University of East Anglia. BALE / C mice were purchased from Chades River and maintained in-house. Animals used throughout were age 8-12 weeks and were randomly mixed between cages prior to experiment onset. All animal experiments were performed in accordance with UK Home Office regulations and the European Legal Framework for the Protection of Animals used for Scientific Purposes (European Directive 86 / 609 / EEC).

[0111] 2.1.1. Orthotopic breast tumour models

[0112] Synergic breast cancer cells were injected in 50μl of a 1:1 mixture of phosphate buffered saline (PBS) and Matrigel (Coming Life Sciences, Coming, USA) into the left inguinal mammary fat pad of age-matched female mice. PyMT-BO1, E0771 and 4T1 cells were each injected 1x105and BRPKp110 cells were injected at 5xl05. The experimental duration for each breast model used is outlined in the associated figures. In situ tumour volumes were measured thrice weekly with digital calipers from the onset of a palpable tumour, using the formula: length x width2x 0.52(7).

[0113] 2.2. In vivo experiments

[0114] 2.2.1. Bacteria and bacterial product administration

[0115] Animals were orally administered thrice weekly with live Bifidobacterium strains (1x1010CFU / 200μl) or isolated EPS (80μg / 200μl) from the onset of a palpable tumour to endpoint. Treatments were administered on tumour model-specific outlines as detailed in the experimental workflow figures.

[0116] 2.2.2. Cyclophosphamide chemotherapy experiments

[0117] Following the formation of palpable tumours, cyclophosphamide (Sigma) was administered by intraperitoneal injection at 100mg / kg on days 10 and 17 (for BRPKp110 experiment) or days 7 and 14 (PyMT-BO1 experiment).

[0118] 2.2.3. aPD-1 immune checkpoint experiments

[0119] Tumour-bearing animals were intraperitoneally administered 20mg / kg anti-PD-1 mAb (Clone J 43, BioXCell) or matched isotype control on day 7, 10, 13 (4T1 experiment) or day 10, 13, 16, 19 (BRPKp110 experiment). 2.2.4. Dendritic cell adoptive transfer experiment

[0120] Dendritic cells were generated as outlined in 2.5.4. before 24 hour co-culture of 1x106DCs per well in a 6 well plate with 10μl PBS (Control), live B. pseudocatenulatum LH663 (at a multiplicity of infection of 5:1 bacterial : mammalian cells), 100 μg B. pseudocatenulatum LH663-EPS, or 100 pg B. longum B71-EPS. BRPKp110 tumour-bearing animals were injected with 1x105pre-treated DCs intravenously on day 10, 14, and 18.

[0121] 2.3. Bacterial culture and preparation

[0122] 2.3.1. Bifidobacterium

[0123] All Bifidobacterium strains used for this study were isolated previously by the laboratory of Prof. Lindsay Hall. The strains were cultured at 37°C in MRS broth with L-cysteine (50mg / L) (Sigma) in an anaerobic chamber (Don Whitley Scientific, Bingley, UK). Strains were cultured for one week and then preserved by lyophilisation in the exponential phase of growth. Lyophilised bacteria were equally distributed across individual dosage vials which were stored at -80°C. To ensure accurate dosing, at least 3x vials from each lyophilisation batch were enumerated by counting CPUs on MRS agar plates across serial PBS dilutions. The mean CPU was calculated for each batch and vials were resuspended in PBS (to 1x1010CFU / 200μl / animal) immediately prior to experimental administrations.

[0124] 2.3.2. Peracetic acid preparation of killed bacteria

[0125] For acid-killed experiments, peracetic acid pre-treatment of bacteria was performed as previously described(2 ). Briefly, lyophilised bacteria were reconstituted in 10ml of sterile PBS at a concentration of ~1x1010CFU / ml. Peracetic acid (Sigma) was added to a final concentration of 0.4% and bacteria were incubated at RT for 1 hour. The bacteria were washed three times in sterile PBS and then resuspended to the appropriate final concentration for animal administrations. Bacterial killing was confitmed by culturing 100μl of acid-killed bacteria on MRS agar under aerobic conditions (outlined in section 2.3.1.) and validating the absence of bacterial growth.

[0126] 2.3.3. Exopolysaccharide isolation and purification

[0127] The Hall laboratory (Technical University of Munich, Munich, Germany) conducted EPS isolation and purification based on the protocol by Ruas-Madiedo (2021) (3). A 20ml bacterial suspension in MRS broth was cultured overnight at 37°C under anaerobic conditions. 200μl of the liquid culture were plated on MRS + L-cysteine (50mg / L), MRS + L-cysteine (50mg / L) + fructose, or MRS + L- cysteine (50mg / L) + arabinose agar plates and incubated for 96 hours anaerobically at 37°C. The bacterial lawn was harvested by adding 1ml of MilliQ water to the plate and scraping with a spreader. This step was repeated as required. One volume of 2M NaOH was added to the harvested bacterial biomass and the resulting solution was stirred gently for 16 h at 150rpm Next, the bacterial biomass solution was centrifuged for 25 minutes at 9200 rpm at 4°C and the resulting supernatant was collected. EPS was precipitated by adding two volumes of ice-cold absolute ethanol to the supernatant and storing at 4°C for at least 48 hours. The precipitated mixture was centrifuged for 25 minutes at 9200 tpm at 4°C. The supernatant was discarded, the precipitate was dissolved in 10ml MilliQ water and transferred to a pre-soaked Spectra / Por® Dialysis Membrane (Spectrum Laboratories, Inc., USA) with a molecular weight cut off of 8,000 or 10,000 Da. The dissolved precipitate was dialysed against MilliQ water with a daily water change for at least 48 hours. The dialysis product was transferred to sterile empty petti dishes (5ml / petri dish), covered with parafilm which was then poked, and lyophilized overnight using Alpha 1-4; CHRIST LOC-lm (Christ, Germany) lyophilizer. The lyophilized crude EPS (cEPS) was harvested using a 10 μl sterile plastic loop, transferred to a sterile cryotube and stored at 4°C. 20mg of crude EPS were dissolved in 4ml of Buffer I (50 mM Tris-HCl pH 7.5, 10 mM MgSO4*7H2O), after which 100Ox stock of DNase I (dissolved in Buffer I) was added to a final concentration of 5.5μg / ml. The solution was gently stirred on a shaker at 37°C for 6 hours. Then, stock of 100x Pronase E dissolved in Buffer II (50 mM Tris-HCl pH 7.5, 2% EDTA pH 7-8) was added to a final concentration of 50μg / ml. The solution was gently stirred on a shaker for 18 hours at 37°C. As a next step, 60% TCA were added to a final concentration of 12%. The solution was transferred to 2ml Eppendorf tubes and incubated on a shaker at 21.5°C for 30 minutes at 350rpm, after which it was centrifuged at 13000 g for 25 minutes at 4°C. The supernatant was collected and adjusted to a pH of 5 with 10M NaOH. The solution was transferred to a pre-soaked Spectra / Por® Dialysis Membrane (Spectrum Laboratories, Inc., USA) with a molecular weight cut off of 10,000 Da and dialysed against MilliQ H2O at 4°C for 48 hours with a daily water exchange. The dialysed sample was then lyophilised and transferred to a sterile cryotube and stored at 4°C.

[0128] 2.4. Cancer cell culture

[0129] All tumour cell lines were cultured in high glucose DMEM (Thermofisher) supplemented with 10% foetal bovine serum (FBS) (Hyclone, Thermo fisher) and 100 units / ml penicillin / streptomycin (Thermofisher). Cells were seeded onto flasks coated with 0.1% porcine gelatin (Sigma) and incubated at 37°C and 5% CO2.

[0130] 2.5. In vitro experiments

[0131] 2.5.1. THPl-Blue reporter cell culture and assays

[0132] THPl-Blue cells were purchased from Invivogen and cultured in RPMI1640 (Sigma) supplemented with 10% FBS (Hyclone, Thermofisher); 1% Penicillin- Streptomycin (Thermofisher), 100 μg / mL Normocin (Invivogen) and 10 μg / mL blasticidin (Invivogen). For NF-KB activity assays, 1x105cells were seeded in 96- well plates and cultured for 24 hours with excipients and LPS positive control. NF- KB activity was measured using QUANTI-Blue detection medium (Invivogen) according to the manufacturer’s instructions. Absorbance was read at 650nm in a microplate reader (Biotec, USA).

[0133] 2.5.2. HEK-Blue hTLR reporter cell culture and assays

[0134] HEK-Blue hTLR2, HEK-Blue hTLR4, and HEK-Blue hTLR5 cells were all purchased from Invivogen. Cells were cultured in RP MI 1640 supplemented with 10% FBS (Hyclone, Thermofisher); 1% Penicillin-Streptomycin (Thermo fisher), 100 μg / mL Normocin (Invivogen). TLR2 and TLR4 reporter cells were additionally supplemented with lx HEK-Blue™ Selection cocktail, and TLR5 reporter cells were supplemented with 30 μg / ml of blasticidin (Invivogen) and 100 |ig / ml of Zeocin (Invivogen). For TLR activity assays, reporter cells were seeded at 5x104cells in 96-well plates in HEK-Blue detection media with experimental interventions and respective positive controls; LTA-BS (TLR2), LPS (TLR4), and recFLA-ST (TLR5), all purchased from Invivogen. Cells were incubated overnight, and absorbance measured at 650nm in a microplate reader (Biotec, USA).

[0135] 2.5.3. CD8+T cell co-cultures

[0136] CD8+T cells were isolated from spleens (see section 2.7.1. for spleen cell preparation) using the CD8a+T Cell Isolation Kit, mouse (Mfltenyi Biotec) according to the manufacturer’s instructions. Purified CD8+T cells were cultured in RP MI 1640 medium supplemented with 10% FBS, 1% penicillin / streptomycin (Gibco). Cells were seeded at 2x105cells per well of a 96-well plate in combination with the indicated treatment. After 24 hours, supernatants were analysed for cytokine levels by MSD multiplex assay.

[0137] 2.5.4. Bone marrow dendritic cell (BMDC) co-cultures

[0138] BMDCs were generated from flushed bone marrow from tibias and femurs of C57 BL / 6 mice. Isolated bone marrow was treated with red blood cell lysis buffer for 5 minutes, washed twice in PBS, and cultured overnight in RP MI 1640 medium supplemented with 10% FBS, 1% penicillin / streptomycin (Gibco), 20ng / ml IL-4 (R&D) and lOng / ml GM-CSF (R&D). On the second day, supernatant with non- adherent cells was removed and culture media replaced. Cell culture medium was replaced again on the fifth day, and semi-adherent BMDCs were collected on the seventh day for experimental use. Purified BMDCs were seeded at 2.5xl05cells per well in 24 well plates, incubated for 48 hours with indicated treatments and LPS positive control, then analysed by flow cytometry for markers of cell maturation.

[0139] 2.6. Organ histology and immunostaining

[0140] 2.6.1. Formaldehyde fixed paraffin embedded tissue for histology

[0141] Harvested organs were incubated overnight in 4% paraformaldehyde (PFA) at 4°C and processed with the Leica Tissue Processor ASP-300-S (Leica Biosystems, Milton Kenes, UK). The tissues were incubated in formalin, dehydrated through increasing concentrations of ethanol (from 70% to 100%), washed in three changes of xylene (Sigma-Aldrich) and then embedded in paraffin (Sigma- Aldrich). Paraffin blocks were sectioned at 6μm using a rotary microtome (Leica Biosystems, RM2235), mounted onto positively charged glass slides (Thermofisher), and incubated overnight at 37°C.

[0142] 2.6.2. Haematoxylin & eosin (H&E) staining

[0143] Prior to histological staining, FFPE tissue sections were deparaffinised in xylene (Sigma) and rehydrated through sequentially decreasing concentrations of ethanol (100%-70%), then into water. H&E staining was performed using a Leica ST5020 tissue multi-stainer (Leica Biosystems, Nussloch, Germany) and sections were then mounted with coverslips with Neo-Mount™ (Sigma). Images were captured using the Olympus BX60 microscope (Olympus, Southend-on-Sea, UK) with a microscope camera Jenoptik C10 and ProgRes CapturePro software v.2.10. 2.7. Flow cytometry

[0144] 2.7.1. In vivo tissue single cell isolation

[0145] Tumours and lungs were excised and mechanically homogenised using scalpels. Tumours were digested in 0.2% collagenase IV (Thermofisher) and lungs in 0.2% collagenase I (Thermofisher), with 0.01% hyaluronidase (Sigma) and 0.01% DNase I (in HBBS) at 37°C under agitation for 60 minutes (tumours) or 30 minutes (lungs). Tissues were then passed through a 70μm filter (Thermofisher) and washed in PBS before further staining. Spleens and lymph nodes were mechanically dissociated through 70μm strainers. Lymph nodes were immediately processed for staining and tumours, lungs and spleens were resuspended red blood cell lysis buffer (Thermofisher) for 5 minutes. Cells were washed in PBS and counted, with 1 million cells being stained per sample in FACS buffer (2% FBS in PBS).

[0146] 2.7.2. Cell staining protocol

[0147] For intracellular cytokine analysis, cells were resuspended in 200μl RPMI supplemented with 10% FBS, 50pM 2-Mercaptoethanol, 50ng / ml Phorbol 12- Myristate 13-Acetate (PMA), 750ng / ml lonomycin and 10μg / ml Brefeldin-A (all purchased from Sigma) and incubated in a 96-well U-bottom plate (Sigma) at 37°C, 5% CO2for 4 hours. Cells were blocked with Fc-receptor blocking reagent (Thermofisher) and incubated in relevant antibody (Table 2.1) and fixable Live / Dead Red (Invitrogen) solutions for 30 minutes (at 4°C in the dark). Cells were washed twice in FACS buffer, fixed with 4% PFA for 30 minutes, and then resuspended in FACS buffer prior to analysis. Where intracellular staining was required, cells were treated using the eBioscience™ Foxp3 / Transcription Factor Staining Buffer Set, as per the manufacturer’s instructions, then stained with intracellular antibodies.

[0148] 2.7.3. Data collection and analysis Data collection was conducted on the BD LSR Fortessa cell analyser and analysed using Flowjo software (BD). All samples were initially gated using FSC-A vs. FSC- H to identify single cells (Singlets), which were then gated for Live / Dead Red negative. Major immune populations were identified using the marker profiles outlined in Table 2.2.

[0149] 2.8. Mesoscale discovery (MSD) multiplex cytokine arrays

[0150] Tissue samples were weighed into a MPBio Lysing Matrix E bead beating tube (MPBio) with 1ml of homogenisation buffer (150 mmol / L NaCl, 20 mmol / L Tris, 1 mmol / L EDTA, 1 mmol / L EGTA, 1% Triton X-100, pH 7.5+ cOmplete™ protease inhibitor (Roche)). An MPBio Fast Prep bead beater (MPBio) was used to homogenise the tissues at a speed 4.0 for 40 seconds, followed by speed 6.0 for 40 seconds. Samples were centrifuged at 12,000 x g for 12 minutes (4°C and then stored at — 80°C until analysed. Samples were run on a custom Mesoscale Discovery U-PLEX Mouse Kit (MSD) according to the manufacturer’s instructions. Plate was read using an MSD QuickPlex SQ 120 imager (MSD, Rockville, MD, USA).

[0151] 2.9. Bifidobacterium exopolysaccharide structural analyses 2.9.1. Monosaccharide composition analysis by alditol acetate derivatisation

[0152] Glycosyl composition analysis was performed by combined gas chromatography- mass spectrometry (GC-MS) of the alditol acetates (AAs) as described by Pena et al. (2012)(4). The exopolysaccharide sample was hydrolysed in 2M trifluoroacetic acid (TFA) for 2 hours in a sealed tube at 120 °C, reduced with NaBD4, and acetylated using acetic anhydride / TFA. The resulting alditol acetates were analysed on an Agilent 7890A GC interfaced to a 5975C MSD, electron impact ionisation mode. Separation was performed on a 30-m Supelco SP-2331 bonded phase fused silica capillary column.

[0153] 2.9.2. Glycosyl linkage analysis

[0154] Glycosyl linkage analysis was performed by combined gas chromatography-mass spectrometry (GC-MS) of the partially methylated alditol acetate (PMAA) derivatives produced from the sample. The procedure is described by Anumula and Taylor (1992) (J). Briefly, permethylation of the sample was achieved by two rounds of treatment with sodium hydroxide (15 minutes) and methyl iodide (30 minutes). The sample was then hydrolysed using 2M TFA (2 hours in sealed tube at 120 °C , reduced with NaBD4, and acetylated using acetic anhydride / TFA. The resulting PMAAs were analyzed on an Agilent 7890A GC interfaced to a 5975C MSD (mass selective detector, electron impact ionisation mode); separation was performed on a 30 m Supelco SP-2331 bonded phase fused silica capillary column for the neutral residues and an EC-1 column for the amino containing residue.

[0155] 2.10. NMR analysis 2.11. EPS size analysis by QToF mass spectrometry

[0156] Sample was analysed using a Waters Cyclic Select Series QTof. The sample was infused directly into the mass spectrometer at a rate of 10 μl / min with electrospray ionisation in positive mode. The mass spectrometer was used in profile scanning positive mode from 50-1200 Da at 1.2 scans per second at a resolution of 50,000.

[0157] 2.12. Statistical analysis

[0158] Statistical analyses were performed using GraphPad Prism 9 software. Unless otherwise stated, Kolgorov-Smimov tests were performed to confirm normality of data and Student’s / -test (unpaired, two-tailed, at 95% confidence interval) were using to generate P-values. Where multiple t-tests were performed, a false discovery rate (FDR) of q<0.05 was used. Significant observations are represented according to the following annotation: ****P < 0.001, ***P < 0.001, **P < 0.01, *P < 0.05. In some figures, raw P-values which did not meet statistical significance are presented next to the corresponding data. Where no P-value is presented, statistical analyses did not identify a significant observation. Full details of specific statistical tests performed corresponding to each dataset can be found the respective figure legends.

[0159] 1. M. M. Tomayko, C. P. Reynolds, Determination of subcutaneous tumor size in athymic (nude) mice. Cancer Chem other Pharmacol TA, 148-154 (1989).

[0160] 2. K. Moor et al, Peracetic Acid Treatment Generates Potent Inactivated Oral Vaccines from a Broad Range of Culturable Bacterial Species. Pront Immunol 1, 34 (2016). 3. P. Ruas-Madiedo, Detection, Isolation, and Purification of Bifidobacterial Exopolysaccharides. Methods Mol Biol 2218, 101-115 (2021).

[0161] 4. M. J. Pena, S. T. Tuomivaara, B. R Urbanowicz, M. A. O'Neill, W. S. York,

[0162] Methods for structural characterization of the products of cellulose- and xyloglucan-hydrolyzing enzymes. Methods Engymol 510, 121-139 (2012).

[0163] 5. K R Anumula, P. B. Taylor, A comprehensive procedure for preparation of partially methylated alditol acetates from glycoprotein carbohydrates. Anal Biochem 203, 101-108 (1992).

[0164] Figure 1. Bifidobacterium is abundant in patients and can induce anti- tumour efficacy in mouse breast models. (A) Relative abundance of the top ten bacterial genera in faecal microbiomes of breast cancer patients from the BEAM cohort. n= 15. Samples represent pooled abundances from baseline and post- treatment timepoints. (B) Species-level composition of the Bifidobacterium genus within the same cohort. Only species with >1% relative abundance in at least one sample are shown. (C) Primary tumour size of Bif cocktail- treated animals compared with PBS vehicle control. n=7-8. C57 BL / 6 mice were orally dosed with ~lxl010 CFU / ml Bifidobacterium thrice weekly upon the onset of a palpable tumour (day 10). (D) Representative flow cytometry plots of treatment and control groups for the assessment of GFP+BRPKp110 cell infiltration into the lungs of primary tumour-bearing animals. (E) Quantification of GFP+BRPKp110 cell infiltration to the lungs following treatment with the Bif cocktail, n=4. (F) Endpoint PyMT-BO1 tumour volumes following administration of various unique strains of Bifidobacterium, or the four-strain consortia (Bif cocktail) comprised of each the individual strains. n=9-10. (G) Endpoint BRPKp110 tumour volumes following administration of Bifidobacterium strains. n=7-9. (H) Quantification of CD8+effector-memory polarisation following administration of Bifidobacterium treatments. n=4-5. Statistical significance calculated by (C and E) two-tailed impaired t test or (F-H) one- way AN OVA with Tukey’s multiple comparisons test. **P < 0.01, *P < 0.05.

[0165] Figure 2. B. pseudocatenulatum 210 treatment inhibits breast tumour progression and enhances response to standard of care therapies. (A) Experimental outlines and tumour growth responses to 210 treatment in the BRPKp110 (n=17-18, N=2), PyMT-BO1 (n=8-9), and 4T1 (n=9) orthotopic tumour models. (B) Quantification showing the average size (n=6-24) and (C) number (n=9) of macromolecular metastatic nodules observable in the lungs of 4T1- bearing animals visualised following histological H&E staining. (D) MMTV- PyMT spontaneous tumour burden growth and tumour onset time following 210 administration. n=12. (E) Experimental outline and associated BRPKp110 tumour growth responses of animals administered combinations of 210 and cyclophosphamide chemotherapy (n=6-8), and (F)

[0166] 4T1 tumour-bearing animals treated with combinations of 210 and anti PD-1 immunotherapy. Statistical significance calculated by (A, D, E, F) two-tailed unpaired t test, (B) Mann-Whitney test, (C) two-tailed unpaired t test with Welch’s correction, and (D) Kaplan-Mieier survival analysis of tumour free animals. ****P < 0.0001, **P < 0.01, *P < 0.05.

[0167] Figure 3. B. pseudocatenulatum 210 anti-tumour efficacy is dependent on CD8+T cell activity. (A) Representative flow cytometry plot showing with quantification showing effector-memory CD8+T cell polarisation in PyMT-BO1 primary tumours following 210 administration. n=ll-12, n=5. (B) Quantification showing CD8+T cell central-memory cell polarisation (n=6, n=4) and (C) CD44+expression (n=8, n=7) in the indicated tumour models and tissues. (D) Quantification of IFNy and TNFa levels (n=4-5) in primary tumours and (E) IFNy levels in the sera (n=5) of BRPKp110-bearing animals measured by MSD multiplex cytokine analysis. (F) Co-expression of IFNy and TNFa (n=6-7) by BRPKp110 primary tumour CD8+T cells following 210 treatment. (G) Representative flow cytometry plot with quantification of IFNy and TNFa co- expression by 4T1 primary tumour CD8+T cells following 210 treatment. n=6. (H) Expression of granzyme B (n=ll-12), (I) CD107a (n=6-7) by BRPKp110 primary tumour CD8+T cells following 210 treatment. Q) BRPKp110 mean tumour growth over time following administration of vehicle control or 210 in combination with either anti CD8-depleting antibody or IgG isotype control (n=9), with quantification of the depletion of primary tumour CD8+T cells following antibody administration. Statistical significance calculated by (A-F, H-J) two-tailed impaired t test and (G) Mann-Whitney test.

[0168] **P < 0.01, *P < 0.05.

[0169] Figure 4. B. pseudocatenulatum 210 programmed dendri3c cells induce anti-tumour CD8+T cell immunity.

[0170] (A) Quantification of the infiltration of dendritic cells and (B) eDC subtype cells in the blood of BRPKp110-bearing animals (n=5). (C) Quantification of CCR7 expression on dendritic cells and (D) cDC1 cells in the blood of BRPKp110- bearing animals (n=6-7). (E) Quantification of the infiltration of cDC1 cells (n=8), and (F) the percentage of CCR7+expressing dendritic cell and cDC1 cells (n=6-7), and (G) CD80+CD86+mature dendritic cells (n=7-8) within BRPKp110 tumour-draining lymph nodes following 210 administrations. (H) Quantification of infiltration of dendritic cell subsets (n=7-8) and (I) CD80+CD86+mature dendritic cells (n=7-8) within BRPKp110 primary tumours following 210 administrations. (J) Experimental outline of 210- treated dendritic cell adoptive transfers to naive BRPKp110 tumour bearing animals, with quantification of tumour growth (n=8- 10) and (K) intratumoural CD8+T cell IFNy expression (n=7-9) following the indicated treatments. Statistical significance was calculated by two- tailed unpaired t test, with Welch’s correction applied to (A) and (B). *P < 0.05. Figure 5. Administra3on of B. pseudocatenulatum 210 does not significantly alter serum metabolite levels in untargeted analyses. (A) 2D score plots of serum metabolite profiles of BRPKp110 (n=5-6) and PyMT-BO1 (n=4-5) tumour- bearing animals; shaded circles indicate 95% confidence intervals. (B) Plots showing significance scores for the differential expression of 500 serum metabolites following 210 therapeutic intervention, n=5. (C) Heatmaps showing the top 12 differentially expressed serum metabolites between vehicle control and B. pseudocatenulatum 210 treated animals. Statistical differences were assessed by two- tailed unpaired t test with an FDR applied at P < 0.05

[0171] Figure 6. B. pseudocatenulatum 210 anti-tumour immunity is induced by cell surface capsular exopolysaccharide. (A) BRPKp110 primary tumour growth following administration with live 210 cells or peracetic acid-killed 210 cells. n=8-9. (B) BRPKp110 primary tumour growth following administration of live 210 or isolated 210 exopolysaccharide (210 EPS) solution (80μg / dose). n=20- 23, N=3. (C) Quantification showing BRPKp110 primary tumour CD8+T cell IFNy and TNFa cytokine expression (n=11-13), effector-memory polarisation (n=ll-13), and (D) macrophage MHCII+ / CD206+infiltration (n=6-7) following administration with 210 EPS. (E) Quantification of IFNy in the serum of BRPKp110-bearing animals, measured by MSD multiplex cytokine analysis. n=5-6. (F) Quantification of the infiltration of dendritic cells and eDC subtype cells in the blood of BRPKp110-bearing animals. n=5. (G) Experimental outline of B. pseudocatenulatum 210 EPS and B. longum B71 EPS-treated dendritic cell adoptive transfers to naive BRPKp110 tumour bearing animals. Data shows quantification of tumour growth (n=18-19, N=2) and intratumoural CD8+T cell IFNy expression (n=12-16) following the indicated dendritic cell treatments. (A-H) Statistical significance was calculated by two-tailed impaired t test. ***P < 0.001, **P < 0.01, *P < 0.05. Figure 7. B. pseudocatenulatum exopolysaccharide function is strain specific and associated with genomic homology of encoded eps clusters. (A) Proposed architecture of putative B. pseudocatenulatum 210 EPS enzymatic clusters. Gene functions were predicted based on combined results from blastp searches (blastp e-value = le-50) against previously described bifidobacterial eps cluster sequences Ferrario et at, 2016(39) and Wang, et at, 2019(40) the NCBI reference protein database (refseq_protein, default seangs), and results generated with dbCAN3 server (hmmer E-value < le-15, coverage > 0.35). (B) Homology maps between the putative B. pseudocatenulatum 210 eps clusters and those of other human- associated strains used in this study, as well as selected reference strains (B. pseudocatenulatum DSM20438 and B. breve IwOl). Gene functions were predicted as above. (C) BRPKp110 tumour growth in animals treated with B. pseudocatenulatum 210 and DSM 20438 (n=8-9), or (D) B. pseudocatenulatum LH14 (n=6-7), or (E) B. lonpum B71 (n=8). (F) In vitro MHCII expression and CD80+CD86+maturation of BMDCs stimulated for 24 h with purified EPS from the indicated strains. Cellular expression was assessed by flow cytometry on CDllc+cells.

[0172] References

[0173] 1. R. Bai et al, Mechanisms of Cancer Resistance to Immunotherapy,Front Oncol 10, 1290 (2020).

[0174] 2. M. R. Fernandes, P. Aggarwal, R. G. F. Costa, A. M. Cole, G. Trinchieri, Targeting the gut microbiota for cancer therapy. Nat. Rev. Cancer 22, 703-722 (2022).

[0175] 3. M. Vetizou et al, Anticancer immunotherapy by CTLA-4 blockade relies on the gut microbiota. Science 350, 1079-1084 (2015). 4. A. Sivan et aL, Commensal Bifidobacterium promotes antitumor immunity and facilitates anti-PD-Ll efficacy. Science 350, 1084-1089 (2015).

[0176] 5. N. lida et aL, Commensal bacteria control cancer response to therapy by modulating the tumor microenvironment. Science 342, 967-970 (2013).

[0177] 6. E. N. Baruch et aL, Fecal microbiota transplant promotes response in immunotherapy-refractory melanoma patients. Science 371, 602-609 (2021).

[0178] 7. D. Davar et aL, Fecal microbiota transplant overcomes resistance to anti- PD-1 therapy in melanoma patients. Science 371, 595-602 (2021).

[0179] 8. N. Dizman et aL, Nivolumab plus ipilimumab with or without live bacterial supplementation in metastatic renal cell carcinoma: a randomized phase 1 trial. Not. Med. 28, 704-712 (2022).

[0180] 9. Q. Zhang et aL, Lactobacillus plantarum-derived indole-3-lactic acid ameliorates colorectal tumorigenesis via epigenetic regulation of CDS. Cell Metab 35, 943-960.e949 (2023).

[0181] 10. L. F. Mager et aL, Microbiome-derived inosine modulates response to checkpoint inhibitor immunotherapy. Science 369, 1481-1489 (2020).

[0182] 11. K. A. Lee et aL, Cross-cohort gut microbiome associations with immune checkpoint inhibitor response in advanced melanoma. Nat. Med. 28, 535-544 (2022).

[0183] 12. R. C. Simpson et al, Diet-driven microbial ecology underpins associations between cancer immunotherapy outcomes and the gut microbiome. Nat. Med 28, 2344-2352 (2022).

[0184] 13. J. A. McCulloch et aL, Intestinal microbiota signatures of clinical response and immune-related adverse events in melanoma patients treated with anti-PD-1. Nat. Med 28, 545-556 (2022).

[0185] 14. L. Derosa et aL, Intestinal Akkermansia muciniphila predicts clinical response to PD-1 blockade in patients with advanced non-small-cell lung cancer. Nat. Med 28, 315-324 (2022). 15. S. H. Lee et aL, Bifidobacterium bifidum strains synergize with immune checkpoint inhibitors to reduce tumour burden in mice. Nat Microbiol 6, 277-288 (2021). 16. G. Longhi, D. van Sinderen, M. Ventura, F. Turroni, Microbiota and

[0186] Cancer The Emerging Beneficial Role of Bifidobacteria in Cancer Immunotherapy. Front Microbiol XX, 575072 (2020).

[0187] 17. M. J. Bender et aL, Dietary tryptophan metabolite released by intratumoral

[0188] Lactobacillus reuteri facilitates immune checkpoint inhibitor treatment. Cell 186, 1846-1862.el 826 (2023).

[0189] 18. Q. Li et aL, Oral administration of Bifidobacterium breve promotes antitumor efficacy via dendritic cells-derived interleukin 12. Oncobnmunology 10, 1868122 (2021).

[0190] 19. Y. Yoon, G. Kim, B. N. Jeon, S. Fang, H. Park, Strain-Specific Enhances the Efficacy of Cancer Therapeutics in Tumor-Bearing Mice. Cancers (Basel) 13, (2021).

[0191] 20. M. Luu et aL, Microbial short-chain faoy acids modulate CDS + T cell responses and improve adoptive immunotherapy for cancer. Nat Commun 12, 4077 (2021).

[0192] 21. M. Uribe-Herranz et aL, Gut microbiota modulate dendritic cell antigen presentation and radiotherapy-induced antitumor immune response. J. Clin. Invest. 130, 466-479 (2020).

[0193] 22. G. Sharma et aL, A dietary commensal microbe enhances antitumor immunity by activating tumor macrophages to sequester iron. Nat. ImmunoL 25, 790-801 (2024).

[0194] 23. A. Hickey et at, Exopolysaccharide Blocks Dendritic Cell Maturation and Activation of CD4. F font Microbiol 12, 653587 (2021).

[0195] 24. N. Castro-Bravo, J. M. Wells, A. Margolles, P. Ruas-Madiedo, Interactions of Surface Exopolysaccharides From Bifidobacterium and Lactobacillus Within the Intestinal Environment. Front Microbiol 9, 2426 (2018). 25. J. D. Baker etal, CoronaHiT: high-throughput sequencing of SARS-CoV-2 genomes. Genome Medicine 13, (2021).

[0196] 26. M. M. Tomayko, C. P. Reynolds, Determination of subcutaneous tumor size in athymic (nude) mice. Cancer Chemother Pharmacol 2A, 148-154 (1989).

[0197] 27. K. Moor et al, Peracetic Acid Treatment Generates Potent Inactivated Oral Vaccines from a Broad Range of Culturable Bacterial Species. Pront Immunol 1, 34 (2016).

[0198] 28. J. Chong, D. S. Wishart, J. Xia, Using MetaboAnalyst 4.0 for Comprehensive and Integrative Metabolomics Data Analysis. Curr Protoc BioinformaGcs 68, e86 (2019).

[0199] 29. Z. Pang et al, Using MetaboAnalyst 5.0 for LC-HRMS spectra processing, multi-omics integration and co variate adjustment of global metabolomics data. Nat Protoc 11, 1735-1761 (2022).

[0200] 30. J. Junick, M. Blaut, Quantification of human fecal bifidobacterium species by use of quantitative real-time PCR analysis targeting the groEL gene. AppL

[0201] Environ. Microbiol 78, 2613-2622 (2012).

[0202] 31. P. Ruas-Madiedo, Detection, Isolation, and Purification of Bifidobacterial Exopolysaccharides. Methods Mol Biol 2278, 101-115 (2021).

[0203] 32. M. A. E. Lawson et al, Breast milk-derived human milk oligosaccharides promote Bifidobacterium interactions within a single ecosystem. ISME J 14, 635- 648 (2020).

[0204] 33. M. Kujawska et al, Succession of Bifidobacterium longum Strains in Response to a Changing Early Life Nutritional Environment Reveals Dietary Substrate Adaptations. iScience 23, 101368 (2020).

[0205] 34. S. Chen, Y. Zhou, Y. Chen, J. Gu, fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinforma Ges 34, i884-i890 (2018).

[0206] 35. R. R. Wick, L. M. Judd, C. L. Gorrie, K. E. Holt, Unicycler: Resolving bacterial genome assemblies from short and long sequencing reads. PLoS Comput Biol 13, 61005595 (2017). 36. D. H. Parks, M. Imelfort, C. T. Skennerton, P. Hugenholtz, G. W. Tyson, CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res. 25, 1043-1055 (2015).

[0207] 37. T. Seemann, Prokka: rapid prokaryotic genome annotation. BioinformaGcs

[0208] 30, 2068-2069 (2014).

[0209] 38. C. Camacho et al, BLAST+: architecture and applications. BMC BioinformaGcs 10, 421 (2009).

[0210] 39. C. Ferrario et al, Modulation of the eps-ome transcription of bifidobacteria through simulation of human intestinal environment. FEMS Microbiol EcoL 92, fiw056 (2016).

[0211] 40. L. Wang et al, Exopolysaccharide, Isolated From a Novel Strain

[0212] Bifidobacterium breve IwOl Possess an Anticancer Effect on Head and Neck

[0213] Cancer. Front Microbiol XB, 1044 (2019).

[0214] 41. C. L. M. Gilchrist, Y. H. Chooi, clinker & clustermap. js: automatic generation of gene cluster comparison figures. BioinformaGcs 37 , 2473-2475 (2021).

[0215] 42. C. Buchta Roseau et al, Preexisting Commensal Dysbiosis Is a Host- Intrinsic Regulator of Tissue Inflammation and Tumor Cell Dissemination in Hormone Receptor-Positive Breast Cancer. Cancer Res 79, 3662-3675 (2019).

[0216] 43. M. J. Allegrezza et al, Trametinib Drives T-cell-Dependent Control of KRAS-Mutated Tumors by Inhibiting Pathological Myelopoiesis. Cancer Res. 76, 6253-6265 (2016).

[0217] 44. X. Su et al, Antagonizing Integrin 03 Increases Immunosuppression in Cancer. Cancer Res. 76, 3484-3495 (2016).

[0218] 45. J. Liu, X. Geng, J. Hou, G. Wu, New insights into M1 / M2 macrophages: key modulators in cancer progression. Cancer Cell Int 21, 389 (2021).

[0219] 46. E. W. Roberts etal, Critical Role for CD103(+) / CD141(+) Dendritic Cells Bearing CCR7 for Tumor Antigen Trafficking and Priming of T Cell Immunity in Melanoma. Cancer Cell SB, 324-336 (2016). 47. M. L. Broz et al, Dissecting the tumor myeloid compartment reveals rare activating antigen-presenting cells critical for T cell immunity. Cancer Cell 26, 638- 652 (2014).

[0220] 48. Q. Yang et al, N Review of Gut Microbiota-Derived Metabolites in Tumor Progression and Cancer Therapy. Adv Sei (Weinh) 10, e2207366 (2023).

[0221] 49. C. Hidalgo-Cantabrana et al, Exopolysaccharide-producing Bifidobacterium animalis subsp. lactis strains and their polymers elicit different responses on immune cells from blood and gut associated lymphoid tissue. Anaerobe 26, 24-30 (2014).

[0222] 50. M. M. Waoenberg et al, Cancer immunotherapy via synergistic coactivation of myeloid receptors CD40 and Dectin-1. Sci bnvHtnolfc, eadj5097 (2023).

[0223] 51. V. Matson et al, The commensal microbiome is associated with anti-PD-1 efficacy in metastatic melanoma patients. Science 359, 104-108 (2018).

Claims

Claims1. A composition for the in vivo stimulation of cancer immunity, said composition including at least one exopolysaccharide (EPS) characterised in that the at least one EPS is a trimeric sugar comprised predominantly of 1-3 and 1-4 linked glucose and galactose residues.

2. A composition according to claim 1 wherein the EPS is included with a pharmaceutical carrier to provide an anti-tumour therapeutic compound or composition.

3. A composition according to claim 1 or 2 wherein the EPS is isolated or cleaved from the bacterium of genus Bifidobacterium.

4. A composition according to claim 3 wherein the EPS trimeric sugars have a molecular weight of at or below 500 da.

5. A composition according to claim 4 wherein the EPS trimeric sugars have a molecular weight at or below 475 da.

6. A composition according to claim 5 wherein the EPS is a capsular EPS.

7. A composition according to claim 6 wherein the EPS is derived or isolated from B.p ■ seudocatenulatum.

8. A composition according to claim 7 wherein the EPS is isolated or derived from Bifidobacterium pseudocatenulatum 210 and / or B. pseudocatenulatum DSM20438.

9. A composition according to any preceding claim wherein the composition induces CD8+T cell-dependent anti-tumour immunity through the activation dendritic cells.

10. A composition according to any preceding claim wherein the composition is active and / or provides anti-tumour activity in respect of breast cancer.

11. A pharmaceutical composition comprising an immunotherapy agent, said agent including a microbial capsular EPS.

12. A composition according to claim 11 wherein the EPS is a Bifidobacterium EPS.

13. A composition according to claim 12 wherein the composition is a chemotherapy adjuvant.

14. A composition according to claim 12 wherein has 80, 85, 90, 95 or 99 % identity to a sequence set out in SEQ ID NO: 1 to 5.

15. A composition according to claim 12 wherein the EPS is obtained from Bifidobacterium pseudocatenulatum 210 and / or B. pseudocatenulatum DSM20438.

Images